Apparent differences in antimicrobial susceptibility as a consequence of national guidelines

Abstract

There is considerable variation in the breakpoint recommendations given by various authorities for classification of antimicrobial susceptibility results. At least six authorities provide guidelines for susceptibility testing and breakpoint criteria in Europe: the British Society for Antimicrobial Chemotherapy (BSAC) [1Supplementary report by the Working Party on antibiotic sensitivity testing of the British Society for Antimicrobial Chemotherapy.J Antimicrob Chemother. 1996; 38: 1103-1105Crossref Scopus (27) Google Scholar], la Société Frańaise de Microbiologie (SFM) [2Clin Microbiol Infect. 1996; 2: 1-48Abstract Full Text Full Text PDF PubMed Google Scholar], Deutsche Industrie Norm - Medizinische Mikrobiologie (DIN) [3Deutsches Institut für Normung Methods for the determination of susceptibility of pathogens (except mycobacteria) to antimicrobial agents. MIC breakpoints of antibacterial agents. DIN, Berlin1998: 58940-58944Google Scholar], Commissie Richtlijnen Gevoeligheidsbepalingen (CRG) in The Netherlands [4Commissie Richtlijnen Gevoeligheidsbepalingen Standaardisatie van gevoeligheids‐bepalingen.Ned Tijdschr Med Microbiol. 1996; 4: 104-106Google Scholar], the Swedish Reference Group for Antibiotics (SRGA) [5Scand J Infect Dis. 1997; : 1-31Google Scholar] and the Norwegian Working Group on Antibiotics (NWGA) [6Bergan T Bruun J Digranes A Lingaas E Melby K Sander J Susceptibility testing of bacteria and fungi. Report from ‘the Norwegian Working Group on Antibiotics’.Scand J Infect Dis Suppl. 1997; 103: 1-36PubMed Google Scholar]. In addition, guidelines issued by the National Committee for Clinical Laboratory Standards (NCCLS) [7National Committee for Clinical Laboratory Standards Performance standards for antimicrobial susceptibility testing. M100–S9. NCCLS, Wayne PA1999Google Scholar] are used in many European countries. In a recent study, we investigated the antibiotic resistance rates of Escherichia coli, Staphylococcus aureus and coagulase-negative staphylococci (CNS) causing bloodstream infections in two Norwegian hospitals [8Leegaard TM Vik E Caugant DA Frøholm LO Høiby Ea Low occurrence of antibiotic resistance in Escherichia coli and staphylococci isolated from blood cultures in two Norwegian hospitals in 1991–92 and 1995–96.APMIS. 1999; 107: 1060-1068Crossref PubMed Scopus (15) Google Scholar]. In that study, we defined resistance according to the minimum inhibitory concentration (MIC) breakpoints of the NCCLS. We then compared the results with those obtained when resistance was defined according to the MIC breakpoints recommended by the NWGA for the same isolates. The percentages of susceptible, intermediate and resistant isolates were very different for certain antibiotics. The breakpoints defined by the various authorities for non-fastidious organisms are presented in Table 1. To illustrate the consequences of breakpoint variations, the percentages of susceptible and resistant isolates from our study of Norwegian bloodstream infections are shown in Table 2 for selected antimicrobials.Table 1Minimum inhibitory concentration breakpoints (mg/L) for susceptible, intermediate and resistant non-fastidious organisms (Enterobacteriaceae and staphylococci) according to the Norwegian Working Group on Antibiotics (NWGA), the National Committee for Clinical Laboratory Standards (NCCLS), the British Society for Antimicrobial Chemotherapy (BSAC), la Société Française de Microbiologie (SFM), Deutsche Industrie Norm-Medizinische Mikrobiologie (DIN), Commissie Richtlijnen Gevoeligheidsbepalingen (CRG) and the Swedish Reference Group for Antibiotics (SRGA)AntibioticNWGANCCLSBSACSFMDINCRGSRGAcPharmacologic MIC breakpointsS (≤)R (≥)S (≤)R (≥)S (≤)R (>)S (≤)R (>)S (≤)R (≥)S (≤)R (>)S (≤)R (≥)Benzylpenicillin1bfor non-β-lactamase-producing isolates32bfor non-β-lactamase-producing isolates0.1β-lactamase0.120.25160.125cfor staphylococci0.25cfor staphylococci0.250.2518Amoxycillin or ampicillin2328321cfor staphylococci/8dfor Enterobacteriaceae4dfor Enterobacteriaceae16dfor Enterobacteriaceae216216116Amoxycillin/clavulanic acidNDND8/416/884162162 + 0.516 + 4NDNDMecillinam1efor urinary tract infections only16efor urinary tract infections onlyNDNDND2efor urinary tract infections only8efor urinary tract infections onlyNDNDNDND116Oxacillin or methicillin142/0.25ffor S. aureus and CNS, respectively4/0.5ffor S. aureus and CNS, respectively4gfor methicillin22124gfor methicillin4gfor methicillin12Cephalothin216832ND832NDND416116Cefuroxime2168hfor intravenous administration32hfor intravenous administration416dfor Enterobacteriaceae8324164164hfor intravenous administration16Cefotaxime21686418cfor staphylococci432216416416Ceftriaxone21686418432432416416Aztreonam2328328dfor Enterobacteriaceae43223248416Imipenem4164164482828416Trimethoprim288efor urinary tract infections only16efor urinary tract infections only0.524efor urinary tract infections only8efor urinary tract infections onlyNDND1228Trimethoprim/sulfamethoxazole2322/384/76ND2/388/15216i1 part trimethoprim and 19 parts sulfamethoxazole128i1 part trimethoprim and 19 parts sulfamethoxazole1i1 part trimethoprim and 19 parts sulfamethoxazole2i1 part trimethoprim and 19 parts sulfamethoxazole3264Erythromycin140.580.5cfor staphylococci1418120.51cfor staphylococciClindamycin140.540.5cfor staphylococci22181428Gentamicin284161448181448Ciprofloxacin0.254141412141218Doxycycline144161jfor tetracycline.48181jfor tetracycline.4jfor tetracycline.14Chloramphenicol288322cfor staphylococci/8dfor Enterobacteriaceae881681648816TeicoplaninNDND8324416NDND2448Vancomycin484324cfor staphylococci4164164848Fusidic acid< 11NDND1cfor staphylococci216NDND1124Rifampicin12141cfor staphylococci416NDND1112ND, not defined.a Pharmacologic MIC breakpointsb for non-β-lactamase-producing isolatesc for staphylococcid for Enterobacteriaceaee for urinary tract infections onlyf for S. aureus and CNS, respectivelyg for methicillinh for intravenous administrationi 1 part trimethoprim and 19 parts sulfamethoxazolej for tetracycline. Open table in a new tab Table 2Percentages of susceptible and resistant isolates for selected antimicrobials classified according to the various guidelines. A Norwegian blood culture material of E. coli (198 isolates), S. aureus (67 isolates) and coagulase-negative staphylococci (65 isolates) was used in the categorizationSpecies/antibioticNWGANCCLSBSACSFMDINCRGSRGASRSRSRSRSRSRSRE. coli Amoxycillin821712172283621828821528 Amoxycillin/clavulanic acidNDND921928571108101NDND Mecillinam7913NDNDNDND8213NDNDNDND7913 Cefuroxime159903653901659653659 Trimethoprim7721792148237921NDND72237721 Trimethoprim–sulfamethoxazole82178317NDND821783082178317aDefined by us as complete resistance in the E test (range 0.002–32 mg/L). Gentamicin971991851991851851991 Doxycycline96074209917420926926960 Chloramphenicol3748698614869861426148614S. aureus Benzylpenicillin3958277334663660346636643960 Oxacillin8539731000973851510008515 Cefuroxime750100010001000100010001000 Erythromycin99190190109919919919010 Gentamicin850100075010007507501000 Ciprofloxacin406936936936936936931 Chloramphenicol09169003169069319316921 Vancomycin1000100010001000100010001000 Fusidic acid5149NDND7822817NDND78228119CNS Benzylpenicillin3248148618821848188218823256 Oxacillin5537386266346337554566345545 Cefuroxime5440603158316025584058315840 Trimethoprim–sulfamethoxazole44454555NDND444557045555743a Ciprofloxacin5417811766346634811781178117 Doxycycline3844643158316431393539353844 Chloramphenicol26669312316931693134316931 Vancomycin982980982980980980982 Fusidic acid3565NDND45554431NDND45554555ND, not defined. See Table 1 for acronyms.a Defined by us as complete resistance in the E test (range 0.002–32 mg/L). Open table in a new tab ND, not defined. ND, not defined. See Table 1 for acronyms. A few examples illustrate the differences. For cefuroxime resistance in E. coli, many isolates are classified as intermediately susceptible (76%) and resistant (9%) when applying NWGA breakpoints, but most of these are susceptible (90%) when applying NCCLS breakpoints. The percentages of resistance using other guidelines are somewhere in between. The susceptibility of S. aureus to oxacillin also varies considerably. Using the British and Dutch breakpoints, 100% of our isolates are classified as susceptible, but applying SRGA breakpoints, 15% of the isolates are classified as resistant. For the CNS, the BSAC breakpoints classify 34% of the isolates as resistant to methicillin, while the NCCLS classify 62% of the same isolates as resistant to oxacillin. For the following antimicrobial agents the differences were minor or absent (mostly because all isolates were susceptible): E. coli—cefotaxime, ceftriaxone, aztreonam and ciprofloxacin; S. aureus—cephalothin, imipenem, trimethoprim-sulfamethoxazole, clindamycin, doxycycline, teicoplanin and rifampicin; CNS—cephalothin, imipenem, erythromycin, clindamycin, gentamicin, teicoplanin and rifampicin. Susceptibility testing is one of the most important tasks performed in a medical microbiological laboratory, in addition to identifying the etiologic agent of an infection. It is therefore essential for susceptibility testing that the breakpoints issued by various national committees are reliable. Breakpoint recommendations consist of two concentrations. One concentration divides the susceptible from the intermediately susceptible bacteria and the other concentration divides the intermediately susceptible from the resistant bacteria (SIR system). The intention has been to: ‘predict the response of the patient treated with the antibacterial drug’ by applying these breakpoints [9British Society for Antimicrobial Chemotherapy A guide to sensitivity testing. Report of the Working Party on Antibiotic Sensitivity Testing of the British Society for Antimicrobial Chemotherapy.J Antimicrob Chemother. 1991; : 1-50Google Scholar]. When determining MIC breakpoints, many variables have to be taken into consideration [10European Society of Clinical Microbiology and Infectious Diseases (ESCMID), European Committee for Antimicrobial Susceptibility Testing (EUCAST) Determination of antimicrobial susceptibility test breakpoints (EUCAST discussion document).Clin Microbiol Infect. 1999; 51: 1-3Google Scholar]. In addition to the distribution of the in vitro antibiotic susceptibility of relevant bacteria, pharmacokinetic properties of the drugs are traditionally important. More recently, clinical efficacy has been the focus of attention. New insights into the pharmacodynamics of antibiotics may also be of importance. In addition, national therapeutic traditions and the level of resistance prevailing in a country or area may be relevant. Our results illustrate the consequences of the different guidelines and MIC breakpoints proposed by the various national authorities. As surveillance networks for collecting susceptibility data at an international level are being established, the differences between guidelines have become evident. While susceptibility data were mainly used locally or nationally, these differences were less obvious. Historically, efforts have been made to set up internationally recognized guidelines, although most of this work has been devoted to the standardization of the disk diffusion test [11Bauer A Kirby W Sherris J Turck M Antibiotic susceptibility testing by a standardized single disk method.Am J Clin Pathol. 1966; 45: 493-496Crossref PubMed Scopus (11769) Google Scholar, 12Ericsson H Sherris J Antibiotic sensitivity testing. Report of an international collaborative study.Acta Pathol Microbiol Scand [B] Suppl. 1971; : 217Google Scholar]. Standardization of the test procedures and the interpretive standards has not been achieved because of the constant addition of new antibiotics and increased knowledge about resistance factors and new or modified pathogens. In addition, the antibiotics licensed and the approved indications for their use in different countries vary. The NCCLS seems to emphasize pharmacokinetics and, more recently, clinical success as the most important criteria for determining breakpoints, as they have wider ranges when defining their breakpoint values, resulting in fewer isolates being classified as resistant. The two Scandinavian countries represented here emphasize the importance of recognizing increasing resistance early, and set their breakpoints close to the native (often susceptible) bacterial population in order to discover emerging resistance early and avoid reporting false susceptibility [5Scand J Infect Dis. 1997; : 1-31Google Scholar]. The advantage of emphasizing clinical success is that clinicians are given guidance as to whether therapy is likely to be successful or not. Classifying fewer isolates as resistant could be an advantage when trying to combat antibiotic resistance. This could prevent clinicians choosing drugs that one would wish them not to use because they select for resistance as well as, on occasion, having an unnecessarily broad spectrum. However, reporting developing resistance early may lead to the earlier application of counter-measures. It would seem to be in the interest of the antibiotic producers that the sensitive category has a wide range, as this results in fewer isolates being classified as resistant. Walker and Thornsberry suggested that authors should report MIC distributions only, instead of sensitivity groups [13Walker R Thornsberry C Decrease in antibiotic susceptibility or increase in resistance?.J Antimicrob Chemother. 1998; 41: 1-4Crossref PubMed Scopus (32) Google Scholar]. By selecting a system of reporting distributions only, the problem of standardization of interpretation is avoided, but this approach is laborious and probably less useful to clinicians who have come to rely upon the SIR system. Another possible solution is to change to species-specific breakpoints, as this would lead to a better correlation between the observed MIC distribution and the resistance problems involved. This has been suggested by the SRGA, and the system is presently being implemented in Sweden [5Scand J Infect Dis. 1997; : 1-31Google Scholar]. In other countries, species-specific breakpoints are used for certain bacteria, such as gonococci and pneumococci. Still, this system involves grouping of bacteria into categories, and comparison of resistance development between areas implementing different guidelines will remain difficult unless these are standardized. Unless species-specific breakpoints are decided in international collaboration studies to make the results obtained internationally valid, they may complicate the situation further. That breakpoints vary considerably between the different guidelines seems to be the rule and not the exception. Even the methods of giving breakpoints used by the different parties differ to a degree that it makes comparison problematic. To agree upon one single set of guidelines seems difficult because of local differences in antibiotic legislation and a number of uncertainties regarding drug dose, drug concentration, MIC and effects of antibiotics. Nonetheless, international standardization is most certainly needed and should be a most important issue for international collaboration between medical microbiologists.