Download Percutaneous left atrial appendage (LAA) occlusion

Technology Note
Percutaneous left atrial appendage (LAA) occlusion for the treatment of
atrial fibrillation (AF) to prevent stroke
March 2013
Introduction
This document synthesises the evidence for the use of percutaneous left atrial appendage (LAA)
occlusion as a method of stroke prevention for patients with atrial fibrillation (AF) – a common
cardiac arrhythmia. AF patients face a five-fold increase in stroke risk over non-AF patients and
LAA occlusion is a way to mitigate that risk (Lever & Larsen, 2011).
The current gold standard treatment for stroke prevention in patients with AF is warfarin. Warfarin,
an anticoagulant, is associated with an estimated prevention of 70% of thromboembolic events.
Though effective, warfarin has a narrow therapeutic range and interacts with a wide variety of other
drugs.
About 50% of high-risk AF patients are not treated with warfarin due to severe
disability/frailty, high risk of falls and limited life expectancy (Salzberg & Tolboom, 2011). These
patients go untreated and are more vulnerable to stroke than warfarin patients. For patients who
are prescribed warfarin and stick to the regimen, the dosage must be monitored/adjusted and
alcohol and diet must be restricted in order for the drug to remain effective.
These factors
contribute to warfarin’s nearly 40% discontinuation rate (Forel, 2010).
For patients unable to take warfarin, the current pharmaceutical alternatives include dabigatran,
aspirin and aspirin plus clopidogrel. Aspirin reduces the risk of stroke by approximately 22% and
although dabigatran is equivalent to warfarin in effectiveness, it is more expensive and there have
been concerns about the inability to reverse anticoagulation in emergencies (Hart, Benavente,
McBride, & Pearce, 1999). Additionally, it is not recommended as a substitute for warfarin in
patients with prosthetic valve implants. Recent warnings from the Federal Drug Administration
(FDA) and Medsafe have added prosthetic valve implants to the list of dabigatran contraindications
because of the results of the failed RE-ALIGN trial (FDA, 2012; Medsafe, 2013) (Table 1).
1
Table 1
As of December 10, 2012. the RE-ALIGN trial data is provided by Pradaxa’s manufacturer (Boehringer Ingelheim
Pharmaceuticals, Inc.); therefore, the data have not undergone quality assurance procedures or verification by FDA.
Death
Stroke
Systemic embolism event (SEE)
Transient ischemic attack (TIA)
Valve thrombosis (VT)
Myocardial infarction (MI)
Composite of events:
Death/stroke/SEE/TIA/VT/MI
Major bleeding
Major bleeding in pericardial location
Any bleeding
Pradaxa
(n=160)*
1 (0.6%)
8 (5.0%)
0
2 (1.3%)
4 (2.5%)
3 (1.9%)
Warfarin
(n=89)*
2 (2.2%)
0 (0%)
0
2 (2.2%)
0
0
16 (10.0%)
4 (4.5%)
6 (3.8%)
5 (3.1%)
36 (22.5%)
1 (1.1%)
0
12 (13.5%)
Source: (FDA, 2012)
These results may apply to percutaneous LAA occlusion patients implanted with a device who are
unable to take warfarin. Because of a lack of suitable alternatives to warfarin in percutaneous LAA
occlusion patients, options are even more limited as safety is still an issue during the recovery
period. Also, these alternative anticoagulants pose similar bleeding risks to warfarin so some
patients are unable to take any of these medications. For them, surgical techniques may be
available. Surgical techniques vary, but they all occlude the LAA from the left atrium using various
clips and sutures such as the AtriClip surgical closure device. Whilst common in rheumatic mitral
valve disease surgery, these LAA surgical procedures vary in safety and require similar resources
as percutaneous implantation of LAA occlusion devices (Forel, 2010).
Percutaneous LAA
occlusion is proposed as an additional alternative treatment for warfarin intolerant AF patients.
According to Dr Gordon Tomaselli, a cardiologist from Johns Hopkins, percutaneous LAA
occlusion may provide an alternative treatment for about 5.0% of the AF population (Phend, 2013).
Of the percutaneous LAA occlusion treatments available, only the WATCHMAN device has been
evaluated thoroughly in a randomised controlled trial (RCT) in humans- the PROTECT-AF trial.
Accordingly, this Technology Note evaluates the evidence for the WATCHMAN device against the
National Health Committee’s four assessment domains to enable evaluation of the intervention
using the Committee’s 11 decision-making criteria.
2
Burden of Disease and Target Population
AF is the most common sustained cardiac arrhythmia with a prevalence of 0.5%-2.0% of the world
population (Go et al., 2003). In New Zealand, this translates to a crude rate of between 50 and
200 cases per 10,000 people. Of those patients with AF, 50.0% develop permanent AF (Levy,
Maarek, Coumel, & al., 1999). Since the prevalence of AF increases with age, the epidemiological
burden in New Zealand is likely to increase with the ageing population. The prevalence rate
roughly doubles with each decade in a person’s lifetime going from 0.5% for ages 50-59 to nearly
9.0% for ages 80-89 (Latif & Messinger-Rapport, 2004). For Maori, AF prevalence is nearly twice
that of the rest of the New Zealand population with younger Maori aged 40-60 facing a 1.78 times
AF prevalence when compared to non-Maori of the same age (Figure 1) (NHC, 2012). With an
additional burden of AF in New Zealand, a five-fold increase in AF-related stroke is expected by
2050 (Miyasaka & al., 2006).
Figure 1. Prevalence of Atrial Fibrillation in New Zealand
The figure shows the relationship between age and rising prevalence of AF. It shows the prevalence of AF is
consistently higher in Maori, whereas the prevalence in Pacific and European patients conforms closely.
Source: NHC executive analysis of national collections.
AF patients face a five-fold increase in stroke over non-AF patients. In 2010-11, approximately
2.3% of AF patients were admitted to hospital with a stroke. The incidence of strokes attributable
to AF increases with age going from 1.5% at ages 50-59 to 23.5% at ages 80-89.
AF is
responsible for 10% of all ischaemic strokes and 50% of all cardioembolic strokes. Strokes affect
an estimated 347,000, or 1.5% of the Australian population, and are a major cause of death and
hospitalisations (35,000) annually. In New Zealand, over 7,000 (0.16% of the population) people
suffer a stroke annually, a lower incidence rate than Australia, yet the rate is higher for Maori and
3
Pacific than New Zealand Europeans (Dyall, Feigin, Brown, & Roberts, 2008). Additionally, since
13.5-18.0% of all strokes have a cardioembolic origin in the LAA, risk mitigation could have a large
impact on stroke prevention for patients unable to take warfarin. Once a patient has had a stroke,
there is a high chance of dependency1 especially for Maori and Pacific groups who have a 300%
higher chance of dependency than Europeans one year after a stroke (Dyall, et al., 2008).
Assuming that 5.0% of the AF population might benefit from WATCHMAN in New Zealand (Phend,
2013), and that the prevalence rate remains constant, this translates to a prevalent target
population of between 1,100 and 4,400 people (or 25 -100 people per 100,000 individuals) in the
total population and an additional 111 cases per annum over the next 10 years as the result of a
65+ population growth rate of 3.5% (NHC, 2012).
Alternative Technology
Besides the WATCHMAN device, the Amplatzer Cardiac Plug and Amplatzer Septal Occlusion
devices are the other main percutaneous devices. The Amplatzer Septal Occlusion device is
targeted for patients with an atrial septal defect (ASD) but has been successfully trialled for LAA
occlusion in one prospective study (Meier et al., 2003). Although it has FDA approval, it has been
superseded by the cardiac plug for LAA occlusion.
The cardiac plug is designed to occlude the LAA but lacks a clinical trial. As of January 2013, the
AMPLATZER Cardiac Plug Clinical Trial is still recruiting patients and final results will not be
available until 2017. The Percutaneous Left Atrial Appendage Occlusion (PLAATO) device had
successful trials but was pulled from the market after it failed to receive FDA approval. Before
production was discontinued, the feasibility of performing percutaneous LAA occlusion in warfarin
intolerant patients was demonstrated in a multi-centre trial. Results from the International MultiCenter Feasibility Trials claimed successful implantation of the PLAATO device in 108 of 111
patients (97.3%, 95% CI: 92.3% to 99.4%) with 5 patients suffering from a major adverse event
through follow-up (Ostermayer et al., 2005).
Other percutaneous devices include the Coherex Wave Crest and the LARIAT device, both of
which are undergoing trials, are less studied and have questions about their clinical safety. Of
these devices, only the WATCHMAN and LARIAT devices are approved by the FDA for use in
stroke prevention for AF patients. However, the LARIAT had very few human subjects in its trial
(13) and is associated with an increased chance of pericarditis (Bartus et al., 2011).
The
advantage to the LARIAT device would be that it does not require warfarin directly after surgery;
however, its lack of clinical evidence makes it unsuitable currently as a warfarin alternative. Table
2 summarises the various devices and their studies:
1
Dependency occurs when a patient requires support from another person during day-to-day activities
4
Table 2. Summary of percutaneous left atrial appendage occlusion devices
Number
Device
Study
Design
of
Inclusion criteria
Mean F/U
Results
Comments
patients
(i) Requires both endocardial and
epicardial access
(ii) Unsuitable in patients with
AF; C/I to warfarin or intolerance
LARIAT
Lee et al.
[39]
Prospective
82
to warfarin or pts who have had
an embolic event on whilst on
3 months
warfarin
96% of patients with
possible pericardial adhesions
successful closure
(e.g., prior history of coronary
continued to have
artery bypass surgery, valvular
complete closure at 1
surgery, pericarditis, and chest
month
radiotherapy)
(iii) Clinical trial data pending
(iv) FDA and CE mark
APPROVED for commercial use
Reddy et al.
WATCHMAN
(PROTECT
RCT
707
AF) [33]
AMPLATZER
Park et al.
cardiac plug
[35]
AMPLATZER
Meier et al.
Permanent or paroxysmal AF;
CHADS2 ≥ 1; suitable for warfarin
18 months
24 hours
Registry
141
Permanent or paroxysmal AF
afterimplantation
Prospective
16
Permanent or paroxysmal AF;
4 months
Probability of
(i) Efficacy demonstrated in
noninferiority of the
clinical trial
intervention was more
(ii) FDA (preliminary) and CE mark
than 99.9%
APPROVED for commercial use
Stroke 2.1% Device
embolisation 1.4%
Pericardial tamponade
3.5%
TIA/stroke 0%
(i) Clinical trial data pending
(ii) CE mark APPROVED for
commercial use
(i) Not a dedicated LAA occluder
5
septal
[28]
C/I to warfarin
Device embolisation 6.3%
occluder
(ii) Has been superseded by the
AMPLATZER Cardiac Plug for
LAA occlusion
(iii) Not in Commercial use
Ostermayer
et al. [30]
Permanent nonrheumatic AF;
Prospective
111
PLAATO
Block et al.
[31]
Prospective
(currently
Wave Crest
recruiting)
10 months
TIA/stroke 2.2%
warfarin
(i) No longer available for clinical
Permanent or paroxysmal AF;
use
CHADS2 ≥ 2;
5 years
Stroke: 3.8%
C/I to warfarin
(i) Retractable coils and anchors
Muller
Coherex
64
patients at risk for stroke; C/I to
52Prospective
[36]
actively
recruiting
Permanent or paroxysmal
Data available on 10 cases
nonvalvular AF; CHADS2 ≥ 1
only, 1 embolic event
to enable optimal device
positioning
(ii) Clinical trial data pending
(iii) Not in commercial use
C/I: contraindication and AF: atrial fibrillation.
Source: (Luis et al., 2012)
6
Technology Status in New Zealand
The WATCHMAN device is not available commercially in New Zealand. However, in 2011, the
Auckland Clinical Practice Committee (CPC) approved a ‘field trial’ of the WATCHMAN device in
order to gain further follow-up data on the procedure’s effectiveness and safety. After considering
various target populations, the CPC endorsed further research into the use of the procedure in 1020 patients with limited warfarin tolerance (up to 45 days for recovery). In order to include enough
people in the research, the procedure needed to be trialled beyond the Auckland catchment.
Because DHBs face the challenge of funding patients from other DHBs with their own funds during
a field evaluation across multiple DHBs, a central funding source is often required to help DHBs
mitigate that problem. On this basis, it was referred to the National Health Committee (NHC) to
consider for public funding, particularly in the absence of alternative treatment for certain patient
groups.
International Status
In 2009, the WATCHMAN device gained FDA approval for the purpose of investigational use2. The
FDA’s advisory panel voted 7 to 5 in favour of approving the WATCHMAN device for
investigational use in the United States (US) pending further trials and procedure training. As a
condition for investigational use, progress reports (i.e. registries) are required at regular intervals to
track the intervention’s long-term safety and effectiveness. These records provide a link between
controlled trials and real world conditions and are important to consider when assessing the value
and safety of a technology. In the US, WATCHMAN patients are being monitored in the Continued
Access PROTECT-AF (CAP) and Aspirin-Plavix (ASAP) registries. Clinical trials are still underway
in the US that will inform these registries so final approval for commercial use is pending.
The WATCHMAN device has CE approval3 for commercial use in Europe and is widely used in
Germany (Forel, 2010). In Australia, the device has Therapeutic Goods Administration (TGA)
approval as a relatively risky Class III device4. HealthPACT reviewed the technology in 2010
recommending that no further assessment be completed on the WATCHMAN device citing a lack
of high-quality evidence in favour of WATCHMAN over warfarin in stroke prevention (Forel 2010).
HealthPACT’s recommendation for further follow-up data has been published from the PROTECTAF RCT, which is included in this assessment. A recent review by DLA-Piper, on behalf of
HealthPACT, found that there is still a paucity of high-quality long-term data surrounding health
outcomes and the need for on-going anticoagulation following LAA occlusion (DLA Piper, 2013).
2
3
4
An investigational device exemption (IDE) allows the investigational device to be used in a clinical study in order to
collect safety and effectiveness data
A CE Marking is a European marking of conformity that indicates a product complies with the essential requirements
of the applicable European laws or directives with respect to safety, health, environment and consumer protection.
See Appendix 1 for definitions of TGA approval classes
7
In 2012, the National Health Service Commissioning Board (NHSCB) in the UK took the following
position: ‘Left Atrial Appendage occlusion devices for stroke prevention in people with atrial
fibrillation will not be routinely commissioned and funded as the published data fails to show
sufficient benefit for patients or good value for the NHS’ (NHSCB, 2012).
Literature Search Strategy
A systematic search was conducted to find clinical and economic evidence of LAA occlusion.
Key Health Technology Assessment (HTA) agencies and repositories were searched using the
search terms: watchman, left atrial appendage occlusion5. MEDLINE, and Cochrane databases
were searched from 1 January 1946 and 21 November 2012. The search terms used were: clinical
trial, meta-analysis, randomized controlled trial, cost effective and systematic review. Results were
limited to either reviews OR clinical studies that used human subjects. No language limits were
applied. The reference lists of identified publications were not searched.
Abstracts were sorted through and categorised, on the basis of their content, into one of two
categories: Clinical or Economic. All abstracts were assessed, but only the full text of relevant
studies where retrieved. A critical appraisal using the GATE methodology for clinical studies and
the CASP checklist (Appendix 2) for review articles and cost effectiveness studies was conducted
to assess the quality of the evidence using the US Preventive Services Task Force (USPSTF)
grading scale. Out of the 26 articles identified in the systematic search, 5 studies and 6 review
articles were identified as relevant and were subsequently appraised. Out of the 5 studies (2
prospective RCTs, 1 prospective single-arm non-randomised trial, 1 retrospective non-randomised
trial and 1 open label non-randomised feasibility pilot study) only 1 was identified as being of ‘fair’
quality and relevant to the questions in the safety/clinical effectiveness domain, the PROTECT-AF
RCT. Additionally, two HealthPACT documents were also cited in subsequent searching. Lastly,
two further recent studies were identified that may provide further clinical evidence, the PREVAIL
RCT and ASAP Registry trial. PREVAIL and ASAP have recently concluded, but final results have
not yet been fully published. Data from these trials may inform more sophisticated economic
analysis in the future. Of the 5 review articles, 1 systematic review and 1 best evidence review
article were deemed suitable and relevant. One cost effectiveness analysis was performed by the
Canadian Health Technology Inquiry Service and is used to approximate the economic burden of
percutaneous LAA occlusion in New Zealand.
5
The full search strategy for both the clinical safety/effectiveness and cost effectiveness search are included in
Appendix 1
8
Clinical Safety and Effectiveness
This section reviews the clinical safety and effectiveness evidence for the WATCHMAN device. As
noted above, only one published study, the PROTECT-AF RCT, was found to be sufficient quality
for data inclusion. Nevertheless preliminary results from two recently completed, but yet
unpublished studies, the PREVAIL RCT and ASAP Registry trial, are also outlined.
PROTECT-AF RCT
The PROTECT-AF RCT is a randomised, non-inferiority trial comparing WATCHMAN to warfarin
alone, and is the only completed and published RCT to date on WATCHMAN. It is approaching its
third year of follow-up and is the basis for the safety and effectiveness of the device. Published
data from the trial claimed WATCHMAN to be as safe and effective as warfarin in the population
studied provided warfarin is given for at least 45 days following implantation of the device
according to the study’s non-inferiority margin. The study’s claim is based on the comparative
incidence of various effectiveness and safety endpoints. The primary endpoints were ischaemic
stroke, cardiovascular disease, haemorrhagic stroke, systemic embolism, all stroke, all mortality
and primary safety. As illustrated in Figure 2, the WATCHMAN device has significantly6 lower rates
of cardiovascular death (CVD) and haemorrhagic stroke than warfarin alone but has a significantly
increased risk of systemic embolism. The incidence of haemorrhagic stroke was lower in the
WATCHMAN group, presumably due to reduced anticoagulant (e.g. warfarin) use.
Overall,
however, there was no significant difference between the incidence of all stroke types together
(ischaemic and haemorrhagic), especially as the study progressed.
There are three other important considerations. The device could not be implanted in 9.0% of
patients randomised to the intervention arm of the trial, 14.0% of WATCHMAN patients had to take
warfarin for 6 months and 8.0% of patients in the WATCHMAN arm had to continue warfarin
indefinitely. Thus, percutaneous LAA occlusion procedures do not completely eliminate the need
for warfarin.
6
Generally, if the confidence intervals of two variables overlap by more than 25.0%, the difference between the two
confidence intervals may be insignificant (Wiley 2011). Otherwise, the difference may be statistically significant. A
confidence interval is the range of values in which the true parameter value of interest could lie given a certain level
of confidence. In this case, the parameter of interest is the true incidence of a particular endpoint and the confidence
level is 95%.
9
Figure 2. Comparison of outcomes
This graph displays the actual values (denoted by a circle) and the associated confidence interval. Primary safety included adverse
events such as bleeding and pericardial effusion.
Source: (Holmes, Reddy, Turi, & al., 2009)
Although the procedure met its pre-specified non-inferiority margin, safety was still a concern in the
PROTECT-AF RCT. The WATCHMAN group had an adverse event rate (primary safety) of 7.4
per 100 patient years compared to 4.4 for the warfarin control group. The relative events rate was
1.69 (95% CI: 1.01-3.19), which represents a significantly higher rate of adverse events in the
WATCHMAN group. The main drivers of this finding are listed in the paragraph below. The device
is also currently being redesigned to help improve safety standards.
The most common adverse event was pericardial effusion (4.8%). There was also a higher risk of
procedure-related ischaemic stroke (1.1%), procedure-related arrhythmia (0.2%), device
embolization (0.6%) and oesophageal tear (0.2%) than in the warfarin control group. Adverse
events decreased as the study progressed yet they still remained higher in the WATCHMAN group
through follow-up. This indicates that there is a learning curve associated with device insertion and
higher risks overall with use of the device when compared to warfarin (Holmes, et al., 2009). For
this reason, the FDA advisory panel recommended a registry for physicians wishing to perform the
procedure as well as further certification as a condition for full FDA approval. These increased
risks may indicate that the WATCHMAN device should only be used after the failure of standard
treatment and not in the first-line.
10
Limitations
PROTECT-AF has several important limitations:

The trial was conducted in patients with paroxysmal, persistent and permanent AF which
has different implications for risk of stroke than a homogenous group of patients with
permanent AF.

There was confounding in the study as a proportion of patients in the intervention arm were
restarted on warfarin therapy for various reasons. With this confounding, it is difficult to
interpret the results based on the established intention-to-treat analysis. For patients in the
intervention arm, this could have overestimated the device’s beneficial effect on CVD and
ischaemic stroke whilst also overestimating the risk of haemorrhagic complications,
including haemorrhagic stroke.

The trial had relatively few high-risk AF patients with an increased risk of stroke. Only
32.0% of patients had a CHADS2 score greater than 2 while nearly a third of patients’
CHADS2 scores were equal to 1 (Holmes, et al., 2009). Without information on riskier
patients, this study is unable to comment on the device’s effectiveness and safety in riskier
patients.

Patients with contraindications to warfarin were excluded from the study. On the basis of
current evidence, WATCHMAN would not, therefore, be indicated for patients who have an
absolute contraindication to warfarin.
As envisaged by the Auckland CPC, the
WATCHMAN device would address a treatment gap for a relatively small patient
population, i.e. those who do not have a contraindication for warfarin but can tolerate only a
limited duration of use. Under this scenario, the WATCHMAN device would be used as an
alternative treatment for patients but still does not fill the current treatment gap for patients
who have an absolute contraindication for warfarin.
The Aspirin Plavix (ASAP) Registry
The ASAP Registry study included 150 patients at three centres in Germany and one in the Czech
Republic. It documents the safety outcomes from the use of the WATCHMAN procedure in patients
who were contraindicated for even short-term warfarin use. The ASAP study reported 4 strokes (3
ischaemic) and 6 device-related thrombus incidents out of the 142 patients (i.e. 7.0% of patients)
with successfully implanted devices (95.0% success rate) in its preliminary abstract from October
2012 (Stiles, 2012). This is a similar rate to that found in the PROTECT-AF trial, which suggests
WATCHMAN may be no less safe in those able to take aspirin and clopidogrel.
The researchers compared this intervention with a control cohort of patients with a similar CHADS2
score, and reported that the incidence of stroke was 77.0% less in the WATCHMAN group than in
11
an aspirin only population. This result must be treated with some trepidation as this was a nonrandomised comparison with a significant and unreported propensity for bias. The study
concluded that the WATCHMAN device may be safe and effective in patients unable to take
warfarin. Although these results suggest similar safety for patients unable to take warfarin as in
patients able to take warfarin, they do not report the adverse events in patients for whom
implantation was unsuccessful. This study should be viewed purely as hypothesis-generating.
Eventually, the study will report follow-up data for at least 250 patient years so it is premature to
make definitive conclusions from these preliminary results. Final results are expected in November
2013.
The PREVAIL RCT
The PREVAIL RCT is a United States based study to test the same safety and efficacy endpoints
as PROTECT-AF but in a riskier AF population. The study has an estimated enrolment of 475
patients.7 Patients in this study were excluded if they were contraindicated for warfarin, but they
must have had a CHADS2 score greater than or equal to 2 in order to participate. Results from this
study are preliminary and unpublished, yet a recent newspaper article has released some of the
results (PR Newswire, 2013). The PREVAIL trial met the pre-specified criteria for the first coprimary endpoint of occurrence of all-cause death, ischaemic stroke, systemic embolism, or device
or procedure-related events requiring open cardiac surgery or major endovascular intervention
(randomisation to seven days post procedure or by hospital discharge, whichever is later). The trial
did not meet the pre-specified criteria for the second co-primary endpoint of the occurrence of all
stroke (ischaemic or haemorrhagic), cardiovascular death and systemic embolism at 18 months.
While the second co-primary efficacy endpoint was not met, the device performed similar to
warfarin with a rate ratio of 1.07. The PREVAIL trial met its pre-specified endpoint for the third coprimary endpoint of the composite of the occurrence of late ischaemic stroke and systemic
embolism (eight days post randomisation and onward) at 18 months. The reported endpoint results
are preliminary and require final validation. Specifically, safety data demonstrated an increase in
implant success rate overall (95.0%), and with new operators (93.2%), compared to PROTECT AF
(90.9%). The overall seven-day serious procedure/device related complication rate was 4.4% in
PREVAIL vs. 8.7% in PROTECT AF, a 49.0% relative reduction. A key result of the PREVAIL trial
was that pericardial effusions requiring intervention occurred at a rate comparable to other left
atrial procedures. PREVAIL reported a 1.9% event rate vs. 4.0% in PROTECT AF, a 52.0%
relative reduction. Additionally, new operators had only one occurrence (1.0%) of pericardial
7
Clinical Trials Gov Site
12
effusion requiring intervention with no device embolization, peri-procedural strokes or cardiac
perforation.
This study combined with the ASAP study may produce evidence of the effectiveness and safety of
WATCHMAN in patients who are at greater risk of stroke and/or have a contraindication to
warfarin.
The PREVAIL study is expected to be reported by June 2017 with data collection
completed in December 2013 (Reddy, Holmes, et al. 2012).
Economic Domain
This section reviews results from the Canadian cost effectiveness model, presents New Zealand
cost effectiveness estimates based on the model and estimates the potential cost impact from the
WATCHMAN device.
Canadian cost effectiveness model
The Canadian study suggests the cost per QALY associated with WATCHMAN could be anywhere
from CA$40,229 to CA$107,239 depending on if the comparator is no treatment or aspirin,
respectively (Health Technology Inquiry Service, 2010). A higher cost per QALY is obtained when
the comparator is aspirin (relative stroke risk ratio of 0.36) compared to no treatment (relative
stroke risk ratio 0.60). A less than doubling of the relative risk ratio more than doubles the cost per
QALY which means there is a relatively smaller marginal health gain for the extra marginal
expenditure, at least in the short-term. The variation in cost-effectiveness is indicated in Table 3.
Cost effectiveness is more favourable for patients when the comparator is aspirin and the time
period is long.
Younger patients unable to take warfarin could potentially benefit from this
treatment if clinical safety and effectiveness evidence improves; however, further analysis would
be needed to assess the validity of this scenario.
Table 3. Canadian cost effectiveness summary
This table highlights the various cost effectiveness potential of LAA occlusion procedures. All scenarios show a reduction in overall
stroke, but the degree to which the procedure is cost effective/saving is highly dependent on the comparator as well as the time horizon.

One way Sensitivity analyses
Incremental (LAA occlusion-No treatment)
Expected Costs
Expected Strokes
Expected QALY’s
Cost per QALY
CHADS2
0
$9,164
-0.048
0.040
$226,392
1
$7,954
-0.072
0.068
$116,804
2
$6,420
-0.103
0.104
$61,980
3
$4,173
-0.146
0.157
$26,617
4
$1,435
-0.198
0.224
$6,415
Time horizon
13
3
$7,907
-0.079
0.050
$158,193
5
$5,270
-0.125
0.131
$40,229
10
-$163
-0.206
0.392
dominates
20
-$5,870
-0.269
0.780
dominates
RR of stroke LAA occlusion
0
$1,213
-0.208
0.221
$5,480
0.2
$3,502
-0.161
0.170
$20,563
0.4
$5,703
-0.116
0.121
$47,182
0.6
$7,818
-0.074
0.073
$107,239
0.8
$9,850
-0.033
0.026
$373,513
Source: (Health Technology Inquiry Service, 2010; Holmes, et al., 2009)
The base-case scenario in the Canadian model considered a time-horizon of 5 years. The
sensitivity analysis varied the time horizons from 3-20 years. The longer the time horizon, the more
cost effective the percutaneous LAA occlusion procedure became and it was cost saving over
longer time horizons. While encouraging, this result must be analysed cautiously since there are
no published clinical studies directly comparing percutaneous LAA occlusion to aspirin or no
treatment on which to accurately base these findings. If safety and effectiveness evidence for
percutaneous LAA occlusion does become available for patients who are unable to take oral
anticoagulants, it may be worth modelling in younger patients with a longer time horizon. Current
evidence, though, can only accurately inform an economic model that compares percutaneous
LAA occlusion to warfarin.
This study did not include societal costs (i.e. lost productivity) and used Canadian estimates for
procedure/device costs based on different devices/procedures to WATCHMAN. Additionally, it
assumed no treatment as the comparator instead of aspirin which greatly reduced the cost per
QALY.
New Zealand Economic Model
A decision analytic model, illustrated in Appendix 3, was used to compare the cost per qualityadjusted life year (QALY) of percutaneous LAA occlusion patients versus warfarin patients. All
costs were obtained using New Zealand sources (Appendix 3) and propensities were obtained
using results from the PROTECT-AF RCT. The model did not take into account societal costs nor
did it model the comparison of percutaneous LAA occlusion to aspirin as was the case in the
Canadian model. The timeframe for the model was 10 years and the discount rate was consistent
with PHARMAC’s 3.5%.
Since there is a paucity of safety evidence for patients unable to take oral anticoagulants, the
comparator in the model was assumed to be warfarin. Using New Zealand cost estimates, the
14
incremental cost effectiveness ratio (ICER) of WATCHMAN versus warfarin was calculated to be
about NZ$101,000 per QALY gained. This figure assumes that stroke risk for WATCHMAN was
equal to that of warfarin. This is roughly equivalent to the results of the Canadian study. The
model was highly sensitive to stroke risk with cost per QALY dropping to around NZ$60,000 when
stroke risk for WATCHMAN patients was higher than for warfarin patients (as was the case in the
PROTECT-AF trial).
Stroke has significant downstream costs, both societal and fiscal. Disability as the result of stroke
leads to lost productivity as well as higher rates of dependency. If the WATCHMAN does reduce
stroke in patients with no alternative treatment, this could lead to savings not only from stroke but
also in rehabilitation costs as well as improved productivity from the reduction of stroke-related
disability. As was the case in both cost models, extending the time horizon increased the cost
effectiveness of WATCHMAN versus warfarin since most of the costs associated with
WATCHMAN are borne upfront and most of the costs associated with warfarin are realised over
time.
The ICERs are based on the assumption that the device’s effectiveness matches that of the
PROTECT-AF trials. Given the lack of experience with the device in New Zealand, it is highly likely
that there would be an increase in adverse events associated with the device such as pericardial
effusion and device embolization.
These outcomes would increase costs and decrease
effectiveness at least until the learning curve flattens out.
Although there is questionable evidence around safety for patients who are contraindicated for
warfarin, it is these patients who may stand to gain the most from WATCHMAN funding.
In
practice, introducing the device for a target population with ‘limited warfarin tolerance’ risks scope
creep where clinicians may stretch the definition of limited warfarin tolerance and prescribe the
device to patients outside of that range. This could greatly affect the cost effectiveness/ clinical
effectiveness of the device. Summarised below in Table 4 are the main results for various target
populations.
15
Table 4. Summary of findings
Population group
Clinical evidence
Economic evidence (cost-effectiveness)
AF patients who discontinue
Increased adverse events related to
ICER of between NZ$139,600 and
warfarin use but are not
device but may be acceptable given
NZ$166,415 per QALY gained; Not cost-
contraindicated
warfarin risks
effective:- WATCHMAN is more expensive
than warfarin and does not provide
greater benefit.
AF patients with no warfarin
None currently; Pending ASAP study
ICER unknown. Potentially cost-effective-
tolerance
results
Overall stroke reduction compared to
alternative; time horizon of at least 10
years/comparator of nil required
AF patients with warfarin
Increased adverse events related to
NZ$101,000 per QALY gained; Not cost-
tolerance
device but may be acceptable given
effective:- WATCHMAN more expensive
warfarin risks
than warfarin alternative and provides no
incremental benefit.
Societal & Ethical Domain
There are no societal and ethical considerations associated with a recommendation not to fund a
procedure with a suitable substitute, thus the decision to not approve its general use in patients
able to take anticoagulation has few ethical risks.
The major ethical consideration is the risk of not funding a potential treatment for patients who are
unable to take anticoagulation. This group will continue to have an increased risk of stroke and will
receive either inferior therapy or no therapy at all.
There is an equally valid ethical risk in approving an expensive treatment in the face of very limited
evidence and which has not been clinically validated in the population of interest. The precedent
set by accepting low-quality evidence could be damaging to the reputation of the NHC.
Feasibility of Adoption Domain
In 2012, there were 241 events involving trans-catheter percutaneous cardiac interventions in New
Zealand for a total estimated cost8 of about NZ$7.5 million with the average cost of an event at
NZ$11,905.
Using New Zealand cost estimates most appropriate for WATCHMAN (NZ$15,000-18,000 per
procedure), as well as the cost of downstream costs (i.e. stroke), it could cost about NZ$67.1
(NZ$20,633 per patient) million more to perform the percutaneous LAA occlusion procedure on the
8
Estimates came from National Minimum Dataset Data Set (NMDS) using DRG code F19Z, trans-catheter percutaneous
cardiac intervention, as well as the national price of NZ$4,614 per costweight.
16
midpoint estimate of 3,250 possible patients compared to warfarin over 10 years (Table 5).
Although the incremental cost of LAA occlusion versus warfarin decreases as the time horizon gets
extended, new cases keep the total costs high. Figure 3 shows the relationship between individual
treatment costs and incremental treatment costs.
Table 5. Cost impact estimates
Costs are in NZ$ millions and are based on the cost estimates in Appendix 3
Year
Incremental
Cumulative
1
$18,320
$6,869,824
2
$18,233
$6,837,258
3
$18,136
$6,801,140
4
$18,035
$6,763,030
5
$17,931
$6,724,131
6
$17,828
$6,685,356
7
$17,726
$6,647,389
8
$17,629
$6,610,727
9
$17,535
$6,575,720
10
$17,447
$6,542,600
Total
$67,057,175
Figure 3. Cost Trends
Because WATCHMAN costs slow down after the initial procedure and warfarin costs increase steadily as stroke and
bleeding risks accumulate, the incremental cost over time between the two decreases.
Cost Trends
30000
25000
20000
15000
10000
5000
0
1
2
3
Warfarin
4
5
WATCHMAN
6
7
8
9
10
Incremental Cost
17
Procedural risks necessitate that tertiary cardiac surgical services also be available to respond to
procedure-related complications which would increase the overall cost of the procedure. This limits
the availability of suitable locations and could put further strain on the demand for cardiac surgery.
So while the procedure itself is performed in a catheterisation laboratory, which already exist in
New Zealand, there would also need to be other tertiary services available in the event of adverse
complications from the procedure. Additionally, there would be an added cost of establishing and
maintaining registries for physicians who perform the procedure as well an additional cost to train
and credential physicians and clinical teams to assess and manage this group of patients and to
perform the procedure safely.
Limitations and Uncertainty
Since the device has not been used publicly in New Zealand, there is uncertainty around the actual
cost of the device and the index admission. In Australia, the cost of the index admission ranges
from AU$6,000 to AU$20,000 yet the costs could be anywhere from NZ$15,000-18,000 per
procedure using New Zealand cost estimates.
There is a scarcity of evidence in the clinical safety and effectiveness domain for the population
unable to take warfarin at all as well as in the cost effectiveness domain. Warfarin is used during
the recovery period to reduce the risk of stroke before the device goes through endotheliasation.
Since WATCHMAN has not been proven to be suitable for patients with contraindications to
warfarin, these patients could suffer from both a risk of procedure-related complications and stroke
concurrently during recovery.
Since WATCHMAN has not been compared to aspirin, it is unknown what the actual reduction in
stroke would be; however, safety has not yet been proven in this population. It is also unrealistic to
assume that the procedure would be performed as an alternative to all AF patients who are able to
tolerate oral anticoagulants since it represents a significant cost increase over current treatment.
Under this scenario, overall stroke incidence would remain constant since the procedure is only
deemed ‘non-inferior’ to warfarin.
Prioritisation
With money and human resources required to fund the use of the procedure, it could detract from
other areas of the health budget. Whilst stroke prevention is a health priority for New Zealand, AF
is only responsible for 10.0% of all ischaemic strokes in New Zealand. The WATCHMAN device is
only intended for use on a subset of AF sufferers, so the potential impact on stroke prevention
would be reduced.
Also, the WATCHMAN device only aims to fix one aspect of stroke risk
(anatomical) and fails to address the underlying, complex nature of stroke. This further reduces its
potential impact on overall stroke rates.
18
The WATCHMAN device could be studied in a New Zealand field evaluation; however, other trials
are currently underway overseas that will address some of the evidence gaps required to assess
the device for use in New Zealand.
19
References
Bartus, K., Bednarek, J., Myc, J., Kapelak, B., Sadowski, J., Lelakowski, J., . . . Lee, R. (2011).
Feasibility of closed-chest ligation of the left atrial appendage in humans. Heart Rhythm,
8(2), 188-193.
DLA Piper. (2013). New and emerging cardiac technologies in Australian and New Zealand public
health services over the nest decade. Melbourne: HealthPact.
Dyall, L., Feigin, V., Brown, P., & Roberts, M. (2008). Stroke: a picture of health disparities in New
Zealand. Social Policy Journal of New Zealand Te Puna Whakaaro(33).
FDA. (2012). FDA Drug Safety Communication: Pradaxa (dabigatran etexilate mesylate) should
not be used in patients with mechanical prosthetic heart valves.
Forel, D. (2010). Horizon scanning technology prioritising summary: watchman left atrial
appendage closure device. Canberra: Commonwealth of Australia.
Go, A. S., Hylek, E. M., Chang, Y., Phillips, K. A., Henault, L. E., Capra, A. M., . . . Singer, D. E.
(2003). Anticoagulation therapy for stroke prevention in atrial fibrillation. JAMA: the journal
of the American Medical Association, 290(20), 2685-2692.
Hart, R. G., Benavente, O., McBride, R., & Pearce, L. A. (1999). Antithrombotic therapy to prevent
stroke in patients with atrial fibrillation: a meta-analysis. Annals of Internal Medicine, 131,
492-501.
Health Technology Inquiry Service. (2010). Left atrial appendage occlusion: cost-effectiveness in a
Canadian setting: Canadian Agency for Drugs and Technologies in Health.
Holmes, D., Reddy, V., Turi, Z., & al., e. (2009). Percutaneous closure of the left atrial appendage
versus warfarin therapy for prevention of stroke in patients with atrial fibrillation; a
randomised non-inferiority trial. Lancet(374), 534-542.
Latif, A. A., & Messinger-Rapport, B. J. (2004). Should nursing home residents with atrial fibrillation
be anticoagulated? Cleveland Clinic journal of medicine, 71(1), 40-44.
Lever, N. A., & Larsen, P. D. (2011). Managing atrial fibrillation: the growing challenge. The New
Zealand Medical Journal, 124(1343), 11-13.
Levy, S., Maarek, M., Coumel, P., & al., e. (1999). Characterisation of different subsets of atrial
fibrillation in general practice in France: the ALFA study Circulation(99), 3028-3035.
Luis, S. A., Roper, D., Incani, A., Poon, K., Haqqani, H., & Walters, D. L. (2012). NonPharmacological therapy for atrial fibrillation: managing the left atrial appendage. [Review
Article]. Cardiology Research and Practice.
Medsafe. (2013). Dabigatran — New Contraindication
Meier, B., Palacios, I., Windecker, S., Rotter, M., Cao, Q. L., Keane, D., . . . Hijazi, Z. M. (2003).
Transcatheter left atrial appendage occlusion with Amplatzer devices to obviate
anticoagulation in patients with atrial fibrillation. Catheterization and cardiovascular
interventions, 60(3), 417-422.
Miyasaka, & al., e. (2006). Secular trends in incidence of atrial fibrillation in Olmstead County,
Minnesota, 1980 to 2000, and implications on the projections for future prevalence.
Circulation, 114, 119-125.
NHC. (2012). Technology Note. Catheter ablation for the treatment of atrial fibrillation.
20
NHSCB. (2012). NHS Commissioning Board Clinical Commissioning Policy Statement: Left Atrial
Appendage (LAA) Occlusion.
Ostermayer, S. H., Reisman, M., Kramer, P. H., Matthews, R. V., Gray, W. A., Block, P. C., . . . Di
Mario, C. (2005). Percutaneous left atrial appendage transcatheter occlusion (PLAATO
system) to prevent stroke in high-risk patients with non-rheumatic atrial fibrillation: results
from the international multi-center feasibility trials. Journal of the American College of
Cardiology, 46(1), 9-14.
Phend, C. (2013). WATCHMAN PREVAILs on Safety for Stroke Prevention in Afib. Medpage
Today. Retrieved from
PR Newswire. (2013). The Boston Scientific Watchman device continues to demonstrate positive
clinical outcomes for patients with atrial fibrillation, CNBC.
Salzberg, S. P., & Tolboom, H. (2011). Management of the left atrial appendage. Multimedia
Manual of Cardio-Thoracic Surgery, 2011(1118).
Stiles, S. (2012). Without warfarin, Watchman still prevents strokes, says registry Paper presented
at the Heart Rhythm Society 2012 Scientific Sessions.
21
Appendix 1: Literature search strategies
Clinical Safety/Effectiveness
Database: Ovid MEDLINE(R) <1946 to November Week 3 2012>
Search Strategy:
-------------------------------------------------------------------------------1 watchman.mp. [mp=title, abstract, original title, name of substance word, subject heading word, protocol
supplementary concept, rare disease supplementary concept, unique identifier] (56)
2 (left atrial appendage adj3 (closure or occlusion)).mp. [mp=title, abstract, original title, name of substance
word, subject heading word, protocol supplementary concept, rare disease supplementary concept, unique
identifier] (172)
3 1 or 2 (213)
4 limit 3 to (clinical trial, all or guideline or meta analysis or randomized controlled trial) (23)
5 3 and systematic review*.mp. [mp=title, abstract, original title, name of substance word, subject heading
word, protocol supplementary concept, rare disease supplementary concept, unique identifier] (3)
6 4 or 5 (26)
Cost Effectiveness
Embase/Medline, adapted for Cochrane Library
Session Results
.......................................................
No. Query Results Results Date
#4. watchman OR 'left atrial appendage occlusion' OR 117 9 Dec 2012
'left atrial appendage exclusion' OR 'left atrial
appendage closure' AND ('cost'/exp OR costs OR
economic* OR 'qaly'/exp OR daly OR 'quality of
life'/exp OR disabilit* OR 'life expectancy' OR
'adverse event' OR 'adverse events' OR 'decision
analysis' OR 'decision analyses' OR 'length of
time' OR 'health outcome' OR 'health outcomes' OR
'stroke risk' OR 'risk of stroke') AND
[2005-2013]/py
#3. watchman OR 'left atrial appendage occlusion' OR 120 9 Dec 2012
'left atrial appendage exclusion' OR 'left atrial
appendage closure' AND ('cost'/exp OR costs OR
economic* OR 'qaly'/exp OR daly OR 'quality of
life'/exp OR disabilit* OR 'life expectancy' OR
'adverse event' OR 'adverse events' OR 'decision
analysis' OR 'decision analyses' OR 'length of
22
time' OR 'health outcome' OR 'health outcomes' OR
'stroke risk' OR 'risk of stroke')
#2. 'cost'/exp OR costs OR economic* OR 'qaly'/exp OR 1,679,033 9 Dec 2012
daly OR 'quality of life'/exp OR disabilit* OR
'life expectancy' OR 'adverse event' OR 'adverse
events' OR 'decision analysis' OR 'decision
analyses' OR 'length of time' OR 'health outcome'
OR 'health outcomes' OR 'stroke risk' OR 'risk of
stroke'
#1. watchman OR 'left atrial appendage occlusion' OR 425 9 Dec 2012
'left atrial appendage exclusion' OR 'left atrial
appendage closure'
.......................................................
*PLAATO/Amplatzer System excluded
23
Appendix 2: USPSTF Grading Criteria and Appraisal Methodology
USPSTF Criteria
Randomized Controlled Trials and Cohort Studies
Criteria
 Initial assembly of comparable groups
 for RCTs: adequate randomization, including first concealment and whether potential
confounders were distributed equally among groups
 for cohort studies: consideration of potential confounders with either restriction or
measurement for adjustment in the analysis; consideration of inception cohorts
 Maintenance of comparable groups (includes attrition, cross-overs, adherence, contamination)
 Important differential loss to follow-up or overall high loss to follow-up
 Measurements: equal, reliable, and valid (includes masking of outcome assessment)
 Clear definition of interventions
 Important outcomes considered
 Analysis: adjustment for potential confounders for cohort studies, or intention to treat analysis for
RCTs.
Definition of ratings based on above criteria
 Good: Meets all criteria: comparable groups are assembled initially and maintained throughout
the study (follow-up at least 80 percent); reliable and valid measurement instruments are used
and applied equally to the groups; interventions are spelled out clearly; important outcomes are
considered; and appropriate attention to confounders in analysis. In addition, for RCTs,
intention to treat analysis is used.
 Fair: Studies will be graded “fair” if any or all of the following problems occur, without the fatal
flaws noted in the “poor” category below: Generally comparable groups are assembled initially
but some question remains whether some (although not major) differences occurred in followup; measurement instruments are acceptable (although not the best) and generally applied
equally; some but not all important outcomes are considered; and some but not all potential
confounders are accounted for. Intention-to-treat analysis is done for RCTS.
Poor: Studies will be graded “poor” if any of the following fatal flaws exists: groups assembled initially are
not close to being comparable or maintained throughout the study; unreliable or invalid measurement
instruments are used or not applied at all equally among groups (including not masking outcome
assessment); and key confounders are given little or no attention. For RCTs, intention-to-treat analysis is
lacking
.
24
GATE Method of RCT appraisal
25
CASP Economic Evaluation Appraisal
26
27
Appendix 3: Decision analytic model tree and cost breakdown
Figure 1: Decision analytic tree
This tree models the comparison between LAA occlusion patients (WATCHMAN) and warfarin patients
28
Table 1: Cost Breakdown
Category
GP Visit
INR Test
Device
Operating Room (1 Hr)
Surgical Staff (1 Hr)
1st Cardiac Consult
1st Anaesthesiologist Consult
Ward Stay (1 Day)
TOE
Warfarin (1 Month)
Cardiac Follow-Up
Procedure Total
Stroke (1st Year)
Stroke (Yrs 2+)
Pericardial Effusion
Bleeding
Cost (lower)
$60.00
$20.00
$8,800.00
$2,100.00
$660.00
$341.00
$341.00
$1,848.00
$571.00
$2.67
$263.00
$15,006.67
Cost (higher)
$11,938.00
$18,144.67
$20,849.00
$3,020.00
$7,392.00
$1,601.00
Source: ADHB’s non-resident price list and clinical consultation.
29
Appendix 4: TGA Approval Classifications
The level of scrutiny by the TGA of a device before it is placed on the ARTG and supplied in
Australia depends on the risk posed by the device. The TGA has adopted a classification system for
devices, based on the level of risk. The lowest risk medical devices, Class 1 devices, are not
assessed by the TGA prior to inclusion on the ARTG.
The TGA medical device classification system
Medical Device Classifications
Examples
Class I
elastic bandages, tongue depressors, cervical collars, slings,
non-sterile dressings
Class IIa
X-ray films, intravenous tubing, contact lenses, catheters
Class IIb
Blood bags, dressings for severe wounds, condoms
Class III
Coronary artery probes, intrauterine contraceptive devices,
medical devices that contain medicines, such as dressings
with an anti-microbial agent
Active implantable medical
devices
Pace makers, cochlear implants
Source: http://www.tga.gov.au/newsroom/devices-basics-regulation.htm
30
National Health Committee (NHC) and Executive
The National Health Committee (NHC) is an independent statutory body which provides advice to
the New Zealand Minister of Health. It was reformed in 2011 to establish evaluation systems that
would provide the New Zealand people and health sector with greater value for the money invested
in health. The NHC Executive are the secretariat that supports the Committee. The NHC
Executive’s primary objective is to provide the Committee with sufficient information for them to
prioritise interventions and make investment and disinvestment decisions. They do this through a
variety of products including Prioritising Summaries, Technology Notes, EpiNotes, CostNotes,
Rapid Reviews, and Health Technology Assessments which are chosen according to the nature of
the decision required and time-frame within which decisions need to be made.
Citation: National Health Committee. 2013. Percutaneous left atrial appendage (LAA) occlusion for
the treatment of atrial fibrillation (AF) to prevent stroke. Wellington: National Health Committee.
Published in March 2013 by the National Health Committee
PO Box 5013, Wellington, New Zealand
This document is available on the National Health Committee’s website:
www.nhc.health.govt.nz
Disclaimer
The information provided in this report is intended to provide general information to clinicians,
health and disability service providers and the public, and is not intended to address specific
circumstances of any particular individual or entity. All reasonable measures have been taken to
ensure the quality and accuracy of the information provided.
If you find any information that you believe may be inaccurate, please email to
[email protected].
The National Health Committee is an independent committee established by the Minister of Health.
The information in this report is the work of the National Health Committee and does not
necessarily represent the views of the Ministry of Health.
The National Health Committee make no warranty, express or implied, nor assumes any legal
liability or responsibility for the accuracy, correctness, completeness or use of any information
provided. Nothing contained in this report shall be relied on as a promise or representation by the
New Zealand government or the National Health Committee.
The contents of this report should not be construed as legal or professional advice and specific
advice from qualified professional people should be sought before undertaking any action following
information in this report.
Any reference to any specific commercial product, process, or service by trade name, trademark,
manufacture, or otherwise does not constitute an endorsement or recommendation by the New
Zealand government or the National Health Committee.
31