BSAC Standardised Disc Susceptibility

1
BSAC Standardised Disc Susceptibility Testing
Method User Group Meeting Newcastle-upon-Tyne Su
rveillance of Resistance Alasdair MacGowan
2

• Topics
• definitions
• why perform surveillance
• what are the attributes of surveillance
• what's currently wrong with resistance
  surveillance
• BSACs role in surveillance
• summary

3
Definition of surveillance ongoing and
systematic collection, analysis
and interpretation of health data essential to
planning, implementation and evaluation of
public health practice closely integrated with
timely dissemination of these data to the
control and prevention of human disease and
injury CDC, 1996, Thacker
4

• Why perform surveillance (1)
• determine the size of the problem
• determine time trends in resistance
• detect previously unknown resistance
• determine geographic spread of resistance
• establish if resistance associated with an
  outbreak
• assess the effectiveness of interventions
• assess the impact of changes in practice (OTC
  drugs)

5

• Why perform surveillance (2)
• regulatory requirement for market authorisation
• often part of post approval surveillance (with
  safety)
• commercial advantage
• inform drug development
• inform production policies/guidelines for
  individuals

6
Assumptions - numerous but most important link
between resistance as defined by
laboratory test and adverse clinical, economic or
Public Health outcome
7
Antibiotic resistance is important (1) Ibrahim
et al, 2000 ICU, St Louis 1997-99
8
Antibiotic resistance is important (2) Harbarth
et al, 2002 ICU, Geneva In multivariate analysis
9
Antibiotic resistance is important
(3) Antibiotic resistance in UTI Talan et al,
2000 Los Angeles, USA as part of a randomised
double blind comparative study of ciprofloxacin
and trimethoprim/sulphametloxazole (n
378) TMP/SMX resistance rate 19 All treated in
Community TMP/SMX associated with higher
bacteriological/clinical failures
10
Antibiotic resistance is important
(4) Penicillin resistance in S.
pneumoniae susceptible ? 0.06mg/L intermedia
te 0.12 - 1.0mg/L resistant ? 2mg/L
11
Penicillin non susceptibility does not impact on
outcome with penicillins paediatric CAP
(Friedland Klugman, 1992) adults with
pneumonia (Sandches et al, 1992) paediatric
pneumococcal bacteraemia (Friedland, 1995) adult
with pneumonia (Pallarves et al, 1995) invasive
pneumococcal disease (Choi Lee,
1998) paediatric invasive pneumococcal disease
(Deeks et al, 1999) hospitalised patients with
pneumococcal bacteraemia (Farinas et al,
2000) CAP - pneumococcal (Bedos et al,
2001) hospitalised patients with pneumococcal
bacteraemia (Moroney, 2001)
12
Penicillin resistance does have an impact on
outcome with penicillins (1) Feikin et al,
2000 excluding early deaths
OR penicillin MIC ? 4mg/L 7.1 (1.7 - 13.0 MIC
? 1mg/L 1.0 (0.3 - 3.0) cefotaxime MIC ?
2mg/L 5.9 (1.1 - 33.0) MIC ? 2mg/L 1.2 (0.3 -
7.4)
13

• Penicillin resistance does have an impact on
  outcome
• with penicillin (2)
• Turrett et al, 1999
• Pneumococcal pneumonia (retrospective study, n
  462)
• multivariate analysis identified the following as
  independent
• predictors of mortality
• old age
• severe disease
• multilobar infiltrate
• effusion
• hispanic
• high level penicillin resistance

14
Conclusion for S. pneumoniae penicillin
resistance probably only clinically significant
in LRTI once MICs ? 2-4mg/L i.e. not the
existing breakpoints except CNS infection
15
Attributes of surveillance organisation study
design data collection susceptibility
testing interpretation dissemination application
Felmingham et al, 2002
16

• Organisation
• sources of funding
• National/International Public Health
  bodies
• pharmaceutical companies
• charities
• examples
• National HPA, Scotland, Wales, CDC in USA for
• community and hospital micro organisms
• International EU/European Antimicrobial
  Resistance
• Surveillance System (EARSS)
• WHO/various
• Industry
• Alexander Project
• PROTEKT study
• SENTRY Antimicrobial Surveillance Programme
• Libra Project
• Charities/learned societies
• Paul Erlich Society

17
Study design (1) many variations - factors are -

• local to international/geographical spread
• single point or longitudinal overtime
• large number of isolates or small
• central lab testing or local
• single or multiple methodology
• manual e transmission of results
• internal and external quality assurance

18
Study design (2)

• phenotypic detection of resistance? Combine with
  typing and genotype detection
• patient or bacteria
• types of patient
• source of isolates - diagnostic specimens or
  screening
• specimen/pathogen transportation, especially with
  central laboratory testing
• linkages to other databases - i.e. antibiotic
  prescribing
• how often to collect strains and test?

19
Geography Influence of population density on
antibiotic resistance compared three cities St
Johns (Canada), Athens (Greece), Groningen
(Netherlands) antibiotic consumption defined
daily dose (DDD) DDD/1000 inhabitants/d
(DID) DDD/Km2/d (DDD/km2) studied
antibiotic resistant E. coli in healthy
volunteers ampicillin resistance in E. coli test
related to DDD/1000 inh/d or DDD/Km2/d Brui
nsma et al, 2003
20

• Data Collection
• population - based sampling of infection vs
  laboratory
• specimens
• difficulties of reaching a microbiological
  diagnosis CAP vs UTI
• community or hospital infection
• patient age, sex
• previous antibiotic use and co-morbidities
• single specimen source or multiple S.
  pneumoniae sputum
• or blood
• remove duplicates

21
Prior antibiotic exposure Prior trimethoprim use
and trimethoprim resistant UTI design nested
case control study using linked database -
Dundee compare trimethoprim resistant and
susceptible patients (n 13,765) logistic
regression showed prior exposure to trimethoprim
OR 4.35 or other antibiotic (OR 1.32) predicted
resistance Steinke et al, 2001
22
Impact of classification of duplicate
isolates VRE isolates at Guys St Thomas
Hospitals resistant 95 CI all
isolates 13.4 12.8 -
14.0 duplicates in 5d excluded 11.6
10.9 - 12.2 duplicates in 30d excluded
10.1 9.4 - 10.8 duplicates in 365d excluded
9.6 8.9 - 10.3 Shannon and French, 2002
23
Susceptibility testing (1)

• disc diffusionsimple, economical, reproducible,
  qualitative or categorical output (S/I/R) done as
  part of care
• Etestquantitative, i.e. MIC but expensive
• national defined methodNCCLS, BSAC, French
  (SFM), German (DIN), Swedish (SRGA)EUCAST -
  microtitre MIC method only
• internal/external standardisation

24
Susceptibility testing (2)
MIC testing agar dilution or microtitre
trayeconomical, reproducible, quality assured,
standardised, quantitative but can be converted
to categoricalregarded as gold standard Good
agreement between national methods when studies
performedi.e. BSAC vs NCCLS BSAC, 2001-02
BSAC vs SRGA vs SFM vs NCCLS BMS, 2001
BSAC vs SRGA vs SFMS vs NCCLS GSK
25
BSAC and NCCLS MIC method comparison
661 S.pneumoniae 938 H.influenzae 421
M.catarrhalis 33 antibiotic ? bacteria
combinations gt90 isolates 1 log gt95 2
logs 2080-89 isolates 1 log gt95 2
logs 5lt80 isolates 1 log lt95 2 logs
8 Reynolds et al, 2001
26
Interpretation MIC breakpoints used to define
S/I/R from MIC distributions main factors in
determining breakpoints - pharmacokinetics pharm
acodynamics MIC distribution clinical
data_____ existing use patterns
(cefuroxime) availability of other drugs
(linezolid) high or low resistance areas (S or N
Europe) cautious or adventurous (Sweden v
USA) commercial interest (high content discs
high breakpoints)
27
Interpretation - example 198 clinical isolates of
E.coli, 67 S.aureus, bloodstream infection in
Norway
Leegard et al, 2000
28

• Dissemination
• publication abstract, poster, paper
• full data esp MIC distributors difficult
• disc derived databases often low scientific
  credibility and very large
• web based presentationAlexander www.alexander
  -network.comLIBRA www.librainitiative.comPROT
  EKT www.protekt.orgESGARS www.ears.rivm.nl
  BSAC www.bsacsurv.org
• feedback to individual hospitalsMRSA bacteraemia
  ? Scotland
• feedback to individual Trust and league tables -
  England

29

• Application
• depends on aims but -

? local and national antibiotic
guidelines ? infection control interventionsi.e.
MRSA in England, S.pneumoniae in
Iceland ? prescribing interventionsi.e.
macrolide resistant Group A Streprococci in
Finland ? public awareness campaigns, i.e. Andy
Biotic in UK
30
Problems with surveillance (1)(DoH sponsored
meeting RSM, 2001)

• too much poor quality data
• poor access to data
• collecting data a substitute for more difficult
  actions
• clear national minimum data set need
• confusion between surveillance and
  epidemiological studies

31
Problems with surveillance (2) Epidemiological
studies Surveillancecomplex simplecomplete
incompletesporadic regulartargeted broad
/inclusivetime limited continuousreal
research not real research
32
Problems with use of routine data(routine
passive surveillance)

• confusion about strain identification
  coliforms, Ent faecalis/faecium
• first line/second line testing makes pooling
  difficult
• poor IT links make collection difficult

Problems with use of sentile surveys

• usually commercial
• concerns about bias in design
• international so individual collections even from
  large countries small, i.e. SENTRY program in
  Europe
• depend on lifecycle of antibiotic, i.e.
  Alexander-network and GSK
• concerns not all data available

All biased by specimen collection
33
Bias in specimen collection 50 patients with UTI
have MSU sent 3 patients who receive antibiotics
for LRTI have specimens sent more specimens sent
in elderly ?lactamase positive H.influenzae22
in lab database cf11 in unselected
patients MacGowan et al, 1998
34
Antibiotic exposure in children compared to those
who submit urine samples from primary care 63
children receive antibiotics in first year of
life 75 in two years Boys earlier exposure than
girls to antibiotics Children who submit urine
samples more socially deprived and more prior to
antibiotic exposure and hospital
admission Steinke et al, 2002
35
BSAC objectives in antimicrobial susceptibility
testing and surveillance

• provide a standard, high quality, reproducible
  testing methodology
• develop surveillance programmes across British
  Isles(respiratory tract, bacteraemia)
• collaborate with HPA, etc., to reduce
  duplication(P.aeruginosa Acinetobacter MRSA)
• provide published outputs
• provide web based outputs

36
BSAC standard disc testing method routine
databases cross-comparison focused/sentile
databases BSAC MIC method BSAC surveillance
programme
37
Summary

• antibiotic resistance surveillance is central to
  managing the issues surrounding resistance
• there is no one way to do resistance multiple
  approaches are essential
• routine data collection and targeted studies
  provide one way forward
• international harmonisation of breakpoints is
  desirable
• present surveillance in the UK has multiple
  problems but internationally is good
• standard laboratory methodologies are vital