Antimicrobial Susceptibility Patterns at AKUHN

The study was carried out at the Aga Khan University Hospital, Nairobi’s (AKUHN) Department of Pathology. Ethical exemption was granted by the AKUHN’s research and ethics committee (Ref 2015/REC-44) since this type of study is low risk and classified as a clinical audit. This was a retrospective study reviewing the MICs of various antimicrobials on one commonly isolated gram positive organism, Staphylococcus aureus (S. aureus), one fermenting gram negative organism, Escherichia coli (E. coli) and one non-fermenting gram negative organism, Pseudomonas aeruginosa (P. aeruginosa). MIC data for E. coli, P. aeruginosa and S. aureus isolates was collected from two Vitek 2 (version 4.01, bioMerieux, Marcy-l’Etoile, France) automated microbiology systems for the period January 2012 to December 2014. The Vitek 2 AST-P580 and AST-GN26/AST-GN83 cards were used for antibiotic susceptibility for S. aureus and E. coli/P. aeruginosa respectively. Ceftriaxone MICs were only available for the year 2014 in the AST-GN83 cards for a total of 1673 E. coli isolates. The data was summarized using Microsoft Excel 2013 and imported into IBM (International Business Machines, Corporation; Armonk, New York, United States of America) SPSS (Statistical Package for the Social Sciences) Version 22.0 that was used for analysis. The MICs were then analyzed using both the CLSI 2015 and EUCAST 2015 guidelines to categorize them as either susceptible, intermediate or resistant [7 , 8 ]. The concordance rate between the two guidelines in percentage was compared. The susceptibility rates for the various antimicrobial agents were also calculated in percentages for each organism. Analysis of the extent of agreement between CLSI 2015 and EUCAST 2015 for the various antimicrobials was carried out using Cohen’s kappa statistics and graded from perfect agreement to poor agreement [9 (link)]. Cohen’s Kappa statistics determines the proportion of agreement over and above chance between two independent observations. This ranges from −1 to 1 and a p value less than 0.05 means that the agreement reported is significantly different from 0 and is not due to chance. For all inferential statistics, a p value less than 0.05 was considered statistically significant.