Download Doctoral thesis (extended summary)

Hanna Woksepp
II. Woksepp H, Ärlemalm A, Schön T, Carlsson B. Simultaneous measurement of
11 antibiotics using ultra performance liquid chromatography tandem mass
spectrometry. Manuscript.
III. Fridlund J, Woksepp H, Schön T. A microbiological method for determining
serum levels of broad spectrum β-lactam antibiotics in critically ill patients.
J Microbiol Methods 2016 Jul 25;129:23-27. doi: 10.1016/j.mimet.2016.07.020
IV. Woksepp H, Jernberg C, Tärnberg M, Ryberg A, Brolund A, Nordvall M, OlssonLiljequist B, Wisell KT, Monstein HJ, Nilsson LE, Schön T. High-resolution
melting-curve analysis of ligation-mediated real-time PCR for rapid evaluation of
an epidemiological outbreak of extended-spectrum-beta-lactamase-producing
Escherichia coli. J Clin Microbiol. 2011 Dec;49(12):4032-9. doi:
10.1128/JCM.01042-11.
V.
Woksepp H, Ryberg A, Billström H, Hällgren A, Nilsson LE, Marklund BI,
Olsson-Liljequist B, Schön T. Evaluation of high-resolution melting curve analysis
of ligation-mediated real-time PCR, a rapid method for epidemiological typing of
ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae,
Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) pathogens.
J Clin Microbiol. 2014 Dec;52(12):4339-42. doi: 10.1128/JCM.02537-14.
VI. Woksepp H, Ryberg A, Schön T, Söderman J. Epidemiological characterization of
a nosocomial outbreak of extended spectrum β-lactamase Escherichia coli ST-131
confirms the clinical value of core genome multi locus sequence typing. Submitted.
All published papers are reprinted with permission from the respective publisher.
Individualized treatment and control of bacterial infections
Woksepp H, Hällgren A, Borgström S, Kullberg F, Wimmerstedt A, Oscarsson A,
Nordlund P, Lindholm ML, Bonnedahl J, Brudin L, Carlsson B, Schön T. High
target attainment for β-lactam antibiotics in intensive care unit patients when actual
minimum inhibitory concentrations are applied. Eur J Clin Microbiol Infect Dis.
2016 Nov 4. doi: 10.1007/s10096-016-2832-4
Nr 274/2017
Hanna Woksepp
List of papers included in this thesis
I.
Linnaeus University Dissertations
Individualized treatment and
control of bacterial infections
Lnu.se
isbn: 978-91-88357-58-8
linnaeus university press
Individualized treatment and control
of bacterial infections
Linnaeus University Dissertations
No 276/2017
I NDIVIDUALIZED TREATMENT AND
CONTROL OF BACTERIAL INFECTIONS
H ANNA W OKSEPP
LINNAEUS UNIVERSITY PRESS
Individualized treatment and control of bacterial infections
Doctoral dissertation, Department of Department of Medicine and
Optometry, Linnaeus University, Kalmar, 2017
ISBN: 978-91-88357-58-8
Published by: Linnaeus University Press, 351 95 Växjö
Printed by: DanagårdLiTHO AB, 2017
Abstract
Woksepp, Hanna (2017). Individualized treatment and control of bacterial infections,
Linnaeus University Dissertation No 276/2017, ISBN: 978-91-88357-58-8.
Written in English.
Infectious diseases cause substantial morbidity and mortality, exacerbated by
increasing antibiotic resistance. In critically ill patients, recent studies indicate
a substantial variability in β-lactam antibiotic levels when standardized dosing
is applied. New methods for characterizing nosocomial outbreaks of bacterial
infections are needed to limit transmission. The goals of this thesis were to
investigate new strategies towards individualized treatment and control of
bacterial infections.
In Paper I we confirmed high variability in β-lactam antibiotic levels among
intensive care unit (ICU) patients from southeastern Sweden, where 45 %
failed to reach treatment targets (100 % fT>MIC). Augmented renal clearance
and establishing the minimum inhibitory concentration of the bacteria were
important for evaluating the risk of not attaining adequate drug levels. In
Paper II a rapid ultra-performance liquid chromatography tandem mass
spectrometry (UPLC-MS/MS) method for simultaneous quantification of 11
commonly used antibiotics was developed and tested in clinical samples.
Performance goals (CV<15%) were reached. A microbiological method for
quantification of β-lactam antibiotics in serum was developed in Paper III.
The method could be important for hospitals without access to an LC-MS
method. Paper IV and Paper V investigated ligation-mediated qPCR with
high resolution melt analysis (LMqPCR HRMA), for transmission
investigation of extended spectrum β-lactamase (ESBL)-producing E. coli and
other common bacterial pathogens. Results comparable to the reference
method (PFGE) could be achieved within one day in a closed system and
confirmed a nosocomial outbreak in Kalmar County. In Paper VI whole
genome sequencing followed by bioinformatic analysis resolved transmission
links within a nosocomial outbreak due to improved discriminatory power
compared to LMqPCR HRMA.
The high proportion of ICU patients with insufficient β-lactam drug levels
emphasizes the need for individualized treatment by therapeutic drug
monitoring (TDM). TDM is enabled by a highly sensitive method, such as
UPLC-MS/MS, but if unavailable, also by a microbial method. Molecular
typing methods used for transmission investigation can detect nosocomial
outbreaks. LMqPCR HRMA can be used for screening purposes. For
enhanced resolution, whole genome sequencing should be used, but always
together with a rigorous epidemiological investigation.
Keywords: therapeutic drug monitoring, intensive care unit, β-lactam
antibiotics, antimicrobial drug resistance, molecular typing, whole-genome
sequencing
”The most difficult thing is the decision to act.
The rest is merely tenacity”
Amelia Earhart
Populärvetenskaplig sammanfattning
Allvarliga bakteriella infektioner orsakar i många fall ökad mortalitet och
morbiditet och utgör tillsammans med ökad antibiotikaresistens ett stort
hälsoproblem. WHO antog år 2015 en global åtgärdsplan med målet att minska
infektionsspridning och att säkerställa effektiv infektionsbehandling. Adekvat
och tidig infektionsbehandling med bredspektrumantibiotika, såsom βlaktamantibiotika, samt en snabb identifiering av bakteriespridning är
avgörande för behandlingsframgång och för att minska bakteriella utbrott.
Målet med denna avhandling var att undersöka nuvarande empirisk
infektionsbehandling och att utveckla metoder som möjliggör individuell
behandling med β-laktamantibiotika samt att utveckla metoder för att snabbt
upptäcka och följa smittspridning.
Upptäckterna som beskrivs i denna avhandling är uppdelade i sex
delarbeten; ett som beskriver empirisk antibiotikabehandling av
intensivvårdade (IVA) patienter, två arbeten som beskriver två olika metoder
för att mäta antibiotikanivåer hos patienter och tre arbeten som fokuserar på
olika tekniker för att upptäcka och följa bakteriers spridning mellan patienter.
I det första arbetet visade vi att nivåerna av β-laktamantibiotika hos IVApatienter varierar mycket trots att de fått samma dos per kilo kroppsvikt. Det
var dessutom relativt vanligt att IVA-patienter inte hade tillräcklig mängd
antibiotika i förhållande till sin infektion. Frekvensen patienter med låga
koncentrationer var högre hos dem med bakterier med sämre känslighetsnivå,
ökad njurfunktion samt de som hade en lägre sjukdomsgrad. I arbete två
presenteras en snabb metod för samtidig koncentrationsmätning av 11 olika
antibiotika. Metoden hade hög känslighet och specificitet men kräver dock dyr
och avancerad utrustning. För att fullt ut validera metoden krävs ett externt
kontrollprogram för samtliga antibiotika. I det tredje arbetet utvecklades en
mikrobiologisk metod för nivåbestämning av tre olika β-laktamantibiotika;
cefotaxim, meropenem och piperacillin. Metoden är inte lika snabb och känslig
som den i arbete två, men kräver ingen särskild utrustning utöver den som redan
finns på ett mikrobiologisk lab. Den kan därför användas när dyr och avancerad
utrustning inte är ett alternativ.
1
I det fjärde och femte arbetet utvecklades en snabb metod för att jämföra om
två eller fler bakterieisolat är olika. Metoden är molekylärbiologisk och jämför
bakteriernas genom (DNA). Resultat jämförbara med tidigare använd
standardmetod erhölls för bakterier som sprids lätt i sjukhusmiljö, såsom
Enterobacteriaceae. Inte heller denna metod kräver annan utrustning än den
som redan finns på ett mikrobiologiskt laboratorium och fungerar därför som
ett alternativ till standardmetoden. I det sjätte arbetet jämfördes metoden från
arbete fyra och fem med en ny metod, som bygger på DNA-sekvensering av
hela bakteriegenom. Med sekvensering erhölls bäst kapacitet för att identifiera
och följa spridning av bakterier.
Som helhet har de olika delarbetena bidragit till att visa på ett behov av att
kontrollera så att svårt sjuka patienter verkligen har den mängd antibiotika som
krävs för att bota infektionen. Denna kontroll kan göras med både avancerade
och enklare metoder, anpassat efter laboratoriets förutsättningar. Kontroll av
bakteriespridning mellan patienter kan också göras med enklare metoder men
för att identifiera och följa spridning av bakterier krävs mer avancerade metoder
där DNA-sekvensering ger bäst upplösning. Det är dessutom viktigt att
bakteriespridning undersöks tillsammans med kliniska uppgifter.
2
List of papers
This thesis is based on the following papers, referred to in the text by their
Roman numerals. All published papers are reprinted with permission from the
respective publisher.
I.
Woksepp H, Hällgren A, Borgström S, Kullberg F, Wimmerstedt A,
Oscarsson A, Nordlund P, Lindholm ML, Bonnedahl J, Brudin L,
Carlsson B, Schön T. High target attainment for β-lactam antibiotics in
intensive care unit patients when actual minimum inhibitory
concentrations are applied. Eur J Clin Microbiol Infect Dis. 2016 Nov
4. doi: 10.1007/s10096-016-2832-4
II.
Woksepp H, Ärlemalm A, Schön T, Carlsson B. Simultaneous
measurement of 11 antibiotics using ultra performance liquid
chromatography tandem mass spectrometry. Manuscript.
III.
Fridlund J, Woksepp H, Schön T. A microbiological method for
determining serum levels of broad spectrum β-lactam antibiotics in
critically ill patients. J Microbiol Methods 2016 Jul 25;129:23-27. doi:
10.1016/j.mimet.2016.07.020
IV.
Woksepp H, Jernberg C, Tärnberg M, Ryberg A, Brolund A, Nordvall
M, Olsson-Liljequist B, Wisell KT, Monstein HJ, Nilsson LE, Schön
T. High-resolution melting-curve analysis of ligation-mediated realtime PCR for rapid evaluation of an epidemiological outbreak of
extended-spectrum-beta-lactamase-producing Escherichia coli. J Clin
Microbiol. 2011 Dec;49(12):4032-9. doi: 10.1128/JCM.01042-11.
V.
Woksepp H, Ryberg A, Billström H, Hällgren A, Nilsson LE,
Marklund BI, Olsson-Liljequist B, Schön T. Evaluation of highresolution melting curve analysis of ligation-mediated real-time PCR,
a rapid method for epidemiological typing of ESKAPE (Enterococcus
faecium,
Staphylococcus
aureus,
Klebsiella
pneumoniae,
Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter
species) pathogens. J Clin Microbiol. 2014 Dec;52(12):4339-42. doi:
10.1128/JCM.02537-14.
VI.
Woksepp H, Ryberg A, Schön T, Söderman J. Epidemiological
characterization of a nosocomial outbreak of extended spectrum βlactamase Escherichia coli ST-131 confirms the clinical value of core
genome multi locus sequence typing. Submitted.
3
Additional work executed during the Ph.D. study but not included in this
thesis:
 Larsson MC, Karlsson E, Woksepp H, Frölander K, Mårtensson A,
Rashed F, Annika W, Schön T, Serrander L. Rapid identification of
pneumococci, enterococci, beta-haemolytic streptococci and S. aureus
from positive blood cultures enabling early reports. BMC Infect Dis.
2014 Mar 19;14:146. doi: 10.1186/1471-2334-14-146.
 Idh J, Andersson B, Lerm M, Raffetseder J, Eklund D, Woksepp H,
Larsson M, Forslund T, Werngren J, Mansjö M, Sundqvist T, Stendahl
O, Schön T. Reduced susceptibility of clinical strains of
Mycobacterium tuberculosis to reactive nitrogen species in activated
macrophages correlates with in vitro efficacy against pretomanid.
Manuscript.
Financial support
Financial support was provided by the research council of Southeast Sweden
(FORSS), the Swedish Heart and Lung Foundation (King Oscar II Jubilee
Foundation), the Marianne and Marcus Wallenberg foundation (2012-0119),
and the research council of Kalmar County.
4
Table of contents
Populärvetenskaplig sammanfattning ................................................................ 1
List of papers ..................................................................................................... 3
Table of contents ............................................................................................... 5
Abbreviations and glossary ............................................................................... 7
Introduction ....................................................................................................... 9
Individualized treatment of bacterial infections .............................................. 11
The intensive care unit ............................................................................... 11
Hospital-acquired infections and bacterial infections in the intensive care
unit ............................................................................................................. 11
Sepsis .................................................................................................... 13
Treatment of bacterial infections in the ICU .............................................. 15
Pharmacokinetics and Pharmacodynamics ........................................... 18
Therapeutic drug monitoring of antibiotics in the ICU .............................. 19
Antibiotics and methods to measure serum drug concentrations .................... 21
Antibiotics .................................................................................................. 21
β-lactam antibiotics ............................................................................... 21
Fluoroquinolones .................................................................................. 22
Oxazolidinones ..................................................................................... 22
Lincosamides ........................................................................................ 22
Ansamycins .......................................................................................... 23
Methods for antibiotic quantification ................................................... 23
Ultra-performance liquid chromatography tandem mass spectrometry 24
Antimicrobial drug susceptibility testing .............................................. 26
Transmission investigation by molecular typing methods .............................. 28
Control of nosocomial outbreaks ............................................................... 28
Especially challenging bacteria ............................................................ 28
Extended-spectrum β-lactamases.......................................................... 29
Methods for bacterial outbreak investigation ............................................. 30
Pulsed field gel electrophoresis ............................................................ 33
Matrix-assisted laser desorption/ionization time of flight .................... 34
Whole genome sequencing ................................................................... 36
Aims of the study ............................................................................................ 39
Materials and methods..................................................................................... 40
Measurement of β-lactam antibiotic concentrations in critically ill
patients ....................................................................................................... 40
Study population ................................................................................... 40
Clinical follow-up ................................................................................. 40
Bacterial isolates and MIC determination............................................. 40
Blood samples....................................................................................... 41
5
Measurement of serum antibiotic concentrations with liquid
chromatography mass spectrometry ..................................................... 41
Development of an ultra-performance liquid chromatography tandem mass
spectrometry method quantifying 11 antibiotics ........................................ 41
Chemicals and reagents ........................................................................ 41
Preparation of calibrators, quality control samples and internal standard
solutions................................................................................................ 42
Sample preparation and extraction ....................................................... 42
Instruments and analytical conditions .................................................. 43
A laboratory comparison of quantification of commonly used
antibiotics ............................................................................................. 44
Development of a microbiological method for quantification of β-lactam
antibiotics concentration ............................................................................ 44
Methods for bacterial outbreak investigation............................................. 46
Study population and bacterial isolates ................................................ 46
Pulse field gel electrophoresis .............................................................. 46
Ligation mediated qPCR high resolution melt analysis........................ 47
Matrix-assisted laser desorption/ionization time of flight .................... 48
Whole genome sequencing ................................................................... 48
Statistical analysis ...................................................................................... 49
Results and Discussion .................................................................................... 50
Managing treatment with β-lactam antibiotics in the ICU ......................... 50
Methods for measurements of β-lactam antibiotics ................................... 54
Development of an ultra-performance liquid chromatography tandem
mass spectrometry method quantifying 11 antibiotics ......................... 54
Development of a microbiological method for quantification of βlactam antibiotics concentration ........................................................... 56
Concluding remarks – methods for measurements of β-lactam
antibiotics ............................................................................................. 57
Strategies for transmission investigation of bacteria ................................. 58
Ligation mediated qPCR high resolution melt analysis........................ 58
Matrix-assisted laser desorption/ionization time of flight .................... 60
Whole genome sequencing ................................................................... 65
Concluding remarks - strategies for transmission investigation of
bacteria ................................................................................................. 67
Conclusions ..................................................................................................... 69
Future perspectives .......................................................................................... 70
Acknowledgments ........................................................................................... 72
References ....................................................................................................... 74
6
Abbreviations and glossary
ACN
ARC
AST
CAI
CAUTI
CI
CL
CLABSI
cg
CTX
ds
ECDC
EMA
ESKAPE
EUCAST
FA
FQ
HAI
HPLC
HRMA
IB
ICU
LC-MSMS
LOS
i.v.
MALDI TOF
MER
MeOH
MIC
MIRU-VNTR
MLST
Acetonitrile
Augmented renal clearance
Antimicrobial drug susceptibility testing
Community acquired infection
Catheter-associated urinary tract infection
Continuous infusion
Clearance
Central-line associated blood-stream infection
Core genome
Cefotaxime
Double stranded
European Centre for Disease Prevention and
Control
European Medicines Agency
Enterococcus faecium, Staphyloccocus aureus,
Klebsiella pneumoniae, Acinetobacter
baumannii, Pseudomonas aeruginosa,
Enterobacter spp.
European Committee on Antimicrobial
Susceptibility Testing
Formic acid
Fluoroquinolones
Hospital acquired infections
High pressure liquid chromatography
High resolution melt analysis
Intermittent bolus dosing
Intensive care unit
Liquid chromatography tandem mass
spectrometry
Length of stay
Intravenous
Matrix-assisted laser desorption/ionization time
of flight
Meropenem
Methanol
Minimum inhibitory concentration
Mycobacterial interspersed repetitive unitsvariable number of tandem repeats
Multi locus sequence type
7
Molecular typing
MPS
MRSA
MSP
Mtb
m/z
NHSN
NGS
PBPs
PD
PFGE
PIP
PIP/taz
PK
RCT
SIRS
SNP
SOFA
ss
TB
TD
TDM
Transmission investigation
Typeability
UPLC
VAP
Vd
VRE
WGS
qPCR
8
Using molecular biology techniques to identify,
classify and compare organisms and their
subtypes
Massive parallel sequencing
Methicillin resistant S. aureus
Main spectra
Mycobacterium tuberculosis
Mass to charge ratio
National Healthcare Safety Network
Next generation sequencing
Penicillin-binding proteins
Pharmacodynamics
Pulsed field gel electrophoresis
Piperacillin
Piperacillin/tazobactam
Pharmacokinetics
Randomized control trial
Systemic inflammatory response syndrome
Single nucleotide polymorphism
Sequential organ failure assessment
Single stranded
Tuberculosis
Denaturation temperature
Therapeutic drug monitoring
Combining molecular typing results and
epidemiological data to analyze bacterial
spread between patients.
Ability of a typing method to clearly
differentiate isolates
Ultra-performance liquid chromatography
Ventilator-associated pneumonia
Volume of distribution
Vancomycin-resistant enterococci
Whole genome sequencing
quantitative Real time polymerase chain
reaction
Introduction
Infectious bacterial diseases are a major cause of morbidity and mortality [3,
8], and are, along with increasing antibiotic resistance a critical health issue [9].
The WHO has adopted a global action plan with five major objectives aimed to
ensure effective and safe treatment, and prevention of infectious diseases [10].
These objectives include i) improvement of the knowledge of antibiotic
resistance through education, ii) surveillance and research, iii) reduction of
infection rates via preventive measures, iv) optimization of treatment and
development of routines, and v) incentives ensuring economic sustainability for
investments in new medicines and other infection interventions [10]. The
ministry of health and social affairs in Sweden has developed a national strategy
to fulfill the global plan of the WHO [11].
To avoid further spread of infectious diseases and antibiotic resistance, all
sections involved in bacterial transmission (Fig 1.) require adequate routines
and guidelines involving vaccination, diagnostics, hygiene, agriculture and
surveillance [10].
Figure 1. Routes of transmission. By using antibiotics, resistant bacteria are
selected for and maintained. To reduce the selective pressure, antibiotics should
only be used to treat infections. From “Antibiotic/Antimicrobial Resistance”, by
Melissa Brower, Centers for Disease Control and Prevention, 2015. In the
public domain. Date of download 11/21/2016. Reprinted with permission.
9
In the health care system, infections may be divided into community
acquired infections (CAI), and hospital acquired infections (HAI), where the
latter are associated with more resistant bacteria [12], increased mortality, and
higher health care costs [13]. HAI are defined as infections occurring later than
48 hours during the course of hospitalization [12].
Severe sepsis and septic shock caused by bacterial infections, affects
millions world-wide every year, with a high mortality rate (>30 %) [8, 14]. In
Sweden, the incidence of severe sepsis in 2010 was 10-45/100 000 inhabitants
[3]. It is possible that this number is highly underestimated due to diagnostic
difficulties, and it has been argued that the incidence of severe sepsis in Sweden
is at least 200/100 000 inhabitants, and the incidence of septic shock >30/100
000 [3]. Early and adequate empirical treatment of severe bacterial infections
with broad spectrum antibiotics as well as early identification of nosocomial
outbreaks is crucial for favorable treatment outcomes [8, 15, 16].
It has been shown that timely administration of antibiotics may not be
enough to ensure a favorable outcome, especially in critically ill patients [17].
This is mainly due to an unpredictable variability in antibiotic concentrations.
Additionally, the clinical efficacy and dosing of antibiotics are rarely tested on
critically ill patients, and dosing regimens are generally developed in a “onesize fits all” manner [18]. Critically ill patients are more likely to become
infected by resistant bacteria, because of the environment in the ICU [19, 20].
Such pathogens include the ESKAPE group (Enterococcus faecium,
Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii,
Pseudomonas aeruginosa and Enterobacter spp.), which is particularly
associated with drug resistance and unfavorable outcomes [21, 22]. When
nosocomial spread of these pathogens occurs, it is essential to quickly identify
the source to limit further spread [22, 23]. Epidemiological typing of the
ESKAPE pathogens has historically been carried out by several different
methods, depending on the species and available equipment. The most
commonly used methods have been multi locus sequence typing (MLST) [24],
and pulsed field gel electrophoresis (PFGE) [25], with separate protocols for
each species. Recent developments of massive parallel sequencing (MPS), and
whole genome sequencing (WGS), have now gradually replaced the earlier
molecular methods [26].
Bearing these challenges, and efforts in controlling infectious diseases and
antibiotic resistance in mind, current treatment and control strategies for
bacterial infections in critically ill patients would benefit from optimization, as
would methods used for surveillance of nosocomial spread, enabling more rapid
interventions.
10
Individualized treatment of bacterial
infections
The intensive care unit
The most critically ill patients with bacterial infections are admitted to an
intensive care unit (ICU). Additionally, and because of high antibiotic pressure
and underlying medical and surgical conditions, patients on an ICU are prone
to develop new infections during their stay. In international studies, more than
20 % of patients were estimated to be admitted to the ICU due to infection [14,
27, 28]. Patients with infections have a longer length of stay (LOS), and an
increased mortality risk [28, 29].
Early and correct assessment of critically ill patients is important in
preventing mortality [3, 8]. For patients with severe infections in the ICU, there
are Swedish national guidelines and international guidelines to ensure proper
assessment [3, 8]. The risk factors for developing bacterial infections in
critically ill patients include comorbidity, immunosuppressive treatment, and
malnutrition, as well as intervention-related factors including invasive devices,
antibiotic exposure, and intensive care-related factors including occupancy,
staff, and inter-hospital transfers [30].
Hospital-acquired infections and bacterial
infections in the intensive care unit
Of all device-associated HAI reported to the National Healthcare Safety
Network (NHSN) among inpatients in the US from January 2006 through
October 2007, 86 % were observed in the ICUs [31]. In a study on ICU patients
in Sweden, several risk factors for developing HAI were identified such as
antibiotic treatment, central-line catheter, urinary catheter, assisted ventilation,
surgery and immune suppression [32]. The rate of HAI in patients with one risk
factor was 15 %, and with several risk factors 22 % [32]. Of those patients
receiving assisted ventilation, the rate of HAI was 27 % [32]. The total number
of reported HAI to NHSN among inpatients in the USA from January 2006
through October 2007 was 28 502, of which 35.3 % were central-line associated
bloodstream infections (CLABSI), 30.1 % were catheter-associated urinary
tract infections (CAUTI), 15.9 % were ventilator-associated pneumonia (VAP),
and 18.6 % were surgical site infections [31]. Nine pathogens accounted for 84
% of all HAI; coagulase-negative Staphylococcus, S. aureus, Enterococcus
spp., Candida spp., E. coli, P. aeruginosa, Klebsiella spp., Enterobacter spp.
and A. baumannii [31]. In Sweden, the rate of HAI in inpatient care has been
estimated by point prevalence measurements on a fixed day and week, twice
11
yearly [32]. In 2014, a total of 38 555 health care events were reviewed, and the
rate of HAI was 5.2 % [32]. Since the start in 2009, the rate has been
approximately 9 % [32]. The most common HAI were VAP, CAUTI, and soft
tissue and skin infections [32].
In an international study (EPIC II) including 75 countries, 1265 ICUs (of
which 22 were Swedish) and 13 796 patients, investigating the prevalence and
outcomes of infection in the ICU, 51 % of patients had an infection, while 71 %
received antibiotics [28], and 70 % had positive cultures [28]. Respiratory tract
infections were the most common, followed by abdominal and bloodstream
infections [28]. The most common isolated organisms were E. coli (16 %), S.
aureus (20 %), and P. aeruginosa (20 %) [28]. The ICU mortality was
approximately twice as high in patients diagnosed with an infection compared
to the remaining patients (25 % versus 11 %, respectively) [28]. In a
multivariate analysis, infections caused by Pseudomonas, Enterococcus or
Acinetobacter spp. were independently associated with increased hospital
mortality [28].
The mortality rate is higher for patients admitted to ICUs in countries with
high levels of antibiotic resistance compared to countries with low levels (Fig.
2) [33]. Bacterial isolates from patients in the ICUs and surgical units are more
resistant to antibiotics than isolates from other hospital wards [20, 34].
Figure 2. Resistance levels in the world for E. coli and K. pneumoniae to
3rd generation cephalosporins. From http://resistancemap.cddep.org/,
created and downloaded 11/21/2016.
12
Sepsis
Sepsis is a syndrome of various manifestations with a dysregulated host
response to pathogens [2, 30]. Septic shock is defined as severe sepsis with
hypotension, unresponsiveness to administration of fluid, together with signs of
hypoperfusion or organ dysfunction [3]. Recently, a new definition has been
established including only sepsis and septic shock [2]. Sepsis is now
internationally defined as a life-threatening organ dysfunction, identified by a
change in total sequential organ failure assessment (SOFA) score by ≥2 points,
caused by a dysregulated host response to infection (Fig. 3) [2]. Septic shock is
sepsis including persisting hypotension despite fluid resuscitation and a serum
lactate level >2 mmol/L (Fig. 3) [2].
Figure 3. Operationalization of clinical criteria identifying patients with sepsis and septic
shock. The baseline Sequential [Sepsis-related] Organ Failure Assessment (SOFA) score
should be assumed to be zero unless the patient is known to have preexisting (acute or
chronic) organ dysfunction before the onset of infection. qSOFA indicates quick SOFA;
MAP, mean arterial pressure, [2]. Date of download 11/21/2016. Reprinted with
permission.
Historically, the definition of sepsis and related syndromes has changed [3].
In 1992 the term systemic inflammatory response syndrome was established
(SIRS), and sepsis was redefined [35]. In the Swedish national guidelines from
2015, severe sepsis is defined as probable infection together with SIRS
alternatively verified infection, with either hypotension, hypoperfusion or organ
dysfunction [3].
13
The host response to bacterial pathogenic factors includes the innate immune
system triggered in response to bacterial endotoxins and exotoxins. This
increases type I interferons and pro-inflammatory cytokines, such as TNF-α,
interleukin (IL)-1 and IL-6, causing an inflammatory response to the infection
[30]. If the inflammatory response is extensive, there is a risk for developing
sepsis. The pathogenesis of sepsis (Fig. 4) includes tissue edema, disseminated
intravascular coagulation, acute respiratory distress syndrome, breakdown of
the endothelial and epithelial barriers due to apoptosis, causing widespread
organ dysfunction and acute kidney injury [30].
There are several studies on biomarkers in sepsis. The most commonly used
in clinical practice are C-reactive protein (CRP) and procalcitonin (PCT) [36].
Figure 4. During sepsis, many types of cells display enhanced apoptosis, leading to various
deleterious consequences. AKI, acute kidney injury; ALI/ARDS, acute lung injury/acute
respiratory distress syndrome; NK, natural killer. Figure is from [5], open access at
openi.nlm.nih.gov. Date of download 11/21/16. Reprinted with permission.
14
The most significant and common pathophysiological changes in critically
ill patients are those that influence the extracellular fluid volume and clearance
pathways (renal and liver) [37]. It has been shown that augmented renal
clearance (ARC) is present in more than 60 % of ICU patients, even if mean
serum creatinine concentrations ranged between 55-79 µmol/L [38]. Other
pathophysiological changes include increased capillary leakage,
hypoalbuminaemia and extracellular compartment expansion that increase the
volume of distribution (VD), and thus decrease the serum level of available
antibiotics [37]. Hypoalbuminaemia may also result in increased availability of
antibiotics, depending on the level of protein binding [39]. Increased cardiac
output and other hyperdynamic changes further decrease the antibiotic levels
due to increased clearance (CL) [37]. On the other hand, the CL can be reduced
due to renal failure [37]. The net sum, of all these pathophysiological changes
in patients suffering from severe bacterial infections in the ICU, determines the
risk for insufficient levels of β-lactam antibiotics. Therefore, these patients
would benefit from individual dosing controlled by therapeutic drug monitoring
(TDM) [42, 53].
Treatment of bacterial infections in the ICU
According to the Swedish national and international guidelines, when
treating critically ill patients with septic shock, close contact with an infectious
disease consultant is recommended, as well as adequate measures to ensure
source control [3, 8]. To achieve adequate empirical treatment, rapid
intravenous (i.v.) antibiotic administration is important, since inadequate
treatment is associated with increased mortality and clinical failure [8, 40, 41].
The choice of treatment is based on the suspected site of infection, local
resistance epidemiology and severity of infection, other diseases, comorbidities,
and recent treatment with antibiotics and epidemiology [3, 8], (Fig 5). The first
choice of treatment in the ICU should be a broad-spectrum agent, ensuring
adequate coverage for potentially causative organisms [8, 42]. Broad-spectrum
therapy should be maintained until the susceptibility pattern of the causative
pathogen is provided, when de-escalation is possible [8]. Treatment time should
be kept as short as possible, and each patient should be assessed individually
[8]. Patients with septic shock may benefit from combination therapy using βlactam antibiotics together with an aminoglycoside, a fluoroquinolone, a
macrolide, or clindamycin [43].
Treatment of critically ill patients with broad-spectrum antibiotics is usually
done according to conventional dosing schemes, which has recently been shown
to result in about 1/3 of patients achieving suboptimal levels, at least for βlactam antibiotics [44-46]. Swedish national and international guidelines for
treatment of critically ill patients recommend higher initial doses than of noncritical patients, for hydrophilic drugs, i.e. β-lactam antibiotics [3, 8].
15
Furthermore, an additional dose given after 50 % of the dosing interval is also
recommended for β-lactam antibiotics [3]. Regarding continuous infusion (CI)
versus intermittent bolus (IB) dosing, a recent systematic review from Sweden
concluded that, to date, scientific evidence is lacking for CI [47]. Other reviews
also question the benefits of CI but conclude its usefulness in severe respiratory
infections caused by pathogens with high MICs [48, 49]. However, this has in
turn been questioned in other studies [50], and it is clearly difficult to design a
randomized controlled trial in ICU patients, with outcomes such as mortality,
to determine the clinical benefit of CI.
16
Figure 5. Algorithm for antibiotic choice in the ICU. Figure adapted by Woksepp
from [3] with permission.
17
Pharmacokinetics and Pharmacodynamics
Early and adequate treatment of critically ill patients improves outcome [51].
Pharmacokinetics (PK) describes the absorption, distribution and elimination of
a drug, and hence the active concentration, whilst pharmacodynamics (PD)
describes the relationship between the concentration and effect of the drug,
which is commonly represented by the MIC for antibiotics [52], (Fig 6).
Figure 6. Pharmacokinetic and pharmacodynamic processes describing the effect of a
drug. After dosage, the drug is processed by the body. The concentration at the site of
infection exerts the anti-infective effects.
Antibiotics have either concentration-dependent killing such as
aminoglycosides and fluoroquinolones, or time-dependent killing such as βlactam antibiotics [53] (Fig 7). Therefore, it is both the type of antibiotic and
minimum inhibitory concentration (MIC) of the pathogen that affects which
PK/PD target that should be chosen to achieve treatment efficacy [52]. The
percentage of free concentration varies between different antibiotics, as well as
between diverse patient groups [39]. The site of infection should also be taken
into consideration, since it is the active free fraction of the drug at the infection
site that can eliminate the pathogen [52, 54].
18
Figure 7. Illustration of the relationship between antibiotic concentration and the minimum
inhibitory concentration (MIC) of the pathogen. AUC=area under the curve;
AUC/MIC=the ratio of the AUC to the time above MIC needed to inhibit microorganisms;
Cmax=maximum serum concentration needed to inhibit microorganisms; Cmax/MIC=
ratio of Cmax to the time above MIC. Figure is from [7]. Open access at openi.nlm.nih.gov
Date of download 01/02/2017. Reprinted with permission.
Therapeutic drug monitoring of antibiotics in the
ICU
Therapeutic drug monitoring of antibiotics can be used to ensure sufficient
drug concentrations in each individual patient. The concept of TDM in
infectious diseases means that the serum antibiotic concentration of the patient
is measured and correlated to the MIC of the pathogen. For β-lactam antibiotics
– time over MIC correlates to efficacy, whereas the efficacy of drugs such as
aminoglycosides and fluoroquinolones correlates to Cmax/MIC [52].
A complicating factor in TDM for β-lactam antibiotics in the ICU is the lack
of a universally accepted PK/PD target. The PK/PD target for β-lactam
antibiotics is under discussion, whereby early animal studies have shown that
40-70 % fT>MIC may be sufficient for clinical cure [52, 55], whereas more recent
retrospective clinical studies propose concentrations of up to 100 % fT>4xMIC,
especially in immunocompromised patients, and to prevent the development of
19
resistance [56, 57]. Most authorities recommend 100 % fT>MIC for ICU patients
[46, 52, 58]. Furthermore, the actual MIC may not be available due to
difficulties in isolating the pathogen. Instead, the target is most often based on
the epidemiological cut-off value (ECOFF) [46, 58], which is the “worst-case”
MIC among susceptible bacteria affecting patients, such as 2mg/L for
meropenem and P. aeruginosa.
In clinical practice, 100 % fT>MIC is often used for TDM [59]. Practically,
the use of TDM-guided dosing means that if the free concentration is above the
set target no dose adjustment would be necessary. On the other hand, if the
concentration is lower than the target, the patient needs dose adjustments, either
by increasing the frequency, dosage, or both.
International studies investigating the empirical treatment of fixed dosing
strategies have found that a substantial portion of the patients are not reaching
adequate concentrations to ensure effective treatment [46, 60, 61]. In the study
Defining Antibiotic Levels in Intensive care unit patients (DALI), 40 % of
patients receiving empirical treatment with β-lactam antibiotics (amoxcillin,
ampicillin, ceftazolin, cefepime, ceftriaxone, doripenem, meropenem, and
piperacillin) did not reach 100 % fT>MIC, which was associated with an impaired
clinical outcome [46]. When using a higher target (100 % fT>4xMIC), 65 % of
patients failed to reach the target [46]. Although all patients were empirically
treated with fixed dosing strategies, a high variation in antibiotic serum
concentration between patients was observed [46, 58]. Approximately the same
rate of target non-attainment (68 % and 71 %), as in the DALI study, was
observed in the control groups of two studies investigating the effect of TDMbased dose optimization, and CI versus IB dosing, respectively [60, 61]. The
study investigating TDM-based dose optimization included four hospitals in
Australia and one in Hong Kong, in total 41 patients [60]. In that study, TDMbased dosing increased target attainment of piperacillin, meropenem, and
ticarcillin [60]. A recent Swedish systematic review concludes that
measurements of β-lactam antibiotic concentrations may be valuable in ICU
patients [47]. The effect of TDM on clinical outcomes in terms of mortality has
not as yet been investigated in randomized controlled trials.
20
Antibiotics and methods to measure
serum drug concentrations
Antibiotics
There are several groups of antibiotics used in clinical practice which
include β-lactam antibiotics, fluoroquinolones, oxazolidinones, lincosamides,
ansamycins, sulphonamides, polypeptides, aminoglycosides, tetracyclines, and
macrolides [62].
β-lactam antibiotics
β-lactam antibiotics are bactericidal drugs that inhibit transpeptidation of the
peptidoglycan layer, thereby disrupting the bacterial cell-wall, causing lysis
[63]. The bacterial proteins involved in the transpeptidation, penicillin-binding
proteins (PBPs), are present in both Gram-negative and Gram-positive bacteria
[63]. The PBPs are divided into two main categories based on molecular mass,
and to subclasses, depending on structure and catalytic activity [63, 64].
Enterococcus spp. have a low intrinsic susceptibility to some β-lactam
antibiotics due to the presence of a low-affinity PBP, capable of taking over
transpeptidation when other PBPs are inhibited [65].
All β-lactam antibiotics contain a β-lactam ring, [66]. β-lactam antibiotics
include penicillins, cephalosporins, carbapenems, and monocyclic β-lactams.
Penicillins are divided into categories based on their structure and activity
[63, 67]. This group has a wide range of activity depending on the type of
penicillin, ranging from Streptococcus pyogenes to Gram-negative bacilli.
Penicillins such as amoxicillin and piperacillin may be combined with a βlactamase inhibitor to achieve activity against β-lactamase-producing
organisms. Piperacillin is a piperazine penicillin, and has, combined with the βlactamase inhibitor tazobactam, activity against many Enterobacteriaceae spp.
and P. aeruginosa [68].
Cephalosporins are divided into generations and cefotaxime, a third
generation cephalosporin, has activity against Staphylococcus spp.,
Streptococcus spp. and many Enterobacteriaceae spp., but lacks activity against
Enterococcus spp., and P. aeruginosa [69].
Carbapenems have the broadest spectrum of the β-lactam antibiotics,
including activity against both Gram-positive and Gram-negative organisms
[70]. Meropenem is stable against many β-lactamases, and has antibacterial
spectrum similar to that of piperacillin/tazobactam [71].
Resistance to β-lactam antibiotics generally occurs through different
mechanisms; enzymatic degradation of the β-lactam ring by β-lactamases,
structural changes of the PBPs decreasing the affinity for antibiotics, cell-wall
21
drug impermeability, and increased efflux of the drug [72, 73]. β-lactamases
are, for example, found in S. aureus and in Enterobacteriaceae spp., particularly
in E. coli and K. pneumoniae [74] and described in more detail below.
In methicillin-resistant S. aureus (MRSA), resistance is conferred by altered
low-affinity PBPs, expressed upon exposure to β-lactam antibiotics [65]. Cellwall impermeability and increased efflux of β-lactam antibiotics due to
increased expression of porin-pumps are mechanisms found in P. aeruginosa
[75]. Due to resistance development, different treatment strategies have been
developed, such as combination therapy with other antibiotics, as well as with
β-lactamase inhibitors [63, 66, 72]. β-lactam antibiotics are sometimes used
together with aminoglycosides for treatment of sepsis with an unknown focus
of infection [3].
Fluoroquinolones
Fluoroquinolones (FQ) are synthetic antibiotics. Compounds belonging to
this class include ciprofloxacin, levofloxacin and moxifloxacin [62]. The major
mechanism is the blocking of bacterial chromosome replication through
inhibition of bacterial DNA gyrase, which prevents DNA supercoiling, thereby
resulting in bacterial killing [76, 77]. FQ are active primarily against Gramnegative but also against Gram-positive bacteria. Resistance to FQ is mediated
through gyrA, gyrB, parC and qnr genes, which reduce the formation of Gyrquinolone-DNA complexes, and thus, the inhibition mediated by the FQ [62].
FQ are used for empirical treatment of febrile upper urinary tract infections [78],
and outpatient management of P. aeruginosa infections. FQ are important parts
in the combination treatment of multi-drug resistant M. tuberculosis [79] and
prosthetic joint infections [80].
Oxazolidinones
Linezolid is an oxazolidinone active against Gram-positive bacteria, by
inhibiting bacterial protein synthesis [81]. Resistance is conferred by alterations
in the binding site of linezolid in the ribosome, changing the peptidyltransferase-center [82]. Linezolid is used restrictively due to its severe side
effects (bone marrow suppression and polyneuropathy) but indications include
complicated skeletal, skin, and soft tissue infections, especially when caused by
resistant Gram-positive cocci, such as MRSA or vancomycin-resistant
enterococcus (VRE) [83], and infections with multi drug resistant M.
tuberculosis [79].
Lincosamides
Lincosamides inhibit bacterial growth by binding to ribosomes, and thereby
inhibiting bacterial protein synthesis [84]. Clindamycin is an example, used for
treatment of infections caused by Gram-positive and anaerobic bacteria [62].
Resistance to lincosamides as well as to macrolides is conferred by mutations
22
and posttranscriptional modifications that alter interaction between the
antibiotic and the peptidyltransferase center of the ribosome [84]. Clindamycin
can be used instead of penicillin in penicillin allergic patients [85]. It is
associated with Clostridium difficile colonization, and should be used
restrictively [85]. In critically ill patients, lincosamides may be used in
combination with other antibiotics to ensure effective treatment of
Staphylococcus or Streptococcus infections [3, 85].
Ansamycins
Rifampicin is active primarily against Gram-positive cocci and M.
tuberculosis [86]. The mechanism of action is by inhibition of bacterial DNAdependent RNA polymerase. Resistance is due to point mutations decreasing
affinity to the RNA polymerase [62]. Rifampicin is used in combination therapy
for M. tuberculosis [79, 87], and for treatment of skeletal, skin, and soft tissue
infections caused by Staphylococcus or Streptococcus and prosthetic joint
infections [80].
Methods for antibiotic quantification
Current methods for quantifying antibiotic concentrations include highpressure liquid-chromatography with ultra-violet detection (HPLC-UV) [88],
liquid chromatography mass spectrometry (LC-MS) [89], immunoassay
methods [90, 91], and microbiological assays [92, 93]. The principle behind
immunoassay methods is the binding of antibodies to the compound of interest.
Detection can be enzymatic (enzyme-linked immunosorbent assay, ELISA) or
via fluorophores (fluoroimmunoassay, FIA) [94]. Immunoassay methods for
detection of β-lactam antibiotics suffer from cross reactivity decreasing
specificity, and many of the methods are detection-only, used for screening of
antibiotic residues in food products [90]. Immunoassays are thus available for
clinical use for aminoglycosides and vancomycin, mainly to limit toxicity [95].
Microbiological methods detect antimicrobial activity by inhibition of bacterial
growth, but is unable to show if the activity is derived from one or several
compounds [94]. Microbiological methods are described for a few antibiotics
for measurements of concentrations in patient samples, such as vancomycin
[92] and teicoplanin [93].
The method with the highest sensitivity and capacity for a high turn-around
for β-lactam antibiotics is LC-MS, which can also analyze multiple compounds
in a single sample [96]. Methodological developments such as ultra-high
pressure liquid chromatography, coupled to tandem mass spectrometry (UPLCMS/MS) allow for detection of low concentrations and several compounds in a
single sample with a short turn-around time [89, 97]. The UPLC allows for
retention times up to eight times lower than HPLC; 2.5 minutes analysis time
versus 20 minutes [88, 89]
23
Ultra-performance liquid chromatography tandem mass
spectrometry
UPLC-MS/MS combines the separation capacity of liquid chromatography
with mass spectrometry’s capability to analyze mass to charge ratio (m/z). The
LC step is used for separation by the difference in the substance distribution
between the stationary and mobile phase using high/ultrahigh pressure
(HPLC/UPLC) [98]. UPLC allows for faster retention times by applying a
higher pressure than HPLC. The stationary phase most often consists of
octadecylsilyl (C18), and the mobile phase, a water-acetonitrile (ACN) or
water-methanol (MeOH) mix [98]. The separation is optimized by varying
organic/water content often by gradient elution, using two separate mobile
phases and by adding low amounts of volatile buffer salts or acids such as
ammonium acetate/formate, formic/acetic acid or ammonia, compatible with
the MS. This set-up is called a reversed-phase (RP) chromatography since the
stationary phase is non-polar (hydrophobic), and the mobile phase is polar
(hydrophilic) [98]. After UPLC-separation, the compounds of the sample are
transferred to the MS, one by one, via ionization by the method of choice;
atmospheric pressure chemical ionization (APCI) or electrospray ionization
(ESI) (Fig. 8), both of which are compatible with LC [98].
Figure 8. Electrospray ionization is created by applying a high voltage on a flow of liquid
at atmospheric pressures. The aerosol is directed to an opening in the vacuum system of
the mass spectrometer, where the droplets are de-solvated by a combination of heat,
vacuum and acceleration into gas. The ions are ejected from the droplets and accelerated
into the mass analyzer by voltage. Figure is from [6]. Open access at openi.nlm.nih.gov
Date of download 11/22/2016. Reprinted with permission.
24
The ions are then separated by mass spectrometry on the basis of their mass
to charge ratio (m/z) [98]. With a tandem MS it is possible to use selected
reaction monitoring (SRM) where the ion of interest (precursor ion) is selected
with the first MS, fragmenting the ion and then measuring the m/z of the
fragments with the second MS [98] (Fig. 9). By fragmentation it is, therefore,
possible to obtain structural information and increase selectivity and sensitivity
[98, 99].
Figure 9. Tandem mass spectrometry using selective reaction monitoring. Figure is from
[4]. Open access at openi.nlm.nih.gov Date of download 10/27/2016. Reprinted with
permission.
Methodological variability associated with LC-MS methods include matrix
interference, variation in temperature and pressure, and thus ionization efficacy
all affecting selectivity and sensitivity [100]. In LC-MS methods three types of
controls are generally included, five to six separate calibration standards, two
to three separate quality controls and an internal standard that is added to each
sample and control [98]. The advantage of using internal standards, is that
quantification is based on ratios between the IS and the sample, and not absolute
responses. Thus, an internal standard similar to the compound to be quantified
will most likely behave similarly, and thus increasing selectivity and sensitivity
[100].
Recently, several methods involving UPLC-MS/MS for the quantification
of antibiotics have been described, including β-lactam antibiotics in various
matrixes, such as serum and urine [89, 97, 101]. In one study, the UPLCMS/MS method was used for quantification of amoxicillin, piperacillin,
meropenem, cefuroxime, and ceftazidime in patient serum samples [89].
Sample preparation was by ACN protein precipitation. Chromatographic
separation was achieved with an ethylene bridged hybrid (BEH) C18 column,
25
and a gradient elution with water and MeOH containing 0.1 % formic acid (FA),
and 2 mM ammonium acetate [89]. Deuterated compounds were used as internal
standards. The range of quantification was 1-100 mg/L for amoxicillin and
cefuroxime, 0.5-80 mg/L for meropenem and ceftazidime, and 1-150 mg/L for
piperacillin, with a total run time of 2.5 minutes [89]. Matrix effects were
detected for amoxicillin (ion suppression) and meropenem (ion enhancement),
which were efficiently compensated by the use of a stable isotopically lableled
internal standards [89]. Inaccuracy and imprecision were, within limits, defined
by the European Medicines Agency (EMA) [89, 102]. In another study using
UPLC-MS/MS, 21 different antibiotics were analyzed at the same time in urine,
serum, cerebrospinal fluid (CSF), and bronchial aspiration samples [97]. Serum
and bronchial aspirates were diluted using ACN and dithiothreitol, respectively,
and centrifuged, after which the supernatant was used for analysis [97]. Urine
and CSF samples were diluted in the mobile phase, and then directly analyzed
[97]. Matrix effects were observed for some or all compounds, mainly from
urine and bronchial aspiration samples, which were managed by using matrixmatched calibration [97]. Total analysis time including sample preparation was
1-3 hours, depending on sample type [97]. For all matrices, recovery was
between 70-120 %, and inter-day variability was <25 % [97].
Antimicrobial drug susceptibility testing
To guide antimicrobial therapy according to the susceptibility pattern of the
pathogen causing the infection, antimicrobial drug susceptibility testing (AST)
is categorized according to the susceptibility (S), intermediate (I), and resistant
(R) system [103]. This categorization is based on clinical breakpoints, and
should separate bacterial isolates with a high (S) from those with a low (R)
likelihood of treatment success [103]. The clinical breakpoints are set by
combining clinical outcome data with data from the wild-type MIC distribution,
epidemiological cut-off values (ECOFF) and PK/PD characteristics [103].
There are several methods for determining antibiotic susceptibility patterns
of bacteria in vitro. The most common methods include the disk diffusion
method [104], and MIC determination using E-tests or broth dilution methods
[105].
In the broth microdilution method (Fig. 10), a bacterial suspension is
dispensed together with serial two-fold dilutions of the antibiotic to be tested.
The inoculum, as well as other conditions should be standardized, as detailed
by reference authorities, such as the European Committee on Antimicrobial
Susceptibility Testing (EUCAST), or the Clinical and Laboratory Standards
Institute (CLSI) [105]. After an overnight incubation, cultures are visually
inspected for growth. The lowest antibiotic concentration that prevents visible
growth is defined as the MIC [105]. According to current standards, the method
variation for MIC-testing is usually within +/- one two-fold MIC-dilution step.
26
Figure 10. MIC determination by the broth dilution method. Isolate the bacterial strain to
be tested (1). Create a bacterial suspension (2). Inoculate the bacterial suspension in a 2step dilution series of the antibiotic to be tested (3). Inoculate the growth control (4).
Incubate overnight at 37 C. Detect the lowest concentration inhibiting visible growth
which is the MIC (5). In the example above the MIC is 4 mg/L.
Two commonly used terms are MIC90 and MICECOFF (Fig. 11). The MIC90
is the concentration needed to inhibit a growth of 90 % of organisms, and is
dependent on the proportion of resistant bacteria [103]. The epidemiological
cut-off value (ECOFF) is defined by the upper part of the wild-type distribution,
or the “worst-case” MIC among susceptible bacteria [46, 58].
Figure 11. MIC distribution of P. aeruginosa exposed to meropenem. The MICECOFF is 2
mg/L, meaning that wild-type organisms have an MIC ≤2 mg/L. The wild-type is by
definition devoid of phenotypically expressed drug resistance. The ECOFF is determined
by a combination of visual and mathematical methods. From EUCAST
[http://mic.eucast.org/Eucast2/regShow.jsp?Id=3208]. Date of download 11/22/2016.
Reprinted with permission.
27
Transmission investigation by molecular
typing methods
Bacterial transmission investigation, or typing, is the identification of
different types of organisms or clones within a species, by combining molecular
typing results and epidemiological data to analyze bacterial spread [106].
Typing can be done phenotypically, but for increased differentiation to identify
transmission, molecular methods are necessary [106].
Control of nosocomial outbreaks
The use of appropriate work clothes, sanitation routines, and basic hygiene
procedures, such as hand hygiene, must be strictly adhered to by all hospital
staff members to prevent hospital-acquired (nosocomial) infections. Care
should be taken regarding cleaning of catheters and tracheal tubes to minimize
transmission of specific nosocomial infections such as VAP, CLABSI and
CAUTI [13]. When outbreaks of bacteria occur, it is crucial to identify and
isolate patients as soon as possible [74].
Especially challenging bacteria
The group of bacteria known as the ESKAPE pathogens [107] are especially
problematic because of their ability to acquire and further develop antibiotic
resistance mechanisms, as well as to spread easily in the health care setting [21,
107]. These pathogens are difficult to treat with available therapeutic
alternatives, and infections caused by these bacteria are associated with
considerable mortality [21, 107]. By using new strategies for rapid molecular
typing of isolates, there will be better possibilities for limiting uncontrolled
spread of these pathogens.
Tuberculosis (TB) is one of the most wide-spread infectious diseases in the
world, caused by Mycobacterium tuberculosis. About one third of the world’s
population is estimated to be infected, and every year, ten million people
develop active tuberculosis, and approximately 1.5 million die from the disease
[108]. Most cases occur in Africa and Asia [108]. In more than 90 % of those
infected, the pathogen is contained as an asymptomatic latent infection, and
controlled by an effective immune response [109]. Effective treatment of TB
requires accurate and early diagnosis together with screening for drug
resistance, the administration of effective regimens under supervision, and the
provision of support to patients for compliance throughout the course of
treatment [110]. Control of M. tuberculosis transmission is important to the
success of the End TB Strategy [108], in which the goal is to end the global TB
endemic, to reduce TB deaths by 95 %, and to decrease the number of new cases
28
by 90 % between 2015 and 2035 [111]. There are several lineages of M
tuberculosis. One of the most successful in terms of spread is the Beijing lineage
[112]. In the UK, WGS is now used for all isolates of M. tuberculosis, enabling
rapid transmission investigation as shown by Walker et al. [113].
Extended-spectrum β-lactamases
Bacteria belonging to Enterobacteriacaeae, such as E. coli and K.
pneumoniae are some of the most common bacteria found in urinary tract
infections, infections after abdominal surgery, and bacteremia [114]. These
Gram-negative bacteria are accumulating resistance mechanisms against βlactam antibiotics, mainly through acquisition of plasmids carrying resistance
genes such as blaCTX-M, blaTEM and blaOXA [114, 115].
β-lactamases can be divided into different classes depending on their
functional characteristics or structure [116]. Based on protein sequence, βlactamases are divided into four classes; A, B, C and D. A, C and D hydrolyze
their substrates via a serine in the active site, whereas class B enzymes have at
least one zink ion in the active site, and are labelled metalloenzymes [116, 117]
(Table 1). Functional classification is made by the ability to hydrolyze β-lactam
classes, together with their susceptibility to β-lactamase inhibitors [116].
Functional classification is divided into 3 groups (1-3), where 2 and 3 are further
divided into subgroups [116] (Table 1).
Table 1. Classification of β-lactamases.
29
According to the Public Health Agency of Sweden, β-lactamases are
classified into three groups; ESBLA, ESBLM and ESBLCARBA [114]. ESBLA,
and ESBLM both inhibit cephalosporins. Resistance to cephalosporins in ESBLA
isolates is inhibited by clavulanic acid, which is not the case for ESBLM isolates,
which can be inhibited by substances such as cefotetan [114]. ESBLCARBA
isolates have, as the name implies, carbapenemase activities. ESBLCARBA are
then further divided into A, B and D. ESBLCARBA A and B inhibit
cephalosporins and carbapenems. ESBLCARBA A can be inhibited by boric acid,
and ESBLCARBA B by dipicolinic acid. ESBLCARBA D inhibits carbapenems, and
there are no known inhibitors [114].
The Public Health Agency of Sweden estimates the cost of healthcare for
dealing with ESBL-producing bacteria to be approximately 75 million SEK
yearly [118]. By the current increasing trends we will have 50 000 new cases of
ESBL carriers, and 25 000 cases of ESBL positive urine cultures by the year
2022, compared to current detection rates of approximately 10 000 yearly [118].
This escalation might increase costs by a factor of 2.5-5 for the healthcare
system to deal with ESBL-producing bacteria [118].
Methods for bacterial outbreak investigation
When outbreaks occur it is important to quickly identify their sources and
organisms, and to take necessary precautions including isolation of high-risk
patients. It is especially important to identify high-risk clones, since spread is
associated with increased hospital mortality [21]. Epidemiological typing
methods are based on genomic, proteomic, and phenotypic principles. Many of
the methods are time-consuming and expensive, and require special equipment.
When designing new typing methods, several factors need to be considered,
depending on their purpose. These factors include stability, discriminatory
power, reproducibility, speed, accessibility, cost-efficiency, and user
friendliness [25, 119]. In addition, its appropriateness in a given situation (e.g.,
an outbreak situation) must be evaluated. To identify potential nosocomial
outbreaks, several molecular methods have been described [119-123].
Until recently, pulsed-field gel electrophoresis (PFGE) was considered the
gold standard for a large number of bacterial species because of its
discriminatory power and high typeability [25, 119]. The drawbacks of PFGE
are that it is laborious and time-consuming, and that interpretation can be
complex, requiring rigorous standardization and experienced personnel to
achieve reproducible results comparable over time and place.
New molecular typing methods are being continually developed. One
approach is repetitive-sequence-based PCR (rep-PCR), where repetitive
sequences in the genome are amplified and analyzed by electrophoresis [124]
or sequencing [125]. Other approaches leading to a higher reproducibility
among laboratories, since they are based on DNA sequencing, include multi
30
locus sequence typing (MLST) [24], and multi locus variable number tandemrepeat analysis (MLVA) [126], (Table 2).
Previously, Masny and Płucienniczak [127] described a novel method based
on ligation-mediated PCR (LM/PCR) using low denaturation temperatures
(TDs), resulting in specific melting-profile DNA product patterns for bacterial
and fungal isolates. The method is based on genomic cleavage by restriction
enzymes, and the amplification of several fragments by ligation-mediated PCR,
followed by an analysis of the DNA fragment patterns by gel electrophoresis
[128]. The method has been previously performed successfully for
Enterobacteriaceae, Candida spp., Staphylococcus aureus, and Enterococcus
spp., with a discriminatory power comparable to that of PFGE [128-131]. For
E. coli, 70 isolates from blood cultures from 37 hospitalized patients with no
epidemiological links were analyzed. LM/PCR identified 36 MP-types as did
the comparative method PFGE [128]. Comparison of LM/PCR with PFGE for
C. albicans was carried out by analyzing 123 isolates including 116 patient
isolates and 7 reference strains [129]. LM/PCR identified 27 unique types
whereas PFGE identified 25 types, and thus the typeability of LM/PCR was
higher (100 %) compared to PFGE (94.2 %) [129]. Analysis of S. aureus based
on 37 isolates from patients suffering from furunculosis, and results from
LM/PCR showed full agreement to PFGE results for 33/37 isolates [130]. Three
of the remaining isolates were sub-grouped using LM/PCR, indicating
microevolution not detected by PFGE [130].
High resolution melting (HRM) of DNA coupled to qPCR is an analysis
method with increasing applications [132]. DNA is melted in the presence of a
double stranded (ds) DNA-binding fluorescent dye. A detector registers the
light emitted from the dye. As the temperature is increased during analysis, the
dsDNA becomes single stranded (ss) DNA, releasing the dye, which stops
emitting light. The point at which 50 % of the DNA has melted is called the
melting temperature (TM), which depends on the length and GC content [132,
133]. Thus, the melting pattern is unique for each DNA sequence of sufficient
length. HRMA has been proposed for species identification and typing of both
isolates and specific genes [134, 135]. HRMA of an SNP in the gene rfbS, used
as a marker for the identification of Salmonella pollorum and S. gallinarum,
showed 100 % specificity and a sensitivity >100-fold higher than the allelic
specific method previously used, when analyzing 15 reference strains and 33
clinical isolates [136].
Today, most molecular typing methods are based on whole genome
sequencing (WGS), generating millions of reads of 100-400 bp, covering nearly
the entire genome [106]. WGS data is analyzed by an assembly of overlapping
reads (de novo), or by comparisons to reference genomes (reference based)
31
[106] (Table 2). The Public Health Agency of Sweden has recently begun to use
WGS routinely for primary molecular typing [137].
Table 2. Methods for molecular typing.
*post culture. **Spa-typing is used as an example for rep-PCR.+poor, ++low,
+++moderate, ++++good, +++++high, ++++++excellent. [Knetsch et al Euro Surveill
2013:18, Mellmann et al. PLoS Med 2006:3].
For Mycobacterium tuberculosis, previous methods used for molecular
typing such as mycobacterial interspersed repetitive units variable number
tandem repeats (MIRU-VNTR) [138], and spacer oligonucleotide typing
(spoligotyping) [139], may be standardized, and the most suitable methods for
long-term surveillance, but slow and resource-demanding when it comes to
early detection of an outbreak [140]. In MIRU-VNTR the variable numbers of
tandem repeat loci or mycobacterial interspersed repetitive units are amplified
by PCR, and the obtained products are sized on agarose gels or capillary gel
electrophoresis, to deduce the number of repeats in each individual locus [138].
In spoligotyping, a single direct repeat locus is amplified by PCR [139]. The
direct repeat harbors 36 bp repeats interspersed with 34-41 bp unique spacer
sequences [139]. The amplified products are hybridized to a membrane with 43
oligonucleotides corresponding to the spacer sequences from M. tuberculosis
H37Rv and M. bovis BCG [139]. Positive and negative hybridization signals
forms the isolate pattern [139].
Today, WGS is becoming the standard for molecular typing of M.
tuberculosis [113]. A particular problem for M. tuberculosis is the lack of a
suitable method to analyze WGS directly from clinical specimens which is
problematic, since it may take 3-4 weeks to achieve sufficient bacterial growth
[141].
32
Pulsed field gel electrophoresis
Pulsed field gel electrophoresis (PFGE) (Fig. 12) is still widely used, and
was until recently, considered the gold standard for molecular typing [25, 119].
During PFGE analysis, bacterial DNA is cut with restriction enzymes, such as
XbaI for E. coli, creating fragments of different lengths. The fragments are then
separated in an agarose gel, based on length using an electric field. Even large
fragments can be separated by changing the direction of the electrical field (Fig.
12) [142]. Disadvantages are that PFGE requires special equipment, the
development and application of standard protocols and training [25, 119, 143146]. PFGE is a labor-intensive method, and interpretation requires
standardization and trained personnel [147].
Figure 12. The PFGE process. From “Pulsed-field Gel Electrophoresis (PFGE)” Centers
for Disease Control and Prevention, 2016. In the public domain. Date of download
11/16/2016. Reprinted with permission.
To standardize the analysis of PFGE results Tenover et al. proposed
guidelines in 1995 with categories for genetic and epidemiological relatedness,
as well as criteria for interpreting the PFGE patterns [148] (Table 3).
33
Table 3. Guidelines for interpretation of PFGE patterns.
Matrix-assisted laser desorption/ionization time of flight
Matrix-assisted laser desorption/ionization time of flight (MALDI TOF) was
introduced in clinical microbiology in the beginning of the 21st century to speed
up bacterial species identification [149, 150]. It became the new gold standard
by enabling bacterial species identification at least one day earlier than
conventional phenotypic methods, and by being a universal method with lower
running costs [150]. Bacterial species identification using MALDI TOF is
carried out by mixing whole cells or cell extracts with a MALDI matrix on a
steel target plate [151]. The sample is ionized by a laser, and charged proteins
and peptides are released. These charged proteins and peptides are accelerated
in a tube, and then detected. The time it takes for each protein and peptide to
reach the detector is dependent on the mass (Fig. 13). All charged proteins and
peptides from a sample are presented in a spectrum, which is then compared to
a database containing reference spectra. Comparison is performed by matching
signals in the reference spectrum to the sample spectrum, and vice versa, and
by analyzing the symmetry of all matching signals [151]. As an example, by
using MALDI TOF, 117 strains of various Yersinia entercolitica bioserotypes
were correctly subtyped to the biotype level [152].
34
Figure 13. The MALDI TOF process. The sample is ionized by a laser (4B). The ionized
proteins and peptides fly through a tube and are detected in sequential order depending on
mass (4C). [http://pe-iitb.vlabs.ac.in/exp8/images/34.JPG] Date of download 11/23/2016.
Reprinted with permission (https://creativecommons.org/licenses/by-nc-sa/3.0/legalcode).
35
Whole genome sequencing
Whole genome sequencing (WGS) by MPS has become the new standard
for molecular typing [143-146, 153]. There are several platforms on the market,
all using different methodologies, producing data of varying length and depth,
as well as of different quality [26]. Examples of platforms are MiSeq from
Illumina, PacBio from Pacific Biosciences, and MinION from Oxford
Nanopore [154, 155]. The most common method is sequencing by creating short
reads, 100-400 bp, covering each base in the genome several times, which are
then assembled by overlapping regions (de novo), or mapped against a reference
genome (reference based assembly) [1, 106] (Fig. 14). There are also methods
creating longer reads with less coverage, which are useful when studying the
complete genome [1] (Fig. 14).
Figure 14. Overview of two different WGS systems. WGS workflow involves DNA
extraction and library preparation (1) followed by sequencing (2) and read assembly (3)
creating contiguous sequences (contigs) (4) arranged into scaffolds. The resulting
assemblies can be used for molecular typing and detection of genes of interest (5). Image is
from [1]. Open access at openi.nlm.nih.gov Date of download 12/30/2016. Reprinted with
permission.
36
When interpreting data generated by WGS, it is evident that there is a need
of standardized analysis, as many different approaches using both different
WGS platforms combined with various bioinformatics pipelines are used.
Single nucleotide polymorphisms (SNPs) and gene-by-gene analyses are used
to investigate epidemiological relatedness. The gene-by-gene approach, core
genome (cg) MLST, extends the MLST analysis to the core genes, and thus
includes hundreds or thousands of conserved genome-wide genes [119]. As for
MLST schemes, also cgMLST schemes are becoming publically available at
www.cgmlst.org/ncs. cgMLST schemes are not available for all species as of
today, and unfortunately not yet standardized.
In a study by Walker et al. [113] investigating cross-sectional, longitudinal,
and household and community spread of M. tuberculosis, 390 isolates from 254
patients were sequenced. All five major lineages were represented, and thus the
results should be applicable even outside the area investigated. The rate of
genomic change was estimated at 0.5 SNP yearly, with a transmission threshold
of ≤5 SNP over 3 years. Epidemiological transmission could be verified, as
could super spreaders [113]. Another study including 344 outbreak isolates
argues that the threshold set by Walker et al. is too high, and that direct
transmission is not necessarily the explanation, even if two isolates are identical
[156]. There is a clear difference in the targeted population in this study where
Walker et al. investigated a low burden setting [113], whereas the study from
Casali et al. focused on a population with several risk factors for transmission
[156].
In another study, an MRSA outbreak was analyzed, and the feasibility of
SNP analysis was tested [157]. The analysis included seven MRSA isolates
thought to be part of a neonatal ICU outbreak, and two MRSA isolates
previously isolated in the same NICU, along with five unrelated MRSA isolates
[157]. MLST analysis was not sufficiently discriminatory to resolve the isolates,
as both outbreak and unrelated isolates were ST-22 [157]. Analysis of the SNPs
in the core genome did, however, reveal a cluster of the suspected outbreak
isolates.
When an outbreak of EHEC O104:H4 causing 834 cases of hemolytic
uremic syndrome (HUS), and 2967 non-HUS cases, was reported in Germany
in 2011, cgMLST was used to propose an evolutionary model for the emergence
of the outbreak strain [158].
The evolutionary relatedness of C. difficile 078 was studied using WGS
followed by SNP analysis [159]. Sixty-five isolates including 19 isolates from
pigs and 15 from asymptotic farmers, of which 12 were paired, and 31 isolates
from hospitalized patients were included [159]. MLST analysis showed that all
isolates were ST-11, and thus not sufficiently discriminatory to allow for
investigation of transmission [159]. The mutation rate was estimated at 1.1
SNPs per genome per year, hence farmers and pigs were colonized with
identical clones (<2 SNPs) [159].
37
Other applications for WGS analysis includes resistance prediction. This is
especially useful for slow-growing pathogens such as M. tuberculosis, where
WGS analysis provides information regarding species, resistance, and
molecular typing faster, or in the same time, as conventional methods,
particularly when sequencing is performed directly on the patient sample [141].
Regarding resistance testing, WGS can only, thus far, be used as a rule in
method for resistance, and not for determining susceptibility [160].
Susceptibility testing by WGS is, however, a complex matter. A recent
report from EUCAST recommends that when assessing presence of a resistance
gene, ECOFF should be used as comparison for WGS-based resistance
prediction [160]. It is, however, more difficult to predict resistance when the
mechanism behind resistance is upregulation of an intrinsic gene, or in cases
where the mechanism is not well understood [160].
When using WGS, the challenges are to create standardized analysis
approaches comparable to previously used methods such as PFGE and MLST,
and that are transportable and globally available [106]. Since the most
frequently used system today creates contigs rather than fully assembled
genomes, in silico PFGE analysis cannot predict the PFGE pattern [106]. In
silico systems for detecting MLST types are available [161], however with
WGS there are possibilities of expanding the analysis beyond a few
housekeeping genes [106]. Maiden et al. proposed in 2013 using an expanded
MLST analysis; whole genome (wg) MLST for highly clonal pathogens, and
cgMLST for more genomically diverse pathogens [162]. The European Centre
for Disease Prevention and Control (ECDC) suggests that standard analysis of
WGS data for surveillance purposes, and to detect outbreaks caused by identical
strains, should include 2 steps [163]. First, cgMLST should be used to identify
clusters and enable comparable nomenclature. The gene-by-gene analysis
should then be followed by an SNP-based analysis to further resolve closely
related cgMLST types [163]. Furthermore, as discussed in the reports from both
ECDC and EUCAST, a comprehensive quality control (QC) program needs to
be developed [160, 163]. The QC program should state which controls to
include (isolate, sequence, DNA), how to overcome sequencing bias created by
the different systems, and how to comprehensively analyze and compare data,
as well as guidelines on storage of data [160, 163].
38
Aims of the study
The overall aim of the study was to investigate strategies to enable
individualized treatment of critically ill patients, and methods to characterize
nosocomial outbreaks of resistant bacteria. This was further specified in the
following research questions:
 What is the distribution of β-lactam drug concentrations in ICU
patients?
 Which risk factors are associated with failure to achieve 100% fT>MIC
for β-lactam antibiotics in ICU patients?
 Is it possible to develop locally adapted methods for determining serum
antibiotic drug concentrations?
 What is the performance of LMqPCR HRMA compared to PFGE and
whole genome sequencing?
39
Materials and methods
Measurement of β-lactam antibiotic concentrations
in critically ill patients
Study population
Participating centers in a multi-center prospective clinical study (Paper I)
were ICUs in southeastern Sweden; Kalmar County Hospital, (lead
investigating site), Linköping University Hospital, Växjö Central Hospital, and
Jönköping County Hospital Ryhov. Ethical approval was granted by the
Regional Ethical Review Board, Linköping University, Sweden (DNR
2014/236-31).
All ICU patients at participating centers were screened during the inclusion
periods. Patients were included during 2-3 months per site from September 2014
to July 2015. Inclusion criteria were treatment with intravenous (i.v.)
antibiotics, and age ≥18. Due to availability of staff (MDs or specially trained
nurses), inclusion took place primarily on weekdays. Clinical follow-up of all
included patients was continued until 30 days after inclusion.
Clinical follow-up
Clinical and demographic variables were collected, including height,
weight, antibiotic treatment, SAPS3 (Simplified Acute Physiology Score 3), use
of assisted ventilation strategies, continuous renal replacement therapy (CRRT),
use of inotropic drugs, and fluid balance. Recorded laboratory variables were
C-reactive protein (CRP), hemoglobin, white blood cell count, and platelet
count, calculated renal clearance by Cockcroft-Gault (eGFR), albumin, urea,
lactate, bilirubin, and alanine aminotransferase (ALAT). Severe sepsis was
defined according to Swedish guidelines [3] as admission to the ICU with
probable or verified infection, and signs of hypotension (need for inotropic
drugs), hypoperfusion (lactate >1 mmol/L above the upper normal limit), or
organ failure (assisted ventilation, eGFR <30 L/min/1.73m2, trombocytopaenia
(<100x109/L), prothrombin complex >1.5 INR or bilirubinaemia (>45µmol/L)).
Septic shock was defined as severe sepsis with hypotension together with
hypoperfusion or organ failure. LOS at ICU and 30-day mortality were
recorded. The source of infection and bacterial etiology based on available
cultures were evaluated by an infectious disease specialist, or attending ICU
physician, blinded to the serum drug concentrations.
Bacterial isolates and MIC determination
Bacterial isolates were collected from patients with positive cultures. MIC
determination was performed with a 1.0 McFarland inoculum on Müller-Hinton
40
agar using E test strips for piperacillin/tazobactam (PIP/taz) (0.016-256/4
mg/L), meropenem (MER) (0.002-32 mg/L) and cefotaxime (CTX) (0.002-32
mg/L). For the analysis of MICACTUAL, only MICs of bacterial isolates where
adequate empirical treatment was identified, were included. The MICECOFF
applied were 16, 2 and 4 mg/L for PIP, MER and CTX as previously described
[46].
Blood samples
Venous blood samples were collected just prior to the next dosing of
antibiotics. From each patient, 1-3 samples were obtained on three consecutive
days. Samples were allowed to coagulate for 50 minutes at room temperature,
and centrifuged at 2000 x g for 10 minutes before storage at -80 C (Paper I).
Shipping of samples for determination of serum antibiotic concentrations
included dry ice.
Measurement of serum antibiotic concentrations with liquid
chromatography mass spectrometry
Measurement of serum antibiotic concentration was performed
retrospectively and separately for PIP, MER and CTX, by using LC-MS at the
accredited Department of Pharmacology, Karolinska University Hospital,
Stockholm, Sweden. Methodological inaccuracy and imprecision were within
±15 % for all drugs. The measurement range was 0.2-100 mg/L for PIP, and
0.2-50 mg/L for MER and CTX. In short, protein precipitation was carried out
with ACN after which the samples were injected into the column (Kinetex 2.6
µm C18 100 A, 50x2.1 mm from Phenomenex) and separated by reversed phase
HPLC (Agilent HP 1100 or 1200). Ionization was achieved by atmospheric
pressure ionization electro spray (API-ES). Detection was by MS (Agilent HP
1100 MSD or 1200 MSD). Mobile phase A and B were 0.1 % FA and ACN,
respectively.
Development of an ultra-performance liquid
chromatography tandem mass spectrometry
method quantifying 11 antibiotics
Chemicals and reagents
All reagents were of analytical grade. ACN, MeOH, and FA of LCMS grade
were from VWR (Stockholm, Sweden). Ultrapure water was obtained from a
Milli-Q water purification system (Millipore, Solna, Sweden). Solvents and
water were degassed in an ultrasonic bath prior to use. The target compound
details are presented in Table 4. Compound specific stable isotopically labeled
IS were used when available, and included meropenem-D6 for MER and CTX,
piperacillin-D5 for PIP, flucloxacillin-13C415NNa for cloxacillin and
41
benzylpenicillin, and linezolid-D3 for linezolid and clindamycin, all from
Alsachim (France) and levofloxacin-D8 for levofloxacin and ciprofloxacin,
moxifloxacin-D4 for moxifloxacin, and rifampicin-D3 for rifampicin, all from
LGC Standards (Germany).
Table 4. Details of the antibiotics investigated: name, chemical abstracts service (CAS)
number, formula, and molecular weight (MW).
Preparation of calibrators, quality control samples and internal
standard solutions
Individual stock solutions (1 mg/mL) of each calibration standard were
prepared in MeOH:water 50:50, except for rifampicin, which was diluted in
pure MeOH. A stock solution containing all compounds was prepared daily, and
further diluted to prepare the calibrators and quality controls. Individual stock
solutions (1 mg/mL) of each internal standard were prepared by adding
MeOH:water 50:50 to meropenem-D6, piperacillin-D5, linezolid-D3,
moxifloxacin-D4, and levofloxacin-D8 and MeOH to flucloxacillin-13C415NNa
and rifampicin-D3. IS working solution (1 µg/mL) was prepared by dilution
with MeOH.
Sample preparation and extraction
During method development different sample preparations were tested.
Final sample extraction was by protein precipitation by adding 100 µl of MeOH
and IS working solution to 50 µl serum sample to be determined. Samples were
mixed and centrifuged at 15 000 x g. Fifty µl of the supernatant was transferred
to a 96 deep-well plate, and 950 µl milliQ-water was added, and after further
mixing, the plate was placed in the autosampler and 10 µl was injected onto the
column.
42
Instruments and analytical conditions
Chromatographic separation was achieved with Acquity UPLC (Waters,
Sollentuna, Sweden) using a kinetex biphenyl reverse phase column (1.7 µm,
2.1x100 mm and security guard ultra-biphenyl 2*2 mm) (Phenomonex,
Denmark). A gradient of mobile phase A (aqueous solution with 0.1 % FA) and
mobile phase B (MeOH 0.1 % FA) was used in the chromatographic separation.
Weak needle wash was ACN:ultrapure water 10:90 and strong needle wash was
ACN:ultrapure water 80:20.
MS detection was performed by a tandem quadrupole mass spectrometer
(Xevo TQ MS, Waters, Sollentuna, Sweden) in positive electrospray ionization
mode, according to each analyte (Table 5). Optimized conditions for data
acquisition were identified by direct infusion of a combined standards solution
(1 µg/mL). Table 5 provides a list of the acquisition conditions including
precursor ions, product ions, cone voltage, collision energies, and retention
times for each antibiotic and internal standard used. The optimal MS/MS setup
parameters were a source temperature of 150 C, cone gas flow of 25 L/h, a
1000 L/h, and 450 °C desolvation gas (N2) flow and temperature, respectively,
and 0.17 mL/min collision gas flow.
Table 5. Settings for mass spectrometry. Electrospray ionization (ESI) and retention time
(RT).
43
A laboratory comparison of quantification of commonly used
antibiotics
The serum samples collected for antibiotic quantification, during Paper I
were analyzed both at the Department of Clinical Pharmacology, Linköping and
as described above at the Department of Pharmacology, Karolinska University
Hospital, Stockholm, Sweden.
Development of a microbiological method for
quantification of β-lactam antibiotics concentration
Serum samples previously collected from ICU patients (Paper I) were
diluted in 24 dilution-steps with bacteria, with known MICs to the antibiotic to
be quantified. To verify the MIC of the bacteria used for the patient sample
dilutions, a dilution series of known antibiotic concentrations were inoculated
(Fig 15). For internal control, the antibiotic to be quantified with a concentration
equal to MICECOFF was included. The internal control was diluted in the same
way as the patient serum. All dilutions of bacteria and antibiotics described
above were made in a 96-well plate. Incubation was done for 20 ±1h at 35-37
°C. The dilution able to inhibit bacterial growth was detected, and the
concentration was calculated from the dilution and the MIC. Criteria for
approved analysis was internal control concentration ±20 % of the target (11.220.8 mg/L for PIP, 1.4-2.6 mg/L for MER and 2.8-5.2 mg/L for CTX), and that
the MICcontrol was within ± 1 MIC dilution.
44
Figure 15. Microbiological method for antibiotic concentration determination. (A) Dilution
of patient serum sample. (B) Dilution of internal control samples as well as internal MIC
control for the bacteria used. (C) Sample plate set up for the microbiological analysis of
drug concentrations. Each serum sample was analyzed using two separate 96-well plates
using two different bacteria to cover for the high and the low concentration range for MER
and CTX depending on the MICs. For PIP, the same bacterium was used for analyzing both
high and low concentrations, but the samples were diluted 1:2 before the 24 dilutions (to
cover the analytical range (4–96 mg/l).
45
Methods for bacterial outbreak investigation
Study population and bacterial isolates
An outbreak of ESBL-producing E. coli was suspected in a hospital in
Kalmar County (2008) because patients were epidemiologically related in time
and place. The majority of the outbreak isolates were urine samples from elderly
patients at a medical department and two nursing homes near a hospital within
Kalmar County. Sixteen isolates, from the outbreak of ESBL-producing E. coli,
16 isolates of ESBL-producing E. coli not belonging to the outbreak, and 13
non ESBL-producing E. coli isolates were obtained from the Department of
Clinical Microbiology, Kalmar County Hospital, Sweden, and The Public
Health Agency of Sweden, Solna, Sweden (Papers IV, V and VI).
Isolates of the ESKAPE pathogens (Paper V) were selected (by The Public
Health Agency) based on PFGE typing results (blinded to the investigator), and
obtained from the Department of Clinical Microbiology, Kalmar County
Hospital, Sweden, the Department of Clinical Microbiology, County Council of
Östergötland, Linköping, Sweden, and The Public Health Agency of Sweden,
Solna, Sweden.
Isolates of M. tuberculosis including both related and unrelated strains, were
obtained and cultured at the Public Health Agency of Sweden, Solna, Sweden.
In general, all bacteria except M. tuberculosis, were grown on blood agar,
and DNA was extracted using a magnetic bead system; MagNA Pure Compact
(Roche, Stockholm, Sweden). M. tuberculosis isolates were grown on
Löwenstein Jensen medium and DNA was extracted manually using a
lysozyme, SDS/proteinase K mix, cetyl trimethylammonium bromide (CTAB)
protocol at the P3 laboratory at The Public Health Agency of Sweden [164].
Pulse field gel electrophoresis
Isolates were analyzed by pulsed field gel electrophoresis at The Public
Health Agency according to previously published methods, using the restriction
enzyme XbaI for E. coli, K. pneumoniae and E. cloacae [123], SpeI for P.
aeruginosa [165], ApaI for Acinetobacter spp. [166], and SmaI for S. aureus
and E. faecium [167, 168]. XbaI-digested DNA from Salmonella enterica
serovar Braenderup H9812 (www.cdc.gov/pulsenet) was included as a
normalization standard when analyzing E. coli, K. pneumoniae, E. cloacae and
P. aeruginosa, and S. aureus NCTC8325 for the Gram-positive bacteria and
Acinetobacter spp.
DNA banding patterns were analyzed by BioNumerics version 6.0 (Applied
Maths NV, Belgium). The Dice similarity coefficient and UPGMA were used
for cluster analyses.
For paper IV, where an epidemiological link was established, isolates with
>90 % similarity were defined as being related, and isolates with >80 %
similarity were defined as probably being related. For paper V, where isolates
46
were clustered based on PFGE results, and not epidemiological linked, a higher
level was used as a threshold for similarity. Isolates with >97 % similarity were
defined as identical, isolates with 90-97 % similarity as being closely related,
and isolates with <90 % similarity defined as unrelated.
Ligation mediated qPCR high resolution melt analysis
DNA concentration was normalized for all samples prior to digestion and up
to 1 µg of DNA was used. Digestion was carried out using HindIII followed by
ligation and subsequent PCR reaction. Analysis was by HRM and agarose gel
electrophoresis (Fig 16). To objectively decide whether the HRM graphs were
different or similar, an algorithm was developed (Paper IV). For M. tuberculosis
HindIII was first tested, and then two other restriction enzymes previously used
in a molecular typing method for M. tuberculosis; SalI in combination with
PvuII [169].
Figure 16. Workflow of LMqPCR HRMA. Culture bacteria of interest and extract DNA (12). Use restriction enzyme HindIII to fragment the DNA (3). Perform ligation of DNAoligos to each fragment (4), enabling simultaneous amplification using an universal primer
and subsequent HRMA (5). Confirm if required, with gel electrophoresis (6).
47
Matrix-assisted laser desorption/ionization time of flight
The E. coli isolates from Paper IV and the ESKAPE isolates (all except for
Enterobacter spp.) from Paper V were subjected to FA extraction prior to
spotting on the target. HCCA (α-cyano-4-hydroxycinnamic acid, Bruker,
Germany) was used as the matrix. The target plate was then analyzed in the
Bruker Microflex LT system (Germany), calibrated according to the
manufacturers instruction with Bacterial Test Standard (Bruker, Germany). A
protein profile with m/z values between 2 and 20 kDa was generated based on
240 laser shot measurements, which is common for bacterial typing. MSPs
(main spectra) were created by the standard MALDI Biotyper MSP creation
method (Bruker Daltonics, Bremen, Germany). For each isolate, both the
protein profiles normally generated for bacterial typing, and the MSP profiles,
were used to evaluate the typing potential of MALDI TOF. Analysis of MALDI
TOF single spectrum and MSPs were performed via hierarchical clustering, and
complete correlation using Maldi Biotyper 3.0 software. Peak data was exported
to Bionumerics (Applied Maths NV) and analyzed as summary spectra,
clustered by the Pearson correlation and UPGMA, or the Dice coefficient and
UPGMA.
Whole genome sequencing
Sixteen isolates, from the clinically well-characterized nosocomial outbreak
of ESBL-producing E. coli of type ST-131, and three isolates of ESBLproducing E. coli not belonging to the outbreak of different MLST types from
the collection from 2008-2009, at the Department of Clinical Microbiology,
Kalmar County Hospital, Sweden, was subjected to WGS.
Following extraction, the DNA concentration was measured and normalized
for every sample by dilution. Sequencing libraries were prepared according to
the manufacturer’s instructions using the Nextera XT DNA Library Preparation
Kit (Illumina, San Diego, CA, USA), indexed, pooled and sequenced as 12plex, using the MiSeq desktop sequencer (Illumina). All libraries were
sequenced using a 2 x 300 bp, paired-end workflow. Base calling,
demultiplexing and adapter trimming, were conducted using MiSeq Reporter
version 2.5.1 (Illumina). Isolates 1, 6, 8 and 13 were subjected to paired-end
WGS using the Illumina HiSeq system (SNP&SEQ Technology Platform,
Uppsala University, Uppsala University Hospital, Science for Life Laboratory,
and the Swedish Research Council) and 2 x 100 bp. Quality control, trimming,
and de novo assembly were carried out using CLC Genomics workbench 7.5
with default settings. Contigs were uploaded and analyzed at the Center for
Genomic Epidemiology (CGE) server (www.genomicepidemiology.org, April
16, 2016). SNP analysis was performed using either a well-characterized ST131 reference sequence (JJ1886, accession number NC_022648.1), or the index
isolate (isolate 14). cgMLST analysis was performed using an ad hoc MLST
48
scheme, including 1602 genes, developed from 181 E. coli genomes in
SeqSphere+ (3.2.1 64-bit, Ridom GmbH, Germany).
Statistical analysis
To compare groups reaching and not reaching the PK/PD target (100 %
fT>MIC), the Mann-Whitney U-test, or chi-squared test with Yate’s correction,
was used (Paper I). A p-value <0.05 was regarded as statistically significant. By
using Liddell’s test a comparison between MICACTUAL and MICECOFF as targets
was performed. For identification of variables independently associated with
not reaching the target (100 % fT>MIC), stepwise, multiple logistic regression
analysis was performed, including variables with a p-value <0.1 in the
univariate analysis. Mean substitution was used and the number of missing data
in the variables was low (range 0-9 missing data entries/variable; median 2).
Analyses were performed by Statistica version 12 (Tulsa, USA), and GraphPad
Prism 6 (La Jolla, USA).
For assessment of imprecision and accuracy of the developed methods for
measurement of β-lactam antibiotics, CV (%) and RE (%) were calculated
(Papers II and III). RE (%) was calculated by dividing the absolute error with
the target concentration. Statistical analysis of the correlation between the
results of the microbiological assay and the LC-MS reference assay (Paper III)
was carried out by using the Spearman Rank Correlation test, in GraphPad
Prism version 6 for Windows, (GraphPad Software, San Diego California
USA).
To objectively determine if HRM curves of two isolates (referred to as
isolate A and isolate B below) were different within the same run, an HRM
evaluation algorithm was developed and applied (Papers IV and V).
WGS data was compared by SNP analysis using Burrows-Wheeler
transform alignment tool v.0.7.2, and cgMLST analysis by creating minimum
spanning trees based on 1561 genes (Paper VI).
49
Results and Discussion
Managing treatment with β-lactam antibiotics in
the ICU
In Paper I, the distribution of serum concentrations of β-lactam antibiotics
following empirical treatment was investigated in four ICU settings in
southeastern Sweden. We enrolled 111 critically ill patients. The ability to reach
the pre-defined target of 100 % fT>MIC was investigated by considering actual
minimum inhibitory concentrations (MICACTUAL). The MICACTUAL is rarely
measured in clinical studies, and we wanted to investigate whether the actual
MIC would affect the rate of patients reaching the target, in contrast to the more
commonly used epidemiological cut-off (ECOFF) values (i.e. the highest MIC
of the susceptible bacterial population or “worst-case MIC”). This is of
particular interest whereby Sweden has a relatively low proportion of resistance
against β-lactam antibiotics (Fig.2) [170]. Furthermore, we also explored the
risk factors associated with not reaching 100% fT>MIC in ICU patients on
empirical treatment with broad-spectrum β-lactam antibiotics. However, it must
be stressed that the correlation between the selected target (100% fT>MIC) and
clinical outcome has not yet been confirmed by controlled clinical studies, and
is, thus, mainly based on expert opinion, observational clinical studies, as well
as animal models and hollow fiber studies [171].
The patients enrolled in our study were older (70 on average, versus 60-65
years old) than in other studies assessing drug distributions of β-lactam
antibiotics [46, 58]. In our study, all ICU patients receiving i.v antibiotics ≥18
years old were included, which is similar to the DALI study [172]. In contrast,
the level of impaired kidney function impairment was higher in our study (mean
eGFR 48 compared to 80 and 91 mL/min) [46, 58]. The mean SOFA score was
8 in our study, compared to a mean SOFA score of 5 in the DALI study [46],
concluding that the patients enrolled in our study were more severely ill [173,
174]. The body mass index was also higher in our study, 27.2 compared to 25.9
kg/m2 [58]. Compared to the baseline characteristics of two recent RCTs
measuring MER and PIP levels, our patients were in general more severely ill,
indicated by lower kidney functions and higher SAPS3 scores [60, 61].
We found a high interpatient variability of β-lactam antibiotic
concentrations (Fig. 17) in agreement with the DALI study [46] and the control
group receiving IB dosing in the β-Lactam Infusion in Severe Sepsis (BLISS)
study [175]. The same interpatient variability was detected in a recent, small
study, including nine patients measuring the pharmacokinetics of PIP [176]. In
this study, the most important determinant for variable antibiotic concentrations
was renal clearance [176]. In our study, renal clearance measured as eGFR
varied between 32-105 mL/min (excluding patients on CRRT), which could
50
explain the high concentration variability of PIP, MER, and CTX. The
concentration in each patient over time was stable around the MICECOFF (88 %,
Paper I), indicating that repeated sampling of patients within target range during
TDM may not be necessary in this setting. This is in contrast to a study with
less severely ill patients (n=11), which showed that 46 % (5/11) had
concentrations below MICECOFF for PIP at least once in seven daily
measurements of the same patient. This difference is probably due to lower
eGFR levels than in our study, although the sample size is limited.
Figure 17. Serum trough antibiotic median concentration for piperacillin, meropenem and
cefotaxime. The free antibiotic median serum concentrations were 30.1 mg/L (IQR 4.863.0), 6.2 mg/L (IQR 2.0-15.2) and 2.0 mg/L (IQR (IQR 0.5-6.6) for piperacillin,
meropenem and cefotaxime, respectively. Individual values day 1, median and IQR are
presented. The y-axis is presented as a log2 scale.
The target attainment for 100 % fT>MICECOFF was 55 %, which is
comparable to the rate observed in the DALI study, 60 % [46]. Similar levels of
target attainment were observed at baseline in an RCT in Belgium (n=41) where
all antibiotics were administered according to an extended infusion protocol (all
patients received a loading dose, 1 g MER or 4 g PIP, followed immediately by
the first dose of either antibiotics in 6 h or 8 h intervals for MER or PIP,
respectively). 71 % of patients receiving PIP, and 46 % of patients receiving
MER, reached 100 % fT>MIC [60]. In contrast, our study showed that the rate of
51
target attainment was highest for the patients receiving MER (80 % reached 100
% fT>MICECOFF). By using TDM in the Belgian study the target attainment was
increased to 95 % compared to 68 % in the control group [60]. In the BLISS
study investigating the effect of CI for cefepime, PIP/taz and MER on target
attainment, the control group receiving empirical IB dosing had lower target
attainment than in our study, and the DALI study, as only 37 % and 36 %
reached 100 % fT>MICECOFF on days one and three, respectively [175]. This
could be partly explained by the fact that the patients in the control group did
not receive a loading dose at the initiation of therapy, as is recommended for
critically ill patients in some guidelines [3, 8, 175]. An even lower rate of target
attainment was detected by Dulhunty et al. in an RCT investigating CI versus
IB for cefepime, MER and PIP/taz, as only 29 % of patients in the IB group
reached 100 % fT>MICECOFF [61].
We found that when the actual MIC was used instead of MICECOFF, the
proportion reaching the target increased from 55 % to 89 % (Paper I), clearly
indicating the importance of correlating drug concentrations to the actual MIC
of the target pathogen. In other studies on ICU patients, just as in ours, a
bacterial pathogen likely to cause the infection could only be isolated in 1/3 of
the patients [46]. When the actual pathogen cannot be isolated, the MICECOFF is
necessary for defining a target. This is also the case when initiating empirical
treatment, as it is impossible to know the MIC and the bacteriological etiology
at that stage. Knowledge of local resistance epidemiology is important in
initially selecting the correct dosing of antibiotics to reach targets. The
proportion of target attainment for broad spectrum β-lactam antibiotics is
dependent on the local resistance epidemiology. In many countries, infections
from resistant bacteria are much more common than in Sweden [170, 177],
which affects the drug concentrations needed to reach 100 % fT>MICACTUAL.
Using the MICECOFF, when MICACTUAL is not available in settings with a high
level of resistant bacterial infections, may thus overestimate target attainment
[178]. In a retrospective study of 23 patients infected with ESBL-producing E.
coli or Klebsiella spp. treated with PIP/taz outcome was successful for 91 % of
patients with pathogens with MICs ≤16 mg/L, while only successful for 20 %
when MICs were >16 mg/L [179].
In our study, increased renal clearance was identified as a risk factor for
target non-attainment. This is in agreement with previously published data that
also show an association between ARC and subtherapeutic β-lactam antibiotic
concentrations [58, 180, 181], underlining the importance of rapidly identifying
individuals with ARC as pointed out by Udy et al. [180] and Lipman et al. [182,
183]. The 30-day mortality was higher (41 % vs. 20 %, p<0.05) for the patients
reaching 100 % fT>MICECOFF. As expected, the patients with the highest SAPS3
score had the highest mortality but also high target PK/PD attainment, which
we interpret mainly as a result of impaired kidney function. Thus, target
attainment is confounded by a high disease severity which makes analysis of
TDM in relation to clinical outcome of this patient group difficult. These results
52
are indicated by earlier studies also showing that increased disease severity is
associated with higher target attainment [46, 58]. Future clinical trials on the
clinical relevance of TDM in ICU patients could preferably focus on younger
age groups without significant comorbidities, but with a risk for target nonattainment due to ARC.
TDM is rarely used for β-lactam antibiotics in clinical practice. In a survey
investigating the use of TDM for β-lactam antibiotics, >90 % of 402
respondents from 328 hospitals in 53 countries answered that they did not use
TDM [184]. It has however been shown in a RCT (n=41) that when using TDM
for ICU patients, an increased dose was needed in 76 %. In the intervention
group guided by TDM, 100 % fT>MIC was significantly more commonly reached
than in the control group (94.7 % versus 68.4 %) [60]. TDM did increase target
attainment, but the study was not sufficiently powered for evaluating mortality
[60]. Even though the data from the literature on TDM in the ICU regarding
patient outcomes is scarce, it is clear that due to increased VD and ARC, the
pharmacokinetics are unpredictable, leading to highly variable levels of target
attainment from consistent dosing regimens [185]. In the ICU, most data at hand
indicates that individual dosing guided by TDM might improve treatment
outcomes by compensating the unpredictable pharmacokinetics [20].
To improve target attainment during TDM, different dosing strategies might
be used, such as more frequent dosing or CI [186]. In our study, 11 patients
received either prolonged or continuous infusion. Nine patients, including all
receiving CI, reached 100 % fT>MICECOFF. Although the data is limited, our
results are in line with other studies investigating whether different dosing
regimens such as CI versus IB could affect target attainment. In the BLISS study
[175], 140 critically ill patients were treated with PIP, MER or cefepime, either
via CI (n=70) or IB (n=70). The target attainment (100 % fT>MIC) was
significantly higher in the CI group compared to the IB group (97 % vs. 70 %,
respectively), but no significant differences in 30-day survival were observed
(74 % vs. 63 %) [175]. Another prospective, multi-center, double blind RCT by
Dulhunty et al. [61] with 22 patients in each group showed a significantly higher
target attainment (82 % vs. 29 %) for CI. Clinical cure was also significantly
higher (70 % vs. 43 %) in the CI group, although there was no difference in
LOS at the ICU or survival [61]. Furthermore, the definition of clinical cure is
difficult, and in the Dulhunty study it was defined as disappearance of all signs
and symptoms related to the infection after 28 days [61]. A meta-analysis of
studies investigating CI and IB dosing [47] concluded that there is no evidence
favoring CI over IB dosing of β-lactam antibiotics with regards to clinical
efficacy. The included studies were highly heterogeneous regarding disease
severity, source of infections, pathogens and dosing regimens [47]. The
recommendations from this meta-analysis is that patients with severe sepsis and
septic shock should be given high and frequent doses, alternatively, prolonged
infusions [47]. Another meta-analysis with more stringent selection criteria
including only studies of patients with severe sepsis or septic shock, and hence
53
more homogenous patient groups, included three RCTs on CI versus IB for
critically ill patients [50]. In that study, hospital mortality (censored at 30 days)
was decreased, and clinical cure was higher with CI versus IB dosing [50]. The
study has some limitations as discussed by Taccone et al., including it being an
open-label clinical trial with an imbalance in disease severity, and MER
administration between the groups [48]. There was no difference in ICU-free
days or ICU mortality, and administration by CI was not independently
associated with clinical cure [50].
In summary, our results confirm the high variability of β-lactam antibiotic
concentrations in ICU patients. TDM could be used to avoid low concentrations
together with other potential strategies, including the use of CI in some patient
groups [60]. Although firm evidence in terms of RCTs with mortality as an
outcome is missing, the studies available point more to a favorable clinical
effect of TDM rather than the opposite. Clearly, patients with very low drug
exposures are unlikely to improve as fast as patients with adequate levels. It is
crucial to the concept of TDM to introduce units of microbiologists,
pharmacologists, infectious disease practitioners, and anesthesiologists to
efficiently use available tools and knowledge clinically.
Methods for measurements of β-lactam antibiotics
Two methods for determination of serum concentration of β-lactam
antibiotics were developed during the work of this thesis. The UPLC-MS/MS
method is more complex, requiring special equipment, and thus more suitable
for reference laboratories (Paper II). The microbiological method (Paper III)
does not require special equipment other than that already present in common
microbiological laboratories and is thus suitable for smaller laboratories.
Development of an ultra-performance liquid chromatography
tandem mass spectrometry method quantifying 11 antibiotics
The first method (Paper II) is based on UPLC-MS/MS where 11 different
compounds (benzylpenicillin, cloxacillin, cefotaxime, piperacillin, meropenem,
linezolid, levofloxacin, ciprofloxacin, moxifloxacin, clindamycin, and
rifampicin) can be quantified in the same sample, simultaneously. Imipenem
was first included in the method, but due to stability issues it was excluded from
further testing. Requirements to ensure stability of imipenem include additives
to the serum at collection, and during sample preparation [187, 188].
Since most antibiotics are hydrophilic and polar [189] a kinetex biphenyl RP
C18 column was chosen. The biphenyl can act as a weak ion-exchanger creating
longer retention times for polar compounds, in contrast to a C18 column alone,
risking co-elution of polar compounds [190]. In the method developed, the
lower limit of quantification (LLOQ) ranged from 0.31-2.34 mg/L, and the
upper limit of quantification ranged from 20-150 mg/L, depending on the
54
compound. The range was chosen from the targeted therapeutically range for
each drug.
Sample preparation was by protein precipitation with MeOH and the total
run time was 5 minutes. The method, has a high sample throughput with
reporting possible within a day. A similar method had a run time of 2.5 minutes,
but measured only five different compounds, all β-lactam antibiotics [89]. All
compounds were measured in ESI+ mode (Table 4). The method published by
Carlier et al. used the same precursor and product ions for meropenem and
piperacillin, as we did, in this method [89]. The precursor and product ions for
ciprofloxacin and levofloxacin were previously used in a study by Chen et al.
[191].
Interferences, co-eluting peaks or matrix effects were not observed for any
compound, or at any concentration level, except for detection of ion-suppression
for rifampicin. The matrix effect observed for rifampicin was corrected for by
the use of IS. Others have shown varied levels of matrix effects, such as ion
enhancement for MER when measured in serum, dependent on both the sample
and the compound analyzed [89, 97].
Inaccuracy and imprecision were within 15 % of target concentration for all
QCs, except for benzylpenicillin which had an inter-day imprecision of 16.1 %
(Table 6). Imprecision for all calibration standards was within limits (<15 %),
except for one of the benzylpenicillin standards (40 mg/L, CV 34.6 %). So,
overall, the validation requirements set by the EMA were reached [102].
Table 6. Inaccuracy and imprecision for the QCs for all included compounds. Standard
deviation SD, coefficient of variation (CV), relative error (RE) and quality control (QC).
55
An analysis of storage stability shows the importance of sample handling
and that all samples should be kept at -80 °C if not analyzed immediately,
especially for MER and PIP, as has also been demonstrated by others [192, 193].
Inaccuracy for was <20 % for PIP and CTX, but not MER for all the serum
samples collected from ICU patients during the study in Paper I analyzed with
both UPLC-MS/MS (Paper II) and LC-MS (Paper I). Agreement around clinical
cut was 97 % for CTX and PIP, 100 % for MER. There was a significant
correlation between both methods (r≥0.95) for PIP, CTX and MER.
There is clearly a need for a reference laboratory and a comprehensive QC
program for determination of antibiotic concentrations. This issue should be
addressed within the European quality assurance authorities as antibiotic
concentration analyses are now increasingly used [184].
Compound-specific deuterated internal standards were not used for each
compound in our method which could decrease accuracy and precision.
Furthermore, the samples collected from critically ill patients [178] were
analyzed using a highly diluted IS for MER and CTX, which may explain the
relatively high imprecision and poor accuracy. Future validation should include
systematic assessment of intraday inaccuracy, imprecision, and evaluation of
matrix effect.
Development of a microbiological method for quantification of
β-lactam antibiotics concentration
A microbiological method was developed with a performance sufficient for
the demands of clinical routine (Paper III). Both when analyzing spiked serum
samples (PIP 16 mg/L, MER 2 mg/L and CTX 4 mg/L) and clinical serum
samples, the inter-assay variation RE (%) and CV (%) were within the target,
set at ±20 %. There was a significant correlation between the microbiological
assay and LC-MS, used as the reference method for all compounds (r ≥0.86).
Around the clinical cut-offs the methods were in agreement for 87, 100 and 90
% for PIP (34/39), MER (18/18) and CTX (28/31), respectively.
Since this method detects antibiotic activity, it is limited to one antibiotic
and thus should not be used for patients treated with more than one antibiotic,
because this will lead to an overestimation of serum concentrations.
Furthermore, the method requires approximately one hour of hands-on time, but
more importantly, overnight incubation, thus delaying reporting when
compared to LC-MS-based methods.
Since LC-MS-based methods are not always readily available, the
microbiological assay could be of use, speeding up the process by at least 3-4
days, avoiding transportation, and analysis at reference centers. Additionally, it
can be used in low resource areas.
To our knowledge, there are, surprisingly, no other microbial methods for
measuring β-lactam antibiotics in patient serum samples. The two other studies
describing microbiological assays for measurements of antibiotics focused on
detection in food products [194, 195]. The aim of those methods was to detect
56
concentrations around the maximum residue limit (MRL) in food products [194,
195]. Both microbiological assays using conductimetric detection, as well as
luminescent bacterial biosensors, could be applied for TDM, although adaption
to serum samples, and assessment of performance around clinically important
cut-offs, are required.
The LOD for the method developed in Paper III was not systematically
tested, thus further analysis is needed. This also includes the antibacterial effects
of drugs other than antibiotics used for patient treatment. However, this was not
considered a major issue in the validation process of the current samples from
ICU patients. The method described by us is, thus, to our knowledge the first
microbiological assay tested for TDM of β-lactam antibiotics in ICU patients.
Concluding remarks – methods for measurements of β-lactam
antibiotics
Comparing the three different systems; microbiological assays, LC-MS/MS
methods, and immunoassays, all have their advantages (Table 7).
Table 7. Comparison between methods available for measurement of antibiotic
concentrations.
*Immunoassays are representative for ELISA, since immunochromatographic assays are
used for detection only.
Immunoassays are fast, and do not require special equipment. They are easy
to interpret, and can be automated, although they suffer from cross-reactivity,
especially for the detection of β-lactam antibiotics. Specificity is an issue and
immunoassays are more suitable for detection than quantification. In a recent
report using class-specific anti-β-lactam receptors in a gold immune-
57
chromatographic assay (GICA), 15 different β-lactam antibiotics could be
detected in milk samples below MLR [91], but could not be quantified and are
probably not applicable for TDM at this stage of development [91].
Microbiological assays are time consuming, and typically show results after 1620 hours, but do not require special equipment other than that available in a
routine microbiological laboratory. They are more difficult to interpret than
immunoassays, but measure the actual activity, and are thus suitable for
quantification. As microbiological methods do not need special equipment, the
method can be set up at any microbiological lab, including low-resource
settings. LC-MS/MS has high turn-around and minimal hands-on time, but
requires expensive special equipment, and specially trained staff. With LCMS/MS it is possible to measure several compounds in the same analysis using
the same sample. The method is highly accurate, with excellent reproducibility.
Strategies for transmission investigation of bacteria
In this thesis, the performance of an in-house, molecular typing method
designed as a closed tube system (LMqPCR HRMA), was investigated. The
method with HRMA was compared to the resolution of the gold standard
method PFGE at the time of the study, and in subsequent experiments to the
new gold standard MPS. Only when results from HRMA showed similarity
between isolates, gel electrophoresis was performed in an effort to enhance
resolution. The molecular method (LMqPCR HRMA) was developed from a
previously described method, and adapted for quantitative PCR and high
resolution melting curve analysis, instead of conventional gel electrophoresis
[127, 128] (Papers VI and V).
Ligation mediated qPCR high resolution melt analysis
Using LMqPCR HRMA, outbreak isolates could be separated from nonoutbreak isolates for a number of species including E. coli (Paper VI) and the
ESKAPE pathogens (Paper V). The results from LMqPCR HRMA were
compared to PFGE. For E. coli 2/15 outbreak isolates had discrepant results,
with lower clustering compared to the other isolates (81% and 68 %, versus >90
% similarity) using PFGE. Furthermore, four controls could not be clearly
separated by HRMA using the algorithm. Evaluating the complete HRM
spectrum resolved two of the isolates. For the remaining two, differentiation
could be achieved after analysis with capillary gel electrophoresis. For species
with a genomic G+C content above 50 %, results were easier to interpret. This
is due to a higher G+C content resulting in larger differences in TM [132, 133]
causing better separation of melting peaks.
LMqPCR HRMA was, however, unable to detect differences between M.
tuberculosis isolates (Fig 18). Another cleavage and ligation protocol was tested
and adapted to HRM analysis [169] without success, whereby no differences
58
between isolates could be detected. This might be due to poor restriction
enzyme activity, or a limited number of sites. However, the most likely
explanation is that the genetic set up of M. tuberculosis is too similar among
isolates, requiring other methods with higher resolution, such as WGS analysis.
Figure 18. LMqPCR HRMA for M. tuberculosis. Results from LMqPCR HRMA (A) and
subsequent agarose gel electrophoresis (B).
59
LMqPCR HRMA could work as a first screening tool, selecting isolates
warranting further investigation for the most common bacterial species causing
outbreaks in a hospital environment, such as the ESKAPE pathogens [21]. By
screening with LMqPCR HRMA, the reporting time will be greatly reduced,
whereby results are available within a day, compared to sending the isolates for
molecular typing in a reference laboratory using PFGE or MPS. LMqPCR
HRMA requires no special equipment other than that already present in most
microbiology laboratories equipped with a qPCR with HRM. For laboratory
workers experienced in molecular biology, the hands-on procedure is straightforward.
There are no other studies using HRMA on complete bacterial genomes for
transmission investigation. However, spa-typing was performed using HRM for
44 MRSA isolates, which resulted in 20 different HRM profiles from 22 spa
types [196]. The utility of HRMA in spa-typing was further demonstrated by
Chen et al. analyzing 55 clinical MRSA isolates with 12 different spa types,
with complete agreement between HRMA and the conventional method [197].
HRMA was also tested as an alternative to sequencing in MLST typing of
Campylobacter jejuni, and resolved 92 % of alleles tested [198]. The
discriminatory power of HRMA is, however, dependent on the type of SNP to
be detected, since SNPs not changing the number of bindings (i.e. A to T, T to
A or G to C, C to G) will not result in large enough TM differences to be detected
by HRMA [132].
Matrix-assisted laser desorption/ionization time of flight
Recent studies indicated that MALDI TOF could be used for transmission
investigation, thus we explored this possibility. The same set of isolates (E. coli
and the ESKAPE pathogens) used for LMqPCR HRMA were analyzed with
MALDI TOF. We failed to adequately differentiate between outbreak and nonoutbreak isolates. Both the single protein profile and MSP resulted in
inconclusive typing, since both control and outbreak isolates clustered together
(Fig. 19 A-B). The Maldi biotyper software v 3.0 used for analyzing isolate
clustering did not allow for a comparison, whereby clustering completely
changed when adding subsequent isolates. No peaks could be identified as
cluster-specific, or used as biomarkers for a certain isolate (Fig. 19 C).
60
Figure 19. Results from epidemiological typing of E. coli using MALDI TOF. Isolates were
run in triplicates and analyzed as single spectrum via hierarchical clustering and complete
correlation using Maldi Biotyper software v 3.0 (A). MSP were created from each isolate
and compared by complete correlation using Maldi Biotyper software v 3.0 (B). Analysis of
peaks specific for outbreak isolates (green) and control isolates (blue) (C).
Further analysis was performed by importing the spectra into Bionumerics
and creating summary spectra of a single spectra (analogues to MSP).
Dendrograms were created using the Pearson correlation or Dice and UPGMA
61
clustering (Fig. 20). There was a poor correlation between PFGE as the
reference method and MALDI TOF, irrespective of which software or
clustering algorithm were used. Results from MALDI TOF typing compared to
PFGE and HRMA (without gel electrophoresis), of E. faecium, S. aureus, K.
pneumoniae and P. aeruginosa are summarized in Table 8 (p.64). Cluster
analysis of A. baumannii was not performed by MALDI TOF due to difficulties
in retrieving high quality spectra.
Figure 20. Results from epidemiological clustering of E. coli using MALDI TOF and
analysis by Bionumerics. Summary spectra are clustered by Pearson correlation and
UPGMA (A) and Dice coefficient and UPGMA (B). O=outbreak isolate.
62
Our results are confirmed by Lasch et al., who investigated the typeability
of MALDI TOF using TFA extraction and spectra collection between 2 and 20
kDa for E. faecium (n=112), and S. aureus (n=59). The results showed
insufficient resolution for both species [199]. They also concluded that clustertype specific biomarkers (peaks) were the exception rather than the rule [199].
However, for Salmonella, genus-, species-, and subspecies-specific markers in
the range between 20-30 kDa have been identified [200]. Furthermore, a peak
specific for the EHEC pathotype was identified, but in the same work no m/z
ranges were found to be absolutely diagnostic of an E. coli pathotype, serotype
or ST [201]. In Vibrio cholera, a higher mass range than is used during normal
bacterial typing (2-20 kDa) was analyzed, and a 34 kDa marker, specific for
discrimination between toxic and epidemic, and non-epidemic strains was
found [202]. In the light of studies on Salmonella and Vibro spp., investigating
a higher mass range might identify markers that can be used for differentiation
between clones, even for E. coli or the ESKAPE pathogens. For E. coli the mass
range was expanded to 140 kDa using a pretreatment protocol, including
surfactants, thus enabling a search for other markers [203].
63
Table 8. Comparison of typing results between PFGE, HRMA and MALDI TOF.
Ef=Enterococcus faecium, Sa=Staphylococcus aureus, Kp=Klebsiella pneumoniae,
Ps=Pseudomonas aeruginosa.
aArbitrary
classification of PFGE types within each species. A, B, C etc. denote specific
types, >97 % similarity, whereas A1, A2 etc. denote closely related patterns, 90-97 %
similarity. bSum spectra and Pearson correlation curve based analysis. cSum spectra and
Pearson correlation peak based analysis.
64
MALDI TOF may be used for transmission investigation for some species
such as C. difficile, although algorithms to differentiate between isolates need
to be developed, taking into account unstable peaks as well as peak height and
width. Karger et al. listed three main areas for increasing the typeability of
MALDI TOF, the sample preparation, the mass spectrometric process, and the
statistical evaluation [150]. By alternating the sample preparation, and mass
spectrometric settings, different mass ranges can be detected [200, 202, 203].
The last issue to solve is the statistical evaluation where previous attempts,
including our work, have not as yet been successful [199, 201]. One possibility
could be to analyze mass shifts for clonal differentiation, since they indicate
mutations in the underlying protein, as suggested by Josten et al. [204].
Combining methods is also a possibility, and amplification of MLST genes
followed by fragmentation and detection with MALDI TOF [205] eliminates
the need for a sequencing step, thus making the method more available. This
will, however, only generate a typeability similar to that of MLST.
Whole genome sequencing
Due to extensive technical development in the last 10 years, MPS has been
made available to almost every lab, and thus allowing for whole genome
sequencing, since several isolates can be analyzed at the same time. The current
dilemma with MPS is that there is no consensus concerning quality controls,
and how to analyze and report MPS data.
By using MPS, followed by cgMLST and SNP analysis when investigating
a clinically identified nosocomial outbreak of E. coli ST-131, we were able to
clearly identify two isolates of 16 not belonging to the outbreak (Paper VI). By
including the genome sequence of the non-related reference isolate JJ1886, with
the same MLST type (ST-131) in the cgMLST analysis, it was possible to
suggest a threshold for differing alleles within a cluster for this ST. In the SNP
analysis however, using JJ1886 instead of the outbreak index isolate for
comparison, resulted in fewer differences between the two discrepant isolates
and the rest of the ST-131 isolates. This was probably due to the fact that a large
number of the sequences from the outbreak suspected ST-131 isolates are not
present in JJ1886. Hence, the SNP analysis using JJ1886 as a reference was
based on a smaller amount of data. Thus, using a universal reference could
decrease resolution while using only local isolates would allow for only local
comparisons, and not between different outbreaks and laboratories. The
conclusion that a closely matched reference genome is important was also
emphasized in a study investigating molecular typing of Legionella
pneumophila [206]. In that study, sequence reads were mapped to a closely
related reference genome to minimize the number of unmapped reads, or a high
number of false positive SNPs. As a consequence, isolates mapped using
different reference genomes were not comparable, but the use of different
references is by itself indicative of a distant, rather than a close, relationship
[206].
65
In our study (Paper VI), both cgMLST and SNP analysis resulted in the same
clusters, with higher differentiation within the cluster of outbreak isolates using
SNP analysis. Nevertheless, the reference-based SNP analysis of E. coli might
not be the most comprehensive approach, since E. coli is a very diverse
organism with a high level of recombination [207]. It may result in false
negative clustering, because short non-allelic sequences may be incorrectly
incorporated during assembly, and then misinterpreted as polymorphisms [208,
209]. As discussed by Leopold et al. cluster analysis by cgMLST instead of SNP
analysis is preferable since allele-based comparisons treat each change, both
SNP and recombination, causing several nucleotide changes, as one
evolutionary event [210]. Others have also compared cgMLST to SNP analysis.
Willems et al. investigated clusters of A. baumannii using cgMLST and SNP
analysis, both analyses gave the same results [211]. For 335 isolates of L.
pneumophilia, the best compromise between discriminatory power and
epidemiological concordance was achieved using a cgMLST, including 50
alleles [206]. SNP analysis and cgMLST schemes including 500, 1455, and
1521 alleles not only resulted in complete differentiation between unrelated
isolates, but also differentiation of clearly epidemiologically linked isolates,
depending on the threshold [206]. These results underline the importance of
clearly set thresholds for differentiating between unrelated and
epidemiologically linked isolates. These thresholds must be developed for every
species in agreement with epidemiological data.
Even if WGS-based methods lack standardized analysis protocols, cgMLST
and SNP analyses finally resolved the epidemiological links within the
nosocomial outbreak, due to improved discriminatory power (Paper IV). The
cgMLST approach has been used for ESBL-producing Enterobacteriaceae spp.
further stressing the usefulness of this analysis method for epidemiological
typing, as they also conclude that cgMLST has higher discriminatory power
than in silico MLST [212].
Concluding the results from the experiments on molecular typing the
possibilities for a pan-species-consensus for WGS molecular typing appear to
be slim, depending on the diversity in genetic make-up between bacterial
species. It should be possible, however, to develop a scheme where the level of
discriminatory power increases by every step, beginning with regular MLST
and finishing with an SNP analysis. The molecular typing methods to be used
would depend on the specific typing question, as well as the bacteria of interest.
M. tuberculosis would benefit from SNP analysis, since the rate of genomic
change is estimated to be 0.5 SNP per year by Walker et al. [113], while for
species belonging to Enterobacteriaceae cgMLST would, in most cases, be
enough, as proposed by Kluytmans-van den Bergh et al. [212]. The ECDC has
also proposed the same approach, were cgMLST should be used as a first step,
and if further discrimination is needed, SNP analysis should be applied [163].
66
Concluding remarks - strategies for transmission investigation
of bacteria
Based on our work and others’, two possible scenarios for transmission
investigation are proposed. For laboratories with resources for instrumentation
and personnel, real time tracking is possible for isolated bacteria using MPS and
cgMLST, and/or SNP analysis. Real time molecular typing has been
successfully implemented, together with ceased isolation of patients with
suspected ESBL Gram-negative infection in risk wards, with no increase and in
transmission rate [213]. Whether such strategies are possible in routine care in
smaller hospitals remains to be investigated. For other laboratories, in hospitals
dealing mainly with E. coli and ESKAPE pathogens, LMqPCR HRMA will
work as a primary screening tool, which may exclude the possibility of
nosocomial outbreaks and even detect similarity with a high performance. In
some cases where the clinical link is less well established, similar isolates
require further investigation using MPS followed by either cgMLST or SNP
analysis. PFGE still constitutes a useful tool when comparing new isolates to
already PFGE-typed isolates (Table 9).
Table 9. Comparison between methods for epidemiological typing of bacteria.
*post culture. +poor, ++low, +++moderate, ++++good, +++++high, ++++++excellent.
One of the challenges of molecular typing methods is to more clearly show
their value for limiting outbreaks in real-time, and the additive effect on the
tools and epidemiological investigations already at hand. If more clinical
microbiological laboratories start using WGS routinely, the information
retrieved could be utilized in clinical decision-making. Importantly, this
67
requires close collaboration between microbiologists, molecular biologists,
hospital hygiene departments, and hospital wards.
WGS by MPS has other applications, and for some species the entire
diagnostic process could be replaced. One such species is M. tuberculosis,
which is tedious to culture, delaying microbial diagnostics and AST by weeks.
During a study the capacity of WGS as a tool for diagnosis, molecular typing,
and AST of M. tuberculosis, was assessed [214]. The conclusions were that
WGS provided the diagnosis and AST at the same rate as routine methods, with
93 % concordance. On the other hand, using WGS was 7 % cheaper, and also
provided information about transmission [214]. The role of WGS for AST
against M. tuberculosis, however, in settings where resistance is uncommon is
not clear, and the method cannot as yet replace current phenotypic susceptibility
testing.
68
Conclusions
 45% of patients from ICUs in Southeast Sweden do not reach 100 %
fT>MIC for β-lactam antibiotics. The target attainment is highly
dependent on the MIC of the pathogen.
 Increased renal clearance was identified as a main risk factor for target
non-attainment.
 A rapid UPLC-MS/MS method for simultaneous quantification of 11
different antibiotics was developed, and showed an inter-day precision
and accuracy within the target of ±15%.
 A microbiological method for quantification of β-lactam antibiotics in
serum was developed, and showed an agreement with an LC-MS
method around the clinical cut-off for clinical samples of 90 %, 100 %,
and 87 % for CTX (4.0 mg/L), MER (2 mg/L) and PIP (16 mg/L),
respectively.
 A new method based on melting curve analysis (LMqPCR HRMA) was
developed and characterized a nosocomial outbreak of ESBLproducing E. coli. The performance for the ESKAPE pathogens was
comparable to PFGE, which was used as a reference method at that
time.
 Whole genome sequencing followed by cgMLST or SNP analysis
resulted in enhanced discriminatory power compared to PFGE and
LMqPCR HRMA within the nosocomial outbreak of ESBL-producing
E. coli.
69
Future perspectives
Empirical treatment of bacterial infection results in lower antibiotic serum
concentrations than expected in critically ill patients. Patients with increased
renal clearance are particularly at risk for insufficient antibiotic serum
concentrations. Different strategies can be applied to ensure higher serum
concentration of antibiotics, such as increased frequency of dosing or CI. Since
it is difficult to predict which patients will develop unstable pharmacokinetics,
TDM is a helpful tool for guiding treatment of critically ill patients. To
individualize and optimize treatment, a patient with a suspected infection should
receive aggressive empiric antibiotic treatment based on the suspected etiology
and site of infection, including local resistance epidemiology. Within two days,
efforts to identify the actual pathogen should be made, and if possible, MIC
determined. Serum levels of antibiotics should be analyzed and as soon as
possible and if available be combined with MIC and susceptibility patterns of
the bacterial pathogen. By this strategy, information from both microbiology
and pharmacology laboratories will provide important guidance back to the
clinic and patient, resulting in possibilities to optimize treatment (Fig. 21).
Figure 21. The aspects of therapeutic drug monitoring.
70
The clinical applications of molecular typing methods are detection of
outbreaks, the possibility to distinguish relapse from re-infection, identification
of clones with increased transmission capacity, and virulence, as well as new
emerging clones. Typing data should be evaluated together with a rigorous
epidemiological investigation (Fig. 22). For successful transmission
investigation areas within the health care sector need to work together enabling
efficient sampling, isolation, analysis, and results.
Figure 22. Clinical applications of bacterial molecular typing.
71
Acknowledgments
I would like to take this opportunity to thank the people who have supported
and aided me in the work leading up to this thesis. I would also like to thank
those who have helped me to forget, from time to time, the struggles associated
with clinical research.
First and foremost, huge thanks to my main supervisor, Thomas Schön.
Thank you for giving me the opportunity to work with you, and for believing in
me. Thank you for always being available, for guidance, and navigation, and
for completely scrutinizing every manuscript and abstract, sharing your
exceptional skills within clinical research.
I would also like to express my gratitude to my co-supervisors Björn
Carlsson and Britt-Inger Marklund for scientific and academic input,
knowledge, and support through-out my work.
My completion of this thesis would not have been possible without the
support of my co-authors of the six papers. Thank you for your constructive
criticism and support. A special thanks to Jan Söderman for his valuable input
and to Michela Jonsson Nordvall and Salomon Ghebremichael for teaching
me the work in a P3-lab.
Thanks to all the people who have contributed with samples to the studies in
this thesis, to the staff members who have coordinated the studies in the
participating departments, and to those who have obtained compliance data and
collected samples day and night.
Thanks to Serhiy Souchelnytskyi and the former TGF-β group at Ludwig
Institute for Cancer Research in Uppsala for introducing me to the field of
biomedical research.
Thank you, Anna Ryberg for all your wisdom and friendship.
72
I would also like to thank all my co-workers at Clinical Microbiology,
Kalmar County Hospital and at Clinical Pharmacology, Linköping University
Hospital for making my work days more fun, and for always creating a place
for my research. A special thanks to Mattias, Heléna, Anna, Ulf, Berit and
Marcus.
Tack till alla er som hjälpt mig att glömma allt vad forskning och avhandling
heter. Tack till mina föräldrar, Anita och Thomas som ständigt skämmer bort
mig och min familj. Tack till min syster och hennes sambo, Camilla och
Torbjörn för träningspass, skvaller och Amarone. Tack till klanen Woksepp
för de alltid så trevliga sammankomsterna. Tack till Denise och Anders, Sofia
och Anders, Monika, Jane och Daniel och Tobias.
Tack till alla medlemmar i KalmarSim och mastersgruppen för
gemenskap och för att åter väckt min glädje till simning.
Slutligen, ett stort tack till min man Jim och vår underbara dotter Emilia för
att ni påminner mig om vad som är viktigast i världen. Jag älskar er!
73
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
74
Kwong JC, McCallum N, Sintchenko V, Howden BP (2015)
Whole genome sequencing in clinical and public health
microbiology. Pathology 47:199-210
Singer M, Deutschman CS, Seymour CW, Shankar-Hari M,
Annane D, Bauer M, Bellomo R, Bernard GR, Chiche JD,
Coopersmith CM, Hotchkiss RS, Levy MM, Marshall JC,
Martin GS, Opal SM, Rubenfeld GD, van der Poll T, Vincent
JL, Angus DC (2016) The Third International Consensus
Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA
315:801-810
Andersson M, Brink M, Cronqvist J, Furebring M, GilleJohansson P, Gårdlund B, Lanbeck P, Ljungström L, Mehle
C, Otto G, Sjölin J, Svefors J, Vikerfors T (2015)
Vårdprogram Svår sepsis och septisk chock - tidig
identifiering och initial handläggning. Svenska
Infektionsläkarföreningen: 95
Yan Y, Zhang S, Wu FX (2011) Applications of graph
theory in protein structure identification. Proteome Sci 9
Suppl 1:S17
Cavaillon JM, Eisen D, Annane D (2014) Is boosting the
immune system in sepsis appropriate? Crit Care 18:216
Banerjee S, Mazumdar S (2012) Electrospray ionization
mass spectrometry: a technique to access the information
beyond the molecular weight of the analyte. Int J Anal Chem
2012:282574
Al-Dorzi HM, Al Harbi SA, Arabi YM (2014) Antibiotic
therapy of pneumonia in the obese patient: dosing and
delivery. Curr Opin Infect Dis 27:165-173
Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H,
Opal SM, Sevransky JE, Sprung CL, Douglas IS, Jaeschke
R, Osborn TM, Nunnally ME, Townsend SR, Reinhart K,
Kleinpell RM, Angus DC, Deutschman CS, Machado FR,
Rubenfeld GD, Webb S, Beale RJ, Vincent JL, Moreno R
(2013) Surviving Sepsis Campaign: international guidelines
for management of severe sepsis and septic shock, 2012.
Intensive Care Med 39:165-228
Spellberg B, Bartlett JG, Gilbert DN (2013) The future of
antibiotics and resistance. N Engl J Med 368:299-302
World Health Organization (2015) Global Action Plan on
Antimicrobial Resistance. Geneva, Switzerland: 28
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Ministry of Health and Social Affairs (2016) Swedish
strategy to combat antibiotic resistance. Stockholm, Sweden:
24
Siegman-Igra Y, Fourer B, Orni-Wasserlauf R, Golan Y,
Noy A, Schwartz D, Giladi M (2002) Reappraisal of
community-acquired bacteremia: a proposal of a new
classification for the spectrum of acquisition of bacteremia.
Clin Infect Dis 34:1431-1439
Mehta Y, Gupta A, Todi S, Myatra S, Samaddar DP, Patil V,
Bhattacharya PK, Ramasubban S (2014) Guidelines for
prevention of hospital acquired infections. Indian J Crit Care
Med 18:149-163
Moreno RP, Metnitz B, Adler L, Hoechtl A, Bauer P,
Metnitz PG, Investigators S (2008) Sepsis mortality
prediction based on predisposition, infection and response.
Intensive Care Med 34:496-504
Kumar A, Roberts D, Wood KE, Light B, Parrillo JE,
Sharma S, Suppes R, Feinstein D, Zanotti S, Taiberg L,
Gurka D, Cheang M (2006) Duration of hypotension before
initiation of effective antimicrobial therapy is the critical
determinant of survival in human septic shock. Crit Care
Med 34:1589-1596
Zaragoza R, Artero A, Camarena JJ, Sancho S, Gonzalez R,
Nogueira JM (2003) The influence of inadequate empirical
antimicrobial treatment on patients with bloodstream
infections in an intensive care unit. Clin Microbiol Infect
9:412-418
Burnham JP, Lane MA, Kollef MH (2015) Impact of Sepsis
Classification and Multidrug-Resistance Status on Outcome
Among Patients Treated With Appropriate Therapy. Crit
Care Med 43:1580-1586
Udy AA, Roberts JA, De Waele JJ, Paterson DL, Lipman J
(2012) What's behind the failure of emerging antibiotics in
the critically ill? Understanding the impact of altered
pharmacokinetics and augmented renal clearance. Int J
Antimicrob Agents 39:455-457
Abdul-Aziz MH, Lipman J, Mouton JW, Hope WW, Roberts
JA (2015) Applying pharmacokinetic/pharmacodynamic
principles in critically ill patients: optimizing efficacy and
reducing resistance development. Semin Respir Crit Care
Med 36:136-153
Lamoth F, Wenger A, Prod’hom G, Vallet Y, Plüss-Suard C,
Bille J, Zanetti G (2010) Comparison of hospital-wide and
75
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
76
unit-specific cumulative antibiograms in hospital- and
community-acquired infection. Infection 38:249-253
Boucher HW, Talbot GH, Bradley JS, Edwards JE, Gilbert
D, Rice LB, Scheld M, Spellberg B, Bartlett J (2009) Bad
bugs, no drugs: no ESKAPE! An update from the Infectious
Diseases Society of America. Clin Infect Dis 48:1-12
Pogue JM, Kaye KS, Cohen DA, Marchaim D (2015)
Appropriate antimicrobial therapy in the era of multidrugresistant human pathogens. Clin Microbiol Infect 21:302312
Vandecandelaere I, Coenye T (2015) Microbial composition
and antibiotic resistance of biofilms recovered from
endotracheal tubes of mechanically ventilated patients. Adv
Exp Med Biol 830:137-155
Maiden MC, Bygraves JA, Feil E, Morelli G, Russell JE,
Urwin R, Zhang Q, Zhou J, Zurth K, Caugant DA, Feavers
IM, Achtman M, Spratt BG (1998) Multilocus sequence
typing: a portable approach to the identification of clones
within populations of pathogenic microorganisms. Proc Natl
Acad Sci U S A 95:3140-3145
van Belkum A (1994) DNA fingerprinting of medically
important microorganisms by use of PCR. Clin Microbiol
Rev 7:174-184
Gilchrist CA, Turner SD, Riley MF, Petri WA, Jr., Hewlett
EL (2015) Whole-genome sequencing in outbreak analysis.
Clin Microbiol Rev 28:541-563
Alberti C, Brun-Buisson C, Burchardi H, Martin C,
Goodman S, Artigas A, Sicignano A, Palazzo M, Moreno R,
Boulme R, Lepage E, Le Gall R (2002) Epidemiology of
sepsis and infection in ICU patients from an international
multicentre cohort study. Intensive Care Med 28:108-121
Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin
CD, Moreno R, Lipman J, Gomersall C, Sakr Y, Reinhart K,
Investigators EIGo (2009) International study of the
prevalence and outcomes of infection in intensive care units.
JAMA 302:2323-2329
Burgmann H, Hiesmayr JM, Savey A, Bauer P, Metnitz B,
Metnitz PG (2010) Impact of nosocomial infections on
clinical outcome and resource consumption in critically ill
patients. Intensive Care Med 36:1597-1601
Gotts JE, Matthay MA (2016) Sepsis: pathophysiology and
clinical management. BMJ 353:i1585
Hidron AI, Edwards JR, Patel J, Horan TC, Sievert DM,
Pollock DA, Fridkin SK, National Healthcare Safety
32.
33.
34.
35.
36.
37.
38.
39.
Network T, Participating National Healthcare Safety
Network F (2008) NHSN annual update: antimicrobialresistant pathogens associated with healthcare-associated
infections: annual summary of data reported to the National
Healthcare Safety Network at the Centers for Disease
Control and Prevention, 2006-2007. Infect Control Hosp
Epidemiol 29:996-1011
Socialstyrelsen (2016) Lägesrapport inom
patientsäkerhetsområdet 2016. 123
Hanberger H, Antonelli M, Holmbom M, Lipman J, Pickkers
P, Leone M, Rello J, Sakr Y, Walther SM, Vanhems P,
Vincent JL (2014) Infections, antibiotic treatment and
mortality in patients admitted to ICUs in countries
considered to have high levels of antibiotic resistance
compared to those with low levels. BMC Infect Dis 14:513
Zhanel GG, DeCorby M, Laing N, Weshnoweski B, Vashisht
R, Tailor F, Nichol KA, Wierzbowski A, Baudry PJ,
Karlowsky JA, Lagace-Wiens P, Walkty A, McCracken M,
Mulvey MR, Johnson J, Canadian Antimicrobial Resistance
A, Hoban DJ (2008) Antimicrobial-resistant pathogens in
intensive care units in Canada: results of the Canadian
National Intensive Care Unit (CAN-ICU) study, 2005-2006.
Antimicrob Agents Chemother 52:1430-1437
Bone RC, Sprung CL, Sibbald WJ (1992) Definitions for
sepsis and organ failure. Crit Care Med 20:724-726
Honore PM, Jacobs R, Hendrickx I, De Waele E, Van Gorp
V, Joannes-Boyau O, De Regt J, Boer W, Spapen HD (2016)
Biomarkers in critical illness: have we made progress? Int J
Nephrol Renovasc Dis 9:253-256
Pea F, Viale P, Furlanut M (2005) Antimicrobial therapy in
critically ill patients: a review of pathophysiological
conditions responsible for altered disposition and
pharmacokinetic variability. Clin Pharmacokinet 44:10091034
Udy AA, Baptista JP, Lim NL, Joynt GM, Jarrett P, Wockner
L, Boots RJ, Lipman J (2014) Augmented renal clearance in
the ICU: results of a multicenter observational study of renal
function in critically ill patients with normal plasma
creatinine concentrations*. Crit Care Med 42:520-527
Wong G, Briscoe S, Adnan S, McWhinney B, Ungerer J,
Lipman J, Roberts JA (2013) Protein binding of beta-lactam
antibiotics in critically ill patients: can we successfully
predict unbound concentrations? Antimicrob Agents
Chemother 57:6165-6170
77
40.
41.
42.
43.
44.
45.
46.
47.
48.
78
Muscedere JG, Shorr AF, Jiang X, Day A, Heyland DK,
Canadian Critical Care Trials G (2012) The adequacy of
timely empiric antibiotic therapy for ventilator-associated
pneumonia: an important determinant of outcome. J Crit
Care 27:322 e327-314
Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH
(2000) The influence of inadequate antimicrobial treatment
of bloodstream infections on patient outcomes in the ICU
setting. Chest 118:146-155
Vincent JL, Bassetti M, Francois B, Karam G, Chastre J,
Torres A, Roberts JA, Taccone FS, Rello J, Calandra T, De
Backer D, Welte T, Antonelli M (2016) Advances in
antibiotic therapy in the critically ill. Crit Care 20:133
Kumar A, Zarychanski R, Light B, Parrillo J, Maki D, Simon
D, Laporta D, Lapinsky S, Ellis P, Mirzanejad Y, Martinka
G, Keenan S, Wood G, Arabi Y, Feinstein D, Kumar A,
Dodek P, Kravetsky L, Doucette S, Cooperative
Antimicrobial Therapy of Septic Shock Database Research G
(2010) Early combination antibiotic therapy yields improved
survival compared with monotherapy in septic shock: a
propensity-matched analysis. Crit Care Med 38:1773-1785
Goncalves-Pereira J, Povoa P (2011) Antibiotics in critically
ill patients: a systematic review of the pharmacokinetics of
beta-lactams. Crit Care 15:R206
Trotman RL, Williamson JC, Shoemaker DM, Salzer WL
(2005) Antibiotic dosing in critically ill adult patients
receiving continuous renal replacement therapy. Clin Infect
Dis 41:1159-1166
Roberts JA, Paul SK, Akova M, Bassetti M, De Waele JJ,
Dimopoulos G, Kaukonen KM, Koulenti D, Martin C,
Montravers P, Rello J, Rhodes A, Starr T, Wallis SC,
Lipman J, Study D (2014) DALI: defining antibiotic levels in
intensive care unit patients: are current beta-lactam antibiotic
doses sufficient for critically ill patients? Clin Infect Dis
58:1072-1083
Leander G, Eliasson E, Hanberger H, Giske C (2015) [Beta
lactam antibiotics and the question of dose regimen for
severe infection. Prolonged infusion theoretically appealing-yet no evidence of clinical benefit]. Lakartidningen 112
Taccone FS, Laupland KB, Montravers P (2016) Continuous
infusion of beta-lactam antibiotics for all critically ill
patients? Intensive Care Med 42:1604-1606
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
Grupper M, Kuti JL, Nicolau DP (2016) Continuous and
Prolonged Intravenous beta-Lactam Dosing: Implications for
the Clinical Laboratory. Clin Microbiol Rev 29:759-772
Roberts JA, Abdul-Aziz MH, Davis JS, Dulhunty JM, Cotta
MO, Myburgh J, Bellomo R, Lipman J (2016) Continuous
versus Intermittent beta-Lactam Infusion in Severe Sepsis. A
Meta-analysis of Individual Patient Data from Randomized
Trials. Am J Respir Crit Care Med 194:681-691
Ferrer R, Martin-Loeches I, Phillips G, Osborn TM,
Townsend S, Dellinger RP, Artigas A, Schorr C, Levy MM
(2014) Empiric antibiotic treatment reduces mortality in
severe sepsis and septic shock from the first hour: results
from a guideline-based performance improvement program.
Crit Care Med 42:1749-1755
Craig WA (1998) Pharmacokinetic/pharmacodynamic
parameters: rationale for antibacterial dosing of mice and
men. Clin Infect Dis 26:1-10; quiz 11-12
Craig WA (1998) Choosing an antibiotic on the basis of
pharmacodynamics. Ear Nose Throat J 77:7-11; discussion
11-12
Cars O (1997) Efficacy of beta-lactam antibiotics: integration
of pharmacokinetics and pharmacodynamics. Diagn
Microbiol Infect Dis 27:29-33
Drusano GL (2004) Antimicrobial pharmacodynamics:
critical interactions of 'bug and drug'. Nat Rev Microbiol
2:289-300
Li C, Du X, Kuti JL, Nicolau DP (2007) Clinical
pharmacodynamics of meropenem in patients with lower
respiratory tract infections. Antimicrob Agents Chemother
51:1725-1730
McKinnon PS, Paladino JA, Schentag JJ (2008) Evaluation
of area under the inhibitory curve (AUIC) and time above the
minimum inhibitory concentration (T>MIC) as predictors of
outcome for cefepime and ceftazidime in serious bacterial
infections. Int J Antimicrob Agents 31:345-351
De Waele JJ, Lipman J, Akova M, Bassetti M, Dimopoulos
G, Kaukonen M, Koulenti D, Martin C, Montravers P, Rello
J, Rhodes A, Udy AA, Starr T, Wallis SC, Roberts JA (2014)
Risk factors for target non-attainment during empirical
treatment with beta-lactam antibiotics in critically ill
patients. Intensive Care Med 40:1340-1351
Wong G, Brinkman A, Benefield RJ, Carlier M, De Waele
JJ, El Helali N, Frey O, Harbarth S, Huttner A, McWhinney
B, Misset B, Pea F, Preisenberger J, Roberts MS, Robertson
79
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
80
TA, Roehr A, Sime FB, Taccone FS, Ungerer JP, Lipman J,
Roberts JA (2014) An international, multicentre survey of
beta-lactam antibiotic therapeutic drug monitoring practice in
intensive care units. J Antimicrob Chemother 69:1416-1423
De Waele JJ, Carrette S, Carlier M, Stove V, Boelens J,
Claeys G, Leroux-Roels I, Hoste E, Depuydt P,
Decruyenaere J, Verstraete AG (2014) Therapeutic drug
monitoring-based dose optimisation of piperacillin and
meropenem: a randomised controlled trial. Intensive Care
Medicine 40:380-387
Dulhunty JM, Roberts JA, Davis JS, Webb SA, Bellomo R,
Gomersall C, Shirwadkar C, Eastwood GM, Myburgh J,
Paterson DL, Lipman J (2013) Continuous infusion of betalactam antibiotics in severe sepsis: a multicenter doubleblind, randomized controlled trial. Clin Infect Dis 56:236244
Aminov R (2016) History of antimicrobial drug discovery Major classes and health impact. Biochem Pharmacol
Bush K, Bradford PA (2016) beta-Lactams and betaLactamase Inhibitors: An Overview. Cold Spring Harb
Perspect Med 6
Sauvage E, Kerff F, Terrak M, Ayala JA, Charlier P (2008)
The penicillin-binding proteins: structure and role in
peptidoglycan biosynthesis. FEMS Microbiol Rev 32:234258
Zapun A, Contreras-Martel C, Vernet T (2008) Penicillinbinding proteins and beta-lactam resistance. FEMS
Microbiol Rev 32:361-385
King DT, Sobhanifar S, Strynadka NC (2016) One ring to
rule them all: Current trends in combating bacterial
resistance to the beta-lactams. Protein Sci 25:787-803
Wright AJ (1999) The penicillins. Mayo Clin Proc 74:290307
Perry CM, Markham A (1999) Piperacillin/tazobactam: an
updated review of its use in the treatment of bacterial
infections. Drugs 57:805-843
Patel KB, Nicolau DP, Nightingale CH, Quintiliani R (1995)
Pharmacokinetics of cefotaxime in healthy volunteers and
patients. Diagn Microbiol Infect Dis 22:49-55
Papp-Wallace KM, Endimiani A, Taracila MA, Bonomo RA
(2011) Carbapenems: past, present, and future. Antimicrob
Agents Chemother 55:4943-4960
Craig WA (1997) The pharmacology of meropenem, a new
carbapenem antibiotic. Clin Infect Dis 24 Suppl 2:S266-275
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
Karam G, Chastre J, Wilcox MH, Vincent JL (2016)
Antibiotic strategies in the era of multidrug resistance. Crit
Care 20:136
Chambers HF (1999) Penicillin-binding protein-mediated
resistance in pneumococci and staphylococci. J Infect Dis
179 Suppl 2:S353-359
Falagas ME, Karageorgopoulos DE (2009) Extendedspectrum beta-lactamase-producing organisms. J Hosp Infect
73:345-354
Livermore DM (2002) Multiple mechanisms of antimicrobial
resistance in Pseudomonas aeruginosa: our worst nightmare?
Clin Infect Dis 34:634-640
Hiasa H, Shea ME (2000) DNA gyrase-mediated wrapping
of the DNA strand is required for the replication fork arrest
by the DNA gyrase-quinolone-DNA ternary complex. J Biol
Chem 275:34780-34786
Maxwell A (1992) The molecular basis of quinolone action.
J Antimicrob Chemother 30:409-414
Rotschafer JC, Ullman MA, Sullivan CJ (2011) Optimal use
of fluoroquinolones in the intensive care unit setting. Crit
Care Clin 27:95-106
(2013) Guidelines for Clinical and Operational Management
of Drug-Resistant Tuberculosis. Paris, France
Berdal JE, Skramm I, Mowinckel P, Gulbrandsen P,
Bjornholt JV (2005) Use of rifampicin and ciprofloxacin
combination therapy after surgical debridement in the
treatment of early manifestation prosthetic joint infections.
Clin Microbiol Infect 11:843-845
Leach KL, Swaney SM, Colca JR, McDonald WG, Blinn JR,
Thomasco LM, Gadwood RC, Shinabarger D, Xiong L,
Mankin AS (2007) The site of action of oxazolidinone
antibiotics in living bacteria and in human mitochondria. Mol
Cell 26:393-402
Jones RN, Kohno S, Ono Y, Ross JE, Yanagihara K (2009)
ZAAPS International Surveillance Program (2007) for
linezolid resistance: results from 5591 Gram-positive clinical
isolates in 23 countries. Diagn Microbiol Infect Dis 64:191201
Eckmann C, Dryden M (2010) Treatment of complicated
skin and soft-tissue infections caused by resistant bacteria:
value of linezolid, tigecycline, daptomycin and vancomycin.
Eur J Med Res 15:554-563
Tenson T, Lovmar M, Ehrenberg M (2003) The mechanism
of action of macrolides, lincosamides and streptogramin B
81
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
82
reveals the nascent peptide exit path in the ribosome. J Mol
Biol 330:1005-1014
Smieja M (1998) Current indications for the use of
clindamycin: A critical review. Can J Infect Dis 9:22-28
Sensi P (1983) History of the development of rifampin. Rev
Infect Dis 5 Suppl 3:S402-406
Forrest GN, Tamura K (2010) Rifampin combination therapy
for nonmycobacterial infections. Clin Microbiol Rev 23:1434
Verdier MC, Tribut O, Tattevin P, Le Tulzo Y, Michelet C,
Bentue-Ferrer D (2011) Simultaneous determination of 12
beta-lactam antibiotics in human plasma by highperformance liquid chromatography with UV detection:
application to therapeutic drug monitoring. Antimicrob
Agents Chemother 55:4873-4879
Carlier M, Stove V, De Waele JJ, Verstraete AG (2015)
Ultrafast quantification of beta-lactam antibiotics in human
plasma using UPLC-MS/MS. J Chromatogr B Analyt
Technol Biomed Life Sci 978-979:89-94
Chen Y, Chen Q, Han M, Liu J, Zhao P, He L, Zhang Y, Niu
Y, Yang W, Zhang L (2016) Near-infrared fluorescencebased multiplex lateral flow immunoassay for the
simultaneous detection of four antibiotic residue families in
milk. Biosens Bioelectron 79:430-434
Chen Y, Wang Y, Liu L, Wu X, Xu L, Kuang H, Li A, Xu C
(2015) A gold immunochromatographic assay for the rapid
and simultaneous detection of fifteen beta-lactams.
Nanoscale 7:16381-16388
Crossley KB, Rotschafer JC, Chern MM, Mead KE, Zaske
DE (1980) Comparison of a radioimmunoassay and a
microbiological assay for measurement of serum
vancomycin concentrations. Antimicrob Agents Chemother
17:654-657
Erickson RC, Hildebrand AR, Hoffman PF, Gibson CB
(1989) A sensitive bioassay for teicoplanin in serum in the
presence or absence of other antibiotics. Diagn Microbiol
Infect Dis 12:235-241
Cháfer-Pericás C, Maquieira Á, Puchades R (2010) Fast
screening methods to detect antibiotic residues in food
samples. Trends in Analytical Chemistry 29:1038-1049
Wilson JF, Davis AC, Tobin CM (2003) Evaluation of
commercial assays for vancomycin and aminoglycosides in
serum: a comparison of accuracy and precision based on
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
external quality assessment. J Antimicrob Chemother 52:7882
Lara F dO-IM, Cruces-Blanco C, Quesada-Molina C, GarcíaCampaña A. (2012) Advances in the determination of βlactam antibiotics by liquid chromatography. . vol. 38: 38523866
Cazorla-Reyes R, Romero-Gonzalez R, Frenich AG,
Rodriguez Maresca MA, Martinez Vidal JL (2014)
Simultaneous analysis of antibiotics in biological samples by
ultra high performance liquid chromatography-tandem mass
spectrometry. J Pharm Biomed Anal 89:203-212
Grebe SK, Singh RJ (2011) LC-MS/MS in the Clinical
Laboratory - Where to From Here? Clin Biochem Rev 32:531
Lange V, Picotti P, Domon B, Aebersold R (2008) Selected
reaction monitoring for quantitative proteomics: a tutorial.
Mol Syst Biol 4:222
Stokvis E, Rosing H, Beijnen JH (2005) Stable isotopically
labeled internal standards in quantitative bioanalysis using
liquid chromatography/mass spectrometry: necessity or not?
Rapid Commun Mass Spectrom 19:401-407
Colin P, De Bock L, T'Jollyn H, Boussery K, Van Bocxlaer J
(2013) Development and validation of a fast and uniform
approach to quantify beta-lactam antibiotics in human
plasma by solid phase extraction-liquid chromatographyelectrospray-tandem mass spectrometry. Talanta 103:285293
European Medicines Agency (2011) Guideline on
bioanalytical method validation. London, United Kingdom:
23
Mouton JW, Brown DF, Apfalter P, Canton R, Giske CG,
Ivanova M, MacGowan AP, Rodloff A, Soussy CJ,
Steinbakk M, Kahlmeter G (2012) The role of
pharmacokinetics/pharmacodynamics in setting clinical MIC
breakpoints: the EUCAST approach. Clin Microbiol Infect
18:E37-45
European Committee on Antimicroibal Susceptibility Testing
(2015) Antimicrobial susceptibility testing EUCAST disk
diffusion method. v 5.0 edn
Andrews JM (2001) Determination of minimum inhibitory
concentrations. J Antimicrob Chemother 48 Suppl 1:5-16
Sabat AJ, Budimir A, Nashev D, Sa-Leao R, van Dijl J,
Laurent F, Grundmann H, Friedrich AW, Markers ESGoE
(2013) Overview of molecular typing methods for outbreak
83
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
84
detection and epidemiological surveillance. Euro Surveill
18:20380
Rice LB (2008) Federal funding for the study of
antimicrobial resistance in nosocomial pathogens: no
ESKAPE. J Infect Dis 197:1079-1081
Organization WH (2015) Global tuberculosis report 2015. In:
Global tuberculosis report 2015. 20th edn. Geneva: World
Health Organization: 1-204
Ewer K, Millington KA, Deeks JJ, Alvarez L, Bryant G,
Lalvani A (2006) Dynamic antigen-specific T-cell responses
after point-source exposure to Mycobacterium tuberculosis.
Am J Respir Crit Care Med 174:831-839
Organization WH (2010). In: Treatment of Tuberculosis:
Guidelines. 4th edn. Geneva
Borgdorff MW, van Soolingen D (2013) The re-emergence
of tuberculosis: what have we learnt from molecular
epidemiology? Clin Microbiol Infect 19:889-901
Coscolla M, Gagneux S (2014) Consequences of genomic
diversity in Mycobacterium tuberculosis. Semin Immunol
26:431-444
Walker TM, Ip CL, Harrell RH, Evans JT, Kapatai G,
Dedicoat MJ, Eyre DW, Wilson DJ, Hawkey PM, Crook
DW, Parkhill J, Harris D, Walker AS, Bowden R, Monk P,
Smith EG, Peto TE (2013) Whole-genome sequencing to
delineate Mycobacterium tuberculosis outbreaks: a
retrospective observational study. Lancet Infect Dis 13:137146
Folkhälsomyndigheten (2014) ESBL-producerande
tarmbakterier. Stockholm: 104
Carattoli A (2009) Resistance plasmid families in
Enterobacteriaceae. Antimicrob Agents Chemother 53:22272238
Bush K, Jacoby GA (2010) Updated functional classification
of beta-lactamases. Antimicrob Agents Chemother 54:969976
Ambler RP (1980) The structure of beta-lactamases. Philos
Trans R Soc Lond B Biol Sci 289:321-331
Folkhälsomyndigheten (2016) Samhällsekonomiska
konsekvenser av antibiotikaresistens.
van Belkum A, Tassios PT, Dijkshoorn L, Haeggman S,
Cookson B, Fry NK, Fussing V, Green J, Feil E, GernerSmidt P, Brisse S, Struelens M, European Society of Clinical
M, Infectious Diseases Study Group on Epidemiological M
(2007) Guidelines for the validation and application of
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
typing methods for use in bacterial epidemiology. Clin
Microbiol Infect 13 Suppl 3:1-46
van Embden JD, Cave MD, Crawford JT, Dale JW, Eisenach
KD, Gicquel B, Hermans P, Martin C, McAdam R, Shinnick
TM, et al. (1993) Strain identification of Mycobacterium
tuberculosis by DNA fingerprinting: recommendations for a
standardized methodology. J Clin Microbiol 31:406-409
Supply P, Mazars E, Lesjean S, Vincent V, Gicquel B, Locht
C (2000) Variable human minisatellite-like regions in the
Mycobacterium tuberculosis genome. Mol Microbiol
36:762-771
Supply P, Lesjean S, Savine E, Kremer K, van Soolingen D,
Locht C (2001) Automated high-throughput genotyping for
study of global epidemiology of Mycobacterium tuberculosis
based on mycobacterial interspersed repetitive units. J Clin
Microbiol 39:3563-3571
Brolund A, Haeggman S, Edquist PJ, Gezelius L, OlssonLiljequist B, Wisell KT, Giske CG (2010) The DiversiLab
system versus pulsed-field gel electrophoresis:
characterisation of extended spectrum beta-lactamase
producing Escherichia coli and Klebsiella pneumoniae. J
Microbiol Methods 83:224-230
Versalovic J, Koeuth T, Lupski JR (1991) Distribution of
repetitive DNA sequences in eubacteria and application to
fingerprinting of bacterial genomes. Nucleic Acids Res
19:6823-6831
Mellmann A, Friedrich AW, Rosenkotter N, Rothganger J,
Karch H, Reintjes R, Harmsen D (2006) Automated DNA
sequence-based early warning system for the detection of
methicillin-resistant Staphylococcus aureus outbreaks. PLoS
Med 3:e33
Lindstedt BA (2005) Multiple-locus variable number tandem
repeats analysis for genetic fingerprinting of pathogenic
bacteria. Electrophoresis 26:2567-2582
Masny A, Plucienniczak A (2003) Ligation mediated PCR
performed at low denaturation temperatures--PCR melting
profiles. Nucleic Acids Res 31:e114
Krawczyk B, Samet A, Leibner J, Sledzinska A, Kur J
(2006) Evaluation of a PCR melting profile technique for
bacterial strain differentiation. J Clin Microbiol 44:23272332
Krawczyk B, Leibner-Ciszak J, Mielech A, Nowak M, Kur J
(2009) PCR melting profile (PCR MP)--a new tool for
85
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
86
differentiation of Candida albicans strains. BMC Infect Dis
9:177
Krawczyk B, Leibner J, Baranska-Rybak W, Samet A,
Nowicki R, Kur J (2007) ADSRRS-fingerprinting and PCR
MP techniques for studies of intraspecies genetic relatedness
in Staphylococcus aureus. J Microbiol Methods 71:114-122
Stojowska K, Kaluzewski S, Krawczyk B (2009) Usefulness
of PCR melting profile method for genotyping analysis of
Klebsiella oxytoca isolates from patients of a single hospital
unit. Pol J Microbiol 58:247-253
Erali M, Voelkerding KV, Wittwer CT (2008) High
resolution melting applications for clinical laboratory
medicine. Exp Mol Pathol 85:50-58
Tong SY, Giffard PM (2012) Microbiological applications of
high-resolution melting analysis. J Clin Microbiol 50:34183421
Ruskova L, Raclavsky V (2011) The potential of high
resolution melting analysis (hrma) to streamline, facilitate
and enrich routine diagnostics in medical microbiology.
Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub
155:239-252
Vossen RH, Aten E, Roos A, den Dunnen JT (2009) Highresolution melting analysis (HRMA): more than just
sequence variant screening. Hum Mutat 30:860-866
Ren X, Fu Y, Xu C, Feng Z, Li M, Zhang L, Zhang J, Liao
M (2016) High resolution melting (HRM) analysis as a new
tool for rapid identification of Salmonella enterica serovar
Gallinarum biovars Pullorum and Gallinarum. Poult Sci
Folkhälsomyndigheten Next Generation Sequencing och
bioinformatik (NGS). vol. 2016
Nikolayevskyy V, Trovato A, Broda A, Borroni E, Cirillo D,
Drobniewski F (2016) MIRU-VNTR Genotyping of
Mycobacterium tuberculosis Strains Using QIAxcel
Technology: A Multicentre Evaluation Study. PLoS One
11:e0149435
Driscoll JR (2009) Spoligotyping for molecular
epidemiology of the Mycobacterium tuberculosis complex.
Methods Mol Biol 551:117-128
Maslow JN, Mulligan ME, Adams KS, Justis JC, Arbeit RD
(1993) Bacterial adhesins and host factors: role in the
development and outcome of Escherichia coli bacteremia.
Clin Infect Dis 17:89-97
Brown AC, Bryant JM, Einer-Jensen K, Holdstock J,
Houniet DT, Chan JZ, Depledge DP, Nikolayevskyy V,
142.
143.
144.
145.
146.
147.
148.
149.
150.
Broda A, Stone MJ, Christiansen MT, Williams R,
McAndrew MB, Tutill H, Brown J, Melzer M, Rosmarin C,
McHugh TD, Shorten RJ, Drobniewski F, Speight G, Breuer
J (2015) Rapid Whole-Genome Sequencing of
Mycobacterium tuberculosis Isolates Directly from Clinical
Samples. J Clin Microbiol 53:2230-2237
Schwartz DC, Cantor CR (1984) Separation of yeast
chromosome-sized DNAs by pulsed field gradient gel
electrophoresis. Cell 37:67-75
Bertelli C, Greub G (2013) Rapid bacterial genome
sequencing: methods and applications in clinical
microbiology. Clin Microbiol Infect 19:803-813
Koser CU, Ellington MJ, Cartwright EJ, Gillespie SH,
Brown NM, Farrington M, Holden MT, Dougan G, Bentley
SD, Parkhill J, Peacock SJ (2012) Routine use of microbial
whole genome sequencing in diagnostic and public health
microbiology. PLoS Pathog 8:e1002824
Reuter S, Ellington MJ, Cartwright EJ, Koser CU, Torok
ME, Gouliouris T, Harris SR, Brown NM, Holden MT, Quail
M, Parkhill J, Smith GP, Bentley SD, Peacock SJ (2013)
Rapid bacterial whole-genome sequencing to enhance
diagnostic and public health microbiology. JAMA Intern
Med 173:1397-1404
Torok ME, Peacock SJ (2012) Rapid whole-genome
sequencing of bacterial pathogens in the clinical
microbiology laboratory--pipe dream or reality? J
Antimicrob Chemother 67:2307-2308
Goering RV (2010) Pulsed field gel electrophoresis: a review
of application and interpretation in the molecular
epidemiology of infectious disease. Infect Genet Evol
10:866-875
Tenover FC, Arbeit RD, Goering RV, Mickelsen PA, Murray
BE, Persing DH, Swaminathan B (1995) Interpreting
chromosomal DNA restriction patterns produced by pulsedfield gel electrophoresis: criteria for bacterial strain typing. J
Clin Microbiol 33:2233-2239
Didelot X, Bowden R, Wilson DJ, Peto TE, Crook DW
(2012) Transforming clinical microbiology with bacterial
genome sequencing. Nat Rev Genet 13:601-612
Karger A (2016) Current developments to use linear
MALDI-TOF spectra for the identification and typing of
bacteria and the characterization of other cells/organisms
related to infectious diseases. Proteomics Clin Appl
87
151.
152.
153.
154.
155.
156.
157.
158.
159.
88
Sandrin TR, Goldstein JE, Schumaker S (2013) MALDI
TOF MS profiling of bacteria at the strain level: a review.
Mass Spectrom Rev 32:188-217
Stephan R, Cernela N, Ziegler D, Pfluger V, Tonolla M,
Ravasi D, Fredriksson-Ahomaa M, Hachler H (2011) Rapid
species specific identification and subtyping of Yersinia
enterocolitica by MALDI-TOF mass spectrometry. J
Microbiol Methods 87:150-153
Dark MJ (2013) Whole-genome sequencing in bacteriology:
state of the art. Infect Drug Resist 6:115-123
Loman NJ, Constantinidou C, Chan JZ, Halachev M,
Sergeant M, Penn CW, Robinson ER, Pallen MJ (2012)
High-throughput bacterial genome sequencing: an
embarrassment of choice, a world of opportunity. Nat Rev
Microbiol 10:599-606
Loman NJ, Pallen MJ (2015) Twenty years of bacterial
genome sequencing. Nat Rev Microbiol 13:787-794
Casali N, Broda A, Harris SR, Parkhill J, Brown T,
Drobniewski F (2016) Whole Genome Sequence Analysis of
a Large Isoniazid-Resistant Tuberculosis Outbreak in
London: A Retrospective Observational Study. PLoS Med
13:e1002137
Koser CU, Holden MT, Ellington MJ, Cartwright EJ, Brown
NM, Ogilvy-Stuart AL, Hsu LY, Chewapreecha C, Croucher
NJ, Harris SR, Sanders M, Enright MC, Dougan G, Bentley
SD, Parkhill J, Fraser LJ, Betley JR, Schulz-Trieglaff OB,
Smith GP, Peacock SJ (2012) Rapid whole-genome
sequencing for investigation of a neonatal MRSA outbreak.
N Engl J Med 366:2267-2275
Mellmann A, Harmsen D, Cummings CA, Zentz EB,
Leopold SR, Rico A, Prior K, Szczepanowski R, Ji Y, Zhang
W, McLaughlin SF, Henkhaus JK, Leopold B, Bielaszewska
M, Prager R, Brzoska PM, Moore RL, Guenther S, Rothberg
JM, Karch H (2011) Prospective genomic characterization of
the German enterohemorrhagic Escherichia coli O104:H4
outbreak by rapid next generation sequencing technology.
PLoS One 6:e22751
Knetsch CW, Connor TR, Mutreja A, van Dorp SM, Sanders
IM, Browne HP, Harris D, Lipman L, Keessen EC, Corver J,
Kuijper EJ, Lawley TD (2014) Whole genome sequencing
reveals potential spread of Clostridium difficile between
humans and farm animals in the Netherlands, 2002 to 2011.
Euro Surveill 19:20954
160.
161.
162.
163.
164.
165.
166.
167.
Ellington MJ, Ekelund O, Aarestrup FM, Canton R, Doumith
M, Giske C, Grundman H, Hasman H, Holden M, Hopkins
KL, Iredell J, Kahlmeter G, Koser CU, MacGowan A,
Mevius D, Mulvey M, Naas T, Peto T, Rolain JM,
Samuelsen O, Woodford N (2016) The Role of Whole
Genome Sequencing (WGS) in Antimicrobial Susceptibility
Testing of Bacteria: Report from the EUCAST
Subcommittee. Clin Microbiol Infect
Larsen MV, Cosentino S, Rasmussen S, Friis C, Hasman H,
Marvig RL, Jelsbak L, Sicheritz-Ponten T, Ussery DW,
Aarestrup FM, Lund O (2012) Multilocus sequence typing of
total-genome-sequenced bacteria. J Clin Microbiol 50:13551361
Maiden MC, Jansen van Rensburg MJ, Bray JE, Earle SG,
Ford SA, Jolley KA, McCarthy ND (2013) MLST revisited:
the gene-by-gene approach to bacterial genomics. Nat Rev
Microbiol 11:728-736
European Centre for Disease Prevention and Control
(ECDC) (2016) Expert opinion on whole genome sequencing
for public health surveillance. Stockhol, Sweden: 20
van Soolingen D, Hermans PW, de Haas PE, Soll DR, van
Embden JD (1991) Occurrence and stability of insertion
sequences in Mycobacterium tuberculosis complex strains:
evaluation of an insertion sequence-dependent DNA
polymorphism as a tool in the epidemiology of tuberculosis.
J Clin Microbiol 29:2578-2586
Johansen HK, Moskowitz SM, Ciofu O, Pressler T, Hoiby N
(2008) Spread of colistin resistant non-mucoid Pseudomonas
aeruginosa among chronically infected Danish cystic fibrosis
patients. J Cyst Fibros 7:391-397
Seifert H, Dolzani L, Bressan R, van der Reijden T, van
Strijen B, Stefanik D, Heersma H, Dijkshoorn L (2005)
Standardization and interlaboratory reproducibility
assessment of pulsed-field gel electrophoresis-generated
fingerprints of Acinetobacter baumannii. J Clin Microbiol
43:4328-4335
Petersson AC, Olsson-Liljequist B, Miorner H, Haeggman S
(2010) Evaluating the usefulness of spa typing, in
comparison with pulsed-field gel electrophoresis, for
epidemiological typing of methicillin-resistant
Staphylococcus aureus in a low-prevalence region in Sweden
2000-2004. Clin Microbiol Infect 16:456-462
89
168.
169.
170.
Saeedi B, Hallgren A, Jonasson J, Nilsson LE, Hanberger H,
Isaksson B (2002) Modified pulsed-field gel electrophoresis
protocol for typing of enterococci. APMIS 110:869-874
Zaczek A, Brzostek A, Kuron A, Wojtasik A, Sajduda A,
Dziadek J (2014) Development of a new ligation-mediated
PCR method for the differentiation of Mycobacterium
tuberculosis strains. Int J Tuberc Lung Dis 18:302-309
The European Centre for Disease Prevention and Control
(2015) Antimicrobial resistance surveillance in Europe 2014.
Annual Report of the
European Antimicrobial Resistance Surveillance Network (EARSNet). . Stockholm
171. Velkov T, Bergen PJ, Lora-Tamayo J, Landersdorfer CB, Li
J (2013) PK/PD models in antibacterial development. Curr
Opin Microbiol 16:573-579
172. Roberts JA, De Waele JJ, Dimopoulos G, Koulenti D, Martin
C, Montravers P, Rello J, Rhodes A, Starr T, Wallis SC,
Lipman J (2012) DALI: Defining Antibiotic Levels in
Intensive care unit patients: a multi-centre point of
prevalence study to determine whether contemporary
antibiotic dosing for critically ill patients is therapeutic.
BMC Infect Dis 12:152
173. Vincent JL, Moreno R (2010) Clinical review: scoring
systems in the critically ill. Crit Care 14:207
174. Mbongo CL, Monedero P, Guillen-Grima F, Yepes MJ,
Vives M, Echarri G (2009) Performance of SAPS3,
compared with APACHE II and SOFA, to predict hospital
mortality in a general ICU in Southern Europe. Eur J
Anaesthesiol 26:940-945
175. Abdul-Aziz MH, Sulaiman H, Mat-Nor MB, Rai V, Wong
KK, Hasan MS, Abd Rahman AN, Jamal JA, Wallis SC,
Lipman J, Staatz CE, Roberts JA (2016) Beta-Lactam
Infusion in Severe Sepsis (BLISS): a prospective, two-centre,
open-labelled randomised controlled trial of continuous
versus intermittent beta-lactam infusion in critically ill
patients with severe sepsis. Intensive Care Med
176. Tsai D, Stewart P, Goud R, Gourley S, Hewagama S,
Krishnaswamy S, Wallis SC, Lipman J, Roberts JA (2016)
Pharmacokinetics of piperacillin in critically ill Australian
Indigenous patients with severe sepsis. Antimicrob Agents
Chemother
177. Kohner PC, Robberts FJ, Cockerill FR, 3rd, Patel R (2009)
Cephalosporin MIC distribution of extended-spectrum-
90
178.
179.
180.
181.
182.
183.
184.
185.
186.
{beta}-lactamase- and pAmpC-producing Escherichia coli
and Klebsiella species. J Clin Microbiol 47:2419-2425
Woksepp H, Hallgren A, Borgstrom S, Kullberg F,
Wimmerstedt A, Oscarsson A, Nordlund P, Lindholm ML,
Bonnedahl J, Brudin L, Carlsson B, Schon T (2016) High
target attainment for beta-lactam antibiotics in intensive care
unit patients when actual minimum inhibitory concentrations
are applied. Eur J Clin Microbiol Infect Dis
Gavin PJ, Suseno MT, Thomson RB, Jr., Gaydos JM,
Pierson CL, Halstead DC, Aslanzadeh J, Brecher S, Rotstein
C, Brossette SE, Peterson LR (2006) Clinical correlation of
the CLSI susceptibility breakpoint for piperacillintazobactam against extended-spectrum-beta-lactamaseproducing Escherichia coli and Klebsiella species.
Antimicrob Agents Chemother 50:2244-2247
Udy AA, Varghese JM, Altukroni M, Briscoe S, McWhinney
BC, Ungerer JP, Lipman J, Roberts JA (2012)
Subtherapeutic initial beta-lactam concentrations in select
critically ill patients: association between augmented renal
clearance and low trough drug concentrations. Chest 142:3039
Carlier M, Carrette S, Stove V, Verstraete AG, De Waele JJ
(2014) Does consistent piperacillin dosing result in
consistent therapeutic concentrations in critically ill patients?
A longitudinal study over an entire antibiotic course. Int J
Antimicrob Agents 43:470-473
Lipman J, Crewe-Brown HH, Saunders GL, Gous AG (1996)
Subtleties of antibiotic dosages--do doses and intervals make
a difference in the critically ill? S Afr J Surg 34:160-162
Lipman J, Roberts J (2015) Does Appropriate Antibiotic
Therapy Mean Only Adequate Spectrum and Timing? Crit
Care Med 43:1773-1774
Tabah A, De Waele J, Lipman J, Zahar JR, Cotta MO,
Barton G, Timsit JF, Roberts JA (2015) The ADMIN-ICU
survey: a survey on antimicrobial dosing and monitoring in
ICUs. J Antimicrob Chemother 70:2671-2677
Tsai D, Lipman J, Roberts JA (2015)
Pharmacokinetic/pharmacodynamic considerations for the
optimization of antimicrobial delivery in the critically ill.
Curr Opin Crit Care 21:412-420
Lipman J (2000) Towards better ICU antibiotic dosing. Crit
Care Resusc 2:282-289
91
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
92
Swanson DJ, DeAngelis C, Smith IL, Schentag JJ (1986)
Degradation kinetics of imipenem in normal saline and in
human serum. Antimicrob Agents Chemother 29:936-937
Hu ZY, Boucher BA, Laizure SC, Herring VL, Parker RB,
Hickerson WL (2013) Nonvolatile salt-free stabilizer for the
quantification of polar imipenem and cilastatin in human
plasma using hydrophilic interaction
chromatography/quadrupole mass spectrometry with
contamination sensitive off-axis electrospray. J Mass
Spectrom 48:945-950
Robinson HJ (1952) General pharmacology of antibiotics.
Ann N Y Acad Sci 55:970-982
Sander LC, Wise SA (1995) Influence of Stationary Phase
Chemistry on Shape Recognition in Liquid Chromatography.
Analytical Chemistry 67:3284-3292
Chen D, Yu J, Tao Y, Pan Y, Xie S, Huang L, Peng D, Wang
X, Wang Y, Liu Z, Yuan Z (2016) Qualitative screening of
veterinary anti-microbial agents in tissues, milk, and eggs of
food-producing animals using liquid chromatography
coupled with tandem mass spectrometry. J Chromatogr B
Analyt Technol Biomed Life Sci 1017-1018:82-88
Nickolai DJ, Lammel CJ, Byford BA, Morris JH, Kaplan
EB, Hadley WK, Brooks GF (1985) Effects of storage
temperature and pH on the stability of eleven beta-lactam
antibiotics in MIC trays. J Clin Microbiol 21:366-370
Patel PR, Cook SE (1997) Stability of meropenem in
intravenous solutions. Am J Health Syst Pharm 54:412-421
Myllyniemi AL, Sipila H, Nuotio L, Niemi A, HonkanenBuzalski T (2002) An indirect conductimetric screening
method for the detection of antibiotic residues in bovine
kidneys. Analyst 127:1247-1251
Virolainen NE, Pikkemaat MG, Elferink JWA, Karp MT
(2008) Rapid Detection of Tetracyclines and Their 4-Epimer
Derivatives from Poultry Meat with Bioluminescent
Biosensor Bacteria. Journal of Agricultural and Food
Chemistry 56:11065-11070
Stephens AJ, Inman-Bamber J, Giffard PM, Huygens F
(2008) High-resolution melting analysis of the spa repeat
region of Staphylococcus aureus. Clin Chem 54:432-436
Chen JH, Cheng VC, Chan JF, She KK, Yan MK, Yau MC,
Kwan GS, Yam WC, Yuen KY (2013) The use of highresolution melting analysis for rapid spa typing on
methicillin-resistant Staphylococcus aureus clinical isolates.
J Microbiol Methods 92:99-102
198.
199.
200.
201.
202.
203.
204.
205.
206.
Levesque S, Michaud S, Arbeit RD, Frost EH (2011) Highresolution melting system to perform multilocus sequence
typing of Campylobacter jejuni. PLoS One 6:e16167
Lasch P, Fleige C, Stammler M, Layer F, Nubel U, Witte W,
Werner G (2014) Insufficient discriminatory power of
MALDI-TOF mass spectrometry for typing of Enterococcus
faecium and Staphylococcus aureus isolates. J Microbiol
Methods 100:58-69
Dieckmann R, Helmuth R, Erhard M, Malorny B (2008)
Rapid classification and identification of salmonellae at the
species and subspecies levels by whole-cell matrix-assisted
laser desorption ionization-time of flight mass spectrometry.
Appl Environ Microbiol 74:7767-7778
Clark CG, Kruczkiewicz P, Guan C, McCorrister SJ, Chong
P, Wylie J, van Caeseele P, Tabor HA, Snarr P, Gilmour
MW, Taboada EN, Westmacott GR (2013) Evaluation of
MALDI-TOF mass spectroscopy methods for determination
of Escherichia coli pathotypes. J Microbiol Methods 94:180191
Paauw A, Trip H, Niemcewicz M, Sellek R, Heng JM, MarsGroenendijk RH, de Jong AL, Majchrzykiewicz-Koehorst
JA, Olsen JS, Tsivtsivadze E (2014) OmpU as a biomarker
for rapid discrimination between toxigenic and epidemic
Vibrio cholerae O1/O139 and non-epidemic Vibrio cholerae
in a modified MALDI-TOF MS assay. BMC Microbiol
14:158
Meetani MA, Voorhees KJ (2005) MALDI mass
spectrometry analysis of high molecular weight proteins
from whole bacterial cells: pretreatment of samples with
surfactants. J Am Soc Mass Spectrom 16:1422-1426
Josten M, Reif M, Szekat C, Al-Sabti N, Roemer T, Sparbier
K, Kostrzewa M, Rohde H, Sahl HG, Bierbaum G (2013)
Analysis of the matrix-assisted laser desorption ionizationtime of flight mass spectrum of Staphylococcus aureus
identifies mutations that allow differentiation of the main
clonal lineages. J Clin Microbiol 51:1809-1817
Tagg KA, Ginn AN, Partridge SR, Iredell JR (2015)
MALDI-TOF Mass Spectrometry for Multilocus Sequence
Typing of Escherichia coli Reveals Diversity among Isolates
Carrying blaCMY(-)(2)-Like Genes. PLoS One
10:e0143446
David S, Mentasti M, Tewolde R, Aslett M, Harris SR,
Afshar B, Underwood A, Fry NK, Parkhill J, Harrison TG
(2016) Evaluation of an Optimal Epidemiological Typing
93
support. Perspectives on Psychological Science, 4(3), 236-255. doi: 10.1111/j.17456924.2009.01122.x
Scheme for Legionella pneumophila with Whole-Genome
Van Humbeeck,
L., Piers,
R. D.,
Van Camp,
S., DillenGuidelines.
L, Verhaeghe, S.
T., van den
Sequence
Data
Using
Validation
J Clin
Noortgate
N. J. (2013).
Aged parents’ experiences during a critical illness trajectory
Microbiol
54:2135-2148
after the death of an adult child: A review of the literature. Palliative Medicine,
207.and
Ben Zakour NL, Alsheikh-Hussain AS, Ashcroft MM,
27(7):583-95. doi: 10.1177/0269216313483662
Khanh Nhu NT, Roberts LW, Stanton-Cook M, Schembri
MA, Beatson SA (2016) Sequential Acquisition of Virulence
Wei, M., Liao, K. Y. H., Ku, T. Y., & Shaffer, P. A. (2011). Attachment, self‐compassion,
and Fluoroquinolone Resistance Has Shaped the Evolution of
empathy, and subjective well‐being among college students and community adults.
Escherichia
ST131.
7
Journal
of Personality,coli
79(1),
191-221.MBio
doi: 10.1111/j.1467-6494.2010.00677.x
208. Chen TW, Gan RC, Chang YF, Liao WC, Wu TH, Lee CC,
Huang PJ, Lee CY, Chen YY, Chiu CH, Tang P (2015) Is the
whole greater than the sum of its parts? De novo assembly
Weiss, A., & Costa, P. T. (2005). Domain and facet personality predictors of all-cause
strategies
for bacterial
genomes
based
on paired-end
mortality
among medicare
patients aged
65 to 100.
Psychosomatic
medicine, 67(5),
sequencing.
BMC Genomics 16:648
724-733.
doi: 10.1097/01.psy.0000181272.58103.18
209. Kaas RS, Friis C, Ussery DW, Aarestrup FM (2012)
Estimating variation within the genes and inferring the
phylogeny
186 M.
sequenced
Escherichia
coliBout, J., van
Wijngaards-de
Meij, L., of
Stroebe,
S., Schut, H.diverse
A. W., Stroebe,
W., van den
der Heijden,
P. G. M.
(2005).
Couples at risk
following the death of their child:
genomes.
BMC
Genomics
13:577
of griefSR,
versus
depression.
Journal
of Consulting
and Clinical
Psychology,
210.Predictors
Leopold
Goering
RV,
Witten
A, Harmsen
D, Mellmann
73(4),A
617-623.
10.1037/0022-006x.73.4.617
(2014)doi:Bacterial
whole-genome sequencing revisited:
portable, scalable, and standardized analysis for typing and
Wijngaards-De
Meij, L.,
M., Stroebe,
W., Schut, H.,
Van den Bout,
J., Van
Der
detection
ofStroebe,
virulence
and antibiotic
resistance
genes.
J Clin
Heijden,
P.G., et al. (2008).
The impact of circumstances surrounding the death of a
Microbiol
52:2365-2370
child on parents’ grief. Death Studies, 32(3), 237-252. doi:
211.10.1080/07481180701881263
Willems S, Kampmeier S, Bletz S, Kossow A, Kock R, Kipp
F, Mellmann A (2016) Whole-Genome Sequencing
Elucidates Epidemiology of Nosocomial Clusters of
World Health Organization (WHO). (2006). Constitution of the World Health
Acinetobacter
baumannii.
J Clin Microbiol 54:2391-2394
Organization.
Available at
http://www.who.int/governance/eb/who_constitution_en.pdf
212. Kluytmans-van den Bergh MF, Rossen JW, BruijningVerhagen PC, Bonten MJ, Friedrich AW, VandenbrouckeZimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The multidimensional
Grauls
CM,
Willems
RJ, Kluytmans
JA,
So Msg52(1),
(2016)
scale of
perceived
social
support. Journal
of Personality
Assessment,
30-41. doi:
Whole
genome
multilocus
sequence
typing
of
extended10.1207/s15327752jpa5201_2
spectrum beta-lactamase-producing Enterobacteriaceae. J
Clin Microbiol
213. Mellmann A, Bletz S, Boking T, Kipp F, Becker K, Schultes
A, Prior K, Harmsen D (2016) Real-Time Genome
Sequencing of Resistant Bacteria Provides Precision
Infection Control in an Institutional Setting. J Clin Microbiol
214. Pankhurst LJ, Del Ojo Elias C, Votintseva AA, Walker TM,
Cole K, Davies J, Fermont JM, Gascoyne-Binzi DM, Kohl
TA, Kong C, Lemaitre N, Niemann S, Paul J, Rogers TR,
Roycroft E, Smith EG, Supply P, Tang P, Wilcox MH,
Wordsworth S, Wyllie D, Xu L, Crook DW, Group C-TS
(2016) Rapid, comprehensive, and affordable mycobacterial
diagnosis with whole-genome sequencing: a prospective
study. Lancet Respir Med 4:49-58
62
94