Ampicillin susceptibility testing of Haemophilus influenzae in the routine clinical laboratory by the EUCAST methodology compared to broth microdilution and the presence of ftsI gene mutations.

ABSTRACT

OBJECTIVES: To investigate the prevalence of ampicillin resistance in Haemophilus influenzae and the diagnostic accuracy of the EUCAST recommended disc diffusion method to detect the increasingly prevalent ampicillin resistance due to the presence of PBP3 alterations based on mutations in the ftsI gene. METHODS: During a 6-month period all consecutive non-duplicate H. influenzae isolates were prospectively collected and stored. MICs of ampicillin were determined by broth microdilution (BMD). PCR was performed to detect mutations in the ftsI gene. Results of routine disc diffusion susceptibility testing, including the penicillin screening test in accordance with the current EUCAST methodology, as well as additional Etest results, were compared to the BMD as the reference method. RESULTS: In 102 isolates, the prevalence of ampicillin resistance was 28% (29/102) by BMD. There was a good correlation between MICs of ampicillin and the presence of a ß-lactamase and/or an ftsI gene mutation. The prevalence of ampicillin resistance was overestimated using the EUCAST method (33% (34/102)) and underestimated when an additional Etest was used (24% (24/102)) (not significant). The sensitivity and specificity of the EUCAST methodology for the detection of ampicillin resistance were 97% ((28/29); 95% CI, 82-100%) and 92% ((67/73); 95% CI, 83-97%), respectively. CONCLUSIONS: The prevalence of ampicillin resistance was 28%, as determined by BMD. Although the overall diagnostic accuracy of the EUCAST ampicillin disc diffusion was high, misclassification of ampicillin susceptibility may still occur.