Methicillin-resistant Staphylococcus aureus vancomycin susceptibility testing: methodology correlations, temporal trends and clonal patterns.

OBJECTIVES To determine the correlation between various vancomycin MIC testing methodologies and explore the phenomenon of MIC creep. METHODS A total of 417 consecutive non-duplicate methicillin-resistant Staphylococcus aureus (MRSA) bloodstream isolates from Liverpool Hospital between 1997 and 2008 were retrieved. All isolates were classified using PFGE and underwent susceptibility testing for vancomycin using a standard Etest (AB bioMérieux, Solna, Sweden), Vitek2(®) (AST-P612; bioMérieux, Inc., Durham, NC, USA) and broth microdilution (BMD) performed as per the CLSI method. RESULTS Over the 12 years, 78% (n = 326) of the isolates were multiresistant MRSA (ST239-like by PFGE, where ST stands for sequence type). The correlation between MIC testing methods was moderate with Spearman's correlation coefficients of 0.50 for BMD versus Etest (P < 0.001), 0.33 for BMD versus Vitek2(®) (P < 0.001) and 0.42 for Etest versus Vitek2(®) (P < 0.001). In general, Etest results were 1 dilution higher while the Vitek2(®) results were 1 dilution lower than the BMD MIC result. MIC creep was dependent on the MIC testing method and the measurement used for analysis (geometric mean MIC versus modal MIC versus frequency analysis), with creep detected for Etest regression analysis only. In contrast, the proportion of isolates with a BMD MIC ≥2 mg/L decreased from 16% to 9% in the latter half of the study. Modal MIC was stable over the 12 years at 1 mg/L irrespective of MIC method used. CONCLUSIONS Correlation between vancomycin MIC methodologies remains suboptimal. Temporal MIC trends should be interpreted with caution as these are dependent on the testing methodology and the measurement used for analysis.