This article appears in the following Journal of Antimicrobial Chemotherapy issue: The British Society for Antimicrobial Chemotherapy Resistance Surveillance Project 1999/2000-2006/7 [View the issue table of contents]
Articles |
Survey, laboratory and statistical methods for the BSAC Resistance Surveillance Programmes
1 Department of Medical Microbiology, Southmead Hospital, Southmead Road, Bristol BS10 5NB, UK 2 Health Protection Agency Centre for Infections, 61 Colindale Avenue, London NW9 5EQ, UK 3 Quotient Bioresearch Ltd, Microbiology, 7-9 William Road, London NW1 3ER, UK
* Corresponding author. Tel: +44-117-959-4080; Fax: +44-117-959-3154; E-mail: rreynolds{at}bsac.org.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Objectives: The British Society for Antimicrobial Chemotherapy (BSAC) Bacteraemia and Respiratory Resistance Surveillance Programmes are designed for long-term surveillance of antimicrobial resistance in key pathogens of bloodstream and community-acquired respiratory infection in the UK and Ireland. This paper describes their methods in detail.
Methods: Sentinel laboratories across the UK and Ireland contributed up to a fixed quota of isolates of defined bacterial groups. Collecting laboratories were compared with national benchmarks for size of Hospital Trust and distribution of bacteraemia pathogens. A central laboratory for each programme confirmed the identification of isolates, measured MICs by the BSAC agar dilution method and undertook further testing by standard methods. The variability of the MIC method was assessed by repeated annual testing of a panel of control isolates. Classification as susceptible, intermediate or resistant was by BSAC and European Committee on Antimicrobial Susceptibility Testing breakpoints. Statistical analysis was adjusted for inter-centre variation using random effects logistic regression.
Results: Thirty-two laboratories contributed 16 550 respiratory isolates from 1999–2000 to 2006–07; 30 laboratories contributed 15 812 bacteraemia isolates from 2001 to 2006. Although large and teaching hospitals were over-represented, the pattern of bacteraemia organisms seen in the collecting laboratories in England and Wales was similar to that in national data reported to the Health Protection Agency. Replicate MIC measurements showed that 
90% agreed within ±1, and 98% within ±2, doubling dilutions.
Conclusions: These surveillance programmes have provided reliable information on antimicrobial susceptibility in the UK and Ireland over six and eight seasons, respectively, so far. Detailed results showing non-susceptibility trends, and relationships with potential predictive factors, are presented in six linked papers in this Supplement.
Keywords: MIC , breakpoint , agar dilution , statistical analysis , antimicrobial resistance
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
The British Society for Antimicrobial Chemotherapy (BSAC) Resistance Surveillance Project was initiated in response to widespread concerns about rising antimicrobial resistance in the UK in the late 1990s, and the lack of consistent long-term surveillance of resistance, as described earlier.1 The Project currently comprises two surveillance programmes. The first began in 1999 and focuses on the three main bacterial species causing community-acquired lower respiratory tract infections. The second, covering the much wider range of pathogens involved in bacteraemia, began in 2001.
This paper describes the implementation of the BSAC Respiratory and Bacteraemia Resistance Surveillance Programmes. It covers the collection of isolates and the subsequent laboratory testing to confirm their identification, assess their susceptibility to antimicrobial agents and gain additional information about mechanisms of resistance. The methods used for statistical analysis of the data are described.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Survey methods
Both the BSAC Respiratory and Bacteraemia Resistance Surveillance Programmes are based on a set of 20–25 sentinel clinical laboratories. The collecting laboratories, originally selected to give wide geographical coverage across the UK and Ireland, include both urban and rural settings and areas with a range of deprivation scores, but they are not a random sample and may not be strictly representative of all clinical laboratories of the British Isles. The bacteraemia collection runs in calendar years, each laboratory collecting up to 10 consecutive isolates, taken from blood and considered to be clinically significant, of each of 12 organism groups (Table 1); these groups comprise those species ranked as the 10 commonest agents of bacteraemia in each of the 10 years for which information was available before the start of the study [Health Protection Agency (HPA) LabBase data].2 For the respiratory study, laboratories collect up to 50 consecutive isolates each of Streptococcus pneumoniae and Haemophilus influenzae and up to 25 of Moraxella catarrhalis each winter between 1 October and 30 April. The isolates are from lower respiratory sources of patients with presumed lower respiratory infection, excluding those with cystic fibrosis and those who have been in hospital for >48 h. Duplicate isolates, defined as isolates of the same species collected within 14 days (respiratory) or 7 days (bacteraemia) of a previous isolate of the same species from the same patient and site (blood or respiratory), are excluded. Additional anonymized information about patient characteristics is collected, as shown in Table 1.
|
Most laboratories in England and Wales reported results for clinically significant bacteraemias to the HPA LabBase/CoSurv system2 throughout the BSAC surveillance period. The distribution of organisms reported to LabBase by laboratories contributing to the BSAC Surveillance was compared with the overall national distribution. Laboratories in Scotland, Ireland and Northern Ireland were asked to supply figures from their internal records for the total number of isolates in each of the BSAC bacteraemia organism collection groups in order to attempt a similar comparison.
Each BSAC collecting laboratory in England was linked with an NHS Hospital Trust. The sizes of those Trusts were noted from publications of the mandatory reporting system for methicillin-resistant Staphylococcus aureus (MRSA) bacteraemia and compared with the national distribution of NHS Trust sizes.3
All microbiological testing was conducted centrally. Respiratory isolates were tested by Quotient Bioresearch Ltd, Microbiology (formerly GR Micro Ltd), London. Bacteraemia isolates were tested at the Antibiotic Resistance Monitoring and Reference Laboratory and the Respiratory and Systemic Infections Laboratory of the HPA Centre for Infections in London.
Storage and transport of isolates
Collecting laboratories stored isolates using local methods or, in the respiratory study, MicrobankTM bead storage vials (Pro-Lab Diagnostics, Neston, UK), until sufficient numbers were available for transport to the central laboratory. They were then subcultured on non-selective medium to provide luxuriant growth after overnight incubation, checked for purity and prepared for transport from fresh growth. Respiratory isolates were sent as a heavy inoculum on a transport swab suspended in Amies transport medium (Technical Service Consultants Ltd, Heywood, UK). Bacteraemia isolates were sent on nutrient agar slopes or, in the case of streptococci, on Dorset Egg slopes. Transport was in accordance with prevailing advice and regulations.4,5 Following checking of purity and regrowth from single colonies at the central laboratories, isolates were stored indefinitely at –70°C as a dense suspension in a high-protein matrix of undiluted horse serum (respiratory study), or in glycerol (bacteraemia) or blood glycerol broth (fastidious bacteraemia organisms). Original bacteraemia samples received from collecting centres were also retained as slopes or agar stabs for 2 years.
Identification of isolates by species and serotype
Isolates were identified by standard methods at the central laboratories. Isolates not belonging to a protocol-defined group were discarded without further testing.
Identification of respiratory isolates.
Isolates were confirmed as S. pneumoniae if they were Gram-positive diplococci growing as 
-haemolytic colonies on horse blood agar, and if they were inhibited by a 5 µg optochin (ethylhydrocupreine hydrochloride) disc (Oxoid, Basingstoke, UK), lysed within 30 min in the presence of 2% sodium deoxycholate and were catalase-negative when tested with a 3% hydrogen peroxide solution. Isolates were confirmed as H. influenzae if they were Gram-negative coccobacilli requiring both haematin (factor X) and nicotinamide adenine dinucleotide (factor V) when grown on brain heart infusion agar using pre-impregnated X, V and XV factor discs (medium and discs, Oxoid). M. catarrhalis isolates were identified as Gram-negative cocci that were positive for both oxidase (using a 1% solution of tetramethyl-p-phenylenediamine dihydrochloride) and butyrate esterase (using Tributyrin diagnostic tablets, Bioconnections, Leeds, UK).
Identification of bacteraemia isolates.
Staphylococci were identified by tube coagulase tests and, until 2005, coagulase-negative organisms were identified to species level by PCR of DNA encoding 16S ribosomal RNA as described previously.6 Pneumococci were identified by optochin susceptibility and serotyping (see below), while other - and non-haemolytic streptococci were identified with the API rapid ID32 Strep system (bioMérieux UK Ltd, Basingstoke, UK) and with additional biochemical tests.7–10 β-Haemolytic streptococci were identified using Lancefield group antisera (Prolex Streptococcal Grouping Latex Kits, Pro-Lab Diagnostics), and enterococci were identified using a previously described multiplex PCR for D-Ala:D-Ala ligase and related genes.11 Primary identification of Escherichia coli was based on pink growth in chromogenic urinary tract infection medium (CM0949, Oxoid); other Enterobacteriaceae and aberrant E. coli that failed to give pink growth on this medium were identified with API20E strips. Pseudomonas aeruginosa isolates were identified as oxidase positive and giving blue/green growth on King’s A medium.12 Pseudomonas isolates not positively identified in this first screen were identified with API20NE, as were isolates submitted as other non-fermenters. Exceptionally, the identity of unusual isolates was obtained by referral to the Molecular Identification Services at the HPA National Collection of Type Cultures.
Serotypes of S. pneumoniae from bacteraemia from 2001 to 2004 were determined by conventional capsular pneumococcal serotyping via slide agglutination test, using Latex Antisera (Pneumotest-Latex kit) and Pneumococcal Antisera (both from Statens Serum Institute, Copenhagen, Denmark). From 2005 onwards, a multiplex ELISA using Luminex xMap technology (Bioplex system, Bio-Rad, Hemel Hempstead, UK) was used as an initial identification screen as described elsewhere.13 Isolates not identified in this initial screen but that were C-polysaccharide-positive were presumed to be pneumococci and were subjected to the slide agglutination tests as in previous years. Respiratory S. pneumoniae (2005–06 winter season only) were identified in a two-stage process, initially to serogroup level using the Pneumotest-Latex kit and then to serotype level by Quellung reaction using pneumococcal capsular antisera (kit and antisera both from Statens Serum Institute). Isolates that gave no capsular reaction to any of the available antisera were sent to the Statens Serum Institute for further examination.
In general, MICs were determined by the BSAC agar dilution method,14,15 as summarized in Table 2, which has remained essentially unchanged throughout the period of the BSAC surveillance studies to date. Internal control strains were included in all test runs, as specified in the BSAC method. Agar plates of 10 x 10 cm (respiratory) or 8 x 12 cm (bacteraemia) containing 50 mL of agar allowed the simultaneous testing of 100 or 96 isolates, respectively, including controls. Plates were read visually in the respiratory study, and using a Sorcerer Image Analysis System (Perceptive Instruments Ltd, Haverhill, UK) in the bacteraemia study, with visual confirmation where required. The density of bacterial suspensions was checked by dilution and spiral plating to ensure the correct inoculum size of 104 (or, for M. catarrhalis with β-lactams, 106) cfu/spot. The few anaerobes and category 3 pathogens were tested using Etests (AB Biodisk Solna, Sweden) by the BSAC method as described previously.16 Testing ranges (Tables 3 and 4) have been refined over the course of the studies to provide full endpoints for the large majority of isolates and to minimize the reporting of censored MICs (those recorded only as 
x or >x mg/L). Some of the agents tested are of limited therapeutic use but aid the inference of resistance mechanisms; others serve as indicators, for example, respiratory H. influenzae and M. catarrhalis were examined for reduced susceptibility to fluoroquinolones by measuring the zone of inhibition around a 30 µg nalidixic acid disc (Oxoid). Clavulanic acid was supplied by GlaxoSmithKline (Brentford, UK), continuity agents (intended to be tested for the full term of the programme) by Sigma-Aldrich (Poole, UK) and additional agents (tested in some seasons only) by their respective manufacturers.
|
|
|
Variability of MIC determination. A set of clinical isolates of relevant species and with varied MICs was selected from the collections of an external laboratory for each programme. The respiratory panel comprised 12 isolates (5 S. pneumoniae, 4 H. influenzae and 3 M. catarrhalis); the bacteraemia panel comprised 50 isolates [9 E. coli, 9 Klebsiella spp., 3 Proteeae, 9 P. aeruginosa, 10 S. aureus, 5 coagulase-negative staphylococci (CoNS) and 5 Enterococcus spp.]. These isolates were supplied each season, under different codes, to be tested alongside that season’s surveillance collection. The repeated results for each strain and antimicrobial agent were compared to assess the variability of MIC measurements, by calculating and summarizing all possible pairwise differences between results. Occasional obvious errors were excluded when results were so divergent across a range of agents as to indicate that they referred to a different isolate from that intended.
Testing for specific mechanisms of resistance and interpretative reading. Interpretative reading of susceptibility patterns in bacteraemia isolates was based upon previously published criteria, but the final interpretation was informed by the experience and judgement of the scientists performing the analysis (D. M. L. and R. H.).17
β-Lactamases.
β-Lactamase was sought in respiratory H. influenzae and M. catarrhalis using the chromogenic cephalosporin, nitrocefin (Oxoid). Extended-spectrum β-lactamase (ESBL) detection in Enterobacteriaceae has been gradually refined throughout the study. In 2001, Enterobacteriaceae, except Serratia, resistant to ceftazidime were tested for ESBL production using ceftazidime and cefotaxime, alone and with 4 mg/L clavulanate; cefotaxime was not included in the screening panel at that time. In 2002, Enterobacteriaceae, except Serratia, resistant to cefotaxime or ceftazidime were subjected to MIC determination with ceftazidime, cefotaxime and cefepime, all ±4 mg/L clavulanate; in this year only, the initial cefotaxime screen was at a single breakpoint concentration of 1 mg/L. In 2003, all isolates resistant to cefotaxime or ceftazidime, including Serratia, were re-tested for clavulanate synergy as in 2002. From 2004, clavulanate synergy testing was extended to include all isolates on the breakpoints for cefotaxime (2 mg/L) or ceftazidime (1 mg/L) as well as resistant isolates, and from 2005, cefpirome replaced cefepime in synergy testing owing to supply constraints. In all cases, a 8-fold reduction of MIC in the presence of clavulanate at 4 mg/L was taken as a positive ESBL synergy result. Klebsiella oxytoca were considered to have false-positive results and excluded from the ESBL count if they were highly resistant (commonly, MIC 64 mg/L) to piperacillin/tazobactam and cefuroxime but not to cefotaxime and ceftazidime, as this pattern indicates hyperproduction of chromosomal K1 β-lactamase rather than ESBL production.17,18 ESBL-positive Enterobacteriaceae with cefotaxime MICs 4-fold higher than for ceftazidime, or cefotaxime MICs >256 mg/L, were subject to a multiplex PCR for CTX-M alleles.19 Isolates resistant to cefotaxime, ceftazidime and cefoxitin, but susceptible to cefepime or cefpirome, and without cephalosporin/clavulanate synergy, were interpreted as AmpC hyperproducers. Isolates with high-level resistance to oxyimino-cephalosporins, resistant also to carbapenems, but lacking clavulanate synergy, were investigated in depth by whatever phenotypic and molecular means were available and deemed appropriate.
Molecular testing of staphylococci and recognition of methicillin resistance
PCR testing for the mecA gene was introduced for S. aureus from 2005 and for all staphylococci from 2006.20,21 Staphylococci were considered to be ‘methicillin-resistant’ if they were phenotypically resistant to oxacillin or if they had the mecA gene. From 2006, the mupA gene, encoding high-level mupirocin resistance, was sought in all staphylococci by PCR.22
Investigation of unusual results. Isolates showing unusual resistance patterns were subject to repeated and further investigations, including, in some cases, more rigorous investigation of species identification. Occasional isolates found by this process to have been allocated to the wrong organism collection group were discarded. Where MIC results varied from those found originally, the later results were used.
Isolates were classified as susceptible, intermediate or resistant using BSAC breakpoints (version 6.1, February 2007),15 or, where BSAC breakpoints were unavailable, European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints.23 EUCAST breakpoints have been adopted by the BSAC progressively since 2006, but the process is not yet complete. Intermediate and resistant isolates were grouped as non-susceptible for some purposes.
Statistical analysis was conducted using Stata versions 9.2 and 10.0 (StataCorp LP, College Station, TX, USA). Analyses for trend and association of either non-susceptibility or resistance with other potential predictors used logistic regression with a random effect for centre to account for inter-centre variation.24 Where a potential predictor had many categories, for example, hospital specialty, the analysis made use of the commonest categories for the organism concerned and grouped the remainder as ‘other’. Pneumococcal serotypes were divided into the 10 commonest serotypes and ‘other’, and analysed by multinomial logistic regression, with robust standard errors to adjust for inter-centre variation. Predictors found significant in univariate analysis were included in regression models in pairwise combinations to test for pairwise independence, but the sheer number of organism–agent combinations precluded the building of full multiple regression models for every case. P values are reported without adjustment in the papers describing the results,25–30 but should be interpreted cautiously as there were multiple tests for each organism, typically around 100, owing to the number of antimicrobials tested and the number of potential predictors. As the outcomes are correlated, because resistance to certain antimicrobials is likely to be associated with resistance to others, a simple Bonferroni correction (e.g. multiplying all P values by 100) would be over-conservative. As a rule of thumb, we therefore considered P values of <0.0001 as very highly significant, P < 0.001 as probably significant and P > 0.01 as non-significant. The power of the study was investigated for representative cases by simulation.24 Comparisons of potency between antimicrobials of the same class were made, always using isolates tested with both agents, by summarizing the paired (isolate-by-isolate) differences in log2(MIC).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
The substantive results of the surveillance project to date are described in six related publications25–30 and, in very large part, are freely available on the project website (www.bsacsurv.org) or via the BSAC’s homepage (www.bsac.org.uk).
A total of 32 collecting laboratories contributed a total of 16 550 isolates to the BSAC Respiratory Resistance Surveillance Programme over the eight winter seasons from 1999–2000 to 2006–07 (Table 5). From a total of 170 centre-seasons of participation, 14 withdrew, giving an annual respiratory laboratory turnover rate of 8%. A total of 30 laboratories contributed a total of 15 812 isolates to the BSAC Bacteraemia Resistance Surveillance Programme over the six calendar years 2001–06 (Table 6). The annual bacteraemia laboratory turnover was 4%, with six withdrawals in 150 centre-years of participation.
|
|
Isolates and tests
The Respiratory Programme collected 5810 isolates of S. pneumoniae, 7371 H. influenzae and 3369 M. catarrhalis (Table 7). MICs were not measured for the 840 M. catarrhalis collected in 2001–02 and 2003–04, for financial reasons (Table 8). The number of bacteraemia isolates collected varied among organism groups, from 1015 for - and non-haemolytic streptococci to 1480 for E. coli (Table 9), and the number tested against each antimicrobial is shown in Tables 10 and 11. Overall, 3.5% of the respiratory isolates supplied (2.5% of those supplied as H. influenzae and M. catarrhalis and 5.5% of those supplied as S. pneumoniae) were excluded from the study because the collecting laboratory identification was not confirmed by the central laboratory. Bacteraemia isolates with identification results differing between the collecting and central laboratory could often be included in the study, either in the intended organism group (for example, Streptococcus anginosus locally identified as Streptococcus oralis) or a closely related one (for example, Klebsiella spp. locally identified as Enterobacter spp.).
|
|
|
|
|
Representativeness of the collecting laboratories
In England, where the Department of Health classifies Health Trusts by size and type for the purpose of reporting mandatory MRSA surveillance, the majority (over 85%) of laboratories contributing isolates to the BSAC Resistance Surveillance Programmes are in Health Trusts classified as ‘large acute’ or ‘acute teaching’, compared with 40% of all Health Trusts in England in these categories. There is not a simple one-to-one correspondence between Trusts, hospitals and laboratories, but this result does suggest a probable imbalance in the types of hospitals collecting isolates for the BSAC Surveillance Project.
Nonetheless, compared with all laboratories contributing information to the HPA LabBase system in England and Wales, the laboratories taking part in the BSAC Bacteraemia Resistance Surveillance Programme reported a broadly similar distribution of organisms in bacteraemia. They had a slightly greater proportion of CoNS and enterococci among their bacteraemias, and a slightly lower proportion of E. coli and S. pneumoniae (Tables 12 and 13), consistent with the high proportion of large and teaching Trusts. The clear decrease in the proportion of bacteraemias caused by S. aureus seen in the national LabBase data between 2001 and 2006 was reflected in the BSAC contributing laboratories’ LabBase reports. The BSAC collecting laboratories also mirrored the general increase of 60% in the total number of reports to LabBase from 2001 to 2006. Among laboratories from Ireland, Northern Ireland and Scotland, where isolates were counted in-house rather than via LabBase, there was a strikingly different distribution with a much higher proportion of CoNS (45%), and the total number of isolates identified was stable. This probably represents a difference in ascertainment, as LabBase reports are restricted to clinically significant isolates and duplicates are excluded.
|
|
Overall, the collected bacteraemia isolates represented 6% of the total seen in the collecting centres for the commonest groups, such as E. coli and S. aureus, and up to 30% to 40% for the least common, such as Enterobacter and Proteeae. The BSAC collecting laboratories generated 21% of all the reports to LabBase (in England and Wales), and the collected isolates are thus estimated to represent between
1% and 7% of all similar isolates across these countries.
Repeated MIC measurements were available for all eight respiratory seasons and five bacteraemia seasons (2002–06). Results for staphylococci for 2006 were unavailable owing to delay in testing, and all CoNS for 2003 were excluded owing to a probable batch-labelling error evidenced by widespread and widely discrepant results. Three individual errors were excluded—one contaminated Proteus in Bacteraemia 2005, one S. pneumoniae in Respiratory 2004–05 and one S. aureus in Bacteraemia 2005. Isolate–agent combinations that were tested only once (usually with developmental drugs) had to be excluded as there was no repeated measurement for comparison; this affected 27/1085 remaining respiratory tests and 127/3218 bacteraemia tests. In addition, if any MIC measurement for a particular strain–agent combination was censored (reported as or ), it could not be accurately compared with other measurements and so all replicates of that combination were excluded; 122/1085 respiratory and 642/3218 bacteraemia tests were excluded for this reason. In the Respiratory Programme, 936 MIC measurements on 19 antimicrobials remained for analysis, representing 183 isolate–agent combinations with 2–8 replicates each, and giving 2446 pairwise comparisons of replicate MICs. In the Bacteraemia Programme, analysis included 2449 MIC measurements on 29 antimicrobials (2–5 replicates for each of 611 isolate–agent combinations), giving 3967 pairwise MIC comparisons. Across both programmes and all test strains and antimicrobials, 50% of all repeated MICs agreed exactly, 90% agreed within ±1 doubling dilution and 98% agreed within ±2 doubling dilutions (Table 14). The standard deviation of the MIC differences was 0.97 in the respiratory study and 0.91 in bacteraemia.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Sampling of isolates
The advantages of antimicrobial resistance surveillance based on sentinel laboratories and centralized testing are many, among them being higher quality, more reproducible testing, consistent core panels of antimicrobials, the ability to investigate unusual isolates in detail and the possibility of testing new and developmental agents. Set against those advantages is the disadvantage of sampling the isolates. First, the number of isolates sampled and tested is necessarily small compared with the national total, reducing the power to detect subtle trends and changes in resistance. Secondly, practical considerations make it extremely difficult to obtain a truly random and representative sample of clinical isolates for central testing. The sampling design of the BSAC Resistance Surveillance Programmes has been pragmatic, and consequently large and specialist/teaching hospital Trusts are over-represented. If large specialist hospitals have a different pattern of antimicrobial resistance compared with smaller general hospitals, the resistance rates measured in the BSAC Programmes will be biased away from the true overall level. However, comparison with voluntarily reported LabBase data, which now captures 70% of clinically significant bacteraemias in England and Wales, shows that neither the proportions of the various pathogens nor their susceptibility rates are grossly different in the BSAC sentinel laboratories from the national average; thus although there may be some bias, it does not appear to be overwhelming. Whether it is nationally representative or not, the panel of sentinel laboratories has been quite stable, and where laboratories have withdrawn they have been replaced by similar centres in the same region. This provides a secure platform for the analysis of trends in antimicrobial resistance over time and, similarly, for the investigation of relationships between resistance and potential predictive factors such as patient age.
The BSAC agar dilution method for MIC measurement is an established and standardized technique. The analysis of repeated MIC measurements showed that it gave reasonably repeatable results over several years within each central laboratory, with 50%, 90% and 98% of repeated MIC measurements agreeing exactly, within ±1 dilution and within ±2 dilutions, respectively. A previous pilot study gave very similar results for within-laboratory repeatability over a shorter timescale.31 Conversely, however, the results indicate the variability of the method: over 40% of repeated MIC measurements did not agree exactly, and this degree of experimental variation could be important, for example, if part of it arises between experimental runs, rather than between individual isolates within runs. In that case, in surveillance programmes like these, where a whole season’s isolates of a particular species are tested in only two or three runs over a short period at the end of the season, between-run variation could translate into experimental variation between seasons, and this could be mistaken for true changes in the susceptibility of the bacteria over time, particularly when the breakpoint lies close to the main MIC distribution. Externally supplied control isolates are too few to clearly detect such variations between years, but close inspection of the MIC distributions of the surveillance isolates themselves suggests that wholesale shifts between years of around half a doubling dilution in MIC for particular organism–agent combinations are not uncommon. Further investigation of the source of experimental variation and its potential impact on the interpretation of surveillance studies would be worthwhile.
It is essential to take account of variation between collecting centres when analysing resistance surveillance studies, as otherwise invalid conclusions may be reached unacceptably often.24 We have used random effects logistic regression models to meet this need, or cluster-robust standard errors in cases such as multinomial regression where random effects models were not readily available in software. We have also considered the issue of multiple testing arising from the many outcomes (non-susceptibility to different antimicrobials) and potential predictors investigated for each species or organism group. We have been appropriately cautious in the interpretation of tests in order to avoid excessive levels of type 1 error (concluding that a trend or relationship exists when, in fact, it does not). The corollary may be that some true trends and relationships have been missed. Simulation studies, reported in more detail separately,24 showed that the BSAC Surveillance Programmes have adequate power, in most cases, to detect a doubling of the odds of resistance over 5 years, or in one group compared with another e.g. those in hospital >48 h compared with other patients: in favourable circumstances they could identify smaller effects. The number of S. aureus and E. coli collected will be doubled from 2008 onwards to increase the ability of the bacteraemia study to recognize more subtle trends among the two most important causative agents of bloodstream infection.
Each of the two BSAC Resistance Surveillance Programmes provides consistent and reliable information on antimicrobial susceptibility in the UK and Ireland by using an adequate number of contributing centres and isolates, combined with standardized microbiological methods applied by a central laboratory, and suitable methods of statistical analysis. The information is more detailed than that available from routine data collections because MICs, and not just susceptible/intermediate/resistant categories, are recorded for every organism and agent; this allows re-analysis in the light of changing circumstances, for example, changes in breakpoints. The results to date are presented on the project web site at www.bsacsurv.org and in six related papers in this Supplement.25–30 The collected isolates are archived and can be made available for further academic research.
|
Funding |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
The BSAC Resistance Surveillance Programmes up to 2006 (bacteraemia) and 2006–07 (respiratory) have received financial support from Abbott, AstraZeneca, Aventis, Basilea, Bayer, Cubist, GeneSoft, GlaxoSmithKline, Johnson & Johnson, Merck Sharp & Dohme, Novartis, Pfizer, Theravance, Wyeth or their predecessors. The BSAC funds the work of the Resistance Surveillance Coordinator (R. R.) and Resistance Surveillance Working Party.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
This article is part of a Supplement sponsored by the British Society for Antimicrobial Chemotherapy.
All authors have no conflicts of interest to declare.
|
Acknowledgements |
|---|
Thanks to Mark Lillie (HPA, London) for supplying the LabBase data (Tables 12 and 13) for comparison of sentinel and national bacteraemia reports. We are grateful to all who have contributed to the success of the BSAC Resistance Surveillance Project, in particular the many laboratories that have collected isolates and all who have played a part in testing them [see page ii10 (Acknowledgements)]. Additional information on the isolates collected in the Project is available on the BSAC Surveillance web site (www.bsacsurv.org) or through a link on the BSAC homepage (www.bsac.org.uk). See page ii12 (Publications) for a full list of previous publications from the Project, which may include parts of the information presented here.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
1 White AR, on behalf of the BSAC Working Parties on Resistance Surveillance. The British Society for Antimicrobial Chemotherapy Resistance Surveillance Project: a successful collaborative model. J Antimicrob Chemother (2008) 62(Suppl 2):ii3–14.
2
Reacher MH, Shah A, Livermore DM, et al. Bacteraemia and antibiotic resistance of its pathogens reported in England and Wales between 1990 and 1998: trend analysis. BMJ (2000) 320:213–6.
3 Health Protection Agency (2008). Annual counts and rates of MRSA bacteraemia April 2001 to September 2007. [Online.] http://www.hpa.org.uk/web/HPAwebFile/HPAweb_C/1202115525736 (26 June 2008, date last accessed).
4 Advisory Committee on Dangerous Pathogens (2005). Biological agents: managing the risks in laboratories and healthcare premises. [Online.] http://www.hse.gov.uk/biosafety/biologagents.pdf (3 July 2008, date last accessed).
5 IATA. Dangerous Goods Regulations 2007 (2006) 48th edn. International Air Transport Association.
6
Gribaldo S, Cookson B, Saunders N, et al. Rapid identification by specific PCR of coagulase-negative staphylococcal species important in hospital infection. J Med Microbiol (1997) 46:45–53.
7
Beighton D, Hardie JM, Whiley RA. A scheme for the identification of viridans streptococci. J Med Microbiol (1991) 35:367–72.
8
Facklam R. What happened to the streptococci: overview of taxonomic and nomenclature changes. Clin Microbiol Rev (2002) 15:613–30.
9
Whiley RA, Fraser H, Hardie JM, et al. Phenotypic differentiation of Streptococcus intermedius, Streptococcus constellatus, and Streptococcus anginosus strains within the ‘Streptococcus milleri group. J Clin Microbiol (1990) 28:1497–501.
10
Whiley RA, Hall LM, Hardie JM, et al. A study of small-colony, β-haemolytic, Lancefield group C streptococci within the anginosus group: description of Streptococcus constellatus subsp. pharyngis subsp. nov. associated with the human throat and pharyngitis. Int J Syst Bacteriol (1999) 49:1443–9.
11 Dutka-Malen S, Evers S, Courvalin P. Detection of glycopeptide resistance genotypes and identification to the species level of clinically relevant enterococci by PCR. J Clin Microbiol (1995) 33:24–7.[Abstract]
12 King E, Ward M, Raney D. Two simple media for the demonstration of pyocyanin and fluorescin. J Lab Clin Med (1954) 44:301–7.[Medline]
13 Sheppard C. The Detection and Characterisation of Pneumococci in Culture Negative Infections. London: Queen Mary College, University of London. PhD Thesis.
14 Andrews JM. Determination of minimum inhibitory concentrations. J Antimicrob Chemother (2001) 48(Suppl S1):5–16.[Abstract]
15 BSAC Working Party on Susceptibility Testing (2007). BSAC disc diffusion method for antimicrobial susceptibility testing. Version 6.1, February 2007. [Online.] http://www.bsac.org.uk/_db/_documents/version_6.1.pdf (7 July 2008, date last accessed).
16 BSAC Working Party on Susceptibility Testing. Use of gradient tests for determination of MICs by BSAC methodology. [Online.] http://www.bsac.org.uk/_db/_documents/Use_of_Etest_for_Determination_of_MICs_July_2007.pdf (3 July 2008, date last accessed).
17 Livermore DM, Winstanley TG, Shannon KP. Interpretative reading: recognizing the unusual and inferring resistance mechanisms from resistance phenotypes. J Antimicrob Chemother (2001) 48(Suppl 1):87–102.[Abstract]
18
Potz NAC, Colman M, Warner M, et al. False-positive extended-spectrum β-lactamase tests for Klebsiella oxytoca strains hyperproducing K1 β-lactamase. J Antimicrob Chemother (2004) 53:545–7.
19
Woodford N, Fagan EJ, Ellington MJ. Multiplex PCR for rapid detection of genes encoding CTX-M extended-spectrum β-lactamases. J Antimicrob Chemother (2006) 57:154–5.
20
Bignardi GE, Woodford N, Chapman A, et al. Detection of the mec-A gene and phenotypic detection of resistance in Staphylococcus aureus isolates with borderline or low-level methicillin resistance. J Antimicrob Chemother (1996) 37:53–63.
21
Geha DJ, Uhl JR, Gustaferro CA, et al. Multiplex PCR for identification of methicillin-resistant staphylococci in the clinical laboratory. J Clin Microbiol (1994) 32:1768–72.
22
Woodford N, Watson AP, Patel S, et al. Heterogeneous location of the mupA high-level mupirocin resistance gene in Staphylococcus aureus. J Med Microbiol (1998) 47:829–35.
23 EUCAST Clinical Breakpoints. [Online.] http://www.srga.org/eucastwt/MICTAB/index.html (17 March 2008, date last accessed).
24
Reynolds R, Lambert P, Burton P, et al. Analysis, power and design of antimicrobial resistance surveillance studies, taking account of inter-centre variation and turnover. J Antimicrob Chemother (2008) 62(Suppl 2):ii29–39.
25
Brown DFJ, Hope R, Livermore D, et al. Non-susceptibility trends among enterococci and non-pneumococcal streptococci from bacteraemias in the UK and Ireland, 2001 to 2006. J Antimicrob Chemother (2008) 62(Suppl 2):ii75–85.
26
Farrell D, Felmingham D, Shackcloth J, et al. Non-susceptibility trends and serotype distributions among Streptococcus pneumoniae from community-acquired respiratory tract infections and from bacteraemias in the UK and Ireland, 1999 to 2007. J Antimicrob Chemother (2008) 62(Suppl 2):ii87–95.
27
Hope R, Livermore DM, Brick G, et al. Non-susceptibility trends among staphylococci from bacteraemias in the UK and Ireland, 2001 to 2006. J Antimicrob Chemother (2008) 62(Suppl 2):ii65–74.
28
Livermore DM, Hope R, Brick G, et al. Non-susceptibility trends among Enterobacteriaceae from bacteraemias in the UK and Ireland, 2001 to 2006. J Antimicrob Chemother (2008) 62(Suppl 2):ii41–54.
29
Livermore DM, Hope R, Brick G, et al. Non-susceptibility trends among Pseudomonas aeruginosa and other non-fermentative Gram-negative bacteria from bacteraemias in the UK and Ireland, 2001 to 2006. J Antimicrob Chemother (2008) 62(Suppl 2):ii55–63.
30
Morrissey I, Maher K, Williams L, et al. Non-susceptibility trends among Haemophilus influenzae and Moraxella catarrhalis from community-acquired respiratory tract infections in the UK and Ireland, 1999 to 2007. J Antimicrob Chemother (2008) 62(Suppl 2):ii97–103.
31
Reynolds R, Shackcloth J, Felmingham D, et al. Comparison of BSAC agar dilution and NCCLS broth microdilution MIC methods for in vitro susceptibility testing of Streptococcus penumoniae, Haemophilus influenzae and Moraxella catarrhalis: the BSAC Respiratory Resistance Surveillance Programme. J Antimicrob Chemother (2003) 52:925–30.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Reynolds Antimicrobial resistance in the UK and Ireland J. Antimicrob. Chemother., September 1, 2009; 64(suppl_1): i19 - i23. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. White and on behalf of the BSAC Working Parties on Resistanc The British Society for Antimicrobial Chemotherapy Resistance Surveillance Project: a successful collaborative model J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii3 - ii14. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Reynolds, P. C. Lambert, P. R. Burton, and on behalf of the BSAC Extended Working Parties on Analysis, power and design of antimicrobial resistance surveillance studies, taking account of inter-centre variation and turnover J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii29 - ii39. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Livermore, R. Hope, G. Brick, M. Lillie, R. Reynolds, and on behalf of the BSAC Working Parties on Resistanc Non-susceptibility trends among Enterobacteriaceae from bacteraemias in the UK and Ireland, 2001-06 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii41 - ii54. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Livermore, R. Hope, G. Brick, M. Lillie, R. Reynolds, and on behalf of the BSAC Working Parties on Resistanc Non-susceptibility trends among Pseudomonas aeruginosa and other non-fermentative Gram-negative bacteria from bacteraemias in the UK and Ireland, 2001-06 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii55 - ii63. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Hope, D. M. Livermore, G. Brick, M. Lillie, R. Reynolds, and on behalf of the BSAC Working Parties on Resistanc Non-susceptibility trends among staphylococci from bacteraemias in the UK and Ireland, 2001-06 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii65 - ii74. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. F. J. Brown, R. Hope, D. M. Livermore, G. Brick, K. Broughton, R. C. George, R. Reynolds, and on behalf of the BSAC Working Parties on Resistanc Non-susceptibility trends among enterococci and non-pneumococcal streptococci from bacteraemias in the UK and Ireland, 2001-06 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii75 - ii85. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. J. Farrell, D. Felmingham, J. Shackcloth, L. Williams, K. Maher, R. Hope, D. M. Livermore, R. C. George, G. Brick, S. Martin, et al. Non-susceptibility trends and serotype distributions among Streptococcus pneumoniae from community-acquired respiratory tract infections and from bacteraemias in the UK and Ireland, 1999 to 2007 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii87 - ii95. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Morrissey, K. Maher, L. Williams, J. Shackcloth, D. Felmingham, R. Reynolds, and on behalf of the BSAC Working Parties on Resistanc Non-susceptibility trends among Haemophilus influenzae and Moraxella catarrhalis from community-acquired respiratory tract infections in the UK and Ireland, 1999-2007 J. Antimicrob. Chemother., November 1, 2008; 62(suppl_2): ii97 - ii103. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||