Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Comparison of Clinical Laboratory Standards Institute and European Committee on Antimicrobial Susceptibility Testing guidelines for the interpretation of antibiotic susceptibility at a University teaching hospital in Nairobi, Kenya: a cross-sectional study

Comparison of Clinical Laboratory Standards Institute and European Committee on Antimicrobial... Background: The Clinical Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Sus- ceptibility Testing (EUCAST ) guidelines are the most popular breakpoint guidelines used in antimicrobial susceptibility testing worldwide. The EUCAST guidelines are freely available to users while CLSI is available for non-members as a package of three documents for US $500 annually. This is prohibitive for clinical microbiology laboratories in resource poor settings. We set out to compare antibiotic susceptibility determined by the two guidelines to determine whether adoption of EUCAST guidelines would significantly affect our susceptibility patterns. Methods: We reviewed minimum inhibitory concentrations (MIC) of various antibiotics routinely reported for Escherichia coli (E. coli), Staphylococcus aureus (S. aureus) and Pseudomonas aeruginosa (P. aeruginosa) isolates from an automated microbiology identification system ( VITEK-2) at the Aga Khan University Hospital Nairobi’s Pathology department. These MICs were then analyzed using both CLSI 2015 and EUCAST 2015 guidelines and classified as resistant, intermediate or susceptible. We compared the susceptibility and agreement between the CLSI and EUCAST categorizations. Results: Susceptibility data from a total of 5165 E. coli, 1103 S. aureus and 532 P. aeruginosa isolates were included. The concordance rates of the two guidelines for E. coli, S. aureus and P. aeruginosa ranged from 78.2 to 100 %, 94.6 to 100 % and 89.1 to 95.5 % respectively. The kappa statistics for E. coli MICs revealed perfect agreement between CLSI and EUCAST for cefotaxime, ceftriaxone and trimethoprim–sulfamethoxazole, almost perfect agreement for ampicillin, ciprofloxacin, cefuroxime, gentamicin and ceftazidime, substantial agreement for meropenem, moderate agreement for cefepime and amoxicillin-clavulanate, fair agreement for nitrofurantoin and poor agreement for amikacin. For S. aureus the kappa statistics revealed perfect agreement for penicillin, trimethoprim–sulfamethoxazole, levofloxacin, oxacillin, linezolid and vancomycin, almost perfect agreement for clindamycin, erythromycin and tetracycline and moderate agreement for gentamicin. For P. aeruginosa the kappa analysis revealed moderate to almost perfect agree- ment for all the anti-pseudomonal antibiotics. *Correspondence: alliassimo@gmail.com Department of Pathology, Aga Khan University Hospital, P.O. Box 30270, Nairobi 00100, Kenya © 2016 Kassim et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 2 of 7 Conclusion: The results show comparable antibiotic susceptibility patterns between CLSI and EUCAST breakpoints. Given that EUCAST guidelines are freely available, it makes it easier for laboratories in resource poor settings to have an updated and readily available reference for interpreting antibiotic susceptibilities. Keywords: European Committee on Antimicrobial Susceptibility Testing (EUCAST ), Clinical Laboratory Standards Institute (CLSI), Antimicrobial susceptibility testing (AST ), Minimum inhibitory concentrations (MIC) decision making process are not accessible to the public. Background Thirdly, the FDA has a major influence in determining Over the last two decades, there has been emergence and official clinical breakpoints before they are adopted by spread of antibiotic resistance in many bacterial clinical CLSI and this raises major concerns on the influence of pathogens [1]. Categorization of minimum inhibitory pharmaceutical industries in establishing the guidelines. concentrations (MIC) of various antibiotics in antimicro- CLSI has a voting committee comprising representa- bial susceptibility testing (AST) depends on breakpoints tives of both the profession and the industry and hence set by various international agencies. These breakpoints the industry plays a role in the decision making process. affect clinical decision making by determining whether As for EUCAST, the industry only plays a consultative an antibiotic is reported as susceptible or not. One of role and is not allowed to finance or participate in deci - the most popular guidelines used worldwide is from sion making [3]. EUCAST encourages the formation of the Clinical and Laboratory Standards Institute (CLSI) National Antimicrobial Susceptibility Testing Commit- whose interpretive cut offs for antibiotics is based on tees (NAC) that can send a representative to sit in its MIC distributions, pharmacokinetic–pharmacodynamic General Committee (GC) thus fostering inclusivity in the (PK-PD) properties and the mechanisms of resistance decision making process [3]. Finally, antibiotics that are [1]. The adoption of antibiotic susceptibility guidelines in not registered in the US may not be included in the CLSI the US is regulated by the Food and Drug Administration guidelines. (FDA). In 1997 various national agencies in Europe came For resource poor settings like Africa, there is need for together to form the European Committee on Antimicro- guidelines that are accessible and affordable while still bial Susceptibility Testing (EUCAST) that has since har- maintaining quality of reported susceptibility. Given that monized antibiotic interpretive breakpoints throughout antibiotic susceptibility is quite variable across different Europe and most European countries have since switched settings, we set out to compare the susceptibility rates of from CLSI and their local guidelines to EUCAST guide- selected antibiotics based on CLSI 2015 and EUCAST lines [2, 3]. EUCAST bases its clinical breakpoints on 2015 AST guidelines and the level of agreement between epidemiological MIC cut-offs (ECOFFS) and PK-PD the two guidelines. properties. All documents on MIC distributions and ECOFFS are freely available on the EUCAST website [4]. Methods Polsfuss et al. compared EUCAST with CLSI in screen- The study was carried out at the Aga Khan University ing for extended-spectrum beta-lactamase (ESBL) pro- Hospital, Nairobi’s (AKUHN) Department of Pathology. ducing Enterobacteriaceae isolates. They found no Ethical exemption was granted by the AKUHN’s research significant difference in the sensitivity of the two guide - and ethics committee (Ref 2015/REC-44) since this type lines in the detection of ESBL-producing isolates [5]. A of study is low risk and classified as a clinical audit. This more recent study by Hombach et  al. demonstrated sig- was a retrospective study reviewing the MICs of various nificant differences in the susceptibility rates for drugs antimicrobials on one commonly isolated gram posi- including cefepime, ceftazidime and cefotaxime in the tive organism, Staphylococcus aureus (S. aureus), one detection of ESBLs between the CLSI 2013 and EUCAST fermenting gram negative organism, Escherichia coli (E. 2013 AST guidelines. This study recommended adjust - coli) and one non-fermenting gram negative organism, ments to the clinical breakpoints to further harmonize Pseudomonas aeruginosa (P. aeruginosa). MIC data for the two guidelines [6]. Since then 2014 and 2015 guide- E. coli, P. aeruginosa and S. aureus isolates was collected lines for both systems have been released in an effort to from two Vitek 2 (version 4.01, bioMerieux, Marcy- harmonize the clinical breakpoints [7, 8]. l’Etoile, France) automated microbiology systems for the CLSI guidelines have a number of disadvantages. period January 2012 to December 2014. The Vitek 2 AST- First, it is based on annual subscriptions of US $350 for P580 and AST-GN26/AST-GN83 cards were used for members and a cost of US $500 to non-members annu- antibiotic susceptibility for S. aureus and E. coli/P. aerugi- ally and this may be a problem for microbiology labora- nosa respectively. Ceftriaxone MICs were only available tories in resource poor settings. Secondly, details on the Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 3 of 7 for the year 2014 in the AST-GN83 cards for a total of gentamicin and meropenem between the two guidelines. 1673 E. coli isolates. The data was summarized using The concordance between the two guidelines ranged Microsoft Excel 2013 and imported into IBM (Interna- from 78.2 to 100 %. Table 1 summarizes the susceptibili- tional Business Machines, Corporation; Armonk, New ties, concordance and kappa statistics between the two York, United States of America) SPSS (Statistical Pack- guidelines for E. coli. age for the Social Sciences) Version 22.0 that was used The E. coli susceptibility patterns achieved after anal- for analysis. The MICs were then analyzed using both ysis using EUCAST 2015 and CLSI 2015 guidelines are the CLSI 2015 and EUCAST 2015 guidelines to catego- similar except for amoxicillin–clavulanate, amikacin and rize them as either susceptible, intermediate or resistant nitrofurantoin. Analysis year by year did not show any [7, 8]. The concordance rate between the two guidelines difference in the trends and overall susceptibilities. The in percentage was compared. The susceptibility rates for kappa analysis revealed perfect agreement for ceftriax- the various antimicrobial agents were also calculated in one, cefotaxime and trimethoprim–sulfamethoxazole percentages for each organism. Analysis of the extent of with kappa (κ) of 1 (p < 0.000). An almost perfect agree- agreement between CLSI 2015 and EUCAST 2015 for ment was noted with ampicillin κ = 0.985 (95 % CI 0.979, the various antimicrobials was carried out using Cohen’s 0.991), p < 0.000, ciprofloxacin κ  = 0.969 (95 % CI 0.963, kappa statistics and graded from perfect agreement to 0.975), p  <  0.000, cefuroxime κ  =  0.924 (95  % CI 0.914, poor agreement [9]. Cohen’s Kappa statistics determines 0.934), p  <  0.000, ceftazidime κ  =  0.859 (95  % CI 0.843, the proportion of agreement over and above chance 0.85), p < 0.000 and gentamicin κ = 0.979 (95 % CI 0.973, between two independent observations. This ranges from 0.985), p  <  0.000. Substantial agreement was noted with −1 to 1 and a p value less than 0.05 means that the agree- meropenem κ  =  0.724 (95  % CI 0.573, 0.875), p  <  0.000 ment reported is significantly different from 0 and is not while moderate agreement was noted with amoxicillin- due to chance. For all inferential statistics, a p value less clavulanate κ  =  0.581 (95  % CI 0.567, 0.595), p  <  0.000 than 0.05 was considered statistically significant. and cefepime κ = 0.600 (95 % CI 0.578, 0.622), p < 0.000. Fair agreement was seen with nitrofurantoin κ  =  0.351 Results (95  % CI 0.314, 0.388), p  <  0.000 while poor agreement A total of 5165 E. coli, 1103 S. aureus and 532 P. aerugi- was noted with amikacin κ  =  0.112 (95  % CI 0.079, nosa non-duplicate isolates were included in the analysis. 0.145), p < 0.000. Of the 5165 E. coli MICs analyzed, there was compara- Of the 1103 S. aureus MICs analyzed, susceptibility ble susceptibility to most antibiotics including ampicillin, to penicillin, oxacillin, levofloxacin, linezolid, trimetho - amoxicillin–clavulanate, cefuroxime, cefotaxime, ceftri- prim–sulfamethoxazole, vancomycin, clindamycin, axone, trimethoprim–sulfamethoxazole, ciprofloxacin, erythromycin and tetracycline was comparable between Table 1 Susceptibilities of E. coli to various antibiotics, concordance and kappa statistics between CLSI 2015 and EUCAST 2015 guidelines Antibiotic CLSI (%); n = 5165 EUCAST (%); n = 5165 Concordance (%) Kappa, κ (95 % CI) S I R S I R Ampicillin 13.8 0.5 85.7 13.8 0 86.2 99.5 0.985 (0.979, 0.991) Amox-Clav 55.8 23.8 20.4 55.8 0 44.2 78.2 0.581 (0.567, 0.595) TMP/SMX 0 0 100 0 0 100 100 1 Nitrofurantoin 84.6 11.9 3.5 96.5 0 3.5 87.6 0.351 (0.314, 0.388) Ciprofloxacin 57.3 0.2 42.5 55.9 1.4 42.7 98.4 0.969 (0.963, 0.975) Cefuroxime 63.7 3.5 32.8 63.7 0 36.3 96.5 0.924 (0.914, 0.934) Gentamicin 78.5 0.2 21.3 78.1 0.4 21.5 99.5 0.979 (0.973, 0.985) Amikacin 99.3 0.3 0.4 90.5 8.8 0.7 91.1 0.112 (0.079, 0.145) Cefotaxime 69.7 0.4 29.9 69.7 0.4 29.9 100 1 Ceftazidime 76.3 0.9 22.8 71.5 4.8 23.7 93.7 0.859 (0.843, 0.85) Ceftriaxone 67.7 0.1 32.2 67.7 0.1 32.2 100 1 Cefepime 80.5 10.0 9.5 72.9 9.8 17.3 84.9 0.600 (0.578, 0.622) Meropenem 99.7 0.1 0.2 99.8 0.1 0.1 99.7 0.724 (0.573, 0.875) Amox-Clav amoxicillin–clavulanate, TMP/SMX trimethoprim–sulfamethoxazole, S susceptible, I intermediate, R resistant Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 4 of 7 the two guidelines. The susceptibilities, concordance and ceftazidime, ciprofloxacin, cefepime, gentamicin, mero - kappa statistics between the two guidelines are shown in penem and piperacillin–tazobactum. The susceptibili - Table 2. ties, concordance and kappa statistics between the two For S. aureus the susceptibilities are generally very guidelines are shown in Table 3. Year by year analysis did similar between the two guidelines. Year by year analy- not show any differences in trends and susceptibility pat - sis did not show any differences in trends and suscepti - terns. The kappa analysis for CLSI 2015 and EUCAST bility patterns. The kappa analysis for CLSI 2015 and 2015 guidelines revealed moderate to almost perfect EUCAST 2015 guidelines revealed perfect agreement for agreement for all the anti-pseudomonal antibiotics. P levofloxacin, linezolid, vancomycin, oxacillin, penicillin values for the kappa statistics for all the antibiotics were and trimethoprim–sulfamethoxazole with a kappa sta- statistically significant at p < 0.000. Overall, the antibiotic tistic of 1 (p  <  0.000). An almost perfect agreement was susceptibility patterns were quite similar regardless of noted with clindamycin κ = 0.904 (95 % CI 0.826, 0.982), whether CLSI or EUCAST 2015 guidelines were used. p < 0.000, erythromycin κ = 0.978 (95 % CI 0.960, 0.996), p  <  0.000 and tetracycline κ  =  0.962 (95  % CI 0.940, Discussion 0.984), p  <  0.000. Moderate agreement was noted with The morbidity and mortality associated with communi - gentamicin κ = 0.537 (95 % CI 0.441, 0.633), p < 0.000. cable diseases including bacterial infections is quite sig- For the 532 P. aeruginosa analyzed there were similar nificant in developing countries [10]. Antibiotics play a susceptibility patterns noted between CLSI 2015 and critical role in treating such infections especially when EUCAST 2015. Antibiotics analyzed included amikacin, instituted in a timely fashion more so when the bacteria Table 2 Susceptibilities of  S.aureus to  various antibiotics and  the concordance and  kappa statistics between  CLSI 2015 and EUCAST 2015 guidelines Antibiotic CLSI (%); n = 1103 EUCAST (%); n = 1103 Concordance (%) Kappa, κ (95 % CI) S I R S I R Penicillin 10.6 0 89.4 10.6 0 89.4 100 1 TMP/SMX 0 0 100 0 0 100 100 1 Clindamycin 98.0 0.3 1.7 98.0 0.1 1.9 99.5 0.904 (0.826, 0.982) Erythromycin 85.6 0.5 13.9 86.0 0 14.0 99.4 0.978 (0.960, 0.996) Gentamicin 96.3 0.5 3.2 91.5 0 8.5 94.6 0.537 (0.441, 0.633) Levofloxacin 90.8 0.3 8.9 90.8 0.3 8.9 100 1 Linezolid 100 0 0 100 0 0 100 1 Oxacillin 93.3 0 6.7 93.3 0 6.7 100 1 Tetracycline 83.2 0.1 16.7 82.4 0.8 16.8 98.9 0.962 (0.940, 0.984) Vancomycin 100 0 0 100 0 0 100 1 TMP/SMX trimethoprim–sulfamethoxazole, S susceptible, I intermediate, R resistant Table 3 Susceptibilities of  P.aeruginosa to  various antibiotics and  the concordance and  kappa statistics between  CLSI 2015 and EUCAST 2015 guidelines CLSI (%); n = 532 EUCAST (%); n = 532 Concordance (%) Kappa, κ (95 % CI) S I R S I R Amikacin 79.5 3.0 17.5 72.2 7.3 20.5 89.7 0.734 (0.703; 0.765) Ceftazidime 70.9 4.7 24.4 70.9 0 29.1 95.3 0.890 (0.870; 0.910) Cefepime 72.6 4.7 22.7 72.6 0 27.4 95.3 0.886 (0.866; 0.906) Gentamicin 72.6 6.2 21.2 72.6 0 27.4 93.8 0.851 (0.829; 0.873) Meropenem 64.1 8.3 27.6 64.1 12.8 23.1 95.5 0.912 (0.895; 0.929) Pip-taz 64.5 8.3 27.3 64.5 0 35.5 91.7 0.830 (0.808; 0.852) Ciprofloxacin 71.8 5.3 22.9 66.2 5.6 28.2 89.1 0.762 (0.736; 0.788) Pip-Taz piperacillin–tazobactum, CI confidence interval, S susceptible, I intermediate, R resistant Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 5 of 7 is susceptible to the antibiotic given. The determina - of isolates labeled as intermediate by EUCAST were all tion of accurate antibiotic susceptibility is therefore an categorized as susceptible by CLSI. EUCAST guidelines important cog in the clinical care of bacterial infections eliminated the intermediate category for some antibiot- especially in organisms that possess acquired resistance ics and this explains the reduced level of agreement for mechanisms and careful consideration should be given some of the antibiotics. For example, CLSI categorized when deciding how to interpret phenotypic susceptibil- 23.8  % of E. coli as having intermediate susceptibility to ity data [11]. In Kenya, many laboratories have adopted amoxicillin-clavulanate but were all categorized as resist- CLSI guidelines as a basis for interpreting their suscep- ant by EUCAST while 11.9 % categorized as intermediate tibility data despite the fact that it costs between US susceptibility to nitrofurantoin by CLSI were categorized $300–$500. These guidelines are updated annually and as susceptible by EUCAST. From a clinical stand point, therefore require laboratories to keep on purchasing reclassifying amoxicillin-clavulanate from intermedi- them at a cost that is prohibitive to most public labo- ate to resistant is unlikely to adversely affect the patient ratories. Failure to stay updated may result in misinter- as it simply removes it from being a therapeutic consid- pretation of susceptibility. For example in 2013, CLSI eration. As for nitrofurantoin, its ability to concentrate abandoned the use of oxacillin disc diffusion in determin - in urine enables it to achieve significant concentrations ing whether a S. aureus isolate is methicillin resistant in and eliminate isolates that may have intermediate sus- favour of cefoxitin because it was more accurate in deter- ceptibility. Therefore, the reclassification of isolates that mining the presence of a mecA mediated mechanism of are intermediate by CLSI to susceptible by EUCAST is resistance [12]. In 2012, it adopted a new disc diffusion unlikely to contribute to adverse outcomes for patients susceptibility cut off for ciprofloxacin of 31 mm up from with urinary tract infections. The EUCAST guidelines 21  mm for non-typhoidal Salmonella (NTS) isolated have slightly more stringent breakpoints for some antibi- from invasive specimens and all Salmonella Typhi iso- otics in an effort to curb the inappropriate use of antibi - lated from both invasive and non-invasive specimens. otics and control the rising rates of antibiotic resistance. The MIC for the same was reduced from 1 to 0.06 μg/mL For S. aureus, all except one of the antibiotics had mod- resulting in many Salmonella spp. previously reported erate to perfect agreement and three quarters of them as susceptible to fall into the intermediate category, and had almost perfect or perfect agreement with kappa many that were intermediate were now categorized as values ranging from 0.862 to 1. The two guidelines per - resistant [13]. In 2013, CLSI recommended that these cut formed equally in the detection of the rate of methicillin offs should apply to all Salmonella spp. including NTS resistant Staphylococcus aureus (MRSA) of 7.2  %. Van- from non-invasive specimens [12, 14]. In the same year comycin susceptible Staphylococcus aureus (VSSA) were new levofloxacin and ofloxacin breakpoints were intro - also detected equally by the two guidelines. EUCAST duced for Salmonella spp. including Salmonella Typhi. eliminated the intermediate category for vancomycin These changes were motivated by an increased risk of in a bid to discourage the reporting of Glycopeptide- treatment failure in patients with decreased ciprofloxa - intermediate Staphylococcus aureus (GISA) due to poor cin susceptible Salmonella spp [15]. In 2012 again, CLSI response even to increased doses of vancomycin [8, 16]. reduced meropenem and piperacillin-tazobactum MIC In view of the fact that no vancomycin resistant Staphy- breakpoints for P. aeruginosa from  ≤4 to  ≤2  ug/mL lococcus aureus (VRSA) has been identified in our set up and  ≤64  ug/mL to  ≤16  ug/mL respectively [13]. These as yet, these changes are unlikely to influence interpre - examples highlight the importance of remaining up-to- tation of breakpoints. The moderate level of agreement date and emphasizes the need for an affordable, up-to- for gentamicin is due to the more stringent breakpoints date and readily available guideline for the interpretation by EUCAST leading to a much higher resistance rate of of antibiotic susceptibility. 8.6 % compared to 3.2 % by CLSI. The difference between In our comparison of CLSI and EUCAST guidelines the two susceptibility cut-offs is two dilutions and this for interpretation of antibiotic susceptibility, for E. coli, may require further harmonization. Gentamicin is rarely most of the antibiotics had moderate to perfect agree- used as monotherapy in treating gram positive bacteria ment between the two guidelines with kappa values and as such this difference in MIC cut off is unlikely to be ranging from 0.581 to 1 and two-thirds of them having clinically significant. almost perfect or perfect agreement with kappa values For P. aeruginosa, five out of seven of the antibiotics ranging from 0.859 to 1. Poor agreement was noted with had almost perfect agreement with kappa values ranging amikacin with EUCAST having a more stringent break- from 0.830 to 0.912 with the remaining two having mod- point for susceptibility of ≤8 mg/L compared to the CLSI erate agreement with kappa values of 0.762 and 0.734. breakpoint of  ≤16  mg/L. The major discrepancy was EUCAST abolished the intermediate category for ceftazi- in the intermediate and susceptible categories as 8.8  % dime, cefepime, gentamicin and piperacillin-tazobactum Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 6 of 7 Food and Drug Administration; GC: General Committee; GISA: Glycopeptide- reclassifying the MICs as resistant. This accounts for intermediate Staphylococcus aureus; IBM: International Business Machines the slightly reduced level of agreement. For amikacin Corporation (Armonk, New York, United States of America); MIC: minimum and ciprofloxacin, with moderate agreement, EUCAST inhibitory concentration; MRSA: methicillin resistant Staphylococcus aureus; NAC: National Antimicrobial Susceptibility Testing Committees; NTS: non- has more stringent breakpoints for susceptibility of  ≤8 typhoidal salmonella; Pip-Taz: piperacillin–tazobactum; PK-PD: pharmacoki- and  ≤0.5  ug/mL compared to CLSI breakpoints of  ≤16 netic–pharmacodynamics; SPSS: statistical package for the social sciences; S. and  ≤1  ug/mL respectively. These accounted for the aureus: Staphylococcus aureus; TMP/SMX: trimethoprim–sulfamethoxazole; US $: United States of America dollar; VRSA: vancomycin resistant Staphylococcus reduced level of agreement and may require further aureus; VSSA: vancomycin susceptible Staphylococcus aureus. harmonization between the two breakpoints. For mero- penem, EUCAST uses a resistant breakpoint of  >8  ug/ Authors’ contributions AK and GO contributed equally in conceiving and designing the study, data mL while CLSI uses  ≥8  ug/mL. In effect this has led to collection, data analysis and drafting the manuscript. ZP and GR participated a slightly higher meropenem resistance of 27.6  % com- in study design, coordination and critical revision of the manuscript. All pared to 23.1 % for CLSI and EUCAST respectively. This authors read and approved the final manuscript. is another aspect that will require harmonization. Our study is limited by the fact that we only compared Acknowledgements the susceptibility for three bacteria whose results may not None. necessarily be generalizable to all clinically relevant gram Competing interests positive and negative bacteria. However, these three bac- The authors declare that they have no competing interests. teria represent a significant proportion of common bac - Received: 22 January 2016 Accepted: 25 March 2016 terial pathogens both in the community and healthcare settings namely Enterobacteriaceae, non-fermenting gram negative bacteria and Staphylococci. The results obtained are also limited to MICs generated by an auto- mated bacterial identification system which is not widely References 1. Marchese A, Esposito S, Barbieri R, Bassetti M, Debbia E. Does the adop- used in developing countries. For P. aeruginosa, Colis- tion of EUCAST susceptibility breakpoints affect the selection of antimi- tin had not yet been included in the gram negative AST crobials to treat acute community-acquired respiratory tract infections? cards being used at the time the study was being con- BMC Infect Dis. 2012;12(1):181. 2. EUCAST. http://www.eucast.org/. Accessed 31 Aug 2015. ducted. Colistin Etest (bioMerieux, Durham, NC, USA) 3. Kahlmeter G. The 2014 Garrod Lecture: EUCAST—are we heading towards was only run for multi-drug resistant isolates of clinical international agreement? J Antimicrob Chemother. 2015;70(9):2427–39. significance but this data was not available for analysis. http://www.jac.oxfordjournals.org/lookup/doi/10.1093/jac/dkv145. 4. Wolfensberger A, Sax H, Weber R, Zbinden R, Kuster SP, Hombach M. In most laboratories in sub-Saharan Africa, disk diffusion Change of antibiotic susceptibility testing guidelines from CLSI to is the preferred mode of antibiotic susceptibility testing. EUCAST: influence on cumulative hospital antibiograms. PLoS One. However, disk diffusion cut offs generally approximate 2013;8(11):1–8. 5. Polsfuss S, Bloemberg GV, Giger J, Meyer V, Hombach M. Comparison of MIC cut offs fairly well and we think a similar compari - European Committee on Antimicrobial Susceptibility Testing (EUCAST ) son based on disc diffusion cut offs would yield similar and CLSI screening parameters for the detection of extended-spectrum results. β-lactamase production in clinical enterobacteriaceae isolates. J Antimi- crob Chemother. 2011;2012(67):159–66. 6. Hombach M, Mouttet B, Bloemberg GV. Consequences of revised CLSI Conclusion and EUCAST guidelines for antibiotic susceptibility patterns of ESBL- and Our study shows acceptable level of agreement between AmpC β-lactamase-producing clinical Enterobacteriaceae isolates. J Antimicrob Chemother. 2013;68:2092–8. EUCAST and CLSI 2015 AST guidelines for E. coli, S. 7. CLSI. M100-S25 performance standards for antimicrobial susceptibility aureus and P. aeruginosa and laboratories with similar testing; Twenty-fifth informational supplement; 2015. antibiotic susceptibility patterns may choose to adopt 8. EUCAST: European Committee on Antimicrobial Susceptibility Testing breakpoint tables for interpretation of MICs and zone diameters. 2015. either guideline without fear of significantly altering http://www.eucast.org/fileadmin/src/media/PDFs/EUC. reported antibiotic susceptibility. With EUCAST guide- 9. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa lines being freely available it should be considered as statistic. Fam Med. 2005;37(5):360–3. 10. GBD_report_2004update_full.pdf. http://www.who.int/healthinfo/ an alternative especially in resource poor settings in global_burden_disease/GBD_report_2004update_full.pdf. Accessed 22 order to maintain up-to-date antibiotic susceptibility Dec 2015. interpretation. 11. Jorgensen JH, Ferraro MJ. Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clin Infect Dis. 2009;49(11):1749–55. 12. CLSI. M100-S23 performance standards for antimicrobial susceptibility Abbreviations testing; twenty-third informational supplement; 2013. AKUHN: Aga Khan University Hospital, Nairobi; Amox-Clav: amoxicillin–cla- 13. CLSI. M100-S22: performance standards for antimicrobial susceptibil- vulanate; AST: antimicrobial susceptibility testing; CLSI: Clinical Laboratory ity testing; twenty-second informational supplement—M100S22E.pdf. Standards Institute; ECOFFS: epidemiological cut-offs; ESBL: extended http://www.antimicrobianos.com.ar/ATB/wp-content/uploads/2012/11/ spectrum beta lactamase; Etest: epsilometer test; EUCAST: European Com- M100S22E.pdf. Accessed 22 Dec 2015. mittee on Antimicrobial Susceptibility Testing; E. coli: Escherichia coli; FDA: Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 7 of 7 14. Humphries RM, Fang FC, Aarestrup FM, Hindler JA. In vitro susceptibility 16. Uçkay I, Bernard L, Buzzi M, Harbarth S, François P, Huggler E, et al. High testing of fluoroquinolone activity against Salmonella: recent changes to prevalence of isolates with reduced glycopeptide susceptibility in CLSI standards. Clin Infect Dis. 2012;55(8):1107–13. persistent or recurrent bloodstream infections due to methicillin-resistant 15. Crump JA, Kretsinger K, Gay K, Hoekstra RM, Vugia DJ, Hurd S, et al. Clini- Staphylococcus aureus. Antimicrob Agents Chemother. 2012;56(3):1258– cal response and outcome of infection with Salmonella enterica serotype 64. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3294919 typhi with decreased susceptibility to fluoroquinolones: A United States &tool=pmcentrez&rendertype=abstract. FoodNet multicenter retrospective cohort study. Antimicrob Agents Chemother. 2008;52(4):1278–84. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Clinical Microbiology and Antimicrobials Springer Journals

Comparison of Clinical Laboratory Standards Institute and European Committee on Antimicrobial Susceptibility Testing guidelines for the interpretation of antibiotic susceptibility at a University teaching hospital in Nairobi, Kenya: a cross-sectional study

Loading next page...
 
/lp/springer-journals/comparison-of-clinical-laboratory-standards-institute-and-european-jfohF5tVnh

References (15)

Publisher
Springer Journals
Copyright
Copyright © 2016 by Kassim et al.
Subject
Biomedicine; Medical Microbiology; Infectious Diseases
eISSN
1476-0711
DOI
10.1186/s12941-016-0135-3
pmid
27068515
Publisher site
See Article on Publisher Site

Abstract

Background: The Clinical Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Sus- ceptibility Testing (EUCAST ) guidelines are the most popular breakpoint guidelines used in antimicrobial susceptibility testing worldwide. The EUCAST guidelines are freely available to users while CLSI is available for non-members as a package of three documents for US $500 annually. This is prohibitive for clinical microbiology laboratories in resource poor settings. We set out to compare antibiotic susceptibility determined by the two guidelines to determine whether adoption of EUCAST guidelines would significantly affect our susceptibility patterns. Methods: We reviewed minimum inhibitory concentrations (MIC) of various antibiotics routinely reported for Escherichia coli (E. coli), Staphylococcus aureus (S. aureus) and Pseudomonas aeruginosa (P. aeruginosa) isolates from an automated microbiology identification system ( VITEK-2) at the Aga Khan University Hospital Nairobi’s Pathology department. These MICs were then analyzed using both CLSI 2015 and EUCAST 2015 guidelines and classified as resistant, intermediate or susceptible. We compared the susceptibility and agreement between the CLSI and EUCAST categorizations. Results: Susceptibility data from a total of 5165 E. coli, 1103 S. aureus and 532 P. aeruginosa isolates were included. The concordance rates of the two guidelines for E. coli, S. aureus and P. aeruginosa ranged from 78.2 to 100 %, 94.6 to 100 % and 89.1 to 95.5 % respectively. The kappa statistics for E. coli MICs revealed perfect agreement between CLSI and EUCAST for cefotaxime, ceftriaxone and trimethoprim–sulfamethoxazole, almost perfect agreement for ampicillin, ciprofloxacin, cefuroxime, gentamicin and ceftazidime, substantial agreement for meropenem, moderate agreement for cefepime and amoxicillin-clavulanate, fair agreement for nitrofurantoin and poor agreement for amikacin. For S. aureus the kappa statistics revealed perfect agreement for penicillin, trimethoprim–sulfamethoxazole, levofloxacin, oxacillin, linezolid and vancomycin, almost perfect agreement for clindamycin, erythromycin and tetracycline and moderate agreement for gentamicin. For P. aeruginosa the kappa analysis revealed moderate to almost perfect agree- ment for all the anti-pseudomonal antibiotics. *Correspondence: alliassimo@gmail.com Department of Pathology, Aga Khan University Hospital, P.O. Box 30270, Nairobi 00100, Kenya © 2016 Kassim et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 2 of 7 Conclusion: The results show comparable antibiotic susceptibility patterns between CLSI and EUCAST breakpoints. Given that EUCAST guidelines are freely available, it makes it easier for laboratories in resource poor settings to have an updated and readily available reference for interpreting antibiotic susceptibilities. Keywords: European Committee on Antimicrobial Susceptibility Testing (EUCAST ), Clinical Laboratory Standards Institute (CLSI), Antimicrobial susceptibility testing (AST ), Minimum inhibitory concentrations (MIC) decision making process are not accessible to the public. Background Thirdly, the FDA has a major influence in determining Over the last two decades, there has been emergence and official clinical breakpoints before they are adopted by spread of antibiotic resistance in many bacterial clinical CLSI and this raises major concerns on the influence of pathogens [1]. Categorization of minimum inhibitory pharmaceutical industries in establishing the guidelines. concentrations (MIC) of various antibiotics in antimicro- CLSI has a voting committee comprising representa- bial susceptibility testing (AST) depends on breakpoints tives of both the profession and the industry and hence set by various international agencies. These breakpoints the industry plays a role in the decision making process. affect clinical decision making by determining whether As for EUCAST, the industry only plays a consultative an antibiotic is reported as susceptible or not. One of role and is not allowed to finance or participate in deci - the most popular guidelines used worldwide is from sion making [3]. EUCAST encourages the formation of the Clinical and Laboratory Standards Institute (CLSI) National Antimicrobial Susceptibility Testing Commit- whose interpretive cut offs for antibiotics is based on tees (NAC) that can send a representative to sit in its MIC distributions, pharmacokinetic–pharmacodynamic General Committee (GC) thus fostering inclusivity in the (PK-PD) properties and the mechanisms of resistance decision making process [3]. Finally, antibiotics that are [1]. The adoption of antibiotic susceptibility guidelines in not registered in the US may not be included in the CLSI the US is regulated by the Food and Drug Administration guidelines. (FDA). In 1997 various national agencies in Europe came For resource poor settings like Africa, there is need for together to form the European Committee on Antimicro- guidelines that are accessible and affordable while still bial Susceptibility Testing (EUCAST) that has since har- maintaining quality of reported susceptibility. Given that monized antibiotic interpretive breakpoints throughout antibiotic susceptibility is quite variable across different Europe and most European countries have since switched settings, we set out to compare the susceptibility rates of from CLSI and their local guidelines to EUCAST guide- selected antibiotics based on CLSI 2015 and EUCAST lines [2, 3]. EUCAST bases its clinical breakpoints on 2015 AST guidelines and the level of agreement between epidemiological MIC cut-offs (ECOFFS) and PK-PD the two guidelines. properties. All documents on MIC distributions and ECOFFS are freely available on the EUCAST website [4]. Methods Polsfuss et al. compared EUCAST with CLSI in screen- The study was carried out at the Aga Khan University ing for extended-spectrum beta-lactamase (ESBL) pro- Hospital, Nairobi’s (AKUHN) Department of Pathology. ducing Enterobacteriaceae isolates. They found no Ethical exemption was granted by the AKUHN’s research significant difference in the sensitivity of the two guide - and ethics committee (Ref 2015/REC-44) since this type lines in the detection of ESBL-producing isolates [5]. A of study is low risk and classified as a clinical audit. This more recent study by Hombach et  al. demonstrated sig- was a retrospective study reviewing the MICs of various nificant differences in the susceptibility rates for drugs antimicrobials on one commonly isolated gram posi- including cefepime, ceftazidime and cefotaxime in the tive organism, Staphylococcus aureus (S. aureus), one detection of ESBLs between the CLSI 2013 and EUCAST fermenting gram negative organism, Escherichia coli (E. 2013 AST guidelines. This study recommended adjust - coli) and one non-fermenting gram negative organism, ments to the clinical breakpoints to further harmonize Pseudomonas aeruginosa (P. aeruginosa). MIC data for the two guidelines [6]. Since then 2014 and 2015 guide- E. coli, P. aeruginosa and S. aureus isolates was collected lines for both systems have been released in an effort to from two Vitek 2 (version 4.01, bioMerieux, Marcy- harmonize the clinical breakpoints [7, 8]. l’Etoile, France) automated microbiology systems for the CLSI guidelines have a number of disadvantages. period January 2012 to December 2014. The Vitek 2 AST- First, it is based on annual subscriptions of US $350 for P580 and AST-GN26/AST-GN83 cards were used for members and a cost of US $500 to non-members annu- antibiotic susceptibility for S. aureus and E. coli/P. aerugi- ally and this may be a problem for microbiology labora- nosa respectively. Ceftriaxone MICs were only available tories in resource poor settings. Secondly, details on the Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 3 of 7 for the year 2014 in the AST-GN83 cards for a total of gentamicin and meropenem between the two guidelines. 1673 E. coli isolates. The data was summarized using The concordance between the two guidelines ranged Microsoft Excel 2013 and imported into IBM (Interna- from 78.2 to 100 %. Table 1 summarizes the susceptibili- tional Business Machines, Corporation; Armonk, New ties, concordance and kappa statistics between the two York, United States of America) SPSS (Statistical Pack- guidelines for E. coli. age for the Social Sciences) Version 22.0 that was used The E. coli susceptibility patterns achieved after anal- for analysis. The MICs were then analyzed using both ysis using EUCAST 2015 and CLSI 2015 guidelines are the CLSI 2015 and EUCAST 2015 guidelines to catego- similar except for amoxicillin–clavulanate, amikacin and rize them as either susceptible, intermediate or resistant nitrofurantoin. Analysis year by year did not show any [7, 8]. The concordance rate between the two guidelines difference in the trends and overall susceptibilities. The in percentage was compared. The susceptibility rates for kappa analysis revealed perfect agreement for ceftriax- the various antimicrobial agents were also calculated in one, cefotaxime and trimethoprim–sulfamethoxazole percentages for each organism. Analysis of the extent of with kappa (κ) of 1 (p < 0.000). An almost perfect agree- agreement between CLSI 2015 and EUCAST 2015 for ment was noted with ampicillin κ = 0.985 (95 % CI 0.979, the various antimicrobials was carried out using Cohen’s 0.991), p < 0.000, ciprofloxacin κ  = 0.969 (95 % CI 0.963, kappa statistics and graded from perfect agreement to 0.975), p  <  0.000, cefuroxime κ  =  0.924 (95  % CI 0.914, poor agreement [9]. Cohen’s Kappa statistics determines 0.934), p  <  0.000, ceftazidime κ  =  0.859 (95  % CI 0.843, the proportion of agreement over and above chance 0.85), p < 0.000 and gentamicin κ = 0.979 (95 % CI 0.973, between two independent observations. This ranges from 0.985), p  <  0.000. Substantial agreement was noted with −1 to 1 and a p value less than 0.05 means that the agree- meropenem κ  =  0.724 (95  % CI 0.573, 0.875), p  <  0.000 ment reported is significantly different from 0 and is not while moderate agreement was noted with amoxicillin- due to chance. For all inferential statistics, a p value less clavulanate κ  =  0.581 (95  % CI 0.567, 0.595), p  <  0.000 than 0.05 was considered statistically significant. and cefepime κ = 0.600 (95 % CI 0.578, 0.622), p < 0.000. Fair agreement was seen with nitrofurantoin κ  =  0.351 Results (95  % CI 0.314, 0.388), p  <  0.000 while poor agreement A total of 5165 E. coli, 1103 S. aureus and 532 P. aerugi- was noted with amikacin κ  =  0.112 (95  % CI 0.079, nosa non-duplicate isolates were included in the analysis. 0.145), p < 0.000. Of the 5165 E. coli MICs analyzed, there was compara- Of the 1103 S. aureus MICs analyzed, susceptibility ble susceptibility to most antibiotics including ampicillin, to penicillin, oxacillin, levofloxacin, linezolid, trimetho - amoxicillin–clavulanate, cefuroxime, cefotaxime, ceftri- prim–sulfamethoxazole, vancomycin, clindamycin, axone, trimethoprim–sulfamethoxazole, ciprofloxacin, erythromycin and tetracycline was comparable between Table 1 Susceptibilities of E. coli to various antibiotics, concordance and kappa statistics between CLSI 2015 and EUCAST 2015 guidelines Antibiotic CLSI (%); n = 5165 EUCAST (%); n = 5165 Concordance (%) Kappa, κ (95 % CI) S I R S I R Ampicillin 13.8 0.5 85.7 13.8 0 86.2 99.5 0.985 (0.979, 0.991) Amox-Clav 55.8 23.8 20.4 55.8 0 44.2 78.2 0.581 (0.567, 0.595) TMP/SMX 0 0 100 0 0 100 100 1 Nitrofurantoin 84.6 11.9 3.5 96.5 0 3.5 87.6 0.351 (0.314, 0.388) Ciprofloxacin 57.3 0.2 42.5 55.9 1.4 42.7 98.4 0.969 (0.963, 0.975) Cefuroxime 63.7 3.5 32.8 63.7 0 36.3 96.5 0.924 (0.914, 0.934) Gentamicin 78.5 0.2 21.3 78.1 0.4 21.5 99.5 0.979 (0.973, 0.985) Amikacin 99.3 0.3 0.4 90.5 8.8 0.7 91.1 0.112 (0.079, 0.145) Cefotaxime 69.7 0.4 29.9 69.7 0.4 29.9 100 1 Ceftazidime 76.3 0.9 22.8 71.5 4.8 23.7 93.7 0.859 (0.843, 0.85) Ceftriaxone 67.7 0.1 32.2 67.7 0.1 32.2 100 1 Cefepime 80.5 10.0 9.5 72.9 9.8 17.3 84.9 0.600 (0.578, 0.622) Meropenem 99.7 0.1 0.2 99.8 0.1 0.1 99.7 0.724 (0.573, 0.875) Amox-Clav amoxicillin–clavulanate, TMP/SMX trimethoprim–sulfamethoxazole, S susceptible, I intermediate, R resistant Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 4 of 7 the two guidelines. The susceptibilities, concordance and ceftazidime, ciprofloxacin, cefepime, gentamicin, mero - kappa statistics between the two guidelines are shown in penem and piperacillin–tazobactum. The susceptibili - Table 2. ties, concordance and kappa statistics between the two For S. aureus the susceptibilities are generally very guidelines are shown in Table 3. Year by year analysis did similar between the two guidelines. Year by year analy- not show any differences in trends and susceptibility pat - sis did not show any differences in trends and suscepti - terns. The kappa analysis for CLSI 2015 and EUCAST bility patterns. The kappa analysis for CLSI 2015 and 2015 guidelines revealed moderate to almost perfect EUCAST 2015 guidelines revealed perfect agreement for agreement for all the anti-pseudomonal antibiotics. P levofloxacin, linezolid, vancomycin, oxacillin, penicillin values for the kappa statistics for all the antibiotics were and trimethoprim–sulfamethoxazole with a kappa sta- statistically significant at p < 0.000. Overall, the antibiotic tistic of 1 (p  <  0.000). An almost perfect agreement was susceptibility patterns were quite similar regardless of noted with clindamycin κ = 0.904 (95 % CI 0.826, 0.982), whether CLSI or EUCAST 2015 guidelines were used. p < 0.000, erythromycin κ = 0.978 (95 % CI 0.960, 0.996), p  <  0.000 and tetracycline κ  =  0.962 (95  % CI 0.940, Discussion 0.984), p  <  0.000. Moderate agreement was noted with The morbidity and mortality associated with communi - gentamicin κ = 0.537 (95 % CI 0.441, 0.633), p < 0.000. cable diseases including bacterial infections is quite sig- For the 532 P. aeruginosa analyzed there were similar nificant in developing countries [10]. Antibiotics play a susceptibility patterns noted between CLSI 2015 and critical role in treating such infections especially when EUCAST 2015. Antibiotics analyzed included amikacin, instituted in a timely fashion more so when the bacteria Table 2 Susceptibilities of  S.aureus to  various antibiotics and  the concordance and  kappa statistics between  CLSI 2015 and EUCAST 2015 guidelines Antibiotic CLSI (%); n = 1103 EUCAST (%); n = 1103 Concordance (%) Kappa, κ (95 % CI) S I R S I R Penicillin 10.6 0 89.4 10.6 0 89.4 100 1 TMP/SMX 0 0 100 0 0 100 100 1 Clindamycin 98.0 0.3 1.7 98.0 0.1 1.9 99.5 0.904 (0.826, 0.982) Erythromycin 85.6 0.5 13.9 86.0 0 14.0 99.4 0.978 (0.960, 0.996) Gentamicin 96.3 0.5 3.2 91.5 0 8.5 94.6 0.537 (0.441, 0.633) Levofloxacin 90.8 0.3 8.9 90.8 0.3 8.9 100 1 Linezolid 100 0 0 100 0 0 100 1 Oxacillin 93.3 0 6.7 93.3 0 6.7 100 1 Tetracycline 83.2 0.1 16.7 82.4 0.8 16.8 98.9 0.962 (0.940, 0.984) Vancomycin 100 0 0 100 0 0 100 1 TMP/SMX trimethoprim–sulfamethoxazole, S susceptible, I intermediate, R resistant Table 3 Susceptibilities of  P.aeruginosa to  various antibiotics and  the concordance and  kappa statistics between  CLSI 2015 and EUCAST 2015 guidelines CLSI (%); n = 532 EUCAST (%); n = 532 Concordance (%) Kappa, κ (95 % CI) S I R S I R Amikacin 79.5 3.0 17.5 72.2 7.3 20.5 89.7 0.734 (0.703; 0.765) Ceftazidime 70.9 4.7 24.4 70.9 0 29.1 95.3 0.890 (0.870; 0.910) Cefepime 72.6 4.7 22.7 72.6 0 27.4 95.3 0.886 (0.866; 0.906) Gentamicin 72.6 6.2 21.2 72.6 0 27.4 93.8 0.851 (0.829; 0.873) Meropenem 64.1 8.3 27.6 64.1 12.8 23.1 95.5 0.912 (0.895; 0.929) Pip-taz 64.5 8.3 27.3 64.5 0 35.5 91.7 0.830 (0.808; 0.852) Ciprofloxacin 71.8 5.3 22.9 66.2 5.6 28.2 89.1 0.762 (0.736; 0.788) Pip-Taz piperacillin–tazobactum, CI confidence interval, S susceptible, I intermediate, R resistant Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 5 of 7 is susceptible to the antibiotic given. The determina - of isolates labeled as intermediate by EUCAST were all tion of accurate antibiotic susceptibility is therefore an categorized as susceptible by CLSI. EUCAST guidelines important cog in the clinical care of bacterial infections eliminated the intermediate category for some antibiot- especially in organisms that possess acquired resistance ics and this explains the reduced level of agreement for mechanisms and careful consideration should be given some of the antibiotics. For example, CLSI categorized when deciding how to interpret phenotypic susceptibil- 23.8  % of E. coli as having intermediate susceptibility to ity data [11]. In Kenya, many laboratories have adopted amoxicillin-clavulanate but were all categorized as resist- CLSI guidelines as a basis for interpreting their suscep- ant by EUCAST while 11.9 % categorized as intermediate tibility data despite the fact that it costs between US susceptibility to nitrofurantoin by CLSI were categorized $300–$500. These guidelines are updated annually and as susceptible by EUCAST. From a clinical stand point, therefore require laboratories to keep on purchasing reclassifying amoxicillin-clavulanate from intermedi- them at a cost that is prohibitive to most public labo- ate to resistant is unlikely to adversely affect the patient ratories. Failure to stay updated may result in misinter- as it simply removes it from being a therapeutic consid- pretation of susceptibility. For example in 2013, CLSI eration. As for nitrofurantoin, its ability to concentrate abandoned the use of oxacillin disc diffusion in determin - in urine enables it to achieve significant concentrations ing whether a S. aureus isolate is methicillin resistant in and eliminate isolates that may have intermediate sus- favour of cefoxitin because it was more accurate in deter- ceptibility. Therefore, the reclassification of isolates that mining the presence of a mecA mediated mechanism of are intermediate by CLSI to susceptible by EUCAST is resistance [12]. In 2012, it adopted a new disc diffusion unlikely to contribute to adverse outcomes for patients susceptibility cut off for ciprofloxacin of 31 mm up from with urinary tract infections. The EUCAST guidelines 21  mm for non-typhoidal Salmonella (NTS) isolated have slightly more stringent breakpoints for some antibi- from invasive specimens and all Salmonella Typhi iso- otics in an effort to curb the inappropriate use of antibi - lated from both invasive and non-invasive specimens. otics and control the rising rates of antibiotic resistance. The MIC for the same was reduced from 1 to 0.06 μg/mL For S. aureus, all except one of the antibiotics had mod- resulting in many Salmonella spp. previously reported erate to perfect agreement and three quarters of them as susceptible to fall into the intermediate category, and had almost perfect or perfect agreement with kappa many that were intermediate were now categorized as values ranging from 0.862 to 1. The two guidelines per - resistant [13]. In 2013, CLSI recommended that these cut formed equally in the detection of the rate of methicillin offs should apply to all Salmonella spp. including NTS resistant Staphylococcus aureus (MRSA) of 7.2  %. Van- from non-invasive specimens [12, 14]. In the same year comycin susceptible Staphylococcus aureus (VSSA) were new levofloxacin and ofloxacin breakpoints were intro - also detected equally by the two guidelines. EUCAST duced for Salmonella spp. including Salmonella Typhi. eliminated the intermediate category for vancomycin These changes were motivated by an increased risk of in a bid to discourage the reporting of Glycopeptide- treatment failure in patients with decreased ciprofloxa - intermediate Staphylococcus aureus (GISA) due to poor cin susceptible Salmonella spp [15]. In 2012 again, CLSI response even to increased doses of vancomycin [8, 16]. reduced meropenem and piperacillin-tazobactum MIC In view of the fact that no vancomycin resistant Staphy- breakpoints for P. aeruginosa from  ≤4 to  ≤2  ug/mL lococcus aureus (VRSA) has been identified in our set up and  ≤64  ug/mL to  ≤16  ug/mL respectively [13]. These as yet, these changes are unlikely to influence interpre - examples highlight the importance of remaining up-to- tation of breakpoints. The moderate level of agreement date and emphasizes the need for an affordable, up-to- for gentamicin is due to the more stringent breakpoints date and readily available guideline for the interpretation by EUCAST leading to a much higher resistance rate of of antibiotic susceptibility. 8.6 % compared to 3.2 % by CLSI. The difference between In our comparison of CLSI and EUCAST guidelines the two susceptibility cut-offs is two dilutions and this for interpretation of antibiotic susceptibility, for E. coli, may require further harmonization. Gentamicin is rarely most of the antibiotics had moderate to perfect agree- used as monotherapy in treating gram positive bacteria ment between the two guidelines with kappa values and as such this difference in MIC cut off is unlikely to be ranging from 0.581 to 1 and two-thirds of them having clinically significant. almost perfect or perfect agreement with kappa values For P. aeruginosa, five out of seven of the antibiotics ranging from 0.859 to 1. Poor agreement was noted with had almost perfect agreement with kappa values ranging amikacin with EUCAST having a more stringent break- from 0.830 to 0.912 with the remaining two having mod- point for susceptibility of ≤8 mg/L compared to the CLSI erate agreement with kappa values of 0.762 and 0.734. breakpoint of  ≤16  mg/L. The major discrepancy was EUCAST abolished the intermediate category for ceftazi- in the intermediate and susceptible categories as 8.8  % dime, cefepime, gentamicin and piperacillin-tazobactum Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 6 of 7 Food and Drug Administration; GC: General Committee; GISA: Glycopeptide- reclassifying the MICs as resistant. This accounts for intermediate Staphylococcus aureus; IBM: International Business Machines the slightly reduced level of agreement. For amikacin Corporation (Armonk, New York, United States of America); MIC: minimum and ciprofloxacin, with moderate agreement, EUCAST inhibitory concentration; MRSA: methicillin resistant Staphylococcus aureus; NAC: National Antimicrobial Susceptibility Testing Committees; NTS: non- has more stringent breakpoints for susceptibility of  ≤8 typhoidal salmonella; Pip-Taz: piperacillin–tazobactum; PK-PD: pharmacoki- and  ≤0.5  ug/mL compared to CLSI breakpoints of  ≤16 netic–pharmacodynamics; SPSS: statistical package for the social sciences; S. and  ≤1  ug/mL respectively. These accounted for the aureus: Staphylococcus aureus; TMP/SMX: trimethoprim–sulfamethoxazole; US $: United States of America dollar; VRSA: vancomycin resistant Staphylococcus reduced level of agreement and may require further aureus; VSSA: vancomycin susceptible Staphylococcus aureus. harmonization between the two breakpoints. For mero- penem, EUCAST uses a resistant breakpoint of  >8  ug/ Authors’ contributions AK and GO contributed equally in conceiving and designing the study, data mL while CLSI uses  ≥8  ug/mL. In effect this has led to collection, data analysis and drafting the manuscript. ZP and GR participated a slightly higher meropenem resistance of 27.6  % com- in study design, coordination and critical revision of the manuscript. All pared to 23.1 % for CLSI and EUCAST respectively. This authors read and approved the final manuscript. is another aspect that will require harmonization. Our study is limited by the fact that we only compared Acknowledgements the susceptibility for three bacteria whose results may not None. necessarily be generalizable to all clinically relevant gram Competing interests positive and negative bacteria. However, these three bac- The authors declare that they have no competing interests. teria represent a significant proportion of common bac - Received: 22 January 2016 Accepted: 25 March 2016 terial pathogens both in the community and healthcare settings namely Enterobacteriaceae, non-fermenting gram negative bacteria and Staphylococci. The results obtained are also limited to MICs generated by an auto- mated bacterial identification system which is not widely References 1. Marchese A, Esposito S, Barbieri R, Bassetti M, Debbia E. Does the adop- used in developing countries. For P. aeruginosa, Colis- tion of EUCAST susceptibility breakpoints affect the selection of antimi- tin had not yet been included in the gram negative AST crobials to treat acute community-acquired respiratory tract infections? cards being used at the time the study was being con- BMC Infect Dis. 2012;12(1):181. 2. EUCAST. http://www.eucast.org/. Accessed 31 Aug 2015. ducted. Colistin Etest (bioMerieux, Durham, NC, USA) 3. Kahlmeter G. The 2014 Garrod Lecture: EUCAST—are we heading towards was only run for multi-drug resistant isolates of clinical international agreement? J Antimicrob Chemother. 2015;70(9):2427–39. significance but this data was not available for analysis. http://www.jac.oxfordjournals.org/lookup/doi/10.1093/jac/dkv145. 4. Wolfensberger A, Sax H, Weber R, Zbinden R, Kuster SP, Hombach M. In most laboratories in sub-Saharan Africa, disk diffusion Change of antibiotic susceptibility testing guidelines from CLSI to is the preferred mode of antibiotic susceptibility testing. EUCAST: influence on cumulative hospital antibiograms. PLoS One. However, disk diffusion cut offs generally approximate 2013;8(11):1–8. 5. Polsfuss S, Bloemberg GV, Giger J, Meyer V, Hombach M. Comparison of MIC cut offs fairly well and we think a similar compari - European Committee on Antimicrobial Susceptibility Testing (EUCAST ) son based on disc diffusion cut offs would yield similar and CLSI screening parameters for the detection of extended-spectrum results. β-lactamase production in clinical enterobacteriaceae isolates. J Antimi- crob Chemother. 2011;2012(67):159–66. 6. Hombach M, Mouttet B, Bloemberg GV. Consequences of revised CLSI Conclusion and EUCAST guidelines for antibiotic susceptibility patterns of ESBL- and Our study shows acceptable level of agreement between AmpC β-lactamase-producing clinical Enterobacteriaceae isolates. J Antimicrob Chemother. 2013;68:2092–8. EUCAST and CLSI 2015 AST guidelines for E. coli, S. 7. CLSI. M100-S25 performance standards for antimicrobial susceptibility aureus and P. aeruginosa and laboratories with similar testing; Twenty-fifth informational supplement; 2015. antibiotic susceptibility patterns may choose to adopt 8. EUCAST: European Committee on Antimicrobial Susceptibility Testing breakpoint tables for interpretation of MICs and zone diameters. 2015. either guideline without fear of significantly altering http://www.eucast.org/fileadmin/src/media/PDFs/EUC. reported antibiotic susceptibility. With EUCAST guide- 9. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa lines being freely available it should be considered as statistic. Fam Med. 2005;37(5):360–3. 10. GBD_report_2004update_full.pdf. http://www.who.int/healthinfo/ an alternative especially in resource poor settings in global_burden_disease/GBD_report_2004update_full.pdf. Accessed 22 order to maintain up-to-date antibiotic susceptibility Dec 2015. interpretation. 11. Jorgensen JH, Ferraro MJ. Antimicrobial susceptibility testing: a review of general principles and contemporary practices. Clin Infect Dis. 2009;49(11):1749–55. 12. CLSI. M100-S23 performance standards for antimicrobial susceptibility Abbreviations testing; twenty-third informational supplement; 2013. AKUHN: Aga Khan University Hospital, Nairobi; Amox-Clav: amoxicillin–cla- 13. CLSI. M100-S22: performance standards for antimicrobial susceptibil- vulanate; AST: antimicrobial susceptibility testing; CLSI: Clinical Laboratory ity testing; twenty-second informational supplement—M100S22E.pdf. Standards Institute; ECOFFS: epidemiological cut-offs; ESBL: extended http://www.antimicrobianos.com.ar/ATB/wp-content/uploads/2012/11/ spectrum beta lactamase; Etest: epsilometer test; EUCAST: European Com- M100S22E.pdf. Accessed 22 Dec 2015. mittee on Antimicrobial Susceptibility Testing; E. coli: Escherichia coli; FDA: Kassim et al. Ann Clin Microbiol Antimicrob (2016) 15:21 Page 7 of 7 14. Humphries RM, Fang FC, Aarestrup FM, Hindler JA. In vitro susceptibility 16. Uçkay I, Bernard L, Buzzi M, Harbarth S, François P, Huggler E, et al. High testing of fluoroquinolone activity against Salmonella: recent changes to prevalence of isolates with reduced glycopeptide susceptibility in CLSI standards. Clin Infect Dis. 2012;55(8):1107–13. persistent or recurrent bloodstream infections due to methicillin-resistant 15. Crump JA, Kretsinger K, Gay K, Hoekstra RM, Vugia DJ, Hurd S, et al. Clini- Staphylococcus aureus. Antimicrob Agents Chemother. 2012;56(3):1258– cal response and outcome of infection with Salmonella enterica serotype 64. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3294919 typhi with decreased susceptibility to fluoroquinolones: A United States &tool=pmcentrez&rendertype=abstract. FoodNet multicenter retrospective cohort study. Antimicrob Agents Chemother. 2008;52(4):1278–84. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Annals of Clinical Microbiology and AntimicrobialsSpringer Journals

Published: Apr 11, 2016

There are no references for this article.