Download INTERNAL MEDICINE JOURNAL

previous
Total-pages:
Paper-type:
issue 3
Editorial
253
March 2016
332
Associations of demographic and behavioural
factors with glycaemic control in young adults
with type 1 diabetes mellitus
J. K. Osan, J. D. Punch, M. Watson, Y. X. Chan, P. Barrie,
P. G. Fegan and B. B. Yeap
339
Inaccurate risk perceptions contribute to
treatment gaps in secondary prevention of
cardiovascular disease
J. Thakkar, E. L. Heeley, J. Chalmers and C. K. Chow
The threat of litigation as a possible barrier
to innovation
J. Coates
Review
255
Cardiac resynchronisation therapy in 2015:
keeping up with the pace
A. Voskoboinik, A. D. McGavigan and J. A. Mariani
Brief Communications
347
Clinical Perspectives
266
Characterisation and therapeutic manipulation
of the gut microbiome in inflammatory bowel
disease
J. Schulberg and P. De Cruz
Original Articles
273
281
Regional challenges: evaluation of a hepatitis
outreach programme using transient
elastography (FibroScan) in Victoria
M. C. Thurnheer, T. R. Schulz, T. Nguyen, J. MacLachlan
and J. Sasadeusz
Baseline abnormal liver function tests are more
important than age in the development of
isoniazid-induced hepatoxicity for patients
receiving preventive therapy for latent
tuberculosis infection
E. L. Gray and H. F. Goldberg
Home medicines reviews in Australian war
veterans taking warfarin do not influence
international normalised ratio control
L. R. E. Bereznicki, E. C. van Tienen and A. Stafford
295
Do community hospice programmes reduce
hospitalisation rate in patients with
advanced chronic obstructive pulmonary
disease?
S. P. M. Iupati and B. R. Ensor
301
307
315
325
IMJ_46_3_cover.indd 1
352
Substantial variation in post-engraftment
infection prophylaxis and revaccination practice
in autologous stem cell transplant patients
H. Y. Lim and A. Grigg
Safety and efficacy of a novel short occlusive
regimen of imiquimod for selected non-melanotic
skin lesions in renal transplant patients
G. Das, B. Tan and K. Nicholls
356
Acute adrenal insufficiency: an aide-memoire
of the critical importance of its recognition and
prevention
A. Gargya, E. Chua, J. Hetherington, K. Sommer and
M. Cooper
360
Severe hypocalcaemia and hypophosphataemia
following intravenous iron and denosumab:
a novel drug interaction
B. Smyth and S. Ong
364
Cardiovascular risk reduction in hypertension:
angiotensin-converting enzyme inhibitors,
angiotensin receptor blockers. Where are
we up to?
A. Sindone, J. Erlich, C. Lee, H. Newman, M. Suranyi
and S. D. Roger
Letters to the Editor
Clinical-scientific notes
373
Screening for coeliac disease in elderly inpatients
with minimal trauma fracture is not indicated
D. I. Clayton-Chubb, M. Wilson, E. Seal and P. W. Lange
Alcohol use disorder hospitalisations over the
last two decades: a population-based cohort
study
W. Liang and T. Chikritzhs
374
Association between hepatitis B virus infection
and risk of multiple myeloma: a systematic
review and meta-analysis
Y. Li, O. Bai, C. Liu, Z. Du, X. Wang, G. Wang
and W. Li
General correspondence
376
Estimation of clinical and economic effects of
prophylaxis against venous thromboembolism
in medical patients, including the effect of
targeting patients at high-risk
J. A. Millar and A. L. K. Gee
Liquid biopsies: advancing cancer research
through drops of blood
A. Khattak
377
Conflict of interest: real and perceived – a more
mature consideration is needed
J. M. Andrews, S. P. Costello, A. K. Agarwal, P. Bampton,
L. Beswick, S. Connor, S. Ghaly, S. O’Connor,
A. Pudipeddi, A. Sechi, M. Sparrow and A. J. Walsh
A tertiary hospital audit of opioids and
sedatives administered in the last 24 h of life
C. Douglas, M. Clarke, S. Alexander and M. Khatun
Total-pages:136
INTERNAL
MEDICINE
JOURNAL
Spine-width: 4.4mm
Paper-type: UPM Classic 64
Volume
46
Issue
3
March
2016
ISSN 1444-0903
Personal Viewpoint
380
Traumatic superior gluteal artery pseudoaneurysm
following a bone marrow biopsy
I. R. Caldwell, B. T. Buckley, R. Rajagopal,
R. Doocey and L. Pemberton
INTER NAL MEDICINE JOUR NAL
288
Current
Volume 46, Issue 3, March 2016, Pages 245–380
VOLUME 46
Spine-width:
Cardiac resynchronisation therapy update
The gut microbiome in inflammatory bowel disease
Community hospice programmes and COPD
Type 1 diabetes control in young adults
Alcohol use disorder hospitalisations
Opioids and sedatives in the last 24 hours of life
Acute adrenal insufficiency: recognition and prevention
Errata
08/03/16 10:38 PM
bs_bs_banner
Honorary Advisory Board
Editor-in-Chief
Jeff Szer, Melbourne
Continuing Education
Deputy Editor-in-Chief
Deputy Editor-in-Chief
Zoltan Endre, Sydney
Paul Bridgman, Christchurch
Subspecialty Editors
Cardiology (General)
Paul Bridgman, Christchurch
Infectious Diseases
David Gordon, Adelaide
Cardiology (Arrhythmias)
Andrew McGavigan, Adelaide
Intensive Care
Michael O’Leary, Sydney
Clinical Genetics
Les Sheffield, Melbourne
Internal Medicine
Ian Scott, Brisbane
Clinical Pharmacology
Jenny Martin, Newcastle
Yvonne Bonomo, Melbourne
(Addiction Medicine)
Nephrology
Zoltan Endre, Sydney
Continuing Education
(Clinical Perspectives)
Christopher Pokorny, Sydney
Emergency Medicine
Paul Middleton, Sydney
Endocrinology
Morton Burt, Adelaide
Anthony Russell, Brisbane
Ethics
Paul Komesaroff, Melbourne
Gastroenterology
David M. Russell, Melbourne
Geriatric Medicine
Leon Flicker, Perth
Haematology (General)
Peter Browett, Auckland
Peter Doherty, Melbourne
Kar Neng Lai, Hong Kong
Richard Larkins, Melbourne
Sir Gustav Nossal, Melbourne
Lawrie W. Powell, Brisbane
Nicholas Saunders, Newcastle
John Shine, Sydney
Chorh Chuan Tan, Singapore
Sir David Weatherall, Oxford
Judith Whitworth, Canberra
Editorial Ombudsman
Graham Macdonald, Sydney
Manager
Virginia Savickis, Sydney
Neurology
David Blacker, Perth
Nuclear Medicine
Frederick A. Khafagi, Brisbane
Occupational and
Environmental Medicine; Health
Economics; Editorials Editor
Des Gorman, Auckland
Oncology
Damien Thomson, Brisbane
Palliative Medicine
Janet Hardy, Brisbane
Editorial Assistant
Aparna Avasarala, Sydney
Previous Editors-in-Chief
Internal Medicine Journal
Edward Byrne (1999–2004)
The Australian and New Zealand
Journal of Medicine
Graham Macdonald (1989–1999)
Michael O'Rourke (1981–1989)
Akos Z. Gyory (1975–1981)
Charles Kerr (1970–1975)
The Australasian Annals of Medicine
Ronald Winton (1957–1970)
Mervyn Archdall (1952–1956)
Public Health Medicine
Mark Ferson, Sydney
Respiratory Medicine
Matthew Naughton, Melbourne
Haemostasis/Thrombosis
Claire McLintock, Auckland
Rheumatology
Peter Youssef, Sydney
Immunology and Allergy
Marianne Empson, Auckland
Sexual Health Medicine
Darren Russell, Cairns
IMJ.JEBJan16
IMJ_13045_JEB.indd fm_iii
27/02/16 9:43 PM
Save Time and Let the
Research Come to You
Sign up for new content alerts for
all of your favorite journals:
9 Be the first to read Early View articles
9 Get notified of Accepted Articles when they
appear online
9 Receive table of contents details each time a
new issue is published
9 Never miss another issue!
Follow these 3 easy
y steps
p to register
g
for alerts online:
Select “Get New Content Alerts” from Journal Tools
on the top left menu on any journal page or visit the
Publications page to view all titles.
Submit your preferences and you are done.
You will now receive an email when a new issue of
the journal publishes.
wileyonlinelibrary.com
12 - 4 6 0 6 0
1
2
3
Log into Wiley Online Library. If you are not already a
registered user, you can create your profile for free.
bs_bs_banner
aims and scope
The Internal Medicine Journal, formerly known as the
Australian and New Zealand Journal of Medicine, is the
official journal of the Adult Medicine Division of
The Royal Australasian College of Physicians (RACP).
Its purpose is to publish high-quality internationally
competitive peer-reviewed original medical research, both
laboratory and clinical, relating to the study and research
of human disease. Papers will be considered from all areas
of medical practice and science. The Journal also has a
major role in continuing medical education and publishes
review articles relevant to physician education. Except
where otherwise stated, articles are peer reviewed.
abstracting and indexing
This journal is indexed by Abstracts on Hygiene and
Communicable Diseases, AgBiotech News and Information,
AIDS Abstracts, Australian Medical Index, BIOBASE,
Biological Abstracts (BIOSIS), Biomedical Reference
(EBSCO), Cambridge Scientific Abstracts, Chemical
Abstracts Service, Current Contents/Clinical Medicine
(an ISI product), Derwent Biotechnology Abstracts,
EMBASE/Excerpta Medica, Environmental Sciences and
Pollution Management, Health and Safety Science Abstracts
(Online version), Helminthological Abstracts, InPharma
Weekly, International Pharmaceutical Abstracts (IPA),
Journals @ Ovid, MEDLINE, Nutrition Abstracts and
Reviews, Pharmacoeoconomics and Outcomes News,
Reactions Weekly, Science Citation Index, SCOPUS, Tropical
Diseases Bulletin, Vitis-Viticulture and Oenology Abstracts
(Online Edition), World Agricultural Economics and Rural
Sociology Abstracts, and CINAHL.
address for editorial correspondence
Editor-in-Chief, Internal Medicine Journal, The Royal
Australasian College of Physicians, 145 Macquarie Street,
Sydney, NSW 2000, Australia (tel: +61 2 9256 5431; fax:
+61 2 9252 3310). For enquiries regarding ScholarOne
Manuscripts (formerly known as ManuscriptCentral)
submissions please email [email protected]
(e.g. IMJ-0000-2014).
General enquiries should be directed to Virginia Savickis,
the Editorial Office, Internal Medicine Journal, using imj@
racp.edu.au
Comments on published papers are welcomed. Authors
are offered right of reply (no more than 500 words) at the
discretion of the Editor and discussion will not be entered
into. Given the current pressures on editorial space,
however, invited comments are restricted to one reply.
disclaimer
The Publisher, RACP and Editors cannot be held
responsible for errors or any consequences arising from the
use of information contained in this journal; the views and
opinions expressed do not necessarily reflect those of the
Publisher, RACP and Editors, neither does the publication
of advertisements constitute any endorsement by the
Publisher, RACP and Editors of the products advertised.
Copyright © 2016 Royal Australasian College of
Physicians.
For submission instructions, subscription and all other information visit
www.wileyonlinelibrary.com/journal/imj
This journal is available online at Wiley Online Library. Visit www.onlinelibrary.wiley.com to search the articles and
register for table of contents and email alerts.
Wiley’s Corporate Citizenship initiative seeks to address the environmental, social, economic, and ethical challenges
faced in our business and which are important to our diverse stakeholder groups. We have made a long-term commitment
to standardise and improve our efforts around the world to reduce our carbon footprint. Follow our progress at
www.wiley.com/go/citizenship
Access to this journal is available free online within institutions in the developing world through the HINARI initiative
with the WHO. For information, visit www.healthinternetwork.org
ISSN 1444-0903 (Print)
ISSN 1445-5994 (Online)
IMJ.JEBJan16
IMJ_13044_aim&scope.indd fm_iii
27/02/16 9:44 PM
hell is the pl ace where nothing connects — T.S. Eliot
reliable reference linking,
powered by CrossRef
Linking to and from this journal is powered by CrossRef.
CrossRef links millions of articles and chapters, from
thousands of publications, from hundreds of publishers.
So your reference linking works. Permanently. Take a
tour at www.crossref.org. And make sure all your
published work is CrossRef-enabled.
crossref
C R O S S R E F. O R G | T H E C I TAT I O N L I N K I N G B A C K B O N E
4 0 S A L E M S T R E E T, LY N N F I E L D M A 0 1 9 4 0 • 7 8 1 - 2 9 5 - 0 0 7 2
bs_bs_banner
March 2016, Volume 46, Issue 3
Editorial
253
The threat of litigation as a possible
barrier to innovation
Do community hospice programmes
reduce hospitalisation rate in patients
with advanced chronic obstructive
pulmonary disease?
J. Coates
S. P. M. Iupati and B. R. Ensor
Review
255
295
301
Cardiac resynchronisation therapy
in 2015: keeping up with the pace
A. Voskoboinik, A. D. McGavigan and
J. A. Mariani
W. Liang and T. Chikritzhs
307
Clinical Perspectives
266
Characterisation and therapeutic
manipulation of the gut microbiome
in inflammatory bowel disease
315
Original Articles
Regional challenges: evaluation of a
hepatitis outreach programme using
transient elastography (FibroScan) in
Victoria
Baseline abnormal liver function tests are
more important than age in the
development of isoniazid-induced
hepatoxicity for patients receiving
preventive therapy for latent tuberculosis
infection
325
Home medicines reviews in Australian war
veterans taking warfarin do not influence
international normalised ratio control
L. R. E. Bereznicki, E. C. van Tienen
and A. Stafford
IMJ_13046_toc1.indd fm_v
A tertiary hospital audit of opioids and
sedatives administered in the last 24 h
of life
C. Douglas, M. Clarke, S. Alexander and
M. Khatun
332
Associations of demographic and
behavioural factors with glycaemic
control in young adults with type 1
diabetes mellitus
J. K. Osan, J. D. Punch, M. Watson,
Y. X. Chan, P. Barrie, P. G. Fegan and
B. B. Yeap
E. L. Gray and H. F. Goldberg
288
Estimation of clinical and economic
effects of prophylaxis against venous
thromboembolism in medical patients,
including the effect of targeting patients
at high-risk
J. A. Millar and A. L. K. Gee
M. C. Thurnheer, T. R. Schulz, T. Nguyen,
J. MacLachlan and J. Sasadeusz
281
Association between hepatitis B virus
infection and risk of multiple myeloma:
a systematic review and meta-analysis
Y. Li, O. Bai, C. Liu, Z. Du, X. Wang,
G. Wang and W. Li
J. Schulberg and P. De Cruz
273
Alcohol use disorder hospitalisations over
the last two decades: a population-based
cohort study
339
Inaccurate risk perceptions contribute to
treatment gaps in secondary prevention
of cardiovascular disease
J. Thakkar, E. L. Heeley, J. Chalmers
and C. K. Chow
08/03/16 10:38 PM
bs_bs_banner
March 2016, Volume 46, Issue 3
Brief Communications
347
Substantial variation in post-engraftment
infection prophylaxis and revaccination
practice in autologous stem cell transplant
patients
Letters to the Editor
Clinical-scientific notes
373
H. Y. Lim and A. Grigg
352
Safety and efficacy of a novel short
occlusive regimen of imiquimod for
selected non-melanotic skin lesions in
renal transplant patients
D. I. Clayton-Chubb, M. Wilson, E. Seal and
P. W. Lange
374
G. Das, B. Tan and K. Nicholls
356
Acute adrenal insufficiency: an
aide-memoire of the critical importance
of its recognition and prevention
A. Gargya, E. Chua, J. Hetherington,
K. Sommer and M. Cooper
360
Severe hypocalcaemia and
hypophosphataemia following
intravenous iron and denosumab:
a novel drug interaction
Screening for coeliac disease in elderly
inpatients with minimal trauma fracture is
not indicated
Traumatic superior gluteal artery
pseudoaneurysm following a bone
marrow biopsy
I. R. Caldwell, B. T. Buckley, R. Rajagopal,
R. Doocey and L. Pemberton
General correspondence
376
Liquid biopsies: advancing cancer research
through drops of blood
A. Khattak
377
Conflict of interest: real and perceived – a
more mature consideration is needed
J. M. Andrews, S. P. Costello, A. K. Agarwal,
P. Bampton, L. Beswick, S. Connor, S. Ghaly,
S. O’Connor, A. Pudipeddi, A. Sechi,
M. Sparrow and A. J. Walsh
B. Smyth and S. Ong
Personal Viewpoint
380
364
Errata
Cardiovascular risk reduction in
hypertension: angiotensin-converting
enzyme inhibitors, angiotensin receptor
blockers. Where are we up to?
A. Sindone, J. Erlich, C. Lee, H. Newman,
M. Suranyi and S. D. Roger
IMJ_13047_toc2.indd fm_v
04/03/16 7:06 PM
WE MAKE YOUR RESEARCH EASY.
NOW WE MAKE JOB HUNTING EASY.
Let your partners in research energize your career.
Drawing on our expertise and relationships in the healthcare industry, Wiley-Blackwell invites you
to join Wiley Healthcare Jobs, the definitive job site for healthcare professionals.
đƫ premium jobs from the most respected names in healthcare
đƫ hundreds of healthcare-industry recruiters and employers
đƫ job alerts that match your criteria
đƫ expert career advice and candidate resources
Register and upload your resume/CV now to begin your job search!
Part of
wileyhealthcarejobs.com
Internal Medicine Journal (2016)
E D I TO R ’ S N OT E
The following editorial by Dr Jonathan Coates, a lawyer who specialises in medicolegal law, is the last in the series on the barriers
to innovation in healthcare. Although the risk of litigation is cited as a stimulant for defensive medical practice and as an inhibitor
of innovation in high-tort-risk countries such as the US, England and Australia, the underpinning data in this regard are not
robust. Dr Coates’ consideration of the situation in New Zealand provides an interesting insight in this context, given that
New Zealand has a no fault compensation system, which makes it very difficult for anyone to sue a doctor for a treatment injury. He
quite appropriately identifies alternative drivers of defensive medical behaviour such as reputational risk and fear of complaint.
As an aside, comparisons of treatment injury rates between New Zealand and tort-rich countries are difficult given that the
monopoly corporation that exists in New Zealand enables a comprehensive and inclusive data set. It is nevertheless of some concern
that within a decade, that treatment injury will be the single highest category of costs to New Zealand’s accident compensation
scheme.
doi:10.1111/imj.13021
D. Gorman
IMJ Editor (Editorials)
E D I TO R I A L
The threat of litigation as a possible barrier to innovation
The suggestion that the threat of litigation against health
professionals may present a barrier to innovation in
health is not new. Headlines in the mainstream media
such as ‘The fear of being sued is ruining modern medicine’1 have been supported, at least in part, by research
that has suggested that defensive medical practice is
common among doctors.2
But to what extent are such claims supported by the
experience in New Zealand, where there are extremely
limited rights to sue for medical malpractice? Conversely, can it be claimed that innovation in health
flourishes in New Zealand because of the inability
to sue?
The threat of litigation and defensive
medicine
Defensive medicine occurs where a health professional
practises in a manner that is different from their usual
practice, or different from accepted good practice, in
order to reduce or prevent complaints, criticism or malpractice liability.2 In one UK study, 78% of doctors
reported practising defensive medicine.2 Defensive medicine can often involve over-prescription of treatment or
diagnostic procedures with obvious impacts on scare
medical resources.
© 2016 Royal Australasian College of Physicians
Defensive medicine is claimed to be an ‘innovation
inhibitor’ with doctors unwilling to try new techniques
and methods because of the fear of being sued if things
go wrong.3 There have been moves in the United Kingdom to legislate to encourage innovation and advancement in medicine without the threat of being sued –
most notably Lord Saatchi’s unsuccessful Medical Innovation Bill and its more recent successor the Access to
Medical Treatment (Innovation) Bill 2015. Regardless of
how the more recent Bill progresses, it seems clear that
there is considerable support for moves to address what
many regard as the impediment of defensive medicine.
Others note that the underlying cause, being the fear of
litigation, can only be removed by removing any possibility of litigation.3
That brings us to consider the New Zealand model;
where the possibility of litigation has been removed.
New Zealand’s model
The right to sue in New Zealand for medical malpractice
was given up in exchange for a comprehensive no-fault
compensation scheme run by the Accident Compensation Corporation (ACC).4 The ACC was established in
1974. The effect is to provide compensation and other
support through the ACC to people who suffer personal
253
Editorial
injury, with the social contract being that victims of personal injury give up the right to sue.
While there have been several amendments to the
scheme, the key principles remain. This means that, in
its current form, patients who suffer a ‘treatment injury’
look to the ACC for compensation. It is only in
extremely rare situations that doctors, other health professionals and healthcare provider organisations can
be sued.
Perhaps in part because of the lack of clinical negligence litigation, New Zealand’s complaints resolution
process and competence assurance regime are well
entrenched. The Health and Disability Commissioner
receives and investigates complaints about the provision
of health services and has been an active and high profile
office for two decades.5 New Zealand’s Health Practitioners Competence Assurance Act 2003 established a
clear framework for the regulators to investigate and act
on concerns and notifications about health practitioners’
competence, conduct and health.
Despite the no-fault compensation model, New
Zealand has not been immune to claims of defensive
medicine. Defensive medicine in New Zealand
has tended to be linked to doctors changing their usual
practice to avoid a complaint, rather than to avoid litigation.6 Defensive medicine has been described as an
undesirable outcome of the complaints process in
New Zealand.7
It has been argued in New Zealand that defensive
medicine is driven less by externalised factors and more
by internalised mechanisms, such as the notion of
shame.7 There is some intuitive attraction in that; a
health professional’s reputation is everything – and likely
more important to the individual than whether an
indemnity insurer needs to pay out significant sums to
settle litigation. The claims of defensive medicine in
References
1 www.telegraph.co.uk [cited 2012 Dec 9].
2 Ortashi O, Virdee J, Hassan R,
Mutrynowski T, Abu-Zidan F. The
practice of defensive medicine among
hospital doctors in the United Kingdom.
BMC Med Ethics 2013; 14: 42.
3 Mills & Reeve LLP. What’s the problem
with defensive medicine? 2014 [cited
2015 Nov 2]. Available from URL: http://
254
New Zealand have provoked an equally strong response,
including questioning the validity of some of the
research.8
Conclusion
There are a lot of positives about the New Zealand nofault compensation model and the regulatory framework
for health services. However, it cannot clearly be concluded that the lack of malpractice litigation, and therefore the removal of the ‘fear of being sued’, has created
an innovation nirvana where new models of healthcare
are flourishing. Some might argue that the ‘fear of being
sued’ in New Zealand has been replaced by the ‘fear of
being complained about’. However, while there may
have been some historical validity to such claims, as
health professionals get more confidence in the integrity
of the complaints process and regulatory system, it
would seem harder to blame such fears for stifling
innovation.
The practical steps needed to encourage and incentivise innovation will be multifaceted. Certainly,
New Zealand does not have all the answers. However,
what can be learned from New Zealand is that the lack
of malpractice litigation will not eliminate the suggestion
of defensive medicine, and that it is not at all clear that
the threat of litigation itself is an obvious and major
impediment to innovation in health. It would appear to
be much more complicated than that.
Received 24 November 2015; accepted 16 December 2015.
doi:10.1111/imj.12994
J. Coates
NZ’s Health Sector Lawyers, Claro Law, Christchurch, New Zealand
www.lexology.com/library/detail.aspx?
g=df0d8e74-289f-43dd-b121975f51d8dcce
4 The current legislation is the Accident
Compensation Act 2001.
5 Health & Disability Commissioner. [cited
2015 Nov 24]. Available from URL:
www.hdc.org.nz
6 Cunningham W, Wilson H. Complaints,
shame and defensive medicine. Postgrad
Med J 2011; 87: 837–40.
7 Cunningham W. Defensive Practice and
the Impact of Complaints Processes on
Providing Quality Patient Care. 11th
Annual Medical Law Conference; 2010
Apr; New Zealand.
8 Paterson R. The bogeyman of defensive
medicine. NZ Doctor 2006.
© 2016 Royal Australasian College of Physicians
REVIEW
Cardiac resynchronisation therapy in 2015: keeping up with
the pace
A. Voskoboinik,1,2 A. D. McGavigan3 and J. A. Mariani2,4
1
Department of Cardiology, Western Hospital, 2Department of Cardiology, Alfred Hospital, 4Cardiac Investigation Unit, St Vincent' Hospital, Melbourne,
Victoria, and 3Department of Cardiovascular Medicine, Adelaide, South Australia, Australia
Key words
cardiac resynchronisation therapy, biventricular
pacing, heart failure, dyssynchrony.
Correspondence
Alex Voskoboinik, Department of Cardiology,
Western Hospital, Melbourne, Vic. 3011,
Australia.
Email: [email protected];
[email protected]
Received 17 February 2015; accepted 24
March 2015.
Abstract
Despite improved understanding of the pathophysiology of heart failure (HF) and availability of better medical therapies, HF continues to grow as a cause of morbidity and mortality in
Australia and worldwide. Over the past decade, cardiac resynchronisation therapy (CRT), or
biventricular pacing, has been embraced as a powerful weapon against this growing
epidemic. However, much has changed in our understanding of dyssynchrony in HF, and
this has led to a change in guidelines to ensure more appropriate selection of CRT candidates
to improve the ‘non‐response’ rate. More data have also emerged about the use of CRT in
atrial fibrillation and in pacemaker‐dependent patients. There has also been a growing focus
on multimodality imaging to guide patient selection and lead positioning. Exciting new lead
technologies are also emerging, with the potential to improve CRT outcomes further.
doi:10.1111/imj.12774
Introduction
Dyssynchrony in heart failure (HF)
As HF progresses, pathological remodelling of the heart
results in cellular and interstitial changes that also
adversely affect the conduction system. Electrical
dyssynchrony causes mechanical dyssynchrony, which
refers to differences in the timing of contraction between
different myocardial regions. This occurs at multiple
levels in the failing heart, particularly atrioventricular
(AV), interventricular and intraventricular. Intraventricular dyssynchrony is often manifested by QRS prolongation and occurs in approximately one third of
cardiomyopathy patients. A left bundle branch block
(LBBB) is the most common disturbance.1
The presence of an LBBB results in delayed activation
of the lateral wall relative to the right ventricle and
septum, leading to reduced mechano‐energetic efficiency
of the heart. This is similar to the effects of chronic right
ventricular (RV) pacing, with a slow depolarisation
wavefront propagating through the myocardium, rather
than the Purkinje system. There are numerous adverse
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
haemodynamic sequelae of LBBB, including abnormal
septal motion causing decreased regional ejection fraction,
redistribution of myocardial fibre strain with altered
regional blood flow, diastolic dysfunction due to shortened
left ventricular (LV) filling time, depressed dP/dt and worsening mitral regurgitation.2–4 Further, LBBB, as either a
contributor to, or simply a surrogate marker of, progressive
cardiac dysfunction, is associated with increased morbidity
and mortality.5
Cardiac resynchronisation therapy (CRT)
CRT aims to reverse the deleterious effects of LBBB, as
the right and left ventricular pacing leads generate two
ventricular activation wavefronts that move towards each
other. The benefit of CRT lies in the effective fusion of
these two depolarisation wavefronts, synchronising the
walls of the LV (Fig. 1).
Since implantation of the first transvenous CRT system
in 1998, numerous large randomised multicentre trials
have demonstrated significant benefit of CRT in appropriately selected HF patients with respect to multiple
end‐points. These have included functional class, exercise
capacity, hospitalisation, quality of life, and reverse
remodelling with improvement in ejection fraction, and
reduction in LV size and mitral regurgitation.6,7 Since the
255
Voskoboinik et al.
Figure 1 Pacing the septal and lateral walls with
cardiac resynchronisation therapy (CRT) (leads
denoted by yellow stars) resynchronises the
heart, negating the adverse effects of left bundle
branch block (LBBB).
Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure (COMPANION) and Cardiac
Resynchronization – Heart Failure (CARE‐HF) trials,
there has also been a reproducible demonstration of
mortality benefit (Table 1).
Patient selection
Current guidelines
While CRT is ‘life‐changing’ for some, an estimated one
third of patients derive little or no benefit.20 Given the
heterogeneity of the various studies, it has taken time to
develop a better understanding of which patients are
most likely to benefit to improve patient selection. This
has led to a change in the recommendations for CRT
since 2006, when this topic was last reviewed in this
journal.21 At that time, the widely accepted Class I indication was ‘severe systolic dysfunction’ (LV ejection fraction less than 35%) in sinus rhythm (SR) with QRS
duration of at least 120 ms, and moderate‐severe HF
symptoms (New York Heart Association (NYHA) Class
III–IV) despite optimal medical therapy.
Increasing data have emerged regarding the critical
importance of a LBBB pattern in deriving clinical
benefit from CRT. Patients with right bundle branch
block (RBBB) do not have a dyssynchrony pattern suitable for CRT (activation of their lateral wall is already
early), and often have concomitant RV dysfunction, pulmonary hypertension and more extensive conduction
disease conferring a worse prognosis. Patients with RBBB
undergoing CRT have significantly lower rates of symptomatic and echocardiographic response, and a lower
survival free from heart transplantation or ventricular
256~
assist device placement.22 A recent meta‐analysis
reinforced this lack of benefit of CRT in patients with
either RBBB or non‐specific intraventricular conduction
delay.23 In the recently published Multicenter Automatic
Defibrillator
Implantation
Trial
with
Cardiac
Resynchronization Therapy (MADIT‐CRT) long‐term follow up,24 there was even the suggestion that CRT without LBBB was associated with harm (adjusted hazard
for all‐cause mortality 1.57; 95% confidence interval
(CI): 1.03–2.39).
While the lack of CRT benefit in narrow QRS width
(<120 ms) is widely accepted, there has been increasing
debate about its benefit in those with ‘intermediate QRS
width’ (120–149 ms), who have less mechanical intra‐
and interventricular dyssynchrony than those with QRS
width >150 ms.25 In a large meta‐analysis of 5813
patients, there was no statistically significant reduction in
death and HF hospitalisation with CRT in this ‘intermediate QRS width’ group.26 Only those with QRS duration
≥150 ms now receive a Class IA indication in major
society guideline statements6,7 (Fig. 2). Unfortunately,
studies utilising various imaging modalities to assess
dyssynchrony have been largely disappointing, and
electrocardiogram (ECG) still remains the gold standard
for this purpose.
Also now receiving a Class I indication for CRT are
mildly symptomatic patients (NYHA Class II). The
Resynchronization–Defibrillation for Ambulatory Heart
Failure Trial (RAFT)27 was the first study to show a significant reduction in mortality in this group (hazard ratio
(HR) 0.75; 95% CI: 0.62–0.91) in addition to reduced
hospitalisation for HF. This was reinforced by MADIT‐
CRT,28,29 the largest CRT trial to date, of whom 85% of
patients were NYHA II, with significant reductions seen
© 2016 Royal Australasian College of Physicians
10
© 2016 Royal Australasian College of Physicians
≤35%
≤30%
≤30%
≤40%
≤35%
≤35%
≤35%
≤35%
≤35%
III/IV
II/III
I/II
I‡/II
III
III/IV
III/IV
II
II/III/IV
III/IV
III/IV
≤35%
≤35%
NYHA class
LVEF
<130†
≥120 or paced >200
≥130
≥120
<130†
≥150 or 120–149†
≥120
≥130
≥130
≥130
≥130
QRS (ms)
(primary end‐point) – trial terminated early
CRT arm: increased all-cause mortality or HF hospitalisation
increased device- or implantation‐related complications
CRT arm: reduced all-cause mortality and HF hospitalisation, but
arm with LBBB, no benefit (perhaps harm) in non-LBBB
Long-term follow up (7-years): lower all-cause mortality in CRT
years
end-point), improved LVEF and reduction in LV volumes at 2.4
CRT arm: lower all-cause mortality or non-fatal HF event (primary
delayed time-to-first heart failure hospitalisation
(primary end-point), but greater improvement in LVESVI and
CRT arm: no difference in heart failure composite response
120–130 ms
end-point); improvement in subgroup with QRS width
CRT arm: no difference in peak oxygen consumption (primary
MR, improved NYHA class, LVEF and QOL
CRT arm: reduced all-cause mortality or hospitalisation, LVESVI,
(combined primary end‐point)
CRT arm(s): reduced all-cause mortality or hospitalisation
dioxide output, reduced LV systolic and diastolic volumes
CRT arm: improved LVEF, NYHA class, ventilation to carbon
reduced LV systolic and diastolic dimensions
CRT arm: improved peak O2 consumption, 6-MWD, LVEF,
CRT arm: improved QOL, NYHA class, peak O2 consumption
hospitalisations
CRT arm: improved 6-MWD, NYHA class, QOL, LVEF, fewer
Significant findings
medical therapy; QOL, quality of life; RETHINQ, Resynchronization Therapy in Normal QRS.
defibrillator; LBBB, left bundle branch block; LVEF, left ventricular ejection fraction; LVESVI, left ventricular end‐systolic volume index; MR, mitral regurgitation; NYHA, New York Heart Association; OMT, optimal
†With echocardiographic evidence of dyssynchrony. ‡Previously symptomatic. 6‐MWD, 6‐min walk distance; CRT, cardiac resynchronisation therapy; CRT‐D, CRT with defibrillator; ICD, implantable cardioverter
CRT‐on versus CRT‐off
Double-blinded, parallel arm, 19 months
2013
CRT‐D versus ICD
ECHO‐CRT19
809
2010
Double-blinded, parallel arm, 40 months
RAFT18
1798
CRT‐D versus ICD
and 201417
Single-blinded, parallel arm
MADIT‐CRT 200916
1820
CRT‐on versus CRT‐off
2008
Double-blinded, parallel arm, 12 months
610
REVERSE15
Double-blinded, parallel arm, 6 months
CRT‐on versus CRT‐off
172
CRT versus OMT
2007
RETHINQ
14
Double-blinded, parallel arm, 29 months
2005
CRT‐D versus CRT versus OMT
813
Double-blinded, parallel arm, 15 months
CRT‐D versus ICD
CARE‐HF13
1520
2004
COMPANION12
Double-blinded, parallel arm, 6 months
186
MIRACLE‐ICD II112004
Double-blinded, crossover, 6 months
CRT‐D versus ICD
490
CRT‐D versus ICD
2003
CONTAK‐CD
2003
Double-blinded, parallel arm, 6 months
369
MIRACLE‐ICD9
Double-blinded, parallel arm, 6 months
CRT versus OMT
453
MIRACLE8
Study design and follow up
2002
n
Study and year(s)
Table 1 Largest published randomised trials of cardiac resynchronisation therapy
Cardiac resynchronisation therapy
257
Voskoboinik et al.
Figure 2 Major society guidelines for cardiac
resynchronisation therapy. EF, ejection fraction;
LBBB, left bundle branch block; NYHA, New York
Heart Association; SR, sinus rhythm.
in all‐cause mortality and non‐fatal HF events again (particularly in those with QRS duration ≥150 ms and LBBB).
Atrial fibrillation (AF) and CRT
Atrial tachyarrhythmias remain common in the HF
population and interfere with CRT response by two mechanisms. First, they worsen the course of HF in general, by
reducing cardiac output through loss of atrial contribution, while rapid ventricular rates interfere with diastolic
filling and can result in a tachycardia‐mediated cardiomyopathy. Second, critical to the efficacy of CRT is that the
ventricles are solely activated by biventricular (BiV) pacing,
with conduction through the atrioventricular node
(AVN) completely pre‐empted. Conduction of atrial
tachyarrhythmias
to
the
ventricles
results
in
dyssynchronous activation in patients with underlying
LBBB, while competition with BiV capture can lead to
fusion or pseudo‐fusion beats.
While there are no large randomised trials looking
directly at CRT in the AF population, a large metaanalysis30 of 23 observational studies involving 7494 CRT
recipients showed that AF was associated with an
increased risk of non‐response to CRT (34.5% vs 26.7%;
pooled relative risk 1.32, P = 0.01). AVN ablation (AVNA)
significantly reduced the risk of clinical non‐response and
death in this population. In a systematic review of 768 AF
patients undergoing CRT, Ganesan et al. demonstrated
significant reductions in mortality and improvement in
symptoms with AVNA compared with pharmacological
therapy.31 In the CERTIFY study,32 AVNA proved superior
to pharmacological rate control at maximising CRT
benefit, with the rate control group having a higher total
and cardiac mortality than both the SR and AVNA groups.
Interestingly, there was no mortality difference between
the SR group and those with AF and AVNA.
Presently, the AHA guidelines give a Class IIa recommendation for CRT in persistent AF patients who
otherwise meet CRT criteria in whom AVNA or pharmacological rate control will allow ‘near 100% pacing’ with
CRT.7 Particularly in AF patients with QRS width
>150 ms who otherwise satisfy the criteria, CRT is still
better than no CRT (just not as beneficial as for SR
258~
patients with QRS width >150 ms). Unfortunately,
Medicare does not currently provide a rebate for CRT
in AF patients, regardless of ejection fraction or QRS
duration. The Australian prospective randomised Cardiac
Resynchronisation Therapy and AV Nodal Ablation Trial
in Atrial Fibrillation (CAAN‐AF)33 is currently recruiting
patients to assess further the role of AVNA in AF patients
with broad QRS and severe systolic dysfunction.
Pacemaker induced dyssynchrony and the role
of CRT
RV apical pacing for patients with symptomatic AV block
creates dyssynchrony, with QRS prolongation similar to
an LBBB morphology, and results in adverse ventricular
remodelling.34–36 Importantly, LV systolic dysfunction is
seen in up to 40% of these patients and is related to the
percentage of RV pacing.24 The detrimental effects of RV
pacing on LV systolic function is well documented, with
increased AF,37 as well as HF hospitalisation and mortality. In the Dual Chamber and VVI Implantable Defibrillator (DAVID) trial of 506 HF patients with implantable
cardioverter defibrillator (ICD), patients randomised to a
backup mode (VVI‐40) had significantly superior outcomes (mortality and hospitalisation) compared with
those with a dual‐chamber pacing mode (DDDR‐70).38
Various strategies have been employed to minimise RV
apical pacing. These have included bradycardia prevention, programming long AV delays to promote intrinsic
conduction and alternate pacing sites, such as the high
RV septum or His bundle.
While undoubtedly the most physiological site, His‐
bundle pacing (HBP) is more technically challenging.
Sharma et al. recently reported an 80% success rate in
achieving permanent HBP (without a mapping catheter
or backup RV lead) using the Medtronic Select Secure
active fixation lead delivered through a fixed‐shaped
catheter. The majority of patients in this study had
>40% ventricular pacing, and HF hospitalisation was
significantly reduced in the HBP group compared
with the RV‐paced group (2% vs 15%; P = 0.02).39
Disappointingly, the randomised Protection of Left Ventricular Function During Right Ventricular Pacing
© 2016 Royal Australasian College of Physicians
Cardiac resynchronisation therapy
Figure 3 Major society guidelines for CRT in patients with conventional indications for anti‐bradycardia pacing. CRT, cardiac resynchronisation
therapy; EF, ejection fraction; ICD, implantable
cardioverter defibrillator; PPM, permanent pacemaker; RV, right ventricular.
(PROTECT‐PACE) study of 240 patients did not demonstrate any difference in ejection fraction drop between
RV apical and high RV septal positions at 2 years.40 However, in recent years, CRT has emerged as a promising
option in these patients.
The largest trial of its kind, Biventricular versus Right
Ventricular Pacing in Heart Failure Patients with Atrioventricular Block (BLOCK‐HF),41 randomised 918 patients
with AV block (with a need for significant ventricular pacing), ejection fraction (EF) <50% and NYHA I–III to receive
either RV pacing or CRT. The primary composite outcome
(all‐cause mortality or HF requiring intravenous therapy
or >15% increase in the left ventricular end‐systolic volume index) occurred in 55.6% of RV‐paced patients,
compared with 45.8% of BiV‐paced patients (HR 0.74;
95% CI: 0.60–0.90). Certainly, there remains a paucity of
data to support the practice of routinely implanting CRT
for all patients with indications for a pacemaker. Most
will not develop LV dysfunction or HF, and subjecting
them to a longer procedure with increasing peri-procedural complications is inappropriate. However, in
RV-paced patients who, during follow up, develop symptoms of HF and have a deteriorating EF should be
upgraded to CRT to reduce hospitalisation, and improve
symptoms and cardiac function.6 Current international
guidelines (Fig. 3) advocate ‘up‐front’ implantation of a
CRT device in patients with severe pre‐existing systolic
dysfunction (EF <35%) who are likely to require significant RV pacing (>40% RV‐paced was found to be predictive of adverse outcomes in a substudy of the DAVID
trial42). While there are no Australian guidelines for this,
the results of BLOCK‐HF suggest this practice should
even be considered in those with milder degrees of systolic dysfunction.
Role of multimodality imaging
As stated, electrical dyssynchrony results in mechanical
dyssynchrony, and multiple imaging modalities have
been used to confirm its presence and determine the
site of latest LV activation. Studies have shown that up
© 2016 Royal Australasian College of Physicians
to 50% of patients with a narrow QRS (<120 ms) have
echocardiographic evidence of dyssynchrony,43 and
multiple early single‐centre studies44,45 (employing predominantly M‐mode and tissue Doppler) were able to
variably demonstrate reverse remodelling with CRT in
these patients. This sparked considerable interest in
expanding the existing CRT indications, which mandated
a broad QRS.
The larger Resynchronization Therapy in Normal QRS
(RETHINQ)46 and Evaluation of CRT in Narrow QRS Patients With Mechanical Dyssynchrony (ESTEEM‐CRT)47
multicentre studies soon cast doubt on this notion as
they failed to show improvement in exercise capacity
in narrow QRS patients with echocardiographic evidence of dyssynchrony. More recently, the multicentre
EchoCRT randomised trial48 assessed patients with QRS
width <130 ms and echocardiographic dyssynchrony
(defined by tissue Doppler opposing‐wall delay in peak
systolic velocity >80 ms or speckle tracking radial strain delay in anteroseptal‐to‐posterior wall >130 ms). After following 809 patients for a mean 19.4 months, the study was
stopped early due to increased mortality in the CRT arm
(11.1% vs 6.4%). This reinforced the notion that current
dyssynchrony indices (apart from the ECG) are ineffective at selecting narrow QRS patients for CRT, and
may worsen outcomes in patients with narrow QRS.
Accordingly, QRS width (with or without mechanical
dyssynchrony) remains the key determinant of clinical
response to CRT, and the only Class I indication.
While availability is currently limited in many Australasian centres, cardiac MRI (CMR) enables the assessment
of both dyssynchrony and scar burden in a single study.
It has been well established that pacing in regions of
scar, as detected by delayed enhancement MRI, predicts
lack of response to CRT.49,50 Combined with numerous
validated measures of dyssynchrony (cine MRI, myocardial tagging, strain imaging), CMR is gaining increasing
attention due to its excellent tissue characterisation,
high spatial resolution and reproducibility.51 Like
multidetector cardiac computed tomography (CT), CMR
also enables the assessment of cardiac venous anatomy.
259
Voskoboinik et al.
In patients with significant scar who may exhibit less
extensive venous anatomy, this may mean a minimally
invasive surgical approach rather than a transvenous
approach. Notably, delineation of the precise number and
location of tributaries of the coronary sinus by CT minimises peri‐procedural venography, and has been shown
to shorten procedural times and limit contrast and radiation exposure.52
Looking towards the future, myocardial imaging
has the potential to individualise CRT. It can confirm
candidacy by demonstrating significant mechanical
dyssynchrony in an otherwise borderline candidate on
ECG criteria. It can also guide lead placement by targeting
an appropriate vein adjacent to the latest activated region
of myocardium and away from scar.
CRT leads: optimal position and new
technologies
Lead positioning
Appropriate positioning of the LV lead is a key determinant of CRT success. In the absence of additional information regarding activation sequence, the lead should be
positioned in the lateral or posterolateral walls, which are
usually activated last and are furthest away from the RV
lead. In a retrospective analysis of 457 CRT recipients,
Dong et al. demonstrated greater improvement in NYHA
class with lateral LV lead positions compared with anterior locations.53 Anterior LV lead positions also have a
significantly higher risk of ventricular arrhythmias than
posterior or lateral positions.54 Basal or mid‐ventricular
LV lead positions are superior to apical locations. In an
analysis of 799 patients in the MADIT‐CRT study, apical
LV lead position was associated with a significantly
increased risk of HF (HR 1.72; 95% CI: 1.09–2.71) and
death (HR 2.91; 95% CI: 1.42–5.97).55
The location of the RV lead is less important, and there
is certainly no clinical benefit of non‐apical RV lead locations (i.e. septum, RV outflow tract) compared with
apical locations.56 There is also some evidence that a
larger horizontal inter‐lead distance (as measured on a
lateral chest X‐ray), which correlates with a longer LV
lead electrical delay, can improve LV anatomic reverse
remodelling post‐CRT.57 The amount of QRS shortening
post‐CRT (DeltaQRS), which relates to both inter‐lead
distance and LV lead electrical delay, is a strong predictor
of improved outcomes post‐CRT. Molhoek et al. demonstrated that a DeltaQRS >10 ms has a high sensitivity
(73%) while a DeltaQRS >50 ms has a high specificity
(88%) for CRT response.58
Showing promise is the use of echocardiography to
guide lead placement. Unlike tissue Doppler, speckle
260
tracking radial strain can distinguish between active and
passive motion (scar tethering) and is able to determine
the site of latest contraction. The Targeted Left Ventricular
Lead Placement to Guide Cardiac Resynchronization Therapy (TARGET) study59 randomised 220 patients to either
standard unguided CRT or LV lead positioning at the site
of latest peak contraction with an amplitude of >10% to
signify freedom from scar. The latter group had a greater
response rate (70% vs 55%, P = 0.031) with respect to
reverse remodelling, as well as lower all‐cause mortality
and hospitalisation.
However, significant variability in coronary venous
anatomy exists between patients, and the ability to
target directly LV lead position is a limitation of the
transvenous route. In a large series investigating distribution of venous anatomy and its effects on LV lead
targeting, Khan et al. demonstrated that limitations of
coronary venous anatomy restrict LV lead placement to a
single vein with little scope for site selection in almost
half of all patients.60 In patients with multiple target
veins, intraprocedural coronary venous electroanatomic
mapping has been assessed as a means of guiding LV lead
placement to the latest activated region free from phrenic
nerve stimulation. In a recently published series, Rad
et al. demonstrated that this latest activated region
(region with an electrical delay >75% of total QRS duration) was more often located anterolaterally, rather than
the empirically targeted inferolateral vein.28
Epicardial leads
Direct surgical epicardial approach may overcome these
limitations, and should be considered in non‐responders
due to suboptimal lead placement, and in those with high
thresholds or intractable phrenic nerve stimulation. Previously only performed by thoracotomy (or sternotomy),
newer techniques have emerged to limit surgical morbidity, including left lateral mini‐thoracotomy, video‐
assisted thoracoscopy and the use of robotic technology.
While long‐term outcomes with respect to mortality and
functional improvement are similar to the transvenous
route, these patients often have longer in‐hospital length
of stay29 and higher perioperative complications, including renal failure and infection.27
Emerging lead technologies
Compared with isolated RV pacing, patients undergoing
CRT with standard bipolar leads have higher rates of LV
lead complications, including lead dislodgement (5.7%)
and coronary sinus complications (2.0%).61 In addition to
loss of capture related to these problems, some patients
also require reoperation due to phrenic nerve stimulation
or increased LV pacing thresholds without obvious
© 2016 Royal Australasian College of Physicians
Cardiac resynchronisation therapy
dislocation. New quadripolar leads have four independent electrodes, which enable programming of additional
different vectors for LV pacing. This provides the operator
greater flexibility, as they can programme different electrode configurations, rather than having to reposition an
entire lead. Down the track, they are more likely to find
a stable position with reasonable pacing thresholds while
minimising phrenic nerve stimulation.46,62
Quadripolar leads also enable multipoint pacing, which
provides greater flexibility for optimal pacing configurations. This technology (MultiPoint Pacing, St. Jude
Medical) delivers two LV pulses per pacing cycle from a
single quadripolar lead. Early studies show some promise
in further optimising haemodynamics, improving contractility with echocardiographic guidance63 and reducing
the rate of CRT non‐response.64
Newer leads are also now being designed to monitor
haemodynamic parameters, which will hopefully enable
device optimisation, guide LV lead positioning and aid in
clinical management. Intracardiac impedance reflects LV
volume changes well in animal models, and early studies
involving Biotronik's intracardiac impedance sensing
technology have demonstrated good correlation with LV
stroke volume.65 Sorin's SonR system utilises a sensor
embedded in the lead tip that detects cardiac muscle
vibrations (peak endocardial acceleration signal system).
In a recent randomised study, automatic optimisation of
AV and ventriculo‐ventricular (VV) delays based on this
system resulted in improved CRT response, with significant improvement in NYHA class.48
Device programming and follow up
Improving %BiV pacing:
arrhythmia suppression
The greatest benefit is derived from CRT when it is used
maximally. Hayes et al. demonstrated a 24% reduction
in mortality in patients receiving above 98.5% BiV
pacing, as compared with those receiving <95% BiV
pacing.66 Incremental increases in mortality benefit are
seen with an increasing percentage of BiV pacing. Atrial
tachyarrhythmias, predominantly AF, are significant contributors to loss of pacing. Santini et al. demonstrated that
even a small burden of atrial arrhythmia in CRT patients
resulted in higher rates of death or HF hospitalisation.67
Aggressive suppression of atrial tachyarrhythmias (with
antiarrhythmics, cardioversion as required, and/or catheter ablation) to maintain SR may be beneficial in ensuring AV synchrony and achieving a high %BiV pacing.
Frequent premature ventricular contractions (PVC) also
contribute to CRT non‐response. Lakkireddy et al. demonstrated significant improvements to LV function, size
© 2016 Royal Australasian College of Physicians
and NYHA class by ablating PVC foci in CRT non‐responders with >10 000 PVC in a 24‐h period.68
Improving %BiV pacing: pacing algorithms
While routine optimisation of AV delay is controversial,
an inappropriately programmed long AV delay may result
in loss of BiV pacing in many patients.69 Avoidance of
right atrial pacing is preferred, because pacing the right
atrial appendage delays left atrial activation, which may
impair LV preload and worsen myocardial performance.
This can be achieved by programming a VDD mode,
rather than a DDD mode.70
A novel adaptive CRT algorithm (AdaptivCRT,
Medtronic) in patients with SR and normal AV conduction is also showing great promise. This algorithm provides LV pacing synchronised to produce fusion with intrinsic RV activation (to minimise RV pacing), with
dynamic optimisation of AV and VV delays. It has been
shown to be safe and at least as effective as BiV pacing
with echocardiographic optimisation,71 with the suggestion of better clinical outcomes in those with a higher
percentage of LV pacing.72
Echocardiographic optimisation
The use of echocardiography to assess mechanical
dyssynchrony and select patients for CRT has been disappointing. However, modern CRT devices enable
programming of AV and VV delays, which can be
optimised based on echocardiographic evaluation of
haemodynamics. With respect to optimising AV delay,
numerous Doppler echocardiographic parameters have
been studied. These include LV outflow tract velocity
time integral (VTI), diastolic filling time and transmitral
flow (EA VTI), with the latter correlating best with invasive measurements (LV dP/dtmax).73 However, the
SmartDelay Determined AV Optimization (SMART‐AV)
randomised trial of 980 patients demonstrated no clinical benefit of an echocardiography‐optimised AV delay
over a fixed empirical AV delay of 120 ms.74 While
many of the landmark CRT trials incorporated some
form of echocardiography‐guided AV delay optimisation
(based on small haemodynamic studies), long‐term
randomised data supporting this practice are lacking.
Several small studies of echocardiography‐guided VV
optimisation have demonstrated acute haemodynamic
improvements compared with simultaneous BiV pacing.
Bordachar et al. demonstrated that improvements in
intraventricular dyssynchrony (using tissue Doppler) correlated with an increase in cardiac output, with simultaneous BiV pacing being the optimal setting in only 15%
of patients.75 In a randomised study of 238 patients with
261
Voskoboinik et al.
VV intervals adjusted to minimise septal to posterior wall
delay, significantly more patients improved based on a
clinical composite score; however, no significant differences were found with respect to exercise capacity,
quality of life or HF event rate. Again, simultaneous BiV
pacing was the optimal setting in only 18%.76 In addition
to a paucity of consistent clinical data demonstrating
clear benefit, echocardiographic optimisation of AV and
VV delays is time‐consuming and resource‐intensive, and
requires significant skill. However, it may be useful
in CRT non‐responders.
Integrated teams and remote monitoring
With the burden of hospital readmission rates for HF
rapidly increasing, multidisciplinary clinics encompassing
nurse practitioners, physiotherapists, social workers,
dieticians, psychologists and specialists are being
developed to deliver holistic, patient‐centred care. A
multidisciplinary approach, coupled with individualised
CRT optimisation (incorporating echocardiography‐
guided adjustment in device settings and optimisation
of medical therapy), led to significant improvements in
LV function and reduction in adverse events in 75 CRT
non‐responders. A reason for CRT non‐response was
identified in most patients, and multidisciplinary recommendations led to a change in treatment in 74%.77
Remote monitoring of CRT devices, with the regular
transmission of important data such as atrial and ventricular arrhythmia burden, heart rate histograms, device
safety issues and ‘%BiV‐paced’, enables regular ‘contact’
with cardiologists and earlier optimisation of therapy.
References
1 Kashani A, Barold SS. Significance of QRS
complex duration in patients with heart
failure. J Am Coll Cardiol 2005; 46:
2183–92.
2 Grines CL, Bashore TM, Boudoulas H,
Olson S, Shafer P, Wooley CF.
Functional abnormalities in isolated left
bundle branch block. The effect of
interventricular asynchrony. Circulation
1989; 79: 845–53.
3 Zhou Q, Henein M, Coats A, Gibson D.
Different effects of abnormal activation
and myocardial disease on left
entricular ejection and filling times.
Heart 2000; 84: 272–6.
4 Bramlet DA, Morris KG, Coleman RE,
Albert D, Cobb FR. Effect of
rate-dependent left bundle branch block
on global and regional left
262
In the Implant-based Multiparameter Telemonitoring of
Patients with Heart Failure (IN‐TIME) randomised trial
of 716 patients, most of whom had CRT‐D (CRT and
defibrillator) devices, those receiving automatic, daily
multi‐parameter tele‐monitoring had a significantly
lower incidence of the primary end‐point (combined
all‐cause mortality, hospitalisation, change in NYHA
class and symptoms). This was largely driven by earlier
recognition of suboptimal device function, detection of
arrhythmias and patient interview prompted by remote
monitoring findings.78 Early results from the Patient
Related Determinants of ICD Remote Monitoring Utilization and Outcomes (PREDICt‐RM) study, which analysed
registry data from 37 742 patients with new ICD or CRTD implants, demonstrated lower mortality (HR = 0.67;
95% CI: 0.64–0.7) and hospitalisation (HR = 0.81; 95%
CI: 0.79–0.83) in those who used remote monitoring.
Despite its many benefits, remote monitoring is currently
underutilised in Australia.79
Conclusion
CRT remains one of the key successes of the ‘device era’
of HF therapy. The past decade has given us a greater
understanding about the importance of appropriate
patient selection, correct lead positioning and device
optimisation to ensure a high percentage of BiV pacing.
With novel indications, new lead technologies, better
remote monitoring tools and integrated multimodality
imaging showing great promise, CRT remains an exciting
and rapidly evolving field.
ventricular function. Circulation 1983; 67:
1059–65.
5 Zannad F, Huvelle E, Dickstein K, van
Veldhuisen DJ, Stellbrink C, Køber L
et al. Left bundle branch block as a risk
factor for progression to heart failure.
Eur J Heart Fail 2007; 9: 7–14.
6 Brignole M, Auricchio A,
Baron‐Esquivias G, Bordachar P, Boriani
G, Breithardt OA et al. ESC guidelines on
cardiac pacing and cardiac
resynchronization therapy. Eur Heart J
2013; 2013: 2281–329.
7 Tracy CM, Epstein AE, Darbar D,
Dimarco JP, Dunbar SB, Estes NA et al.
ACCF/AHA/HRS focused update of
the 2008 guidelines for device‐based
therapy of cardiac rhythm abnormalities. J
Am Coll Cardiol 2012; 60:
1297–313.
8 Abraham WT, Fisher WG, Smith AL,
Delurgio DB, Leon AR, Loh E et al.
Cardiac resynchronization in chronic
heart failure. N Engl J Med 2002; 346:
1845–53.
9 Young JB, Abraham WT, Smith AL,
Leon AR, Lieberman R, Wilkoff B et al.
Combined cardiac resynchronization and
implantable cardioversion defibrillation
in advanced chronic heart failure: the
MIRACLE ICD Trial. JAMA 2003; 289:
2685–94.
10 Higgins SL, Hummel JD, Niazi IK,
Giudici MC, Worley SJ, Saxon LA et al.
Cardiac resynchronization therapy for
the treatment of heart failure in patients
with intraventricular conduction delay
and malignant ventricular
tachyarrhythmias. J Am Coll Cardiol
2003; 42: 1454–9.
© 2016 Royal Australasian College of Physicians
Cardiac resynchronisation therapy
11 Abraham WT, Young JB, León AR,
Adler S, Bank AJ, Hall SA et al. Effects of
cardiac resynchronization on disease
progression in patients with left
ventricular systolic dysfunction, an
indication for an implantable
cardioverter‐defibrillator, and mildly
symptomatic chronic heart failure.
Circulation 2004; 110: 2864–8.
12 Bristow MR, Saxon LA, Boehmer J,
Krueger S, Kass DA, De Marco T et al.
Cardiac‐resynchronization therapy with
or without an implantable defibrillator
in advanced chronic heart failure. N Engl
J Med 2004; 350: 2140–50.
13 Cleland JG, Daubert JC, Erdmann E,
Freemantle N, Gras D, Kappenberger L
et al. The effect of cardiac
resynchronization on morbidity and
mortality in heart failure. N Engl J Med
2005; 352: 1539–49.
14 Beshai JF1, Grimm RA, Nagueh SF,
Baker JH 2nd, Beau SL, Greenberg SM
et al. Cardiac‐resynchronization therapy
in heart failure with narrow QRS
complexes. N Engl J Med 2007; 357:
2461–7.
15 Linde C, Abraham WT, Gold MR,
St John Sutton M, Ghio S, Daubert C.
Randomized trial of cardiac
resynchronization in mildly symptomatic
heart failure patients and in
asymptomatic patients with left
ventricular dysfunction and previous
heart failure symptoms. J Am Coll Cardiol
2008; 52: 1834–43.
16 Moss AJ, Hall WJ, Cannom DS, Klein H,
Brown MW, Daubert JP et al.
Cardiac‐resynchronization therapy for
the prevention of heart‐failure events.
N Engl J Med 2009; 361: 1329–38.
17 Goldenberg I, Kutyifa V, Klein HU,
Cannom DS, Brown MW, Dan A et al.
Survival with cardiac‐resynchronization
therapy in mild heart failure. N Engl J
Med 2014; 370: 1694–701.
18 Tang AS, Wells GA, Talajic M, Arnold
MO, Sheldon R, Connolly S et al.
Cardiac‐resynchronization therapy for
mild-to-moderate heart failure. N Engl J
Med 2010; 363: 2385–95.
19 Ruschitzka F, Abraham WT, Singh JP,
Bax JJ, Borer JS, Brugada J et al.
Cardiac‐resynchronization therapy in
heart failure with a narrow QRS
complex. N Engl J Med 2013; 369:
1395–405.
20 Yu CM, Hayes DL. Cardiac
resynchronization therapy. Eur Heart J
2013; 34: 1396–403.
© 2016 Royal Australasian College of Physicians
21 Mariani JA, Gould PA, Broughton A,
Kaye DM. Cardiac resynchronisation
therapy for heart failure. Intern Med J
2006; 36: 114–23.
22 Adelstein EC, Saba S. Usefulness of
baseline electrocardiographic QRS
complex pattern to predict response to
cardiac resynchronization. Am J Cardiol
2009; 103: 238–42.
23 Sipahi I, Chou JC, Hyden M, Rowland
DY, Simon DI, Fang JC. Effect of QRS
morphology on clinical event reduction
with cardiac resynchronization therapy:
meta‐analysis of randomized controlled
trial. Am Heart J 2012;
163: 260–7.
24 Begg GA, Gierula J, Waldron ZL, Witte
KK. Patients receiving standard
pacemaker generator replacements
frequently have impaired left ventricular
function and exercise intolerance,
related to the percentage of right
ventricular pacing. Heart 2011; 97: A86.
25 Haghjoo M, Bagherzadeh A, Fazelifar
AF, Haghighi ZO, Esmaielzadeh M,
Alizadeh A et al. Prevalence of
mechanical dyssynchrony in heart
failure patients with different QRS
durations. Pacing Clin Electrophysiol 2007;
30: 616–22.
26 Sipahi I, Carrigan TP, Rowland DY,
Stambler BS, Fang JC. Impact of QRS
duration on clinical event reduction
with cardiac resynchronization therapy:
meta‐analysis of randomized controlled
trials. Arch Intern Med 2011; 171:
1454–62.
27 Ailawadi G, Lapar DJ, Swenson BR,
Maxwell CD, Girotti ME, Bergin JD et al.
Surgically placed left ventricular leads
provide similar outcomes to
percutaneous leads in patients
with failed coronary sinus lead
placement. Heart Rhythm 2010; 7:
619–25.
28 Rad MM, Blaauw Y, Dinh T, Pison L,
Crijns HJ, Prinzen FW et al. Left
ventricular lead placement in the latest
activated region guided by coronary
venous electroanatomic mapping.
Europace 2015; 17: 84–93.
29 Koos R, Sinha AM, Markus K,
Breithardt OA, Mischke K, Zarse M et al.
Comparison of left ventricular lead
placement via the coronary venous
approach versus lateral thoracotomy in
patients receiving cardiac
resynchronization therapy. Am J Cardiol
2004; 94: 59–63.
30 Wilton SB, Leung AA, Ghali WA, Faris
P, Exner DV. Outcomes of cardiac
resynchronization therapy in patients
with versus those without atrial
fibrillation: a systematic review and
meta‐analysis. Heart Rhythm 2011; 8:
1088–94.
31 Ganesan AN, Brooks AG,
Roberts‐Thomson KC, Lau DH, Kalman
JM, Sanders P. Role of AV nodal
ablation in cardiac resynchronization in
patients with coexistent atrial fibrillation
and heart failure a systematic review. J
Am Coll Cardiol 2012; 59: 719–26.
32 Gasparini M, Leclercq C, Lunati M,
Landolina M, Auricchio A, Santini M
et al. Cardiac resynchronization therapy
in patients with atrial fibrillation. JACC
Heart Fail 2013; 1: 500–7.
33 National Institute of Mental Health,
University of Virginia. Cardiac
resynchronisation therapy and AV nodal
ablation trial in atrial fibrillation patients
(CAAN‐AF). In: Sanders P, ed.
ClinicalTrials.gov [Internet]. Bethesda
(MD): National Library of Medicine
(US); 2000 [cited 2015 Jan 20]. Available
from URL: https://clinicaltrials.gov/ct2/
show/NCT01522898
34 Nielsen JC, Kristensen L, Andersen HR,
Mortensen PT, Pedersen OL, Pedersen
AK. A randomized comparison of atrial
and dual‐chamber pacing in 177
consecutive patients with sick sinus
syndrome: echocardiographic and
clinical outcome. J Am Coll Cardiol 2003;
42: 614–23.
35 Karpawich PP, Rabah R, Haas JE.
Altered cardiac histology following
apical right ventricular pacing in patients
with congenital atrioventricular block.
Pacing Clin Electrophysiol 1999; 22:
1372–1377.
36 Vernooy K, Dijkman B, Cheriex EC,
Prinzen FW, Crijns HJ. Ventricular
remodeling during long‐term right
ventricular pacing following His bundle
ablation. Am J Cardiol 2006; 97: 1223–7.
37 Sweeney MO, Hellkamp AS, Ellenbogen
KA, Greenspon AJ, Freedman RA, Lee KL
et al. Adverse effect of ventricular pacing
on heart failure and atrial fibrillation
among patients with normal baseline QRS
duration in a clinical trial of pacemaker
therapy for sinus node dysfunction
(MOST). Circulation 2003; 107: 2932–7.
38 Wilkoff BL, Cook JR, Epstein AE,
Greene HL, Hallstrom AP, Hsia H et al.
Dual‐chamber pacing or ventricular
263
Voskoboinik et al.
backup pacing in patients with an
implantable defibrillator (DAVID). JAMA
2002; 288: 3115–23.
39 Sharma PS, Dandamudi G, Naperkowski
A, Oren JW, Storm RH, Ellenbogen KA
et al. Permanent His‐bundle pacing is
feasible, safe, and superior to right
ventricular pacing in routine clinical
practice. Heart Rhythm 2015; 12: 305–12.
40 Kaye GC, Linker NJ, Marwick TH,
Pollock L, Graham L, Pouliot E et al.
Effect of right ventricular pacing lead
site on left ventricular function in
patients with high‐grade atrioventricular
block: results of the Protect‐Pace study.
Eur Heart J 2015; 36: 856–62.
41 Curtis AB, Worley SJ, Adamson PB,
Chung ES, Niazi I, Sherfesee L et al.;
Biventricular versus Right Ventricular
Pacing in Heart Failure Patients with
Atrioventricular Block (BLOCK HF) Trial
Investigators. Biventricular pacing for
atrioventricular block and systolic
dysfunction. N Engl J Med 2013; 368:
1585–93.
42 Sharma AD, Rizo‐Patron C, Hallstrom
AP, O'Neill GP, Rothbart S, Martins JB
et al. Percent right ventricular pacing
predicts outcomes in the DAVID trial.
Heart Rhythm 2005; 2: 830–4.
43 Yu CM, Lin H, Zhang Q, Sanderson JE.
High prevalence of left ventricular
systolic and diastolic asynchrony in
patients with congestive heart failure
and normal QRS duration. Heart 2003;
89: 54–60.
44 Pitzalis MV, Iacoviello M, Romito R,
Massari F, Rizzon B, Luzzi G et al.
Cardiac resynchronization therapy
tailored by echocardiographic evaluation
of ventricular asynchrony. J Am Coll
Cardiol 2002; 40: 1615–22.
45 Pitzalis MV, Iacoviello M, Romito R,
Guida P, De Tommasi E, Luzzi G et al.
Ventricular asynchrony predicts a better
outcome in patients with chronic heart
failure receiving cardiac
resynchronization therapy. J Am Coll
Cardiol 2005; 45: 65–9.
46 Burger H, Schwarz T, Ehrlich W, Sperzel
J, Kloevekorn WP, Ziegelhoeffer T. New
generation of transvenous left
ventricular leads – first experience with
implantation of multipolar left
ventricular leads. Exp Clin Cardiol 2011;
16: 23–6.
47 Donahue T, Niazi I, Leon A, Stucky M,
Herrmann K. Acute and chronic
response to CRT in narrow QRS patients
264
(ESTEEM‐CRT study). J Cardiovasc Transl
Res 2012; 5: 232–41.
48 Ritter P, Delnoy PP, Padeletti L, Lunati
M, Naegele H, Borri‐Brunetto A et al. A
randomized pilot study of optimization
of cardiac resynchronization therapy in
sinus rhythm patients using a peak
endocardial acceleration sensor vs.
standard methods. Europace 2012; 14:
1324–33.
49 White JA, Yee R, Yuan X. Delayed
enhancement magnetic resonance
imaging predicts response to cardiac
resynchronization therapy in
patients with intraventricular
dyssynchrony. J Am Coll Cardiol 2006;
48: 1953–60.
50 Bleeker GB, Kaandorp TA, Lamb HJ.
Effect of posterolateral scar tissue on
clinical and echocardiographic
improvement after cardiac
resynchronization therapy. Circulation
2006; 113: 969–76.
51 Delgado V, Bax JJ. Assessment of systolic
dyssynchrony for cardiac
resynchronization therapy is clinically
useful. Circulation 2011; 123: 640–55.
52 Girsky MJ, Shinbane JS, Ahmadi N,
Mao S, Flores F, Budoff MJ. Prospective
randomized trial of venous cardiac
computed tomographic angiography for
facilitation of cardiac resynchronization
therapy. Pacing Clin Electrophysiol 2010;
33: 1182–7.
53 Dong YX1, Powell BD, Asirvatham SJ,
Friedman PA, Rea RF, Webster TL et al.
Left ventricular lead position for cardiac
resynchronization: a comprehensive
cinegraphic, echocardiographic, clinical,
and survival analysis. Europace 2012; 14:
1139–47.
54 Kutyifa V, Zareba W, McNitt S, Singh J,
Hall WJ, Polonsky S et al. Left
ventricular lead location and the risk of
ventricular arrhythmias in the
MADIT‐CRT trial. Eur Heart J 2013; 34:
184–90.
55 Singh JP, Klein HU, Huang DT, Reek S,
Kuniss M, Quesada A et al. Left
ventricular lead position and clinical
outcome in the multicenter automatic
defibrillator implantation trial–cardiac
resynchronization therapy (MADIT‐CRT)
Trial. Circulation 2011; 123: 1159–66.
56 Singh JP, Klein HU, Huang DT, Reek S,
Kuniss M, Quesada A et al. Impact of the
right ventricular lead position on clinical
outcome and on the incidence of
ventricular tachyarrhythmias in patients
with CRT‐D. Heart Rhythm 2013; 10:
1770–7.
57 Merchant FM, Heist EK, Nandigam KV,
Mulligan LJ, Blendea D, Riedl L et al.
Interlead distance and left ventricular
lead electrical delay predict reverse
remodeling during cardiac
resynchronization therapy. Pacing Clin
Electrophysiol 2010; 33: 575–82.
58 Molhoek SG, Van Erven L, Bootsma M,
Steendijk P, Van Der Wall EE, Schalij MJ.
QRS duration and shortening to
predict clinical response to cardiac
resynchronization therapy in patients
with end‐stage heart failure. Pacing Clin
Electrophysiol 2004; 27: 308–13.
59 Khan FZ, Virdee MS, Palmer CR, Pugh
PJ, O'Halloran D, Elsik M et al. Targeted
left ventricular lead placement to guide
cardiac resynchronization therapy: the
TARGET study: a randomized, controlled
trial. J Am Coll Cardiol 2012; 59:
1509–18.
60 Khan FZ, Virdee MS, Gopalan D, Rudd
J, Watson T, Fynn SP et al.
Characterization of the suitability of
coronary venous anatomy for targeting
left ventricular lead placement in
patients undergoing cardiac
resynchronization therapy. Europace
2009; 11: 1491–5.
61 Van Rees JB, de Bie MK, Thijssen J,
Borleffs CJ, Schalij MJ, van Erven L.
Implantation‐related complications of
implantable cardioverter‐defibrillators
and cardiac resynchronization therapy
devices: a systematic review of
randomized clinical trials. J Am Coll
Cardiol 2011; 58: 995–1000.
62 Forleo GB, Della Rocca DG,
Papavasileiou LP, Molfetta AD, Santini
L, Romeo F. Left ventricular pacing with
a new quadripolar transvenous lead for
CRT: early results of a prospective comparison with conventional implant
outcomes. Heart Rhythm 2011; 8: 31–7.
63 Rinaldi CA, Leclercq C, Kranig W, Kacet
S, Betts T, Bordachar P et al.
Improvement in acute contractility and
hemodynamics with multipoint pacing
via a left ventricular quadripolar pacing
lead. J Interv Card Electrophysiol 2014; 40:
75–80.
64 Pappone C, Ćalović Ž, Vicedomini G,
Cuko A, McSpadden LC, Ryu K et al.
Multipoint left ventricular pacing in a
single coronary sinus branch improves
mid‐term echocardiographic and clinical
response to cardiac resynchronization
© 2016 Royal Australasian College of Physicians
Cardiac resynchronisation therapy
therapy. J Cardiovasc Electrophysiol 2015;
26: 58–63.
65 Bocchiardo M, Meyer zu Vilsendorf D,
Militello C, Lippert M, Czygan G, Gaita
F et al. Intracardiac impedance monitors
stroke volume in resynchronization
therapy patients. Europace 2010; 12:
702–7.
66 Hayes DL, Boehmer J, Day J. Cardiac
resynchronization therapy and
relationship of percent biventricular
pacing. Heart Rhythm 2011; 8: 1469–75.
67 Santini M, Gasparini M, Landolina M,
Lunati M, Proclemer A, Padeletti L et al.
Device‐detected atrial tachyarrhythmias
predict adverse outcome in real‐world
patients with implantable biventricular
defibrillators. J Am Coll Cardiol 2011; 57:
167–72.
68 Lakkireddy D, Di Biase L, Ryschon K,
Biria M, Swarup V, Reddy YM et al.
Radiofrequency ablation of premature
ventricular ectopy improves the efficacy
of cardiac resynchronization therapy in
nonresponders. J Am Coll Cardiol 2012;
60: 1531–9.
69 Cheng A, Landman SR, Stadler RW.
Reasons for loss of cardiac
resynchronization therapy pacing:
insights from 32 844 patients. Circ
Arrhythm Electrophysiol 2012; 5: 884–8.
70 Bernheim A, Ammann P, Sticherling C,
Burger P, Schaer B, Brunner‐La Rocca
© 2016 Royal Australasian College of Physicians
HP et al. Right atrial pacing impairs
cardiac function during
resynchronization therapy: acute effects
of DDD pacing compared to VDD pacing.
J Am Coll Cardiol 2005; 45: 1482–7.
71 Martin DO, Lemke B, Birnie D, Krum H,
Lee KL, Aonuma K et al. Investigation of a
novel algorithm for synchronized left‐
ventricular pacing and ambulatory optimization of cardiac resynchronization
therapy. Heart Rhythm 2012; 9: 1807–14.
72 Birnie D1, Lemke B, Aonuma K, Krum
H, Lee KL, Gasparini M et al. Clinical
outcomes with synchronized left
ventricular pacing: analysis of the
adaptive CRT trial. Heart Rhythm 2013;
10: 1368–74.
73 Jansen AH, Bracke FA, van Dantzig JM,
Meijer A, van der Voort PH, Aarnoudse
W et al. Correlation of echo‐Doppler
optimization of atrioventricular delay in
cardiac resynchronization therapy with
invasive hemodynamics in patients with
heart failure secondary to ischemic or
idiopathic dilated cardiomyopathy. Am J
Cardiol 2006; 97: 552–7.
74 Ellenbogen KA, Gold MR, Meyer TE,
Fernndez Lozano I, Mittal S, Waggoner
AD et al. Primary results from the
SmartDelay determined AV optimization:
a comparison to other AV delay methods
used in cardiac resynchronization therapy
(SMART‐AV) trial: a randomized trial
75
76
77
78
79
comparing empirical, echocardiography‐
guided, and algorithmic atrioventricular
delay programming in cardiac
resynchronization therapy. Circulation
2010; 122: 2660–8.
Bordachar P, Lafitte S, Reuter S, Sanders P,
Jaïs P, Haïssaguerre M et al. Echocardiographic parameters of ventricular
dyssynchrony validation in patients with
heart failure using sequential biventricular
pacing. J Am Coll Cardiol 2004; 44: 2157–65.
Abraham WT, León AR, St John Sutton
MG, Keteyian SJ, Fieberg AM, Chinchoy
E et al. Randomized controlled trial comparing simultaneous versus optimized
sequential interventricular stimulation
during cardiac resynchronization therapy.
Am Heart J 2012; 164: 735–41.
Mullens W, Grimm RA, Verga T, Dresing
T, Starling RC, Wilkoff BL et al. Insights
from a cardiac resynchronization optimization clinic as part of a heart failure
disease management program. J Am Coll
Cardiol 2009; 53: 765–73.
Hindricks G, Taborsky M, Glikson M,
Heinrich U, Schumacher B, Katz A et al.
Implant‐based multiparameter
telemonitoring of patients with heart failure (IN‐TIME): a randomised controlled
trial. Lancet 2014; 384: 583–90.
Akar JG. Abstract LB03‐03. Presented at:
Heart Rhythm Society Annual Scientific
Sessions; May 7–10, 2014, San Francisco.
265
CLINICAL PERSPECTIVES
Characterisation and therapeutic manipulation of the gut
microbiome in inflammatory bowel disease
J. Schulberg1 and P. De Cruz1,2
1
Department of Gastroenterology, Austin Hospital, and 2Department of Medicine, The University of Melbourne, Melbourne, Victoria, Australia
Key words
Crohn disease, ulcerative colitis, microbiota,
antibiotics, prebiotics, probiotics, faecal
microbiota transplantation.
Correspondence
Peter De Cruz, Department of Medicine, Austin
Academic Centre, The University of Melbourne,
Melbourne, Vic. 3084, Australia.
Email: [email protected]
Received 28 August 2015; accepted
4 November 2015.
doi:10.1111/imj.13003
Abstract
Inflammatory bowel diseases are thought to develop as a result of dysregulation of the
relationship that exists between the gut microbiota, host genetics and the immune system. The advent of culture-independent techniques has revolutionised the ability to
characterise the role of the gut microbiota in health and disease based on the microbiota’s genetic make-up. Inflammatory bowel diseases are characterised by dysbiosis which
is an imbalance between pro- and anti-inflammatory bacteria and a reduction in bacterial diversity. Emerging data suggest that it is not only the presence of the gut microbiota but the functional activity of the microbiota that appears to play an important role
in health and disease. Current strategies to manipulate therapeutically the gut microbiota using dietary modification, prebiotics, probiotics, antibiotics and faecal microbiota
transplantation aim to restore the balance to a state of normobiosis. However, the ability of such strategies to correct dysbiosis and thereby achieve therapeutic benefit is yet
to be fully characterised.
Introduction
Characterisation of the intestinal microbiome
From culture to metagenomics
The gut microbiota is a complex ecological environment,
with each human harbouring up to 100 trillion (1014)
bacteria.1,2 Man can be considered a ‘supra-organism’ –
a composite of human and microbial genes (the
microbiome).1 The microbiome is estimated to encode
100-fold more genes than the human genome, and is
currently being characterised as part of the Human Intestinal Tract (MetaHit) project.2 Seventy percent of the gut
microbiota has not been cultured by standard, culturebased techniques.3 Metagenomics, the study of microbiota using culture-independent techniques,4 has greatly
propelled forward our understanding of the role of the
gut microbiota in health and inflammatory bowel disease
(IBD). Metagenomics has been enabled by the presence
of the bacterial 16S ribosomal (r)RNA gene in all bacteria
Funding: P. De Cruz is supported by a Gastroenterological
Society of Australia Bushell Post Doctoral Award and University
of Melbourne David Bickart Clinician Research Award.
Conflict of interest: None.
266
which has revolutionised bacterial taxonomy through
the identification of both known and novel bacteria
based on sequence similarity to previously identified
bacteria.
Owing to the large inter-individual variation, the normal spectrum of gut microbiota composition in healthy
individuals is still unknown. There appears to be a
small number of bacterial genes shared among individuals, referred to as the ‘phylogenetic core’.5 Although
the phylogenetic core describes what is commonly ‘present’ in the gut among individuals, it does not explain
the role of these bacteria. There is increasing focus
on the elements of the microbiome that represent a
functional core, a ‘core microbiome’, at a gene expression and organismal lineage level, related to the
products and effects of the microbiota.5 With more than
1000 humans sampled from three different continents,
the metagenomics integrated gene catalogue is
now numbered at almost 10 million genes with data
demonstrating some country-specific gut microbial
signatures.2
Characterisation of the gut microbiota in IBD
Two main approaches have been used to characterise
the gut microbiota in IBD. Characterisation of the
© 2016 Royal Australasian College of Physicians
The gut microbiome in IBD
microbial community as a whole has been referred to as
the ‘global description strategy’ and led to the concept of
‘dysbiosis’ which refers to an imbalance between beneficial and harmful bacteria. The alternative approach has
focused on single specific microorganisms that are
thought to be pathogenic and this has been referred to
as the ‘candidate microorganism strategy’. Dysbiosis
in IBD is characterised by an overall reduction in microbial diversity, reduced abundance of the phylum
Firmicutes, including Faecalibacterium prausznitzii as well
as concurrent increases in Bacteroidetes.6 Differences
in the bacterial communities between Crohn disease
and ulcerative colitis (UC) suggest that the microbiome
plays a different role in the pathogenesis of these
diseases.
The search for a single specific pathogen as per the
‘candidate microorganism strategy’ in Crohn disease has
largely focused on two organisms, Mycobacterium avium
subsp. Paratuberculosis (MAP) and Escherichia coli. MAP
has been suggested due to the similarity of Johne disease
in cattle caused by MAP and Crohn disease in humans.
However, results of studies have been inconsistent.7
Escherichia coli is a commensal organism in the gut; however, adherent invasive E. coli (AIEC) is a pathogenic
strain of E. coli which has developed virulence factors
that allow it to adapt, survive and colonise intestinal
mucosa in up to a third of those with ileal Crohn disease
and also is associated with early post-operative Crohn
disease recurrence.8 Other organisms have been implicated and research assessing non-bacterial members of
the microbiome including fungi and viruses suggests that
these organisms too may have a significant role in gastrointestinal disease.9,10
There is an ongoing debate as to whether changes in
the intestinal microbiota precede or follow the development of colitis in IBD. Early work in mice has shown
altered microbiota and composition changes preceding
the onset of colitis in interleukin 10 knockout mice, thus
suggesting that dysbiosis may in fact precede the development of IBD.11
(metatranscriptomics), protein expression (metaproteonomics) and metabolites (metabolomics).12
Functional analysis: metagenomics,
metatranscriptomics, metaproteonomics
and metabolomics
Metatranscriptomics: RNA
Although phylogenetic analysis using 16S rRNA technology provides us with information on the specific
microbiota present in a given microbial community,
it does not explain the specific functional role of the
bacteria. Functional analyses of the microbiota are a
rapidly evolving area of investigation, comprising
various ‘omics’ technologies that characterise microbial
activity at the level of DNA (metagenomics); RNA
© 2016 Royal Australasian College of Physicians
Metagenomics: DNA
Metagenomic sequencing characterises the metabolic
and physiologic potential of the microbiota based on
detailed DNA information. It is able to help identify
which genes are present and therefore provides an
insight into the functional capability of the microbiome
by virtue of gene expression.13 Next-generation sequencing technologies have enabled widespread analyses
beyond the 16S rRNA gene to sequencing the wholegenome of microbial communities en masse referred to
as ‘shotgun metagenomics’. However, despite providing
extremely detailed information on the microbiome and
its DNA, the main current limitation of next generation
metagenomics is the inability to correlate directly isolated bacterial DNA with gene expression. This in part
relates to the absence of matches in the worldwide databases of the human microbiome which are still in their
infancy.14 Nonetheless, it is anticipated that with the current microbiome sequencing projects the composition of
human gut metagenomes will fast approach a saturation
point whereby the level of characterisation involved in
their ‘profiling’ will extend beyond a select number of
taxonomic marker genes, and include a much greater
depth of functional complexity, including the approaches
described below. The vast majority of recent studies of
the gut microbiota are based on 16S sequencing with
only limited studies utilising shotgun metagenomics. Initial studies comparing 16S rRNA sequencing with shotgun metagenomics in IBD patients have found that
microbial genetic expression, particularly in relation to
fundamental microbial metabolic pathways, are seen to
be far more perturbed than shifts in the microbiota at
the genus level.15 In particular in IBD, genes involved in
amino acid biosynthesis and carbohydrate metabolism
appear to be down-regulated relative to genes involved
in nutrient uptake.
Metatranscriptomics seeks to characterise the gene
expression patterns of the microbiome and is thereby
used to determine the activity of genes present. Significant qualitative compositional differences in rRNA
expression have been seen in Crohn disease patients
compared with controls, with Bacteroidetes being the most
active in contrast to Actinobacteria and Firmicutes which
were transcriptionally inactive in Crohn disease
patients.16 Although the metatranscriptome provides a
functional analysis of gene expression, it relies on mRNA
267
Schulberg & De Cruz
stability which is extremely low in prokaryotes, thus limiting its use.17
composition and metabolites contributing to microbial
community architecture.21
Metaproteonomics: proteins
Variation in microbiota throughout life
Metaproteonomics refers to the set of all expressed proteins in a cell, tissue or organism. The gut proteome is
dynamic and subject to interactions between genes,
environment and the microbiome. There is distinct fluctuation in the expression of proteins in response to
health and disease, thus making them attractive biomarkers and potential targets for molecular therapeutics.11
Most commonly, mass spectrometry is the technique
used, measuring the mass-to-charge ratio of charged particles. These proteins, peptides or metabolites are then
separated according to this ratio. Patients with ileal
Crohn disease have been found to have a different metaproteome to healthy controls, including a generalised
depletion of many proteins, alterations in bacterial carbohydrate metabolism, bacterial–host interactions, as
well as human host-secreted enzymes. However, Gramnegative bacterial outer membrane proteins have been
found to have a higher representation in the ileal microbiota of those with Crohn disease compared with healthy
controls.18 Taken together these data illustrate that the
alteration in the metaproteome may be reflective of
underlying dysbiosis. DNA sequence databases that are
being rapidly produced should serve to improve the
interrogative potential of proteomic methods, ensuring
that more of these types of analyses successfully result in
the identification of the specific gene (and microbe
[s]) encoding that particular protein.
It has been traditionally thought that from birth a newborn’s gut is sterile and that following birth there is a
rapid colonisation of microbes from the mother and surrounding environment in the newborn’s gastrointestinal
tract.22 Initially, the gut microbiota of infants are unstable and highly variable. Abrupt changes in taxonomic
groups can be linked to illness, dietary change and antibiotic therapy.23,24 A number of maternal and environmental conditions surrounding the birth, including
before, during and early after birth has additionally
resulted in differences in the composition of the gut
microbiota in children. These include the time and mode
of delivery, diet, mother’s age, body mass index, smoking status, household milieu, socio-economic status,
breastfeeding and mother’s antibiotic use.25 This early
period in the development of the gut microbiota is
thought to be crucial in determining subsequent health
or disease including IBD and other conditions such as
eczema and asthma.11 Over time the phylogenetic diversity of the infant’s microbiome increases gradually with
changes in microbial community composition increasing
along a steady gradient.23 By the age of 2 years, an
infant’s microbiota is established and thought to remain
relatively stable throughout the remainder of life.26 It is
likely that a so-called phylogenetic core exists among
healthy humans of heterogeneous populations and this
core microbiota likely has a key structural and therefore
functional role in maintaining the health of the host. In
contrast, individuals with disease may have aberrant gut
microbial patterns leading to an unhealthy state of
dysbiosis.27
Metabolomics: metabolites
Metabolomics refers to the quantification of metabolites present in cells or organisms that participate in
the metabolic reactions required for growth and maintenance of normal function. It not only includes the
metabolites of the gut microbiota but also metabolites
ingested from the external environment. Metabolome
wide association studies have recently been shown to
have potential in linking metabolic phenotypes with
disease risk19 just as genome-wide association studies
have found associations between genotype and disease
phenotypes.20 Significant interdependence of the
mucosal metabolome and microbiome has been found
in patients with Crohn disease by McHardy et al.;
however, it is not entirely clear whether the microbiota may be the source of and/or dependent upon gut
epithelial metabolites. Current evidence suggests that
the microbiome and metabolome have bi-directional
influence, with bacteria influencing metabolite
268
Host–bacterial mutualism
Man’s co-evolution with microbiota over time has led to
a finely balanced relationship which is referred to as
host–bacterial mutualism.28 The host influences or determines the make-up and activity of the microbiota and
the microbiota influences or determines aspects of host
immunity.3
Host effects on the microbiota
The genetics of the host have been shown in several
studies to have an effect on the gut microbial composition. Analysis of faecal samples from children has shown
a greater similarity in the faecal-associated microbiota of
monozygotic twins versus dizygotic twins and similarly
© 2016 Royal Australasian College of Physicians
The gut microbiome in IBD
unrelated children have been shown to have greater differences than both mono and dizygotic twin groups.29
Host genetic mutations, especially those genes involved
in the innate immune system, have been shown to influence the response to and relationship with the gut
microbiota. This is the case in IBD whereby host innate
immune genetic mutations have been associated with
dysbiosis. In a recent study of the >160 IBD susceptibility
loci, several of the genes implicated in disease pathogenesis were involved in detecting, processing and responding to the gut microbiota.30 In particular, genes including
NOD2, ATG16L1, IRGM and OCTN1/2 have been linked to
the innate immune response to host bacteria and the
development of Crohn disease. Moreover, significant
shifts in the gut microbial composition have been seen in
IBD patients who have the NOD2 composite genotype
and ATG16L1 genotype.31,32
Microbiota and its effect on host immunity
It is not only the host genotype that dictates the development of the immune system. From early after birth, the
gut microbiota also impacts on the development of the
host intestinal immune system. The signals generated by
exposure to the microbiota, influence maturation of the
gut-associated lymphoid tissues. Lymphoid follicles
in the gut mature following multiple signals from Gramnegative bacteria including that of Gram-negative
peptidoglycans.33 The microbiota has also been shown to
generate signals that promote recruitment of immunoglobulin A-secreting plasma cells and T-cells to the gut
mucosa.34 This facilitates crosstalk between dendritic
cells (DCs) and gut epithelial cells, thereby regulating the
intensity of B- and T-cell responses.35,36 In addition to
intestinal immunity, there is likely a significant modulatory effect of the intestinal microbiome on systemic
immunity. This research is still in its infancy; however,
systemic antibody production and cytokine levels have
been shown to be affected by the nature of the commensal gut microbiome.36,37
Environmental influences on gut microbiota
Geography and lifestyle
The gut microbiome has been compared in multiple
studies between people of different countries and/or different lifestyles.27,38 In European infants 6 weeks from
birth, country of birth had greater influence than both
delivery mode and feeding method on the faecal microbiota.39 The microbiome of 6-month only infants living
in rural Malawi has been compared with children of the
same age living in urban Finland. Significant differences
© 2016 Royal Australasian College of Physicians
were seen in the relative abundance and species diversity
across the two groups. In a large worldwide metagenomic study of adults by Li et al. looking for country specific differences, Chinese and Danish cohorts could be
readily distinguished from each other based on specific
signatures associated with their respective gut microbiomes.2 In data from our own group, geography and
ethnicity both appeared to play a significant role in
the composition of bacteria among people with and
without IBD.38
Diet
The influence of diet on the gut microbiota is often confounded by variation in the genetic make-up of the host
as well as other environmental exposures and geography.38 Nevertheless, there is accumulating evidence that
long-term diet is a primary driver of gut microbiota composition and function as demonstrated by Wu et al. in a
study of 98 American subjects.40,41 The microbiota of
American subjects was found to be clustered into ‘enterotypes’ distinguished primarily by levels of Bacteroides
and Prevotella. Bacteroides was associated with a typical
‘Western’ diet high in protein and animal fat whereas
Prevotella was enriched in people with high-fibre and
high-carbohydrate diets.41 As part of that study, 10 subjects were placed for 10 days on a high-fat/low-fibre or
low-fat/high-fibre controlled diet. There were detectable
changes in the microbiome composition within 24 h of
initiating the dietary modification; however, these
changes were small and overall the ‘enterotype’ identity
remained stable during the 10-day study. Long-term diet
was therefore found to have a far stronger influence on
the microbiome than short-term diet. Epidemiologic
trends suggest that the rising incidence of IBD in Asia
correlates with an increasing consumption of a ‘western’
diet in the East. However, despite the likely impact of
diet on the microbiota and IBD pathogenesis in Crohn
disease, the only dietary therapy that has been shown
definitively to induce remission has been exclusive
enteral nutrition therapy, the efficacy of which has been
largely limited to paediatric patients.
Numerous observational studies have attempted to
identify specific dietary patterns that may contribute to
the risk of IBD. There is some suggestion of increased
risk of IBD among those who consume greater quantities
of meat and fats, particularly polyunsaturated fatty acids
and omega-6 fatty acids, and lower risk among people
with diets high in fibre, vegetables and fruits. However,
clinical studies to date investigating various dietary interventions have had multiple limitations including recall
bias, lack of placebo control and insufficient molecular
microbiologic correlation.42
269
Schulberg & De Cruz
Manipulation of the gut microbiota
Just as diet has been shown to enrich and influence
the microbiome significantly, pharmacological and
other interventions have aimed to ameliorate
dysbiosis associated with intestinal and non-intestinal
diseases.
Probiotics
Probiotics involve the use of live microorganisms that
are believed to provide health benefits when consumed.
Probiotics mediate their effect through exclusion of
pathogens, maintenance of epithelial barrier function
and induction of adaptive immunity. Probiotics have
demonstrated efficacy in both induction and maintenance of remission in UC as well as maintenance of
remission in pouchitis.43,44 A number of strains has been
included in studies including Lactobacillus spp, Bifidobacteria spp, E. coli Nissle 1917 and Saccharomyces boulardii;
however, the best evidence for efficacy in UC has been
found in a combination probiotic VSL #3 which comprises eight separate organisms: three Bifidobacteria, one
Streptococcus and four Lactobacilli.44,45 Meta-analyses have
also demonstrated that probiotics are effective for the
prevention of Clostridium difficile infection.46 However,
there is no proven benefit for the use of probiotics in
patients with Crohn disease.
Prebiotics
Prebiotics are generally regarded as non-digestible food
ingredients that are fermented by intestinal bacteria in a
selective manner which promote changes in the gut ecosystem that benefit the host. However, it should be noted
that the literature contains no formal consensus regarding the specific definition of prebiotics.47 Prebiotics
include oligofructose and inulin which are polymers of
fructose found naturally and have been shown to
increase commensal anti-inflammatory faecal and mucosal Bifidobacteria and Faecalibacterium prausnitzii in healthy
humans.48 Prebiotic fermentation results in the production of short-chain fatty acids such as butyrate which
appears to enhance epithelial integrity and promote regulatory DC function in vitro.49 While the use of prebiotics such as fructo-oligosaccharides in animal models of
colitis and in healthy human subjects has shown multiple benefits, there have been conflicting results in two
recent randomised placebo-controlled trials of patients
with Crohn disease.50,51
Antibiotics
The use of antibiotics has a substantial impact on the
microbiome. During short-term antibiotic use, gut bacteria respond early with a significant reduction in
270
susceptible organisms, an overall reduction in diversity
and an increased likelihood of colonisation by ‘presumptive’ naturally resistant bacteria.52 The microbiome
responds by activating systems to avoid the antimicrobial
effects of the drugs. Major metabolic changes occur rapidly which reduce capacity to transport and metabolise
bile acid, cholesterol, hormones and vitamins. However,
following completion of antibiotic therapy, microbial
interactions improve significantly. The long-term data on
the effect of a course of antibiotics on the human microbiome are limited; however, recent animal studies have
shown that antibiotic exposure early in life induces lasting effects on the microbiome, which may alter body
composition and play a role in the development of obesity.53 In Crohn disease, modification of the microbiota
with long-term antibiotics has proven efficacy in treatment of perianal Crohn disease as well as prevention of
post-operative recurrence.43 In contrast it has been
shown that prior exposure to antibiotics early in childhood predisposes to the risk of developing Crohn disease
but not UC later in life.54 The evidence for antibiotic use
in UC is relatively lacking except for in patients with
acute severe colitis or induction of remission of
pouchitis.43
Faecal microbiota transplantation
Faecal microbiota transplantation (FMT) is highly efficacious in patients with refractory C. difficile infection.55,56
There have been two recent randomised controlled trials
which have evaluated FMT in UC with conflicting
results. In the study by Moayyedi et al. there was only a
modest benefit of weekly FMT over a period of 6 weeks
for inducing remission compared with placebo (24 vs
5%); however, most of the benefit was seen in those
patients who received FMT from one individual donor
perhaps suggesting that microbial characteristics
from the donor was a significant factor in its efficacy.57
In contrast, Rossen et al. were not able to show a significant difference in outcome between FMT recipients and
placebo-treated patients; only 41% of patients overall in
the per protocol analysis who received donor faeces
achieved clinical and endoscopic remission, suggesting
that the two infusion protocols used in the study were
insufficient to alter the microbiota in the majority of
patients. Despite FMT’s relative lack of efficacy in UC,
Rossen et al. found that at baseline UC patients had a significantly lower abundance of members of Clostridium
clusters IV, XIVa and XVIII and higher abundance of Bacteroidetes, Bacilli, Proteobacteria and Clostridium clusters IX
and XI compared with healthy donors. In those patients
with UC who responded to the FMT infusions, the
microbiota similarity shifted from the UC patient genotype (similar to UC patients who were non-FMT
© 2016 Royal Australasian College of Physicians
The gut microbiome in IBD
responders) at baseline to a microbiome which was similar to the healthy donors, suggesting that FMT may have
the potential to correct dysbiosis in a proportion of
patients with UC.58 There have been no randomised
trials of FMT to date in Crohn disease.
Conclusion
The advent of new high-throughput DNA sequencing
technologies as well as functional analysis has revolutionised the characterisation of the microbiome in health
References
1 Turnbaugh PJ, Ley RE, Hamady M,
Fraser-Liggett CM, Knight R, Gordon JI.
The human microbiome project. Nature
2007; 449: 804–10.
2 Li J, Jia H, Cai X, Zhong H, Feng Q,
Sunagawa S et al. An integrated catalog
of reference genes in the human gut
microbiome. Nat Biotechnol 2014; 32:
834–41.
3 Sokol H, Seksik P. The intestinal
microbiota in inflammatory bowel
diseases: time to connect with the host.
Curr Opin Gastroenterol 2010; 26:
327–31.
4 Ley RE, Lozupone CA, Hamady M,
Knight R, Gordon JI. Worlds within
worlds: evolution of the vertebrate gut
microbiota. Nat Rev Microbiol 2008; 6:
776–88.
5 Qin J, Li R, Raes J, Arumugam M,
Burgdorf KS, Manichanh C et al. A
human gut microbial gene catalogue
established by metagenomic
sequencing. Nature 2010; 464: 59–65.
6 Walker AW, Sanderson JD, Churcher C,
Parkes GC, Hudspith BN, Rayment N
et al. High-throughput clone library
analysis of the mucosa-associated
microbiota reveals dysbiosis and
differences between inflamed and noninflamed regions of the intestine in
inflammatory bowel disease. BMC
Microbiol 2011; 11: 7.
7 Abubakar I, Myhill D, Aliyu SH,
Hunter PR. Detection of Mycobacterium
avium subspecies paratuberculosis from
patients with Crohn’s disease using
nucleic acid-based techniques: a
systematic review and meta-analysis.
Inflamm Bowel Dis 2008; 14: 401–10.
8 Darfeuille-Michaud A, Boudeau J,
Bulois P, Neut C, Glasser AL, Barnich N
et al. High prevalence of adherent-
© 2016 Royal Australasian College of Physicians
9
10
11
12
13
14
15
16
and IBD. The microbiome and its relationship with the
host play a central role in the pathogenesis of both
Crohn disease and UC. The huge datasets that have
emerged over the past few years will continue to grow
and be refined to further help us evaluate the interaction
between the host, diet, environment and microbiota.
Therapeutic manipulation of the microbiome with
bacteria-specific targeted therapies will now continue to
be a focus of future therapeutic trials in IBD, which were
previously dominated by strategies targeting the immune
system alone.
invasive Escherichia coli associated with
ileal mucosa in Crohn’s disease.
Gastroenterology 2004; 127: 412–21.
Norman JM, Handley SA, Virgin HW.
Kingdom-agnostic metagenomics and
the importance of complete
characterization of enteric microbial
communities. Gastroenterology 2014; 146:
1459–69.
Ott SJ, Kuhbacher T, Musfeldt M,
Rosenstiel P, Hellmig S, Rehman A et al.
Fungi and inflammatory bowel diseases:
alterations of composition and diversity.
Scand J Gastroenterol 2008; 43: 831–41.
De Cruz P, Prideaux L, Wagner J,
Ng SC, McSweeney C, Kirkwood C et al.
Characterization of the gastrointestinal
microbiota in health and inflammatory
bowel disease. Inflamm Bowel Dis 2012;
18: 372–90.
Li M, Wang B, Zhang M,
Rantalainen M, Wang S, Zhou H et al.
Symbiotic gut microbes modulate
human metabolic phenotypes. Proc Natl
Acad Sci USA 2008; 105: 2117–22.
Wang WL, Xu SY, Ren ZG, Tao L,
Jiang JW, Zheng SS. Application of
metagenomics in the human gut
microbiome. World J Gastroenterol 2015;
21: 803–14.
Qin J, Li R, Raes J, Arumugam M,
Burgdorf KS, Manichanh C et al. A
human gut microbial gene catalogue
established by metagenomic
sequencing. Nature 2010; 464: 59–65.
Morgan XC, Tickle TL, Sokol H,
Gevers D, Devaney KL, Ward DV et al.
Dysfunction of the intestinal
microbiome in inflammatory bowel
disease and treatment. Genome Biol
2012; 13: R79.
Rehman A, Lepage P, Nolte A,
Hellmig S, Schreiber S, Ott SJ.
Transcriptional activity of the dominant
gut mucosal microbiota in chronic
17
18
19
20
21
22
23
24
inflammatory bowel disease patients.
J Med Microbiol 2010; 59:
1114–22.
Redon E, Loubiere P, CocaignBousquet M. Role of mRNA stability
during genome-wide adaptation of
Lactococcus lactis to carbon starvation. J
Biol Chem 2005; 280: 36380–5.
Erickson AR, Cantarel BL,
Lamendella R, Darzi Y, Mongodin EF,
Pan C et al. Integrated metagenomics/
metaproteomics reveals human hostmicrobiota signatures of Crohn’s
disease. PLoS One 2012; 7: e49138.
Bictash M, Ebbels TM, Chan Q, Loo RL,
Yap IK, Brown IJ et al. Opening up the
"Black Box": metabolic phenotyping
and metabolome-wide association
studies in epidemiology. J Clin Epidemiol
2010; 63: 970–9.
Wellcome Trust Case Control
Consortium. Genome-wide association
study of 14,000 cases of seven common
diseases and 3,000 shared controls.
Nature 2007; 447: 661–78.
McHardy IH, Goudarzi M, Tong M,
Ruegger PM, Schwager E, Weger JR
et al. Integrative analysis of the
microbiome and metabolome of the
human intestinal mucosal surface
reveals exquisite inter-relationships.
Microbiome 2013; 1: 17.
Mackie RI, Sghir A, Gaskins HR.
Developmental microbial ecology of the
neonatal gastrointestinal tract. Am J Clin
Nutr 1999; 69: 1035S–45.
Koenig JE, Spor A, Scalfone N,
Fricker AD, Stombaugh J, Knight R
et al. Succession of microbial consortia
in the developing infant gut
microbiome. Proc Natl Acad Sci USA
2011; 108: 4578–85.
Savino F, Roana J, Mandras N,
Tarasco V, Locatelli E, Tullio V. Faecal
microbiota in breast-fed infants after
271
Schulberg & De Cruz
25
26
27
28
29
30
31
32
33
34
35
antibiotic therapy. Acta Paediatr 2011;
100: 75–8.
Munyaka PM, Khafipour E, Ghia JE.
External influence of early childhood
establishment of gut microbiota and
subsequent health implications. Front
Pediatr 2014; 2: 109.
Zoetendal EG, Akkermans AD, De
Vos WM. Temperature gradient gel
electrophoresis analysis of 16S rRNA
from human fecal samples reveals
stable and host-specific communities of
active bacteria. Appl Environ Microbiol
1998; 64: 3854–9.
Zhang J, Guo Z, Xue Z, Sun Z, Zhang M,
Wang L et al. A phylo-functional core of
gut microbiota in healthy young
Chinese cohorts across lifestyles,
geography and ethnicities. ISME J 2015;
9: 1979–90.
Backhed F, Ley RE, Sonnenburg JL,
Peterson DA, Gordon JI. Host-bacterial
mutualism in the human intestine.
Science 2005; 307: 1915–20.
Stewart JA, Chadwick VS, Murray A.
Investigations into the influence of host
genetics on the predominant eubacteria
in the faecal microflora of children. J
Med Microbiol 2005; 54: 1239–42.
Van Limbergen J, Radford-Smith G,
Satsangi J. Advances in IBD genetics.
Nat Rev Gastroenterol Hepatol 2014; 11:
372–85.
Frank DN, Robertson CE, Hamm CM,
Kpadeh Z, Zhang T, Chen H et al.
Disease phenotype and genotype are
associated with shifts in intestinalassociated microbiota in inflammatory
bowel diseases. Inflamm Bowel Dis 2011;
17: 179–84.
Knights D, Silverberg MS, Weersma RK,
Gevers D, Dijkstra G, Huang H et al.
Complex host genetics influence the
microbiome in inflammatory bowel
disease. Genome Med 2014; 6: 107.
Bouskra D, Brezillon C, Berard M,
Werts C, Varona R, Boneca IG et al.
Lymphoid tissue genesis induced by
commensals through NOD1 regulates
intestinal homeostasis. Nature 2008;
456: 507–10.
Hooper LV, Macpherson AJ. Immune
adaptations that maintain homeostasis
with the intestinal microbiota. Nat Rev
Immunol 2010; 10: 159–69.
Duerkop BA, Vaishnava S, Hooper LV.
Immune responses to the microbiota at
272
36
37
38
39
40
41
42
43
44
45
46
the intestinal mucosal surface. Immunity
2009; 31: 368–76.
Cerf-Bensussan N, GaboriauRouthiau V. The immune system and
the gut microbiota: friends or foes? Nat
Rev Immunol 2010; 10: 735–44.
Schachtschneider KM, Yeoman CJ,
Isaacson RE, White BA, Schook LB,
Pieters M. Modulation of systemic
immune responses through commensal
gastrointestinal microbiota. PLoS One
2013; 8: e53969.
Prideaux L, Kang S, Wagner J,
Buckley M, Mahar JE, De Cruz P et al.
Impact of ethnicity, geography, and
disease on the microbiota in health and
inflammatory bowel disease. Inflamm
Bowel Dis 2013; 19: 2906–18.
Fallani M, Young D, Scott J, Norin E,
Amarri S, Adam R et al. Intestinal
microbiota of 6-week-old infants across
Europe: geographic influence beyond
delivery mode, breast-feeding, and
antibiotics. J Pediatr Gastroenterol Nutr
2010; 51: 77–84.
Xu Z, Knight R. Dietary effects on
human gut microbiome diversity. Br J
Nutr 2015; 113: S1–5.
Wu GD, Chen J, Hoffmann C,
Bittinger K, Chen YY, Keilbaugh SA
et al. Linking long-term dietary patterns
with gut microbial enterotypes. Science
2011; 334: 105–8.
Lee D, Albenberg L, Compher C,
Baldassano R, Piccoli D, Lewis JD et al.
Diet in the pathogenesis and treatment
of inflammatory bowel diseases.
Gastroenterology 2015; 148: 1087–106.
Sartor RB. Therapeutic manipulation of
the enteric microflora in inflammatory
bowel diseases: antibiotics, probiotics,
and prebiotics. Gastroenterology 2004;
126: 1620–33.
Mimura T, Rizzello F, Helwig U,
Poggioli G, Schreiber S, Talbot IC et al.
Once daily high dose probiotic therapy
(VSL#3) for maintaining remission in
recurrent or refractory pouchitis. Gut
2004; 53: 108–14.
Sood A, Midha V, Makharia GK,
Ahuja V, Singal D, Goswami P et al. The
probiotic preparation, VSL#3 induces
remission in patients with mild-tomoderately active ulcerative colitis. Clin
Gastroenterol Hepatol 2009; 7: 1202–9,
1209.e1.
Goldenberg JZ, Ma SS, Saxton JD,
Martzen MR, Vandvik PO,
47
48
49
50
51
52
53
54
55
Thorlund K et al. Probiotics for the
prevention of Clostridium difficileassociated diarrhea in adults and
children. Cochrane Database Syst Rev
2013; 5: CD006095.
Hutkins RW, Krumbeck JA, Bindels LB,
Cani PD, Fahey G Jr, Goh YJ et al.
Prebiotics: why definitions matter. Curr
Opin Biotechnol 2015; 37: 1–7.
Gibson GR, Beatty ER, Wang X,
Cummings JH. Selective stimulation of
bifidobacteria in the human colon by
oligofructose and inulin. Gastroenterology
1995; 108: 975–82.
Millard AL, Mertes PM, Ittelet D,
Villard F, Jeannesson P, Bernard J.
Butyrate affects differentiation,
maturation and function of human
monocyte-derived dendritic cells and
macrophages. Clin Exp Immunol 2002;
130: 245–55.
Joossens M, De Preter V, Ballet V,
Verbeke K, Rutgeerts P, Vermeire S.
Effect of oligofructose-enriched inulin
(OF-IN) on bacterial composition and
disease activity of patients with Crohn’s
disease: results from a double-blinded
randomised controlled trial. Gut 2012;
61: 958.
Benjamin JL, Hedin CR,
Koutsoumpas A, Ng SC, McCarthy NE,
Hart AL et al. Randomised, doubleblind, placebo-controlled trial of fructooligosaccharides in active Crohn’s
disease. Gut 2011; 60: 923–9.
Perez-Cobas AE, Gosalbes MJ,
Friedrichs A, Knecht H, Artacho A,
Eismann K et al. Gut microbiota
disturbance during antibiotic therapy: a
multi-omic approach. Gut 2013; 62:
1591–601.
Cox LM, Yamanishi S, Sohn J,
Alekseyenko AV, Leung JM, Cho I et al.
Altering the intestinal microbiota
during a critical developmental window
has lasting metabolic consequences. Cell
2014; 158: 705–21.
Ungaro R, Bernstein CN, Gearry R,
Hviid A, Kolho KL, Kronman MP et al.
Antibiotics associated with increased
risk of new-onset Crohn’s disease but
not ulcerative colitis: a meta-analysis.
Am J Gastroenterol 2014; 109: 1728–38.
van Nood E, Vrieze A, Nieuwdorp M,
Fuentes S, Zoetendal EG, de Vos WM
et al. Duodenal infusion of donor feces
for recurrent Clostridium difficile.
N Engl J Med 2013; 368: 407–15.
© 2016 Royal Australasian College of Physicians
56 Kelly CR, Kahn S, Kashyap P, Laine L,
Rubin D, Atreja A et al. Update on fecal
microbiota transplantation 2015:
indications, methodologies,
mechanisms, and outlook.
Gastroenterology 2015; 149: 223–37.
57 Moayyedi P, Surette MG, Kim PT,
Libertucci J, Wolfe M, Onischi C et al.
Fecal microbiota transplantation
induces remission in patients with
active ulcerative colitis in a randomized
controlled trial. Gastroenterology 2015;
149: 102–9.e6.
58 Rossen NG, Fuentes S, van der
Spek MJ, Tijssen J, Hartman JH,
Duflou A et al. Findings from a
randomized controlled trial of fecal
transplantation for patients with
ulcerative colitis. Gastroenterology
2015; 149: 110–8.e4.
O R I G I N A L A RT I C L E
Regional challenges: evaluation of a hepatitis outreach
programme using transient elastography (FibroScan) in Victoria
M. C. Thurnheer,1 T. R. Schulz,1,2 T. Nguyen,1 J. MacLachlan3 and J. Sasadeusz1
1
Victorian Infectious Diseases Service, 2Department of Medicine, and 3Victorian Infectious Diseases Reference Laboratory, The Royal Melbourne
Hospital, Doherty Institute, Melbourne, Victoria, Australia
Key words
hepatitis C, hepatitis B, outreach programme,
transient elastography, liver fibrosis, rural.
Correspondence
Joseph Sasadeusz, Victorian Infectious Diseases
Service, The Royal Melbourne Hospital, Doherty
Institute, 792 Elizabeth Street, Melbourne, Vic.
3000, Australia.
Email: [email protected]
Received 5 August 2015; accepted
18 November 2015.
doi:10.1111/imj.12963
Abstract
Background: Evaluation of an outreach programme using a mobile transient elastography (TE) device (FibroScan) to improve liver disease assessment in different clinical
settings.
Aims: To evaluate a programme of liver fibrosis assessment by TE and to compare
fibrosis scores between different sites and patient groups.
Methods: Prospective cohort study. TE was conducted at a tertiary hospital and during
outreach clinics in three different settings: community clinics, clinics for people who
use drugs (PWUD) and regional clinics in rural Victoria. All patients referred for TE at
the participating locations were eligible during the study period.
Results: A total of 200 of 623 patients was assessed and evaluated during outreach sessions (regional 100; PWUD 18; community 82). While the majority of patients in community centres were infected with hepatitis B (68%), most patients in regional clinics
and in PWUD settings had hepatitis C virus (HCV) (81 and 100%, respectively). Significantly more patients assessed at regional clinics and PWUD settings presented with
severe fibrosis (F3-F4, F4): regional clinics 39%; PWUD 31%; tertiary 11%; community
7%, (P <0.001). Multivariable logistic regression analysis revealed that older age, alcohol
consumption, male sex, increased alanine transferase levels, HCV infection and importantly, evaluation at regional sites were independently associated with severe fibrosis.
Conclusions: A TE-based outreach programme allows for assessment of liver fibrosis
in varied and regional populations. The finding that patients in regional settings and
PWUD presented with more advanced fibrosis should prompt improvements in healthcare to improve access for these populations.
Introduction
Funding: This study was supported by a grant from the Royal
Melbourne Hospital and The University of Melbourne Centre
for Excellence in Rural Sexual Health (CERSH). The grant
included purchase of the mobile FibroScan device used in the
study.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
Despite concerted efforts to increase screening and notifications for viral hepatitis in Australia, it is estimated
that around 45% of 218 000 patients living with chronic
hepatitis B virus (HBV)1 are unaware of their disease
and 40 000–50 000 of 230 000 people infected with
chronic hepatitis C virus (HCV), remain undiagnosed.1,2
273
Thurnheer et al.
Follow-up and management of patients with chronic
hepatitis is often insufficient. It is estimated that in 2013
only 10% of the patients eligible for HBV treatment
received antiviral medication and a total of 2550 patients
with chronic HCV was treated.1,2
In Australia, the majority of individuals with chronic
HBV are either migrants, originating from high prevalence countries or Aboriginal and Torres Strait Islanders.3
The majority of people with HCV have a current or prior
history of intravenous drug use (IDU).4 For these vulnerable populations, access to medical care can be limited
due to current or previously experienced stigmatisation,
cultural differences, language difficulties and low health
literacy.5–8
No studies have specifically addressed the provision of
hepatitis care for people living in rural, regional or
remote regions of Victoria, although several authors
have emphasised the need for improved hepatitis care in
regional and remote settings in Australia.9–11
In recent years, transient elastography (TE) has
become the preferred method for non-invasive fibrosis
screening allowing for prompt, point of care
assessment.12 Outreach programmes using a mobile
FibroScan are successful, can be integrated in harmreduction strategies and engage patients who might otherwise not be able to commit to medical care.13,14 In
2012, a physician-initiated project was launched in Victoria, with the aim to provide onsite fibrosis assessment
and disease evaluation for patients with chronic liver disease (primarily viral hepatitis) attending tertiary and
non-tertiary healthcare facilities including clinics located
in rural and regional Victoria. The aim of this study was
to describe the specific characteristics of the populations
reached through the programme, to identify risk factors
associated with worse outcomes and to define specific
needs for further shaping the programme.
Methods
This study was approved via the Melbourne Health
human research and ethics committee as a quality assurance project with number QA2015105. Commencing in
2012, a mobile device for TE was used (FibroScan402,
Echosens, Paris, France) at hepatitis clinics held at a tertiary hospital, a community clinic, a health centre servicing
people who use drugs (PWUD) and at four regional clinic
locations.
Using a standardised case report form, demographic
(age, sex, region of origin), epidemiological (cause of
liver disease, history of drug use, self-reported increased
alcohol intake (not quantified)) and clinical information
(height, weight, laboratory parameters including alanine
transferase (ALT), results of TE) were collected during
274
the clinical sessions. The database was censored on
31 May 2014, and patients with a complete set of data
were included in the current analysis.
Results for median stiffness were used to categorise
the level of fibrosis based on a published scoring system.15,16 For patients with underlying chronic hepatitis
C: median stiffness 0–7 kPa (F0–1), median stiffness
>7–9.5 kPa (F1–F2), median stiffness >9.5–14.5 kPa
(F3–F4), median stiffness >14.5 (F4). For patients with
chronic hepatitis B: median stiffness 0–7.3 (F0–F1),
median stiffness 7.3–11 (F2–F3), median stiffness 11–18
(F3–F4), median stiffness >18 (F4). ALT levels were categorised as more than twice the upper limit of normal
(UN) (>112 U/L) and less than 2× UN (<112 U/L).
To estimate the impact of age on the degree of liver
fibrosis, age groups were defined based on age quartiles
calculated for the entire population: age group 1:
<33.7 years; age group 2: 33.7– <44.47 years; age group
3: 44.7– < 53.94 years; age group 4: 53.94–84 years.
Data analysis was performed using STATA SE, v12. Due to
non-normal distribution and the presence of outliers,
continuous numerical variables (age, body mass index,
median stiffness) were compared using either two-tailed
Wilcoxon rank tests or Kruskal–Wallis tests. Depending
on the sample size, categorical data were compared using
either Fisher’s exact t test or Chi-squared test.
Results
Distinct populations are reached at different
clinical sites
During the study period, a total of 680 patients was
assessed for liver fibrosis using TE. Complete data collection was available for 623 patients and these were
included in the analysis (Table 1).
The majority of patients (423 (68%)) were seen at a
tertiary hospital; 254 (60%) of these were male and the
median age was 45 years. Most were of either Asian or
Australian origin (121 (28.6%) and 180 (43%), respectively) and were referred for evaluation of chronic hepatitis B (214, 51%) or for assessment of chronic hepatitis
C 175 (41%).
Eighty-two patients were evaluated in urban community clinics, most of them were male (59 (72%)) and
slightly younger (median age of 38 years). They were
mostly of either Asian or African origin (45 (55%) and
20 (24%), respectively). The majority of these patients
were referred for assessment of chronic hepatitis B infection (56 (68%)) (Table 1).
Of the 18 patients assessed at a health centre for
PWUD, 11 (61.1%) were male, the median age was
39 years. Half of these patients were Australians and one
© 2016 Royal Australasian College of Physicians
© 2016 Royal Australasian College of Physicians
P50
44
24
214 (50.6%)
175 (41.4%)
8 (1.9%)
21 (5.0%)
0
5 (.2%)
106 (25.06%)
39 (9.2%)
279 (44.8%)
285 (45.8%)
13 (2.1%)
27 (4.3%)
3 (0.5%)
16 (2.6%)
169 (27.1%)
81 (13.0%)
P50
45
24
180 (42.5%)
121 (28.6%)
28 (6.6%)
27 (6.4%)
2 (0.5%)
65 (15.4%)
Max
84
52
423
254 (60.1%)
P25 P75 Min
34
55
18
22
28
16
Max
84
52
Tertiary hospital n (%)
238 (38.2%)
207 (33.2%)
51 (8.2%)
33 (5.3%)
3 (0.5%)
91 (14.6%)
623
390 (62.6%)
P25 P75 Min
34
53
16
21.6 27.5 16
Total n (%)
P50
38
23
9 (11.0%)
6 (7.3%)
56 (68.3%)
16 (19.5%)
3 (3.7%)
1 (1.2%)
0
6 (7.3%)
45 (54.9%)
8 (9.8%)
20 (24.4%)
2 (2.4%)
0
7 (8.5%)
82
59 (72.0%)
P25 P75 Min
29
46
16
22
27
16
Community n (%)
Max
71
32
P50
39
25
15 (83.3%)
5 (27.8%)
0
13 (71.2%)
0
5 (27.8%)
0
0
6 (33.3%)
9 (50.0%)
1 (5.6%)
2 (11.1%)
0
0
18
11 (61.1%)
P25 P75 Min
32
53
25
20
28
17
PWUD n (%)
Max
60
38
P50
47
25
39 (39.0%)
31 (31.0%)
9 (9.0%)
81 (81.0%)
2 (2.0%)
0
3 (3.0%)
5 (5.0%)
7 (7.0%)
69 (69.0%)
2 (2.0%)
2 (.0%)
1 (1.0%)
19 (19.0%)
100
66 (66.0%)
P25
P75 Min
40
55
18
22
28 16.5
Regional clinic n (%)
Max
72
37
<0.001
<0.001
<0.001
<0.001
0.692
<0.001
0.001
0.005
<0.001
<0.001
<0.001
0.127
0.790
0.065
0.0001‡
0.259‡
0.191
P†
†Comparing teritary hospital, community, PWUD and remote clinics. ‡Kruskal–Wallis test. BMI, body mass index; HBV, hepatitis B virus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IDU,
intravenous drug use; PWUD, people who use drugs.
Age (years)
BMI
Region of origin
Asia
Australia
Africa
Europe
America
Unknown
Cause of hepatitis
HBV
HCV
HBV+ HCV
HIV (+HCV/+HBV)
Alcohol
Other
Risk behaviour
IDU ever
Alcohol
Number of patients
Sex male
Table 1 Population characteristics
Outreach transient elastography
275
Thurnheer et al.
clinics (median 6.2 and 6 kPa, respectively, P < 0.001,
Fig. 1A).
Levels of median stiffness were lower in patients
with chronic hepatitis B compared with patients not
infected with HBV (median 5.3 vs 6.3 kPa, P < 0.001).
No significant differences were observed when comparing HBV-infected patients assessed at different clinic
locations (Kruskal–Wallis analysis of variance,
P = 0.696) (Fig. 1B).
Levels of liver stiffness were higher in HCV-infected
individuals, compared with patients not infected with
HCV (median 6.4 vs median 5.3 kPa, P < 0.001) with
highest values found in patients evaluated at regional
clinics (8.6 vs 6.3 kPa, Mann–Whitney test P < 0.001)
(Fig. 1C).
For the 564 patients infected with either hepatitis B or
hepatitis C (HBV n = 279 and HCV n = 285) disease-specific levels of fibrosis were calculated based on TE results
(Fig. 2). Of all patients included in this analysis,
88 (15.6%) were diagnosed with severe fibrosis (F3–F4,
F4) and 41 (7.3%) of these were cirrhotic. This proportion was markedly higher in the population evaluated at
regional clinics: in 35 patients (38.9%) TE results were
compatible with fibrosis levels of F3–F4, and
19 (21.11%) were cirrhotic (Fig. 2A).
third originated from Asia (9 (50%) and 6 (33%),
respectively. Hepatitis C was the predominant cause for
referral (13 (71%)); of note five (28%) were co-infected
with human immunodeficiency virus (Table 1).
Of the 100 patients screened at regional clinics,
66 were male. They were older than patients in the
other three settings (median age 47 years). The majority
of these patients were of Australian origin (69 patients,
including 15 individuals who identified as indigenous)
and chronic hepatitis C was the predominant cause for
liver disease (81 patients).
Compared with the two other non-PWUD-locations,
more patients reported a history of IDU (39) and
increased alcohol consumption (31) (Table 1).
Higher median liver stiffness in patients with
chronic hepatitis C and in patients evaluated
at regional clinics
Results of TE showed a wide range (2.5–72 kPa) with a
median of 5.7 kPa for the entire population (Fig. 1A).
Patients seen at the tertiary hospital and in community
clinics had significantly lower levels of liver stiffness
(median 5.6 and 5.3 kPa, respectively) compared with
patients assessed at the PWUD centre or at regional
(A)
(B)
(C)
All etiologies
80
HBV
HCV
*
60
p<0.0001 (Kruskal Wallis)
*
**
p=0.696 (Kruskal Wallis)
p=0.0008 (Kruskal Wallis)
Fibroscan median stiffness (kPa)
40
20
10
9
8
7
6
5
4
3
5.7
All
5.6
5.3
Tertiary Community
6.2
6.8
5.3
PWUD
Regional
All
5.2
5.3
4.9
Tertiary Community Regional
6.4
All
6.3
5.5
Tertiary Community
6.6
8.6
PWUD
Regional
Figure 1 Transient elastography results (median stiffness in kilopascal) by location and cause of hepatitis. (A) Median stiffness for combined causes of
underlying liver disease. All patients: tertiary patients evaluated at a tertiary hospital; community patients assessed at community clinics; PWUD
patients examined at PWUD healthcare centre; regional patients examined at regional clinic location. (B) Patients mono-infected with hepatitis B.
(C) Patients mono-infected with hepatitis C. * indicates median stiffness HBV (median 5.3 kPa) versus non-HBV (median 6.4 kPa): P < 0.0001 (Mann–
Whitney test), median stiffness HCV (median: 6.4 kPa) versus non-HCV (median: 5.3 kPa): P < 0.0001 (Mann–Whitney test), median stiffness HCV
(median: 6.4 kPa) versus HBV (median: 5.3 kPa): P < 0.0001 (Mann–Whitney test). ** indicates median stiffness HCV remote clinic (median 8.6 kPa)
versus HCV non-remote clinic (median 6.3 kPa): P = 0.0002 (Mann–Whitney test).
276
© 2016 Royal Australasian College of Physicians
Outreach transient elastography
Proportion of total per location
(A)
Level of Fibrosis: all patients, adjusted for underlying disease
100
n=564
n=389
n=72
n=13
n=90
All
Tertiary
Community
PWUD
Regional Clinic
50
0
Proportion of total per location
(B)
Level of Fibrosis: HBV infection
100
n=279
n=214
n=56
All
Tertiary
Community
50
0
(C)
Proportion of total per location
n=9
PWUD
Regional Clinic
Level of Fibrosis: HCV Infection
100
n=285
n=175
n=16
n=13
n=81
All
Tertiary
Community
PWUD
Regional Clinic
50
0
Figure 2 Levels of liver fibrosis based on results of transient elastography. (A) Levels of liver fibrosis in combined assessment of all causes of
hepatitis by clinic location. ( ), F0; ( ), F1-2; ( ), F3-F4; ( ), F4.
(B) Levels of liver fibrosis in hepatitis B infected patients by clinic location. ( ), F0; ( ), F1-2; ( ), F3-F4; ( ), F4. (C) Levels of liver fibrosis
in hepatitis C infected patients by clinic location. ( ), F0; ( ), F1-2;
( ), F3-F4; ( ), F4.
Advanced fibrosis was rare among the patients
infected with hepatitis B. Only a minority of 14 (5%)
patients had TE results compatible with severe fibrosis;
overt cirrhosis was found in four individuals (2.2%)
(Fig. 2B); most patients had TE results compatible with
no or only minimal liver fibrosis (F0–1: 241 individual (86.4%)).
In contrast, a significant proportion of 26% (74 individuals) of all patients infected with HCV showed signs
of advanced fibrosis (F3–4, F4) (Fig. 2B), and
35 (12.3%) had TE results compatible with cirrhosis
(Fig. 3C). This finding was most pronounced at regional
clinic locations: 35 (43.2%) showed advanced liver
© 2016 Royal Australasian College of Physicians
Figure 3 Degree of liver fibrosis progresses with age but is significantly
higher in patients infected with hepatitis C when compared with HBVinfected patients of the same age group. (A) Age correlation of median
liver stiffness (R2: 0.03999, P < 0.001). (B) Age-stratified comparison of
median stiffness between HBV- and HCV-infected patients. *Age groups
based on age quartiles for entire population: Age group 1: age
<33.7 years; age group 2: age 33.7 years–age <44.7 years; age group
3: age 44.7 years–age <53.9 years; age group 4: age >53.9 years.
fibrosis, and 19 of these (23.5%) had TE results compatible with cirrhosis.
Distinct populations are infected with
hepatitis B or hepatitis C
In our study, people infected with chronic hepatitis C
were significantly older compared with patients infected
with chronic hepatitis B (median age 49 vs 39 years),
results for median stiffness were higher (6.4 vs 5.3 kPa),
alcohol consumption was more frequent (23.5 vs 2.5%)
277
Thurnheer et al.
Table 2 Population characteristics, HBV versus HCV
Total
Sex male
Age (years)
BMI (kg/m2)
ALT† (U/L)
Median stiffness
Fibrosis
Region of origin
Treatment location
ATSI
IDU ever
ALT >2UN†
Alcohol
P50
39
24
30
5.3
F0-1
F2-3
F3-4
F4
Asia
Australia
Africa
Europe
Americas
Unknown
Tertiary
Community
PWUD
Remote
P25
31
22
20
4.2
HBV
HCV
279
160 (57.4%)
P75
50
27
45
6.2
241 (86.4%)
24 (8.6%)
8 (2.9%)
6 (2.2%)
190 (68.1%)
15 (5.4%)
39 (14%)
9 (3.2%)
0
26 (9.3%)
214 (76.7%)
56 (12.7%)
0
9 (3.23%)
0
2 (0.7%)
11 (3.94%)
7 (2.5%)
285
183 (64.2%)
P75
56
28
113
9.8
162 (56.8%)
49 (17.2%)
39 (13.7%)
35 (12.3%)
33 (11.6%)
172 (60.4%)
6 (2.1%)
21 (7.4%)
2 (0.7%)
51 (17.9%)
175 (61.4%)
16 (5.6%)
13 (4.6%)
81 (28.4%)
19 (6.67%)
152 (53.3%)
60 (21.5%)
67 (23.5%)
Min
16
16
7
2.5
Max
82
52
261
33
P50
49
25
59
6.4
P25
39
22
39
5.1
P
0.095
Min
22
16
10
2.5
Max
84
38
510
72
<0.001
0.064
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.028
0.161
0.003
<0.001
<0.001
<0.001
<0.001
<0.001
†No information available for 49/285 HCV-infected patients and 28/279 HBV-infected patients. ALT, alanine transferase; BMI, body mass index; HBV,
hepatitis B virus; HCV, hepatitis C virus; IDU, intravenous drug use; PWUD, people who use drugs.
and a large proportion had a prior or current history of
IDU (53.3 vs 0.7%, Table 2). Levels of ALT were significantly higher in HCV-infected patients and a larger proportion had ALT levels of more than 2× UN (Table 2),
compatible with more inflammatory activity and potentially more inflammation-related liver injury in this
group.
It has previously been shown that levels of fibrosis
increase with age and our data were compatible with this
observation (Fig. 3A). Although patients infected with
HCV were markedly older than patients infected with
HBV, this age difference did not fully explain the more
severe fibrosis stage in HCV-infected patients. Agestratified comparison of TE results revealed persistently
increased levels of liver stiffness in HCV-infected individuals compared with HBV-infected patients of the same
age group (Fig. 3B).
Higher probability of cirrhosis in older
patients, in male patients and in patients
with chronic hepatitis C
Overall, 91 of 623 (14.1%) patients assessed in this study
had advanced liver fibrosis or cirrhosis (F3–F4,
F4) (Fig. 2).
278
A multivariate logistic regression model revealed that
for patients infected with HCV, age, increased alcohol
consumption, ALT levels of >2× UN and screening at
outreach clinics were associated with higher odds ratio
(OR) for advanced liver fibrosis (Table 3).
Relevant fibrosis was rare among HBV-infected individuals. Multivariate logistic regression identified age as
the only factor independently associated with increased
OR for relevant fibrosis in this population.
Discussion
We demonstrate the feasibility of assessing liver fibrosis
in distinct populations with chronic liver disease through
an outreach programme using a mobile TE device.
Patients reached by this programme differed from the
population seen at a tertiary hospital. The majority of
people assessed in urban community healthcare settings
had a migration background and a relevant proportion of
patients reached at regional clinic locations had a history
of drug use. Consistent with previous epidemiological
findings, the underlying liver disease was largely determined by ethnic background or prior risk behaviour.1,4
Marked differences in levels of fibrosis were seen between
clinical sites and underlying cause of liver disease.
© 2016 Royal Australasian College of Physicians
Outreach transient elastography
Table 3 Risk factors for advanced fibrosis (F3–F4)
Risk factors for advanced fibrosis in HCV-infected patients
Univariable
Male sex
Age quartile
Outreach†
BMI > 25 kg/m2
ALT > 2UN (U/L)
Alcohol yes
Prior or current IDU
Multivariable
OR
CI
P
OR
CI
P
2.058
1.842
2.557
1.400
4.114
3.831
0.721
(1.13–3.74)
(1.38–2.45)
(1.49–4.39)
(0.81–2.41)
(2.25–7.52)
(2.13–6.88)
(0.42–1.22)
0.018
<0.001
0.001
0.230
<0.001
<0.001
0.227
1.446
2.340
2.743
(0.72–2.88)
(1.64–3.33)
(1.45–5.159)
0.296
<0.001
0.002
5.095
2.844
(2.43–10.64)
(1.46–5.54 )
<0.001
0.002
Risk factors for advanced fibrosis in HBV-infected patients
Univariable
Male sex
Age quartile
Outreach†
BMI > 25 kg/m2
ALT > 2UN (U/L)
Alcohol yes
Multivariable
OR
CI
P
OR
CI
P
4.743
2.396
0.893
2.636
4.741
3.321
(1.04–21.61)
(1.39–4.14)
(0.24–3.30)
(0.81–8.61)
(0.92–24.38)
(0.37–29.64)
0.044
0.002
0.865
0.109
0.063
0.283
4.145
2.460
1.091
(0.85–19.62)
(1.40–4.31)
(0.27–4.33)
0.073
0.002
0.902
4.795
(0.75–30.85)
0.99
†Outreach: patients not assessed at tertiary hospital. ALT, alanine transferase; BMI, body mass index; CI, confidence interval; HCV, hepatitis C virus;
IDU, intravenous drug use; OR, odds ratio.
Most advanced fibrosis was found at regional clinics,
where patients were older and a higher proportion was
infected with HCV. It has been demonstrated that both
age and duration of disease contribute to progression
liver fibrosis,17,18 and in our study, older patients had
more advanced fibrosis. There was a significant age difference of 10 years between patients with HBV and HCV
and the observed higher levels of median stiffness in
HCV-infected patients might be age related. However,
higher levels of median stiffness persisted in the agestratified comparison of HCV- and HBV-infected individuals (Fig. 3B).
Despite the age difference between HCV- and HBVinfected patients, it is unlikely that longer disease duration
would account for the increased levels of fibrosis observed
in the HCV population. In our study, most HBV-infected
patients originated from high-endemic countries and
therefore presumably acquired HBV in early childhood. In
contrast, large proportions of HCV-infected patients had a
history of drug use and were likely infected with HCV
during adult life. Thus, infection with HBV is likely to
occur 15–20 years earlier than infection with HCV, suggesting overall shorter disease duration for HCV-infected
individuals despite their older age.
Interestingly, a recent report comparing large cohorts
of HCV- and HBV-infected patients found higher liverassociated hospital morbidity for HCV-infected patients,
possibly reflecting a higher proportion of advanced liver
fibrosis/cirrhosis in the HCV group.19 These results
© 2016 Royal Australasian College of Physicians
suggest that while age is a driver of fibrosis progression,
additional factors, either disease-specific liver injury or
individual characteristics and risk behaviour, must promote the accelerated fibrosis progression in HCVinfected patients.20
In our study, HCV-infected patients had significantly
higher levels of ALT, compatible with more active inflammation. Elevated ALT levels can increase TE results and
lead to an overestimation of fibrosis.21 In our study, multivariate analysis identified increased ALT levels as independent risk factor for relevant fibrosis in HCV-infected
patients, and it is therefore possible that fibrosis levels
were overestimated in some individuals.
ALT flares are not a common feature of chronic HCV
and it is likely that additional factors such as increased
alcohol intake are the cause of increased inflammation
observed in the HCV population.
In addition to the discussed biological variables such as
age, causative agent (HCV), increased ALT and ongoing
liver injury due to alcohol, we found that assessment at
outreach clinics was also independently associated with
increased OR for advanced fibrosis. It is plausible that limited access to healthcare services, delayed specialist referral and evaluation only at an advanced stage account for
this finding. Consistent with this interpretation, patients
with chronic HCV, seen at a tertiary centre, had much
milder fibrosis, a concerning observation, as patients with
more advanced disease and higher risk for complication
are more likely to benefit from tertiary care. Although
279
Thurnheer et al.
overt cirrhosis was rare in the entire population, 26% of
all HCV-infected people had TE results compatible with
advanced fibrosis/cirrhosis (F3–4, F4). This finding is
important and should impact disease management. In an
era when effective treatment may only be available to a
limited number of patients, treatment access should be
prioritised for those with more advanced disease and it is
important to identify populations at risk.
The study was limited by being conducted in one
Australian State and results may not be generalisable.
Smoking status was not recorded, alcohol consumption
was not evaluated quantitatively and patients were
not routinely fasted prior to scanning. The programme
represents a resource intensive approach to improving
hepatitis care and might not be applicable in other settings. However, this audit illustrates that even in a wellfunctioning healthcare system gaps in care have been
identified and should lead to improvements in care for
neglected patient groups.
No data on disease duration were collected and therefore the impact of disease duration on fibrosis progression
remains speculative. Also, no information on prior care
was available, although for most of the patients seen in
References
1 MacLachlan JH, Allard N, Towell V,
Cowie BC. The burden of chronic
hepatitis B virus infection in Australia,
2011. Aust N Z J Public Health 2013; 37:
416–22.
2 Sievert W, Razavi H, Estes C, Thompson
AJ, Zekry A, Roberts SK et al. Enhanced
antiviral treatment efficacy and uptake
in preventing the rising burden of
hepatitis C-related liver disease and
costs in Australia. J Gastroenterol Hepatol
2014; 29: 1–9.
3 Reekie J, Gidding HF, Kaldor JM, Liu B.
Country of birth and other factors
associated with hepatitis B
prevalence in a population with
high levels of immigration. J
Gastroenterol Hepatol 2013; 28:
1539–44.
4 Dore GJ, Ward J, Thursz M. Hepatitis C
disease burden and strategies to manage
the burden (Guest Editors Mark Thursz,
Gregory Dore and John Ward). J Viral
Hepat 2014; 21: 1–4.
5 Dahl TF, Cowie BC, Biggs BA, Leder K,
MacLachlan JH, Marshall C. Health
literacy in patients with chronic
hepatitis B attending a tertiary
hospital in Melbourne: a questionnaire
based survey. BMC Infect Dis 2014;
14: 537.
280
regional settings and at PWUD clinics the current assessment was their first (personal communication J. Sasadeusz). Treatment data were scarce, and no statement can
be made as to whether treatment might have influenced
the observed fibrosis levels in this study. We cannot comment on the long-term impact of this outreach strategy on
disease management and it remains unclear whether disease evaluation resulted in a higher rate of treatment
uptake. Although PWUD represent the largest group of
people affected by chronic hepatitis C, only a small group
of people was evaluated at the PWUD clinic, reflecting
remaining challenges to reach this patient group.
Conclusion
Overall, we have shown that an outreach programme
using a mobile TE device improves access to evaluation
of liver fibrosis in affected populations who might otherwise not be reached. The finding that patients seen at
regional clinics presented with more advanced disease
should reinforce attempts to further improve access to
care for these individuals.
6 Wallace J, McNally S, Richmond J,
Hajarizadeh B, Pitts M. Managing
chronic hepatitis B: a qualitative study
exploring the perspectives of people
living with chronic hepatitis B in
Australia. BMC Research Notes 2011;
4: 45.
7 Harris M, Rhodes T. Hepatitis C
treatment access and uptake for people
who inject drugs: a review mapping the
role of social factors. Harm Reduction
Journal 2013; 10: 7.
8 Zeremski M, Zibbell JE, Martinez AD,
Kritz S, Smith BD, Talal AH. Hepatitis C
virus control among persons who inject
drugs requires overcoming barriers to
care. World J Gastroenterol 2013; 19:
7846–51.
9 Fisher DA, Huffam SE. Management of
chronic hepatitis B virus infection in
remote-dwelling Aboriginals and Torres
Strait Islanders: an update for primary
healthcare providers. Med J Aust 2003;
178: 82–5.
10 Preston-Thomas A, Fagan P, Nakata Y,
Anderson E. Chronic hepatitis B--care
delivery and patient knowledge in
the Torres Strait region of
Australia. Aust Fam Physician 2013; 42:
225–31.
11 Nazareth S, Kontorinis N,
Muwanwella N, Hamilton A,
Leembruggen N, Cheng WS. Successful
12
13
14
15
16
17
treatment of patients with hepatitis C in
rural and remote Western Australia via
telehealth. J Telemed Telecare 2013; 19:
101–6.
Castera L, Forns X, Alberti A. Noninvasive evaluation of liver fibrosis
using transient elastography. J Hepatol
2008; 48: 835–47.
Foucher J, Reiller B, Jullien V, Leal F, di
Cesare ES, Merrouche W et al.
FibroScan used in street-based outreach
for drug users is useful for hepatitis C
virus screening and management: a
prospective study. J Viral Hepat 2009;
16: 121–31.
Moessner BK, Jorgensen TR,
Skamling M, Vyberg M, Junker P,
Pedersen C et al. Outreach screening of
drug users for cirrhosis with transient
elastography. Addiction 2011; 106:
970–6.
de Ledinghen V, Vergniol J. Transient
elastography (FibroScan). Gastroenterol
Clin Biol 2008; 32: 58–67.
Echosens. Echosens handbook 2015.
[cited 2016 Jan 26]. Available from
URL: http://asp-indus.secure-zone.net/
v2/indexPop.jsp?id=955/1196/3607
Pradat P, Voirin N, Tillmann HL,
Chevallier M, Trepo C. Progression to
cirrhosis in hepatitis C patients: an agedependent process. Liver Int 2007; 27:
335–9.
© 2016 Royal Australasian College of Physicians
18 Poynard T, Mathurin P, Lai C-L,
Guyader D, Poupon R, Tainturier M-H
et al. A comparison of fibrosis
progression in chronic liver
diseases. J Hepatol 2003; 38:
257–65.
19 Gidding HF, Dore GJ, Amin J, Law MG.
Trends in all cause and viral
liver disease-related hospitalizations
in people with hepatitis B or C:
a population-based linkage study.
BMC Public Health 2011; 11: 52.
20 Rueger S, Bochud PY, Dufour JF,
Mullhaupt B, Semela D, Heim MH et al.
Impact of common risk factors of
fibrosis progression in chronic hepatitis
C. Gut 2015; 64: 1605–15.
21 Arena U, Vizzutti F, Corti G, Ambu S,
Stasi C, Bresci S et al. Acute viral
hepatitis increases liver stiffness values
measured by transient elastography.
Hepatology 2008; 47: 380–4.
Baseline abnormal liver function tests are more important than
age in the development of isoniazid-induced hepatoxicity for
patients receiving preventive therapy for latent tuberculosis
infection
E. L. Gray and H. F. Goldberg
1
Department of Respiratory Medicine, Prince of Wales Hospital, Sydney, New South Wales, Australia
Key words
latent tuberculosis, isoniazid, drug-induced liver
injury, patient compliance, rifampicin.
Correspondence
Emma Gray, Department of Respiratory
Medicine, Prince of Wales Hospital, Barker
Street, Randwick, Sydney, NSW 2031, Australia.
Email: [email protected]
Received 20 September 2015; accepted
23 November 2015.
doi:10.1111/imj.12979
Abstract
Background: One of the cornerstones of Australia’s public health programmes to eliminate tuberculosis (TB) is the identification and treatment of latent tuberculosis infection (LTBI).
Aims: The main aim of this study is to determine the demographics, compliance, completion rates and adverse events of patients on preventive therapy (PT) for LTBI at our
institution. The secondary aim is to determine the rates of isoniazid (INH) hepatotoxicity and identify any contributory factors.
Methods: The method used was an audit using medical records of 100 consecutive
patients (2010–2014) treated with PT for LTBI.
Results: Seventy-two patients with confirmed LTBI started 9 months of INH and
22 started 4 months of rifampicin (RIF). The median age was 30 years. Half the patients
were born in high TB-prevalence countries. Fifty-six per cent were contacts of index
cases with confirmed TB, and 26% were pre-immunosuppression. Seventy-seven per
cent completed PT with adequate compliance. Thirty-three per cent on INH and 23%
on RIF experienced some liver function test (LFT) abnormality while on treatment. INH
was ceased in 3% due to asymptomatic hepatic dysfunction (transaminases >5x upper
limit of normal). No patients had permanent liver damage. Significant risk factors for
liver dysfunction during PT were risk factors for liver disease (χ23 = 8.7; P = 0.03) or
abnormal pre-therapy LFT (χ23 = 22.4; P < 0.001). No patients developed active TB.
Conclusion: The completion rate of 77% and rate of INH-induced hepatic dysfunction of
3% is comparable with the literature. We found no age association with the risk of INHinduced hepatic dysfunction; however, there was a significant and linear association with
the degree of liver function abnormality during INH therapy and the presence of abnormal
baseline LFT. Routine LFT monitoring allowed early cessation of INH in those with significant but asymptomatic hepatitis who did not meet criteria for ATS/CDC LFT monitoring.
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
281
Gray & Goldberg
Introduction
Australia has a low incidence of tuberculosis (TB) with
the majority of cases of active disease occurring due to
reactivation of latent infection, most commonly acquired
in high prevalence countries.1 One of the cornerstones
of public health programmes to eliminate TB is the identification and treatment of latent tuberculosis infection
(LTBI).2 Isoniazid (INH) preventive therapy (PT) has
been used for many decades with its effectiveness a function of both duration of treatment and adherence. The
optimal duration of therapy has been determined to be
9 months for immuno-competent adults, which provides
a protective efficacy of up to a maximum of 90% against
reactivation of TB.3 There is a wide range of estimates in
the literature relating to completion rates of INH therapy. These range from 19 to 96% with higher compliance rates seen with 6 months of treatment.4
INH-induced liver injury is idiosyncratic, independent
of dosing and is a diagnosis of exclusion.5 Historically,
the incidence of INH-induced asymptomatic hepatitis has
been reported in up to 10–20% patients and overt hepatitis in 1%.2,5 The incidence is generally reported to be
age-related, with those older than 35 years at increased
risk.6 However, the lack of specific diagnostic criteria
complicates comparisons across studies. The current
American Thoracic Society/Centers for Disease Control
and Prevention (ATS/CDC) recommendations are that
routine blood testing of LFT is indicated only if baseline
transaminases are abnormal or if the patient is at risk of
hepatic disease, defined as having the humanimmunodeficiency virus, chronic liver disease, being
pregnant or postpartum, excessive alcohol consumption
or an active intravenous drug user.
At our institution, PT is dispensed monthly by the TB
clinic, free of charge to patients, and usually comprises
9 months of INH therapy. Patients greater than 30 years
have baseline and then monthly blood tests looking for
hepatotoxicity. Patients 16–30 years of age have baseline
LFT testing; however, further monitoring is at the discretion of the clinician. All patients have clinical reviews at
3-month intervals after commencement of therapy.
Patients are educated as to the side-effect profile of INH,
isigns of hepatotoxicity (anorexia, nausea, vomiting and
jaundice) and peripheral neuropathy in particular, and
are instructed to contact the clinic if these symptoms
occur.
Aims
The primary aim of this audit was to evaluate the demographics, compliance, completion rates and adverse
events of patients on PT for LTBI. Secondly, we wanted
282
to determine the rates of INH hepatotoxicity and to identify any contributory factors.
Methods
A retrospective audit was undertaken of 100 consecutive
patients who commenced PT for LTBI at our
institution from 15 July 2010 to 10 December 2013. The
ethics was submitted but deemed not necessary by
the committee. Patients were identified sequentially
from the database kept in the chest clinic. Their medical
records and pathology results were subsequently
analysed.
At our institution, children less than 5 years of age
who are deemed close TB contacts are offered INH whilst
waiting for their second tuberculin skin test (TST),
10 weeks after the break of contact. If their second TST
is negative, therapy is ceased at this time. During the
study period, five children were commenced on INH and
then ceased due to a second negative TST. As these children did not have LTBI, they were excluded from further
analysis.
A 77-year-old female born in Greece had a negative
TST and an initial positive QuantiFERON-TB Gold (QFTG) prior to tumour necrosis factor (TNF)-alpha blockade.
A decision was made to start INH therapy. Her QFT-G
batch was subsequently recalled with the repeat test
being negative. Therapy was stopped after 89 days of
treatment. The patient experienced no adverse events.
As she did not have LTBI, she was excluded from further
analysis.
INH-induced hepatotoxicity was defined in this
study as:
1. the presence of symptoms of hepatitis and laboratory
evidence of hepatic dysfunction with aminotransferase (alanine aminotransferase (ALT) or aspartate
aminotransferase (AST)) elevation greater than three
times the upper limit of normal (ULN) with or without bilirubin elevation or
2. asymptomatic aminotransferase elevation greater
than five times ULN with or without bilirubin
elevation.
Statistical analysis
Statistical analysis calculations were carried out using
IBM SPSS Statistics Version 22.0 (IBM, Armonk, NY,
USA). Non-parametric analysis was undertaken. The
Pearson chi-test was used to identify any demographic
contributors to those who ceased therapy early, had any
adverse event (grade 1–4 side-effect) or developed either
mild or significant liver function test (LFT) abnormalities.
© 2016 Royal Australasian College of Physicians
Isoniazid for latent tuberculosis
100 patients commenced on LTBI therapy
78 commenced on INH
22 commenced on RIF
72 confirmed LTBI
22 confirmed LTBI
6 excluded as did
not have LTBI on
further
investigation (see
text)
Figure 1 Flow diagram of 100 consecutive
patients screened for LTBI at our chest clinic.
INH, Isoniazid; LTBI, latent tuberculosis infection; RIF, rifampicin.
16 ceased
prematurely
2 ceased
prematurely
1 lost to follow-up
1 lost to follow up
2 transferred
service
53 completed INH
19 completed RIF
Table 1 Clinical characteristics of patients
All P < 0.05
significant.
values
were
considered
statistically
Results
During the period of the audit, a prospective randomised
controlled trial assessing the efficacy of 4 months rifampicin (RIF) therapy versus 9 months INH therapy was
being conducted. Of the 94 patients commenced on PT
for LTBI, 72 (77%) were started on INH and 22 (23%)
on RIF (Fig. 1).
The clinical characteristics of the patients are shown
below in Table 1. Of the 24 patients commencing PT
prior to immunosuppression, 13 were pre-TNF-alpha
therapy, seven were pre-solid-organ transplantation,
two had systemic lupus erythematosus nephritis and two
were pre-chemotherapy.
Patients who were commenced on therapy prior to
immunosuppression were older (median 50 years; range
22–80 years) than those commenced for being close contacts (28 years; 3–69 years) or for other indications
(27 years; 2–57 years) (P < 0.001; independent samples
median test).
Of the 94 patients started on PT for LTBI, 88 met the
appropriate TST criteria for LTBI for their indication
(Table 2). Five patients did not have a TST performed as
they had already tested positive for QFT-G. There were
six discordant results (all TST+/QFT-G-).
Seventy-two (77%) patients completed therapy with
adequate compliance, finishing 9 months of INH therapy
within 10 or 4 months of RIF-therapy within 5 months
of their start date. Twenty patients stopped prematurely
(17 on INH, 3 on RIF) with two transferring to another
© 2016 Royal Australasian College of Physicians
Characteristics
Gender
Male
Female
Age at commencement of treatment
Age <16 years
Age 16–35 years
Age ≥35 years
Country of birth
High prevalence for TB†
Low prevalence for TB
Australia
BCG vaccination status
Yes
No
Unsure
Chest radiograph
Normal
Apical thickening or granuloma
Abnormality unrelated to TB
Indication for treatment
Healthcare worker
Pre-immunosuppression
Close contacts
Australian defence force
Other
Number (%)
42 (45)
52 (55)
30 years (2–80 years)
12 (13)
44 (47)
38 (40)
50 (53)
18 (19)
26 (28)
56 (60)
28 (30)
10 (10)
82 (87)
9 (10)
3 (3)
11 (12)
24 (26)
53 (56)
3 (3)
3 (3)
†Countries with an incidence of greater than or equal to 40 cases per
100 000 persons (WHO Global Tuberculosis Control Report 2013). TB,
tuberculosis.
service (Table 3). Those that stopped did so at a median
of 95 days (7–221 days) after commencing therapy.
Patients taking INH were statistically no more likely
than those taking RIF to cease prematurely (χ21 = 1.5;
P = 0.2). However, the point estimate of early cessation
283
Gray & Goldberg
was higher in the INH group compared with the RIF
group (22 vs 8%; P = 0.08).
Twenty-nine patients (24 of 72 (33%) on INH and
5 of 22 (23%) on RIF) experienced some LFT abnormality on treatment, with the peak occurring at a median of
3 months (0–8 months) after commencement of PT
(Table 4). No patients experienced symptomatic INHinduced hepatotoxicity.
Significant risk factors for having LFT abnormalities
while on INH included having a pre-existing risk factor
for liver disease (χ23 = 8.7; P = 0.03) and having any LFT
abnormality prior to commencing therapy (χ23 = 22.4;
P < 0.001). In addition, both of these had statistically significant linear associations, with the degree of liver function abnormality increasing with the presence of
baseline abnormal LFT (linear-by-linear association chi
square test; χ21 = 22.2; P < 0.001) and risk factors for
liver disease (χ21 = 3.8; P = 0.05). Regarding the risk of
LFT abnormality during PT, there was no association
with gender or concurrent hepatotoxic medications. Age
was not associated with either any or major LFT
abnormalities.
Two of the 19 patients who ceased INH prematurely
did so due to INH-induced hepatotoxicity with the only
significant risk factor being the presence of any baseline
LFT abnormality (χ21 = 15.6; P < 0.001). There was no
association with elevated transaminases (AST/ALT) at
baseline and significant INH-induced hepatotoxicity
(χ21 = 0.2; P = 0.7). No patients experienced RIF-induced
hepatotoxicity.
Of the 24 cases with abnormal LFT on INH therapy,
19 demonstrated complete resolution of hepatic dysfunction at completion or after cessation of therapy. The
remaining five had transaminase improvements to less
than 3 × ULN. Males were twice as likely (95% confidence interval (CI) 1.3–3.0, χ21 = 5.3; P = 0.02) to have
persistent LFT abnormalities. There was no association
with age ≥35 years (χ21 = 1.7; P = 0.2) or degree of LFT
abnormality (χ22 = 3.0; P = 0.2) in predicting the resolution of LFT abnormalities. No patients experienced fulminant hepatotoxicity or were admitted to hospital. No
Table 3 Outcomes of PT for LTBI
Table 2 Tuberculin skin test (TST) and QuantiFERON-TB Gold (QFT-G)
test results
QFT-G result
TST result†
Positive
Negative
Not done
Total
Total
88
1
5
94
Positive Negative
3
6‡
0
0
5
0
8
6
Indeterminate Not done
0
79
1
0
0
0
1
79
†TST considered positive if ≥5 mm for a close contact and ≥10 mm for
all other indications. ‡Batch recalled: initial QFT-G was positive with
repeat QFT-G negative; the patient continued on therapy. INH, isoniazid;
RIF, rifampicin.
Outcome
Treatment completed (%)
Stopped early (%)
Grade 3/4 side-effect
Grade 1/2 side-effect
Became pregnant
Social reasons
Lost to follow up (%)
Transferred service (%)
INH,
number (%)
RIF,
number (%)
Total,
n = 94
53 (74)
16 (22)
3
12
0
1
1 (1)
2 (3)†
19 (86)
2 (9)
0
1
1
0
1 (5)
0
72 (77)
18 (19)
3
13
1
1
2 (2)
2 (2)
†One patient completed 3 months, and one patient completed
7 months prior to transferring to another TB clinic; GP, general practitioner; INH, isoniazid; LTBI, latent tuberculosis infection; PT, preventive
therapy; RIF, rifampicin.
Table 4 INH-associated LFT abnormalities for the 61 patients who had LFT monitoring
Total
Median age (years)
Male
Age ≥ 35 years
Abnormal baseline LFT
Risk factors for liver disease
Hepatotoxic meds
Pre-immunosuppression
Peak month
Resolved
No abnormality
Mild LFT abnormality†
ALT/AST 3–5 × ULN
ALT/AST >5 × ULN
37
33
15 (41%)
17 (46%)
1 (3%)
1 (3%)
7 (19%)
13 (35%)
n/a
n/a
17
41
8 (47%)
10 (59%)
6 (35%)
2 (12%)
3 (18%)
9 (53%)
3
15 (88%)
5
30
3 (60%)
2 (40%)
3 (60%)
2 (40%)
1 (20%)
1 (20%)
4
3 (60%)
2
39
2 (100%)
1 (50%)
2 (100%)
0
1 (50%)
0
0.3
1 (50%)
P-value
0.8
0.4
0.8
<0.001
0.03
0.7
0.3
0.2
†Mild LFT abnormality classified as transaminase elevation 1–3 × ULN (13 patients) or isolated gamma-glutamyl transferase (GGT) elevations
(4 patients). ALT, alanine aminotransferase; AST, aspartate aminotransferase; INH, isoniazid; LFT, liver function test; ULN, upper limit of normal.
284
© 2016 Royal Australasian College of Physicians
Isoniazid for latent tuberculosis
patients developed active pulmonary TB on therapy;
chest radiographs remained stable on completion of
treatment in all patients.
Discussion
Completion of 9 months of INH therapy, with an efficacy
of approximately 90%, currently remains the standard
therapy for the treatment of LTBI.3 Effective treatment is
of particular importance for patients with a higher estimated risk of developing active TB, such as those with
LTBI undergoing transplantation (RR 20–74)7–10 or TNFalpha therapy (RR 1.7–9.0)11–14). However, there remain
two main barriers to the acceptance of therapy, leading
to poor adherence and consequently reduced effectiveness. The first is the long duration of therapy and the
second is INH toxicity, particularly hepatotoxicity, which
can be fatal.
Alternative regimens are available, including
4 months of RIF montherapy and 3 or 4 months of combination RIF–INH. Randomised controlled trials have
already shown that RIF monotherapy is consistently
safer than INH, especially with regards to hepatotoxicity.15,16 Despite this, long-term follow-up data on
both efficacy and acquired RIF resistance are relatively
limited compared with the body of evidence supporting
the use of INH. The results of a randomised
controlled trial of 4 months RIF against 9 months of INH
is due in the next couple of years and may provide some
answers.
Our completion rate of 74%, with adequate compliance, for 9 months of INH therapy is on the higher end
of the range in the literature. On post-hoc analysis, 78%
completed at least 6 months of INH therapy. A systematic review of completion rates from 78 studies on LTBI
PT showed highly variable rates of 19–96%.4 However,
limiting this to the largest studies using INH only, the
rates are more constant at 61–64%.17–21 There is some
suggestion that ‘clinic factors’, such as intensive patient
education, follow up and support, may lead to increased
patient completion; however, this has not been formally
examined.17,22 A recent audit at another Sydney TB
clinic found completion rates of 75% for 6 months INH
therapy; however, those that ceased due to adverse
events were excluded from the analysis.23
For patients commencing RIF, 86% completed therapy. This was not statistically different to the completion
rates for INH; however, our study supports the current
opinion of an improved safety profile of RIF, with a
trend towards more adverse events experienced after
9 months of INH therapy.
© 2016 Royal Australasian College of Physicians
Overall, there were no statistically significant demographic features that predicted either early cessation of
PT or adverse events leading to cessation of PT. Our
overall rates of 17% ceasing due to adverse events
(including 3% grade 3/4 and 14% grade 1/2) are high
compared with the literature (1.4–3.6%).17,22 The relatively high proportion of patients ceasing due to mild
adverse events may be due to a change in the perceived
risk/benefit ratio of PT in these patients, especially as half
of the decisions to cease PT were initiated by the patient.
We are prescribing potentially toxic medications to a
relatively well cohort of asymptomatic patients. This
combined with patients being more attuned to
symptoms due to starting a new medication may have
led to an over-representation of minor adverse events
causing cessation of therapy. In addition, our 2% rate
of loss to follow up is nearly 10 times lower than
other retrospective audits.17 This may be due to the close
monitoring of patients’ compliance and up to
10 phone calls for missed scheduled medication pick-ups
or clinic appointments. Given that our overall noncompletion rates are similar to other studies, approximately 20%, an alternative explanation is that patients
in other studies may have experienced adverse events
and decided not to continue on with therapy or follow
up, leading to the different proportions of those lost
to follow up compared with those experiencing adverse
events.
In the literature, up to 20% of individuals treated with
INH alone experience low-grade, asymptomatic and
transient elevations of transaminases.24,25 INH-induced
hepatotoxicity is reported in 1–4% of patients treated
with LTBI in smaller studies,26,27 with larger reviews
quoting the rates of 0.1–0.56%.20,28 Some of this difference pertains to the definition used and monitoring
practices of the clinics involved. For example, rates of
hepatotoxicity in a Seattle centre with careful patient
selection (i.e. few patients older than 35 years) and
monthly clinical monitoring were as low as 0.1%.20
However, the definition of hepatotoxicity used was
based on aspartate aminotransferase (AST) levels only
despite alanine aminotransferase (ALT) being more specific for hepatocellular injury.29,30 In addition, hepatotoxicity rates are higher (1.7%) in patients 35 years and
older compared with younger patients (0.2%).31 Despite
our older population group, our results are comparable
with the existing body of literature.
In our study, we found that abnormal baseline LFT
were associated with INH-induced hepatotoxicity. There
was a significant linear association between the degree
of hepatic dysfunction observed during treatment and
the presence of both abnormal baseline LFT and risk
factors for liver disease. We did not find an age or
285
Gray & Goldberg
gender association, which previous studies have
shown.2,3,17,20
Due to the study design, there are limitations in attributing causation of adverse events to INH given the
lack of a control arm. The small cohort size may have
been underpowered to detect patient characteristics
that led to premature cessation or adverse events. Compliance and completion data were based on patient
reporting and recording of monthly pick-up dates. There
is a possibility that patients were picking up their medication, attending appointments but failing to take the
medications.
Conclusion
Our cohort of patients undergoing 9 months of INH
therapy had respectable completion rates compared with
other retrospective audits. This may be related to the
strong emphasis on close monitoring and patient education as well as the free provision of medication, blood
tests and follow up.
The rates of INH-induced asymptomatic hepatitis leading to cessation of therapy occurred in 3% of those
References
1 Denholm J, McBryde E. Management of
latent tuberculosis infections in
Australia and New Zealand: a review of
current practice. Tuber Res Treat 2010;
2010: 5.
2 Centers for Disease Control and
Prevention. Targeted tuberculin skin
testing and treatment of latent
tuberculosis infection. Morb Mortal Wkly
Rep 2000; 49: 1–54.
3 Comstock G. How much isoniazid is
needed for prevention of tuberculosis in
immunocompetent adults. Int J Tuber
Lung Dis 1999; 3: 847–50.
4 Hirsch-Moverman Y, Daftary A,
Franks J, Colson P. Adherence to
treatment for latent tuberculosis
infection: systematic review of studies in
the US and Canada. Int J Tuber Lung Dis
2008; 12: 1235–54.
5 American Thoracic Society. An official
ATS statement: hepatotoxicity of
antituberculosis therapy. Am J Respir Crit
Care Med 2006; 174:
935–52.
6 Kucers A, Crowe S, Grayson M, Hoy J.
The Use of Antibiotics: A Clinical Review of
Antibacterial, Antifungal and Antiviral
286
7
8
9
10
11
commencing INH. This value is comparable with smaller
studies reported in the literature. The ATS/CDC guidelines state that monitoring of patients is only warranted
if ALT/AST at baseline is elevated or there are risk factors
for liver disease. Although our study agrees that preexisting risk factors for liver disease were associated
with INH-induced hepatotoxicity, we found that any
abnormal baseline LFT, not just abnormal tranaminases,
were associated with INH-induced hepatotoxicity. Following these guidelines, we would have missed one of
our patients. Gone unchecked, asymptomatic liver
enzyme elevations may develop into fulminate
hepatotoxicity.
Despite the small numbers of this audit, we
have shown that we can safely administer INH to a relatively older cohort of patients. This is of particular interest given the increasing age of patients being
offered solid-organ transplantation and TNF-alpha therapy in other areas of medicine. Although the safety
data on RIF are promising, the long-term efficacy data
will not be available for some years. In the interim, we
have available a medication with a long history of use
and efficacy.
Drugs, 5 edn. London: ButterworthHeinemann; 1997.
Sakhuja V, Jha V, Varna P, Joshi K,
Chugh K. The high incidence of
tuberculosis among renal transplant
recipients in India. Transplantation 1996;
61: 211–15.
Aguado J, Herrero J, Gavalda J, TorreCisneros J, Blanes M, Rufi G et al.
Clinical presentation and outcome of
tuberculosis in kidney, liver and heart
transplant recipients in Spain. Spanish
Transplantation Infection Study Group.
GESITRA. Transplantation 1997; 63:
1278–86.
Miller R, Lanza L, Kline J, Geist L.
Mycobacterium tuberculosis in lung
transplant recipients. Am J Respir Crit
Care Med 1995; 152: 374–6.
Meyers B, Halpern M, Sheiner P,
Mendelson M, Neibart E, Miller C.
Tuberculosis in liver transplant patients.
Transplantation 1994; 58: 301–6.
Carmona L, Gomez-Reino J, RodriguezValverde V, Montero D, PascualGomez E, Mola E et al. Effectiveness of
recommendations to prevent
reactivation of latent tuberculosis
infection in patients treated with
tumour necrosis factor antagonists.
12
13
14
15
16
Arthritis Rheum 2005; 52:
1766–72.
Keane J, Gershon S, Wise R, MirabileLevens E, Kasznica J, Schwieterman W
et al. Tuberculosis associated with
infliximab, a tumour necrosis factor a –
neutralising agent. N Engl J Med 2001;
345: 1098–104.
Brassard P, Kezouh A, Suissa S.
Antirheumatic drugs and the risk of
tuberculosis. Clin Infect Dis 2006; 43:
717–22.
Wolfe F, Michard K, Anderson J,
Urbansky K. Tuberculosis infection in
patients with rheumatoid arthritis and
the effect of infliximab therapy. Arthritis
Rheum 2010; 50: 372–9.
Menzies D, Dion M, Rabinovitch B,
Mannix S, Brassard P, Schwartzman K.
Treatment completion and costs of a
randomised trial of rifampin for
4 months versus isoniazid for 9 months.
Am J Respir Crit Care Med 2004; 170:
445–9.
Menzies D, Long R, Trajman A, Dion M,
Yang J, Al Jajdali H et al. Adverse events
with 4 months of rifampin therapy or
9 months of isoniazid therapy for latent
tuberculosis infection: a randomised
trial. Ann Intern Med 2008; 149: 689–97.
© 2016 Royal Australasian College of Physicians
Isoniazid for latent tuberculosis
17 LoBue P, Moser K. Use of isoniazid for
latent tuberculosis infection in a public
health clinic. Am J Respir Crit Care Med
2003; 168: 443–7.
18 Jereb J, Etkind S, Joglar O, Moore M,
Taylor Z. Tuberculosis contact
investigations: outcomes in selected
areas of the United States, 1999. Int J
Tuberc Lung Dis 2003; 7: S384–90.
19 Lobato M, Leary L, Simone P.
Treatment for latent TB in correctional
facilities: a challenge for TB elimination.
Am J Prev Med 2003; 24: 249–53.
20 Nolan C, Goldberg S, Buskin S.
Hepatotoxicity associated with isoniazid
preventive therapy: a 7-year survey
from a public health tuberculosis clinic.
JAMA 1999; 11: 1014–18.
21 Sprinson J, Flood J, Fan C, Shaw T,
Pascopella L, Young J et al. Evaluation
of tuberculosis contact investigations in
California. Int J Tuberc Lung Dis 2003; 7:
S363–8.
22 Jasmer R, Snyder D, Chin D,
Hopewell P, Cuthbert S, Antonio Paz E
et al. Twelve months of isoniazid
compared with four months of rifampin
for persons with radiographic evidence
© 2016 Royal Australasian College of Physicians
23
24
25
26
27
of previous tuberculosis: an outcome
and cost-effectiveness analysis.
Am J Respir Crit Care Med 2000; 162:
1648–52.
Dobler C, Marks G. Completion of
treatment for latent tuberculosis
infection with monthly drug
dispensation directly through the
tuberculosis clinic. PLoS One 2012; 7:
e48900.
Mitchell J, Zimmerman H, Ishak K,
Thorgerisson U, Timbrell J,
Snodgrass W et al. Isoniazid liver injury:
clinical spectrum, pathology and
probable pathogenesis. Ann Intern Med
1976; 84: 181–92.
Scharer L, Smith J. Serum transaminase
elevations and other hepatic
abnormalities in patients receiving
isoniazid. Ann Intern Med 1969; 71:
1113–20.
McNeill L, Allen M, Estrada C, Cook P.
Pyrazinamide and rifampin vs isoniazid
for the treatment of latent tuberculosis:
improved completion rates but more
hepatoxicity. Chest 2003; 123: 102–6.
Millard P, Wilcosky T, ReadeChristopher S, Weher D. Isoniazid-
28
29
30
31
related fatal hepatitis. West J Med 1996;
164: 486–91.
Fountain F, Tolley E, Chrisman C,
Self T. Isoniazid hepatotoxicity
associated with treatment of latent
tuberculosis infection: a 7 year
evaluation from a public health
tuberculosis clinic. Chest 2005; 128:
116–23.
Dufour D, Lott J, Nolte F, Gretch D,
Koff R, Seeff L. Diagnosis and
monitoring of hepatic injury.
II. Recommendations for use of
laboratory tests in screening, diagnosis
and monitoring. Clin Chem 2000; 46:
2050–68.
Dufour D, Lott J, Nolte F, Gretch D,
Koff R, Seeff L. Diagnosis and
monitoring of hepatic injury:
I. Performance characteristics of
laboratory tests. Clin Chem 2000; 46:
2027–49.
Kunst H, Khna K. Age-related risk of
hepatoxicity in the treatment of latent
tuberculosis infection: a systematic
review. Int J Tuberc Lung Dis 2010; 14:
1374–81.
287
Home medicines reviews in Australian war veterans taking
warfarin do not influence international normalised ratio control
L. R. E. Bereznicki,1 E. C. van Tienen1 and A. Stafford2
2
WA Dementia Training Study Centre, School of Pharmacy, Curtin University, Perth, Western Australia, Australia
Key words
anticoagulants, elderly, warfarin, home-based
team care, medication review.
Correspondence
Luke Ryan Elliot Bereznicki, Department of
Pharmacy, School of Medicine, University of
Tasmania, Private Bag 26, Hobart, TAS 7001,
Australia.
Email: [email protected]
Received 17 July 2015; accepted
6 November 2015.
doi:10.1111/imj.12964
Abstract
Background: The clinical outcomes of warfarin are largely dependent on the international normalised ratio (INR) control achieved, and strategies to improve the time in
therapeutic range (TTR) should be identified and widely implemented in practice.
Aims: To investigate the influence of pharmacist-led medication reviews on INR control and observe the quality of INR control in Australian veterans who take warfarin.
Methods: We undertook a retrospective cohort study using administrative claims data
for Australian veterans and war-widows identified by the Department of Veterans’
Affairs who were regularly dispensed warfarin and invited them to contact the research
team. Pathology providers were subsequently contacted to provide INR results.
Results: INR data were available for 344 of 818 (42.1%) veterans who consented to
participate in the study; 64.4% were male and the median age was 83 years. The overall TTR for the veteran cohort during the study period was 64.0%. There was no difference in the TTR in the 6 months following home medicines review (HMR) compared
with the control group (63.0% vs 67.0%, P = 0.27), with the TTR in patients with INR
data available in the 6 months prior to, and the 6 months following HMR, remaining
high (67.9% vs 69.6% P = 0.63). Approximately, one-third of veterans in this study
had a percentage TTR below 60%.
Conclusions: INR was well-controlled in this elderly cohort, comparable to that
achieved in recent randomised trials involving warfarin. Pharmacist-led medication
reviews were not associated with a change in INR control.
Introduction
Most individuals who receive warfarin therapy are elderly patients with atrial fibrillation and acute or recurrent venous thromboembolism. Anticoagulation in
elderly people poses unique challenges because they are
simultaneously at higher risk of recurrent thromboembolism and major bleeding, including catastrophic intracranial haemorrhage. The effectiveness of warfarin
therapy is strongly linked to the proportion of time that
patients spend in the target international normalised
ratio (INR) range (time in therapeutic range, TTR).1,2
The risk of death, myocardial infarction, major bleeding
Funding: This article was funded by the Commonwealth
Department of Veterans’ Affairs under the Applied Research
Program. L. R. E. Bereznicki has received consultancy research
funding and speaker honoraria from Roche Diagnostics Australia,
and consultancy funding from Aspen Pharmacare Australia.
Conflict of interest: None.
288
and stroke or systemic embolism are all related to INR
control.3
In Australia, the Department of Veterans’ Affairs
(DVA) introduced a system of providing formal medication reviews for Australian veterans in 1999, which was
followed by the home medicines review (HMR) program
(available to all members of the Australian public) in
2001. For veterans taking warfarin, these reviews provide an opportunity for patient education and review of
warfarin management in the community setting. General practitioners refer patients to an accredited pharmacist who undertakes a home visit, identifying any
medication-related problems, including potential underuse, overuse, adverse effects, compliance and knowledge
problems, or hoarding. The pharmacist provides a report
to the doctor who has responsibility for follow-up with
the patient. The potential benefits of patients receiving a
pharmacist-conducted medication review were established in several large research projects performed in the
© 2016 Royal Australasian College of Physicians
Medication reviews and INR control
late 1990s.4,5 These studies found that HMR resulted in
the resolution of medication-related problems and
showed trends in reduced medication costs.
A study by Roughead et al. assessed the effect of HMR
in Australian veterans and war widows taking warfarin
retrospectively using administrative claims data.6 The
study identified a 79% reduction in the likelihood of
hospitalisation for bleeding between 2 and 6 months following the HMR (hazard ratio 0.21, 95% CI 0.05–0.87).
This beneficial effort was not evident for 6–12 months
following review. As INR testing and TTR were not
assessed in the study, it is unclear whether improved
INR control occurred as a result of the HMR, or whether
the benefits occurred independently of improved INR
control.
We aimed to determine whether HMRs are associated
with improved INR control, and observe the degree of
INR control in this population.
The DVA identified a list of veterans who met the
inclusion criteria and who had also had an HMR prior to
30 June 2009. This allowed for data to be available for at
least 6 months following the HMR in the group exposed
to HMR. The DVA then randomly selected a matching
number of veterans who met the inclusion criteria who
had not been exposed to an HMR in the study period.
The identified veterans were sent an information sheet
and consent form. A list of veterans who consented to be
involved in the study was generated and sent to the
DVA for data extraction.
Data provided by the DVA included age, sex,
remoteness,10 number of co-morbidities, dispensed medications, dates of HMR claims and health services utilisation
(hospitalisations
including
diagnoses
and
procedures; general practitioner visits; and specialist visits). The research team then contacted pathology laboratories that had claimed payment from DVA for
measuring the veteran’s INR.
Methods
Formation of HMR and control groups
The DVA database
The Australian DVA claims databases contain details of
all prescription medicines, medical and allied health services and hospitalisations provided for which DVA pay a
subsidy. At the time of the study, the data file contained
140 million pharmacy records, 200 million medical and
allied health service records and over 6 million hospital
records for a treatment population of 310 000 veterans.
The DVA maintains a client file, which includes data on
gender, date of birth, date of death and family status.
Medicines are coded in the dataset according to the
World Health Organization (WHO) Anatomical Therapeutic Chemical (ATC) classification7 and the Schedule
of Pharmaceutical Benefits item codes.8 Hospitalisations
are coded according to the WHO International Classification of Diseases (ICD) classification.9
The ‘HMR’ group were those who met the eligibility criteria, had received an HMR, and had at least two dispensings of warfarin in the 6 months prior to the HMR. The
‘control’ group comprised veterans who met the eligibility criteria and who had an average of at least two dispensings of warfarin per 6 months between the first and
final dispensing during the study period (to identify regular warfarin use), had at least 3 months between their
first and final dispensings of warfarin (to identify longterm warfarin use) and had a first warfarin dispensing
date before 1 January 2009 (to allow adequate time for a
minimum of 6 months follow-up within the study
period).
Eligible veterans in the control group were randomly
allocated to an index month in the study period to match
the time of the HMR in the HMR group. Control group
veterans were only matched once in the study period.
Study design and data collection
A retrospective cohort study was undertaken to compare
the degree of INR control in veterans taking warfarin
who were exposed and not exposed to HMR. Eligible
veterans were initially identified and selected by the
DVA based on data from their patient database. To be eligible for inclusion into this study, veterans were
screened by the DVA to meet the following inclusion criteria: possess a Gold repatriation benefit card (eligibility
for full entitlements), dispensed warfarin during the
study period (1 January 2007 to 31 December 2009),
and residing at home (as opposed to a residential agedcare facility).
© 2016 Royal Australasian College of Physicians
Data handling and statistics
Data were analysed using SPSS 19.0 for Windows (IBM
Corporation, New York, NY, USA). Demographic variables were compared between the HMR and control
groups using the following methods: paired and
unpaired t tests were used for normally distributed continuous variables; the non-parametric Mann–Whitney
test was used for non-normal data. Categorical variables
were analysed using the Chi-squared test. Fisher’s exact
test was used when at least one of the variables had
fewer than five patients or events. Statistical significance
was set at P < 0.05.
289
Bereznicki et al.
The primary outcome was the percentage TTR, calculated using Rosendaal’s linear interpolation method,11
for the 6 months prior to the HMR or index date, compared to the 6 months following the HMR or index date.
An INR target range of 2.0–3.0 was assumed for this
analysis as specific diagnoses for the condition requiring
anticoagulation were not available for the majority of
patients and a range of 2.0–3.0 was considered to be
appropriate for most elderly patients. The literature suggests that patients in the community spend 50–60% of
their time within the target range. At a power of 80%
and statistical significance set at 0.05, a minimum of
75 patients analysed before and after HMR was required
to detect a 10% difference in the percentage TTR.
The composite incidence of major bleeding and major
thrombotic events resulting in hospitalisation occurring
within 6 months of the HMR or index date was a secondary outcome. The ICD codes used to determine
the primary diagnosis of hospitalisation associated with
a bleeding event or thrombotic event are available in
Table S1, Supporting Information.
Ethics
Ethical approval for this project was provided by the
DVA Human Research Ethics Committee (Reference
No. E009-010) and the Tasmanian Health and Medical
Human Research Ethics Committee (Reference
No. H0010963) prior to the commencement of the
study.
Results
Figure 1 shows a flow diagram of recruitment of veterans into the study. The DVA selected 3884 veterans
according to the project methodology, from whom the
research team received 1213 replies (31.2% response
rate). A total of 1029 veterans who replied provided
their consent and was eligible for inclusion in the project. There was a total of 818 veterans who were allocated to the HMR (n = 281) or control groups (n = 537).
INR data were available for a total of 344 of 818 (42.1%)
veterans. At least two INR results were required to allow
calculation of the TTR; veterans with only one INR result
in the specified timeframe were excluded from the
respective analyses. A total of 321 veterans had at least
two INR results within the 12-month study period;
265 veterans had two or more INR results recorded in
the 6-month baseline period prior to their first HMR or
index date, 279 veterans had two or more INR results
recorded in the 6-month period following their first
HMR or index date and 229 veterans had two or more
INR results in each 6-month period.
290
Characteristics of the groups are shown in Table 1. The
groups were well matched with respect to gender, prior
hospitalisations, prior bleeding and thrombotic events
and region. The median number of co-morbidities was
statistically significantly higher in the HMR group. The
median age was also 1 year older in the HMR group,
which was of marginal statistical significance.
In the overall study cohort, the median testing interval
was approximately 16 days (range 1.0–65.7) and the
mean TTR was 64.0 27.3% (n = 321). The proportion
of veterans whose percentage TTR was >60% and >70%
were 64.5 and 49.2% respectively. The mean percentage
TTR following HMR and index date was 63.0 30.1%
(n = 98) and 67.0 27.7% (n = 181) respectively
(P = 0.27). There was no significant change in the TTR
in either of the groups or the overall veteran cohort in
the period following the HMR or index date from the 6month baseline period (Table 2).
Veterans living in outer regional and remote areas had
significantly poorer INR control than those living in
inner regional areas and major cities (mean TTR
49.9 30.7% vs 65.3 30.7% for those living in outer
regional/remote areas and those living in inner regional
areas/major cities respectively; P < 0.01).
There was no change in the combined number of
bleeding and thrombotic events leading to hospitalisation
(4/281 (1.4%) in the HMR group vs 6/537 (1.1%) compared with the control group; P = 0.74). In the HMR
group, there was no significant change in the combined
number of bleeding and thrombotic events leading to
hospitalisation before and after the HMR (1/281 (0.4%)
vs 3/281 (1.1%) respectively; P = 0.62).
Discussion
The main finding of this study was that HMRs were not
associated with a change in INR control, as estimated by
the TTR. The degree of INR control in this study (mean
TTR 64%) compares well with the mean TTR achieved
in recent randomised trials comparing warfarin to the
new anticoagulants dabigatran (64%), rivaroxaban
(55%) and apixaban (62%)12–14 in a much older population, and exceeds the usual level of INR control achieved
in the primary care setting in many countries.15
A systematic review reported that INR control differed
based on study site in community-based studies, anticoagulation clinics and RCTs; mean TTR was 56.7, 65.6 and
66.4% respectively.15 In the literature, appropriate TTR
benchmarks for patients taking warfarin are suggested to
be 60–70%.16 In a study comparing the outcomes of
patients randomised to dual antiplatelet therapy or warfarin, the benefits of warfarin were predicted to be lost
using a population model when the TTR fell below
© 2016 Royal Australasian College of Physicians
Medication reviews and INR control
3,884 veterans dispensed
p
warfarin
wa
(1,942
non-HMR)
4 HMR, 1,942 non-H
42
HMR)
2,671 did not respond
1 213 responded
1,213
d d
184 did not
n meet eligibility criteria:
C
- Consent
received following
closing
date (43)
cl
V
- Veteran
deceased (18)
D
- Declined
consent (123)
1,029 eligible
211 excluded
exclu
as warfarin commenced
following HMR or index date
818 incl
iincluded
luded
d d
n = 281 (HMR group))
n = 537 (Control)
474 veter
veterans’ INR data unavailable
344 veterans
e
erans
with
ith INR
INR data
d t available
avvailable
n = 121 (HMR group))
groupp)
n = 223 (Control group)
23 veterans
vetera with only a single INR
result ava
available excluded from INR
analysis
321 veterans
t
terans
with useable INR
IN
NR data
n = 115 (HMR group))
group
n = 206 (Control group)
Figure 1 Patient recruitment flow diagram.
58%.2 If we accept the TTR benchmarks that have
emerged internationally since our study, a TTR of 60%
as a lower benchmark for acceptable INR control, around
35% of veterans in our study were below this figure. If
upper benchmark of 70% is used, 51% of veterans were
below this mark.
There is a strong correlation between TTR and clinical
outcomes for patients taking warfarin.3,17–19 The generalisability of the trial results comparing new anticoagulants to warfarin depends to a large extent on the TTR
achieved in the trials, as it has been established that the
efficacy, safety and cost-effectiveness of comparators to
warfarin changes depend on the quality of INR control.20,21 In the RE-LY trial, a 10% increase in TTR independently predicted a 20% lower rate (P < 0.001) of the
composite clinical outcome (stroke, systemic embolism
© 2016 Royal Australasian College of Physicians
or major haemorrhage).19 The number of clinical events
in this study was too low to enable a statistical comparison of the clinical event rate and the degree of INR
control.
Noting that the study was underpowered to detect differences in hospitalisation for major bleeding and thrombosis, we did not identify any effect of HMR on hospital
admission resulting from complications associated with
warfarin therapy in veterans who had been taking warfarin for a period of at least 6 months.
Two studies in Australia involving a combination of a
series of medication review and point-of-care INR monitoring have found that this combination reduces the risk
of complications with warfarin therapy in the early postdischarge period.22,23 However, most of the beneficial
effect of these interventions was due to a reduction in
291
Bereznicki et al.
Table 1 Patient characteristics
HMR group (n = 281)
Control group (n = 537)
181 (64.4)
84.0 (56.0–93.0)
9.0 (1.0–28.0)
349 (65.0)
83.0 (41.0–94.0)
7.0 (1.0–30.0)
200 (71.2)
44 (15.7)
37 (13.2)
3 (1.1)
390 (72.6)
96 (17.8)
51 (9.5)
8 (1.5)
1 (0.4)
2 (0.7)
2 (0.4)
6 (1.1)
Male gender (%)
Median age (range) (years)
Median number of co-morbidities (range)
Prior hospitalisations†
0 (%)
1 (%)
2 or more (%)
Prior hospitalisation for a bleeding or thrombotic event
(%)†
Prior hospitalisation for a bleeding event (%)†
Prior hospitalisation for a thrombotic event (%)†
Region‡
Major city (%)
Inner regional (%)
Outer regional (%)
Remote (%)
Mean percentage TTR (SD)§
Median number of tests (range)§
Median testing interval (range) (days)§
175 (62.3)
79 (28.1)
23 (8.2)
3 (1.1)
64.8 (25.2)
8.0 (2.0–30.0)
12.6 (1.0–76.0)
347 (64.6)
142 (26.4)
39 (7.3)
6 (1.1)
68.3 (28.4)
7.0 (2.0–52.0)
17.4 (1.0–70.5)
P-value
0.87
0.047
0.003
0.24
0.76
0.99
0.72
0.56
0.32
0.33
0.002
†In the 6 months prior to the HMR (HMR group) or index date (control group). ‡The region was unknown for four veterans. §Veterans with only one
international normalised ratio value recorded for the 6 months prior to the HMR or index date were excluded (n = 94 for the HMR group and n = 171
for the control group). HMR, home medicines review; TTR, time in therapeutic range.
Table 2 Percentage TTR following HMR or index date†
HMR group (n = 78)
Control group (n = 151)
Overall (n = 229)
Before
After
Mean difference
67.9 23.3
70.4 26.1
69.6 25.1
69.6 25.2
68.5 26.1
68.9 25.8
+1.7
−1.9
+0.7
P-Value
0.63
0.40
0.72
†Only veterans with more than one international normalised ratio value available for both the 6 months before and the 6 months following the HMR
or index date were included. HMR, home medicines review; TTR, time in therapeutic range
minor bleeding, and not events that resulted in hospitalisation. In the more recent study,23 the intervention was
associated with significantly decreased rates of combined
major and minor haemorrhagic events to day 90 compared to usual care. However, there were no significant
differences in readmission and death rates in INR control
between the groups. Furthermore, significant reductions
in complications associated with warfarin therapy only
occurred in the group of patients newly initiating, rather
than continuing, warfarin therapy.
It is therefore difficult to extrapolate the benefits of
these interventions involving multiple home visits to the
effectiveness of a single HMR on people who are relatively stable on warfarin therapy (vs those recently discharged from hospital and/or recently initiated on
warfarin). Additionally, the focus of the medication
review in the studies was directly on warfarin, while this
may often not be the reason for the initiation of a HMR
in people who are stabilised on warfarin therapy. In a
retrospective study by Roughead et al, using
292
administrative claims data from the DVA, the effect of a
single HMR on Australian veterans and war widows
65 years and older who were taking warfarin was investigated.6 The study identified a 79% reduction in the
likelihood of hospitalisation for bleeding between 2 and
6 months following the HMR, but as this analysis did not
include INR data, it was unclear whether the HMR influenced INR control. The results of the present study,
albeit in a smaller subset of the veteran population, suggest that the reduction in hospitalisation occurred independently of improved INR control.
In order to obtain INR histories for patients included
in the present study, the investigators were required to
obtain consent from veterans. This meant that it was not
possible to obtain information from veterans who had
died either during the study period or in the time following the study period prior to the DVA mailout, and
excluded veterans who were readmitted to hospital with
a major bleeding or thrombotic event and subsequently
died. Therefore, we were not able to investigate the data
© 2016 Royal Australasian College of Physicians
Medication reviews and INR control
of the entire veteran cohort who were taking warfarin
during the study period. It is possible that the more seriously ill veterans and perhaps those most likely to suffer
from adverse events related to warfarin were not
included as a result of this methodology. This may have
underestimated any influence that HMR may have had
on the clinical outcomes of warfarin therapy or on INR
control. The low rate of major bleeding in this study
(equivalent to 2% p.a.) might be explained by careful
selection of veterans who are candidates for warfarin by
prescribers, the relatively high standard of INR control,
the exclusion of veterans who suffered major bleeding
events and subsequently died (due to the nature of study
methodology) or a combination of these factors.
Limitations
There were several methodological limitations to the
study, which have largely been previously discussed. It
should be acknowledged that while the percentage TTR
is strongly associated with bleeding and thromboembolism in people taking warfarin, TTR is a surrogate measure of these clinical outcomes. Additionally, the data
available from the DVA were limited in respect to the
documentation of co-morbidities, which meant that it
was impossible to determine the indication for warfarin
from the data available. Therefore, a target INR of
2.0–3.0 was assumed for all veterans. It is likely that this
would have resulted in an under-estimation of the TTR
rather than an over-estimation of the degree of INR control. INR data were only available for approximately
40% of the included veterans. In some cases, the pathology provider did not comply with the joint request from
the research team, the DVA and the veteran for the data
to be released. In other cases the pathology provider only
held a proportion of the INR data available; the veteran
may have changed provider or office-based INR testing
was used (in which case it was not available to the
pathology provider). It was not possible to identify which
veterans may have received office-based point of care
References
1 Oden A, Fahlen M. Oral
anticoagulation and risk of death: a
medical record linkage study. BMJ 2002;
325: 1073–5.
2 Connolly SJ, Pogue J, Eikelboom J,
Flaker G, Commerford P, Franzosi MG
et al. Benefit of oral anticoagulant over
antiplatelet therapy in atrial fibrillation
depends on the quality of international
normalized ratio control achieved by
centers and countries as measured by
© 2016 Royal Australasian College of Physicians
INR testing during the study period and it is therefore
unknown whether their INR control is comparable to
that of the veterans included in this study.
Conclusions
The overall level of INR control in the veterans participating in the study was good, and comparable to that
achieved in RCT, which generally involve a younger,
healthier cohort. However, there appear to be a relatively large group of patients who would benefit from
interventions to improve their INR control. HMR did not
appear to influence the INR control of veterans whose
pathology data were available for analysis. The previously reported effect of pharmacist-led medication
review on reducing hospital admission is likely to be due
to other factors. It is clear there is an ongoing need to
audit regularly INR control in veterans taking warfarin
and intervene as appropriate to maximise the benefits
and minimise the risks of warfarin therapy.
Acknowledgements
This research was supported by the Australian Government Department of Veterans’ Affairs. The Department
of Veterans’ Affairs reviewed the content. The authors
acknowledge the support and assistance of staff who
assisted with this project. In particular, we thank Mimi
van Duren (Primary Care Policy Support, Department of
Veterans’ Affairs), Francene Benness (Assistant Director,
Pharmacy and Health Technology Policy, Department of
Veterans’ Affairs) and Will Hanham (Director, Pharmacy
and Health Technology Policy, Department of Veterans’
Affairs) for their commitment and assistance throughout
the project. We also thank the veterans who consented
to be involved in the project and those who contacted us
but were unfortunately not able to participate. We also
thank the pathology providers who assisted us by providing INR data.
time in therapeutic range. Circulation
2008; 118: 2029–37.
3 White HD, Gruber M, Feyzi J, Kaatz S,
Tse HF, Husted S et al. Comparison of
outcomes among patients randomized
to warfarin therapy according to
anticoagulant control. Arch Intern Med
2007; 167: 239–45.
4 Gilbert AL, Roughead EE, Beilby J, Mott
K, Barratt JD. Collaborative medication
management services: improving
patient care. Med J Aust 2002; 177:
189–92.
5 Krass I, Smith C. Impact of medication
regimen reviews performed by
community pharmacists for ambulatory
patients through liaison with general
medical practitioners. Int J Pharm Pract
2000; 8: 111–20.
6 Roughead EE, Barratt JD, Ramsay E,
Pratt N, Ryan P, Peck R et al.
Collaborative home medicines
review delays time to next
hospitalization for warfarin
associated bleeding in Australian
293
Bereznicki et al.
7
8
9
10
11
12
13
war veterans. J Clin Pharm Ther 2010;
36: 27–32.
World Health Organization
Collaborating Centre for Drug Statistics
Methodology. Anatomical Therapeutic
Chemical Code Classification Index with
Defined Daily Doses. 2008 [cited
2015Jul 2]. Available from URL: http://
www.whocc.no/atc_ddd_index/
Australian Government Department of
Health and Ageing. Schedule of
Pharmaceutical Benefits. [cited 2015 Jul
2]. Available from URL: http://www.
pbs.gov.au
World Health Organization.
International Statistical Classification of
Diseases and Related Health Problems:
10th Revision. 2002. [cited 2015 Jul 2].
Available from URL: http://www.who.
int/classifications/icd/en/
Australian Institute of Health and
Welfare. Rural, Regional and Remote
Health: A Guide to Remoteness
Classifications. 2004. [cited 2015 Jul 2].
Available from URL: http://www.aihw.
gov.au/rural-health-remotenessclassifications/
Rosendaal FR, Cannegieter SC, van der
Meer FJ, Briët E. A method to
determine the optimal intensity of oral
anticoagulant therapy. Thromb Haemost
1993; 69: 236–9.
Connolly SJ, Ezekowitz MD, Yusuf S,
Eikelboom J, Oldgren J, Parekh A et al.
Dabigatran versus warfarin in patients
with atrial fibrillation. N Engl J Med
2009; 361: 1139–51.
Granger CB, Alexander JH,
McMurray JJ, Lopes RD, Hylek EM,
14
15
16
17
18
Hanna M et al. Apixaban versus
warfarin in patients with atrial
fibrillation. N Engl J Med 2011; 365:
981–92.
Patel MR, Mahaffey KW, Garg J, Pan G,
Singer DE, Hacke W et al. Rivaroxaban
versus warfarin in nonvalvular atrial
fibrillation. N Engl J Med 2011; 365:
883–91.
van Walraven C, Jennings A, Oake N,
Fergusson D, Forster AJ. Effect of study
setting on anticoagulation control: a
systematic review and metaregression.
Chest 2006; 129: 1155–66.
Camm AJ, Lip GY, De Caterina R,
Savelieva I, Atar D, Hohnloser SH et al.
2012 focused update of the ESC
Guidelines for the management of atrial
fibrillation: an update of the 2010 ESC
Guidelines for the management of atrial
fibrillation. Eur Heart J 2012; 33:
2719–47.
Veeger NJ, Piersma-Wichers M,
Tijssen JG, Hillege HL, van der Meer J.
Individual time within target range in
patients treated with vitamin k
antagonists: main determinant of
quality of anticoagulation and predictor
of clinical outcome—a retrospective
study of 2300 consecutive patients with
venous thromboembolism. Br J
Haematol 2005; 128: 513–9.
Wan Y, Heneghan C, Perera R, Roberts
N, Hollowell J, Glasziou P et al.
Anticoagulation control and prediction
of adverse events in patients with atrial
fibrillation: a systematic review. Circ
Cardiovasc Qual Outcomes 2008; 1: 84–91.
19 Van Spall HG, Wallentin L, Yusuf S,
Eikelboom JW, Nieuwlaat R, Yang S
et al. Variation in warfarin dose
adjustment practice is responsible for
differences in the quality of
anticoagulation control between centers
and countries: an analysis of patients
receiving warfarin in the randomized
evaluation of long-term anticoagulation
therapy (RE-LY) trial. Circulation 2012;
126: 2309–16.
20 Wallentin L, Yusuf S, Ezekowitz MD,
Alings M, Flather M, Franzosi MG et al.
Efficacy and safety of dabigatran
compared with warfarin at different
levels of international normalised
ratio control for stroke prevention
in atrial fibrillation: an analysis of
the RE-LY trial. Lancet 2010; 376:
975–83.
21 Shah SV, Gage BF. Cost-effectiveness of
dabigatran for stroke prophylaxis in
atrial fibrillation. Circulation 2011; 123:
2562–70.
22 Jackson SL, Peterson GM, Vial JH, Jupe
DM. Improving the outcomes of
anticoagulation: an evaluation of home
follow-up of warfarin initiation. J Intern
Med 2004; 256: 137–44.
23 Stafford L, Peterson GM, Bereznicki LR,
Jackson SL, van Tienen EC, Angley MT
et al. Clinical outcomes of a
collaborative, home-based
postdischarge warfarin management
service. Ann Pharmacother 2011; 45:
325–34.
Supporting Information
Additional supporting information may be found in the online version of this article at the publisher’s web-site:
Table S1 ICD-10 primary diagnosis codes used to identify hospitalisation due to haemorrhagic and thromboembolic
events.
294
© 2016 Royal Australasian College of Physicians
Do community hospice programmes reduce hospitalisation
rate in patients with advanced chronic obstructive pulmonary
disease?
S. P. M. Iupati1 and B. R. Ensor2
1
Te Omanga Hospice, and 2Mary Potter Hospice, Wellington, New Zealand
Key words
COPD, hospices, hospitalisation, terminal
care, utilisation.
Correspondence
Salina Iupati, Te Omanga Hospice, PO Box 30
814, Lower Hutt 5040, New Zealand.
Email: [email protected]
Received 3 June 2015; accepted 3 November
2015.
doi:10.1111/imj.12947
Abstract
Background: Since Hinton first published his observations on the distress of patients dying
on a medical ward in 1963, there has been increasing awareness of the palliative care needs
in patients who have non malignant diseases. Patients with advanced chronic obstructive
pulmonary disease (COPD) are known to have comparable symptom burden to lung cancer
patients and are more likely receive invasive treatment at the end of life than patients with
end stage lung cancer. They are also less likely to receive hospice services, and the benefit of
such programmes in this key group of patients remain largely unknown, in particular what
effect hospice programmes have on hospitalisation.
Aims: (i) To examine any effect of community hospice programmes on hospitalisation in
patients with advanced COPD. (ii) To identify any association between utilisation of specific
hospice services with hospitalisation. (iii) To describe key peri-mortem outcomes.
Methods: This was a retrospective study of consecutive patients with COPD admitted into
community hospice programmes in the greater Wellington region, New Zealand between 1
October 2007 and 31 October 2013.
Results: A mean decrease of 2.375 (median decrease of 2; 95% confidence interval 1, 3)
hospital admissions over a 12-month period was found after admission into hospice
programme (P < 0.0005).
Conclusion: Community hospice programmes may be associated with reduction in
hospitalisation in patients with advanced COPD.
Introduction
Chronic obstructive pulmonary disease (COPD) is a major
cause of disability, hospital admission and premature death
in Australia and New Zealand,1 and the cost of hospital care
for people with COPD is high.2 Factors that have been
shown to be associated with healthcare use, such as hospital
admissions are: previous hospital admissions,3 dyspnoea4
and symptom cluster.5
Several international guidelines have highlighted the importance of palliative care in the management of patients with
advanced COPD.1,6,7 The benefit of specialist palliative care in
these patients is being explored in the literature; in particular,
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
it remains largely unknown whether these programmes have
an effect on hospitalisation rates in these patients.8
End-of-life use of medications from this study will add to
the wider literature as there is no acceptable dosage agreed
currently. This is of particular importance when the use palliative sedation and hastening of death is in question.
Methods
Patients
Consecutive patients who were enrolled with Mary Potter
Hospice and Te Omanga Hospice Programmes with a
primary diagnosis of COPD between 1 October 2007 and
31 October 2013 were identified. Patients with concomitant
cancer diagnosis were excluded. Health and Disability
Ethics Committee online questionnaire was completed.
Written permissions were obtained from the Ethics
295
S. P. M. Iupati et al.
Committee of Mary Potter Hospice, Te Omanga Hospice,
Capital and Coast District Health Board and Hutt Valley District Health Board prior to data collection.
Programme outline
The combined serviced population of Mary Potter Hospice
and Te Omanga hospice is 430 000, each hospice covering
different geographical areas of the greater Wellington region. Both hospices provide in-patient-unit beds and community programmes, and are both staffed by specialist
palliative care physicians and palliative care nurses. At
Mary Potter hospice, each patient is designated to a palliative care coordinator (nurse) whose role is primarily that
of case management during office hours with phone advice
out of hours and working with other community services,
such as primary care, district nursing services and ambulance service, all of whom contribute to out of hours care
in the home. Each patient is seen by a palliative care physician as required after enrolment on designated ‘clinic’ days.
There are two community programmes at Te Omanga
Hospice, General Practitioner (GP) Led and Hospice Led. Patients’ needs and their GP’s wishes and availability are the
determinants for which programme the patient is to be enrolled in. Patients can move between the two programmes
and patients at the end of life are usually on the Hospice
Led programme with very few exceptions. Patients on hospice led programmes have 24-h access to both nursing and
medical team advice and consultations. Patients on GP led
programmes are case managed by a designated palliative
care nurse and they have 24-h access to hospice nursing advice and visits. Palliative care physicians’ consultations are
on request by their GPs or palliative care nurses.
Referrals were accepted if patients have a prognosis of
12 months or less and they have complex physical or
psychosocial palliative needs.
There were no acceptance criteria specific to COPD
during the study period.
visits, allied health engagement, number of in-patient-unit (IPU) admissions and reasons for admissions.
References to resuscitation status or ceiling of medical
care for future exacerbations were noted.
Numbers of all hospital admissions 12 months before
enrolment and the entire follow-up period up to 31 October
2013 were collected. For patients who survived for
12 months or more on the programme, the total number
of hospitalisations up to 12 months after admission into
hospice programme was also collected for comparison.
Key peri-mortem data includes survival on the programme,
places of death and end-of-life medication use. Only deidentified data were collected on the data form during the
chart review.
Statistical analysis
Data were analysed using RStudio statistical software
v0.98.1103. The number for all hospital admissions for
patients before and after they were admitted onto the programme was analysed. For patients who survived on the
programme for 12 months of more, we used a paired sample t test to compare the number of hospital admissions
12 months before and after patients’ enrolment at the hospice. The 95% confidence interval for the difference in median was calculated using an empirical bootstrap method.
Pearson’s product moment correlation coefficients were
calculated to determine any relationship between the number of all hospital admissions after enrolment at the hospice
(follow-up period until 31 October 2013) and number of
IPU admissions, nursing visits and medical visits separately.
Welch two sample t test was used to determine any difference in hospital admissions between groups 1: with or without ceiling-of-care discussion; 2: with or without moving to
higher level of care.
Results
Baseline patients’ characteristics
Data collected
Data were retrospectively collected from electronic
clinical records at the two hospices, Capital and Coast
and Hutt Valley DHB. This comprises of patients’
demographics, baseline COPD characteristics and comorbidity. Outpatient management by specialist respiratory services (respiratory specialists and/or clinical
nurse specialists) were noted. Patients who were on
pulmonary rehabilitation programme only were not
counted as receiving respiratory specialist services in
this study. Reasons for referral to hospice service were
identified, and service utilisation data were collected
from patients’ enrolment to death or discharge. These
include: total number of hospice doctors and nurses’
296
A total of 73 patients with a primary diagnosis of COPD was
identified. Baseline patients’ demographics and characteristics are summarised in Table 1.
Reasons for referral were documented in 67 patients, and
these include: symptom control (n = 46), patient did not
wish to return to hospital (n = 9), end-of-life care (EOLC)
(n = 10) and others (n = 2).
At the time of first assessment, the levels of support
for all patients were documented, and these include:
‘Lives with others who provide support’ (n = 50 68%),
‘Lives alone with external professional support’ (n = 7,
9%), ‘Lives at home with others with external professional support’ (n = 6, 8%), ‘Lives alone with external
non professional support’ (n = 2, 3%), ‘Lives alone with
© 2016 Royal Australasian College of Physicians
Hospice and hospitalisation in COPD
Table 1 Baseline patients’ characteristics
Patients
n = 73
Age (range)
Mean 76.1 years (50–91)
Sex
Male 37 (51%)
Female 36 (49%)
Ethnicity
European 64 (87.7%)
Maori 5 (6.8%)
PI 3 (4.1%)
Asian 1 (1.4%)
FEV1
Median 30.6% (14–59%)
MRCDS (Medical Research Council Dyspnoea Score)
Median 5 (3–5)
Domiciliary oxygen
n = 33 (44.6%)
Significant co-morbidity (≥1)
n = 40 (55%)
Congestive heart failure (CHF)
n = 19
Ischaemic heart disease (IHD)
n = 14
CHF and IHD
n=2
Others
n=5
NIV (within 12 months prior to enrolment)
n = 11 (15%)
Specialist respiratory services
n = 56 (76.7%)
At time of enrolment
n = 42 (57.5%)
Within 12 months prior
n = 14 (19.2%)
no care/support’ (n = 4, 5%), ‘rest home level care’
(n = 4, 5%) and ‘hospital level care’ (n = 1,1%).
Results for baseline medication use in COPD were
collected for all patients: short-acting bronchodilators
93% (n = 68), inhaled corticosteroid 88% (n = 64), long
acting beta-agonists 85% (n = 62), long acting muscarinic antagonists 66% (n = 48) and long-term systemic
steroid 19% (n = 14).
Service utilisation
The mean number of hospice doctors’ visits and nurses
visits was 2.2 (median = 1) and 20 (median = 12) respectively. There was a total of 97 IPU admissions over the 6year period. Forty-three patients utilised in-patient service.
The mean number of admission was 1.3 per patient (median = 1). The most common reason for admission was respite (n = 45), followed by symptom control (n = 25),
EOLC (n = 17) and others (n = 4).
Fifty-seven out of 73 patients were noted to have received services from members of the hospice allied health.
hospital admissions over a 12-month period prior to enrolment was 3.6 (median 3). There were 31 patients who
remained on the hospice programme for 12 months or
more. We used a paired sample t test to determine whether
the number of hospital admissions before patients’ enrolment at the hospice differed from the number of admissions
post hospice enrolment. Following admission into the hospice programme, a mean decrease of 2.375 (median 2; 95%
confidence interval 1, 3) hospital admissions over a 12month period was found. The difference was highly statistically significant, t (31) = 4.3813, P < 0.0005.
Pearson’s product moment correlation coefficients
(95% confidence interval) for the number of all hospital admissions after enrolment with IPU admissions,
nursing visits and medical visits were calculated, and
no correlation between hospital admission and the
number of IPU admissions, nursing visits or medical
visits were found in all patients and in the subgroup
of those who survived for more than 12 months.
The mean number of all hospital admissions after
enrolment in patients with and without documented
ceiling-of-care discussion was 2.0 and 1.63 respectively. The difference between the two groups was
not statistically significant (P-value = 0.5386).
Similarly the differences in the number of admissions
12 months prior and after enrolment for patients who
moved to higher level of care and those who did not were
2.67 and 2.32 respectively (P-value 0.77).
Peri-mortem outcomes
The number of patients who died or were discharged from
the programme was 64 (87.7%) and 5 (6.8%) respectively.
The median survival on the hospice programme was
8 months (range 1 day to 45 months). Place of death is
summarised in Table 2. There were no hospital deaths
amongst patients who were referred for EOLC and those
who did not wish to return to hospital.
Use of regular medications around the time of death was
known in 46 (62%) patients and is summarised in Table 3.
The median and mean regular total 24-h oral morphine
equivalent dose was 25 mg and 41 mg respectively (range:
Table 2 Place of death
Hospital admissions
Place of death
Number of patient (n = 64)
The total number of hospital admissions within the 12month period before and after enrolment for the 73 patients was 267 and 97 respectively. The latter includes three
of the 10 patients who did not wish to return to hospital on
referral to the hospice programme. The mean number of all
Home
15 (23.4%)
© 2016 Royal Australasian College of Physicians
Hospice
22 (34.4%)
Hospital
12 (18.8%)
Residential care facilities
15 (23.4%)
297
S. P. M. Iupati et al.
Table 3 End-of-life use of medications
Use of:
Number of patients (n = 46)
Syringe driver
23 (50%)
Regular opioid
33 (72%)
Regular benzodiazepine
25 (54%)
Antibiotics
23 (50%)
Systemic steroid
24 (52%)
0–240 mg). The median and mean regular total 24-h subcutaneous midazolam equivalent dose was 3.75 mg and 6 mg
respectively (range: 0–30 mg). PRN use of opioids and benzodiazepines was documented in 25 (34%) patients. The
median and mean 24-h PRN oral morphine equivalent dose
was 7.5 mg and 11 mg respectively (range 0–45 mg). The
median and mean 24-h PRN subcutaneous midazolam
equivalent dose was 5 mg and 6.12 mg respectively (range
0–38 mg).
Discussion
The number of patients aged ≥40 years with COPD in the
population captured by the two hospices is estimated to be
28 000.9,10 With the average of 12 patients per year, only
a fraction of these patients received hospice care during
the study period. This may be because of difficulty in prognostication resulting in physicians’ reluctance to discuss
end-of-life issues, advance care planning or hospice referral.11,12 Alternatively patients may not want to commit
themselves to a possibly dying trajectory.13
Fifty percent of patients were in the severe COPD category according to Global Initiative for Chronic Obstructive
Lung Disease classification14 and yet the Medical Research
Council Dyspnoea Score indicates much higher symptom
burden than what the forced expiratory volume in 1 s
(FEV1) measurements otherwise suggest. This is consistent
with findings of other studies that showed a poor correlation between physiological measurements, such as FEV1
and severity of symptoms.,15,16 Fifty five percent of patients
have at least one co-morbidity which may account for some
of the symptom burden of breathlessness and the lower
than expected median survival observed if based on FEV1
alone.17
Results for baseline medications indicate the use of
long acting muscarinic antagonists could be improved
prior to referral. One possible explanation could be
the requirement of application of special funding for
these medications by providers in New Zealand.
298
Patients were most commonly referred from their GP and
general physicians. Respiratory services were either involved at the time of or within 12-month period prior to referral in 76.7% of patients, which may imply patients
referred having refractory symptoms despite management
of the underlying disease by specialists. It is consistent with
the finding of symptom management being most common
reason for referral. Another underlying reason for referral
not overtly identified could be carers’ stress given that
68% of patients were living at home with others. Unfortunately it was not possible to identify the actual triggers for referral because of lack of details in the referral data.
To the authors’ knowledge, this is the first study that describes hospice services utilisation for patients in New
Zealand. The need for respite was found to be high, and this
result is consistent with qualitative studies which
highlighted the burden on caregivers.18,19
The median number of hospital admissions per patient
over a 12-month period prior to enrolment into hospice
programme was 3 (mean 3.6) and is consistent with New
Zealand Palliative Council referral criteria.20 A decrease of
2.375 hospital admissions over a 12-month period was
found for patients who remained on the programme for
12 months or longer. The 12-month period was selected
to minimise confounding effect from seasonal variations
in COPD exacerbations. The reduction is both highly statistically and clinically significant and could be attributable to
hospice intervention. Results of other studies have been
mixed. Young and Simpson observed a substantial drop in
Emergency room visits and hospitalisations in patients with
COPD who received home-based education and advanced
care planning.21 In contrast, a randomised controlled study
in COPD and chronic heart failure patients who received
domiciliary palliative care service, while exhibited better
outcomes on self management of illness and improved
symptom control but failed to show a reduction in emergency department utilisation rate. Hospitalisation rate was
not analysed in that study.22 Results of studies that examined the cost effectiveness of hospice programmes in nonmalignant conditions have been mixed.22,23
One may argue the inclusion of patients who did not
wish to return to hospital or were referred for EOLC for
analysis could affect the results in hospitalisation. The effect
of the latter group is offset by analysis of the subset of patients who survived for ≥12 months. As this study is an
attempt to explore the effect of a community intervention
in helping patients to stay in the community regardless of
the reasons for referral and in providing an alternative to
admission to hospital when they no longer were able to
cope at home particularly around the time of an exacerbation, it is paramount such patients are not excluded from
analysis as they constitute a significant group hospices generally cater for. Moreover patients’ wishes may change with
© 2016 Royal Australasian College of Physicians
Hospice and hospitalisation in COPD
time as found in this study where 3 of such patients
returned to hospital after enrolment.
No single component of hospice intervention alone was
significantly associated with the reduction of hospital admissions, which could be because of the small study population. It may be that the broader psychosocial support
makes a significant contribution that cannot be quantified
in terms of medical or nursing interventions.24 Adequately
powered controlled prospective studies are required to confirm the result of reduction in hospitalisation observed in
this study and to identify determinants or specific hospice
intervention(s) associated with hospitalisation. Results of a
prospective study measuring outcomes of a communitybased integrated palliative care programme for patients
with COPD were inconclusive because of poor patient enrolment and incomplete data collection8 highlighting recruitment challenges in this area.
Our peri-mortem data help to fill the gap in current literature about important palliative care outcomes in advanced
COPD. The mean and median survival on the hospice programme was 12 and 8 months respectively. In contrast, a
subgroup analysis of COPD patients in an American hospice
study reports a median survival of only 76.5 days.25 Almost
half of the patients in this study died at their usual abode
(25% at home and 23% at residential care facilities). Only
18% died in hospital, and by comparison a rate of 72%
was reported in an Australian non hospice cohort.26 A potential explanation being patients who accept hospice care may
also be more inclined to forego life-prolonging measures that
can only be offered in hospitals. Future studies are required
to confirm whether hospice care could reduce deaths occurring in hospitals in patients with end-stage COPD.
A significant influence of palliative care is the introduction of opioid for dyspnoea and benzodiazepines for dyspnoea related anxiety. The role of opioids and
benzodiazepines in severe COPD has been extensively examined in the literature,27–30 and results of this study serve
to describe their use outside research protocol. Furthermore
there remains a paucity of end-of-life data.
The median 24-h oral morphine equivalent found in
this study was 25 mg which is similar to the highest effective dosage of 30 mg demonstrated in the only dose
ranging study published to date.31 Given that many patients are receiving greater than 30 mg, further consideration of whether this is being used appropriately for
pain control, or for other reasons is warranted. Similar
dosage has also been reported in studies of patients
References
1 The COPD-X Plan: Australian and New
Zealand Guidelines for the management
© 2016 Royal Australasian College of Physicians
with motor neuron disease where the median dose
for oral regular morphine 24 h before death ranged between 20 mg–30 mg.32,33
Unlike opioid, the benefit of benzodiazepine in alleviating dyspnoea is not established, and its use in severe COPD
has been associated with increased mortality.30 The median
midazolam dose found in this study was 3.75 mg/24 h
which is considerably less than 20 mg/24 h found in the
above mentioned motor neuron disease study.33
To our knowledge our study is the first to report on medication use in dying patients with COPD in New Zealand. Future benchmarking studies are important particularly when
the issue of palliative sedation or hastening of death is raised.
Conclusions and limitations
Community hospice programmes may be associated with a
reduction in hospitalisation. Most patients who did not wish
to return to hospital managed to avoid hospitalisation, and
there was no hospital death in this sub-group. No correlation was found between the numbers of hospital admissions after enrolment and utilisation of individual aspects
of hospice services, including IPU, nurses and doctors’ visits.
Neither did ceiling-of-care discussion nor moving patients
to higher level of care during the enrolment period appears
to have an effect on hospitalisation. This could be because of
the small sample size in this study. This is a retrospective
study, and the number of patients is very small and they
are highly selected. Furthermore this study is limited by
only two regional hospices being included, and
community-based programmes may vary significantly between hospices. Therefore one would be cautious to generalise results from this study to a wider group of patients with
COPD and other geographical areas. Further research in patients with advanced COPD receiving hospice care is
warranted.
Acknowledgements
We thank clerical staff at Te Omanga Hospice, Mary Potter
Hospice for their assistance for this study. We are grateful
for statistical advice from Dr Dalice Sim; School of
Mathematics, Statistics and Operations Research, Victoria
University and Dr James Stanley; Department of Public
Health, Wellington School of Medicine, University of
Otago.
of Chronic Obstructive Pulmonary Disease
2015 June 2015 version 2.42 ed: Australia
Lung Foundation, The Thoracic Society of
Australia and New Zealand 2015.
2 (National Institutes of Health NH, Lung and
Blood Institute). Morbidity and Mortality:
2009 Chart Book on Cardiovascular, Lung
and Blood Diseases. 2013 [updated 2013;
299
S. P. M. Iupati et al.
3
4
5
6
7
8
9
10
11
cited 2014 Sep 29]. Available from URL:
http://www.nhlbi.nih.gov/resources/docs/
2009_chartbook.pdf
Almagro P, Ochoa de Echaguen A,
Quintana S, Rodriguez Carballeria M,
Heredia J, Garau J. Risk factors for hospital
readmission in patients with chronic pulmonary disease. Respiration 2006; 73:
311–7.
Schembri S, Anderson W, Morant S, Winter J, Thompson P, Pettitt D et al. A
predictive model of hospitalization and
death from chronic obstructive pulmonary
disease. Respir Med 2009; 103: 1461–7.
Park SK, Larson JL. Symptom cluster,
healthcare use and mortality in patients
with severe chronic obstructive pulmonary
disease. J Clin Nurs 2014; 23: 2658–71.
Lanken P, Terry P, Delisser HM, Fahy B,
Hansen-Flaschen J, Heffner JE. An official
American Thoracic Society clinical policy
statement: palliative care for patients with
respiratory diseases and critical illnesses.
Am J Respir Crit Care Med 2008; 177:
912–27.
O’Donnell D, Aaron S, Bourbeau J,
Hernandez P, Marciniuk D, Balter M.
Canadian Thoracic Society recommendations for management of chronic
obstructive pulmonary disease, 2003. Can
Respir J 2003; 10(Suppl A): 11A–65A.
Horton R, Rocker G, Dale A, Young J,
Hernandez P, Sinuff T. Implementing a palliative care trial in advanced COPD: a
feasibility assessment (The COPD IMPACT
study). J Palliat Med 2013; 16: 67–73.
The New Zealand census of population and
dwelling. Statistic New Zealand; 2013 [updated 2013; cited 2014 Apr 18]; Available
from URL: http://www.stats.govt.nz/
Census/2013
Shirtcliffe P, Weatherall M, Marsh S,
Travers J, Hansell A, McNaughton A et al.
COPD prevalence in a random population
survey: a matter of definition. Eur Respir J
2007; 30: 232–9.
Thomas J, O’Leary J, Fried TR. Understanding their options: determinants of
hospice discussion for older persons with
advanced illness. J Gen Intern Med 2009;
24: 923–8.
300
12 Fitzsimons D, Mullan D, Wilson J. The
challenge of patients’ unmet palliative care
needs in the final stages of chronic illness.
Palliat Med 2007; 21: 313–22.
13 Macpherson A, Walshe C, O’Donnell V,
Vyas A. The views of patients with severe
chronic obstructive pulmonary disease on
advance care planning: a qualitative study.
Palliat Med 2012; 27: 265–72.
14 Rabe K, Hurd S, Anzueto A, Barnes P, Buist
S, Calverley P et al. Global strategy for the
diagnosis, management, and prevention
of chronic obstructive pulmonary disease;
GOLD executive summary. Am J Respir Crit
Care Med 2007; 176: 532–55.
15 Nishimura K, Izumi T, Tsukino M, Oga T.
Dyspnea is a better predictor of 5-year survival than airway obstruction in COPD.
Thorax 2002; 121: 1434–40.
16 Wedzicha J, Bestall J, Garrod R. Randomized controlled trial of pulmonary
rehabilitation in severe chronic obstructive
pulmonary disease, stratified with the
MRC dyspnoea scale. Eur Respir J 1998;
12: 363–9.
17 Celli B, Snider G, Heffner J. Standards for
the diagnosis and care of patients with
chronic obstructive pulmonary disease.
Am J Respir Crit Care Med 1995; 152:
S77–120.
18 Bailey P. Death stories: acute exacerbations
of chronic obstructive pulmonary disease.
Qual Health Res 2001; 11: 322–38.
19 Seamark D, Blake S, Seamark C, Halpin D.
Living with severe chronic obstructive pulmonary disease (COPD): perception of
patients and their carers. An interpretative
phenomenological analysis. Palliat Med
2004; 18: 619–25.
20 Referral criteria for adult palliative care services in New Zealand. In: Zealand PCCoN,
editor. Wellington; 2014.
21 Young J, Simpson A, Demmons J, Conrad
W, Rocker G. Evaluating the impacts of
“INSPIRED”, a new outreach program for
patients and families living with advanced
chronic obstructive pulmonary disease
(COPD). Am J Respir Crit Care Med 2012;
185: A3732.
of a randomized trial of the PhoenixCare Intervention:
programme
of
case
management and coordinated care for the seriously chronically ill. J Palliat Med 2006; 9:
111–26.
23 Lorenz K, Shugarman L, Lynn J. Health
care policy issues in end of life care. J Palliat
Med 2006; 9: 731–48.
24 Simpson AC, Rocker GM. Advanced
chronic obstructive pulmonary disease: rethinking models of care. Q J Med 2008;
101: 697–704.
25 Christakis N, Escarce J. Survival of medicare patients after enrolment in hospice
programmes. N Engl J Med 1996; 335:
172–8.
26 Philip J, Lowe A, Gold M, Brand C, Miller
B, Douglass J et al. Palliative care for
patients with chronic obstructive pulmonary disease: exploring the landscape.
Intern Med J 2012; 42: 1053–7.
27 Jennings A, Davies A, Higgins J, Gibbs J,
Broadley K. A systematic review of the
use of opioids in the management of dyspnoea. Thorax 2002; 57: 939–44.
28 Rocker G, Horton R, Currow D, Goodridge
D, Young J, Booth S. Palliation of dyspnea
in advanced COPD: revisiting a role for opioids. Thorax 2009; 64: 910–5.
29 Currow D, Ekstrom M, Abernathy A. Opioids for chronic refractory breathlessness:
right patient, right route? Drugs 2014; 74:
1–6.
30 Ekstrom M, Bornefalk-Hermansson A, Abernathy A, Currow D. Safety of
benzodiazepines and opioids in very severe
respiratory disease: national prospective
study. BMJ 2014; 348: 1–10.
31 Currow DSQ, Green A. The longitudinal
pattern of response when morphine is used
to treat chronic refractory dyspnoea. J
Palliat Med 2013; 16: 881–6.
32 O’Brien T, Kelly M, Saunders C. Motor
neurone disease: a hospice perspective.
BMJ 1992; 304: 471–3.
33 Ryan R, Tracey G, Lawlor P, O’Siorain L,
Higgins S. A retrospective review of specialist palliative care involvement in motor
neurone disease. Ir Med J 2012; 105: 335–7.
22 Aiken L, Butner J, Lockhart C, Volk-Craft B,
Hamilton G, Williams F. Outcome evaluation
© 2016 Royal Australasian College of Physicians
Alcohol use disorder hospitalisations over the last two
decades: a population-based cohort study
W. Liang and T. Chikritzhs
National Drug Research Institute, Curtin University of Technology, Perth, Western Australia, Australia
Key words
alcohol-related disorders, cohort studies,
hospitalisation, health status indicators.
Correspondence
Wenbin Liang, National Drug Research Institute,
Curtin University of Technology, GPO Box
U1987, Perth, WA 6845, Australia.
Email: [email protected]
Received 9 September 2015; accepted 3
December 2015.
doi:10.1111/imj.12980
Abstract
Background: Alcohol use disorders are risk factors for almost all health conditions due
to heavy alcohol use. The epidemiology of alcohol use disorders can be used to monitor
harm from heavy alcohol consumption.
Aim: To estimate changes in the risk of alcohol use disorders over the last two decades
among the Western Australian adult population.
Methods: This population-based cohort study used hospital separation records for
Western Australian residents aged 18 years and older that occurred between 1990 and
2013 with a primary diagnosis of alcohol use disorder and annual estimated residential
population to estimate the annual gender- and age-specific incidence rate. A random
sample of emergency presentations to public hospitals in Western Australia between
2002 and 2013 was used to account for confounding effects, such as changes in patient
access to medical care and overall improvement in healthcare service in the multivariable Poisson regression model.
Results: The risk of alcohol use disorder hospitalisations among the Western Australia
population has increased considerably since 1998 with a decline in 2012 and 2013. The
average rate remained significantly higher from 2010 to 2013 compared with previous
years.
Conclusions: The trend of alcohol use disorder hospitalisations is indicative of an
increase in harm due to heavy alcohol use in the population.
Introduction
Alcohol, a widely used psychoactive substance with
dependence-producing properties, is a major risk factor
for global mortality and morbidity.1,2 Alcohol has been
widely consumed in Australia for more than 200 years,3
and about 80% of the population currently drink alcohol.4 Health conditions that can be caused by alcohol
include various cancers, liver disease, injuries from falls,
physical assaults, road traffic crashes and mental health
disorders.2 Mental and behavioural disorders due to use
of alcohol (defined by the International Classification of
Diseases (ICD) developed by the World Health Organization)5 or alcohol-related disorders (defined by the Diagnostic and Statistical Manual of Mental Disorders
Funding: This work was supported by the Australian
Government Department of Health and Ageing under the
National Drug Strategy’s funding of the National Drug Research
Institute. This work was also supported by Healthway and the
Drug and Alcohol Office, Western Australia.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
developed by the American Psychiatric Association)6,7
are a cluster of neuropsychiatric conditions,5–7 and they
are direct contributors to the alcohol-related burden.8 In
2010, the prevalence of harmful use of alcohol and alcohol dependence combined among the population
15 years and older was estimated to be 4.1% worldwide
and 3.5% in Australia.2 To establish a diagnosis of an
alcohol-related disorder requires evidence of alcohol
use.5–7 For example, diagnostic criteria of alcohol use disorder includes ‘alcohol is often taken in larger amounts
or over a longer period than was intended’, whereas a
diagnosis of alcohol intoxication requires ‘clinically significant problematic behavioural or psychological
changes after recent ingestion of alcohol’. Alcohol use is
assessed based on self-reported data and/or objective
analysis of specimens, such as breath, blood or urine.5–7
Because of the connection with heavy alcohol use, the
epidemiology of alcohol-related disorders can therefore
be used to monitor the harmful effects of heavy alcohol
consumption8 including injuries and road traffic crashes;
alcohol-induced mental disorders, such as depression;
301
Liang & Chikritzhs
multiple organ damage, such as liver disease; hypertension; development of cancer; unsafe sex; adverse social
behaviours (such as violence) and suicidal behaviour.6,9
This study aimed to estimate changes in the risk of alcohol use disorders (mental and behavioural disorders due
to use of alcohol) over the last two decades among the
Western Australian adult population using a populationbased approach.
previous 4-year period (e.g. 1994–1997 vs 1990–1993).
As noted previously in the Methods, a multivariable
Poisson regression model was employed to model
changes in risk over time adjusting for changes in gender
and age distribution of the Western Australian population and potential confounding effects.
Methods
This study has been approved by the Department of
Health Western Australia Human Research Ethics Committee and by the Human Research Ethics Committee,
Curtin University (approval number NDRI-04-2013).
The overall annual risk of hospitalisation for alcohol
disorders was first estimated by calculating the annual
age- and gender-specific incidence rate. In order to
investigate changes in overall risk over time, a multivariable Poisson regression model was employed to adjust
for changes in gender and age distribution of the Western Australian population. To account further for confounding effects, such as changes in patient access
to medical care and improvement in medical treatments,
a newly developed proxy outcome approach10–13 was
used to offset possible biases using a random sample of
emergency department admissions.
Three datasets were used in this study: (i) hospital
morbidity data; (ii) emergency department data; and
(iii) estimated residential population. The hospital morbidity data included all hospital admissions in Western
Australia since 1970. This study includes all hospital separation records for Western Australian residents aged
18 years and older that occurred between 1990 and
2013 with a primary diagnosis of alcohol use disorder.
The hospital morbidity data are coded using ICD-9-CM
between 1990/1991 and 1998/1999 (fiscal year) and
using the ICD-10-CM classification system between 1999
and 2013 in the data. The following codes were to
extract alcohol use disorder hospitalisation records: ICD9-CM codes 291.0–291.9, 303.0–303.9, 305.0 and ICD10-AM codes F10.0–F10.9. Emergency department data
covered emergency presentations to public hospitals in
Western Australia from 2002. A random sample of 1.6
million presentations for Western Australian residents
aged 18 years and older occurred between 2002 and
2013 was used in the analysis.
Data analysis
Gender- and age-specific (grouped into: 18–24 years
then 10-year age groups from 25–74 years and 75+
years) as well as overall incidence rates of alcohol
use disorder hospitalisations were estimated for each
calendar year between 1990 and 2013. Gender- and agespecific incidence rate ratios were calculated to investigate changes in risk between each 4-year period and its
302
Ethics
Results
A total of 3 526 563 person-years at risk was accumulated from the Western Australian adult population
between 1990 and 2013, incurring 58 924 hospitalisations for alcohol use disorders. The risks (95% confidence intervals) per 1000 person-years of hospitalisation
for alcohol use disorders were estimated to be 16.7 overall (16.6, 16.8) and 21.7 (21.5, 22.0) and 11.7 (11.5,
11.8) for males and females respectively. Figure 1 shows
age- and gender-specific annual incidence rates. The risk
of alcohol use disorders has increased from the late
1990s among both genders and all ages with the exception of those aged over 75 years.
Incidence rate ratios of alcohol use disorders were further estimated to compare the risk between each 4-year
period and its previous 4-year period for each gender
and age group (Fig. 2). In males, significant increases
were observed for periods and ages as follows:
2002–2013 for 18–24 years; 2006–2013 for 25–35 years;
1998–2006 and 2010–2013 for 35–44 years; 1998–2001,
2010–2013 for 45–55 years; 2002–2009 for 55–64 years;
2010–2013 for 65–74 years; and no significant increase
observed for 75+ years. In females, significant increases
were observed for periods and ages as follows:
2002–2013 for 18–24 years; 1998–2001, 2006–2013 for
25–34 years; 1994–2013 for 35–44 years; 1998–2013 for
45–54 years; 2006–2013 for 55–64 years; 2002–2013 for
65 years; and no significant increase observed for 75+
years. Significant decrease in risk of alcohol use disorders
was observed for males but not females and mainly
observed in 1994–1997 (compared with 1990–1993) for
25–34, 35–34, 45–54 and 75+ years and in 2002–2005
for the 25–34 years age group.
Moreover, as shown in Figure 1, the risk of alcohol
use disorders decreased in both genders and in most age
groups in the last 2 years (2012 and 2013) within the latest 4-year period, 2010–2013. It was noted that the average risk was higher than the previous 4-year period,
© 2016 Royal Australasian College of Physicians
Alcohol use disorders over last 20 years
2006–2009. Annual changes in the risk of alcohol use
disorders for the whole population were further calculated for each year between 2011 and 2013 using a multivariable Poisson regression model. In keeping with
Figure 1, significant decreases in risk from the previous
year were observed in 2012 (5.5%) and 2013 (11%).
The average risk of alcohol use disorder for the whole
Western Australia population has increased considerably
in the last 10 years (Table 1), and the estimates
remained similar after controlling for unmeasured confounding effects (Table 2). It was observed that participants aged 35–64 years have a higher risk of
hospitalisation for alcohol use disorder.
Incidence Rate
0
10
20
30
18 - 24yrs
0 10 20 30 40
25 - 34yrs
Discussion
0 10 20 30 40 50
35 - 44yrs
0 10 20 30 40 50
45 - 54yrs
0
10 20 30 40
55 - 64yrs
0
10 20 30 40
65 - 74yrs
0 5 10 15 20 25
75yrs +
1990
1995
2000
2005
2010
2013
year
Figure 1 Annual age- and gender-specific incidence rate (per 10 000
person-years) of alcohol use disorder hospitalisations between 1990
and 2013. (
), Male; (
), female.
© 2016 Royal Australasian College of Physicians
The incidence rate of hospitalisation for alcohol use disorders decreased in 1994–1997 then increased from
1998, peaked in 2011 and declined in 2012 and 2013.
However, the average rate remained significantly higher
from 2010 to 2013 compared with previous years
(Tables 1 and 2). The incidence rate trend is in relatively
good agreement with the national trend in per capita
alcohol consumption among the 15+ year population,
which showed a decrease in 1990–1996 and increased
afterwards until about 2008,14 whereas the decrease in
the risk of alcohol use disorders in 2012 and 2013
appeared to reflect the gradual decline in per capita alcohol consumption since 2009.15 In addition, a similar
trend has been reported among national hospitalisations
for alcohol-caused liver disease in Australia.16 The
decline in incidence in 2012 and 2013 is in good agreement with data from the National Drug Strategy Household Survey (NDSHS) 2013, which showed that the
prevalence of lifetime risky drinkers (≥2 standard drinks
per day on average) among the 14+-year Australian
population reduced to 18.2% (21.6% in WA) in 2013
from 20.0% (23.0% in WA) in 2010, and the prevalence
of consuming ≥4 drinks on most days/everyday reduced
to 6.7% in 2013 from 7.9% in 2010.4 The increased risk
of hospitalisation for alcohol use disorders among the
35–64 year age group compared with the younger age
groups appeared to be against the age-specific prevalence
of lifetime risky drinkers (≥2 standard drinks per day on
average) observed in the NDSHS, which was the highest
in the 18–24 years age group.4 Alcohol use disorders are
most likely developed among people who frequently use
alcohol at the highest level, and data from the same survey showed that the prevalence of consuming ≥4 drinks
on most days/everyday was higher among middle-age
and older people compared with the youngest age
groups.4 Nevertheless, it is possible that the age difference in the risk of alcohol use disorder hospitalisation
303
Liang & Chikritzhs
Female, 18-24yrs
Male, 25-34yrs
Female, 25-34yrs
Male, 35-44yrs
Female, 35-44yrs
Male, 45-54yrs
Female, 45-54yrs
Male, 55-64yrs
Female, 55-64yrs
Male, 65-74yrs
Female, 65-74yrs
Male, 75yrs+
Female, 75yrs+
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
.5
1
1.5
2
2.5
Male, 18-24yrs
94-97
98-01
02-05
06-09
10-13 94-97
98-01
Denoted four-year period
304
02-05
06-09
10-13
Figure 2 Changes in risk of alcohol use disorder hospitalisation between each 4-year period
(measured as incidence rate ratio of each
denoted 4-year period to its previous 4–year
period) by age and gender.
© 2016 Royal Australasian College of Physicians
Alcohol use disorders over last 20 years
Table 1 Changes in the risk of alcohol use disorder hospitalisations
over the last two decades and estimates from multivariable Poisson
regression with adjustment for age and gender
Observation period
1990–1993
1994–1997
1998–2001
2002–2005
2006–2009
2010–2013
Age group (years)
18–24
25–34
35–44
45–54
55–64
65–74
75+
Sex
Male
Female
IRR
P value
95% confidence interval
0.89
0.84
0.95
1.00
1.25
1.87
<0.001
<0.001
<0.001
0.86
0.81
0.92
0.92
0.86
0.98
<0.001
<0.001
1.22
1.82
1.29
1.92
1.00
1.36
1.89
1.87
1.46
0.89
0.37
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
1.32
1.83
1.81
1.41
0.85
0.34
1.41
1.95
1.92
1.51
0.93
0.39
1.00
0.55
Table 2 Changes in the risk of alcohol use disorders hospitalisations for
2002–2013, estimates from multivariable Poisson regression with
adjustment for age and gender and underline trend of ED admissions
Observation period
2002–2005
2006–2009
2010–2013
Age group (years)
18–24
25–34
35–44
45–54
55–64
65–74
75+
Sex
Male
Female
IRR
P value
95% confidence interval
1.00
1.10
1.52
<0.001
<0.001
1.07
1.48
1.13
1.56
1.00
1.32
2.12
2.31
1.67
0.74
0.14
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
1.27
2.04
2.22
1.60
0.70
0.13
1.37
2.20
2.39
1.74
0.78
0.16
1.00
0.55
<0.001
0.54
0.56
IRR, incidence rate ratio.
<0.001
0.54
0.56
IRR, incidence rate ratio.
was at least partly due to the delay between the onset of
alcohol use disorders and treatment, which has been
estimated to be about 18 years in Australia.17 Also noteworthy is that this study used unlinked hospitalisation
data, and thus, repeated hospitalisation of the same
patients are possible, that is, admission due to relapse in
later life. In addition, the frequency of visits to general
practitioners increases with age,18 and therefore, as has
been observed for other medical conditions,18 alcohol
use disorders may be more likely to be diagnosed among
middle-aged and older people compared with younger
populations.
The trends for risk of alcohol use disorder hospitalisation measured in this study indicated that the harmful
effects caused by heavy alcohol use have been rising
throughout most of the last two decades among Western
Australians. Another finding with potential importance
is that the risk of alcohol use disorders among women
aged 35–54 years appeared to keep rising until 2011
with small reductions in 2012 and 2013. Similar findings
have been reported in a US study, which showed that
References
1 WHO Collaborative Study Group on
Alcohol and Injuries. WHO Collaborative
Study on Alcohol and Injuries: Final Report.
Geneva: Department of Mental Health
and Substance Abuse, Department of
Injuries and Violence Prevention, World
Health Organization; 2007.
© 2016 Royal Australasian College of Physicians
several older people requiring treatment for alcohol problems has been increasing in recent years.19 Concerns
about heavy alcohol use behaviours and alcohol use disorders among middle-aged and older women in
Australia have been discussed in recent studies.20,21 Findings from this study indicate that further research targeting middle-aged female daily alcohol users is important
for the development of future prevention and treatment
strategies.
Conclusion
The risk of alcohol use disorder hospitalisations among
the Western Australia population has increased considerably over the last two decades. This is indicative of an
increase in the harmful effects of heavy alcohol use in
the population.
Acknowledgements
The authors thank the Western Australia Department of
Health and Australian Bureau of Statistics for providing
the data used in this study.
2 World Health Organization. Global Status
Report on Alcohol and Health. Geneva:
WHO; 2014.
3 National Expert Advisory Committee on
Alcohol (NEACA). Alcohol in Australia:
Issues and Strategies. A Background Paper
to the National Alcohol Strategy: A Plan for
Action 2001 to 2003/04. Ministerial
Council on Drug Strategy,
Department of Health and Ageing:
Canberra; 2001.
4 AIHW. National Drug Strategy Household
Survey Key Findings: Alcohol Use, Vol.
2014. Canberra: Australian Institute of
Health and Welfare; 2013.
5 World Health Organization. The ICD-10
Classification of Mental and Behavioural
Disorders: Clinical Descriptions and
305
Liang & Chikritzhs
6
7
8
9
10
11
Diagnostic Guidelines. Geneva: World
Health Organization; 1992.
American Psychiatric Association D. S.
M. Task Force. Diagnostic and Statistical
Manual of Mental Disorders: DSM-5, 5th
edn. Washington, DC: American
Psychiatric Association; 2013.
American Psychiatric Association.
Diagnostic and Statistical Manual of Mental
Disorders: DSM-IV, 4th edn. Washington,
DC: American Psychiatric Association;
1994.
Rehm J, Mathers C, Popova S,
Thavorncharoensap M,
Teerawattananon Y, Patra J. Global
burden of disease and injury and
economic cost attributable to alcohol
use and alcohol-use disorders. Lancet
2009; 373: 2223–33.
Schuckit MA. Alcohol-use disorders.
Lancet 2009; 373: 492–501.
Liang W, Zhao Y, Lee AH. A proxy
outcome approach for causal effect in
observational studies: a simulation
study. Biomed Res Int 2014; 2014: 8.
Liang W, Chikritzhs T. Observational
research on alcohol use and chronic
306
12
13
14
15
16
disease outcome: new approaches to
counter biases. Sci World J 2013;
2013: 14.
Tchetgen TE. The control outcome
calibration approach for causal
inference with unobserved
confounding. Am J Epidemiol 2014; 179:
633–40.
Liang W, Chikritzhs T. Alcohol
consumption and health status of family
members: health impacts without
ingestion. Intern Med J 2013; 43:
1012–16.
Chikritzhs T, Allsop SJ, Moodie AR,
Hall WD. Per capita alcohol
consumption in Australia: will the real
trend please step forward? Med J Aust
2010; 193: 594–7.
Australian Bureau of Statistics.
4307.0.55.001 – Apparent Consumption of
Alcohol, Australia, 2012-13. Canberra:
ABS; 2014.
Liang W, Chikritzhs T, Pascal R,
Binns CW. Mortality rate of alcoholic
liver disease and risk of hospitalisation
for alcoholic liver cirrhosis, alcoholic
hepatitis, and alcoholic liver failure in
17
18
19
20
21
Australia between 1993 and
2005. Intern Med J 2011;
41: 34–41.
Chapman C, Slade T, Hunt C,
Teesson M. Delay to first treatment
contact for alcohol use disorder. Drug
Alcohol Depend 2015; 147: 116–21.
Australian Bureau of Statistics. Health
Services: Patient Experiences in Australia.
Canberra: Australian Bureau of
Statistics; 2009.
Han B, Gfroerer JC, Colliver JD,
Penne MA. Substance use disorder
among older adults in the United States
in 2020. Addiction 2009; 104: 88–96.
Wilkinson C, Allsop S, Dare J. Alcohol,
ageing and Australia. Drug Alcohol Rev
2015. doi:10.1111/dar.12301
Clemens SL, Matthews SL, Young AF,
Powers JR. Alcohol consumption of
Australian women: results from the
Australian Longitudinal Study on
Women’s Health. Drug Alcohol Rev 2007;
26: 525–35.
© 2016 Royal Australasian College of Physicians
Association between hepatitis B virus infection and risk of
multiple myeloma: a systematic review and meta-analysis
Y. Li, O. Bai, C. Liu, Z. Du, X. Wang, G. Wang and W. Li
Cancer Center, First Hospital of Jilin University, Changchun, Jilin Province, China
Key words
hepatitis B virus, multiple myeloma,
meta-analysis.
Correspondence
Wei Li, Cancer Center, First Hospital of Jilin
University, 71 Xinmin Street, Changchun
130021, Jilin Province, China.
Email: [email protected]
Received 3 December 2014; accepted 3
December 2015.
doi:10.1111/imj.12981
Abstract
Background: Hepatitis B virus (HBV) infection is a major global public health concern.
Although recent findings suggest an inverse relationship between HBV infection and
multiple myeloma (MM), the true relationship between these two conditions remains
unclear.
Aim: The primary aim of this meta-analysis was to evaluate the association between
HBV infection, defined as hepatitis B surface antigen positivity, and the incidence of MM.
Methods: We searched the PubMed/Medline, Cochrane Library and EMBASE databases from January 1975 to July 2014 and reviewed the reference lists of all retrieved
articles. Odds ratios (OR) and 95% confidence intervals (CI) were calculated using
fixed- and random-effects models.
Results: We identified nine case–control studies involving 30 646 patients with MM
and 379 837 controls. HBV infection was not significantly associated with the development of MM (OR = 1.3; 95% CI: 0.92–1.82; P = 0.14). A similar risk of developing MM
was present in different HBV-prevalent countries. However, significant heterogeneity
was observed among studies (P = 0.01). A statistically significant relationship between
HBV infection and increased MM risk was detected in sub-analyses evaluating highquality studies and those with hospital-based controls (P < 0.05).
Conclusion: HBV infection may be associated with an increased risk of MM. However,
confirmation of this relationship and the specific molecular mechanisms involved in the
association between HBV infection and the development of MM require further
exploration.
Introduction
Multiple myeloma (MM) is a B-cell malignancy characterised by the proliferation of clonal plasma cells within
the bone marrow with the production of an abnormal
monoclonal paraprotein and evidence of end-organ
damage. On a global scale, an estimated 86 000 incident
cases occur annually,1 accounting for approximately 1%
of all cancers and 13% of all haematologic malignancies.2 Although the exact cause of MM is unknown,
older age, obesity, male sex, exposure to ionising radiation, African ancestry and a history of monoclonal gammopathy of undetermined significance are established
risk factors for MM.3 Some studies have suggested that
human immunodeficiency virus and hepatitis C virus
infections are related to an elevated risk of developing MM.4–8
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
The hepatitis B virus (HBV), a small DNA virus, is a
member of the Hepadnaviridae family. HBV is the causative pathogen in most cases of acute and chronic hepatitis worldwide. The World Health Organization has
estimated that more than 2 billion people exhibit serologic evidence of HBV infection and that 360 million of
these people have chronic HBV infection.9 A mathematical model estimated that, globally, around 600 000 individuals die of HBV-related liver disease every year.10
HBV not only exhibits hepatotropic characteristics, but
also a lymphotropic virus.11 HBV infection of and replication within lymphocytes is speculated to contribute to
the development of lymphoproliferative disorders. Several studies have evaluated the potential associations
between the presence of HBV infection and the development of lymphoproliferative disorders.12–14 A recent
meta-analysis demonstrated an increased prevalence of
HBV infection in patients with non-Hodgkin lymphoma.15 Although several recent studies have suggested
307
Li et al.
an association between the presence of HBV and the
development of MM,16,17 this association remains controversial. The aim of the present meta-analysis of epidemiologic studies was to evaluate the association between
HBV infection and MM.
Methods
Literature search
Two authors (Y.Y.L. and O.B.) independently searched
the electronic databases of PubMed/Medline, EMBASE
and the Cochrane Library for relevant studies published
in English from January 1975 to July 2014. The combination term ‘hepatitis B and myeloma’ was used in the
search. The titles and abstracts of all studies were
reviewed to determine whether an article was relevant
to the present meta-analysis. Relevant papers were
retrieved and assessed, and the reference lists in these
articles were screened to identify additional relevant
studies.
Study selection
We included only case–control studies that investigated
the association between the presence of HBV infection
and the development of MM. Case series, case reports
and review articles were excluded. HBV infection was
defined as the presence of hepatitis B surface antigen
(HBsAg) positivity. The definition of MM was based on
either that described in the World Health Organization
classification18 or by the International Myeloma Working
Group.19
Any disagreements between the two reviewers were
resolved by discussion among all of the authors. When
the same study population was included in multiple publications, only the most informative study was selected
for inclusion in the meta-analysis. Full-text versions of
all eligible studies were obtained for quality assessment
and data extraction.
Data extraction
Data extraction was independently performed by two
reviewers (Y.Y.L. and O.B.). The previously designed
data extraction form included the following items:
authors, title, date of publication, country of origin, years
of inclusion, method of determining HBV status, method
of MM diagnosis, sources and definitions of patients and
controls, sample size and sex and age of study participants. Disagreements between the two reviewers were
resolved by consensus among all of the authors.
308
Quality assessment
The quality of each study was independently assessed by
two reviewers (Y.Y.L. and O.B.) using the Newcastle–
Ottawa scale (NOS)20 (Table S1, Supporting information). The NOS includes three quality parameters: study
group selection (0–4 points), study group comparability
(0–2 points) and determination of either the exposure or
outcome of interest (0–3 points). The highest possible
score is 9 points and is assigned to studies of the highest
quality. In the present analysis, NOS scores of 1–3, 4–6
and 7–9 indicated low-, intermediate- and high-quality
studies respectively.
Data analysis
Review Manager 5.3 software (RevMan 5.3; The Nordic
Cochrane Center, The Cochrane Collaboration, Copenhagen, Denmark) and Stata 12.0 (StataCorp, College
Station, TX, USA) were used to perform all statistical
analyses. Heterogeneity between studies was assessed
using the Q test (P < 0.10 indicated statistically significant heterogeneity) and I2 test (values of 25, 50 and
75% indicated mild, moderate and severe heterogeneity
respectively).21 Either a random- or fixed-effects model
was used for the meta-analysis depending on the heterogeneity between studies. The fixed-effects model was
used in the absence of significant heterogeneity
(I2 < 50%, P > 0.10); otherwise, the random-effects
model was used.
A contour-enhanced funnel plot22 and Egger’s test23
were used to evaluate publication bias. For Egger’s test,
P < 0.05 was considered to indicate statistically significant publication bias. The primary outcome in this metaanalysis was the prevalence of HBV infection compared
between patients with and without MM. Comparisons
were made by calculating odds ratios (OR) and 95%
confidence intervals (CI). Subgroup analysis of the HBV
prevalence by country was performed. Countries were
considered to have low, low-intermediate, highintermediate and high HBV prevalence based on an HBV
prevalence report previously published by the Centers
for Disease Control and Prevention.24 Sensitivity analyses were performed to assess the stability of the results.
For these analyses, individual studies were eliminated in
turn in order to determine the influence of any given
individual data set to the pooled OR.
Results
Search results
Our electronic database search initially identified
949 citations. We excluded 938 articles because they did
© 2016 Royal Australasian College of Physicians
Systematic review of HBV infection and multiple myeloma
Figure 1 Search results.
not meet the inclusion criteria. Nine case–control
studies12,13,16,17,25–29 were selected for the meta-analysis.
A flow chart of the search process is shown in Figure 1.
in four studies were hospital-based controls,13,16,26,27 and
five studies used population-based controls.12,17,25,28,29
Outcomes
Characteristics of studies
Table 1 lists the main characteristics of all nine case–
control studies included in the meta-analysis. All studies
were published from 1975 to 2012. The nine studies
involved 30 646 patients with MM and 379 857 controls. One study originated from Africa,29 one from the
United States,25 three from Europe12,26,28 and four from
Asia.13,16,17,27 One study was prospective12 and eight
were
retrospective.13,16,17,25–29
Seven
stud12,13,16,17,26,27,29
ies
(78%) determined patients’ HBV status by seropositivity of at least the HBsAg. The other
two studies25,28 (22%) did not test the HBsAg level;
one determined patients’ HBV status based on their
Medicare diagnosis and/or billing records,25 and the
other used data obtained from the Swedish Inpatient
Register.28
Quality assessment
According to the NOS, six studies12,13,16,26,27,29 (63%)
were of high quality and three17,25,28 (33%) were of
intermediate quality. Two studies17,29 (22%) did not
describe how patients and controls were matched at the
time of the trial design or whether confounders were
adjusted for and during data analysis. The control groups
© 2016 Royal Australasian College of Physicians
The OR of detecting HBV infection in patients with MM
versus controls using a random-effects model was 1.30
(95% CI: 0.92–1.82, P = 0.14), suggesting an increased
prevalence of HBV infection in patients with MM, however the difference was not statistically significant
(Fig. 2). A moderate degree of heterogeneity was
observed (I2 = 60%, P = 0.01). No evidence of publication bias was found on the funnel plots (Fig. 3) or by
Egger’s test (P = 0.13).
Subgroup analysis
The prevalence of HBV infection varied geographically.
Countries with high (>8%), low-intermediate (2–4%),
high-intermediate (5–7%) and low (< 2%) HBsAg prevalence were identified.24,30 Subgroup analyses were performed in the countries with low, high and intermediate
HBsAg prevalence. Among the studies performed in
countries
with
high
and
intermediate
HBV
infection,13,16,17,27,29 the adjusted OR of MM in patients
with HBV was 1.26 (95% CI: 0.72–2.19, P = 0.41)
(Fig. 4A). A high degree of heterogeneity was observed
(I2 = 76%, P < 0.01), but no publication bias was evident (Egger’s test, P = 0.68). Studies performed in
Europe and the United States,12,25,26,28 regions with a
309
310
Nigeria (high)
intermediate)
Taiwan (high-
intermediate)
China (high-
Retrospective
Retrospective
Retrospective
Retrospective
Cases
Controls
Centre, age, sex,
time of blood
technique
Reverse passive
haemagglutination
ELISA
ELISA
1987–1989
2003–2008
Ibadan, Nigeria
Hospital, Ibadan,
Nigeria
Hospital blood bank,
Healthy voluntary donors to
the University College
University College
16
scale survey in Taiwan
malignancies at
Patients with
1ymphoproliferative
General Hospital
with myeloma at
Taipei Veterans
3
Guangzhou
Historical data from a large-
Guangzhou
155
Affiliated Hospital of Sun
Yat-sen University in
of Sun Yat-sen
University in
17
leukaemia at the First
First Affiliated Hospital
Patients newly diagnosed
diagnosed acute
and Spain
Patients with newly
with myeloma at the
Patients newly diagnosed
Republic, Ireland, Italy
438
27 210
5690
157 720
Not reported
Not reported
Age and sex
lymphoma in Czech
and Spain
Age, sex, country
of residence
Sex, age and country
Age, sex and year of
serum collection
level
299
1375
75 408
15 562
hospitalised patients with
diseases other than
36
170
4
636
Republic, France,
Germany, Ireland, Italy
Population registers in
Database
Swedish Population
Asan Medical Center
Patients undergoing general
medical examinations at
immunoassay
299
171
19 112
593
and education
58
27
1
40
Germany and Italy;
Hospital- or clinic-based
Cancer Register
Nationwide Swedish
Asan Medical Center
pathology records
fasting status
470
collection and
3
available
Cohort participants who
were cancer-free, alive
Age and sex
selection year
Age, sex and
Matching and
adjustments
and had blood samples
238
802
12,2531
N
pathology registries
6
40
242
n
records and
Regional cancer registry,
health insurance
disorder
lymphoproliferative
blood centre
Patients recruited using
Medicare Beneficiary
Database
5% random sample from
Source of controls
patients newly
diagnosed with
67
9995
N
records from university
2
20
n
Gaziantep University:
Haematology Clinic at
SEER-Medicare dataset
Source of cases
patients in Czech
1996–2008
1998–2004
1965–2004
1997–2008
1993–1998
2003–2007
1993–2002
Evaluation
period
microparticle
Chemiluminescent
NA
ELISA
ELISA
ELISA and PCR
claim for HBV
NA/single Medicare
HBsAg detection
Pathologic
confirmation
Working Group
Myeloma
by the
International
Criteria proposed
confirmation
Pathologic
confirmation
Swedish Cancer
Register
Pathologic
Cases in nationwide
Pathologic
confirmation
Pathologic
confirmation
confirmation
Pathologic
ICD-03 coding
MM assessment
7
6
7
7
6
7
8
7
6
NOS score
ELISA, enzyme-linked immunosorbent assay; HBsAg, hepatitis B surface antigen; ICD, International Classification of Diseases; NOS, Newcastle–Ottawa Scale; PCR, polymerase chain reaction; SEER, Surveillance, Epidemiology, and End Results.
Olatunji,
199129
Teng, 201117
201216
Huang,
201226
Becker,
Europe (low)
Retrospective
Sweden (low)
Lindqvist,
201128
Retrospective
South Korea (lowintermediate)
Kang, 201127
Prospective
Retrospective
Europe (low)
intermediate)
Retrospective
Type of study
Franceschi,
201112
Okan, 200813
Turkey (low-
United States (low)
Anderson,
200825
Country (HBV
prevalence)
References
Table 1 Main characteristics of studies evaluating the association between hepatitis B virus (HBV) infection and multiple myeloma (MM)
Li et al.
© 2016 Royal Australasian College of Physicians
Systematic review of HBV infection and multiple myeloma
Figure 2 Estimated odds ratios of developing multiple myeloma in patients with hepatitis B virus infection using a random-effects model.
Figure 3 Funnel plot for publication bias with
95% pseudoconfidence limits.
low prevalence of HBV infection, had an adjusted OR of
MM in patients with HBV of 1.23 (95% CI: 0.91–1.67,
P = 0.17) (Fig. 4B). No significant heterogeneity was
observed (I2 = 17%, P = 0.30), and no publication bias
was found (Egger’s test, P = 0.25).
Sensitivity analysis and assessment
of heterogeneity
Significant heterogeneity was present in the overall analysis. Thus, we performed the sensitivity analysis by
excluding one study at a time and recalculating the OR
in an attempt to detect the source of heterogeneity. Similar results were obtained after sequentially excluding
each study, with exception of the Teng et al. study,17
where a significant relationship between HBV infection
and increased risk of MM was observed when it was
excluded (OR = 1.48, 95% CI: 1.16–1.88, P < 0.01).
© 2016 Royal Australasian College of Physicians
The results were altered when we performed subgroup analyses according to the quality of the studies.
After exclusion of the three studies with NOS scores
<7,17,25,28 the pooled results showed that HBV infection
was significantly associated with increased risk of
MM (OR = 1.63, 95% CI: 1.30–2.05, P < 0.01). No heterogeneity (I2 = 4%, P = 0.39) or obvious publication
bias was found (Egger’s test, P = 0.33). A similar outcome was observed after excluding the two studies
that did not describe the comparability of patients and
controls.17,29 The OR of MM in patients with HBV
was 1.44 (95% CI: 1.13–1.85, P < 0.01), the degree of
heterogeneity was not significant (I2 = 23%, P = 0.25),
and no publication bias was identified (Egger’s test,
P = 0.14). HBV infection was significantly associated
with MM risk in the hospital-based control13,16,26,27
(OR = 1.57, 95% CI: 1.25–1.96, P < 0.01), but not in the
population-based controls12,17,25,28,29 (OR = 1.20, 95%
311
Li et al.
Figure 4 Estimated odds ratios of developing multiple myeloma in patients with hepatitis B virus (HBV) infection in (A) countries with a high and intermediate prevalence of HBV infection and (B) countries with a low prevalence of HBV infection.
CI: 0.64–2.28, P = 0.57). There was no heterogeneity
among the studies that included hospital-based controls
(I2 = 0, P = 0.42), but significant heterogeneity was
found in those using population-based controls
(I2 = 62%, P = 0.03).
Discussion
The main finding of this meta-analysis is that a relatively
higher rate of seropositivity for HBV infection is observed
among patients with than without MM. Patients with
HBV infection have 1.3 times higher risk of developing
MM than controls. This finding suggests that HBV infection may play a role in the pathogenic development of
MM. The subtype analysis of studies performed in countries with a high, intermediate or low prevalence of HBV
infection suggested that those infected in all countries
had a similar increased risk of developing MM. However,
the differences were not statistically significant, and high
heterogeneity was present. Heterogeneity was reduced
when evaluating studies with consideration of their
quality and the source of controls, and the relationship
between HBV infection and the risk of MM development
achieved significance among the high quality and
hospital-based control studies.
Another important point is that the high prevalence of
HBV infection in patients with MM may cause a
312
therapeutic dilemma in the clinical setting. Previous data
show that HBV reactivation commonly occurs following
chemotherapy, occurring in 21–53% of HBsAg carriers.31
All patients with HBV infection who subsequently
develop MM are at risk for reactivation following chemotherapy, particularly following immunosuppressive chemotherapy. Several recent studies found a significantly
higher occurrence of HBV reactivation and/or hepatic
damage after patients with HBV infection underwent chemotherapy, especially protocols containing high doses of
glucocorticoids or proteasome inhibitor bortezomib,32–35
which are the commonly used agents in the treatment of
MM. The mechanism of HBV reactivation in association
with chemotherapy is unclear. HBV DNA contains a
glucocorticoid-responsive element that reportedly facilitates HBV replication.36,37 Bortezomib may alter the number and function of CD8 T cells and CD56 natural killer
cells,38 and HBV reactivation may be attributable to the
effect of bortezomib on cell-mediated immunity.32 In
addition, patients with concurrent MM and HBV infection
reportedly experience more and earlier hepatic adverse
events while undergoing chemotherapy.17,39 The HBV status of patients with MM should be determined before
chemotherapy or immunosuppressive therapy is initiated.
Evidence suggests that patients with MM undergoing chemotherapy would benefit more from antiviral prophylaxis
than from serial monitoring of HBV DNA.40
© 2016 Royal Australasian College of Physicians
Systematic review of HBV infection and multiple myeloma
Our meta-analysis has certain limitations. First, a significant degree of heterogeneity was detected among the
studies evaluated. Moreover, some of the original studies
had their own limitations, such as too few controls,12 relatively low sensitivity methods for detection of HBV29 or
used hospital-based controls,13,16,26,27 which may have
influenced the result of the meta-analysis. Indeed, a significant relationship was observed between HBV infection
and the increased risk of MM when controlling for these
factors. Second, we did not assess serum markers that are
used to identify occult HBV infections, such as HBeAg or
HBV DNA. Several recent studies showed a significant correlation between the presence of occult HBV infection and
the development of chronic lymphocytic leukaemia/small
lymphocytic lymphoma.41 These findings indicate that the
presence of HBsAg seropositivity may lead clinicians to
underestimate the risk of MM in patients with HBV infection. Further studies are needed to determine the relationship between the presence of occult HBV infection and
the development of MM. Third, as mentioned above, MM
patients are at a risk for HBV reactivation following treatment. The time interval between the HBsAg test and MM
diagnosis may have affected the results. However, we
could not assess the relationship due to the lack of accurate descriptions of this time interval in most of the studies. Finally, this meta-analysis included only studies
published in English. HBV infection is highly endemic in
Asia, Africa and the Middle East, and the populations in
References
1 Becker N. Epidemiology of multiple
myeloma. Recent Results Cancer Res 2011;
183: 25–35.
2 Palumbo A, Anderson K. Multiple
myeloma. N Engl J Med 2011; 364:
1046–60.
3 Alexander DD, Mink PJ, Adami HO,
Cole P, Mandel JS, Oken MM et al.
Multiple myeloma: a review of the
epidemiologic literature. Int J Cancer
2007; 120(Suppl): 40–61.
4 Dal Maso L, Franceschi S. Hepatitis C
virus and risk of lymphoma and other
lymphoid neoplasms: a meta-analysis of
epidemiologic studies. Cancer Epidemiol
Biomarkers Prev 2006; 15: 2078–85.
5 Duberg AS, Nordstrom M, Torner A,
Reichard O, Strauss R, Janzon R et al.
Non-Hodgkin’s lymphoma and other
nonhepatic malignancies in Swedish
patients with hepatitis C virus infection.
Hepatology 2005; 41: 652–9.
6 Goedert JJ, Cote TR, Virgo P,
Scoppa SM, Kingma DW, Gail MH et al.
Spectrum of AIDS-associated malignant
© 2016 Royal Australasian College of Physicians
7
8
9
10
11
most regions of these countries do not speak English. This
factor may have influenced the results.
Despite the above-mentioned limitations, our study has
several strengths. First, to the best of our knowledge, this
is the first meta-analysis to evaluate the association
between the presence of HBV infection and the development of MM. Second, the NOS score indicated that most
of the included studies were of high quality. In addition,
the main association between HBV infection and MM
became statistically significant after excluding studies with
an NOS score of <7. Finally, our study showed that the
association between the presence of HBV infection and
the development of MM might not vary between countries with a high and low prevalence of HBV.
Conclusion
The present meta-analysis has shown that the presence of
HBV infection may be associated with an increased risk of
developing MM. However, the relatively low qualities of
enrolled studies have affected the reliability of the results.
Future studies on the risk of MM in patients with HBV
infection are needed to elucidate fully the relationship
between HBV infection in the whole population, including those with occult HBV infection, and the development
of MM. Importantly, further studies should focus on elucidating the biologic relationship between the presence of
HBV infection and the development of MM.
disorders. Lancet 1998; 351:
1833–9.
Grulich AE, Wan X, Law MG, Coates M,
Kaldor JM. Risk of cancer in people
with AIDS. AIDS 1999; 13:
839–43.
Montella M, Crispo A, Russo F,
Ronga D, Tridente V, Tamburini M.
Hepatitis C virus infection and new
association with extrahepatic disease:
multiple myeloma. Haematologica 2000;
85: 883–4.
World Health Organization. Hepatitis B
vaccines. Wkly Epidemiol Rec 2009; 84:
405–19.
Goldstein ST, Zhou F, Hadler SC,
Bell BP, Mast EE, Margolis HS. A
mathematical model to estimate global
hepatitis B disease burden and
vaccination impact. Int J Epidemiol 2005;
34: 1329–39.
Yoffe B, Noonan CA, Melnick JL,
Hollinger FB. Hepatitis B virus DNA in
mononuclear cells and analysis of cell
subsets for the presence of replicative
intermediates of viral DNA. J Infect Dis
1986; 153: 471–7.
12 Franceschi S, Lise M, Trepo C,
Berthillon P, Chuang SC, Nieters A et al.
Infection with hepatitis B and C viruses
and risk of lymphoid malignancies in
the European Prospective Investigation
into Cancer and Nutrition (EPIC). Cance
Epidemiol Biomarkers Prev 2011; 20:
208–14.
13 Okan V, Yilmaz M, Bayram A,
Kis C, Cifci S, Buyukhatipoglu H
et al. Prevalence of hepatitis B and C
viruses in patients with
lymphoproliferative disorders.
Int J Hematol 2008; 88: 403–8.
14 Pinato DJ, Rossi D, Minh MT,
Toniutto P, Boccato E, Minisini R et al.
Hepatitis B virus and lymphomagenesis:
novel insights into an occult
relationship. Dig Liver Dis 2012; 44:
235–8.
15 Dalia S, Chavez J, Castillo JJ, Sokol L.
Hepatitis B infection increases the risk
of non-Hodgkin lymphoma: a metaanalysis of observational studies. Leuk
Res 2013; 37: 1107–15.
16 Huang B, Li J, Zhou Z, Zheng D, Liu J,
Chen M. High prevalence of hepatitis
313
Li et al.
17
18
19
20
21
22
23
24
B virus infection in multiple
myeloma. Leuk Lymphoma 2012; 53:
270–4.
Teng CJ, Liu HT, Liu CY, Hsih CH,
Pai JT, Gau JP et al. Chronic hepatitis
virus infection in patients with multiple
myeloma: clinical characteristics and
outcomes. Clinics (Sao Paulo) 2011; 66:
2055–61.
McKenna RW, Kyle RA, Kuehl WM,
Grogan TM, Harris NL, Coupland RW.
Plasma cell neoplasms. In: Swerdlow
SH, Campo E, Harris NL, Jaffe ES, Pileri
SA, Stein H et al., eds. WHO Classification
of Tumours of Haematopoietic and
Lymphoid Tissues, 4th edn. Lyon: IARC
Press; 2008, 200–13.
International Myeloma Working Group.
Criteria for the classification of
monoclonal gammopathies, multiple
myeloma and related disorders: a report
of the International Myeloma Working
Group. Br J Haematol 2003; 121:
749–57.
Wells G, Shea B, O’Connell D,
Peterson J. The Newcastle-Ottawa Scale
(NOS) for assessing the quality if
nonrandomized studies in metaanalyses. Ottawa Health Research
Institute [cited 2014 Nov 30]. Available
from URL: http://www.ohri.ca/
programs/clinical_epidemiology/
oxford.asp
Higgins JP, Thompson SG. Quantifying
heterogeneity in a meta-analysis. Stat
Med 2002; 21: 1539–58.
Peters JL, Sutton AJ, Jones DR,
Abrams KR, Rushton L. Contourenhanced meta-analysis funnel plots
help distinguish publication bias from
other causes of asymmetry. J Clin
Epidemiol 2008; 61: 991–6.
Egger M, Davey Smith G, Schneider M,
Minder C. Bias in meta-analysis
detected by a simple, graphical test. BMJ
1997; 315: 629–34.
Francisco A. Chapter 3: Infectious Diseases
Related to Travel, Hepatitis B. Centers for
Disease Control and Prevention; 2014
[cited 310 Nov 2014]. Available from
URL: http://wwwnc.cdc.gov/travel/
25
26
27
28
29
30
31
32
yellowbook/2016/infectious-diseasesrelated-to-travel/hepatitis-b
Anderson LA, Pfeiffer R, Warren JL,
Landgren O, Gadalla S, Berndt SI et al.
Hematopoietic malignancies associated
with viral and alcoholic hepatitis. Cancer
Epidemiol Biomarkers Prev 2008; 17:
3069–75.
Becker N, Schnitzler P, Boffetta P,
Brennan P, Foretova L, Maynadie M
et al. Hepatitis B virus infection and risk
of lymphoma: results of a serological
analysis within the European casecontrol study Epilymph. J Cancer Res
Clin Oncol 2012; 138:
1993–2001.
Kang J, Cho JH, Suh CW, Lee DH,
Oh HB, Sohn YH et al. High prevalence
of hepatitis B and hepatitis C virus
infections in Korean patients with
hematopoietic malignancies. Ann
Hematol 2011; 90: 159–64.
Lindqvist EK, Goldin LR, Landgren O,
Blimark C, Mellqvist UH, Turesson I
et al. Personal and family history of
immune-related conditions increase the
risk of plasma cell disorders: a
population-based study. Blood 2011;
118: 6284–91.
Olatunji PO, Okpala IE, Sorunmu MA.
Hepatitis B surface antigenaemia in
patients with malignant
lymphoproliferative disorders. Tokai J
Exp Clin Med 1991; 16: 171–3.
Ott JJ, Stevens GA, Groeger J,
Wiersma ST. Global epidemiology of
hepatitis B virus infection: new
estimates of age-specific HBsAg
seroprevalence and endemicity. Vaccine
2012; 30: 2212–19.
Yeo W, Chan PK, Zhong S, Ho WM,
Steinberg JL, Tam JS et al. Frequency of
hepatitis B virus reactivation in cancer
patients undergoing cytotoxic
chemotherapy: a prospective study of
626 patients with identification of risk
factors. J Med Virol 2000; 62: 299–307.
Hussain S, Jhaj R, Ahsan S, Ahsan M,
Bloom RE, Jafri SM. Bortezomib
induced hepatitis B reactivation. Case
Rep Med 2014; 2014: 964082.
33 Matsui T, Kang JH, Nojima M,
Tomonari A, Aoki H, Yamazaki H et al.
Reactivation of hepatitis B virus in
patients with undetectable HBsAg
undergoing chemotherapy for
malignant lymphoma or multiple
myeloma. J Med Virol 2013; 85:
1900–6.
34 Tanaka H, Sakuma I, Hashimoto S,
Takeda Y, Sakai S, Takagi T et al.
Hepatitis B reactivation in a multiple
myeloma patient with resolved hepatitis
B infection during bortezomib therapy:
case report. J Clin Exp Hematop 2012;
52: 67–9.
35 Yeo W, Johnson PJ. Diagnosis,
prevention and management of
hepatitis B virus reactivation during
anticancer therapy. Hepatology 2006; 43:
209–20.
36 Liaw YF. Hepatitis viruses under
immunosuppressive agents. J
Gastroenterol Hepatol 1998; 13: 14–20.
37 Tur-Kaspa R, Burk RD, Shaul Y,
Shafritz DA. Hepatitis B virus DNA
contains a glucocorticoid-responsive
element. Proc Natl Acad Sci U S A 1986;
83: 1627–31.
38 Mya DH, Han ST, Linn YC, Hwang WY,
Goh YT, Tan DC. Risk of hepatitis B
reactivation and the role of novel agents
and stem-cell transplantation in
multiple myeloma patients with
hepatitis B virus (HBV) infection. Ann
Oncol 2012; 23: 421–6.
39 Takai S, Tsurumi H, Ando K,
Kasahara S, Sawada M, Yamada T et al.
Prevalence of hepatitis B and C virus
infection in haematological
malignancies and liver injury following
chemotherapy. Eur J Haematol 2005; 74:
158–65.
40 Lok AS, McMahon BJ. Chronic hepatitis
B: update 2009. Hepatology 2009; 50:
661–2.
41 Chen MH, Hsiao LT, Chiou TJ, Liu JH,
Gau JP, Teng HW et al. High prevalence
of occult hepatitis B virus infection in
patients with B cell non-Hodgkin’s
lymphoma. Ann Hematol 2008; 87:
475–80.
Supporting Information
Additional supporting information may be found in the online version of this article at the publisher’s web-site:
Table S1 Methodologic quality of studies included in the final analysis based on the Newcastle–Ottawa scale of quality
of case–control studies.
314
© 2016 Royal Australasian College of Physicians
Estimation of clinical and economic effects of prophylaxis
against venous thromboembolism in medical patients, including
the effect of targeting patients at high-risk
J. A. Millar1,2 and A. L. K. Gee3†
1
Department of Medicine, Albany Regional Hospital, Albany, 2Medical Education, Curtin University, and 3Department of Medicine, Royal Perth Hospital,
Perth, Western Australia, Australia
Key words
thromboprophylaxis, medical patients,
guidelines, selection criteria, cost-effectiveness.
Abstract
Correspondence
J. Alasdair Millar, Department of Medicine,
Albany Regional Hospital, Albany, WA 6331,
Australia.
Email: [email protected]
Aim: To estimate the effects of MT in Australia.
Methods: A decision tree model of MT was populated with national data for medical
Background: The clinical and economic effects of medical thromboprophylaxis (MT)
using low molecular weight heparin in Australia are unknown.
Received 7 December 2014; accepted
2 January 2016.
doi:10.1111/imj.12995
admissions. The Prevention of Recurrent Venous Thromboembolism (PREVENT) trial
was chosen as the primary data source because its design uniquely avoided bias caused
by treatment of sub-clinical events. Clinical efficacy and costs were estimated compared
with no prophylaxis, assuming full compliance and according to three definitions of eligibility. Effectiveness was estimated as thrombotic events saved, mortality from bleeding or pulmonary embolus (PE), cost and $/year of life saved. Model outputs were
subjected to sensitivity analysis.
Results: MT decreased thrombotic events, and the numbers avoided increased as eligibility broadened (deep vein thrombosis (DVT): 2597, 2771 and 3232 at restricted, intermediate and broad eligibility; PE: 454, 484 and 565 respectively). The annual cost of no
prophylaxis was $88.7 m. Costs were reduced at most restricted eligibility (−$7.9 m),
but increased by $3.0 and $32.1 m at broader eligibility. PE deaths declined, but this
was offset by deaths from haemorrhage, causing a net increase (158, 299 and
672 respectively). Estimates were sensitive to the incidence of venous thromboembolic
event (VTE), case-fatality rates for PE and bleeds and the relative risk reduction for PE
with prophylaxis.
Conclusions: Under PREVENT trial conditions, MT avoids up to 3200 DVT and 565 PE
events annually, but may increase mortality.
Introduction
Thromboprophylaxis using low molecular weight heparin (LMWH) is widely promoted and practised in medical
wards across Australia, but this activity occurs without
evidence of benefit for clinical end-points in randomised
controlled trials, all of which were powered to detect
sub-clinical events detected by imaging.1,2 Obtaining
secure knowledge for clinical end-points in the near
future is unlikely without a major collaborative effort,
given a sample size requirement of 200 000.3 In these
circumstances, we considered whether data on clinical
event rates reported in published trials could be used as
†Present address: Department of Public Health, Geraldton
Regional Hospital, Geraldton, Western Australia, Australia.
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
an indicator of the likely clinical benefit of medical
thromboprophylaxis (MT). However, in addition to
being underpowered for clinical events, all trials save
one were adversely affected by anticoagulant treatment
given in response to the detection of sub-clinical
thrombi, thereby altering the natural history. The exception is the Prevention of Recurrent Venous Thromboembolism (PREVENT) trial,4 which avoided this source
of bias by deferring imaging until at least day 14 of the
study. It is the largest trial of LMWH prophylaxis reporting clinical thrombotic end-points, but was underpowered for these events.
Several studies5–7 claim that MT is cost effective, but
these estimates are affected by the same problem, being
based mainly on the Prophylaxis in Medical Patients
with Enoxaparin (MEDENOX) study, a randomised controlled trial of enoxaparin versus placebo8 in which the
primary end-point was asymptomatic thromboses. The
315
Millar & Gee
Figure 1 Decision tree model of venous thromboembolic event events
with prophylaxis in eligible medical patients. Patients who do not have a
major bleed caused by prophylaxis may have a clinical deep venous
thrombosis (DVT), pulmonary embolus or no event. Post-thrombotic syndrome (PTS) may develop in patients after a DVT. The model was populated by clinical event rates from the Prevention of Recurrent Venous
Thromboembolism (PREVENT) study4 except for PTS incidence (not studied in the PREVENT trial). In the case of major bleeds, the rate in the
untreated group was set at zero and the net rate (prophylaxis less placebo rates) was applied in the treated group. Except for major bleeding,
the branching pattern is the same for non-eligible patients.
main incremental cost-effectiveness ratio obtained in
these analyses (cost per asymptomatic venous thromboembolic event (VTE) avoided) is of questionable clinical
significance and is not comparable with other medical
interventions.
Australian guidelines for MT9,10 promote ‘consideration’
of MT in patients with any risk factor and no bleeding
risk. However, an excess of major bleeding with LMWH
has been reported in most trials in spite of the exclusion
of patients at risk of bleeding.3 Thus broad eligibility for
MT, which by definition includes patients with weak risk
factors, may have a low benefit to risk ratio.11 Amending
eligibility for MT to patients with only strong risk factors
has been suggested,2,11–13 but the quantitative clinical and
economic effects of such a change are uncertain. If eligibility is restricted, more VTE events will occur in noneligible patients because their number will increase and
those with weaker risk factors achieve sub-group membership, and this could compromise the potential gain in
therapeutic efficiency.
With these points in view, and because of the importance of obtaining further information on the true value
of MT, we decided to model its risks and benefits based
on the PREVENT trial alone and to explore alternative
methods to minimise its uncertainties.
Methods
We constructed a decision tree model of thromboembolic
outcomes in adult medical patients admitted to all
Australian public hospitals14 in 2011–2012, given or not
316
given LMWH prophylaxis against VTE (Fig. 1), using an
Excel spreadsheet (Microsoft Corporation, Redmond,
WA, USA). The model treated deep vein thrombosis
(DVT) and pulmonary embolus (PE) as separate events
because they are reported as such in clinical trials,
including PREVENT. Cost-effectiveness was assessed
from the point of view of the public health system in
Australia, assuming full compliance in order to estimate
the maximum possible effects. The clinical outputs were
DVT, PE, post-thrombotic syndrome (PTS), major bleeds
and deaths from acute PE and major bleeds, and
corresponding costs.
The model contained three levels of eligibility for MT
according to the statistical weighting of known risk factors (Table 1): a relatively small population (25% of the
total) carrying the strongest risk factors for thrombosis
(malignancy, especially with chemotherapy; previous
history of thromboembolism and some rarer high-risk
conditions, such as acute inflammatory bowel disease)15;
an intermediate population with strong plus moderate
risk factors, such as cardiac or respiratory failure, sepsis
or inflammation (40%)15; and a larger group containing
all putative risk factors (80%)16 equivalent to Australian
guidelines9,10 and consisting essentially of the intermediate group plus patients satisfying an age-related criterion,
such as ‘age > 40’9 or ‘age > 60’.10 Thus, differences
between estimates from the intermediate and broad eligibility criteria are a guide to the effect of including an
age-related risk factor for MT. We refer to these subgroups by the proportion of the general inpatient medical population they represent, as above, or by the terms
© 2016 Royal Australasian College of Physicians
Medical thromboprophylaxis in Australia
Table 1 Eligibility for thromboprophylaxis used in the economic analyses†
Model
No.
1
2
3
Guidelines source
Eligible (%) (rounded
values used in the model)
Maximum DVT risk
relative to population
Absolute DVT risk in
eligible patients
Absolute DVT risk
(low-risk group)
Australia and New Zealand Working
Party on Thrombosis and
Haemostasis10
Australian NHMRC (without age
criterion)9
Millar et al.15
82% (80%)
1.25
0.00502
0.00143
37% (40%)
2.50
0.00860
25%
4.00
0.01290
†Absolute DVT risks in eligible patients (high-risk groups) were estimated as described in the text, using the base-case population clinical DVT risk
(0.0043). The ‘absolute DVT risk in eligible patients’ is the absolute baseline population DVT risk (0.0043) multiplied by the adjusted relative risk (see
text); DVT, deep vein thrombosis.
‘restricted’, ‘intermediate’ and ‘broad’. As no guidelines
currently recommend MT after discharge in medical
patients, our study was restricted to the inpatient phase
of management. The comparator was no prophylaxis
because enoxaparin is the only LMWH registered for MT
in Australia and is the drug of choice in most Australian
hospitals. In its absence, prophylaxis is unlikely. A
minority of hospitals may use unfractionated heparin,
but it is less effective17 and carries a higher risk of bleeding and heparin-induced thrombocytopenia.18 Orally
active thrombin and factor Xa inhibitors (dabigatran,
apixaban and rivaroxaban) are not registered for MT in
Australia.
Costs in the model (AU$) were of acquisition and
administration of enoxaparin, and treatment costs for
DVT, PE, PTS and of major bleeds (Table 2). The model
was populated by clinical outcome data from the PREVENT trial4 for reasons already stated and amplified in
the Discussion. Administration of enoxaparin was
assumed to occupy 5 min of nursing time, costed at
$2.73 using level 1.5 of the WA nurses award. We
assumed PTS did not curtail life expectancy, but PE was
estimated to shorten survival from 13 to 11 years, using
annual survival probabilities given in Australian Life
tables19 after applying a survival hazard ratio of 1.4.22
Cost estimates for PTS obtained from non-Australian
studies were sensitive to the time point of currency
exchange because of appreciation of the Australian versus the US dollar up to 2013 and were taken as the average of estimates in the year of publication and in 2013.
The health system cost of each death was taken as zero,
and hospital costs preceding each death were ignored,
leading to a possible modest cost underestimate. The net
present value of PTS and the loss of life expectancy for
non-fatal PE were obtained using a discount rate of 3%.
Economic effectiveness was estimated as dollars per year
of life saved ($/YOLS), estimated from the actuarial life
expectancy in Australia19 (13 years) at the mean age of
medical inpatients on admission (74 years),15 except
when net life years saved was negative (i.e. mortality
© 2016 Royal Australasian College of Physicians
increased with prophylaxis). Cost-effectiveness ratios for
morbid events (e.g. $/DVT avoided) were estimated in
all cases. No quality of life adjustments were made.
We assumed a baseline clinical DVT incidence across
the general medical inpatient population of 0.0043
(0.43%) as reported by us23 for ‘VTE’ (PE + DVT). For
the purposes of this study, we regarded this incidence as
being due to DVT only. This compares with the PREVENT study placebo-group (a selected sub-population of
the above) DVT incidence of 0.0063. The baseline PE
incidence was estimated from the ratio of symptomatic
PE to DVT incidence reported in the PREVENT study,4
giving a value of 0.0023. The relative risk reductions
(RRR) for symptomatic DVT and PE with prophylaxis,
major bleeding rates and case-fatality rates for PE and
major bleeding were as reported in the PREVENT study.
We included an assessment of the contribution to
overall clinical performance from non-eligible patients at
each definition of eligibility. This in turn requires an estimate of the VTE risk in non-eligible patients, which was
obtained from the known reciprocal relationship,
reported by us, between the proportion of patients in a
population selected by risk factor analysis and the maximum supplemental risk in that population.24 The derivation of this relationship assumes a limiting VTE risk of
zero in the low-risk group, but since events must occur
in that group, the actual supplemental risk in high-risk
groups must be less than the theoretical maximum, to
an extent that is unknown. We assumed arbitrarily that
the risk was two-thirds of the theoretical maximum.
With this adjustment, risk in non-eligible patients was
calculated using the formula for weighted average.24
The model was challenged by sensitivity analysis for
important assumptions, including baseline VTE risk up to
the highest value reported,25 and likely sources of uncertainty (Table 4), including the uncertainty from the
immediately above procedure and the effect of substituting independent literature values, such as the casefatality rates for PE26–28 and major bleeding,29–31 where
these values differed markedly from the PREVENT trial
317
Millar & Gee
Table 2 Demographic and incidence and costs (AU$) of disease conditions included in the economic model, and sources†
Factor or condition
Enoxaparin 40 mg/day
Rivaroxaban/day (30% of market)
Warfarin 5 mg (70% of market)
Demographic variables
Clinical DVT
Pulmonary embolus
Major bleed
PTS
Component
Source
Unit cost or value
Prophylaxis
Treatment of DVT
WA state drug tender
PBS
Mean age
Actuarial life expectancy (male and female
average)
Relative risk reduction (RRR)
Rate
Hospital costs
Post discharge anticoagulation (warfarin, 70%;
rivaroxaban, 30%) (4.5 m)
GP visits (3)
INR weekly
Millar et al.15
Australian Life tables19
RRR
Case-fatality rate after PE
Life years lost after PE
Hospital costs
Post discharge warfarin (6 months)
GP visits post discharge (4)
INR weekly
Net rate (LMWH minus placebo)
Mortality rate after major bleed
Hospital costs
Incidence after clinical DVT
Diclofenac + paracetamol
GP visits (4/year)
Topical relief (ointments, etc.)
PREVENT study4
PREVENT study
DOHA20
PBS
$4.66/day
$3.43/day
$0.28/day
74
13.5 years
0.56
0.0063
$6347.00
$1.23
MBS
MBS
Total DVT
PREVENT study
PREVENT study
See text
DOHA20
PBS
MBS
MBS
Total PE
PREVENT study
PREVENT study
Muntz et al.21
Total major haemorrhage
Assumed
PBS
MBS
Assumed, per week
Total PTS
Total cost
$36.30
$13.70
0.176
0.177
1.14
$8973.00
$0.28
$60.00
$13.70
0.0039
0.20
$14 289
0.3
$1.07
$36.30
$20.00
$6347.00
$165.38
$108.90
$172.62
$6793.90
$8973.00
$50.40
$240.00
$328.80
$9592.20
$14 289
$14 289
$4256.64
$1582.55
$11335.06
$17174.25
†Costs for PTS and future life years lost after PE were discounted at 3%. DOHA, Department of Health and Ageing, Canberra; DVT, deep vein thrombosis; GP, general practitioner; INR, international normalised ratio; MBS, Medicare Benefits Schedule, Australia (www.mbsonline.gov.au); PBS, Pharmaceutical Benefits Schedule, Australia (www.pbs.gov.au); PE, pulmonary embolus; PREVENT, Prevention of Recurrent Venous Thromboembolism; PTS, postthrombotic syndrome.
values or biased the results in a particular direction
(Table 3). This research did not involve access to individual patients. Ethical approval was not required.
Results
The number of overnight medical admissions in
Australia reported for 2012 by the Australian Institute of
Health and Welfare were 1 458 600.14 Patients eligible
for MT (Table 1) therefore numbered 364 650, 583 440
and 1 166 880 as eligibility criteria broadened.
Base-case model
Clinical outcomes
Estimated annual numbers of DVT, PE and PTS across
Australia and deaths from PE and major bleeding, with
318
associated costs and at 100% compliance are shown in
Table 3 for the base-case from each data source and at
each level of eligibility (25, 40 and 80%) and the comparator (no prophylaxis). This order of reporting is used
throughout the paper. The model predicted that 6272
DVT and 3385 PE will occur annually without prophylaxis and that MT decreases these numbers under all eligibility criteria (Table 3). However, the model predicted
an excess of major bleeds with prophylaxis and that
deaths from major bleeds exceeded those saved by
decreased PE under all definitions of eligibility under
PREVENT conditions.
As expected, events including death in non-eligible
patients increased as eligibility was progressively
restricted (Table 3), because that group contains more
untreated patients with risk factors for thrombosis, and
the numbers of events increase. This effect contributed
to the modestly greater total events seen under broad
© 2016 Royal Australasian College of Physicians
Medical thromboprophylaxis in Australia
Table 3 Detail of base-case estimates from the model for clinical end-points, end-points avoided by prophylaxis and corresponding costs, according
to eligibility expressed as the percentage of the inpatient medical population across Australia per year (n = 1 458 600)18 who are eligible according to
the criteria shown in Table 1, applying the output clinical data reported in the PREVENT study3 and assuming full compliance
Comparator (no prophylaxis)
0
Eligibility (%)
Measure
DVT
Foetal + non-fatal PE
PTS
PE deaths
Prophylaxis
Yes
No
Total
DVT avoided
Yes
No
Total
PE avoided
Yes
No
Total
PTS avoided
Yes
No
Bleeding deaths
Total deaths
25
40
80
n
$
n
$
n
$
n
0
6272
6272
$0
$42 611 205
$42 611 205
$14 314 438
$10 651 736
$24 966 174
$0
$26 559 632
$26 559 632
0
1882
1882
$0
$19 537 077
$19 537 077
0
616
0
616
—
—
—
2247
1254
3501
2771
2224
677
2901
484
674
376
1050
832
405
123
387
915
−299
2275
—
—
—
$15 268 617
$8 521 389
$23 790 006
0
3385
3385
2107
1568
3675
2597
2085
846
2931
454
632
470
1102
780
379
153
242
774
−158
1422
—
—
—
2622
418
3040
3232
2594
226
2820
565
787
125
912
970
472
41
775
1288
−672
3850
—
—
—
Deaths avoided
Major bleeds and costs
Drug and nursing costs
Total cost
Cost saving
Prophylaxis
0
—
—
$0
$0
$88 707 914
—
$16 359 920
$6 648 689
$23 008 609
$6 563 116
$4 883 781
$11 446 897
—
—
—
$10
$10
$80
$7
656
779
856
851
062
054
796
118
$17 450 448
$5 313 196
$22 763 644
$7 000 603
$3 907 025
$10 907 628
—
—
—
$17
$17
$91
−$3
054
246
761
053
135
486
899
985
$
$17 813 083
$2 840 463
$20 653 546
$20 358 510
$1 771 065
$22 129 575
$8 167 232
$1 302 342
$9 469 574
—
—
—
$34
$34
$120
−$32
097
492
842
134
182
973
850
935
Drug acquisition costs shown are for enoxaparin 40 mg daily. Events avoided are the difference between the total number of events at each level of
eligibility and the comparator. Negative results for ‘deaths avoided’ and ‘cost saving’ indicate an estimated net increase in deaths or cost, respectively,
with prophylaxis; DVT, deep vein thrombosis; PE, pulmonary embolus; PTS, post-thrombotic syndrome.
eligibility, but did not reverse measures of performance
expressed as total mortality, cost or $/YOLS.
eligibility each VTE avoided came at a cost ($938 and
$8463 respectively).
Cost outcomes
Annual cost of treating thrombotic events without MT
was $88.7 m. That cost was reduced modestly by MT by
$7.9 m at 25% eligibility, but increased by $3.0 and
$32.1 m at 40 and 80% eligibility. These trends were
due to increases in drug acquisition and administration
costs plus the cost of treating non-fatal major bleeds, at
each level of eligibility ($21.4, $34.3 and $68.6 m/year
respectively), offsetting the reduced cost of treating VTE
events (Fig. 2).
Economic performance
The outcome variable $/YOLS was indeterminate in the
base-case because of the excess of deaths and life years
lost. In some runs of the model, YOLS was positive in
spite of a small increase in mortality, because of the contribution to YOLS from the effect of MT on PE (Table 4).
Negative costs (i.e. cost savings) per non-fatal VTE event
(DVT plus PE) avoided were seen at 25% eligibility in
the base-case (−$2573/VTE avoided), but at 40 and 80%
© 2016 Royal Australasian College of Physicians
Sensitivity analyses
Table 4 displays the results of sensitivity analyses. We
noted a general tendency for declining performance as
eligibility was broadened, also seen clearly in Figure 2.
Of particular interest were:
1. Baseline incidence of VTE. The VTE benefit of MT
and aggregate costs was directly related to baseline VTE
incidence. At the maximum reported VTE incidence of
1.58%,25 performance was improved compared with the
base-case, but net mortality remained higher than comparator at intermediate (+93) and broad (+433) eligibility and showed only a modest decline (−34) at restricted
eligibility. Costs increased substantially but the qualitative pattern in favour of restricted eligibility remained (−
$1426, $841 and $51 308/YOLS). At an arbitrary value
of half the base-case VTE rate (which is below the least
reported incidence15 of 0.34%), total costs fell because
there are fewer VTE events, but the pattern in mortality
was exaggerated (Table 4). These trends arise because
319
Millar & Gee
MT in the base-case. The same manoeuvre applied to
major bleeds29–31 reduced deaths from major bleeding
and hence had the opposite effect. Under an assumption
of zero deaths from major bleeding, MT caused a reduction in deaths (Table 4), as expected, but increasing survivorship entailed an increase in costs from treatment of
non-fatal major bleeds.
3. Assumption regarding VTE risk in non-eligible
patients. Variation in this assumption altered the estimates of VTE events avoided, but had a modest effect
only on mortality and no qualitative effect on costeffectiveness expressed as $/YOLS (Table 4).
Discussion
The provenance of the PREVENT study
Figure 2 Base-case estimates for fatal and non-fatal pulmonary embolus (PE) and deep venous thrombosis (DVT) (upper panel), deaths from
PE and major bleeds (middle panel) and costs of treatment of DVT, nonfatal PE, post-thrombotic syndrome (PTS), major bleeds, prophylactic
drugs and nursing (lower panel) under three conditions of eligibility
(restricted, moderate and broad, equivalent to 25, 40 and 80% of the
inpatient medical population) as explained in the text, based on output
data from the PREVENT study4 and assuming 100% compliance.
the number of events, and events prevented by MT at a
fixed RRR, increases as the baseline incidence increases,
but the cost and clinical effects of major bleeding are
augmented as the treated population grows.
2. Case-fatality rates for PE and major bleeding. Both
these rates appeared high in the PREVENT study compared with literature values. Substituting more likely
(lower) values for PE case-fatality26–28 worsened the
increased mortality against the comparator seen with
320
The main limitation of our study is that the precision of
the PREVENT trial clinical data is low because of low
event numbers, indicating the possibility of chance variation from reality. The study was underpowered for clinical events, and the differences in event rates with
prophylaxis were not statistically significant. In addition,
some PREVENT trial outcomes were inconsistent with
results from independent studies. These facts make
extrapolation to a relatively large population hazardous,
because the uncertainty in the trial results is subject to a
multiplier effect. However, another design characteristic
of the PREVENT trial means, in our view, that it is the
only valid data source for our study. Of several randomised controlled trials of MT using LMWH, it alone
avoided bias caused by anticoagulant therapy for subclinical events, at least up to the 14th day of admission.
These features make our study controversial. We make
the following points:
1. The PREVENT study was a model randomised controlled trial, and the largest to date reporting clinical data
on DVT, PE and major haemorrhage. The data represent
real scientific measurements that confirm the very low
frequency of clinical thrombotic events in medical
patients. The trial DVT incidence is consistent with independent study estimates, and the reported RRR for DVT
with prophylaxis is similar to the RRR for sub-clinical
end-points. PE events were rare but, at about half the
DVT rate, consistent with other studies, although ascertainment bias may apply to PE data. We believe that the
PREVENT study clinical data deserve to be analysed and
disseminated. On the whole, these results appear acceptable for modelling.
2. The preceding claim cannot be made with confidence
for other PREVENT results used in our model. The incidence of major bleeding was greater than in other trials,
and the case-fatality rate was at the upper range of
© 2016 Royal Australasian College of Physicians
© 2016 Royal Australasian College of Physicians
25
Rationale
50% of base-case
Ref 13
Ref 32
Exploratory
Ref 32
Exploratory
50% of base-case
See legend
Twice base-case
PREVENT RRR for
DVT
Major bleed incidence 0.0012 and CFR
11.1%; PE RRR 32% and CFR 4.6%
33%
See Methods
Variation
0.00215
0.0158
0.0012
Zero
11.1%
Zero
9.05%
4.6%‡
32%
55.5%
1957
3255
3051
1468
10 786
2928
2923
2936
2936
2973
2991
3268
3746
VTE
−192
105
−158
−204
34
−14
74
−58
75
−204
−222
−94
12
Deaths
25%
NC
−$6896
NC
NC
−$1426
−$6387
−$3531
−$10 668
−$3500
NC
NC
−9647
−$2890
$/YOLS
2349
3480
3255
1566
11 506
3123
3118
3131
3131
3171
3191
3485
3996
VTE
−327
78
−299
−347
−93
−61
79
−133
80
−347
−367
−229
−116
Deaths
40%
NC
−$4515
NC
NC
$841
$4830
$1523
NC
$1509
NC
NC
NC
$2556
$/YOLS
3392
4079
3797
1827
13 423
3643
3643
3653
3853
3700
3723
4066
4661
VTE
Percentage of population eligible for prophylaxis
−688
7
−672
−728
−433
−188
92
−333
94
−728
−751
−590
−458
Deaths
80%
NC
$4918
NC
NC
$51 308
NC
$13 285
NC
$13 114
NC
NC
NC
NC
$/YOLS
†The upper limit of symptomatic DVT risk is the highest value reported to date and the lower limit is one-half the base-case value. For the other variables, the analysis was based on values present in
the literature,3 or were exploratory values designed to establish a frame of reference. Otherwise, input variables have base-case values. Negative values for deaths avoided indicate an increase in
deaths with thromboprophylaxis, and for $/YOLS, a cost saving per YOLS; ‘NC’ indicates that the ICER was not able to be calculated because MT caused an increase in life years lost; CFR, case-fatality
rate; DVT, deep vein thrombosis; ICER, incremental cost-effectiveness ratio; MT, medical thromboprophylaxis; PE, pulmonary embolus; RRR, relative risk reduction; VTE, venous thromboembolic event;
YOLS, year of life saved. ‡Weighted average of case-fatality rates reported.26–28
Risk adjustment for non-eligible group (66%)
Combined plausible or literature-supported values
RRR for PE with prophylaxis
Case-fatality rate after PE (17.7%)
Case-fatality rate after major bleed (18%)
Major bleed incidence (0.0033)
Variable (base-case value)
Population risk of DVT (0.0043)
Base-case result
Table 4 Estimates of VTE (DVT + PE) events avoided in survivors, net deaths avoided and economic performance ($/YOLS) according to eligibility for VTE prophylaxis (see Methods) in sensitivity
analysis for selected variables†
Medical thromboprophylaxis in Australia
321
Millar & Gee
reported values. However, all trials have reported an
added major bleeding risk,17,33 and in a recent Cochrane
meta-analysis,3 it is reported as statistically significant.
Major bleeding is unlikely to have a case-fatality rate of
zero, but the effect of removing its contribution is nevertheless shown in Table 4. PE events were low, as was the
reported RRR with prophylaxis, and the case-fatality
rates were substantially higher than independently
reported values.26–28 These are substantial difficulties
with a potential to bias model outputs for reasons given
above. Fortunately, independent high-quality epidemiological data exist for both PE and major bleeding.29–31
We felt that it was reasonable to apply the independent
data in sensitivity analysis to determine the likely range
of outputs (Table 4).
3. If the independent studies are reliable (which there is
no reason to doubt), the model outputs should move
closer to reality. When variables supporting reduced
mortality are deployed together in the model, a net
reduction is shown at 25 and 40% eligibility, but a finely
balanced effect occurs at 80% eligibility, which we note
is the eligibility described in current Australian
guidelines.9,10
4. The value of tentative extrapolation from uncertain
data lies in the capacity to generate new hypotheses, a
time-honoured scientific approach and indeed the main
objective of a study such as ours. As a result of this study,
we are able to confirm that morbid VTE events are likely
to be reduced by MT, but also to express a novel hypothesis that it may be associated with net mortality due to
major bleeding that offsets the expected mortality gain
from reduced fatal PE, especially if broad eligibility criteria are applied. This hypothesis provides a possible
alternative explanation of why no reduction in mortality
from MT has so far been documented.32,34 We note that
an identical offsetting and previously unrecognised effect
of haemorrhage has been found recently in the related
field, also involving LMWH, of bridging anticoagulation
for surgery in patients with atrial fibrillation.35
Qualitatively similar clinical details have been reported
in various meta-analyses, including a recent Cochrane
meta-analysis,3,17,33 which we initially considered as a
data source for our study but ultimately rejected. All the
trials but one (PREVENT) included in the Cochrane analysis3 applied treatment anticoagulation if imaging was
positive for DVT. In addition, major bleeding results in
active and placebo arms were in some cases unpaired,
and the selected trials included studies of unfractionated
heparin, which has a higher risk of bleeding than
LMWH. These aspects would have introduced important
additional sources of bias to our study, without easy
remedy. Although we concluded in favour of proceeding
with the PREVENT data, we recognise that the approach
322
has limitations. We regard our results as tentative, to be
interpreted cautiously after reasonable sensitivity analyses, used solely for hypothesis-generating activity, and
confirmed or refuted by a future clinical trial. We are
happy to provide results of any combination of input
variables to our model on request.
Study results
Our modelling suggests that MT is complex and likely to
reduce VTE events, but has a possible adverse effect on
mortality. The results are due in turn to model input
variables of uncertain validity from the PREVENT trial as
discussed above. The effect of most of these values is to
produce model outputs in favour of our emergent
hypothesis, and the effect of substituting more likely
values is to moderate but not eradicate the trends arising
from the PREVENT trial results unless all are deployed
together in the model or extreme assumptions are made,
for example, major bleeds have a case-fatality rate
of zero.
Trends in effectiveness across the range of eligibility
were qualitatively similar in all runs of the model.
Although associated with modestly greater numbers of
VTE events prevented, broad eligibility carries the penalties of higher drug acquisition and nursing costs, with
increasing cost and medical penalties from major haemorrhage, including possible death. Also, as eligibility for
MT is broadened to include progressively weaker risk
factors, one factor (increasing age) becomes important
but is controversial. Although age appears as a risk factor
for thrombosis in univariate analyses, the risk effectively
disappears in multivariate analysis.36,37 Thus, it can be
said that elderly patients have a predisposition to thrombosis because they suffer from diseases that are thrombogenic. On the other hand, age is a risk factor for bleeding
in patients with PE, atrial fibrillation or myocardial
infarction.38–40 It is possible that the increased major
bleeding rates found in clinical trials were due to the
deployment of prophylactic LMWH in elderly patients
who were not at real risk of thrombosis because their
only ‘risk factor’ was advanced age. The incidence of
major bleeding caused by MT is uncertain and may have
been unduly high in the PREVENT trial (0.0033). Major
bleeding compromises the clinical and economic standing of MT because the events are important by definition, treatment is costly and death is a possible outcome.
If age was removed as a risk factor, the proportion of
patients eligible for prophylaxis would be about 40%.15
Our results suggest that patients who are not eligible
for prophylaxis at each level of eligibility contribute to a
burden of VTE, an effect that expands as eligibility is
restricted. In fact, this is an inevitable result in any area
© 2016 Royal Australasian College of Physicians
Medical thromboprophylaxis in Australia
of medical practice where patients are dichotomised on
the basis of risk factor analysis as being at high or low
risk, and treated on that basis. In such a system, the
higher the risk threshold deemed to justify treatment,
the larger the untreated group will be and the more
events that will occur in that group. However, as we
have shown here, the existence of this reciprocal trend
does not have the effect of nullifying the increased efficiency of improved patient selection.
Clinical and economic performance improves as the
baseline incidence of events of interest increases, because
the number of events that can be avoided by prophylaxis
increases, but so do costs. This is noted for VTE in our
modelling even in the relatively narrow range reported
(0.34%15–1.58%25). Therefore, the true performance of
MT remains uncertain, but even at the highest reported
VTE incidence25 excess mortality was found at intermediate and broad eligibility under PREVENT conditions in
spite of greater benefit on morbid events.
An important underlying assumption in this study is
that MT is in fact effective. This has been questioned
recently in the Journal.1 However, it appears likely that
the RRR uniformly observed across all trials using
asymptomatic end-points will apply to symptomatic disease, as one follows the other. Benefit in clinical end-
References
1 Spencer A, Cawood T, Frampton C,
Jardine D. Heparin-based treatment to
prevent symptomatic deep venous
thrombosis, pulmonary embolism or
death in general medical inpatients is
not supported by best evidence. Intern
Med J 2014; 44: 1054–65.
2 Welfare M. NICE’s recommendations
for thromboembolism are not evidence
based. BMJ 2011; 343: d6452.
3 Alikhan R, Bedenis R, Cohen AT.
Heparin for the prevention of
venous thromboembolism in acutely ill
medical patients (excluding stroke
and myocardial infarction).
Cochrane Database Syst Rev 2014; 3:
CD003747.
4 Leizorovicz A, Cohen AT, Turpie AGG,
Olsson C-G, Vaitkus PT, Goldhaber SZ
et al. Randomized, placebo-controlled
trial of dalteparin for the prevention of
venous thromboembolism in acutely ill
medical patients. Circulation 2004; 110:
874–9.
5 Schadlich PK, Kentsch M, Weber M,
Kammerer W, Brecxht J, Nadipelli V
et al. Cost effectiveness of enoxaparin as
prophylaxis against venous
thromboembolic complications in
© 2016 Royal Australasian College of Physicians
6
7
8
9
points has been found in most trials in spite of them
being substantially underpowered, but none has
reported a mortality benefit. We are uncomfortable with
the notion that the clinical data should not be examined,
because this opinion causes the current uncertainty over
the clinical value of MT to be perpetuated, perhaps
indefinitely. We suggest that the necessary ethical equipoise has emerged to justify establishment of a national
study to establish the true clinical effectiveness of MT.
Our study also assumes that MT guidelines are fully
complied with, but in practice non-compliance is common.16,41 Non-compliance will compromise clinical benefits and cost-savings with MT, but will also reduce
detrimental effects.
Conclusion
Our modelling of the PREVENT trial data suggests that
MT decreases VTE events but may have an adverse effect
on mortality. It also suggests that patient selection criteria for MT should be made more restrictive, possibly by
removing current age-related criteria for MT as an independent risk factor.
acutely ill medical patients.
Pharmacoeconomics 2006; 24: 571–90.
Nuijten MJC, Berto P, Kosa J,
Nadipelli V, Cimminiello C, Spreafico A.
Cost-effectiveness of enoxaparin as
thromboprophylaxis in acutely ill
medical patients from the Italian NHS
perspective. Recenti Prog Med 2002; 93:
80–91.
Lamy A, Wang X, Kent R, Smith KM,
Gafni A. Economic evaluation of the
MEDENOX trial: a Canadian
perspective. Can Respir J 2002; 9:
169–77.
Samama MM, Cohen AT, Darmon J-Y,
Desjardins L, Eldor A, Janbon C et al. A
comparison of enoxaparin with placebo
for the prevention of venous
thromboembolism in acutely ill medical
patients. Prophylaxis in Medical Patients
with Enoxaparin Study Group. N Engl J
Med 1999; 341: 793–800.
National Health and Medical Research
Council. Clinical Practice Guideline for the
Prevention of Venous Thromboembolism
(Deep Vein Thrombosis and Pulmonary
Embolism) in Patients Admitted to
Australian Hospitals. Melbourne:
National Health and Medical Research
Council; 2009 [cited 2015 May 5].
Available from URL: http://www.
10
11
12
13
nhmrc.gov.au/_files_nhmrc/
publications/attachments/guideline_
prevention_venous_thromboembolism.
pdf
The Australia and New Zealand
Working Party on the Management and
Prevention of Venous
Thromboembolism. Prevention of Venous
Thromboembolism. Best Practice
Guidelines for Australia and
New Zealand, 4th edn. Sydney: Health
Education & Management Innovations
Pty Ltd; 2007.
Millar JA. Selection of medical patients
for prophylaxis of venous
thromboembolism based on analysis of
the benefit-hazard ratio. Intern Med J
2009; 39: 606–12.
D’Costa DF. Time to reconsider NICE
guidance on heparin prophylaxis in
medical patients. BMJ 2011; 343:
d7886.
Lederle FA, Zylla D, MacDonald R,
Wilt TJ. Venous thromboembolism
prophylaxis in hospitalized medical
patients and those with stroke: a
background review for an American
College of Physicians Clinical Practice
Guideline. Ann Intern Med 2011; 155:
602–15.
323
Millar & Gee
14 Australian Institute of Health and
Welfare, 2013. Admitted Patient Care:
Overnight Acute Care 2011–2012 [cited
2013 Apr 26]. Available from URL:
http://www.aihw.gov.au/haag11-12/
admitted-patient-care-overnight-acute/
15 Millar JA, Lett J, Bagley LJ, Densie IK.
Eligibility for medical
thromboprophylaxis based on riskfactor weights, and clinical thrombotic
event rates. Med J Aust 2012; 196:
457–61.
16 Gibbs H, Fletcher J, Blombery P,
Collins R, Watley D. Venous
thromboembolism prophylaxis
guideline implementation is improved
by nurse directed feedback and audit.
Thromb J 2011; 9: 7.
17 Wein L, Wein S, Haas SJ, Shaw J,
Krum H. Pharmacological venous
thromboembolism prophylaxis in
hospitalised medical patients. Arch
Intern Med 2007; 167: 1476–86.
18 Martel N, Lee J, Wells PS. Risk for
heparin-induced thrombocytopenia
with unfractionated and lowmolecular-weight heparin
thromboprophylaxis: a meta-analysis.
Blood 2005; 106: 2710–15.
19 Australian Bureau of Statistics.
3302.0.55.001 – Life Tables, States,
Territories and Australia, 2009–2011
[cited 2013 Jun 24]. Available from
URL: http://www.abs.gov.au/ausstats/
[email protected]/mf/3302.0.55.001
20 Australian Government Department of
Health and Ageing. Data Collections:
Round 14 (2009–10) National Public
Cost Weight Tables – version 6.0x and
version 5.2. 2012 Nov 9 [cited 2013 Jun
24]. Available from URL: http://www.
health.gov.au/internet/main/
publishing.nsf/Content/Round_14-costreports
21 Muntz J, Scott DA, Lloyd A, Egger M.
Major bleeding rates after prophylaxis
against venous thromboembolism:
systematic review, meta-analysis, and
cost implications. Int J Technol Assess in
Health Care 2004; 20: 405–14.
22 Klok FA, Zondag W, van
Kralingen KW, van Dijk APJ,
Tamsma JT, Heynig FH et al. Patient
outcomes after acute pulmonary
embolism. Am J Respir Crit Care Med
2010; 181: 501–6.
324
23 Millar JA, Lee GE, Ienco R. Prevalence
of venous thromboembolism in medical
inpatients. Med J Aust 2010; 192: 724–6.
24 Millar JA. Measures of risk and their
relationship to the relative size of a
high-risk group: application to medical
thromboprophylaxis. Clin Epidemiol
2012; 4: 275–81.
25 Edelsberg J, Hagiwara M, Taneja C,
Oster G. Risk of venous
thromboembolism among hospitalized
medically ill patients. Am J Health Syst
Pharm 2006; 63: S16–22.
26 Laporte S, Mismetti P, Decoussus H,
Uresandi F, Otero R, Lobo JL et al.
Clinical predictors for fatal pulmonary
embolism in 15,520 patients with
venous thromboembolism: findings
from the Registro Informatizado de la
Enfermedad TromboEmbolica Venosa
(RIETE) Registry. Circulation 2008; 117:
1711–16.
27 den Exter PL, van Es J, Klok FA,
Kroft LJ, Kruipe MJHA, Kamphuisen W
et al. Risk profile and clinical outcome of
symptomatic subsegmental acute
pulmonary embolism. Blood 2013; 122:
1144–9.
28 Nijkeuter M, Söhne M, Tick LW,
Kamphuisen PW, Kramer MH,
Laterveer L et al. The natural course of
hemodynamically stable pulmonary
embolism: clinical outcome and risk
factors in a large prospective cohort
study. Chest 2007; 131: 517–23.
29 Majeed A, Hwang HG, Connolly SJ,
Eikelboom JW, EWzekpwitz MD,
Wallentin L et al. Management and
outcomes of major bleeding during
treatment with dabigatran or warfarin.
Circulation 2013; 128: 2235–332.
30 Avas NT, Dodek PM, Wang H,
Fowler R, Wong H, Norena M.
Attributable length of stay and mortality
of major bleeding as a complication of
therapeutic anticoagulation in the
intensive care unit. J Patient Saf 2015;
11: 23–7.
31 Navarro JL, Cesar JM, Fernandez MA,
Fontcuberta J, Reverter JC, GolFreixa J. Morbidity and mortality in
patients treated with oral
anticoagulants. Rev Esp Cardiol 2007; 60:
1226–32.
32 Kakkar AK, Cimminiello CX,
Goldhaber SZ, Parakh R, Wang C,
Bergmann J-F. Low molecular weight
33
34
35
36
37
38
39
40
41
heparin and mortality in acutely ill
medical patients. N Engl J Med 2011;
365: 2463–72.
Dentali F, Douketis JD, Gianni M,
Lim W, Crowther MA. Meta-analysis:
anticoagulant prophylaxis to prevent
symptomatic venous thromboembolism
in hospitalized medical patients. Ann
Intern Med 2007; 146: 278–88.
Mahe L, Bergmann JF, d’Azemar P,
Vaissie JJ, Caulinj C. Lack of effect of a
low-molecular-weight heparin
(nadroparin) on mortality in bedridden
medical inpatients: a prospective
randomised double-blind study. Eur J
Clin Pharmacol 2005; 61: 347–51.
Douketis JD, Spyropoulos AC, Kaatz S,
Becker RC, Caprini JA, Dunn AS et al.
Perioperative bridging anticoagulation
in patients with atrial fibrillation. N Engl
J Med 2015; 373: 823–33.
Alikhan R, Cohen AT, Combe S,
Samama MM, Desjardins L, Eldor A
et al. Risk factors for venous
thromboembolism in hospitalised
patients with acute medical illness:
analysis of MEDENOX Study. Arch
Intern Med 2004; 164: 963–8.
Kobbervig CE, Heit JA, Petterson TM,
Bailey KR, Melton LJ. The effect of
patient age on the incidence of
idiopathic vs secondary venous
thromboembolism: a population-based
cohort study. Blood 2004; 104: 957a.
Hughes M, Lip GY. Risk factors for
anticoagulation-related bleeding
complications in patients with atrial
fibrillation. QJM 2007; 100: 599.
Kadakiab MB, Desai NR, Alexander KP,
Chen AY, Foody JM, Cannon CP et al.
Use of anticoagulant agents and risk of
bleeding among patients admitted with
myocardial infarction. J Am Coll Cardiol
2010; 3: 1166–77.
Mikkola KM, Patel SR, Parker JA,
Goldhaber SZ. Increasing age is a major
risk factor for haemorrhagic
complications after pulmonary embolus
thrombolysis. Am Heart J 1997; 134:
69–72.
Cohen AT, Tapson VF, Bergmann J-F,
Goldhaber SZ, Kakkar AK, Deslandes B
et al. Venous thromboembolism risk and
prophylaxis in the acute hospital care
setting (ENDORSE study): a
multinational cross-sectional study.
Lancet 2008; 371: 387–94.
© 2016 Royal Australasian College of Physicians
A tertiary hospital audit of opioids and sedatives administered
in the last 24 h of life
C. Douglas,1 M. Clarke,2 S. Alexander3 and M. Khatun4
1
Palliative and Supportive Care Service and, 2Internal Medicine, Royal Brisbane and Women’s Hospital, 4School of Public Health, University of
Queensland, Brisbane, and 3Internal Medicine, Rockhampton Hospital, Rockhampton, Queensland, Australia
Key words
opioid, sedative, benchmark, dying, tertiary
hospital.
Correspondence
Carol Douglas, Palliative and Supportive Care
Service, Royal Brisbane and Women’s Hospital,
Ned Hanlon Building, Butterfield Street, Herston,
Qld 4029, Australia.
Email: [email protected]
Received 9 January 2015; accepted 15
December 2015.
doi:10.1111/imj.12985
Abstract
Background/Aim: To audit the doses of opioids and sedatives administered to patients
in the last 24 h of life in an Australian tertiary hospital and compare results with doses
published in New Zealand (NZ) benchmarking studies and to examine the effect of caring for dying patients using a modified version of the Liverpool Care Pathway (mLCP)
in respect to doses of opioids and sedatives.
Methods: A retrospective chart audit of 102 patients who died in a tertiary hospital
was carried out over a 3-month period in 2011. Diagnosis, demographic patient characteristics, use of the mLCP, use of subcutaneous infusions and key symptoms were identified. Chi-squared and the non-parametric Mann–Whitney tests were applied to
compare the group differences for categorical and continuous variables as appropriate.
Parenteral morphine-equivalent daily dose (pMEDD) was calculated. A t-test assessed
the variable mean doses of medication and patient characteristics.
Results: Of the audited patients, 76.5% died of non-malignant disease. The overall
mean dose of midazolam was significantly lower compared with that of the NZ study
pMEDD (6.0 vs 20.7 mg). The overall mean dose of morphine benchmarked closely
with the NZ study (56.5 mg Australian study vs 47.8 mg NZ study). A total of 83% of
patients with a malignant diagnosis was supported with the mLCP compared with 51%
of patients with a non-malignant diagnosis.
Conclusion: The significance of the lower midazolam doses was postulated, including
the possibility of inadequate symptom control for patients with a non-malignant diagnosis. The use of the mLCP did not lead to the provision of higher doses of medications.
Introduction
The provision of palliative care in supporting dying
patients at the end of life is not a matter of debate and is
now endorsed as an essential element of care for all by
the World Health Assembly.1 This includes the recognition of the need to update, as appropriate, national
essential medicines lists in light of the recent addition of
sections on pain and palliative care medicines to the
World Health Organization Model List of Essential Medicines.2 There is a large body of evidence for the use of
opioids for the control of pain and dyspnoea.3 Similarly,
benzodiazepines are considered essential in the practice
of palliative care, and anti-psychotics, such as haloperidol, are commonly used for nausea, vomiting and delirium to best support the dying patient.
There is a paucity of literature regarding the practice
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
of caring for the dying in the acute tertiary setting, with
palliative care literature largely drawing on the population of patients under the care of specialist palliative care
services in dedicated units.
In 2005, a dedicated specialist palliative care service
was established in this 929-bed tertiary referral teaching
hospital. To better support generalists caring for dying
patients across a diverse and large clinical environment,
a clinical pathway for the dying patient (modified from
the Liverpool Care Pathway, mLCP)4 was implemented
across all clinical areas in 2009. The mLCP reached full
compliance with the Marie Curie Institute, Liverpool,
UK and was described as modified due to local formatting requirements for documents. The mLCP utilised full
clinician engagement processes, education and a clear
governance model of support.
Medication guidelines to support the mLCP were
developed by the specialist palliative care service in consultation with broad representation from across the facility to reflect the best practice palliation of dying
325
Douglas et al.
patients.5 After the introduction of the mLCP, and
despite rigorous governance to support the care of the
dying, concerns were ‘voiced’ as to whether patients on
the pathway might be medicated in excess of their needs.
To understand better the validity of this concern, an
audit of opioid and sedative prescribing for the dying
patient was undertaken.
Two recent studies in New Zealand (NZ) across 14 hospices benchmarked the use of opioids and sedatives in
the last 24 h of life to self-assess prescribing patterns.
Patients in the NZ studies were admitted under the care
of specialist palliative care teams. These studies6,7 also
included the use of the LCP to support patients. The
intention of this study was not to compare clinical practices across the two countries but to audit the doses of
opioids and sedatives administered to a patient in the last
24 h of life in an acute tertiary hospital in Australia, with
and without the utilisation of the mLCP, and to review
these doses with those published in the NZ benchmarking studies.
Methods
A retrospective clinical chart audit was carried out over
3 months on 102 patients who died consecutively in a
large tertiary teaching hospital in Australia. Following
the death certification of a patient in the facility, a notification was made to the study team by the Mortality
Review Officer. All deaths were included, regardless of
their setting, including the intensive care unit and emergency department. A low-risk ethics approval was
granted prior to the commencement of the audit. Charts
were audited by a single medical practitioner to maximise reliability of the information obtained.
Data collected included gender, age, dates of admission
and death, length of stay, primary diagnosis, comorbidities, utilisation of mLCP, utilisation of a continuous subcutaneous infusion (CSCI) and identification of clinical
symptoms by descriptors of pain, dyspnoea and delirium.
Demographic characteristics of patients and principal
diagnoses were identified to compare malignant and
non-malignant populations. The Chi-squared and the
Mann–Whitney tests were applied for categorical and
continuous variables as appropriate to compare these
group differences.
All doses of opioids, midazolam and haloperidol
administered in the last 24 h of life were recorded,
including all ‘as required’ or prn doses. No other sedative
medications were included in the audit. The morphineequivalent daily dose (parenteral) (pMEDD) was calculated according to the Edmonton method.8 This method
was chosen to allow comparison with other international
studies. The patients’ characteristics and the variation
326
in the mean doses of haloperidol, midazolam and
parenteral morphine were assessed using the t-test. The
distribution of the opioids was positively skewed. A logarithmic transformation was performed to approximate a
normal distribution, and a parametric t-test was applied
to compare the patients’ characteristics and the variation
in the mean doses of haloperidol, midazolam and parenteral morphine. When patients were prescribed sedatives
but they were not administered during the last 24 h of
life, the doses were treated as a nil dosage and excluded
from the analysis to match the method used in the NZ
study. Geometric mean was estimated and compared
with the NZ benchmarking studies.
Results
Table 1 describes the demographic and clinical features
of the 102 patients. The majority of patients studied died
from non-malignant disease (76.5%).
The median age of the deceased patients was 79 years,
with patients with a malignant diagnosis having a longer
length of stay than those with a non-malignant diagnosis
(13 vs 7 days; P-value = 0.024). Patients with a malignant diagnosis were more likely to be supported by the
mLCP than those with a non-malignant diagnosis (79.2
vs 53.2%). Patients supported with the mLCP were also
more likely to have symptom control medications delivered with a CSCI.
There was significant symptom prevalence in the last
24 h of life, with pain recorded in 46.5%, dyspnoea in
65.3% and signs of terminal agitation with descriptors of
confusion, distress and agitation recorded in 74.3% of
patients. The primary cause of death (from Death Certification) is shown in Figure 1.
Respiratory failure is listed as the major cause of death
(27.8%) followed closely by the group defined as cerebrovascular accident (CVA) or stroke and cerebral haemorrhage (18.6%). The predominant underlying
respiratory disease was pneumonia or chronic obstructive pulmonary disease.
A total of 84% of the patients audited received an opioid in the last 24 h of life. Figure 2 shows that the pattern of opioid use in this study follows patterns that are
very similar to the NZ benchmarking study, noting that
hydromorphone was not available for use in
NZ. Methadone has historically been the second opioid
of choice in NZ with the more recent introduction of fentanyl and oxycodone preparations. Morphine (52.0%)
was the most commonly used opioid in the study facility
followed by fentanyl (14%) and oxycodone (10.8%).
Most of the patients in this study received a single opioid
except one patient who received morphine and fentanyl
simultaneously.
© 2016 Royal Australasian College of Physicians
Opioid and Sedative Audit Dying Patient
Table 1 Patients’ demographic and clinical characteristics
Patients’ characteristics
Age, median (IQ)
Gender†, (%) (n)
Female
Male
Length of stay, median (IQ)
mLCP‡, (%) (n)
CSCI§, (%) (n)
Clinical symptoms, (%) (n)
Dementia
Pain‡
Dyspnoea‡
Delirium/confusion/agitation‡
Overall (n = 102)
Malignant (n = 24)
Non-malignant (n = 78)
P-value
79 (71.0, 86.0)
73.5 (63.5, 83.0)
79.5 (73.0, 86.0)
0.029
50.0 (48)
50.0 (48)
8 (4.8,14.0)
59.4 (60)
58.8 (57)
43.5 (10)
56.5 (13)
13.0 (5.8,36.5)
79.2 (19)
82.6 (19)
52.1 (38)
47.9 (35)
7.0 (4.0,12.5)
53.2 (41)
51.4(38)
0.473
0.024
0.024
0.008
14.7(15)
46.5 (47)
65.3 (66)
74.3 (75)
0.0
66.7 (16)
66.7 (16)
75.0 (18)
19.2 (15)
40.3 (31)
64.9 (50)
74.0 (57)
0.020
0.024
0.876
0.766
†Gender has six missing values. ‡mLCP, pain, dyspnoea, delirium/confusion/agitation each has one missing value. §CSCI has three missing values.
CSCI, continuous subcutaneous infusion; IQ, interquartile range; mLCP, modified Liverpool Care Pathway.
Respiratory Failure/Pneumonia/COPD
25.5
CVA/Cerebral Haem
18.6
Sepsis/Urosepsis
7.8
ACS/MI
6.9
Lung Cancer
6.9
Liver Failure
5.9
Renal Failure
Figure 1 Cause of death (Death Certificate).
Note: Other malignancies including endometrial, colorectal, squamous cell carcinoma and
glioblastoma multiforme. ACS, acute coronary
syndrome; BMT-GVHD, bone marrow transplant
graft versus host disease; COPD, chronic
obstructive pulmonary disease; CVA, cerebrovascular accident; MI, myocardial infarction.
5.9
Other malignancies
4.9
BMT-GVHD/Leukemia
4.9
Multiple Organ Failure
3.9
Miscellaneous
3.9
Melanoma
2.9
Colorectal cancer
2.0
0
5
10
15
Percentage
20
25
30
80.0
70.0
Percentage
60.0
50.0
40.0
30.0
20.0
Figure 2 Pattern of opioid use in an Australian
tertiary
hospital
compared
with
the
New Zealand benchmarking study.
© 2016 Royal Australasian College of Physicians
10.0
0.0
Morphine
Oxycodone
Methadone
Hydromorphone
Fentanyl
327
Douglas et al.
disease was much higher in this study, the actual total
number included in the analysis of sedatives was similar
between Australia and NZ (27 vs 22), but the mean
doses were dissimilar (4.8 vs 12.5 mg).
Midazolam was the preferred benzodiazepine prescribed in association with the mLCP. There was no significant difference between patients managed on the
mLCP who received a mean dose of 6.6 mg compared
with those not managed on the mLCP who
received 5.1 mg.
Haloperidol was the first-line antipsychotic prescribed
at the study facility. No audited patients were prescribed
levomepromazine or other major tranquillisers. The
overall mean dose of haloperidol was very similar in
both studies (2.4 mg NZ study vs 2.7 mg Australian
study) as shown in Table 4. There was no significant difference between the doses of haloperidol for patients
cared for with and without the mLCP (1.3 vs 2.3 mg).
The overall mean dose of pMEDD was close to that of
the NZ benchmarking study (56.5 mg Australian study
vs 47.8 mg NZ study) as shown in Table 2. Both studies
showed that patients with malignant disease received on
an average significantly higher doses of pMEDD than
patients with non-malignant disease (Australian study:
99.1 vs 46.3 mg; (P-value 0.002) and NZ study 53.8 vs
26.0 mg; P-value <0.001). In comparison with the NZ
study, irrespective of diagnosis, all patients were administered higher doses of pMEDD (malignant diagnosis:
99.1 mg Australian study vs 53.8 mg NZ study; nonmalignant diagnosis: 46.3 mg Australian study vs
26.0 mg NZ study). The mean pMEDD for those patients
managed according to the mLCP was 61.1 mg, (range:
47.0–79.5 mg) with no significant difference from the
pMEDD for those not managed on the mLCP (49.6 mg,
range: 32.9–74.7 mg). There was a close correlation of
pMEDD for patients cared for with the mLCP and LCP
(NZ) (61.1 mg Australian study vs 63.1 mg NZ study).
Midazolam was the benzodiazepine provided to the
majority of patients in this study. Two patients received
clonazepam, and two patients received temazepam; both
drugs were excluded from the calculations. As presented
in Table 3, the overall mean midazolam dose in the
Australian study setting was significantly lower than that
in the NZ study (6.0 vs 20.7 mg). Patients with a nonmalignant diagnosis received significantly lower doses of
midazolam compared with patients with a malignant
diagnosis (4.8 vs 9.2 mg; P-value = 0.006). The exclusion of 0 mg doses of medication (i.e. where prescribed
but not administered) had the effect of significantly
reducing the number of patients with non-malignant
conditions included in the analysis of sedatives. Whilst
the proportion of patients dying from non-malignant
Discussion
The majority of Australians die each year in acute tertiary hospitals. In 2011–2012, 51% of all deaths occurred
in an admitted-patient setting, and 39.5% had been a
palliative care patient in their final admission.9
This study was unique in that there has been little evidence regarding the clinical characteristics and diagnoses
of those admitted patients previously. This study
included all consecutive deaths at a single site, an acute
tertiary hospital where there is palliative care consultation only. This is in contrast to the NZ study where the
audited patients were under the care of specialist palliative care teams in multiple settings, including at home
and in-patient palliative care units.
Table 2 Patients’ characteristics and the average doses of parenteral morphine (pMEDD) (in milligrams) comparison between the Australian tertiary
hospital and the New Zealand benchmarking study
Patients’ characteristics
Parenteral morphine daily dose(pMEDD)†
Tertiary hospital Australia
Overall
Gender‡
Female
Male
Malignant
Yes
No
mLCP§
Yes
No
New Zealand benchmarking study
n
Geometric mean (95% CI)
P-value
Geometric mean (95% CI)
P-value
77
56.5 (45.5, 70.1)
—
47.8 (41.9, 54.6)
—
36
35
55.9 (39.8, 78.3)
53.9 (39.4, 73.7)
0.877
51.8 (42.6, 63.0)
44.4 (37.0, 53.3)
0.256
20
57
99.1 (66.6, 147.2)
46.3 (36.3, 59.0)
0.002
53.8 (46.6, 62.1)
26.0 (19.1, 35.3)
<0.001
51
25
61.1 (47.0, 79.5)
49.6 (32.9, 74.7)
0.374
63.1 (47.0, 84.6)
44.4 (38.3, 51.6)
0.035
†pMEDD has one missing value. ‡Gender has six missing cases. §mLCP has one missing value. ; CI, confidence interval; mLCP, modified Liverpool Care
Pathway; pMEDD, parenteral morphine-equivalent daily dose.
328
© 2016 Royal Australasian College of Physicians
Opioid and Sedative Audit Dying Patient
Table 3 Patients’ characteristics and the average doses of Midazolam (in milligrams) comparison between the Australian tertiary hospital and the
New Zealand benchmarking study
Patients’ characteristics
Midazolam
Tertiary hospital Australia
Overall
Gender‡
Female
Male
Malignant§
Yes
No
mLCP¶
Yes
No
n
Geometric mean (95% CI)
41
6.0 (4.8, 7.6)
22
16
5.3 (3.7, 7.5)
7.1 (4.8, 10.2)
14
27
30
10
New Zealand benchmarking study
n
Geometric mean†
P-value
209
20.7
—
0.257
103
106
13.6
13.1
0.823
9.2 (6.3, 13.2)
4.8 (3.6, 6.3)
0.006
135
22
15.2
12.5
0.359
6.6 (4.9, 8.8)
5.1 (3.3, 7.5)
0.320
46
24
15.3
9.5
0.001
P-value
†95% CI of the geometric mean is not available in the New Zealand study. ‡Gender has three missing cases. §In the New Zealand study, malignant and
non-malignant patients died under the care of hospice (variable settings including home). ¶mLCP has one missing case. mLCP, modified Liverpool Care
Pathway.
Table 4 Patients’ characteristics and the average doses of haloperidol (in milligrams) comparison between the Australian tertiary hospital and the
New Zealand benchmarking study
Patients’ characteristics
Haloperidol
Tertiary hospital Australia
Overall
Gender‡
Female
Male
Malignant
Yes
No
mLCP§
Yes
No
New Zealand benchmarking study
n
Geometric mean (95% CI)
P-value
n
Geometric mean†
P-value
53
2.4 (2.0, 2.9)
—
220
2.7
—
27
23
2.4 (1.8, 3.1)
2.4 (1.7, 3.3)
0.969
112
108
1.3
1.3
0.819
20
33
3.0 (2.3, 4.0)
2.1 (1.6, 2.7)
0.064
193
27
1.3
1.0
0.189
36
17
1.3 (1.1, 1.4)
2.3 (1.5, 3.3)
0.616
72
66
1.8
1.0
<0.001
†95% CI of the geometric mean is not available in the New Zealand study. ‡Gender has three missing cases. §In the New Zealand study, malignant and
non-malignant patients died under the care of hospice (variable settings including home). CI, confidence interval; mLCP, modified Liverpool Care
Pathway.
The majority of patients in this study had a nonmalignant primary diagnosis (76.5%), with the majority
dying of respiratory or cerebrovascular disease. Across
Australia in 2011–12, 73.8% of patients with a primary
diagnosis of malignancy received specialist palliative care
in their final admission compared with 24% of patients
with a primary non-malignant diagnosis.9
In both, this study and the NZ study, the mean dose of
opioids for patients with malignant diagnoses is significantly higher than those with non-malignant diagnoses.
This phenomenon might be explained by the longerterm use of opioids in patients with cancer, with the
© 2016 Royal Australasian College of Physicians
inherent pain and symptom burden leading to death.
Whilst such a requirement has not been specifically studied, Sykes et al. have compared the rate of increase in
opioid doses leading to death in four studies of patients
mostly with malignancy in palliative care settings. They
demonstrated that 42–76% of patients were using
opioids prior to the dying phase.10
Patients supported by specialist palliative care were
not specifically identified in this study, but the use of the
mLCP acted as a surrogate marker. In the context of
being dying in an acute hospital, the study examined the
329
Douglas et al.
use of opioids and sedatives and the relationship of the
use of the mLCP to these medications.
There has been more recently intense scrutiny of the
LCP in the UK, resulting in an enquiry and its resultant
phasing out. As a consequence, there have been significant international developments to ensure more robust
and compassionate processes for the care of the dying.11
Of note in this study, having a non-malignant diagnosis
demonstrated a reduced likelihood of care with the
mLCP in comparison with patients with cancer diagnoses. The concerns initially voiced regarding the introduction of the mLCP within the acute facility and the
administration of medications to support care with the
mLCP were not supported by the results in this study. In
relation to the effect on the pMEDD, with the use of the
mLCP, there was no demonstrated difference in pMEDD.
For those patients with malignancy, the mean doses of
pMEDD were very similar to the NZ opioid benchmark
study.
In this study, patients with malignancy were largely
cared for within a designated cancer care unit and were
more likely to have been supported with the mLCP. The
transition to dying is more likely to be recognised in a
patient with malignancy, thus meeting the criteria for
dying with the mLCP.12
There was significant symptom burden in the study
patients in the last 24 h of life. A total of 65.3% of the
study patients experienced dyspnoea. Reassuringly, in
respect to use of opioids, the overall mean pMEDD provided to these patients indicated recognition of the positive effect of an opioid to palliate dyspnoea.13
Evidence of distress and agitation in the last 24 h of
life was predominant, with documentation in 74.3% of
patients.
There were significant limitations with the interpretation of the audit data. Symptom recording and symptom
relief were not able to be linked to the administration of
a medication; hence, relief of a symptom was not able to
be assessed.
The mean dose of midazolam administered in this
study was 6.0 mg compared with 20.7 mg in the NZ
study. It is postulated that this difference may have indicated a reluctance to administer sedatives for symptom
control. Alternatively, the variance may reflect a difference in prescribing patterns between the NZ specialist
palliative care or hospice providers and the generalist
setting within an acute tertiary hospital. However, it is
noted that in an international multi-centre study of sedation by Fainsinger et al.14 and an Australian study by
Good et al.15 the mean midazolam doses were remarkably similar to the mean midazolam doses in NZ. The
results may highlight a stigma related to use of sedatives15 and a resulting reluctance to manage dyspnoea
330
and agitation in patients with a non-malignant diagnosis
with sedatives. However, without more comprehensive
symptom relief data, it is not possible to draw such a
conclusion in this study.
It is recognised that prognostication in the group of
patients with chronic disease or non-malignant diagnoses is more difficult due to the inherently unpredictable nature of end-stage chronic disease.16 The concept of
a transition to dying or ‘diagnosing dying’ has been
described previously as being essential to the provision
of optimal symptom control in the context of treatment
in an acute facility.12 However, for the patient with a
non-malignant condition, this may not be achievable.
Providing care that synchronously supports judicious
active treatment that ameliorates the underlying condition with symptom control would obviate the need for
determining whether a patient was in a terminal phase.
Palliation and clinical management of a condition should
not be mutually exclusive when the treatments are not
burdensome or deemed futile. In this context, the recognition of a transition to dying would not be pivotal to the
provision of a ‘good death’. There might rather be the
recognition of clinical deterioration and the need for palliation, which identifies the palliative phase of the illness.
Other research has demonstrated that there is a need to
identify better the palliative phase of non-malignant
conditions so as to undertake appropriately advance care
planning, ensure access to symptom management and
provide support to patients and their families.17
This study has implications for those caring for patients
with chronic non-malignant disease with the evidence of
significant symptom burden in the last 24 h of life.
Whilst there are significant limitations within this study,
a previous study in a large tertiary hospital in Australia
confirmed that the majority of dying patients were not
referred for palliative care support, and there were
reported concerns regarding evidence of deficits in the
care of these dying patients.18 Given that patients with a
primary non-malignant diagnosis are significantly less
likely to be referred to palliative care for symptom support, the onus to provide the best care of these patients
‘rests’ with the treating physician.
Conclusions
This study was initiated to understand the doses of
opioids and sedatives administered to dying patients in
an Australian tertiary facility and to review the doses
prescribed to dying patients managed on the mLCP.
A comparison was made with previous benchmarking
studies of patients cared for by specialist palliative care.
There was no significant difference of medication doses
with the use of the mLCP. The diagnoses of CVA and
© 2016 Royal Australasian College of Physicians
Opioid and Sedative Audit Dying Patient
respiratory failure were the most common causes of
death. Dyspnoea and distress were highly prevalent in
the last 24 h of life. The most common opioid prescribed
was morphine, and the overall mean pMEDD was comparable with the NZ study.
The lower doses of sedatives prescribed to patients
with non-malignant disease in this study may be interpreted as a reluctance to administer sedatives but may
also reflect differences between the management of
malignant and non-malignant disease and between the
generalist and specialist settings.
The difficulty in diagnosing the transition to dying in
this group of patients, coupled with concerns of inducing
deterioration, may have limited clinicians optimally prescribing and administering sedative medications for the
management of distress and agitation. An alternative
paradigm of dual care is suggested.
References
1 Strengthening of Palliative Care as a
Component of Comprehensive Care
Throughout the Life Course. Geneva:
Sixty-Seventh World Health Assembly
WHA67.19. 2014 May 24.
2 WHO model list of essential medicines
18th list, 2013. Available from URL:
http://www.who.int/medicines/
publications/essentialmedicines/en/
index.html
3 Robert, Twycross, Andrew, Wilcock.
PCF3 Palliative Care Formulary, 3rd edn.
Nottingham: palliativedrugs.com Ltd;
2010.
4 Ellershaw J, Wilkinson S. Care of the
Dying: A Pathway to Excellence, 2nd edn.
Oxford: Oxford University Press; 2011.
5 Therapeutic Guidelines Palliative Care
Version 3. Melbourne: Palliative Care
Expert Group, Therapeutic Guidelines
Limited; 2012.
6 Ensor BR, Middlemiss TP.
Benchmarking opioids in the last
24 hours of life. Intern Med J 2011; 41:
179–85.
© 2016 Royal Australasian College of Physicians
As the majority of Australians die in acute facilities,
the results of this study may be significant. Whilst there
is limited evidence of the effect of medication on symptom relief and the quality of these deaths, there is a need
for further research, particularly in respect to the quality
of deaths of patients with chronic, non-malignant
disease.
Acknowledgements
The authors acknowledge the close support of Carol Parker, Mortality Review Coordinator and Marcella
Choudhury, Interviewing Officer at the Royal Brisbane
and Women’s Hospital
7 Ensor BR, Cohen D. Benchmarking
benzodiazepines and antipsychotics in
the last 24 hours of life. N Z Med J 2012;
125: 1–30.
8 Lawlor P, Pereira J, Bruera E. Dose
ratios among different opioids:
underlying issues and an update on the
use of the equianalgesic table. In: Topics
in Palliative Care. UK: Oxford University
Press; 2001, 247–76.
9 AIHW. Palliative Care Services in Australia
2014, Cat no HWI 128. Canberra,
Australia: AIHW. Australian Institute of
Health and Welfare; 2014.
10 Sykes N, Thorns A. The use of opioids
and sedatives at the end of life. Lancet
Oncol 2003; 4: 312–8.
11 Hardy JR, Good P. A good death in
hospital. Intern Med J 2014; 44: 313–4.
12 Ellershaw J. Care of the dying patient:
the last hours or days of life. BMJ 2003;
326: 30–4.
13 Currow D, Ward AM, Abernethy A.
Advances in the pharmacological
management of breathlessness. Curr
Opin Support Palliat Care 2009; 3: 103–6.
14 Fainsinger RL, Waller A, Bercovici M,
Bengtson K, Landman W, Hosking M
et al. A multicentre international study
of sedation for uncontrolled symptoms
in terminally ill patients. Palliat Med
2000; 14: 257–65.
15 Good PD, Ravenscroft PJ, Cavenagh J.
Effects of opioids and sedatives on
survival in an Australian inpatient
palliative care population. Intern Med J
2005; 35: 512–7.
16 Coventry PA, Grande GE, Richards DA,
Todd CJ. Prediction of appropriate
timing of palliative care for older adults
with non-malignant life-threatening
disease: a systematic review. Age Ageing
2005; 34: 218–7.
17 Luckett T, Philips J, Agar M, Virdun C,
Green A, Davidson PM. Elements of
effective palliative care models: a rapid
review. BMC Health Serv Res 2014;
14: 136.
18 Le BHC, Watt JN. Care of dying in
Australia’s busiest hospital: benefits of
palliative care consultation and
methods to enhance access’. J Palliat
Med 2010; 13: 855–60.
331
Associations of demographic and behavioural factors with
glycaemic control in young adults with type 1 diabetes mellitus
J. K. Osan,1 J. D. Punch,1 M. Watson,1 Y. X. Chan,1,2 P. Barrie,3 P. G. Fegan2 and B. B. Yeap1,2
1
School of Medicine and Pharmacology, University of Western Australia, 2Department of Endocrinology and Diabetes, and 3Diabetes Education Unit,
Fiona Stanley and Fremantle Hospitals, Perth, Western Australia, Australia
Key words
type 1 diabetes, young adult, glycosylated
haemoglobin A, psychosocial factor, behaviour.
Correspondence
Bu B. Yeap, School of Medicine and
Pharmacology, Harry Perkins Institute of
Medical Research, Fiona Stanley Hospital,
Murdoch, WA 6150, Australia.
Email: [email protected]
Received 21 July 2015; accepted 23
December 2015.
doi:10.1111/imj.12991
Abstract
Background: Despite recognised benefits of optimal glycaemic control in patients with
type 1 diabetes mellitus (T1DM), good control is still difficult to achieve, particularly for
adolescents and young adults. Recognition of factors that may assist early optimisation
of glycaemic control is therefore important.
Aims: We explored associations of demographic, social and behavioural factors with
glycosylated haemoglobin (HbA1c) levels in participants with T1DM aged 18–25 years.
Methods: A cross-sectional analysis was performed on young adults attending a dedicated multidisciplinary clinic at Fremantle Hospital, Western Australia from January to
August 2014.
Results: Data from 93 participants were analysed. Mean age was 21.4 2.3 years, and
39.8% of the cohort were female. Longer duration of diabetes was associated with
higher HbA1c (r = 0.25, P = 0.04). Men had lower HbA1c than women (8.2 1.6 vs
9.2 2.0%, P = 0.01). Increased frequency of clinic attendance was associated with
lower HbA1c (r = −0.27, P = 0.02). Those engaged in work or study had better HbA1c
compared with those who were not (8.9 2.1 vs 10.5 2.1%, P = 0.03). Socioeconomic disadvantage, risk-taking behaviour, insulin pump use and distance travelled to
clinic were not associated with differences in HbA1c.
Conclusion: In young adults with T1DM, geographical separation, socioeconomic disadvantage and risk-taking behaviours did not influence glycaemic control. Longer duration of diabetes identifies young adults at higher risk of poor control, while attendance
at a multidisciplinary clinic and engagement in work or study was associated with better glycaemic control. Additional studies are warranted to clarify the role of behavioural
interventions to improve diabetes management in young adults.
Introduction
Australia is ranked amongst the top 10 countries in the
world for its incidence and prevalence of type 1 diabetes
mellitus (T1DM) in children and young adults.1 Diabetes
as a whole represents the fourth largest contributor to
the overall burden of disease in Australia, placing it
within the top eight national health priority conditions.2
In adults with T1DM, glycaemic control reduces diabetic
complications, which is one of the principal goals of diabetes management.3 Studies of young adults with
T1DM have shown that a period of tight glucose control
is associated with long-term reduction in microand macro-vascular complications and improved betacell function.4–6 The Australian Diabetes Society
Funding: None.
Conflict of interest: None.
332
recommends a target glycosylated haemoglobin (HbA1c)
of <7%; however, individualisation based on patient factors needs to be considered.3 Thus, it is important to
identify factors that can positively or negatively influence glycaemic control and in turn modulate risk of
complications and burden of disease.
The period of transition from a paediatric to adult diabetes clinic is marked with uncertainty and inconsistent
behavioural patterns that may impact diabetes self-care.7
Adherence to treatment is suboptimal in the young adult
population who are irregular attendees at outpatient
clinics.7,8 Furthermore, young adults with T1DM have
increased levels of psychological distress.9 Youth of lower
socioeconomic status are less likely to use adaptive coping strategies, suggesting that this demographic may
require additional support and services.10 These factors,
amongst others, contribute to the higher HbA1c seen in
many young adults7,11,12 and increase the likelihood of
© 2016 Royal Australasian College of Physicians
Type 1 diabetes control in young adults
developing long-term complications. The transition
between child and adult services is recognised as a highrisk period, although studies assessing the impact of
healthcare transition on outcomes for young people with
T1DM are limited.8 Thus, determining factors that act as
barriers for achieving optimal glycaemic control in youth
may help to improve the quality of services provided.
The aim of this study was to analyse physical, psychosocial and behavioural factors contributing to glycaemic
control in patients aged 18–25 years with T1DM who
attended a dedicated multidisciplinary diabetes clinic.
We examined associations of socio-demographic
factors, clinic attendance, mode of insulin delivery,
engagement in risk-taking behaviours and occupational
and educational status with glycaemic control in this
population.
Methods
Study cohort
We conducted a systematic review of medical records for
patients aged 18–25 attending the Fremantle Hospital
Young Adult Clinic (YAC) with T1DM from 1 January
2014 to 31 August 2014. Fremantle Hospital is a tertiary
adult teaching hospital that services the Southern Metropolitan Region of Perth, Western Australia. The Fremantle Hospital YAC population comprises direct referrals to
a tertiary adult diabetes centre and the transition patients
from Princess Margaret Hospital, the major tertiary paediatrics centre in Western Australia. The transition
patients from Princess Margaret Hospital make up
approximately 75% of the attendees at the YAC. A standardised data collection tool was used to record medical,
social and behavioural information from medical records
by study investigators. Exclusion criteria included those
aged <18 or >25 and those who did not attend the YAC
within the study time frame. For patients with repeat
visits, data from the most recent visit were analysed. The
study was approved by the Clinical Governance Unit,
Fremantle Hospital.
Risk-taking behaviours (defined as smoking, excessive
alcohol intake or illicit drug use), involvement in physical activity, eating behaviours (recorded as missing
meals, body image issues or advice to change from
healthcare professionals), polypharmacy and the use of
an insulin pump versus injected insulin were assessed.
Excessive alcohol consumption was defined as >14
standard drinks per week for males and >7 standard
drinks per week for females. Illicit drugs were defined as
illegal drugs of abuse. Smoking was recorded if it had
been noted in the patient’s file within the previous
year from the date of data collection. Other variables
© 2016 Royal Australasian College of Physicians
that were assessed include study or employment status,
Index of Relative Socioeconomic Disadvantage (IRSD) –
obtained through the Australian Bureau of Statistics
(ABS)13 and distance of the patient’s primary residence
from the YAC – obtained through Google maps (https://
www.google.com.au/maps). Other social factors, such
as relationship status, driver’s licence status, hobbies
and several first-degree relatives with T1DM, were collated. HbA1c was used as a measure of glycaemic
control.
Data collection and statistical analysis
Data were collected using a dedicated data collection tool
(Appendix I). This data collection tool was developed by
the Fremantle Hospital Department of Endocrinology
and Diabetes and was designed to capture information
regarding glycaemic control, diabetes-related complications, method of insulin administration and behavioural
and social factors affecting diabetes care.14 This was used
by the study investigators (JO, JP and MW) to extract
and record relevant information from hard copy hospital
medical records. Data entry methods were standardised
and tested for reproducibility on a weekly basis by having each recorder assess the same medical record blinded
to the results of the others. There was <10% variance in
data collection between recorders over the course of the
study. We used the Statistical Package for Social Sciences
(version 22.0, IBM Corp, Armonk, NY, USA) to analyse
the data. Comparison of means was performed using
one-way analysis of variance. Bivariate correlations were
performed using Spearman’s rank correlation coefficient.
We considered two-tailed P-values <0.05 to be
significant.
Results
Baseline characteristics of the study cohort
Between January and August, 97 patients attended the
YAC at Fremantle Hospital. Of these patients, four were
excluded as they were >25 years at the time of data collection, leaving 93 who were included in the analyses.
Of the 93 patients, 56 (60.2%) were male (Table 1).
Mean age of participants was 21.4 2.3 years, with the
average duration of diabetes being 10.0 6.0 years.
Twenty patients were receiving insulin through a subcutaneous pump (21.5%). Mean duration of diabetes was
lower in males (8.1 5.7 for males vs 12.9 5.4 for
females). Retinopathy was recorded in three patients
(one male, two female), and risk-taking behaviours,
including excessive consumption of alcohol, smoking
and illicit drug use, were recorded in 23 patients
333
Osan et al.
Table 1 Baseline characteristics of the study cohort. Data expressed as number (%) and mean SD
n (%)
Pump users (%)
Age (years)
BMI (kg/m2)
Time since diagnosis (years)
HbA1c (%)
Retinopathy (%) (n = 65)
Smoking (%) (n = 76)
High-risk alcohol (%) (n = 77)
Drug use (%) (n = 78)
All risk taking (%) (n = 93)
Distance from clinic (km) (n = 93)
Males
Females
Total
56 (60.2)
8 (8.8)
21.5 2.4
24.8 4.5
8.0 5.6
8.24 1.35
1 (1.5)
9 (11.8)
7 (9.1)
7 (8.9)
15 (16.1)
10.65 12.01
37 (39.8)
12 (13.2)
21.2 2.1
23.8 4.1
12.8 5.4
9.23 2.01
2 (3.0)
4 (5.2)
4 (5.2)
3 (3.8)
8 (8.6)
14.64 22.94
93 (100)
20 (22.0)
21.4 2.3
24.4 4.4
9.9 6.0
8.64 1.71
3 (4.6)
13 (17.1)
11 (14.3)
10 (12.8)
23 (24.7)
12.31 17.40
BMI, body mass index; HbA1c, glycosylated haemoglobin.
(excessive alcohol intake in 11, smoking in 13 and illicit
drug use in 10). Participants lived an average of 12.3 km
from the clinic.
Associations of sex, occupational and
educational status and method of insulin
delivery with glycaemic control
HbA1c was lower in males compared with females
(8.2 1.4 vs 9.2 2.0%, P = 0.012) (Table 2). Those
engaged in full- or part-time work and/or study had
lower HbA1c compared with those not engaged in either
(8.4 1.6 vs 9.7 2.0%, P = 0.022). There was no difference in HbA1c for those using insulin pumps compared with those who were on basal-bolus insulin
(Table 2). There was no difference in HbA1c when comparing those reporting risk-taking behaviour to those
who did not (8.4 1.94 vs 8.7 1.7, P = 0.574). The
male and female number of YAC clinic visits was comparable. Males missed fewer clinic visits compared with
females (0.9 1.1 vs 1.5 1.4, P = 0.026). Total daily
insulin doses for those using insulin pumps were lower
than those not using pumps (39.9 19.1 vs 59.6 19.9
units, P = 0.002).
Associations of socio-demographic factors,
duration of diabetes and clinic attendance
with glycaemic control
There was a positive association between duration of diabetes and HbA1c (r = 0.247, P = 0.038) (Table 3). There
was an inverse association between the number of clinic
visits and glycaemic control (r = −0.269, P = 0.020).
Age, distance of residence from clinic, IRSD, travel time
and missed appointments were not associated with
HbA1c (Table 3). Distance from clinic was not associated
with either missed or total visits (r = 0.113, P = 0.291
and r = −0.009, P = 0.519 respectively).
Discussion
In this cross-sectional analysis of T1DM patients aged
18–25 years, several factors were associated with better
glycaemic control. These include engagement in study or
work, shorter duration of diabetes and frequency of clinic
attendance. There was no association between glycaemic
control and socioeconomic status, risk-taking behaviours,
distance travelled to the outpatient clinic and use of insulin pumps.
Studies assessing the association between employment
or education and glycaemic control in young adults with
Table 2 Factors associated with glycaemic control: between-group comparisons. Data are mean SD
Outcome variable
HbA1c (%)
HbA1c (%)
HbA1c (%)
HbA1c (%)
Missed visits
Total visits
Group 1
Group 2
P-value
Male 8.24 1.35
Work/study 8.44 1.60
Pump users 8.54 1.31
Risk-takers 8.39 1.92
Male 0.9 1.0
Male 3.1 2.2
Female 9.23 2.01
No work/study 9.66 1.95
Non-pump users 8.76 1.74
Non-risk-takers 8.69 1.70
Female 1.4 1.3
Female 3.0 2.3
0.012
0.022
0.654
0.574
0.026
0.957
HbA1c, glycosylated haemoglobin.
334
© 2016 Royal Australasian College of Physicians
Type 1 diabetes control in young adults
Table 3 Factors associated with glycaemic control (HbA1c, %): analysed
as continuous variables
Variables
Diabetes duration (years)
Visits (n)
Age (years)
Distance (km)
IRSD
Travel time (h)
Missed visits (n)
Correlation coefficient (r)
P-value
0.247
−0.269
−0.155
−0.153
−0.101
−0.169
0.177
0.038
0.020
0.185
0.199
0.399
0.157
0.317
HbA1c, glycosylated haemoglobin; IRSD, index of relative socioeconomic disadvantage.
T1DM are lacking. Interestingly, we identified an association between engagement in study and/or employment
and better glycaemic control. It is plausible that the regular routine or cognitive functioning required for work or
study facilitates executive functioning and motivation,
lending itself to beneficial self-care behaviours. A systematic review of studies assessing the association of executive functioning and adherence and glycaemic control in
youths with T1DM found a positive association between
these variables.15 Furthermore, a study of patients with
type 2 diabetes mellitus aged more than 25 years found
that employment was associated with better health literacy and self-efficacy, both of which are thought to contribute directly and indirectly to self-care behaviours.16
Therefore, participation in work or study may be associated with psychological or other factors that facilitate
improved self-care. Alternatively, better glycaemic control
may allow for engagement in study or work due to
improved confidence or well being, or patients with good
glycaemic control may be inherently highly motivated
and more likely to be working or studying. It has previously been found that young adults with T1DM find diabetes management difficult in the workplace17; however,
our findings reassure that study and work can be encouraged in young adults with T1DM.
We found that longer duration of diabetes was associated with worse glycaemic control. This is consistent with
findings from another observational study of youths with
T1DM.12 Longer duration of T1DM increases the risk of
developing diabetic complications.4 Pre-pubertal duration
of disease contributes to the development of diabetic retinopathy in the adolescent and young adult years.18 The
implications of this finding are that young adults with
longer duration of T1DM should be monitored, supported
and encouraged to maintain attention to diabetes selfcare. Females in our cohort had worse glycaemic control
compared with males. A systematic review of demographic and personal factors associated with glycaemic
control and self-care in youth with T1DM showed a trend
towards higher HbA1c in adolescent girls.19 Further
© 2016 Royal Australasian College of Physicians
studies are therefore required to clarify the role of gender
on glycaemic control in youths with T1DM.
We found a significant relationship between the number of clinic visits and lower HbA1c. A previous study of
patients with T1DM who were diagnosed before age
40 identified access to nurse educators as a determinant
of glycaemic control.18 While we did not categorise clinic
visits into appointments with the medical practitioner,
nurse educator, dietician or psychologist, many visits to
the clinic involved patient interactions with more than
one member of the diabetes team. Our findings support
the importance of providing access to dedicated multidisciplinary care for young adults with T1DM, with encouragement to keep scheduled appointments.
Socioeconomic status and travel distance was not associated with glycaemic control in our cohort of patients.
This is in contrast to findings from a systematic review
reporting an inverse association between socioeconomic
status and diabetes control in youths with T1DM.19 Similar inverse relationships have been demonstrated in older
patients with T1DM.20,21 Socioeconomic status in our
study was determined based on the IRSD developed by
the ABS through the socio-economic indexes for areas
survey.14 Chaturvedi et al. defined socioeconomic status
based on level of education,20 while Secrest et al. utilised
income level.21 The differences in our results may therefore be due to different definitions of socioeconomic status. IRSD is determined by Australian postcode and
therefore does not describe the individual’s relative disadvantage but rather that of the area they lived in at the
time of the study. Of note, travel distance was not associated with glycaemic control in our cohort. Therefore,
within a metropolitan area, distance and classification of
socioeconomic status based on residential location does
not appear to be a barrier to diabetes care, possibly
because our patients had access to appropriate transportation to attend clinic.
In our study, 20 of the participants received insulin
through an insulin pump. The use of an insulin pump
was not associated with improved glycaemic control compared with multiple daily injections. These findings differ
with previous research. A meta-analysis of 23 studies containing 976 patients comparing continuous subcutaneous
insulin infusion (CSII) and three or more insulin injections per day found a statistically significant difference in
terms of glycaemic control between the two groups.22
Those using CSII had, on average, an HbA1c that was
0.3% less than those on insulin injections.22 Given this
small difference, our cohort size may have been too small
to reflect a subtle effect of pump use on glycaemic control
in our cohort. While we did not find a difference in glycaemic control in patients on an insulin pump versus
patients using multiple daily injections, the use of an
335
Osan et al.
insulin pump was associated with a lower total daily dose
of insulin.
Adolescence is associated with a period of substance
experimentation with the attendant risk of developing a
substance use disorder.23 The developing adolescent brain
may be vulnerable to negative impacts of substance abuse,
and the likelihood of developing an addiction is inversely
associated with the age of first substance use.23 Of note,
risk-taking behaviours defined as a composite of substance
use (smoking, excessive alcohol intake or illicit drug use)
did not appear to affect glycaemic control. Other risktaking behaviours, including insulin mismanagement and
disordered eating behaviours, were not examined in this
study. Previous studies assessing the associations between
smoking and glycaemic control have reported contrasting
results. Two large studies in children and adolescents consisting of 27 561 and 629 participants found that smoking
was associated with worse glycaemic control.24,25 A further study consisting of participants with T1DM aged
18–59 years found that smoking was associated with
higher HbA1c.26 Alcohol intake has been associated with
increased glycaemic variability in adolescents and
increased frequency of hypoglycaemia in adults; however,
studies of alcohol intake and glycaemic control in adolescents are lacking.27 The effects of illicit drug use on diabetic control are unclear.27 Therefore, further studies would
be needed to ascertain whether risk-taking behaviours, in
particular alcohol and illicit drug use, affect glycaemic variability and their impact on other self-care behaviours.
Our population was aged 18–25 years; however, studies
suggest a significant increase in substance use with emerging adulthood and ongoing changes in trends associated
with age.28,29 Furthermore, in industrialised societies,
establishment of life-long behaviours may not occur until
25–30 years.7,30 Thus, risk-taking behaviours and their
associations with diabetes self-care and control may continue to evolve above and beyond the age range that we
studied.
Strengths of our study include the focus on young
adults and the examination of current employment, education and risk-taking behaviours in relation to glycaemic
control, which are relevant to this age group. There are
several limitations to this study. As this was a crosssectional analysis, causation cannot be inferred. Furthermore, this study was a convenience sample of clinic attendees. Patients who lack motivation or capacity to attend
hospital-based clinics may be under-represented. The
study focused on a single site, which was the referral
References
1 Australian Institute of Health and Welfare.
Incidence of Type 1 Diabetes in Australia
336
centre for the Southern Metropolitan area of Perth, providing a reasonably representative sample of this group of
patients. Our sample size was limited, which restricts the
analyses we can conduct. As outpatient notes were the
basis of our data collection, we relied on their accuracy
and completeness. The use of novel instruments, such as
‘Google maps’, for distance calculations may lead to bias.
While the distances were calculated using the same thirdparty software for all patients, comparisons were not
made with other tools designed to make the same
calculations.
Conclusion
The findings of this cohort study have several potential
clinical implications. First, these results support encouraging young patients with T1DM to engage in employment
and/or study given its association with better glycaemic
control. Second, within a metropolitan area, socioeconomic status, travel duration and distance may not be a
factor influencing glycaemic control; rather, the frequency
of clinic appointments was more important. The provision
of care in the form of dedicated multidisciplinary clinics
for young adults with T1DM represents an appropriate
means to support diabetes self-care in this age group.
Therefore, young adults with T1DM should be given
appropriate encouragement to maintain regular contact
with these multidisciplinary clinics. Given the limited
sample size in this study, further studies in other populations are required to confirm these findings and explore
in greater depth the associations of substance use and
gender with glycaemic control. Additional research is
warranted to determine whether the influence of study
and/or work on glycaemic control reflects underlying
behavioural traits or represents an avenue for applied
behavioural interventions to improve glycaemic control in
this population.
Implications for practice
• Where appropriate, encourage young adults with type
1 diabetes to study or work; these behaviours are
associated with better glycaemic control.
• Visits to a multidisciplinary clinic for young adults with
type 1 diabetes are associated with better HbA1c values.
• Longer duration of type 1 diabetes in this age group
identifies patients at risk of poorer control.
2000-2013. Diabetes series no. 23. Cat no.
CVD 69. Canberra: AIHW; 2015.
2 Begg SVT, Barker B, Stevenson C,
Stanley L, Lopez AD. The Burden of
Disease and Injury in Australia 2003.
Canberra: AIHW; 2007.
3 Cheung NW CJ, d’Emden MC,
Gunton JE, Jenkins AJ, Ross GP,
© 2016 Royal Australasian College of Physicians
Type 1 diabetes control in young adults
4
5
6
7
8
9
Sinha AK et al. Australian Diabetes
Society Position Statement:
Individualization of HbA1c Targets for
Adults with Diabetes Mellitus. The
Australian Diabetes Society. 2009.
Available from URL: https://
diabetessocietycomau/downloads/
positionstatements/HbA1ctargetspdf
The Diabetes Control and Complications
Trial Research Group. The effect of
intensive treatment of diabetes on the
development and progression of longterm complications in insulindependent diabetes mellitus. N Engl J
Med 1993; 329: 977–86.
Nathan DM, Cleary PA, Backlund JY,
Genuth SM, Lachin JM, Orchard TJ
et al. Intensive diabetes treatment and
cardiovascular disease in patients with
type 1 diabetes. N Engl J Med 2005; 353:
2643–53.
The Diabetes Control and Complications
Trial Research Group. Effect of intensive
therapy on residual beta-cell function in
patients with type 1 diabetes in the
diabetes control and complications trial.
A randomized, controlled trial. Ann
Intern Med 1998; 128: 517–23.
Peters A, Laffel L, American Diabetes
Association Transitions Working Group.
Diabetes care for emerging adults:
recommendations for transition from
pediatric to adult diabetes care systems:
a position statement of the American
Diabetes Association, with
representation by the American College
of Osteopathic Family Physicians, the
American Academy of Pediatrics, the
American Association of Clinical
Endocrinologists, the American
Osteopathic Association, the Centers for
Disease Control and Prevention,
Children with Diabetes, The Endocrine
Society, the International Society for
Pediatric and Adolescent Diabetes,
Juvenile Diabetes Research Foundation
International, the National Diabetes
Education Program, and the Pediatric
Endocrine Society (formerly Lawson
Wilkins Pediatric Endocrine Society).
Diabetes Care 2011; 34: 2477–85.
Sheehan AM, While AE, Coyne I. The
experiences and impact of transition
from child to adult healthcare services
for young people with type 1 diabetes: a
systematic review. Diabet Med 2015; 32:
440–58.
Hislop AL, Fegan PG, Schlaeppi MJ,
Duck M, Yeap BB. Prevalence and
associations of psychological distress in
© 2016 Royal Australasian College of Physicians
10
11
12
13
14
15
16
17
18
19
young adults with type 1 diabetes.
Diabet Med 2008; 25: 91–6.
Jaser SS, Faulkner MS, Whittemore R,
Jeon S, Murphy K, Delamater A et al.
Coping, self-management, and
adaptation in adolescents with type
1 diabetes. Ann Behav Med 2012; 43:
311–19.
Lancaster BM, Pfeffer B, McElligott M,
Ferguson AT, Miller M, Wallace D et al.
Assessing treatment barriers in young
adults with type 1 diabetes. Diabetes Res
Clin Pract 2010; 90: 243–9.
Galler A, Lindau M, Ernert A,
Thalemann R, Raile K. Associations
between media consumption habits,
physical activity, socioeconomic status,
and glycemic control in children,
adolescents, and young adults with type
1 diabetes. Diabetes Care 2011; 34:
2356–9.
Pink B. Socio-economic Indexes for
Areas (SEIFA) [cited 2014 October].
Available from URL: http://www.
ausstats.abs.gov.au
Burkhardt M, Summers M, Lovell J,
Zhu H, Barrie P, Fegan PG et al.
Documenting Insulin Non-adherence in
Young Adults with Type 1 Diabetes.
Proceedings of the Australian Diabetes
Society Annual Scientific Meeting,
Sydney, Australia; 2013 (abstract).
Duke DC, Harris MA. Executive
function, adherence, and glycemic
control in adolescents with type
1 diabetes: a literature review. Curr Diab
Rep 2014; 14: 532.
Bohanny W, Wu SF, Liu CY, Yeh SH,
Tsay SL, Wang TJ. Health literacy, selfefficacy, and self-care behaviors in
patients with type 2 diabetes mellitus. J
Am Assoc Nurse Pract 2013; 25: 495–502.
Balfe M, Brugha R, Smith D, Sreenan S,
Doyle F, Conroy R. Why do young
adults with type 1 diabetes find it
difficult to manage diabetes in the
workplace? Health Place 2014; 26:
180–7.
Couper JJ, Jones TW, Donaghue KC,
Clarke CF, Thomsett MJ, Silink M. The
Diabetes Control and Complications
Trial. Implications for children and
adolescents. Australasian Paediatric
Endocrine Group. Med J Aust 1995; 162:
369–72.
Neylon OM, O’Connell MA,
Skinner TC, Cameron FJ. Demographic
and personal factors associated with
metabolic control and self-care in youth
with type 1 diabetes: a systematic
20
21
22
23
24
25
26
27
28
29
30
review. Diabetes Metab Res Rev 2013; 29:
257–72.
Chaturvedi N, Stephenson JM,
Fuller JH. The relationship between
socioeconomic status and diabetes
control and complications in the
EURODIAB IDDM Complications
Study. Diabetes Care 1996; 19: 423–30.
Secrest AM, Costacou T, Gutelius B,
Miller RG, Songer TJ, Orchard TJ.
Associations between socioeconomic
status and major complications in type
1 diabetes: the Pittsburgh epidemiology
of diabetes complication (EDC) Study.
Ann Epidemiol 2011; 21: 374–81.
Misso ML, Egberts KJ, Page M,
O’Connor D, Shaw J. Continuous
subcutaneous insulin infusion (CSII)
versus multiple insulin injections for
type 1 diabetes mellitus. Cochrane
Database Syst Rev 2010; 1: CD005103.
Monasterio EB. Adolescent substance
involvement use and abuse. Prim Care
2014; 41: 567–85.
Hofer SE, Rosenbauer J, GrulichHenn J, Naeke A, Frohlich-Reiterer E,
Holl RW et al. Smoking and metabolic
control in adolescents with type
1 diabetes. J Pediatr 2009; 154: 20–30.
Kummer S, Stahl-Pehe A, Castillo K,
Bachle C, Graf C, Strassburger K et al.
Health behaviour in children and
adolescents with type 1 diabetes
compared to a representative reference
population. PLoS One 2014; 9: e112083.
Melin EO, Thunander M, Svensson R,
Landin-Olsson M, Thulesius HO.
Depression, obesity, and smoking were
independently associated with
inadequate glycemic control in patients
with type 1 diabetes. Eur J Endocrinol
2013; 168: 861–9.
Jaser SS, Yates H, Dumser S,
Whittemore R. Risky business: risk
behaviors in adolescents with type
1 diabetes. Diabetes Educ 2011; 37:
756–64.
Palladino DK, Helgeson VS,
Reynolds KA, Becker DJ, Siminerio LM,
Escobar O. Emerging adults with type
1 diabetes: a comparison to peers
without diabetes. J Pediatr Psychol 2013;
38: 506–17.
Australian Institute of Health and
Welfare. Young Australians: Their Health
and Wellbeing 2011. Cat. no. PHE 140.
Canberra: AIHW; 2011.
Arnett JJ. Emerging adulthood: a
theory of development from the late
teens through the twenties. Am Psychol
2000; 55: 469–80.
337
Osan et al.
Appendix I
338
© 2016 Royal Australasian College of Physicians
Inaccurate risk perceptions contribute to treatment gaps
in secondary prevention of cardiovascular disease
J. Thakkar,1,2,3 E. L. Heeley,1,2 J. Chalmers1,2 and C. K. Chow1,2,3
1
The George Institute for Global Health, 2The University of Sydney, and 3Department of Cardiology, Westmead Hospital, Sydney, New South Wales,
Australia
Key words
risk perception, cardiovascular diseases,
secondary prevention.
Correspondence
Jay Thakkar, Level 10, King George V Building,
PO Box M 201, Missenden Road, Camperdown,
NSW 2050, Australia.
Email: [email protected]
Received 28 August 2015; accepted 7
December 2015.
doi:10.1111/imj.12982
Abstract
Background: All patients with cardiovascular disease (CVD) are at high risk of recurrent
events. Despite strong evidence, large treatment gaps exist in CVD secondary prevention.
We hypothesise that patients’ self-perception and general practitioner’s (GP) assessment
of future cardiovascular (CV) risk may influence secondary prevention behaviours.
Aim: To examine in patients with known CVD the perceived risk of future CV events
and its relationship with use of secondary prevention medications and risk factor control.
Methods: We examined patient and practitioner’s perceived risk and its relationship
with the uptake of secondary prevention recommendations in adults with CVD participating in the Australian Hypertension and Absolute Risk Study.
Results: Among the 1453 participants, only 11% reported having a high absolute risk
and 29% reported high relative risk of recurrent events. The GP categorised only 30%
as having a 5-year risk ≥15%. After adjusting for covariates, hospitalisation within the
preceding 12 months was the only significant predictor of patients’ accurate risk perception. Conventional CV risk factors were predictive of the GP’s risk estimates. Patients
who accurately understood their risk reported higher smoking cessation rates (7 vs 3%,
P = 0.003) and greater use of antiplatelet, blood pressure lowering therapy and statins
(P ≤ 0.01). However, there was no relationship between GP’s risk perception and secondary prevention treatments.
Conclusion: Both patients and GP have optimistic bias and underestimate the risk of
future CV events. Patients’ accurate self-perception, but not GP risk perception, was
associated with improved secondary preventative behaviours. This suggests that helping
patients to understand their risk may influence their motivation towards secondary prevention. Providing support to GP or programmes to help patients better understand
their risks could have potential benefit on secondary prevention behaviours.
Funding: The current manuscript has no funding or grant. The
original data collection for the AusHEART study was supported
by an unrestricted educational grant from Servier Australia.
Conflict of interest: The AusHEART study was conducted as a
collaborative project between The George Institute for Global
Health and Servier Australia. Jay Thakkar is a PhD student with
The School of Public Health at The University of Sydney and
recipient of Australian Post Graduate Award Scholarship. Emma
Heeley received a travel grant from Servier to present
AusHEART findings at the European Stroke Conference. John
Chalmers has received research grant from Servier for the
Perindopril Protection Against Recurrent Stroke Study, The
Action in Diabetes and Vascular Disease Preterax and Diamicron
MR Controlled Evaluation and the AusHEART study, and has
received lecture fees for speaking at scientific meetings from
Servier. Clara Chow is a primary supervisor of Jay Thakkar’s
PhD and is funded by a Career Development Fellowship cofunded by the National Health and Medical Research Council
and National Heart Foundation of Australia (1033478) and
Sydney Medical Foundation Chapman Fellowship.
© 2016 Royal Australasian College of Physicians
Introduction
Cardiovascular disease (CVD) is the leading cause of
death globally and affects approximately one in six
Australians.1,2 CVD is the most expensive disease group
in Australia in terms of direct healthcare expenditure
and is a major driver of escalating healthcare costs.3
Patients who have had a prior cardiovascular (CV) event
are at a very high risk for repeat events with an estimated risk that exceeds 20% over 5 years.4 Almost half
of the CV events occur in those with prior CVD,5 yet
many of these events may have been prevented with
intensive secondary prevention therapy.6 Preventative
efforts are most efficient when directed at those patients
with highest risk. Despite the strong evidence that treatment with secondary prevention medications, lifestyle
modification and risk factor control decrease the risk of
339
Thakkar et al.
repeat events, numerous studies have shown a gap in
use of secondary prevention treatments.7,8 We have
shown this in The Australian Hypertension and Absolute
risk study (AusHEART).9 Several studies have suggested
that patient and clinician attitudes towards their treatments may be a barrier to adherence,10 however, there
is a little quantitative evidence that supports this.
The general practitioners (GP) assessment of their
patients’ overall CV risk could impact on the prevention
treatments they advise. Patients’ perceptions of their
own CV risk may influence their own secondary prevention behaviours. The aims of this research paper was to
examine, in patients with known CVD recruited in the
AusHEART study through general practice: (i) the perceived future risk of CV events as estimated by patients
and their GP in absolute and relative terms and (ii) the
relationship between these perceptions and the uptake
of secondary preventive measures.
Methods
Study design and population
AusHEART was a nationally representative, clusterstratified, cross-sectional survey of CV risk management
in patients aged 55 years or older presenting to primary
healthcare centres that has been described previously.9,11 The study was approved by The Royal Australasian college of General Practitioners National Research
and Evaluation Ethics Committee. All patients gave
informed written consent for participation. Expression
of interest letters was sent out to 21 074 GP registered
to practise in Australia at the beginning of 2008. Of the
1416 GP who expressed interest in the study, 534 were
randomly selected stratified by geographic location to
ensure that distribution by state and rural–urban
reflected the population distribution according to the
data from 2004 census. GP were asked to recruit 15–20
consecutive adult patients (aged ≥55 years) presenting
to their practices irrespective of the reasons for their
consultation between April and June 2008. In total,
5293 patients were recruited to AusHEART, 1453
patients with history of CVD are the subjects of the current analysis.
Data collection
The GP completed a one-page questionnaire for each
recruited patient on CV risk factors, use of cardiac medications, checked blood pressure with an electronic
monitor (Omron HEM-907 when available). If certain
key CV risk factors, including fasting blood lipids, blood
glucose, estimated glomerular filtration rate and urine
340
albumin–creatinine ratio, had not been measured
within the National Heart Foundation of Australia
guideline-recommended time frame, the GP were
requested to repeat these tests and record the results, to
allow a ‘forced capture’ of CV risk factor levels in all
registered patients. GP were asked to estimate and record the patient’s absolute risk of a future CV event over
the next 5 years.
Each patient completed a questionnaire that included
questions on socio-economic status and physical activity.
Each patient was asked to estimate their future risk of a
CV event in the next 5 years and report this in absolute
and relative terms. Specifically each patient was asked
firstly ‘What do you think your chances of having a
heart attack or stroke in the next 5 years are?’ and given
six response options to select from – no chance, low,
mild, moderate, high and very high. Secondly, they were
asked – ‘How do you think your chances of having a
heart attack or stroke in the next 5 years compare to
chances of an average person the same age and sex as
you?’ and given a scale from 1 to 5 (where 1–2, lower
chance; 3, same chance; 4–5, higher chance). Finally,
they were requested to provide a numeric risk estimate
(range 0–100%) of the risk of an average person the
same age and sex, and their risk of having a heart attack
or stroke in the next 5 years.
Definitions
CVD was defined as patients with prior stroke, transient
ischaemic attack (TIA), myocardial infarction (MI),
angina or coronary revascularisation. Patients with
established CVD are considered to be at high risk of a
future CV event, that is, they are generally regarded to
have a 5-year risk greater than 15%.12 We categorised
GP’s risk estimates as an ‘accurate risk estimate’ if the GP
reported the absolute CV risk of patients with CVD to be
>15% over 5 years and otherwise an ‘underestimate of
risk’.13 We categorised patients who indicated their
future chances of heart attack or stroke was ‘high’ or
‘very high’ as ‘accurate high self-perception of risk’,
whereas patients who indicated their risk as ‘no chance,
low, mild or moderate’ were categorised into the ‘inaccurate low-moderate self-perception of risk’ group.
We defined good risk factor control as achieving three
or more of the following five parameters: (i) physical
activity of >30 min for 5 or more days per week, (ii) non-smoking status, (iii) LDL <2 mmol/L, (iv) BP <140/
90 mm Hg and (v) body mass index (BMI) <25 kg/m2.
We defined good secondary prevention medication use
as the use of statin, antiplatelet and BP lowering
therapy.
© 2016 Royal Australasian College of Physicians
Risk perceptions in CV disease
Statistical analysis
We compared risk factor levels and use of secondary prevention treatments between groups with ‘accurate’ and
‘underestimated’ GP assessed risk and patients’ selfperception as ‘accurate high risk’ and ‘inaccurate low to
moderate risk’. Descriptive statistics are presented as percentages or mean with standard deviation. Comparisons
between groups were tested using t-tests for continuous
variables and chi-squared tests for categorical variables.
In all cases, a P value of <0.05 was used to reject the null
hypothesis. To examine the relationship of perception of
risk with risk factor control and the use of secondary
prevention treatments, we used a multivariable logistic
regression model. Selection of parameters for inclusion
into the multivariable model was based on clinical judgement and univariate statistical significance (P < 0.05).
The covariates considered were: age, gender, smoking
status, diabetes, obesity, systolic blood pressure control,
diastolic blood pressure control, elevated total cholesterol
(>4 mmol/L) and type of CVD (cardiac or cerebrovascular). All statistical analysis was conducted using STATA
version 13.1 (Stata Corporation, College station,
TX, USA).
Results
Of the 1453 AusHEART patients with a history of CVD,
743 (51%) had a history of coronary heart disease
(CHD), 533 (37%) stroke or TIA and 177 (12%) had
both. The mean age of participants with CVD was
70 years (SD: 9.1, range: 55–98 years) and 43% were
females. Data on GP’s estimate of future risk and
patient’s self-estimate of risk were available for 1260
(90%) and 1399 (96%) respectively. Larger proportion
of patients with the history of CHD when compared with
stroke was perceived to be at higher risk by the GP as
well as patients.
Patients’ perceptions of their health
and future CV risk
Patients with CVD reported their health as excellent
(2%), very good (15%), good (45%), fair (32%) and
poor (6%). With respect to their future chances of a CV
event, 2% reported their risk to be very high, 9% high,
34% moderate, 23% mild, 28% low and 4% no chance.
In relative terms to a person with the same age and sex,
29% reported themselves to be at much higher risk,
43% at the same risk and 28% at lower risk for a future
event.
The risk perception for future events was similar by
gender or age group (defined as age >65 years). Patients
© 2016 Royal Australasian College of Physicians
that accurately perceived their own risk to be high were
more likely to have prior history of diabetes, hypertension, higher BMI, less physically active or had at least
one hospital admission in preceding 12 months
(Table 1). However, after adjusting for covariates recent
hospitalisation within preceding 12 months (Table 2)
remained a significant predictor of patients perceiving
their risk as high.
Perception of risk by GP
GP categorised 30% of the patients with established CVD
as having a 5-year risk ≥15%, and thus 70% with an
estimated absolute risk <15%. GP reported higher mean
risk estimates in patients with CHD compared with
stroke/TIA (17.0 vs 14.6%, P = 0.025). Patients that GP
categorised accurately as high risk (>15%) were older,
male, diabetic, hypertensive, had higher blood pressure,
higher BMI, elevated total cholesterol and smokers
(Table 1). After adjusting for covariates, age, male sex,
diabetes, hypertension, obesity, elevated total cholesterol
and smoking were significant predictors of GP accurately
assessing risk as high (Table 2). There was poor correlation between the GP and patient risk estimates (coefficient 0.175, Spearman’s rho 0.187).
Risk perception and relation to risk factor
control and management
Patients that perceived they were at increased risk for
recurrent events reported a higher smoking cessation
(quit) rate over the preceding 12 months (7 vs 3%,
P = 0.03) (Table 1). There was no significant difference
in the quit rates between patient groups based on GP
perception (5 vs 4%, P = 0.53). Control of other risk factors was similar in patients that perceived their risk as
high versus low/moderate (Figs 1,2).
After adjusting for covariates, patients’ perception of
risk (adjusted Odds Ratio (aOR) 1.00, 95% CI 0.61–1.65,
P = 0.985) and the GP’s estimate of risk (aOR 0.98, 95%
CI 0.68–1.42, P = 0.93) were both not significantly associated with achieving risk factor control (≥3 of 5 risk factors controlled). With respect to the use of all secondary
prevention medications, patients’ perception was significantly associated (aOR 1.64, 95% CI 1.12–2.43,
P = 0.012) with receiving all therapy (aspirin plus BP
lowering plus statin). The GP perception of the patients
risk was not associated with receiving all CV preventative therapy (aOR 0.83, 95% CI 0.63–1.08, P = 0.174)
(Figs 1,2).
341
Thakkar et al.
Table 1 Participant characteristics, risk factor control and treatment in patients with CVD as stratified by patient and GP perception of risk of future
cardiovascular event
Patients
Accurate (high) selfperceived risk (n = 154)
Age ≥ 65 years
101 (66%)
Male
88 (58%)
Current smoker
12 (8%)
Quit smoking in
11 (7%)
past year
Medical history
Disease type
CAD/angina
84 (55%)
Stroke/TIA
44 (29%)
Both
26 (16%)
Diabetes
55 (36%)
Hypertension
131 (86%)
Recent admission
83 (54%)
<12 months
SBP (mm Hg)
136 17
DBP (mm Hg)
74 11
BMI (kg/m2)
29.17 5.7
TC (mmol/L)
4.49 1.26
LDL (mmol/L)
2.37 0.85
HDL (mmol/L)
1.30 0.45
Risk factor control
Not smoking
142 (92%)
Regular exercise‡
26 (17%)
BMI < 25
37 (24%)
LDL < 2 mmol/L
57 (40%)
BP < 140/90
95 (62%)
≥3 risk factors
23 (20%)
controlled
Secondary prevention therapy
Antiplatelet
128 (85%)
BP lowering
133 (90%)
Statin
121 (82%)
GP
Underestimate (low-mod)
perceived risk (n = 1245)†
Pvalue
Accurate
estimate
(n = 375)
Underestimate
(n = 885)
Pvalue
0.667
0.724
0.827
0.003
276 (73%)
231 (62%)
51 (14%)
13 (5%)
573 (64%)
484 (55%)
58 (7%)
29 (4%)
0.001
0.024
<0.001
0.513
206 (55%)
115 (31%)
54 (14%)
142 (38%)
301 (81%)
118 (32%)
441 (50%)
342 (39%)
102 (12%)
207 (23%)
601 (69%)
248 (29%)
<0.001
<0.001
0.266
0.201
0.605
0.033
0.448
0.083
0.177
139 19
76 12
29 5.5
4.62 1.16
2.53 0.05
1.27 0.02
132 + 16
73 10
28 4.9
4.50 1
2.47 0.03
1.36 .01
<0.001
<0.001
0.002
0.07
0.317
<0.001
1137 (92%)
301 (24%)
320 (26%)
387 (33%)
851 (68%)
185 (19%)
0.827
0.041
0.571
0.079
0.118
0.785
323 (86 %)
87 (24%)
89 (24%)
110 (32%)
211 (56 %
54 (20%)
824 (93 %)
211 (25%)
235 (27%)
301 (35%)
653 (74 %)
136 (19%)
<0.001
0.657
0.284
0.273
<0.001
0.767
850 (71%)
934 (77%)
872 (72%)
<0.001
<0.001
0.010
280 (76%)
301 (83%)
269 (74%)
616 (72%)
661 (77%)
631 (73%)
0.110
0.012
0.695
838 (67%)
692 (57%)
103 (8%)
33 (3%)
0.021
0.035
627 (50%)
472 (38%)
146 (12%)
332 (27%)
862 (71%)
336 (27%)
134
74
28.22
4.55
2.51
1.35
17
11
5.6
1.01
0.88
0.01
0.027
<0.001
<0.001
†34% moderate, 23% mild, 28% low, 4% no chance. Data are mean SD or number (%) and based on non-missing value. ‡>30 min per day for >5 days
per week. BMI, body mass index; BP, blood pressure; CHD, coronary heart disease; CVD, cardiovascular disease; DBP, diastolic blood pressure; GP,
general practitioner; HDL, high density lipoprotein cholesterol; HT, hypertension; LDL, low density lipoprotein cholesterol; SBP, systolic blood pressure;
TC, total cholesterol; TIA, transient ischaemic attack.
Discussion
Our study found that the large majority of patients with
established CVD recruited through primary care in
Australia had inaccurate perception of their risk of future
CV events. The GP of these patients tended to underestimate absolute numeric risk. Only 11% of patients with
established CVD reported that they remained at high risk
of future CV events. Nearly 70% of patients perceived
their relative risk to be lower or similar to age and sex
matched individuals. These findings support previous
observations from the UK and USA, where the majority
of participants have optimistic bias and underestimate
342
their risk of having heart attack or stroke in future.14,15
In general, subjects are optimistic and underestimate
their risk, but an overestimate of the risk is also possible
and was reported16 in a postal survey of over 2000
patients with hypertension and diabetes that reported
more than 50% overestimated absolute risk of heart
attack or stroke by more than 20%. The misperception
of risk is evident in hospitalised patients, a study of
79 patients with acute MI admitted to coronary care unit
in Auckland found patients’ risk perception were not
congruent with estimates from established clinical
indicators.17
© 2016 Royal Australasian College of Physicians
Risk perceptions in CV disease
Table 2 Predictors of accurate perception of absolute CVD risk by patients with CVD and their GP as assessed by multivariable logistic regression
model
Variables
Patients perception
Age
Male
Obesity (BMI > 30 kg/m2)
CHD
Diabetes
Hypertension
Elevated total cholesterol†
Regular physical activity
Smoking
Recent hospitalisation‡
GP perception
OR (95% CI)
P-value
OR (95% CI)
P-value
1.02 (0.99–1.03)
1.02 (0.69–1.51)
1.22 (0.82–1.84)
1.13 (0.74–1.72)
1.43 (0.96–2.14)
1.84 (1.11–3.05)
0.85 (0.57–1.23)
0.69 (0.43–1.11)
1.14 (0.56–2.32)
2.81 (1.94–4.06)
0.16
0.91
0.33
0.58
0.08
0.02
0.42
0.13
0.72
<0.01
1.03 (1.02–1.05)
1.37 (1.03–1.84)
1.56 (1.15–2.11)
1.25 (0.92–1.67)
2.16 (1.62–2.92)
1.56 (1.12–2.18)
1.52 (1.12–2.06)
1.02 (0.74–1.41)
2.86 (1.81–4.52)
1.01 (0.75–1.36)
<0.01
0.03
0.01
0.15
<0.01
0.01
0.01
0.86
<0.01
0.95
†>4 mmol/L ‡Hospitalisation in previous 12 months. BMI, body mass index; CHD, coronary heart disease; CVD, cardiovascular disease; GP, general
practitioner; OR, odds ratio.
Non-smoking
p=0.04
Regular exercise
p=0.57
BMI<25 kg/m2
p=0.08
p=0.12
LDL<2 mmol/L
BP<140/90
p=0.79
≥3 Risk factor controlled
p=0.27
All therapy
p<0.001
Antiplatelet
p< 0.001
BP Lowering
p<0.001
Statin
p=0.010
0
10
20
30
Risk under-estimated
40
50
60
70
80
90
100
Accurate risk estimate
Figure 1 Risk factor control and secondary prevention management in patients with cardiovascular disease based on self-perception of absolute risk.
The GP underestimated the numeric risk of future CV
events in patients with established CVD in 70% of cases.
In addition, we noted that GP categorised more patients
with CHD in high-risk subgroup compared with stroke.
This disparity in differential risk perception towards cerebral compared with cardiac CVD may contribute towards
gaps in secondary prevention therapies. Several studies
have shown that doctors do not accurately estimate
risk,18,19 nor do they discuss the CV risk with their
patients.20 Higher risk estimates by the GP’s were associated with the presence of conventional risk factors. However, after adjusting for covariates, our study found that
GP perceptions did not translate into improved control
over their risk factors.
© 2016 Royal Australasian College of Physicians
Our data have highlighted substantial lack of patients’
insight into their future CV risk. Several potential explanations exist. For example, patients’ under recognition
of increased risk may be a symptom of poor health literacy in general and poor awareness of the risk factors.21
Inaccurate risk perception can result from lack of risk
assessment by the physician and/or lack of communication. This can be due to inadequate consultation time,
lack of financial incentive and low physician communication skills.22 A survey of 380 GP in Swiss cantons
(Zurich and Aargau) found that 57% reported that
numerical information resulting from risk prediction
algorithms is not helpful.23 Misrepresentations in the
media and by influential people may confuse the
343
Thakkar et al.
Non-smoking
p=0.66
p=0.28
Regular exercise
p=0.27
BMI<25 kg/m2
LDL<2 mmol/L
p<0.01
p=0.77
BP<140/90
≥3 Risk factor controlled
p = 0.88
Allthearpy
p=0.01
Statin
p = 0.695
BP Lowering
p = 0.01
Antiplatelet
0
10
20
30
Risk under-estimated
40
50
60
70
p = 0.11
80
90
100
Accurate risk estimate
Figure 2 Risk factor control and secondary prevention management in patients with cardiovascular disease stratified by risk estimates by the general
practitioners.
information on CV risk.24 Erroneous perception can
result from denial and tendency to compare themselves
with persons who are worse off rather than an average
matched person.20 Some studies have found that physicians overestimate patients’ knowledge with respect to
their health. For example, in one large survey (REACT
study) involving 754 primary care physicians and 5104
general public participants from five European countries,
92% physicians reported they believed that their patients
associate high cholesterol with CHD,25 yet only 51% of
public participants were aware that high cholesterol
increases CHD risk.26
The health belief model is a psychological health behaviour change theory that suggests that people’s beliefs
about their health problem, for example, with respect to
its severity or perceived benefits, are related to their
engagement or lack of engagement with healthpromoting behaviours.27 These theories underpin the
potential importance of patients needing to have an
accurate perception of their risk in order to engage effectively in behaviours, such as lifestyle change and medical
adherence necessary for improving their condition. Intuitive recognition that a specific action will offer protection is central to theories surrounding health-related
behaviour.28 Accurate perception of illness is probably a
good motivation for self-care and influences the behaviour and risk factor control.29–31 Our study findings are
consistent with this. We found that patients who accurately perceived their future risk as high, reported
greater smoking cessation over the previous 12 months.
344
Medication adherence is an important patient behaviour that is influenced by beliefs. More benign perceptions of illness translate to lower medication
adherence.32 This is reflected in our study. Patients who
accurately perceived future risk as high were significantly more likely to be on secondary prevention medication. Information on health beliefs may be important
at achieving concordance between patient and practitioner and can be a target for intervention to enhance
adherence.33
General practice is an ideal setting for providing preventative care. CV risk assessment in general practice
can be helpful in motivating patients and reinforcing
self-care.34 Even brief advice from the GP can act as a
catalyst for change in an integrated model of disease prevention. One meta-analysis described significant increase
in rates of smoking cessation with only brief advice.35
These highlight the importance of the physician’s role in
risk quantification and knowledge translation.9
Our study has some limitations. The assessment of risk
perception for patients was drawn from self-reported
responses to a Likert scale and quantitative questions
about their risk relative to others in the community. The
associations are estimates from cross-sectional survey
data and as such we are unable to establish the direction
of the associations found. Participants recruited were
patients presenting to primary care for a variety of reasons, and GP were only a small proportion of all GP and
while our study sample is likely to be broadly representative, it may represent a sample of patients that are
© 2016 Royal Australasian College of Physicians
Risk perceptions in CV disease
more motivated about their health as they attend the
doctor and their doctor is interested in patient audits.
Conclusion
This study has identified that both patients and GP
underestimate risk of future CV events. Patients with an
accurate understanding of their CV risk were more likely
to observe treatment recommendations. The GP perceptions were not correlated with patients’ perceptions of
their own risk and could reflect the challenges doctors
face in communicating risk concepts to patients. As the
findings in this study support, patients’ misperceptions of
their risk may present a barrier to the implementation
of secondary prevention. The potential implications of
these findings are that increasing patients’ knowledge of
their risk, for example, through supporting GP with tools
References
1 National Heart Foundation of Australia.
The Shifting Burden of Cardiovascular
Disease in Australia. Access Economics
Pty Ltd. 2005. [cited 2015 Apr 22].
Available from URL: http://www.
heartfoundation.org.au/
sitecollectiondocuments/hfshifting_burden-cvd-accecons-2005may.Pdf
2 Australia’s Health 2012. The Thirteenth
Biennial Health Report of Australian
Institute of Health and Welfare. [cited
2015 Jan 10]. Available from URL:
http://www.aihw.gov.au/publicationdetail/?id=10737422172
3 Goss J. Projection of Australian Health Care
Expenditure by Disease, 2003 to 2033.
Canberra: Australian Institute of Health
and Welfare; 2008.
4 Randomised trial of cholesterol
lowering in 4444 patients with coronary
heart disease: the Scandinavian
Simvastatin Survival Study Group.
Lancet 1994; 344:1383–9.
5 Kerr AJ, Broad J, Wells S, Riddell T,
Jackson R. Should the first priority in
cardiovascular risk management be
those with prior cardiovascular disease?
Heart 2009; 95: 125–9.
6 Fletcher GF, Bufalino V, Costa F,
Goldstein LB, Jones D, Smaha L,
et al. Efficacy of drug therapy in the
secondary prevention of cardiovascular
disease and stroke. Am J Cardiol 2007;
99: S1–35.
7 Chow CK, Jolly S, Rao-Melacini P,
Fox KA, Anand SS, Yusuf S. Association
of diet, exercise, and smoking
© 2016 Royal Australasian College of Physicians
8
9
10
11
12
13
or programmes to help educate patients about their risk,
could translate into improved secondary prevention.
Electronic decision support tools embedded in electronic
health records, M-health tools or counselling programs
may have a role.36–38 Further research, however, is
required to identify effective strategies as well as to evaluate whether these can translate into improved clinical
outcomes for individual patients.
Acknowledgments
We thank all the GP investigators who participated in
the study. Data were collected by Statistical Revelations,
Melbourne. We thank Claire Morgan for the help in setting up the project and Catherine Devlin for the help
with collating data.
modification with risk of early
cardiovascular events after acute
coronary syndromes. Circulation 2010;
121: 750–8.
Yusuf S, Islam S, Chow CK,
Rangarajan S, Dagenais G, Diaz R et al.
Use of secondary prevention drugs for
cardiovascular disease in the
community in high-income, middleincome, and low-income countries (the
PURE Study): a prospective
epidemiological survey. Lancet 2011;
378: 1231–43.
Heeley E, Anderson C, Patel A, Cass A,
Peiris D, Weekes A et al. Disparities
between prescribing of secondary
prevention therapies for stroke and
coronary artery disease in general
practice. Int J Stroke 2012; 7: 649–54.
Broadbent E, Donkin L, Stroh JC.
Illness and treatment perceptions are
associated with adherence to
medications, diet, and exercise in
diabetic patients. Diabetes Care 2011; 34:
338–40.
Heeley EL, Peiris DP, Patel AA, Cass A,
Weekes A, Morgan C et al.
Cardiovascular risk perception and
evidence – practice gaps in Australian
general practice (the AusHEART study).
Med J Aust 2010; 192: 254–9.
Smolina K, Wright FL, Rayner M,
Goldacre MJ. Long-term survival and
recurrence after acute myocardial
infarction in England, 2004 to 2010.
Circ Cardiovasc Qual Outcomes 2012; 5:
532–40.
National Vascular Disease Prevention
Alliance. Guidelines for the assessment of
absolute cardiovascular disease risk.
14
15
16
17
18
19
20
National Health and Medical Research
Council; 2009.
Avis NE, Smith KW, McKinlay JB.
Accuracy of perceptions of heart attack
risk: what influences perceptions and
can they be changed? Am J Public Health
1989; 79: 1608–12.
Marteau TM, Kinmonth AL, Pyke S,
Thompson SG. Readiness for lifestyle
advice: self-assessments of coronary risk
prior to screening in the British family
heart study. Family Heart Study Group.
Br J Gen Pract 1995; 45: 5–8.
Frijling BD, Lobo CM, Keus IM,
Jenks KM, Akkermans RP,
Hulscher ME et al. Perceptions of
cardiovascular risk among patients with
hypertension or diabetes. Patient Educ
Couns 2004; 52: 47–53.
Broadbent E, Petrie KJ, Ellis CJ,
Anderson J, Gamble G, Anderson D
et al. Patients with acute myocardial
infarction have an inaccurate
understanding of their risk of a future
cardiac event. Intern Med J 2006; 36:
643–7.
Grover SA, Lowensteyn I, Esrey KL,
Steinert Y, Joseph L, Abrahamowicz M.
Do doctors accurately assess coronary
risk in their patients? Preliminary
results of the coronary health
assessment study. BMJ 1995; 310:
975–8.
Montgomery AA, Fahey T,
MacKintosh C, Sharp DJ, Peters TJ.
Estimation of cardiovascular risk in
hypertensive patients in primary care.
Br J Gen Pract 2000; 50: 127–8.
Webster R, Heeley E. Perceptions of
risk: understanding cardiovascular
345
Thakkar et al.
21
22
23
24
25
26
disease. Risk Manag Healthc Policy 2010;
3: 49–60.
Karthik S, Tahir N, Thakur B,
Nair U. Risk factor awareness and
secondary prevention of coronary
artery disease: are we doing enough?
Interact Cardiovasc Thorac Surg 2006; 5:
268–71.
Hobbs FD, Jukema JW, Da Silva PM,
McCormack T, Catapano AL. Barriers to
cardiovascular disease risk scoring and
primary prevention in Europe. QJM
2010; 103: 727–39.
Eichler K, Zoller M, Tschudi P,
Steurer J. Barriers to apply
cardiovascular prediction rules in
primary care: a postal survey. BMC Fam
Pract 2007; 8: 1.
Mosca L, Mochari-Greenberger H,
Dolor RJ, Newby LK, Robb KJ.
Twelve-year follow-up of American
women’s awareness of cardiovascular
disease risk and barriers to heart health.
Circ Cardiovasc Qual Outcomes 2010; 3:
120–7.
Hobbs FD, Erhardt L. Acceptance of
guideline recommendations and
perceived implementation of coronary
heart disease prevention among
primary care physicians in five
European countries: the Reassessing
European Attitudes about
Cardiovascular Treatment (REACT)
survey. Fam Pract 2002; 19: 596–604.
Erhardt L, Hobbs FD. Public perceptions
of cardiovascular risk in five European
countries: the react survey. Int J Clin
Pract 2002; 56: 638–44.
346
27 Janz NK, Becker MH. The Health Belief
Model: a decade later. Health Educ Q
1984; 11: 1–47.
28 Prentice-Dunn S, Rogers RW.
Protection motivation theory and
preventive health: beyond the health
belief model. Health Educ Res 1986; 1:
153–61.
29 Aljasem LI, Peyrot M, Wissow L, Rubin
RR. The impact of barriers and selfefficacy on self-care behaviors in type
2 diabetes. Diabetes Educ 2001; 27:
393–404.
30 Mosleh SM, Almalik MM. Illness
perception and adherence to healthy
behaviour in Jordanian coronary heart
disease patients. Eur J Cardiovasc Nurs
2014.
31 Mosnier-Pudar H, Hochberg G,
Eschwege E, Halimi S, Virally ML,
Guillausseau PJ et al. How patients’
attitudes and opinions influence selfcare behaviours in type 2 diabetes.
Insights from the French DIABASIS
survey. Diabetes Metab 2010; 36: 476–83.
32 Rajpura J, Nayak R. Medication
adherence in a sample of elderly
suffering from hypertension: evaluating
the influence of illness perceptions,
treatment beliefs, and illness burden.
J Manag Care Pharm 2014; 20: 58–65.
33 Ross S, Walker A, MacLeod MJ. Patient
compliance in hypertension: role of
illness perceptions and treatment
beliefs. J Hum Hypertens 2004; 18:
607–13.
34 Wan Q, Zwar N, Vagholkar S,
Campbell T, Harris M. Cardiovascular
35
36
37
38
absolute risk assessment and
management – engagement and
outcomes in general practice patients.
Aust Fam Physician 2010; 39: 954–8.
Stead LF, Buitrago D, Preciado N,
Sanchez G, Hartmann-Boyce J,
Lancaster T. Physician advice for
smoking cessation. Cochrane Database
Syst Rev 2013; 5: CD000165.
Peiris D, Usherwood T, Panaretto K,
Harris M, Hunt J, Patel B et al. The
treatment of cardiovascular risk in
primary care using electronic decision
suport (TORPEDO) study-intervention
development and protocol for a cluster
randomised, controlled trial of an
electronic decision support and quality
improvement intervention in Australian
primary healthcare. BMJ Open 2012; 2:
e002359.
Chow CK, Redfern J, Hillis GS,
Thakkar J, Santo K, Hackett ML et al.
Effect of lifestyle-focused text
messaging on risk factor modification in
patients with coronary heart disease: a
randomized clinical trial. JAMA 2015;
314: 1255–63.
Redfern J, Usherwood T, Harris MF,
Rodgers A, Hayman N, Panaretto K
et al. A randomised controlled trial of a
consumer-focused e-health strategy for
cardiovascular risk management in
primary care: the Consumer Navigation
of Electronic Cardiovascular Tools
(CONNECT) study protocol. BMJ Open
2014; 4: e004523.
© 2016 Royal Australasian College of Physicians
B R I E F C O M M U N I C AT I O N S
Substantial variation in post-engraftment infection prophylaxis
and revaccination practice in autologous stem cell transplant
patients
H. Y. Lim and A. Grigg
1
Department of Clinical Haematology, Austin Health, Melbourne, Victoria, Australia
Key words
prophylaxis, autologous stem cell transplant,
Pneumocystis jirovecii, antiviral, revaccination.
Correspondence
Hui Yin Lim, Department of Haematology, Austin
Health, PO Box 5555, Heidelberg, Vic. 3084,
Australia.
Email: [email protected]
Abstract
There is a paucity of evidence supporting the necessity or duration of Pneumocystis jirovecii and antiviral prophylaxis as well as revaccination following autologous stem cell
transplant (ASCT). A survey aimed at evaluating these policies was distributed to
34 ASCT centres across Australasia. The 26 survey respondents demonstrated significant heterogeneity in their infection prophylaxis and revaccination strategy posttransplant despite the availability of consensual guidelines.
Received 19 July 2015; accepted 3
December 2015.
doi:10.1111/imj.12992
Haemopoietic stem cell transplant patients are at risk of
a range of opportunistic infections depending on their
degree of immunosuppression and time since transplant.
The majority of the current literature in this context
focuses on allogeneic patients with a paucity of data on
the risks of post-autologous transplant (ASCT).1–3
Despite this, guidelines for infection prophylaxis generally do not distinguish between these types of transplant.
To obtain an overview of local practice, we conducted
a survey of prophylaxis policy in ASCT patients in Australasian autograft centres. We focussed on three specific
areas: prophylaxis against Pneumocystis jirovecii (JPJ), previously known as Pneumocystis carinii, Herpes simplex
and varicella zoster infections and revaccination policies.
A 15-item survey document (Appendix I) addressing
these policies was emailed to 34 ASCT centres in
Australia and New Zealand between May 2014 and June
2014. The survey allowed for descriptive comments. The
responses were compared with the eviQ guidelines,4 a
New South Wales online cancer service that serves as a
point-of-care evidence-based clinical information
resource relevant to local Australian practice; and international guidelines.5
Approval was obtained from the Austin Human
Research Ethics Committee (LNR/14/Austin/659).
Twenty-six centres (76%) from across Australasia
responded. A median of 30 ASCT (range 15–90) is performed in each centre annually. Table 1 describes the
prophylaxis and revaccination policies in the various
ASCT centres.
PJP prophylaxis was reportedly used routinely in the
majority of centres (77%), most commonly commenced
from the time of engraftment. Sulfamethoxazole/trimethoprim was the drug of choice, although the
strength and dosing regimen varied, as did the duration
of prophylaxis. Peripheral blood CD4+ count influenced
the duration in only three centres. Only nine centres
(35%)
used
prophylaxis
during
maintenance
thalidomide- or prednisolone-based protocols.
Most centres (85%) reported the use of antiviral prophylaxis, with once daily valaciclovir or twice daily aciclovir being the commonest regimens. Most commenced
antiviral prophylaxis around the time of admission and
continued beyond engraftment, but the duration was
heterogeneous, ranging from 1 to 12 months.
Sixteen centres (62%) reported routine implementation of a revaccination policy post-ASCT, generally as
per published European and local guidelines.5–7
Discussion
Funding: None.
Conflict of interest: None.
© 2016 Royal Australasian College of Physicians
The results of this survey demonstrate significant heterogeneity among Australasian autograft centres in
347
Brief Communications
Table 1 Infection prophylaxis and vaccination policy in the 26 responding centres
No. centres utilising
Drug options and dosing regimen
Time of commencement
From admission
From engraftment
Other
Duration of prophylaxis
≤3 months
>3 to ≤6 months
12 months
Recommendations according to
international guidelines5
% centres compliant
eviQ guidelines4
% centres compliant
PJP prophylaxis n (%)
Antiviral prophylaxis n (%)
20 (77)
S/T DS† (18)
BD‡ twice weekly: 10 (50)
OD§ thrice weekly: 7 (32)
OD: 1 (5)
S/T SS¶ (2)
OD: 2
22 (85)
Valaciclovir 500 mg (16)
OD: 14 (64)
BD: 2 (9)
Aciclovir 400 mg (6)
OD: 1 (5)
BD: 5 (23)
1 (5)
18 (90)
1 (5)
17 (77)
3 (14)
2 (9)
11 (55)
9 (45)
0
16 (73)
2 (9)
4 (18)
3–6 months
17 (65)
3–6 months
17 (65)
12 months
4 (18)
6 months
2 (9)
†S/T DS sulfamethoxazole/trimethoprim double strength (800/160 mg). ‡BD twice daily. §OD once daily. ¶S/T SS sulfamethoxazole/trimethoprim single
strength (400/80 mg).
infection prophylaxis and revaccination strategy posttransplant despite the availability of published
guidelines.1–3
The recommended drug option for PJP prophylaxis
post-ASCT is sulfamethoxazole/ trimethoprim, both due
to its efficacy against PJP and its broad spectrum of activity providing protection against other pathogens, including toxoplasma, nocardia and other respiratory
pathogens.4,5,8 While relatively safe, it can be associated
with significant adverse drug effects, including significant
myelosuppression and potentially fatal hypersensitivity
syndromes; hence, its routine use post-ASCT should be
based on evidence demonstrating its necessity. Our
review of the literature, however, could not find convincing evidence to justify its routine use or the optimal
duration of PJP prophylaxis in ASCT patients in the
absence of risk factors, such as concomitant glucocorticoids or other immunosuppressants or previous PJP
infection.
A recent retrospective study reported five cases of PJP
pneumonia in 1191 ASCT patients (0.42%) over a 10year period in the absence of prophylaxis (all cases
occurred in patients receiving concomitant steroids), suggesting that routine prophylaxis may not be warranted.9
Moreover, in patients receiving immunomodulatory
therapy post-ASCT for myeloma, commonly with thalidomide or lenalidomide with or without steroids, it
remains unclear if this represents a higher risk group in
348
whom PJP prophylaxis is clearly justified, although the
reactivation of hepatitis B in patients receiving thalidomide suggest that these patients may be at risk of opportunistic infection.10 In human immunodeficiency virus
(HIV) patients, a CD4+ count of less than 200/mm3 is
associated with increased risk of PJP, and hence, this
threshold determines prophylaxis use and its duration,11
but it remains unknown whether CD4+ can similarly be
used as one of the surrogate laboratory parameters of
immune recovery in a non-HIV setting and a possible
guideline to optimal duration of prophylaxis in high-risk
patients.2,12–15
Similar to antiviral prophylaxis, there was substantial
variability in practice, particularly with respect to the
duration. Previous studies suggest that in the absence of
post-engraftment prophylaxis, the frequency of varicella
zoster infection in ASCT patients is not insubstantial
(16–28%), with the majority occurring within the first
12 months.16 Non-randomised studies suggest a potential benefit of an extended period of prophylaxis against
varicella zoster viral reactivation for up to 12 months,
although to our knowledge, there is no reported randomised trials addressing the issue of optimal
duration.7,15,17,18
Immune reconstitution gradually occurs post-ASCT,
although the CD4+ count remains low for at least the
first year post-transplant despite reconstitution of total B
and T cells to normal levels at 2–4 months.5,13 The
© 2016 Royal Australasian College of Physicians
Brief Communications
impact of this immunodeficiency on susceptibility to
infectious complications, however, is not well documented, although previous studies have shown that antibody
titres to vaccine-preventable diseases decline during the
first decade after transplant.5,6 Moreover, to our knowledge, there are no published randomised trial data to
support the efficacy of revaccination against any viruses,
herpetic or otherwise. Despite this, largely based on
extrapolation from the allogeneic setting, current guidelines advocate for revaccination against pneumococcus,
Haemophilus influenzae type b, meningococcus, measles,
mumps, rubella, hepatitis B, polio, tetanus, diphtheria
and pertussis among others in ASCT patients,5–7 ostensibly with the intention of reducing the risk of disease and
to promote herd immunity where a vulnerable transplant patient could be a potential sentinel case.19,20 In
our survey, not all the centres implemented this policy,
which may reflect both the lack of evidence and the cost,
noting that the vaccinations are not funded under the
Australian Pharmaceutical Benefits Scheme and are estimated to cost more than AU$650 per patient.
Interpreting the results of our survey should be done
with acknowledgement of its limitations. A quarter of
invited centres did not respond, which may have biassed
the results in either direction. However, the centres that
did accounted for approximately 900 out of the reported
1118 ASCT performed (80.5%) in Australasia in 2013,21
which represents the majority of the autograft centres
and gives a good representation of current practice in
this region. In approaching the centres, we directed our
request to the heads of each unit. Accordingly, the survey was not designed to collect variability of individual
physician practices within units without strict adherence
References
1 Thomas CF Jr, Limper AH.
Pneumocystis pneumonia.
N Engl J Med 2004; 350:
2487–98.
2 De Castro N, Neuville S, Sarfati C,
Ribaud P, Derouin F, Gluckman E et al.
Occurrence of Pneumocystis jiroveci
pneumonia after allogeneic stem cell
transplantation: a 6-year retrospective
study. Bone Marrow Transplant 2005; 36:
879–3.
3 Tuan IZ, Dennison D, Weisdorf DJ.
Pneumocystis carinii pneumonitis
following bone marrow transplantation.
Bone Marrow Transplant 1992; 10:
267–72.
4 eviQ Cancer Treatments Online 2011
[homepage on the Internet]. New South
Wales: Cancer Institute NSW [cited
© 2016 Royal Australasian College of Physicians
5
6
7
8
to centralised unit protocols; however, we feel that this
is unlikely to be a major factor influencing the validity of
our results. Furthermore, the reasons behind nonadherence were not addressed in this survey due to the
potential complexity involved in order to simplify the
survey and maximise chances of response.
The heterogeneity within institutions identified from
this practice survey and the not inconsiderable overall
proportion of non-adherence to consensus guidelines
may reflect a lack of strong evidence supporting the policies and hence a healthy scepticism towards their implementation in addition to patient and physician factors
and economic costs. Prospective studies in these areas
appear justified, and such a study investigating the use
of no routine PJP prophylaxis post-ASCT in low-risk
patients is being designed.
Acknowledgements
The authors thank the survey responders – Royal Melbourne, Vic.; The Alfred, Vic.; Eastern Health, Vic.; St
Vincent’s, Vic.; Peter MacCallum Cancer Centre, Vic.;
Royal Prince Alfred, NSW; Nepean Cancer Care Centre,
NSW; Calvary Mater Newcastle, NSW; Gosford, NSW; St
George, NSW; St Vincent’s, NSW; Royal North Shore,
NSW; Westmead, NSW; Concord Repatriation, NSW;
Prince of Wales, NSW; Canberra Hospital, ACT; Princess
Alexandra, QLD; Royal Brisbane, QLD; Townsville Hospital, QLD; Royal Perth, WA; Sir Charles Gairdner, WA;
Canterbury District Health Board, NZ; Waikato DHB, NZ;
Wellington, NZ; Auckland City, NZ.
2015 Apr 8]. Available from URL:
https://www.eviq.org.au/
Tomblyn M, Chiller T, Einsele H,
Gress R, Sepkowitz K, Storek J et al.
Guidelines for preventing infectious
complications among hematopoietic cell
transplantation recipients: a global
perspective. Biol Blood Marrow Transplant
2009; 15: 1143–238.
Australian Government Department of
Health and Ageing. The Australian
Immunisation Handbook, 10th edn.
Canberra: The Australian Government
Department of Health; 2013, 156–7.
Ljungman P, Cordonnier C, Einsele H,
Englund J, Machado CM, Storek J et al.
Vaccination of hematopoietic cell
transplant recipients. Bone Marrow
Transplant 2009; 44: 521–6.
Cooley L, Dendle C, Wolf J, Teh BW,
Chen SC, Boutlis C et al. Consensus
guidelines for diagnosis, prophylaxis
and management of Pneumocystis jirovecii
pneumonia in patients with
haematological and solid malignancies,
2014. Intern Med J 2014; 44: 1350–63.
9 Raser KJ, Yanik GA, Magenau JM,
Goldtein SC, Pawarode A, Kitko C et al.
Pneumocystis Jirovecii pneumonia in
recipients of autologous hematopoietic
stem cell transplantation: a 10-year
cohort study. Blood 2013; 122: 3372.
10 Grigg A, Sasadeusz J. Hepatitis B
reactivation after thalidomide. Intern
Med J 2008; 38: 301–2.
11 Lopez Bernaldo de Quiros JC, Miro JM,
Peña JM, Podzamczer D, Alberdi JC,
Martínez E et al. A randomized trial of
the discontinuation of primary and
secondary prophylaxis against
Pneumocystis carinii pneumonia after
highly active antiretroviral therapy in
349
Brief Communications
12
13
14
15
patients with HIV infection. N Engl J
Med 2001; 344: 159–67.
Mansharamani NG, Balachandran D,
Vernovsky I, Garland R, Koziel H.
Peripheral blood CD4 + T-lymphocyte
counts during Pneumocystis carinii
pneumonia in immunocompromised
patients without HIV infection. Chest
2000; 118: 712–20.
Steingrimsdottir H, Gruber A,
Björkholm M, Svensson A, Hansson M.
Immune reconstitution after autologous
hematopoietic stem cell transplantation
in relation to underlying disease, type of
high-dose therapy and infectious
complications. Haematologica 2000; 85:
832–8.
Overgaard UM, Helweg-Larsen J.
Pneumocystis jiroveci pneumonia (PCP) in
HIV-1-negative patients: a retrospective
study 2002–2004. Scand J Infect Dis
2007; 39: 589–95.
Sandherr M, Einsele H, Hebart H,
Kahl C, Kern W, Kiehl M et al. Antiviral
prophylaxis in patients with
haematological malignancies and solid
tumours: Guidelines of the Infectious
Diseases Working Party (AGIHO) of the
German Society for Hematology and
Oncology (DGHO). Ann Oncol 2006; 17:
1051–9.
16 Schuchter LM, Wingard JR,
Piantadosi S, Burns WH, Santos GW,
Saral R. Herpes zoster infection after
autologous bone marrow
transplantation. Blood 1989; 74: 1424–7.
17 Yahav D, Gafter-Gvili A, Muchtar E,
Skalsky K, Kariv G, Yeshurun M et al.
Antiviral prophylaxis in haematological
patients: systematic review and metaanalysis. Eur J Cancer 2009; 45:
3131–48.
18 Truong Q, Veltri L, Kanate AS, Hu Y,
Craig M, Hamadani M et al. Impact of
the duration of antiviral prophylaxis on
rates of varicella-zoster virus
reactivation disease in autologous
hematopoietic cell transplantation
recipients. Ann Hematol 2014; 93:
677–82.
19 Ariza-Heredia EJ, Gulbis AM, Stolar KR,
Kebriaei P, Shah DP, McConn KK et al.
Vaccination guidelines after
hematopoietic stem cell transplantation:
practitioners’ knowledge, attitudes, and
gap between guidelines and clinical
practice. Transpl Infect Dis 2014; 16:
878–6.
20 Small TN, Cowan MJ. Immunization of
hematopoietic stem cell transplant
recipients against vaccine-preventable
diseases. Expert Rev Clin Immunol 2011;
7: 193–203.
21 Australasian Bone Marrow Transplant
Recipient Registry. Australasian Bone
Marrow Transplant Registry: Annual Data
Summary 2013. Darlinghurst: ABMTRR;
2014.
Appendix I
A Survey of Infection Prophylaxis in Autologous Stem Cell Transplantation (ASCT)
Use of PJP Prophylaxis post-ASCT
1. What is the average number of ASCT performed in your centre per year? ____________
2. Is PJP prophylaxis routinely used post-ASCT?
a. Not routinely used (See Q7)
b. Yes – in all patients
c. Yes – in selected patients. Which groups? ____________
3. Does post-ASCT maintenance (eg thalidomide, thalidomide/steroids, rituximab) influence the use of PJP
prophylaxis?
a. No
b. Yes – which maintenance regimen?
i. Thalidomide alone
ii. Thalidomide/Prednisolone
iii. Rituxumab or other monoclonal antibodies
iv. Other_______________________
4. If used, what is the schedule of the first line PJP prophylaxis used? ____
a. Cotrimoxazole DS BD twice a week
b. Cotrimoxazole DS OD three times a week
c. Cotrimoxazole 80/400 mg i OD daily
d. Others (pls describe) _______
5. If used, how
a. When is it started? _______
b. How long is prophylaxis given for? _______
6. Does CD4+ count influence the duration of prophylaxis (eg in determining when to stop)?
a. If yes (what is the cutoff _________)
b. No
350
© 2016 Royal Australasian College of Physicians
Brief Communications
7. Can you recall any PJP complications in the first 12 months post-ASCT in the absence of other risk factors (eg high
dose steroids)? ____
a. Yes (how many ____ ) (see Q8)
b. No (see Q10)
8. When did the PJP complications occur? ____
a. Within first 6 months post-ASCT
b. 6–12 months post-ASCT
9. Did the PJP complications occur while on PJP prophylaxis? ____
a. Yes (which prophylaxis ____________)
b. No
10. Would your centre be interested in participating in a prospective study to investigate the necessity of PJP prophylaxis post-autograft in the absence of other immunosuppressive therapy? ____
a. Yes
b. No
Use of Antiviral Prophylaxis post-ASCT in the absence of concomitant immunosuppressive therapy
1. Is antiviral prophylaxis used in this context post-engraftment? ____
a. Yes
b. No
2. Which antiviral prophylaxis is used? ____
a. Aciclovir 400 mg oral BD
b. Aciclovir 200 mg tds
c. Valaciclovir 500 mg OD
d. Others (please describe______________)
3. Describe the schedule of antiviral prophylaxis used.
a. Time commenced______________
b. Time ceased______________
Vaccination Policy post-ASCT
1. Is vaccination routinely given to ASCT patients? ____
a. Yes
b. No
2. If yes, which guidelines do you use?
a. Australian Immunisation Handbook in conjunction with the BMT 2009 guidelines (Ljungman et al. Vaccination of
Haematopoetic Cell Transplant Recipients. BMT 2009:44:521–526)
b. Others (please state_________________)
Thank you!
Are you willing for your answers to be anonymously included on a review of the use of infection prophylaxis in the
haematological conditions discussed above?
_______________________
Other comments___________________________________________
© 2016 Royal Australasian College of Physicians
351
Safety and efficacy of a novel short occlusive regimen of
imiquimod for selected non-melanotic skin lesions in renal
transplant patients
G. Das,1 B. Tan2 and K. Nicholls1,3
Departments of 1Nephrology and 2Dermatology, The Royal Melbourne Hospital and 3Department of Medicine, Royal Melbourne Hospital,
The University of Melbourne, Melbourne, Victoria, Australia
Key words
imiquimod, renal transplant, non-melanotic skin
cancer, BCC, Bowen disease, actinic keratoses.
Correspondence
Gayatri Das, The Department of Nephrology,
The Royal Melbourne Hospital, Melbourne, Vic.
3050, Australia.
Email: [email protected]
Abstract
Australian patients remain at very high risk of non-melanotic skin cancer after renal
transplantation. Surgical excision offers a cure but destroys tissue and may jeopardise
function and cosmesis. We report excellent safety and efficacy using topical imiquimod
in a novel short intensive regimen in 10 renal transplant patients with superficial basal
cell carcinomas, Bowen disease or actinic keratosis. Outcomes compare well to those
reported with extended-use imiquimod protocols.
Received 9 June 2015; accepted 4
January 2016.
doi:10.1111/imj.13004
While transplantation is undisputed as the treatment of
choice for end-stage renal disease, chronic immunosuppression induces a significant risk of malignancy.1 While
graft survival has almost doubled over the past 20 years,
Australian renal transplant patients (RTP) remain at high
risk of non-melanotic skin cancer (NMSC), mainly squamous cell carcinoma (SCC).2 In transplant patients, SCC
occurs 65 times more frequently, and basal cell carcinomas (BCC) 10 times more frequently, than in the general
population.3,4 NMSC incidence is cumulative over time;
Australia has the highest incidence in the world at 38%
at 10 years and 70% at 20 years, with the adjusted relative risk for development of NMSC in Australian compared with Dutch patients being 3.6.5 There is no
evidence that newer immunosuppressive regimens have
decreased cancer risks.6
Management of NMSC in RTP is identical to immunocompetent patients; surgery and histological confirmation of complete excision is the mainstay of therapy.
However, due to the increased serial incidence of skin
lesions, good cosmetic and functional results may be difficult to achieve surgically. In addition to destructive
Funding: G. Das was supported by the International Society of
Nephrology Fellowship programme, by the Australian and
New Zealand Society of Nephrology and by the Department of
Nephrology Research, Royal Melbourne Hospital.
Conflict of interest: None.
352
modalities, such as cryotherapy, surgical excision and
curettage and diathermy, superficial malignancies are
amenable to topical treatment, achieving the triple aims
of cure, preservation of function and good cosmetic
results.
Imiquimod is a topical treatment option approved by
the Australian Therapeutic Goods Administration for the
primary treatment of actinic keratosis (AK) and superficial BCC. A small molecule of the imidazoquinoline
group of nucleoside analogues,4,7–9 imiquimod promotes
anti-tumour activity through stimulation of toll-like
receptors 7 and 8 on monocytes/macrophages and dendritic cells. This results in the activation of NF Kappa B
and expression of multiple cytokines, chemokines and
inflammatory mediators. Angiogenesis is inhibited,
apoptosis is promoted and the anti-tumour immune
response is enhanced.7–9 Imiquimod has been proven
safe and efficient in both non-immunosuppressed7–13
and immunosuppressed patients,14–16 with an 83% clearance rate in Phase III studies for superficial BCC.6 In
a randomised control trial of cryotherapy versus topical
5-fluorouracil versus topical imiquimod for multiple AK
in non-immunosuppressed patients,17 imiquimod
achieved comparable initial clearance rates to 5FU (85%
vs 96%) and significantly better sustained clearance rates
for both the lesion (73 vs 54%) and the total treatment
field (73 vs 33%). Both creams out-performed cryotherapy when applied two to three times weekly for
4–8 weeks.17 This duration is shorter than the
© 2016 Royal Australasian College of Physicians
Brief Communications
established regimen of applying two to three times
weekly for 16 weeks.7–16
Imiquimod in Bowen disease (superficial SCC) is less
studied, but efficacy is reported despite concomitant
immunosuppression.15,16,18
Shorter-duration treatment protocols offer financial
advantage and allow earlier recognition of treatment
failure. We report here the safety and efficacy of a shortduration, lower-dose occlusive imiquimod regimen for
treatment of superficial NMSC and AK in RTP.
Patients were selected for imiquimod treatment by a
specialist dermatologist (BT) according to small lesion
size, superficial depth and cosmetically sensitive site. All
RTP treated with topical imiquimod between its first use
in July 2004 and December 2013 were identified
through pharmacy records. Clinical data were retrospectively collected from hospital records and our departmental database. All biopsy-confirmed NMSC treated
with imiquimod were included in this study. Imiquimod
(5% cream, Aldara, Meda Pharmaceuticals, Takeley,
Bishop's Stortford, UK) was applied daily for 7 consecutive days under plastic occlusion or tape. Between daily
applications, the area was washed with mild soap and
water. At 1 week, the dermatologist assessed the local
inflammatory reaction and extended treatment for a further 5–7 days if none was present. If no response was
achieved at 2 weeks, alternative treatment was initiated.
Patients were assessed again at 3 months, when complete clinical response was defined as 100% clearance
and partial response (PR) as clearance of at least 50%.
All patients were followed up for at least 9 months.
Immunosuppression in all patients comprised a standard combination of corticosteroids, anti-proliferative
agent (usually mycophenolate mofetil) and calcineurin
inhibitor (usually Tacrolimus). Estimated GFR (e-GFR,
MDRD equation) for each patient was recorded at baseline and 3 months post treatment with imiquimod, and
patient records and biopsy databases were searched to
identify rejection episodes.
Data were available for 12 NMSC treated with topical
imiquimod in 10 RTP (six females, four males) with
mean age 58.2 12 years (range: 36.9–75 years). Targeted lesions were Bowen disease (five lesions in four
patients), superficial BCC (five in four patients) and AK
(two in two patients). No patient had a history of skin
lesions prior to renal transplantation.
Anatomically, lesions were in sun-exposed areas: six
head and neck lesions (in four patients); three arm
lesions (three patients); two leg lesions (two patients)
and a single trunk lesion in another patient. The mean
follow-up time after imiquimod treatment was 5.2 3.4
years (range: 0.8–8.9 years).
All lesions treated with imiquimod initially responded.
None of the five superficial BCCs recurred, but one of
the five Bowen lesions recurred after 12 months, and
one of the two AK recurred after 4 months. Thus, complete response at 3 months was achieved in 10/12
lesions (83%), with ultimate treatment failure in 2 of
12 lesions (Table 1).
Renal function in all patients was stable throughout
treatment with topical imiquimod (Fig. 1). No patient
experienced any decline in renal allograft function during or after treatment; no graft biopsies were undertaken; and no patient reported any systemic symptoms.
Our results using this novel intensive therapeutic regimen of imiquimod for NMSC in RTP are comparable
with reported data using an extended-use protocol.14
The earlier defined end point allows for timely identification of treatment failure and institution of alternative
therapy if necessary, facilitates monitoring and compliance and reduces the required dose, improving cost
efficacy.
Consistent with literature reports, side-effects were
very well tolerated by all our patients and were confined
to the local inflammatory reactions explained by local
induction of pro-inflammatory cytokines and strongly
linked to drug efficacy. No patient withdrew from therapy. It is possible that immunosuppression may reduce
Table 1 Treatment response
Patient
1
2
3
4
5
6
7
8
9
10
No. tumours
Type of lesions
Response
Recurrence/time of recurrence
2
1
1
1
1
1
2
1
1
1
Bowen disease
Bowen disease
Bowen disease
Bowen disease
sBCC
sBCC
sBCC
sBCC
AK
AK
PR/CR
CR
CR
CR
CR
CR
CR/CR
CR
CR
PR
Yes (>1 year)/No
No
No
No
No
No
No/No
No
No
Yes (>4 months)
Follow-up time (years)
8.6
1
8.4
8.9
8.2
6.6
0.7
1.2
3.6
4.3
AK, actinic keratosis; CR, clinical response; PR, partial response; sBCC, superficial basal cell carcinoma
© 2016 Royal Australasian College of Physicians
353
Brief Communications
100
e-GFR ml/mins / 1.73 m2
90
80
70
60
50
40
30
20
10
0
PRE
IMIQUIMOD
3 M POST
IMIQUIMOD
Figure 1 e-GFR (calculated using the Modification of Diet in Renal Disease equation) baseline and 3 months after imiquimod therapy.
the severity of the inflammatory response to imiquimod,
but immunosuppression itself did not prevent the therapeutic response. None of our patients experienced the
less commonly reported systemic adverse effects of fever,
fatigue, headache, nausea, diarrhoea and muscle pain.
Our patient numbers do not allow for a comment on
the relative efficacy of imiquimod for superficial BCC,
AK and Bowen disease, although our results endorse the
published levels of efficacy across all lesion types. Our
100% clearance rate of superficial BCC, without recurrence, is consistent with the high efficacy reported with
longer-duration imiquimod protocols in both immunocompetent and immunosuppressed patients. Of our two
patients treated for AK, one had a complete and the
other a PR, although the latter had recurrence at
4 months. In our four BD patients, all five lesions initially responded. One lesion recurred at 12 months, did
not then respond to 5-fluorouracil cream and was subsequently surgically managed. This experience is comparable with that of Smith et al.,18 who reported total clinical
and histological clearance of BD lesions in five immunosuppressed patients, using a combination of imiquimod
5% cream and 5% 5-fluorouracil gel. Clearance was
achieved within 9 weeks, with no evidence of
References
1 Gutierrez-Dalmau A, Campistol JM.
Immunosuppressive therapy and
malignancy in organ transplant
recipients: a systematic review. Drugs
2007; 67: 1167–98.
354
recurrence over the follow-up period of 3–15 months.
Two additional studies reported similar results in biopsyproven BD in transplant patients. Prinz et al.15 reported
clearance of all four BD lesions after an average of
6 weeks using imiquimod 5% cream three times weekly,
although one lesion recurred after 10 months, while
Kovak and Stasko19 reported complete clearance of BD
lesions in two transplant patients and PR in a third
patient.
Few studies have comprehensively compared the costeffectiveness of the newer treatment modalities for
NMSC with surgical excision. A Spanish study of
209 patients20 analysed the cost of treating a single
superficial BCC (<2 cm) with either surgery or imiquimod 5% cream, applied five times per week for 6 weeks
and using 36 sachets. The use of imiquimod reduced the
cost per patient cured compared to surgical excision, by
8.1% in dermatology and 36% in non-dermatology services, thus achieving cost-effectiveness.20 The efficacy of
topical imiquimod was 82% at 1 year. However, the
costs of treatment failures and of follow up of possible
failures were not considered, and the study was not
randomised.
As imiquimod works through immune stimulation, its
use in RTP has induced concern regarding rejection risk.
This theoretical concern has not been substantiated in
our study or in the literature to date,4 probably because
systemic drug effects are very low with topical administration. Renal function was stable for all our RTP, with
no episodes of rejection throughout follow up.
Within the limitation of our retrospective study, the
short occlusive regimen of imiquimod was very well tolerated; patient acceptance and compliance was high; and
efficacy was excellent. We conclude that a protocol of 7day self-application of imiquimod with occlusion is a safe
and effective alternative to higher-dose regimens or surgery in selected cases of superficial BCC, Bowen disease
and AK in RTP.
Acknowledgement
The advice and assistance of Ms Michelle Nalder M
Pharm is gratefully acknowledged.
2 Excell L, Russ G, Wride P. ANZOD
Registry Report 2005. Adelaide: Australian
and New Zealand Organ Donation
Registry; 2005.
3 Jensen P, Hansen S, Moller B, Leivestad
T, Pfeffer P, Geiran O et al. Skin cancer
in kidney and heart transplant
recipients and different long-term
immunosuppressive therapy regimens.
J Am Acad Dermatol 1999; 40: 177–86.
4 Bangash HK, Colegio OR. Management
of non-melanoma skin cancer in
immunocompromised solid organ
transplant recipients. Curr Treat Options
Oncol 2012; 13: 354–76.
© 2016 Royal Australasian College of Physicians
Brief Communications
5 Bavinck JNB, Hardie DR, Green A,
Cutmore S, MacNaught A, O’Sullivan B
et al. The risk of skin cancer in renal
transplant recipients in Queensland,
Australia. A follow-up study.
Transplantation 1996; 61: 715–21.
6 Gallagher MP, Kelly PJ, Jardine M,
Perkovic V, Cass A, Craig JC et al. Longterm cancer risk of immunosuppressive
regimens after kidney transplantation.
J Am Soc Nephrol 2010; 21: 852–8.
7 Geisse J, Caro I, Lindholm J, Golitz L,
Stampone P, Owens M. Imiquimod 5%
cream for the treatment of superficial
basal cell carcinoma: results from two
phase III, randomized, vehicle
controlled studies. J Am Acad Dermatol
2004; 50: 722–33.
8 Burns CA, Brown MD. Imiquimod for
the treatment of skin cancer. Dermatol
Clin 2005; 23: 151–64.
9 Vidal D, Alomar A. Mode of action and
clinical uses of imiquimod. Expert Rev
Dermatol 2008; 3: 151–9.
10 Ryu J, Yang FC. A review of topical
imiquimod in the management of basal
cell carcinoma, actinic keratoses, and
other skin lesions. Clin Med Insights Ther
2009; 1: 1557–75.
11 Lebwohl M, Dinehart S, Whiting D, Lee
PK, Tawfik N, Jorizzo J et al. Imiquimod
5% cream for the treatment of actinic
keratosis: results from two phase III,
randomized, double-blind, parallel
© 2016 Royal Australasian College of Physicians
12
13
14
15
group, vehicle-controlled trials. J Am
Acad Dermatol 2004; 50: 714–21.
Alomar A, Bichel J, McRae S. Vehiclecontrolled, randomized, double-blind
study to assess safety and efficacy of
Imiquimod 5% cream applied once
daily 3 days per week in one or two
courses of treatment of actinic keratosis
on the head. Br J Dermatol 2007; 157:
133–41.
Korman N, Moy R, Ling M, Matheson
R, Smith S, McKane S et al. Dosing with
5% Imiquimod cream 3 times per week
for the treatment of actinic keratosis:
results of two phase 3, randomized,
double-blind, parallel-group, vehiclecontrolled trials. Arch Dermatol 2005;
141: 467–73.
Ulrich C, Bichel J, Euvrard S, Guid B,
Proby CM, Van de Kerkof PCM et al.
Topical immunomodulation under
systemic immunosuppression: results of
a multicenter, randomized, placebocontrolled safety and efficacy study of
Imiquimod 5% cream for the treatment
of actinic keratoses in kidney, heart,
and liver transplant patients. Br J
Dermatol 2007; 157: 25–31.
Prinz BM, Hafner J, Dummer R,
Burg G, Bruswanger U, Kempf W.
Treatment of Bowen’s disease with
Imiquimod 5% cream in transplant
recipients. Transplantation 2004; 77:
790–1.
16 Brown VL, Atkins CL, Ghali L, Cerio R,
Harwood CA, Proby CM. Safety and
efficacy of 5% Imiquimod cream for the
treatment of skin dysplasia in high-risk
renal transplant recipients: randomized,
double-blind, placebo-controlled trial.
Arch Dermatol 2005; 141: 985–3.
17 Krawtchenko N, Roewert-Huber J,
Ulrich M, Mann I, Sterry W,
Stockfleth E. A randomised study of
topical 5% imiquimod vs. topical 5fluorouracil vs. cryosurgery in
immunocompetent patients with actinic
keratoses: a comparison of clinical and
histological outcomes including 1-year
follow-up. Br J Dermatol 2007; 157:
34–40.
18 Smith KJ, Germain M, Skelton H.
Squamous cell carcinoma in situ
(Bowen’s disease) in renal transplant
patients treated with 5% Imiquimod
and 5% 5-fluorouracil therapy. Dermatol
Surg 2001; 27: 561–4.
19 Kovach BT, Stasko T. Use of topical
immunomodulators in organ transplant
recipients. Dermatol Ther 2005; 18:
19–27.
20 Vanaclocha F, Dauden E, Badia X,
Guillén C, Conejo-Mir JS, Sainz de Los
Terreros M et al. Cost effectiveness of
treatment of superficial basal cell
carcinoma: surgical excision
vs. Imiquimod 5% cream. Br J Dermatol
2007; 156: 769–1.
355
Acute adrenal insufficiency: an aide-memoire of the critical
importance of its recognition and prevention
A. Gargya,1 E. Chua,2 J. Hetherington,2 K. Sommer3 and M. Cooper4
1
Department of Endocrinology, Royal Prince Alfred Hospital, University of New South Wales, 2Department of Endocrinology, Royal Prince Alfred
Hospital, University of Sydney, 3Department of Endocrinology, Concord Hospital and 4Department of Endocrinology, Concord Hospital, University of
Sydney, Sydney, New South Wales, Australia
Key words
adrenal crisis, adrenal insufficiency, Addison’s
disease, corticosteroids.
Correspondence
Ash Gargya, Department of Endocrinology,
Royal Prince Alfred Hospital, Missenden Road,
Camperdown, NSW 2050, Australia.
Email: [email protected]
Received 9 July 2015; accepted 16
December 2015.
doi:10.1111/imj.12998
Abstract
Adrenal crisis is a life-threatening emergency that causes significant excess mortality in
patients with adrenal insufficiency. Delayed recognition by medical staff of an impending adrenal crisis and failure to give timely hydrocortisone therapy within the emergency department continue to be commonly encountered, even in metropolitan
teaching hospitals. Within the authors’ institutions, several cases of poorly handled
adrenal crises have occurred over the last 2 years. Anecdotal accounts from members
of the Addison’s support group suggest that these issues are common in Australia. This
manuscript is a timely reminder for clinical staff on the critical importance of the recognition, treatment and prevention of adrenal crisis. The manuscript: (i) outlines a case
and the clinical outcome of sub-optimally managed adrenal crisis, (ii) summarises the
clinical features and acute management of adrenal crisis, (iii) provides recommendations on the prevention of adrenal crisis and (iv) provides guidance on the management
of ‘sick days’ in patients with adrenal insufficiency.
Our index patient was a 51-year-old woman with a 7year history of autoimmune primary adrenal insufficiency. She usually took 20 mg hydrocortisone in the
morning and 10 mg at 3 pm and 100 μg fludrocortisone
twice daily. She had a MedicAlert bracelet (MedicAlert
Foundation, Adelaide, SA, Australia) and an understanding of sick day management. She refused to have a parenteral hydrocortisone emergency kit at home, but was
well aware of the need to present to hospital in the event
of an impending crisis. She developed a gastrointestinal
viral infection (of 24-h duration) with vomiting and
diarrhoea. She tripled her oral hydrocortisone dose, but
could not keep the tablets down. The ambulance service
was called. At retrieval, her blood pressure was
80/60 mmHg, and 1 L of Hartmann’s solution was given
en route to hospital. On arrival to the emergency department, her blood pressure was 101/62 mmHg, and she
was labelled ‘normotensive’ and triaged to be medically
reviewed within 60 min. Intravenous fluids and hydrocortisone were not administered despite repeated
requests by the patient’s family. Three hours elapsed,
and her condition deteriorated, eventuating in
Funding: None.
Conflicts of interest: None.
356
cardiorespiratory arrest. Her arterial pH was 7.06. She
was resuscitated, intubated and transferred to the intensive care unit (ICU). Her treating endocrinologist was
first notified of her hospital presentation and admission
5 h after the transfer to ICU. Her hospital length of stay
was 16 days, and her admission was complicated by
stress-induced cardiomyopathy and a broken tooth from
intubation. The actual and optimal management of our
index patient is summarised in Table 1.
Discussion
Adrenal crisis is a life-threatening emergency that causes
significant excess mortality in patients with adrenal
insufficiency. Epidemiological studies indicate an incidence of between 5 and 10 adrenal crises per 100 patient
years each year; 1 in 200 patients will die from an adrenal crisis.1 A recent Australian study reviewed national
databases over a 13-year period and identified an
increase in adrenal crisis hospital admission rates (further highlighting the increasing importance of prompt
medical staff recognition and treatment of adrenal crises)
that was positively correlated with a rise in short-acting
glucocorticoid prescription rates. In lieu of modern-day
lower-dose glucocorticoid replacement regimens, this
study suggests a possible causal relationship between
© 2016 Royal Australasian College of Physicians
Brief Communications
Table 1 Initial in-hospital management of the index patient presenting with adrenal crisis
Triage
Medical
assessment
Parenteral
hydrocortisone
Hydrocortisone
dose
Endocrine
consult
Index patient
Addison’s disease
noted but
triaged as nonurgent
After 3 h
Administered
after cardiac
arrest
100 mg IV
5 h after ICU
admission
Optimal
management
Identify at risk of
adrenal crisis
Rapid, ideally
within 10 min
Administer
immediately
100 mg IM or
IV
Yes
Outcome
ICU admission 7 days
16 days hospital
admission
Complicated by stressinduced
cardiomyopathy and
broken tooth from
intubation
Dissatisfied family
Expect clinical
recovery in
24–48 h, depending
on precipitant
ICU, intensive care unit.
adrenal crisis events and the increasing use of shortacting glucocorticoids +/− reduced effective doses.2
Although the treatment of an adrenal crisis should be
straightforward, delayed recognition by medical staff of
an impending or established adrenal crisis and failure to
give timely hydrocortisone therapy within the emergency department continue to be commonly encountered even in metropolitan teaching hospitals. Within
the authors’ institutions, several cases of poorly handled
adrenal crises, such as the case above, have occurred
over the last 2 years. Anecdotal accounts from members
of the Addison’s support group suggest that these issues
are common throughout Australia. Several clinical and
sociological factors appear to account for these delays
and potentially life-threatening omissions. Clinical staff
commonly fail to recognise the severity of an impending
adrenal crisis as hypotension develops rapidly. Emergency department staff are often concerned about precipitating harmful effects with the administration of
hydrocortisone. However, such effects are unlikely, and
the risks are considerably outweighed by the likely benefits. Perhaps of most concern is that there are many
instances where glucocorticoid treatment is delayed even
when patients (and/or their carers) clearly outline the
need for treatment, and the need is emphasised in written materials or MedicAlert bracelets.
Suspected adrenal crisis requires immediate therapeutic intervention (in undiagnosed adrenal insufficiency,
treatment usually precedes biochemical proof of diagnosis). Treatment is simple and entails parenteral hydrocortisone (an initial bolus dose of 100 mg hydrocortisone)
and isotonic saline to correct volume depletion. Prompt
recognition and treatment of an impending crisis generally results in clinical recovery within 24 h.1 It has been
© 2016 Royal Australasian College of Physicians
observed that clinical recovery from untreated adrenal
insufficiency presenting with confusion and somnolence
can take several days or even up to 1 week. Parenteral
hydrocortisone followed by a timely change to an
increased dose of oral glucocorticoids with subsequent
tapering and overlap (generally 24–48 h following clinical improvement) is recommended, depending on the
precipitant. Treatment of the inter-current illness and/or
precipitant (e.g. with antibiotics etcetera) is required.
Infections, particularly gastroenteritis, are the most frequent causes of an adrenal crisis.1,3 Patients on long-term
exogenous glucocorticoids (for a non-endocrine condition) are at risk of an adrenal crisis.
Patient education is considered critical to prevent
crises.4 Preventative strategies to ensure early recognition and timely intervention of an impending adrenal
crisis are essential to reduce morbidity (including hospital length of stay and time to clinical recovery) and mortality from a crisis. Patients and carers of those with
adrenal insufficiency commonly understand their condition very well, sometimes even better than the health
professionals, and feel distressed and frustrated while
encountering barriers to emergency care. Listening to a
well-informed patient and/or carer in adrenal crisis who
says that he or she needs steroids and taking urgent
action will avoid unnecessary deaths from this treatable
medical problem.5 Proper information on ‘sick day’ management with written instructions on glucocorticoid dose
adjustment during stressful situations is usually provided
(Table 2). Recommendations regarding glucocorticoid
coverage during non-surgical illnesses are largely based
on expert opinion; the guidelines provided herein are in
general
agreement
with
other
published
recommendations.1,6–8
Individuals
with
adrenal
357
358
e.g. mild infection (such as cystitis) with low grade
temperature
e.g. high fevers
e.g. gastroenteritis with vomiting +/− diarrhoea or
pneumonia etc
e.g. mild cold
Symptoms
>38.5 C
Could be normal or raised
No temperature (less than
37.4 C)
37.5–38.5 C
Temperature
Parenteral hydrocortisone (50–100 mg IV bolus at
induction followed by 25–50 mg IV every 8 h till
stable) then 2× normal oral dose
No
Yes
Yes
NSW
†Only dexamethasone (adrenal insufficiency is not stated as an indication for its use).
‡If patient is carrying a letter stating it is required in the appropriate clinical setting.
Are there clinical practice guidelines and/or drug therapy
protocols for acute adrenal insufficiency?
Does the ambulance carry parenteral hydrocortisone?
Can paramedics administer patient’s own emergency parenteral
hydrocortisone?
Table 3 Ambulance service role in managing acute adrenal insufficiency
No†
Yes‡
No
VIC
Yes
Yes
Yes
ACT
Yes
Yes
Yes
QLD
Yes
Yes
No
SA
Yes
Yes
Yes
NT
No†
No
No
WA
Continue for 48 h post
procedure
24 h
No†
Yes‡
No
TAS
Until recovery plus 2 days
Until recovery plus 2 days
Until recovery plus 1–2 days
2× normal oral dose
3× normal oral dose†
Early parenteral hydrocortisone (50–100 mg IV bolus
followed by 25–50 mg IV every 8 h till the condition
stabilises) then 2–3× normal oral dose
2× normal oral dose
N/A
Duration of oral dose
adjustment
No change
Dose adjustment
†The ‘3 × 3 rule’ means tripling the patient’s usual oral dose for 3 days or for the duration of the illness after which the patient resumes their usual dose.6,7
Minor dental procedure (>1 h under local anaesthetic), significant injury
or major emotional stress
Major surgery or procedure with general anaesthetic
Trivial
illness
Mildly
unwell
More
unwell
Issue
Table 2 Sick day management for patients on glucocorticoid therapy
Brief Communications
© 2016 Royal Australasian College of Physicians
Brief Communications
insufficiency are encouraged to carry an emergency card
and/or a MedicAlert bracelet that succinctly reflects glucocorticoid dependency. Every patient should be provided with an emergency kit for parenteral
hydrocortisone self-administration (this injection can be
administered by a trained relative and/or a healthcare
professional, e.g. paramedic, prior to transfer to the hospital emergency department). Table 3 summarises
whether each Australian state’s ambulance service has
clinical practice guidelines and/or drug therapy protocols
for managing acute adrenal insufficiency, whether parenteral hydrocortisone is available in the ambulance
and, if unavailable, whether paramedics are able to
administer the patient’s emergency parenteral hydrocortisone. Patient education group meetings may be an
References
1 Allolio B. Extensive expertise in
endocrinology: adrenal crisis. Eur J
Endocrinol 2015; 172: R115–24.
2 Rushworth RL, Torpy DJ. Adrenal
insufficiency in Australia: is it possible
that the use of lower dose, short-acting
glucocorticoids has increased the risk of
adrenal crisis? Horm Metab Res 2015; 47:
427–32.
3 Johannsson G, Falorni A, Skrtic S,
Lennernas H, Quinkler M, Monson J
et al. Adrenal insufficiency: review of
clinical outcomes with current
© 2016 Royal Australasian College of Physicians
effective intervention.9 Hospital computer systems may
be able to ‘flag’ patients with adrenal insufficiency.
In conclusion, prompt recognition and treatment of an
impending adrenal crisis is critical to reduce its associated
morbidity and mortality. Healthcare workers must be
acutely aware of its presentation (including the initial
false appearance of clinical stability with potential rapid
deterioration) and appreciate the urgency of its treatment. Patients, carers and families must be well educated
and updated on ‘sick day’ management using a structured and quality-controlled approach. A MedicAlert or
emergency card and an emergency hydrocortisone kit
ready for use by self or others are an essential part of
patient management.
glucocorticoid replacement therapy. Clin
Endocrinol (Oxf ) 2014; 82: 1–0.
4 Hahner S, Loeffler M, Bleicken B,
Drechsler C, Milavanovic D, Fassnacht M
et al. Epidemiology of adrenal crisis in
chronic adrenal insufficiency: the need
for new prevention strategies. Eur J
Endocrinol 2010; 162: 597–602.
5 Wass J, Arlt W. How to avoid
precipitating an acute adrenal crisis. BMJ
2012; 345: e6333.
6 Nenke M, Torpy D. Addison’s disease:
managing “sick days” to avoid crises.
Endocrinol Today 2014; 3:
26–31.
7 Nieman L. Treatment of adrenal
insufficiency in adults. In: Lacroix A,
ed. UpToDate. Waltham: Wolters Kluwer;
2013.
8 Jung C, Inder W. Management of
adrenal insufficiency during the stress of
medical illness and surgery. Med J Aust
2008; 188: 409–13.
9 Han J, Repping-Wuts W, Nike M,
Stikkelbroeck M, Noordzij A, Kerstens M
et al. A glucocorticoid education group
meeting: an effective strategy for
improving self-management to prevent
adrenal crisis. Eur J Endocrinol 2013; 169:
17–22.
359
Severe hypocalcaemia and hypophosphataemia following
intravenous iron and denosumab: a novel drug interaction
B. Smyth and S. Ong
Department of Nephrology, St George Hospital, Sydney, New South Wales, Australia
Key words
hypocalcaemia, hypophosphataemia,
intravenous iron, iron polymaltose, denosumab,
chronic kidney disease.
Correspondence
Brendan Smyth, 205D/250 Anzac Parade,
Kensington, NSW 2033, Australia.
Email: [email protected]
Abstract
We present the case of a 59-year-old woman with chronic kidney disease who suffered
severe hypocalcaemia and hypophosphataemia after receiving denosumab and intravenous iron. This potentially life-threatening adverse drug interaction has never been
reported before. We propose a mechanism to explain it with reference to the physiological derangements caused by both agents on calcium and phosphate homeostasis.
Received 18 March 2015; accepted 3
December 2015.
doi:10.1111/imj.13001
A 59-year-old woman who attended our nephrology
clinic for routine review of stage 3 chronic kidney disease (CKD) due to ischaemic nephrosclerosis and type
2 diabetes mellitus was found to have a corrected calcium of 1.81 mmol/L (normal value: 2.20–2.60) and
phosphate of 0.36 mmol/L (normal value: 0.80–1.50).
She was admitted to the hospital and commenced on
intravenous calcium and phosphate replacement. History
revealed that 26 days prior to presentation, she had
received 60 mg denosumab for osteoporosis and 5 days
later she had received 1000 mg intravenous iron polymaltose for iron deficiency. This was her second denosumab infusion (the first had been 6 months previously
and uncomplicated) and her first iron infusion. Her other
comorbidities were dyslipidaemia and peripheral vascular disease. She was not taking calcium or vitamin D supplements. Further investigations on admission included
a parathyroid hormone (PTH) of 40.2 pmol/L (normal
value: 1.1–6.9), 25-OH vitamin D of 55 nmol/L (normal
value: 25–150), creatinine of 166 μmol/L (normal value:
50–90; estimated glomerular filtration rate (eGFR)
27 mL/min/1.73 m2 by CKD-EPI) and an elevated nonfasting urine fractional excretion of phosphate (FePO4)
of 71.6% (normal value: 10–20%). Seven weeks prior to
presentation, her corrected calcium was 2.42 mmol/L,
phosphate 1.41 mmol/L, creatinine 148 μmol/L (eGFR
Funding: None.
Conflict of interest: None.
360
31 mL/min/1.73m2), PTH 4.6 pmol/L and alkaline phosphatase 42 U/L (Figs 1, 2).
Her admission was uncomplicated. She remained
asymptomatic and was discharged after 3 days on calcitriol 0.5 mcg BD, calcium carbonate 1.2 g daily and
phosphate 500 mg daily. Phosphate, calcium and calcitriol replacements were weaned gradually and all supplements were ceased 3 months after the initial
administration of iron and denosumab. At the final visit,
the patient’s FePO4 had reduced to 34.7% and her
serum calcium and phosphate were within the normal
range.
Discussion
Denosumab is a monoclonal antibody to the receptor
activator of nuclear factor-κB ligand (RANK-L) used to
treat osteoporosis and skeletal complications of metastatic solid organ malignancies. RANK-L is expressed by
osteoblasts and stimulates osteoclastogenesis and bone
resorption through interaction with RANK expressed on
osteoclasts. Denosumab inhibits this interaction, thus
suppressing osteoclast activity leading to increased bone
mineral density.1,2 RANK/RANK-L interactions are also
found in the immune system where they have a role in
lymph node growth, dendritic cell survival, T-cell activation, including regulatory T-cells and induction of tolerance.1 The clinical relevance of these observations on the
use of denosumab is debated as some, but not all, studies
have shown a small absolute increase in risk of infection.
© 2016 Royal Australasian College of Physicians
Brief Communications
Figure 1 Serum corrected calcium and phosphate over time. Baseline at day 1 represents
the most recent blood tests prior to the denosumab and iron administration (day 22 and 27).
, Corrected calcium (mmol/L, dotted line indicates reference intervals; , Phospate (mmol,
L, dashed line indicates reference intervals);
red arrow indicates Denosumab; blue arrow
indicates iron polymaltose; green area indicates
hospital admission.
Figure 2 Serum parathyroid hormone (PTH)
and urine fractional excretion of phosphate
(FEPO4) over time. Baseline at day 1 represents
the most recent blood tests prior to the denosumab and iron administration (days 22 and
27). , FE phosphate (%, dotted line indicates
reference intervals); , Parathyroid hormone
(pmol/L, dashed line indicates reference intervals); red arrow indicates Denosumab; blue
arrow indicates iron polymaltose; green area
indicates hospital admission.
Side effects include eczema, hypercholesterolaemia,
musculoskeletal pain, hypocalcaemia, hypophosphataemia and, rarely, osteonecrosis of the jaw. No previous
drug interactions have been reported for denosumab.
Unlike bisphosphonates, denosumab does not require
dose reduction in renal impairment and is not associated
with nephrotoxicity.2
Hypocalcaemia is a common complication of denosumab treatment for bony metastases, occurring in 5–10%
of patients.3,4 It reflects inhibition of bone resorption and
is accompanied by a compensatory increase in PTH.5
Trials of densumab for post-menopausal osteoporosis
have reported a much lower incidence of hypocalcaemia
(<1%)6,7 reflecting both the higher doses used in malignancy and also the differing states of bone turnover in
patients with metastases versus those with osteoporosis.1,8 Impaired renal function is a well-documented risk
© 2016 Royal Australasian College of Physicians
factor for denosumab-induced hypocalcaemia, which
occurs in 15–45% of patients depending on degree of
renal impairment.9,10 A nadir is typically reached around
10 days after the infusion and may be severe, requiring
intravenous replacement. Serum calcium may take up to
8 weeks to return to the previous levels which correlate
roughly with the mean half-life of denosumab of 28–32
days.2,3,8–12 Hypophosphataemia may also occur in
patients with renal impairment given denosumab. In a
series of 22 patients with malignancy and impaired renal
function (eGFR < 30 mL/min/1.73 m2), in addition to
45% of patients with hypocalcaemia, 32% experienced
hypophosphataemia following 120 mg denosumab.10
The increased incidence of hypocalcaemia after denosumab in patients with renal disease is attributable to
CKD-bone and mineral disorder (CKD-MBD). CKDMBD encompasses a wide spectrum of abnormalities of
361
Brief Communications
bone turnover, vascular calcification and calcium
and phosphate homeostasis that progress as renal function declines. Bone pathology (renal osteodystrophy)
varies widely from low turnover (adynamic) bone disease to high turnover (osteitis fibrosa) bone disease
with mixed disorders occurring. As denosumab acts
to inhibit osteoclast activity, it is hypothesised that
the disordered bone metabolism associated with
CKD predisposes patients to varying degrees of ‘hungry
bone
syndrome’
causing
hypocalcaemia
and
hypophosphataemia.9
Disordered phosphate metabolism is central to the
pathophysiology of CKD-MBD. Elevated serum phosphate is not typically seen until the late stages of CKD
(eGFR < 30 mL/min/1.73 m2). Rather, the earliest
abnormality is an elevation in fibroblast growth factor
23 (FGF-23) which can be seen in stage 3 CKD (eGFR
30–60 mL/min/1.73 m2). FGF-23, produced by osteocytes, is the principal negative regulator of phosphate
homeostasis. It acts on the proximal tubule to reduce
renal phosphate reabsorption through the Na-Pi channel
and inhibits 1-α-hydroxylase (thereby inhibiting the conversion of calcidiol to calcitriol). It also acts directly on
the parathyroid to suppress PTH release.13
Hypophosphataemia following intravenous iron is well
described and may be severe. It is thought to be due to
the elevated levels of FGF-23.14–17 A low serum phosphate, high FePO4 and elevated FGF-23 have been
demonstrated in patients treated with intravenous iron
polymaltose, including in haemodialysis patients.18,19 As
would be expected from the physiological actions of
FGF-23, reductions in PTH and serum 1,25(OH) vitamin
D have also been demonstrated following intravenous
iron highlighting the interdependence of calcium and
phosphate metabolism. Hypophosphataemia may be
more common in patients with CKD. Prats et al. demonstrated hypophosphataemia in 35 of 47 (74%) patients
with renal impairment (mean eGFR 26 mL/min/
1.73 m2) treated with ferric carboxymaltose. In contrast
to other studies, however, C-terminal FGF-23 levels
were reduced.20
We believe that the severity of denosumab-induced
hypocalcaemia and the co-existence of severe hypophosphataemia in this patient were out of proportion to the
References
1 Cheng ML, Fong L. Effects of RANKLtargeted therapy in immunity and
cancer. Front Oncol 2014; 3: 1–8.
2 Narayanan P. Denosumab: a
comprehensive review. South Asian J
Cancer 2013; 2: 272–7.
362
probable degree of disordered bone metabolism. While
this cannot be proven in the absence of a bone biopsy,
her normal baseline serum phosphate, calcium, PTH,
alkaline phosphatase and the context of stage 3 CKD
make it unlikely that she had severe renal osteodystrophy. Instead, the constellation of biochemical abnormalities and her clinical course is best explained by the
superposition of the iatrogenic disturbances of phosphate, calcium and bone metabolism by parenteral iron
and denosumab. The pathophysiology of the interaction
requires the pre-existence of CKD-MBD and is hypothesised to be as follows: denosumab caused hypocalcaemia
through inducing a mild form of ‘hungry bone syndrome’ in a patient with abnormal bone metabolism.
Intravenous iron then induced a rise in FGF-23 that
blunted the homeostatic response to hypocalcaemia by
inhibiting PTH secretion and activation of vitamin D. The
result was a marked exaggeration of denosumabinduced hypocalcaemia. The severity of the patient’s
hypophosphataemia is similarly explained by the conjunction of an increase in FGF-23-induced phosphaturia
following the intravenous iron, reduced activation of
vitamin D causing reduced phosphate absorption and a
failure to buffer the fall in phosphate due to the antiresorptive activity of denosumab. Furthermore, the
hypocalcaemia-induced increase in PTH would be
expected to worsen the hypophosphataemia by promoting renal phosphate wasting.
This case demonstrates a novel and potentially serious
drug interaction that is readily explained with reference
to the current understanding of calcium and phosphate
homeostasis. Impaired renal function, iron deficiency
and osteoporosis are common and co-prevalent, meaning the number of patients at risk for similar complications is not insignificant. Furthermore, a similar
interaction could be hypothesised to occur with intravenous bisphosphonates and iron. In summary, we suggest
avoiding denosumab and intravenous iron in close temporal association in patients with impaired renal function. However if required, then provision of calcium and
calcitriol supplementation and close monitoring of serum
calcium and phosphate in the first 2 weeks after treatment are imperative to prevent life-threatening calcium
or phosphate disturbance.
3 Stopeck AT, Lipton A, Body JJ,
Steger GG, Tonkin K, de Boer RH et al.
Denosumab compared with zoledronic
acid for the treatment of bone
metastases in patients with advanced
breast cancer: a randomized, doubleblind study. J Clin Oncol 2010; 28:
5132–9.
4 Henry DH, Costa L, Goldwasser F,
Hirsh V, Hungria V, Prausova J et al.
Randomized, double-blind study of
denosumab versus zoledronic acid in
the treatment of bone metastases in
patients with advanced cancer
(excluding breast and prostate cancer)
© 2016 Royal Australasian College of Physicians
Brief Communications
5
6
7
8
9
10
or multiple myeloma. J Clin Oncol 2011;
29: 1125–32.
McClung MR, Lewiecki EM, Cohen SB,
Bolognese MA, Woodson GC,
Moffett AH et al. Denosumab in
postmenopausal women with low bone
mineral density. N Engl J Med 2006;
354: 821–31.
Cummings SR, Martin JS,
McClung MR, Siris ES, Eastell R,
Reid IR et al. Denosumab for prevention
of fractures in postmenopausal women
with osteoporosis. N Engl J Med 2009;
361: 756–65.
Recknor C, Czerwinski E, Bone HG,
Bonnick SL, Binkley N, Palacios S et al.
Denosumab compared with ibandronate
in postmenopausal women previously
treated with bisphosphonate therapy.
Obstet Gynecol 2013; 121: 1291–9.
Farinola N, Kanjanapan Y. Denosumabinduced hypocalcaemia in high bone
turnover states of malignancy and
secondary hyperparathyroidism from
renal failure. Intern Med J 2013; 43:
1243–6.
Block GA, Bone HG, Fang L, Lee E,
Padhi D. A single-dose study of
denosumab in patients with various
degrees of renal impairment. J Bone
Miner Res 2012; 27: 1471–9.
Watkins KR, Rogers JE, Atkinson B.
Tolerability of denosumab in metastatic
© 2016 Royal Australasian College of Physicians
11
12
13
14
15
solid tumor patients with renal
insufficiency. Support Care Cancer 2015;
23: 1657–62.
Ikesue H, Tsuji T, Hata K, Watanabe H,
Mishima K, Uchida M et al. Time course
of calcium concentrations and risk
factors for hypocalcemia in patients
receiving denosumab for the treatment
of bone metastases from cancer. Ann
Pharmacother 2014; 48:
1159–65.
Sirvent AE, Enriquez R, Sánchez M,
González C, Millán I, Amorós F.
Extreme hypocalcaemia and
hyperparathyroidism following
denosumab. Is this drug safe in chronic
kidney disease? Nefrologia 2014; 34:
542–4.
Lanske B, Razzaque MS. Molecular
interactions of FGF23 and PTH in
phosphate regulation. Kidney Int 2014;
86: 1072–4.
Shimizu Y, Tada Y, Yamauchi M,
Okamoto T, Suzuki H, Ito N et al.
Hypophosphatemia induced by
intravenous administration of
saccharated ferric oxide: another form
of FGF23-related hypophosphatemia.
Bone 2009; 45: 814–16.
Yamamoto S, Okada Y, Mori H,
Fukumoto S, Tanaka Y. Fibroblast
growth factor 23-related osteomalacia
caused by the prolonged administration
16
17
18
19
20
of saccharated ferric oxide. Intern Med
2012; 51: 2375–8.
Fierz YC, Kenmeni R, Gonthier A,
Lier F, Pralong F, Coti BP. Severe and
prolonged hypophosphatemia after
intravenous iron administration in a
malnourished patient. Eur J Clin Nutr
2014; 68: 531–3.
Blazevic A, Hunze J, Boots JMM.
Severe hypophosphataemia after
intravenous iron administration. Neth J
Med 2014; 72: 49–53.
Schouten BJ, Hunt PJ, Livesey JH,
Frampton CM, Soule SG. FGF23
elevation and hypophosphatemia after
intravenous iron polymaltose: a
prospective study. J Clin Endocrinol
Metab 2009; 94: 2332–7.
Takeda Y, Komaba H, Goto S, Fujii H,
Umezu M, Hasegawa H et al. Effect of
intravenous saccharated ferric oxide on
serum FGF23 and mineral metabolism
in hemodialysis patients. Am J Nephrol
2011; 33: 421–6.
Prats M, Font R, Carmen G, Cabré C,
Jariod M, Vea AM. Effect of ferric
carboxymaltose on serum phosphate
and C-terminal FGF23 levels in nondialysis chronic kidney disease patients:
post-hoc analysis of a prospective study.
BMC Nephrol 2013; 14: 167.
363
PERSONAL VIEWPOINT
Cardiovascular risk reduction in hypertension: angiotensinconverting enzyme inhibitors, angiotensin receptor blockers.
Where are we up to?
A. Sindone,1 J. Erlich,2,3 C. Lee,4 H. Newman,5 M. Suranyi6 and S. D. Roger7
1
Heart Failure Unit and Department of Cardiac Rehabilitation, Concord Repatriation General Hospital, Concord, 2Faculty of Medicine, University of NSW,
Department of Nephrology, Prince of Wales Hospital, 4Department of Cardiology, Nepean Hospital and 5Department of Cardiology and 6Liverpool
Renal Clinical Research Centre, Liverpool Hospital, Sydney and 7Department of Renal Medicine, Gosford Hospital, Gosford, New South Wales, Australia
3
Key words
angiotensin-converting enzyme inhibitor,
angiotensin receptor antagonist, cardiovascular
diseases, risk assessment, hypertension.
Correspondence
Andrew Sindone, Heart Failure Unit and
Department of Cardiac Rehabilitation, Concord
Repatriation General Hospital, Hospital Road,
Concord, NSW 2139, Australia.
Email: [email protected]
Received 16 July 2015; accepted 16
November 2015.
doi:10.1111/imj.12975
Abstract
Previously, management of hypertension has concentrated on lowering elevated blood
pressure. However, the target has shifted to reducing absolute cardiovascular (CV) risk.
It is estimated that two in three Australian adults have three or more CV risk factors at
the same time. Moderate reductions in several risk factors can, therefore, be more effective than major reductions in one. When managing hypertension, therapy should be
focused on medications with the strongest evidence for CV event reduction, substituting
alternatives only when a primary choice is not appropriate. Hypertension management
guidelines categorise angiotensin-converting enzyme inhibitors (ACEI) and angiotensin
receptor blockers (ARB) interchangeably as first-line treatments in uncomplicated
hypertension. These medications have different mechanisms of action and quite different evidence bases. They are not interchangeable and their prescription should be based
on clinical evidence. Despite this, currently ARB prescriptions are increasing at a higher
rate than those for ACEI and other antihypertensive classes. Evidence that ACEI therapy prevents CV events and death, in patients with coronary artery disease or multiple
CV risk factors, emerged from the European trial on reduction of cardiac events with
perindopril in stable coronary artery disease (EUROPA) and Heart Outcomes Prevention Evaluation (HOPE) trials respectively. The consistent benefit has been demonstrated in meta-analyses. The clinical trial data for ARB are less consistent, particularly
regarding CV outcomes and mortality benefit. The evidence supports the use of ACEI
(Class 1a) compared with ARB despite current prescribing trends.
Introduction
In the past, patients with hypertension have been managed by concentrating on reducing elevated blood pressure (BP). However, therapy is now targetted at
reducing absolute cardiovascular (CV) risk through
medications with proven reductions in fatal and nonfatal events. Current hypertension management guidelines recommend angiotensin-converting enzyme inhibitors (ACEI) and angiotensin receptor blockers (ARB) as
interchangeable first-line pharmacological treatments in
Funding: None.
Conflict of interest: A. Sindone has received honoraria, speaker fees, consultancy fees, is a member of advisory boards or has appeared
on expert panels for: Abbott, Alphapharm, Aspen, Astra Zeneca, Bayer, Biotronik, Boehringer Ingleheim, Bristol-Myers Squibb, Cube,
CSL, General Electric, Glaxo Smith Kline, Guidant, Janssen-Cilag, Johnson and Johnson, Medimark, Medtronic, Menarini, Merck
Sharp and Dohme, Mylan, Novartis, Ogilvy, Pfizer, Phillips, Roche, Sanofi, Servier, St Jude, Sunshine Heart and Vifor. J. Erlich has
received speaker fees, travel support or has been a member of an advisory board for: Amgen, Sanofi-Aventis, Servier and Solvay. C.
Lee has received an honorarium and sponsorship to cardiology meetings from Servier. H. Newman has received honoraria, speaker
fees and travelling grants from: Astra Zeneca, Bristol Myer Squibb, Novartis, Pfizer, Roche, Sanofi, Servier and Solvay. M. Suranyi has
given talks and presentations on behalf of, received honoraria and travel grants from and participated in financially sponsored
multicentre clinical research funded by: Astra Zeneca, Boehringer Ingelheim, Pfizer, Servier and other pharmaceutical companies
operating in Australia. S. D. Roger has received honoraria, speaker fees and consultancy fees, is a member of advisory boards or has
appeared on expert panels for: Abbott, Amgen, Aspen, Astra Zeneca, Baxter, Bayer, Fresenius Medical, Hoffman-La Roche, JanssenCilag, Medimark, Merck Sharp and Dohme, Novartis, Sandoz, Sanofi-Aventis, Servier, Solvay, Takeda and Vifor.
364
© 2016 Royal Australasian College of Physicians
CV risk reduction, ACEI and ARB
uncomplicated hypertension (alongside calcium channel
blockers (CCB)).1 Hypertension is the only parameter
contributing to CV risk and should be considered as the
‘gateway to a cascade of CV risk’. When assessing
patients with hypertension, absolute CV risk should be
calculated using a risk calculator, as CV risk is often
underestimated and patients with hypertension, or other
CV risk factors, may miss the opportunity to have their
CV risk profile measured and appropriately managed.2
Management of hypertension focusses on improving
CV morbidity and mortality, based on clinical trial data
and meta-analyses. Therapy should concentrate on
agents with the strongest evidence for CV event reduction, substituting alternatives only when a primary
choice is not appropriate. Despite this, ARB prescriptions
are increasing at a higher rate than those for ACEI and
other antihypertensive therapies.3 The evidence base for
ACEI and ARB is not the same, and difficulties may arise
in applying this evidence to managing CV risk.
This article addresses the shift in antihypertensive
management towards CV risk reduction by focusing on
guidelines and their shortcomings, and the impact of
multiple risk factor interventions to reduce CV risk.
Within this context, the role of ACEI and ARB is
explored, particularly their mechanisms of action and
data supporting their use.
Hypertension guidelines: where do
you look?
Clinicians have a plethora of guidelines to assist decisions
about appropriate healthcare. The range of guidelines
makes it difficult for clinicians to know which are the
most appropriate for their patients. Whether regional,
national or international, hypertension guidelines differ
in recommendations for target BP, treatment options
and recommendations for special groups, for example,
the elderly, patients with diabetes or chronic kidney disease (CKD).
Guidelines rely on robust search strategies and methodologies. Disappointingly, some guidelines have fallen
short of these standards and have been criticised.4 A systematic review of guidelines on the diagnosis, assessment
and management of hypertension assessed the quality
and consistency of recommendations of 11 guidelines.4 It
highlighted methodological gaps, including clarifying the
scope and purpose, ensuring representation of all stakeholders, including consumers, scientific rigour, supporting implementation and declaration of editorial
independence. Across the 11 guidelines reviewed, the
highest scores were for clarity of presentation and the
lowest for rigour of development.
Several guidelines group ACEI/ARB agents inappropriately and interchangeably (e.g. National Heart Foundation
of Australia: Guide to Management of Hypertension).1
Extrapolation from one drug class to another may not be
based on clinical trial data. In addition, primary prevention hypertension guidelines and secondary cardiovascular disease (CVD) prevention guidelines differ in their
recommendations. For example, while National Heart
Foundation guides to the management of hypertension
and diabetes, management in general practice guidelines
for type 2 diabetes does not prefer ACEI or ARB choice.1
However, the American secondary CVD prevention
guidelines recommend ACEI as first-line (with ARB utilised for ACEI intolerant patients)5,6 (Table 1).
BP targets worldwide may change following the early
termination of the systolic blood pressure intervention
trial (SPRINT),7 which determined a BP treatment regimen targeting a systolic BP target of 120 mm Hg,
reduced the risk of CV events by 30% and all-cause
Table 1 Summary of key international hypertension/cardiovascular risk reduction guidelines detailing recommendations for treatment and blood
pressure goals in various subgroups
BP in mmHg
‘Hypertension’ threshold
BP targets
Drug treatment
threshold in low risk
patients
BP targets (elderly)
BP targets (diabetics)
NICE
ESH/ESC
AHA/ACC/CDC
JNC 8 and ASH/ISH
140/90 and ABPM 135/85
<140/90
160/100 or ABPM 150/95
140/90
<140/90
140/90
140/90
<140/90
140/90
N/A
<140/90
140/90
<150/90
>80 years
N/A
Systolic BP 140–150
>80 years
<140/85
N/A
<150/90
>60 years
<140/90
<140/90
Lower targets may be considered
ABPM, ambulatory blood pressure monitor; AHA/ACC/CDC, American Heart Association/American College of Cardiology/Centers for Disease Control
and Prevention; ASH, American Society of Hypertension; BP, blood pressure; ESC, European Society of Cardiology; ESH, European Society of Hypertension; ISH, International Society of Hypertension; JNC8, Eighth Joint National Committee; N/A, not applicable; NICE, National Institute for Clinical Excellence (Adapted from Lindholm and Carlberg,6 with permission).
© 2016 Royal Australasian College of Physicians
365
Sindone et al.
100
Proportion of adults (%)
90
80
70
60
50
40
30
20
10
0
Hypertension
High blood
cholesterol
Inactive or
insufficiently
active
Overweight or
obese
Daily smoking Inadequate fruit
and vegetable
consumption
Behavioural and biomedical risk factors
Proportion of adults (%)
35
30
25
20
15
10
5
0
1
2
3
4
5 or 6
Number of risk factors
Figure 2 Several cardiovascular disease risk factors in Australian
adults (including hypertension, high blood cholesterol, inactive or insufficiently active, overweight or obese, daily smoking, inadequate fruit
and vegetable consumption. Source: Australian Institute of Health and
Welfare, 2015).
mortality by nearly 25% when compared with patients
treated to a target of 140 mm Hg in high-risk hypertensive adults ≥50 years of age.7
CV mortality in Australia
Over the past few decades, Australia has achieved major
gains in CV survival, due to improvements in prevention, detection, management and lifestyle modification,
such as dietary intake, physical activity and smoking cessation.8 Death rates have fallen from the late 1960s,
when CVD was responsible for approximately 55% of
annual mortality. Age-adjusted CV mortality fell by 76%
from 1968 to 2007, a potential reduction of 140 000
lives if CV mortality was at 1968 levels.8
Multiple CV risk factor intervention
In Australia, hypertension is one of the most common
risk factors in people with CVD (Fig. 1).9 Two in three
Australian adults have three or more CV risk factors
366
Figure 1 Risk factor profile of Australian adults
with cardiovascular disease (based on
Australian Institute of Health and Welfare
material, 2015).
(Fig. 2).9 Assessment of CV risk leads to a more effective
intervention than assessment based on single risk factors
(Fig. 3).2,4,8,9 Moderate reductions in several risk factors
can, therefore, be more effective than major reductions
in one. This is despite the fact that BP treatment recommendations have traditionally been based on BP targets
rather than adjusting for associated CV risks.1,4 This is in
contrast to more recent lipid lowering recommendations.
The Blood Pressure Lowering Treatment Trialists’ Collaboration has now demonstrated progressively greater
absolute risk reductions for BP treatment, as baseline risk
increases.11
Risk for CV events (including CV death) remains,
despite treatment of hypertension.12 Studies have evaluated the efficacy of statins and their potential limitations
in reducing risk of coronary disease.13 They have
demonstrated the effectiveness of statins in reducing the
combined primary end-point of coronary heart death,
non-fatal myocardial infarction (MI) or cerebrovascular
event. However, they have shown that it is insufficient
to treat individual CV risk factors, because although this
may diminish risk somewhat, risk for further events
remains very high.13
Statin use in patients with ≥15% risk of developing
CVD in 10 years, regardless of lipid concentrations or
BP, would avoid more CV deaths compared with a population health strategy.14 Under a population health strategy, reducing the total cholesterol concentration of each
person in the population by 2% would lead to 42 fewer
deaths per 100 000 people treated for 10 years. Treating
only patients with cholesterol >6.2 mmol/L would prevent 125 deaths per 100 000 people treated for 10 years.
Current guidelines recommend treating patients with
increased CV risk at lower cholesterol thresholds.1,8,15
The most effective strategy is to treat those with 15% or
greater risk of developing CVD in 10 years, which would
prevent 290 deaths per 100 000 people treated for
10 years.13
© 2016 Royal Australasian College of Physicians
CV risk reduction, ACEI and ARB
Figure 3 Reduction in CVD with multiple cardiovascular risk management (adapted from
Emberson et al.10; data reproduced by permission of Oxford University Press/on behalf of
European Society of Cardiology and by Jonathan Emberson). CVD, cardiovascular disease.
Achieved BP treatment targets and the
J-curve
Observational studies have suggested a log-linear relationship between BP and risk of stroke, coronary heart disease
and renal injury, through the normal range to the lowest
levels of BP.16 The Blood Pressure Lowering Treatment
Trialists’ Collaboration demonstrated progressively greater
absolute risk reductions for BP treatment, as baseline risk
increases.10 There was no observed BP below which continuous improvements in BP were not associated with risk
reduction.17 This led to the hypothesis that control of
hypertension with antihypertensive drugs should seek to
achieve BP in the low-normal range.16
The J-curve phenomenon was encountered in interventional studies designed to achieve tight BP control
and describes the increased risk with BP reduction below
the given threshold,18 but remains controversial. The
J-curve was most apparent in studies relating achieved
diastolic BP to outcomes, especially in patients with
known heart disease. The effect may be mediated by
impaired diastolic coronary blood flow in patients with
coronary artery disease (CAD) and left ventricular
hypertrophy.18
Patients with significant postural BP drops, the elderly
and those with low diastolic BP even prior to treatment
may be at particular risk with further lowering of BP.17
The concern over the impact of the J-curve effect has led
to less stringent BP guideline targets in the elderly and
those with diabetes (e.g. in the European Society of
Hypertension/European Society of Cardiology Guidelines).14 However, the SPRINT study results suggest that
tight BP control can have significant benefits without a
J-curve risk, in selected patients without diabetes, prior
stroke or polycystic kidney disease.7
Mechanism of action
The mechanism of action of ACEI and ARB is not the
same. Figure 4 depicts the classical description of the
renin–angiotensin–aldosterone system (RAAS).
ACEI reduce angiotensin-2 (AT-2) synthesis by inhibiting ACE action. ACE has two enzymatic sites, one at
© 2016 Royal Australasian College of Physicians
the N-terminus and one at the C-terminus. The Cterminus is the domain responsible for the reduction in
AT-2-mediated hypertension, and the N-terminus for
the accumulation of bradykinin.19 ACEI block both functions, and consequently increase other bioactive peptides, notably bradykinin and angiotensin 1-7.20
Angiotensin 1-7 acting through the MAS receptor is
important in the kidney, heart and muscle where it has
anti-proliferative,21 anti-thrombotic22 and hypotensive
effects by inducing nitric oxide.23
ARB bind directly to the angiotensin-1 (AT-1) receptor, preventing its activation by AT-2. This results in a
relative increase in circulating AT-2 which may stimulate
ACE expression and inhibit the expression of ACE2.24
ACE2 is different from ACE and is not affected by ACEI.
ACE2 is important under stress and converts AT-2 to
angiotensin 1-7. Circulating AT-2 may then bind to
other angiotensin receptors, particularly the AT-2 receptor. Several effects of AT-1 blockade could be due to the
effects of increased availability of AT-2 to bind to the AT2 receptor.
A further potential mechanism for the pleiotropic
effects of ACEI is that ACEI (but not AT-1 blockers) regulate bradykinin BR2 receptor signalling which has
effects on innate and adaptive immunity25 and renal
tubular salt handling.26
While there are similarities, there are many differences in how ACEI and ARB produce their therapeutic
effects, which may account for the differences in clinical
outcomes.
Effect of RAAS agents on glomerular
filtration rate
The effect of RAAS agents on glomerular filtration rate
(GFR) is context dependent. ACEI therapy is associated
with a reduction in mean arterial pressure and peripheral vascular resistance, including renal vascular resistance. In most patients, there is an increase in renal blood
flow. Despite a reduction in efferent glomerular vascular
resistance and reduced intra-glomerular pressure, the
net effect on GFR is minimal.26
367
Sindone et al.
Figure 4 The renin–angiotensin–aldosterone system. ACE, angiotensin converting enzyme; ACE2, angiotensin converting enzyme 2; ACEI,
angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; AT-1, angiotensin-1; AT-2, angiotensin-2; AT-3, angiotensin-3; AT-4,
angiotensin-4.
While ACEI improve renal blood flow and sodium
excretion rates in chronic heart failure and reduce the
rate of progressive renal injury in CKD, their use can be
associated with a syndrome of ‘functional renal insufficiency’ and/or hyperkalaemia.26 ACEI and ARB therapy
may contribute to acute kidney injury (AKI) and
reduced GFR in patients with volume insufficiency.
In hypertensive patients with normal renal function
(i.e. estimated GFR >90 mL/min/1.73 m2), there is little
change. In patients with vascular disease, particularly
renal artery stenosis, there may be a greater risk of deterioration in renal function. Patients who take nonsteroidal anti-inflammatory agents with ACEI or ARB,
and have pre-existing renal insufficiency, are particularly
at risk of renal impairment.27
A meta-analysis of patients taking ACEI and ARB prior
to coronary angiography found an increased risk for AKI
in patients taking ARB (odds ratio (OR) = 3.31; 95%
confidence interval (CI) = 1.89–5.78; P < 0.0005), but
not in patients taking ACEI. This may represent differential effects of AT-2 on various AT receptors compared
with the inhibition of AT-2 production.28
368
The combined arm of the ONTARGET study using
ACEI and ARB therapy on CV outcomes and BP did not
lead to a further reduction in MI or other CV events
when compared with either agent alone, with some
increase in hypotension.29 ACEI and ARB are, therefore,
not recommended in combination to reduce CV risk.
A meta-analysis of patients receiving combined ACEI/
ARB therapy for CKD showed that there was some benefit in reducing albuminuria, but at the expense of a
small decrease in GFR.30
Evidence for ACEI use
Evidence that ACEI prevent CV events and death in
patients with multiple CV risk factors31 or CAD32
emerged with the Heart Outcomes Prevention Evaluation (HOPE)31 and European trial on reduction of cardiac
events with perindopril in stable coronary artery disease
(EUROPA)32 trials. The consistent benefit has been
demonstrated in meta-analyses.33–35
In patients with high CV risk, a meta-analysis of
26 randomised controlled trials involving 108 112
© 2016 Royal Australasian College of Physicians
CV risk reduction, ACEI and ARB
patients without heart failure reported that ACEI significantly reduced the composite outcome of CV death, MI
and stroke by 17% (P = 0.001), MI by 19% (P < 0.001),
stroke by 20% (P < 0.004), all-cause death by 9%
(P = 0.008), new-onset heart failure by 21% (P = 0.001)
and new-onset diabetes by 14% (P < 0.001).33
van Vark et al. completed a pooled analysis of 20 CV
morbidity–mortality trials using RAAS inhibitors for the
treatment of hypertension.34 They demonstrated that
RAAS inhibitors were associated with a reduction in allcause mortality of 5% (P = 0.032).35 In a post hoc analysis of ACEI and ARB separately, they found that the
observed benefit was largely attributable to the beneficial
effects of ACEI (10% relative risk reduction in all-cause
mortality, P = 0.004). There was a significant statistical
heterogeneity between ACEI and ARB effects on mortality (but not CV mortality).34 A South African study evaluating cost and outcomes found equal input costs for
ACEI and ARB, but significant reduction in downstream
events and costs.35
A meta-analysis of ACEI in patients with diabetes analysed 23 trials in 32 827 patients. ACEI significantly
reduced all-cause mortality by 13% (P = 0.02), CV death
by 17% (P = 0.04), MI by 21% (P = 0.01) and heart failure by 19% (P = 0.002).36
Evidence for ARB use
ARB have less compelling evidence pertaining to CV
morbidity and mortality. A meta-analysis of 37 randomised controlled trials, including 147 020 patients
showed that ARB reduce the relative risk of stroke by
10% (P = 0.007), heart failure by 13% (P < 0.001) and
new-onset diabetes by 15% (P = 0.003) compared with
placebo or an active comparator. Importantly, however,
ARB did not reduce the risk of MI (1% reduction,
P = 0.055), CV death (1% reduction, P = 0.090) or death
from any cause (no reduction, P = 0.482).37 This metaanalysis included the ONTARGET trial that established
that telmisartan (an ARB) was not inferior to ramipril
(an ACEI) in the reduction of major CV events and CV
mortality.29
The meta-analysis by van Vark et al. showed that ARB
did not reduce overall or CV mortality in patients with
hypertension. The paper was limited by the fact that it
was an indirect comparison between studies and possible
confounding attributable to additional parameters,
including adjuvant treatment. Inter-trial comparisons are
prone to epidemiological fallacies. Head-to-head comparisons are preferred. However, this meta-analysis
directly compares these two classes in a predominantly
hypertensive cohort.34
© 2016 Royal Australasian College of Physicians
Direct comparison
The ONTARGET trial directly compared ACEI with ARB
and showed that telmisartan (an ARB) was not inferior
to ramipril (an ACEI) in the risk of major CV events,
including CV death. Patients randomised to the ARB had
less cough.29
In patients with hypertension, a meta-analysis of
head-to-head trials found nine studies of more than
1 year in duration and demonstrated no difference
between ACEI and ARB for total mortality (RR = 0.98),
CV events (RR = 1.07) or CV mortality (RR = 0.98). This
meta-analysis highlighted that ACEI have shown positive improvement in these outcomes over placebo, but
these findings cannot be extrapolated to ARB.38
Two similarly designed studies have been performed
in patients with multiple CV risk factors. The HOPE
study31 evaluated the effect of ACEI and found a significant reduction in CV events. However, the TRANSCEND
study39 assessed the effect of ARB in high-risk patients
intolerant of ACEI and showed no significant reduction
in CV events. The population demographics and concomitant medications varied between the trials.
A Cochrane review of ACEI versus ARB for hypertension noted substitution of an ARB for an ACEI, while
supported by evidence on grounds of tolerability, must
be made in consideration of the weaker evidence for the
efficacy of ARB regarding mortality and morbidity compared with ACEI.38 However, in patients with renal
impairment, diabetes and established vascular risk RAAS
system intervention should be first-line38
Tolerability
The most cited reason for discontinuation of ACEI therapy is cough. A meta-analysis of 27 492 ACEI-naïve
patients with CVD treated with the ACEI perindopril or
placebo found a 3.9% incidence of cough.40 It found no
role for race in determining the incidence of cough for
people of white people or Asian descent. The predictors
of cough were female gender, older age and concomitant
use of lipid lowering agents. This study cast doubt on the
role of ethnicity in the genesis of cough, and further suggested that the switch from ACEI to ARB therapy should
be no more than 4%.40 This is particularly so as ACEI
have Class 1A evidence supporting their use.1
Management of resistant
hypertension
Studies estimate that 10% of hypertensive patients have
resistant hypertension.15 Primary hyperaldosteronism
may occur in a significant proportion of this group.1
369
Sindone et al.
Spironolactone, an aldosterone antagonist, is considered
a fifth-line agent for the treatment of hypertensive
patients resistant to a combination of ACEI/ARB, CCB, a
thiazide diuretic or a beta-blocker. Spironolactone may
potentially benefit hypertensive patients with heart failure.1 Eplerenone, an aldosterone antagonist with less
anti-progesterone and anti-androgenic effects, is effective in hypertension and post-MI heart failure.1
Australian guidelines recommend several other
fifth-line agents, including beta-blockers, clonidine,
hydralazine, methyldopa, moxonidine and prazosin.1
Methyldopa, oxprenolol and labetalol may be used in
pregnancy, whereas diuretics, ACEI and ARB are
contraindicated.1
Timing of antihypertensive therapy
Several studies have produced findings in favour of nocturnal dosing of antihypertensive medications. In
patients with CKD whose BP did not fall during the night
on ambulatory BP monitoring (non-dippers), changing
antihypertensive therapy to the night time decreased
nocturnal BP and decreased proteinuria.41
In two studies by Hermida et al., patients with hypertension were randomised to receive all of their antihypertensive medications in the morning or at least one at
night. In diabetic hypertensive patients, taking at least
one antihypertensive medication at night reduced CV
risk.42 In patients with hypertension and CKD, those
who took at least one BP-lowering medication at night
had an adjusted risk of CV events of less than one third
of those who took all medications in the morning
(adjusted hazard ratio (HR) = 0.31; 95% CI = 0.21–0.46;
P < 0.001). Taking medications at night demonstrated a
similar reduction in a composite outcome of CV death,
MI and stroke (adjusted HR = 0.28; 95% CI = 0.13–0.61;
P < 0.001).43
These results may change timing of antihypertensive
therapy, particularly as the majority of CV events occur
in the morning hours41–43, when night time medications
exert their maximal effect, whereas morning medications may be losing efficacy.
Medication adherence and compliance
There has been a shift from the term ‘compliance’
to ‘adherence’, as the latter reflects the onus placed on
the patient to take prescribed medication and implies the
sharing of responsibility between doctor and patient.44
The primary aim of antihypertensive medication is to
reduce CV events, but non-adherence to medications
may adversely affect the ability to control BP. In hypertensive patients, a meta-analysis demonstrated that the
370
OR of achieving BP with adherence to dosing regimens
versus non-adherence was 3.44 (95% CI = 1.60–7.37).45
Non-adherence may be a major contributor to ‘refractory hypertension’ in a subgroup of hypertensive
patients.44 A middle group, referred to as the ‘whitecoat’ compliers, take their medications prior to the doctor’s appointment, and others adhere on a ‘prn’ basis,
with either wrong dosage or frequency.45
Most hypertensive patients require at least two antihypertensive medications and some guidelines recommend
initiation with dual therapy.11 Single tablet combinations, as opposed to prescription of two medications
given separately (free drug components), were compared in a meta-analysis which found that single tablet
combinations are associated with greater compliance
(OR = 1.21; 95% CI = 1.03–1.43; P = 0.02).46 A reduction in systolic/diastolic BP of 4.1/3.1 mm Hg was seen;
however, the change was not statistically significant.
Conclusion
Reducing CV morbidity and mortality is the focus of
treating patients with hypertension. The means by which
this is achieved is important. It is not simply about reaching a target BP, but about using medications which have
been demonstrated to reduce CV events. The evidence
supports the use of ACEI (Class 1A) in preference to
ARB despite current prescribing trends. These medications have different mechanisms of action and quite
different evidence bases. They are not interchangeable
and their prescription should be based on clinical
evidence.
Doctors are able to prescribe a range of medications
which are able to lower BP. The key is to choose a management strategy which encompasses multiple risk factor
reduction, is well tolerated, effective, economic and one
which is likely to maximise patients’ adherence. This
review has discussed the guidelines and evidence, as well
as challenged current treatment paradigms in light of the
available data to try and achieve maximal benefit for
patients with increased CV risk.
Management of patients to reduce CV risk
should include the evaluation of compliance with medications, whether medications are taken in the morning
or at night and attempts to address modifiable risk
factors, including hypertension, hyperlipidaemia, cigarette smoking, dietary intake and physical activity.
This will enhance management and aims to reduce risk
further.
The goal of CV therapy in hypertension management
should focus on reducing CV events using medications
with proven benefit. ACEI appear to have superior evidence over ARB and should be preferred to ARB.
© 2016 Royal Australasian College of Physicians
CV risk reduction, ACEI and ARB
Prescribers should only cease or change from an ACEI
for a clearly confirmed side effect. Reducing CV morbidity and mortality in hypertension is not just about
achieving a target BP. The class of medication used can
References
1 National Heart Foundation of Australia.
Guide to Management of Hypertension.
2008. [updated 2010 Dec; cited 2015
Oct 26]. Available from URL: http://
www.heartfoundation.org.au/
SiteCollectionDocuments/Hypertension
Guidelines2008to2010Update.pdf
2 National Vascular Disease Prevention
Alliance. Absolute Cardiovascular
Disease Risk Assessment. Quick
Reference Guide for Health
Professionals, 2009 [cited 2015 Oct 26].
Available from URL: http://www.
cvdcheck.org.au
3 Medicare Australia. Pharmaceutical
Benefits Scheme Item Reports [cited
2015 Oct 26]. Available from URL:
http://medicarestatistics.humanservices.
gov.au/statistics/pbs_item.jsp
4 Al-Ansary LA, Tricco AC, Adi Y,
Bawazeer G, Perrier L, Al-Ghonaim M
et al. A systematic review of recent
clinical practice guidelines on the
diagnosis, assessment and management
of hypertension. PLoS One 2013; 8:
e53744.
5 Smith SC, Benjamin EJ, Bonow RO,
Braun LT, Creager MA, Franklin BA
et al. AHA/ACCF secondary prevention
and risk reduction therapy for patients
with coronary and other atherosclerotic
vascular disease: 2011 update.
A guideline from the American Heart
Association and American College of
Cardiology Foundation. Circulation
2011; 124: 2458–73.
6 Lindholm LH, Carlberg B. Guidelines for
hypertension 2011–14. Hypertension
News January 2014; Opus 35: p. 3.
7 National Institutes of Health. Landmark
NIH Study Shows Intensive Blood
Pressure Management May Save Lives.
Press release 2015 Sep 11 [cited 2015
Sep 28]. Available from URL: http://
www.nih.gov/news/health/sep2015/
nhlbi-11.htm
8 Australian Institute of Health and
Welfare. Impact of falling cardiovascular
disease death rates: deaths delayed and years
of life extended. Bulletin no. 70. Cat.
no. AUS 113. Canberra: AIHW; 2009.
9 Australian Institute of Health and
Welfare. Cardiovascular disease, diabetes
© 2016 Royal Australasian College of Physicians
10
11
12
13
14
15
16
17
affect outcomes, even though the same BP has been
achieved. It is not just about the destination, but also
about the journey.
and chronic kidney disease – Australian
facts: risk factors. Cardiovascular, diabetes
and chronic kidney disease series no. 4.
Canberra: AIHW; 2015.
Emberson J, Whincup P, Morris R,
Walker M, Ebrahim S. Evaluating the
impact of population and high-risk
strategies for the primary prevention of
cardiovascular disease. Eur Heart J 2004;
25: 484–91.
Blood Pressure Lowering Treatment
Trialists’ Collaboration, Sundstrom J,
Arima H, Woodward M, Jackson R,
Karmali K et al. Blood pressurelowering treatment based on
cardiovascular risk: a meta-analysis of
individual patient data. Lancet 2014;
384: 591–8.
PROGRESS Collaborative Group.
Randomised trial of perindopril based
blood pressure lowering regimen
among 6105 individuals with previous
stroke or transient ischaemic attack.
Lancet 2001; 358: 1033–41.
Kastelein JJP. The realities of
dyslipidaemia: what do the studies tell
us? Eur Heart J 2005; 7: F27–33.
Manuel DG, Lim J, Tanuseputro P,
Anderson GM, Alter DA, Laupacis A
et al. Revisiting rose: strategies for
reducing cardiovascular disease. Br Med
J 2006; 332: 659–2.
Mancia G, Fagard R, Narkiewicz K,
Redon J, Zanchetti A, Bohm M et al.
The task force for the management of
arterial hypertension of the European
Society of Hypertension (ESH) and of
the European Society of Cardiology
(ESC). 2013 ESH/ESC guidelines for the
management of arterial hypertension.
J Hypertens 2013; 31: 1281–357.
MacMahon S, Peto R, Cutler J,
Collins R, Sorlie P, Neaton J et al.
Epidemiology: blood pressure, stroke
and coronary heart disease. Part
1. Prolonged differences in blood
pressure: prospective observational
studies corrected for the regression
dilution bias. Lancet 1990; 335: 765–4.
Collins R, Peto R, MacMahon S,
Hebert P, Fiebach NH, Eberlein KA et al.
Blood pressure, stroke and coronary
heart disease. Part 2. Short term
reductions in blood pressure: overview
of randomised drug trials in their
18
19
20
21
22
23
24
25
26
epidemiological context. Lancet 1990;
335: 827–35.
Tsika EP, Poulimenos LE, Boudoulas
KD, Manolis AJ. The J-curve in arterial
hypertension: fact or fallacy? Cardiology
2014; 129: 126–35.
Burger D, Reudelhuber TL, Mahajan A,
Chibale K, Sturrock ED, Touyz RM.
Effects of a domain-selective ACE
inhibitor in a mouse model of chronic
angiotensin II-dependent hypertension.
Clin Sci (Lond) 2014; 127: 57–63.
Chappell MC. Emerging evidence for a
functional angiotensin-converting
enzyme 2-angiotensin-(1-7)-MAS
receptor axis: more than regulation of
blood pressure? Hypertension 2007; 50:
596–9.
Tallant EA, Ferrario CM, Gallagher PE.
Angiotensin-(1-7) inhibits growth of
cardiac myocytes through activation of
the MAS receptor. Am J Physiol Heart
Circ Physiol 2005; 289: H1560–6.
Fraga-Silva RA, Da Silva DG,
Montecucco F, Mach F, Stergiopulos N,
da Silva RF et al. The angiotensinconverting enzyme
2/angiotensin-(1-7)/MAS receptor axis:
a potential target for treating
thrombotic diseases. Thromb Haemost
2012; 108: 1089–96.
Heitsch H, Brovkovych S, Malinski T,
Wiemer G. Angiotensin-(1-7)stimulated nitric oxide and superoxide
release from endothelial cells.
Hypertension 2001; 37: 72–6.
Koka V, Huang XR, Chung AC,
Wang W, Truong LD, Lan HY.
Angiotensin II up-regulates angiotensin
I-converting enzyme (ACE), but downregulates ACE2 via the AT1-ERK/p38
MAP kinase pathway. Am J Pathol 2008;
172: 1174–83.
Schmitz V, Almeida LN, Svensjö E,
Monteiro AC, Köhl J, Scharfstein J. C5a
and bradykinin receptor cross-talk
regulates innate and adaptive immunity
in Trypanosoma cruzi infection.
J Immunol 2014; 193: 3613–23.
Katori M, Majima M. Renal (tissue)
kallikrein-kinin system in the kidney
and novel potential drugs for saltsensitive hypertension. Prog Drug Res
2014; 69: 59–109.
371
Sindone et al.
27 Schoolwerth AC, Sica DA,
Ballermann BJ, Wilcox CS, Council on
the Kidney in Cardiovascular Disease
and the Council for High Blood Pressure
Research of the American Heart
Association. Renal considerations in
angiotensin converting enzyme
inhibitor therapy: a statement for
healthcare professionals from the
Council on the Kidney in
Cardiovascular Disease and the Council
for High Blood Pressure Research of the
American Heart Association. Circulation
2001; 104: 1985–91.
28 Peng F, Su J, Lin J, Niu W. Impact of
renin-angiotensin-aldosterone systemblocking agents on the risk of contrastinduced acute kidney injury: a
prospective study and meta-analysis.
J Cardiovasc Pharmacol 2015; 65: 262–8.
29 ONTARGET Investigators, Yusuf S,
Teo KK, Pogue J, Dyal L, Copland I et al.
Telmisartan, ramipril, or both in
patients at high risk for vascular events.
N Engl J Med 2008; 358: 1547–59.
30 Susantitaphong P, Sewaralthahab K,
Balk EM, Eiam-ong S, Madias NE,
Jaber BL. Efficacy and safety of
combined vs. single renin-angiotensinaldosterone system blockade in chronic
kidney disease: a meta-analysis.
Am J Hypertens 2013; 26: 424–1.
31 Yusuf S, Sleight P, Pogue J, Bosch J,
Davies R, Dagenais G. Effects of an
angiotensin-converting-enzyme
inhibitor, ramipril, on cardiovascular
events in high-risk patients. The heart
outcomes prevention evaluation study
investigators. N Engl J Med 2000; 342:
145–53.
32 Fox KM. For the EUROPA investigators.
Efficacy of perindopril in reduction of
cardiovascular events among patients
with stable coronary artery disease:
randomised, double-blind, placebocontrolled, multicentre trial (the
EUROPA study). Lancet 2003; 362:
782–8.
372
33 Savarese G, Costanzo P, Cleland JG,
Vassallo E, Ruggiero D, Rosano G et al.
A meta-analysis reporting effects of
angiotensin-converting enzyme
inhibitors and angiotensin receptor
blockers in patients without heart
failure. J Am Coll Cardiol 2013; 61:
131–42.
34 van Vark LC, Bertrand M,
Akkerhuis KM, Brugts JJ, Fox K,
Mourad JJ et al. Angiotensinconverting enzyme inhibitors reduce
mortality in hypertension: a metaanalysis of randomized clinical trials of
renin-angiotensin-aldosterone system
inhibitors involving 158,998 patients.
Eur Heart J 2012; 33: 2088–97.
35 Makkink JL, Greeff OB. Angiotensin
converting enzyme inhibitors
v. angiotensin receptor blockers in the
management of hypertension: a
funder’s perspective. S Afr Med J 2014;
104: 292–4.
36 Cheng J, Zhang W, Zhang X, Han F,
Li X, He X et al. Effect of angiotensinconverting enzyme inhibitors and
angiotensin II receptor blockers on allcause mortality, cardiovascular
deaths, and cardiovascular events in
patients with diabetes mellitus: a metaanalysis. JAMA Intern Med 2014; 174:
773–85.
37 Bangalore S, Kumar S, Wetterslev J,
Messerli FH. Angiotensin receptor
blockers and risk of myocardial
infarction: meta-analyses and trial
sequential analyses of 147 020 patients
from randomised trials. BMJ 2011; 342:
d2234. doi: 10.1136/bmj.d2234.
38 Li EC, Heran BS, Wright JM.
Angiotensin converting enzyme (ACE)
inhibitors versus angiotensin receptor
blockers for primary hypertension.
Cochrane Database Syst Rev 2014; 8:
CD009096.
39 TRANSCEND Investigators, Yusuf S,
Teo K, Anderson C, Pogue J, Dyal L
et al. Effects of the angiotensin-receptor
40
41
42
43
44
45
46
blocker telmisartan on cardiovascular
events in high-risk patients intolerant to
angiotensin-converting enzyme
inhibitors (TRANSCEND): a randomised
controlled trial. Lancet 2008; 372:
1174–83.
Brugts JJ, Arima H, Remme W,
Bertrand M, Ferrari R, Fox K et al. The
incidence and clinical predictors of
ACE-inhibitor induced dry cough by
perindopril in 27,492 patients with
vascular disease. Int J Cardiol 2014; 176:
718–23.
Minutolo R, Gabbai FB, Borrelli S,
Scigliano R, Trucillo P, Baldanza D
et al. Changing the timing of
antihypertensive therapy to reduce
nocturnal blood pressure in CKD: an
8-week uncontrolled trial. Am J Kidney
Dis 2007; 50: 908–17.
Hermida RC, Ayala DE, Mojon A,
Fernandez JR. Influence of time of day
of blood pressure-lowering treatment
on cardiovascular risk in hypertensive
patients with type 2 diabetes. Diabetes
Care 2011; 34: 1270–6.
Hermida RC, Ayala DE, Mojón A,
Fernández JR. Bedtime dosing of
antihypertensive medications reduces
cardiovascular risk in CKD. J Am Soc
Nephrol 2011; 22: 2313–1.
DiMatteo MR, Giordani PJ, Lepper HS,
Croghan TW. Patient adherence and
medical treatment outcomes: a metaanalysis. Med Care 2002; 40:
794–811.
Burnier M, Santschi V, Favrat B,
Brunner HR. Monitoring compliance in
resistant hypertension: an important
step in patient management. J Hypertens
2003; 21: S37–42.
Gupta AK, Arshad S, Poulter NR.
Compliance, safety, and effectiveness of
fixed-dose combinations of
antihypertensive agents – a metaanalysis. Hypertension 2010; 55:
399–407.
© 2016 Royal Australasian College of Physicians
L E T T E R S TO T H E E D I TO R
Clinical-scientific notes
Screening for coeliac disease in
elderly inpatients with minimal trauma
fracture is not indicated
It has been noted recently that in a Liverpool
coeliac cohort,1 26% of tested coeliac patients had osteoporosis through bone mineral density. Due to this, some
authorities recommend screening orthogeriatric populations for coeliac disease.1,2 However, this advice is
at odds both with other literature3 and with our
clinical experience – orthogeriatric patients with clinically defined osteoporosis (minimal trauma fracture) do
not appear to have an increased incidence of coeliac
disease.
Thus, we undertook a retrospective audit of the results
of screening tests for coeliac disease obtained during
732 consecutive acute admissions from 2007 to 2009
with minimal trauma fracture to the orthogeriatric service of the Royal Melbourne Hospital (72.00% female,
mean age of admission = 81.16 years). Tissue transglutaminase (tTG) levels were assessed in 290 (74.05%
female) of the 732 admissions due to the lack of clear
guidelines and differing practice. Some patients did not
have any screening completed during their acute admission. Testing was not based on the presence or absence
of signs, symptoms or other results suggestive of coeliac
disease. Approval was subsequently obtained as a quality
assurance activity from the Melbourne Health Human
Research Ethics Committee. Testing data were obtained
using the INOVA Diagnostics tTG assay (ref = 0–20
units). Three females and one male who presented with
a fracture were tTG abnormal (1.38%). During their
acute admission, none of the four went on to have gastroscopic confirmation of histologically evident coeliac
disease or HLA DQ2/8 testing, and one died as a result of
the fracture. In addition, none of the patients in our
sample was completely IgA deficient (ref <0.05 g/L).
These results could have been strengthened by additional serologic testing, including HLA DQ2/8 or for
DGP-IgG, but this was not a part of routine screening at
the time of admission.
The serologic prevalence of coeliac disease (1.38%) in
our cohort closely approximates the prevalence of coeliac disease in a general Australian population of 1.3%.4
Recent work5 suggests a high rate of gluten or wheat
exclusion in the general Australian population. While it
© 2016 Royal Australasian College of Physicians
is possible that this could lead to an increased rate of
false negatives in this cohort, there are two potential
mitigating factors. The first is the time this data were collected from – the fractures are from prior to 2009, where
gluten avoidance may have been less frequent. Secondly,
patients with a minimal trauma fracture are not representative of a general Australian population, with differences of age, residence, cognitive status and available
nutrition such that gluten avoidance may have been less
frequent.
In a review of the literature, several other authors
have approached the association between coeliac disease
and minimal trauma fractures with similarly low rates of
association. Of particular note, our findings concord with
a case series of 347 older patients with hip fracture
which found no association with coeliac disease,
although these authors used an older screening test
(endomysial antibody) that has lower reported sensitivity than the test that we used (tTG-IgA).3
Given the low rates of coeliac autoantibody positive
testing, there are at least four possibilities: first, that this
test is for a previously unknown reason not useful in the
diagnosis of coeliac disease in the elderly (despite various
studies to the contrary); second, that coeliac disease is
not associated with minimal trauma fracture; third, that
the elderly represent a population with a low rate of coeliac disease compared with younger cohorts and therefore low rates of association (which is again contradicted
by the literature) or finally that there is significant symptomatic and serologic control of coeliac disease through
the voluntary adoption of a gluten-free diet in this
population.
The lack of an increased autoantibody frequency in
our cohort above general population prevalence supports our clinical impression that coeliac disease is not
more common in the elderly suffering minimal trauma
fracture. However, this finding may not necessarily apply
to younger cohorts, where the contribution of reduced
bone mineral density from coeliac disease may be relatively greater in the absence of other risk factors, such
as age.
In summary then, we recommend against routine
serological screening for coeliac disease in elderly inpatients with minimal trauma fracture. This is important
given the lack of utility as described above, and the cost
of screening a minimal trauma population with
373
Letters to the Editor
relatively reduced QALY gains (compared with only
symptomatic at-risk screening).6
D. I. Clayton-Chubb,1,2 M. Wilson,2 E. Seal2
and P. W. Lange2
1
Eastern Health and 2Department of Geriatrics,
Received 10 June 2015; accepted 20 July 2015.
Royal Melbourne Hospital, Melbourne, Victoria, Australia
doi:10.1111/imj.13000
References
1 Collum J, Smith A, Martin K. PWE-112
Prevalence of osteoporosis in a Liverpool
celiac cohort supports routine use of
bone mineral density assessment. Gut
2012; 61: A342.
2 Stenson WF, Newberry R, Lorenz R,
Baldus C, Civitelli R. Increased
prevalence of celiac disease and need for
routine screening among patients with
osteoporosis. Arch Intern Med 2005; 165:
393–9.
3 Fisher AA, Davis MW, Budge MM.
Should we screen adults with
osteoporotic fractures for coeliac disease?
Gut 2004; 53: 154–5.
4 Anderson RP, Henry MJ, Taylor R,
Duncan EL, Danoy P, Costa MJ et al. A
novel serogenetic approach determines
the community prevalence of celiac
disease and informs improved diagnostic
pathways. BMC Med 2013; 11: 1–3.
5 Golley S, Corsini N, Topping D, Morell
M, Mohr P. Motivations for avoiding
Traumatic superior gluteal artery
pseudoaneurysm following a bone
marrow biopsy
A 62-year-old man with recently diagnosed chronic
myeloid leukaemia (CML) presented to the emergency
department 2 h after a diagnostic bone marrow aspirate
and trephine (BMAT) taken from the right posterior iliac
crest. He described progressive pain and swelling in the
right buttock and difficulty walking due to pain. He
denied weakness or paraesthesia in the leg.
The man’s CML was in chronic phase, and he had not
yet been started on treatment. His presenting blood tests
showed haemoglobin of 86 g/L, platelet count of
616 × 109/L and a total white cell count of
376.50 × 109/L. His Sokal relative risk was high (1.63).1
Examination on admission revealed a superficial haematoma over the right buttock with no external bleeding
from the biopsy site. There was no evidence of neurovascular compromise in the lower limbs. The man was initially treated with analgesia and observed; however, on
the second day of his admission, his right buttock and
leg were markedly more swollen. The man complained
of severe pain mobilising and light-headedness on standing. Repeat blood tests showed a decline in the haemoglobin concentration to 60 g/L.
An urgent computed tomographic scan was requested,
which showed a large haematoma in the right buttock
and thigh with active haemorrhage at the level of the
biopsy track in the right iliac crest.
The patient went forward for urgent endovascular
treatment in view of active bleeding. Right internal iliac
374
wheat consumption in Australia: results
from a population survey. Public Health
Nutr 2015; 18: 490–9.
6 Park KT, Tsai R, Wang L, Khavari N,
Bachrach L, Bass D. Cost-effectiveness of
universal serological screening to prevent
non-traumatic hip and vertebral fractures
in patients with celiac disease. Clin
Gastroenterol Hepatol 2013; 11: 645–53.
artery angiography demonstrated a pseudoaneurysm
arising from a branch of the right superior gluteal artery
(Fig. 1A). The branch of the superior gluteal artery supplying the pseudoaneurysm was accessed using a microcatheter, and the pseudoaneurysm was successfully
coiled (Fig. 1B).
Following successful embolisation, the patient
reported a dramatic improvement in his pain. He was
discharged after 8 days in hospital when he was able to
walk independently with the aid of crutches. His mobility was limited by pain only with no neurological complications detected.
Bone marrow biopsies are used extensively for the
investigation of haematological diseases. They are considered a safe procedure with an adverse event rate of
0.08%.2 Complications are mostly limited to discomfort
at the biopsy site, bruising and occasionally local bleeding. While infrequent, haemorrhage is the most common
serious complication, and patients typically present
within hours of the procedure.3 Haemorrhagic complications requiring surgical or endovascular intervention are
rare with only a small number of cases reported in the
literature.
Patients with myeloproliferative neoplasms have the
greatest risk of haemorrhage following BMAT.2 Other
risk factors for bleeding include antiplatelet and anticoagulant therapy, coagulation defects, platelet disorders,
renal impairment, bone disease and obesity.2,3
Pseudoaneurysms arise from a defect in the artery wall
with active bleeding into surrounding haematoma.4
Superior gluteal artery pseudoaneurysms are a rare but
© 2016 Royal Australasian College of Physicians
Letters to the Editor
Figure 1 (A) Right internal iliac artery angiography demonstrates a pseudoaneurysm arising
from a branch of the right superior gluteal
artery. (B) Post embolisation. The vessel supplying the pseudoaneurysm has been successfully coiled.
known complication of bone marrow biopsies taken
from the posterior iliac crest with a typical presentation
of pain and swelling within the gluteal or retroperitoneal
compartments.4,5 Morbidity following a superior gluteal
artery pseudoaneurysm can be significant with prolonged hospital stays and case reports of patients having
sciatic nerve injury.6 Endovascular treatment of traumatic pseudoaneurysms provides a fast, effective and
minimally invasive method for patients who would otherwise require open surgery.3,4
We report a case of a BMAT resulting in severe haemorrhage from a traumatic superior gluteal artery pseudoaneurysm that was successfully treated with an
embolisation procedure. Clinicians need to be aware of
References
1 Sokal J, Cox E, Baccarani M, Tura S,
Gomez G, Robertson J et al. Prognostic
discrimination in ‘good-risk’ chronic
granulocytic leukemia. Blood 1984; 63:
789–99.
2 Bain BJ. Bone marrow biopsy morbidity:
review of 2003. J Clin Pathol 2005; 58:
406–8.
3 Arellano-Rodrigo E, Real M,
Muntanola A, Burrel M, Rozman M,
© 2016 Royal Australasian College of Physicians
and minimise risk factors for bleeding following BMAT.
In addition, knowledge of complications allows prompt
diagnosis and treatment of patients who may present following a BMAT procedure.
Received 3 July 2015; accepted 19 October 2015.
doi:10.1111/imj.13002
I. R. Caldwell,1 B. T. Buckley,1 R. Rajagopal,2
R. Doocey1 and L. Pemberton1
1
Department of Haematology, Auckland City Hospital, and
2
Department of Haematology, Middlemore Hospital,
Fraire G et al. Successful treatment by
selective arterial embolization of severe
retroperitoneal hemorrhage secondary to
bone marrow biopsy in postpolycythemic myelofibrosis. Ann Hematol
2004; 83: 67–70.
4 Taif S, Derweesh A, Talib M. Superior
gluteal artery pseudoaneurysm
presenting as a gluteal mass: case report
and review of literature. J Clin Imaging Sci
2013; 3: 49.
Auckland, New Zealand
5 Chamisa I. Fatal vascular retroperitoneal
injury following bone marrow biopsy:
clinical images: SAMJ forum. S Afr Med J
2007; 97: 246.
6 Ge PS, Ng G, Ishaque BM, Gelabert H, de
Virgilio C. Iatrogenic pseudoaneurysm of
the superior gluteal artery presenting as
pelvic mass with foot drop and sciatica:
case report and review of literature. Vasc
Endovascular Surg 2010; 44: 64–8.
375
Letters to the Editor
General correspondence
Liquid biopsies: advancing cancer
research through drops of blood
Tissue biopsy has been the gold standard for diagnosis of
cancer for years. Advances in the genotyping of tumour
tissue acquired often through invasive procedures have
led to development of several prognostic and predictive
biomarkers to help facilitate complex decision-making in
oncology clinics. However, the information acquired
from a single biopsy only provides a limited snapshot at
a single time point. We know that tumours evolve with
time especially after exposure to various systemic therapies and can develop secondary mutations leading to
drug resistance.1 Another major challenge in the era of
personalised targeted therapy is that of intra-tumour
heterogeneity. Gerlinger et al. took paired samples from
various sites of the primary tumour specimen and its
metastases and found that there was significant intra and
inter-tumoural heterogeneity.2
Longitudinal tumour biopsies seem to be a rational
approach, but are less practical due to several issues,
including discomfort to the patient, risk of complications
from invasive procedures and associated costs. In contrast, liquid biopsies obtained through a blood sample
offer an attractive alternative despite its limitations. Liquid biopsy can potentially provide information regarding
the genetic makeup of both primary tumour and metastases offering the opportunity to track systematically
genomic evolution of the tumour. Analysis of circulating
tumour cells (CTC) and circulating tumour DNA
(ctDNA) in the plasma samples can provide extremely
useful information through a non-invasive technique.
CTC are cells of solid tumour origin that can be detected
in the blood of cancer patients. ctDNA on the other hand
can be detected in the peripheral blood specimen predominantly as a consequence of passive release of DNA
fragments from necrotic and apoptotic malignant cells.
Serial liquid biopsies can be performed to give us a
better understanding of the evolution of tumour with
time, identify various prognostic and predictive biomarkers and identify mechanism of resistance to targeted
therapy. Several studies in various tumour types have
looked at evaluating the prognostic impact of CTC on
patient outcomes.3 Results have however been conflicting. Studies have revealed a significant correlation
between disease stage and the presence of tumourassociated genetic aberrations (including mutations in
TP53, KRAS and APC) through ctDNA analysis in the
376
blood of patients with breast, ovarian, pancreatic and
colorectal cancer.4
One major issue with CTC and ctDNA analysis is the
lack of standardisation of technique. This has led to conflicting results in different studies evaluating the prognostic impact of CTC and ctDNA.3,4 CellSearch assay is
the most widely used method for CTC quantification.3
However, the CellSearch method of CTC identification
relies heavily on epithelial surface markers (melanoma
cell adhesion molecule (MCAM)) that is likely to be lost
in the context of metastatic disease involving epithelial
to mesenchymal transition meaning certain CTC subtypes can be missed. Others have evaluated a multimarker approach involving stem cell markers (ABCB5,
CD271 and RANK) along with epithelial surface markers
(MCAM and modified citrus pectin (MCP)) that enriches
the isolation of CTC in melanoma.5 Lack of standardised
techniques in ctDNA analysis can be due to differences
in the DNA extraction material (serum vs plasma) or in
the techniques used for the quantification of ctDNA (digital polymerase chain reaction, next generation sequencing, BEAMing).4
Another promising application for liquid biopsies is the
early detection of cancer relapse after curative surgery.
Surveillance protocols vary for different tumour types
and there is variable clinical practice globally. However,
most surveillance protocols will include clinical examination that has a low sensitivity, tumour markers that can
be non-specific or some patients can be non-secretors
and various radiological assessments that are expensive
and require injection of contrast media. A study by Diehl
et al. showed that it was possible to identify disease
recurrence with almost 100% sensitivity and specificity
by monitoring tumour-specific aberrations (including
APC, TP53 and KRAS) in the plasma of patients with
colorectal cancer post-surgery.6
Role of liquid biopsies in situations where a biopsy
might be difficult to obtain like patients with pancreatic
cancer, deep abdominal or pelvic masses or patients with
bony metastases needs to be explored. One of the most
exciting areas of research is to investigate the predictive
significance of CTC and ctDNA in determining response
and resistance to treatment in various tumour types.
Detection of androgen receptor splice variant-7 (AR-V7)
in CTC from men with advanced prostate cancer is associated with resistance to both enzalutamide and abiraterone, as evidenced by inferior prostate specific antigen
responses, progression-free survival and overall
© 2016 Royal Australasian College of Physicians
Letters to the Editor
survival.7 Sequist et al. recently reported the efficacy of
rociletinib (CO-1686) in plasma-genotyped T790Mpositive non-small-cell lung cancer patients and found
T790M decrease in most patients.8
In conclusion, liquid biopsies offer an exciting area of
future research that has the potential to change clinical
practice and make it more efficient. It will assist the clinicians in complex decision-making, and through serial
liquid biopsies, we will be able to track an evolving
tumour and identify predictive and prognostic biomarkers and study mechanisms of drug resistance. Several
challenges remain before this is possible and these
include lack of standardisation of quantification
References
1 Gazdar AF. Activating and resistance
mutations of EGFR in non-small-cell
lung cancer: role in clinical response to
EGFR tyrosine kinase inhibitors. Oncogene
2009; 28: S24–31.
2 Gerlinger M, Rowan AJ, Horswell S,
Larkin J, Endesfelder D, Gronroos E et al.
Intratumor heterogeneity and branched
evolution revealed by multiregion
sequencing. N Engl J Med 2012; 366:
883–92.
3 Khoja L, Lorigan P, Dive C, Keilholz U,
Fusi A. Circulating tumour cells as
tumour biomarkers in melanoma:
Received 19 August 2015; accepted 14 October 2015.
doi:10.1111/imj.12996
A. Khattak
Department of Medical Oncology, Fiona Stanely Hospital,
Perth, Western Australia, Australia
detection methods and clinical relevance.
Ann Oncol 2015; 26: 33–9.
4 Crowley E, Di Nicolantonio F,
Loupakis F, Bardelli A. Liquid biopsy:
monitoring cancer-genetics in the blood.
Nat Rev Clin Oncol 2013; 10: 472–84.
5 Freeman JB, Gray ES, Millward M,
Pearce R, Ziman M. Evaluation of a
multi-marker immunomagnetic
enrichment assay for the quantification
of circulating melanoma cells. J Transl
Med 2012; 10: 192.
6 Diehl F, Li M, Dressman D, He Y,
Shen D, Szabo S et al. Detection and
quantification of mutations in the plasma
Conflict of interest: real and
perceived – a more mature
consideration is needed
We were somewhat bemused to read the letter regarding
conflict of interest (COI) statements in the Journal
recently1 as it appeared to be a direct communication
resulting from our publication,2 yet we had not been
invited to respond prior to its publication, as would usually occur with correspondence arising. COI are an
important issue as they are impossible to avoid completely and so reducing discussion of COI to a letter in
which inaccurate statements are made and deceitful conduct implied is not helpful. We therefore felt compelled
to put our views and open further more nuanced discussion on this important issue.
Niall says that our study exemplifies the fact that COI
statements’ ‘purpose and interpretation remain unclear’.
I fail to see that this is true. The purpose of a COI statement is to make the reader aware of (potential) biases of
the authors (especially, but not only, financial) which
may be of relevance to the subject matter. The
© 2016 Royal Australasian College of Physicians
techniques, logistics of promptly processing the blood
specimen after it is obtained and lack of widespread
availability of laboratories with expertise in these techniques minimising the generalisability and uptake of liquid
biopsies in routine clinical practice.
of patients with colorectal tumors. Proc
Natl Acad Sci USA 2005; 102: 16368–73.
7 Antonarakis E, Lu C, Wang H, Luber B,
Nakazawa M, Roeser J et al. AR-V7 and
resistance to enzalutamide and
abiraterone in prostate cancer. N Engl J
Med 2014; 371: 1028–38.
8 Sequist L, Goldman J, Wakelee H,
Camidge D, Yu H, Varga A et al. Efficacy
of rociletinib (CO-1686) in plasmagenotyped T790M-positive non-small cell
lung cancer (NSCLC) patients. J Clin
Oncol 2015; 33: abstr 8001.
interpretation is up to the readers – and each will have
their own biases (otherwise also known as COI) that will
lead each to a slightly different interpretation. A reader
will see that authors on our paper made detailed disclosures, beyond this it is up to the readers to judge for
themselves – having read the paper – whether they feel
it is unduly biased.
Niall goes on to say that ‘The origin of these funds
(to provide compassionate access therapy) is not clarified’. This is also inaccurate; our article clearly states that
these doses were from ‘hospital or industry funds’. This
means that either the hospital paid or the drug company
forewent payment. We are not sure how this could have
been more clearly written.
In the following paragraph, there is an indirect implication that we had funding to perform this study from
pharma, which we have simply not yet declared. This is
done by joining our study to a completely separate article in the same issue – yet the other article remains
uncited, leaving us coloured by these comments – poor
writing and sloppy editing.
377
Letters to the Editor
Niall then appears to object to the fact that several of
the authors on our paper declared their relationships
with pharma – we can’t win, as they must be declared.
Because we have involvement with pharma, does Niall
propose that we should never publish on any inflammatory bowel disease-related issue?
Niall talks about ‘competing interests… hidden hand
of the drug industry… quest for open trial data… overdiagnosis and overtreatment’. We fail to see how any of
this is related to our paper: interests are declared; there
is no hidden hand; it is openly declared that it is observational clinical audit. Moreover, there was certainly no
issue with over-diagnosis or overtreatment as would be
obvious to any clinician who has cared for someone with
ulcerative colitis unresponsive to standard therapy.
Allowing these comments to be made in reference to our
paper implies that we are guilty of each of these heinous
crimes. If one is to have a sensible discussion of these
issues, one should at least pick discrete examples where
each of these is an issue, rather than implying our paper
is an excellent example of every possible sin where COI
are concerned.
Furthermore, we find it ironic that Niall does not
make any COI statement. Are we to understand he has
never had any dealings with any industry involved in
delivering medical services? This in itself may be an
undeclared COI, as it is possible that this lack of recognition or engagement is resented. We are not implying this
is true, yet rather suggesting that the issue is most constructively viewed from as broad a perspective as possible
and not reduced to a simplistic concept involving only
pharma companies and money.
From the broader view of healthcare, engagement
with industries, such as pharma, is a necessary part of
expert medical practice. Without such engagement in
Australia, we would have submissions to the Pharmaceutical Benefits Advisory Committee and Medical Services Advisory Committee, which lacked any local
clinical input. We would have educational meetings
solely run by company employees (read ‘marketing
departments’) and many patient support initiatives
(some very useful, and not provided elsewhere by our
public health system) would not exist. Perhaps in an
ideal world, all of these roles would be funded without
pharma – at an increased cost to the taxpayer – however,
they are not, and so a lack of engagement by senior clinicians would appear to be counterproductive in our ‘user
pays’ system especially with regard to Pharmaceutical
Benefits Scheme and Medicare Benefits Schedule
listings.
378
Having this sort of engagement does come with a high
level of responsibility and all clinicians in this position do
need to be careful about recognising their own and the
pharmaceutical industry’s motives. Being naïve about
company motives and failing to acknowledge where
their interests diverge from our patients’ best interest are
perhaps the biggest risks. However, the fact that such
engagement entails risk does not mean that the best way
to proceed is to have a lack of engagement. There is no
other sector of work where anyone would genuinely
propose that two of the biggest stakeholders would
attempt to perform their roles without ever directly
speaking to each other; to propose this in medicine
appears illogical.
On a positive note, we wholeheartedly endorse Niall’s
statement that ‘finance for healthcare is finite’ and this is
the prime reason that several of us have been involved
with and published on clinical outcomes through audit
for some years. It is only by actually examining the outcomes of care that we can be informed as to how much
value we get from our care and modify what we do in
the future when deficiencies are discovered.
Personally, we feel that much more could be achieved
by concentrating on evaluating outcomes of care actually
achieved than sniping from the sidelines when people
who are prepared to engage in consultation and declare
their funding sources dare to audit their clinical results.
Our patients knew from whence extra drug doses
came, and all were very grateful for and comfortable
with the arrangement.
Received 30 December 2015; accepted 5 January 2016.
doi:10.1111/imj.12989
J. M. Andrews,1 S. P. Costello,1 A. K. Agarwal,1
P. Bampton,2 L. Beswick,3 S. Connor,4 S. Ghaly,5
S. O’Connor,6 A. Pudipeddi,5 A. Sechi,7 M. Sparrow8
and A. J. Walsh5
1
IBD Service, Department of Gastroenterology and Hepatology,
Royal Adelaide Hospital, Adelaide 2Department of
Gastroenterology, Flinders Medical Centre, Adelaide,
South Australia, 3Department of Gastroenterology,
Barwon Health, Geelong, 8Department of Gastroenterology,
Alfred Hospital, Melbourne, Victoria,
4
7
Department of Gastroenterology and
IBD Service, Liverpool Hospital, 5Department of Gastroenterology,
St Vincent’s Hospital, Sydney, New South Wales, 6Department of
Gastroenterology, Princess Alexandra Hospital, Brisbane,
Queensland, Australia
© 2016 Royal Australasian College of Physicians
Letters to the Editor
References
1 Niall J. Conflict of interest statements:
clarification. Intern Med J 2015; 45: 1095.
© 2016 Royal Australasian College of Physicians
2 Costello SP, Ghaly S, Beswick L,
Pudipeddi A, Agarwal A, Sechi A et al.
Compassionate access anti-tumour
necrosis factor-alpha therapy for
ulcerative colitis in Australia: the benefits
to patients. Intern Med J 2015; 45:
659–66.
379
Internal Medicine Journal (2016)
Errata
The publisher would like to draw the readers’ attention to an error in the following article:
A. Le Viellez, S. P’Ng, S. Buffery, M. Wright, J. Cooney, P. Cannell and D. Purtill. Red cell and platelet transfusion burden following myeloablative allogeneic haemopoietic stem cell transplantation. Intern Med J 2015; 45: 1286–1292. doi:
10.1111/imj.12894
The following running headers were incorrect.
Optical diagnosis of polyp histology
Chandran et al.
They should have read:
Transfusion burden in allo-HSCT
Le Viellez et al.
The publisher apologises for this error.
The publisher would like to draw the readers’ attention to an error in the following article:
P. Rebeiro and J. Moore. The role of autologous haemopoietic stem cell transplantation in the treatment of
autoimmune disorders. Intern Med J 2016; 46: 17–28. doi: 10.1111/imj.12944
The full form of the first acronym in Table 2 was incorrect.
ASSIST, A Stop Smoking in Schools Trial.
It should have read:
ASSIST, American Scleroderma Stem Cell versus Immune Suppression Trial.
The publisher apologises for this error.
380
© 2016 Royal Australasian College of Physicians
to wileyonlinelibrary.com/tax-vat. The price includes online access from
current content and all online back files to January 1st 2012, where
available. For other pricing options, including access information and
terms and conditions, please visit: wileyonlinelibrary.com/access.
Publisher
Internal Medicine Journal is published by John Wiley & Sons Australia, Ltd
155 Cremorne Street
Richmond, Victoria, 3121
Australia
Tel: +61 3 9274 3100
Fax: +61 3 9274 3101
Journal Customer Services
For ordering information, claims and any enquiry concerning your
journal subscription please go to www.wileycustomerhelp.com/ask or
contact your nearest office.
Americas: Email: [email protected]; Tel: +1 781 388 8598 or 1 800 835
6770 (toll free in the USA & Canada).
Europe, Middle East and Africa: Email: [email protected]; Tel:
+44 (0) 1865 778315.
Asia Pacific: Email: [email protected]; Tel: +65 6511 8000.
Japan: For Japanese speaking support, email: [email protected]; Tel
(toll free): 005 316 50 480.
Visit our Online Customer Get-Help available in 7 languages at
www.wileycustomerhelp.com/ask
Production Editor
Yvonne Lim (email: [email protected])
Information for Subscribers
Internal Medicine Journal is published in 12 issues per year. Institutional
subscriptions on prices for 2016 are: Print & Online: US$1852 (US),
US$2241 (Rest of World), €1452 (Europe). Prices are exclusive of tax.
Asia-Pacific GST, Canadian GST/HST and European VAT will be applied at
the appropriate rates. For more information on current tax rates, please go
Delivery Terms and Legal Title
Where the subscription price includes print issues and delivery is to the
recipient’s address, delivery terms are Delivered at Place (DAP); the recipient is responsible for paying any import duty or taxes. Title to all issues
transfers FOB our shipping point, freight prepaid. We will endeavour to
fulfil claims for missing or damaged copies within six months of publication, within our reasonable discretion and subject to availability.
Printing and Despatch
Printed in Singapore by Markono Print Media Pte Ltd.
All journals are normally despatched direct from the country in which
they are printed by surface air-lifted delivery.
Offprints
C.O.S. Printers Pte Ltd, 9 Kian Teck Crescent, Singapore 628875. Fax:
+65 6265 9074. Email: [email protected]
Back Issues
Single issues from current and recent volumes are available at the
current single issue price from [email protected]. Earlier issues
may be obtained from Periodicals Service Company, 11 Main Street,
Germantown, NY 12526, USA. Tel: +1 518 537 4700, Fax: +1 518 537 5899,
Email: [email protected]
Copyright and Photocopying
© 2016 Royal Australasian College of Physicians. All rights reserved. No
part of this publication may be reproduced, stored or transmitted in any
form or by any means without the prior permission in writing from the
copyright holder. Authorisation to photocopy items for internal and
personal use is granted by the copy-right holder for libraries and other
users registered with their local Reproduction Rights Organisation (RRO),
e.g. Copyright Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA
01923, USA (www.copyright.com), provided the appropriate fee is paid
directly to the RRO. This consent does not extend to other kinds of copying
such as copying for general distribution, for advertising or promotional
pruposes, for creating new collective works or for resale. Special requests
should be addressed to [email protected]
IMJ.PI.JAN16
IMJ_46_3_cover.indd 2
08/03/16 10:38 PM
previous
Total-pages:
Paper-type:
issue 3
Editorial
253
March 2016
332
Associations of demographic and behavioural
factors with glycaemic control in young adults
with type 1 diabetes mellitus
J. K. Osan, J. D. Punch, M. Watson, Y. X. Chan, P. Barrie,
P. G. Fegan and B. B. Yeap
339
Inaccurate risk perceptions contribute to
treatment gaps in secondary prevention of
cardiovascular disease
J. Thakkar, E. L. Heeley, J. Chalmers and C. K. Chow
The threat of litigation as a possible barrier
to innovation
J. Coates
Review
255
Cardiac resynchronisation therapy in 2015:
keeping up with the pace
A. Voskoboinik, A. D. McGavigan and J. A. Mariani
Brief Communications
347
Clinical Perspectives
266
Characterisation and therapeutic manipulation
of the gut microbiome in inflammatory bowel
disease
J. Schulberg and P. De Cruz
Original Articles
273
281
Regional challenges: evaluation of a hepatitis
outreach programme using transient
elastography (FibroScan) in Victoria
M. C. Thurnheer, T. R. Schulz, T. Nguyen, J. MacLachlan
and J. Sasadeusz
Baseline abnormal liver function tests are more
important than age in the development of
isoniazid-induced hepatoxicity for patients
receiving preventive therapy for latent
tuberculosis infection
E. L. Gray and H. F. Goldberg
Home medicines reviews in Australian war
veterans taking warfarin do not influence
international normalised ratio control
L. R. E. Bereznicki, E. C. van Tienen and A. Stafford
295
Do community hospice programmes reduce
hospitalisation rate in patients with
advanced chronic obstructive pulmonary
disease?
S. P. M. Iupati and B. R. Ensor
301
307
315
325
IMJ_46_3_cover.indd 1
352
Substantial variation in post-engraftment
infection prophylaxis and revaccination practice
in autologous stem cell transplant patients
H. Y. Lim and A. Grigg
Safety and efficacy of a novel short occlusive
regimen of imiquimod for selected non-melanotic
skin lesions in renal transplant patients
G. Das, B. Tan and K. Nicholls
356
Acute adrenal insufficiency: an aide-memoire
of the critical importance of its recognition and
prevention
A. Gargya, E. Chua, J. Hetherington, K. Sommer and
M. Cooper
360
Severe hypocalcaemia and hypophosphataemia
following intravenous iron and denosumab:
a novel drug interaction
B. Smyth and S. Ong
364
Cardiovascular risk reduction in hypertension:
angiotensin-converting enzyme inhibitors,
angiotensin receptor blockers. Where are
we up to?
A. Sindone, J. Erlich, C. Lee, H. Newman, M. Suranyi
and S. D. Roger
Letters to the Editor
Clinical-scientific notes
373
Screening for coeliac disease in elderly inpatients
with minimal trauma fracture is not indicated
D. I. Clayton-Chubb, M. Wilson, E. Seal and P. W. Lange
Alcohol use disorder hospitalisations over the
last two decades: a population-based cohort
study
W. Liang and T. Chikritzhs
374
Association between hepatitis B virus infection
and risk of multiple myeloma: a systematic
review and meta-analysis
Y. Li, O. Bai, C. Liu, Z. Du, X. Wang, G. Wang
and W. Li
General correspondence
376
Estimation of clinical and economic effects of
prophylaxis against venous thromboembolism
in medical patients, including the effect of
targeting patients at high-risk
J. A. Millar and A. L. K. Gee
Liquid biopsies: advancing cancer research
through drops of blood
A. Khattak
377
Conflict of interest: real and perceived – a more
mature consideration is needed
J. M. Andrews, S. P. Costello, A. K. Agarwal, P. Bampton,
L. Beswick, S. Connor, S. Ghaly, S. O’Connor,
A. Pudipeddi, A. Sechi, M. Sparrow and A. J. Walsh
A tertiary hospital audit of opioids and
sedatives administered in the last 24 h of life
C. Douglas, M. Clarke, S. Alexander and M. Khatun
Total-pages:136
INTERNAL
MEDICINE
JOURNAL
Spine-width: 4.4mm
Paper-type: UPM Classic 64
Volume
46
Issue
3
March
2016
ISSN 1444-0903
Personal Viewpoint
380
Traumatic superior gluteal artery pseudoaneurysm
following a bone marrow biopsy
I. R. Caldwell, B. T. Buckley, R. Rajagopal,
R. Doocey and L. Pemberton
INTER NAL MEDICINE JOUR NAL
288
Current
Volume 46, Issue 3, March 2016, Pages 245–380
VOLUME 46
Spine-width:
Cardiac resynchronisation therapy update
The gut microbiome in inflammatory bowel disease
Community hospice programmes and COPD
Type 1 diabetes control in young adults
Alcohol use disorder hospitalisations
Opioids and sedatives in the last 24 hours of life
Acute adrenal insufficiency: recognition and prevention
Errata
08/03/16 10:38 PM