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.

BACKGROUND: The Clinical Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility 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 agreement for all the anti-pseudomonal antibiotics. 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.