Antimicrobial susceptibility testing of Haemophilus influenzae Improvement of accuracy of the disc diffusion test

Abstract

A national quality control study was performed in 1986 to investigate the standard of performance of the disc diffusion antimicrobial susceptibility testing of Haemophilus influenzae in Sweden. The accuracy of susceptibility interpretations was unacceptably low. A new standardized method for susceptibility testing of H. influenzae was then worked out, and a new method of setting interpretive zone breakpoints was introduced. The susceptibility category of the main population of clinical isolates was determined according to the MIC50 of the strains. The zone histograms of clinical isolates from five reference laboratories were used for the calculation of interpretive breakpoints. For instance, for susceptible strains the mean of the combined zone values from these laboratories +/- 2 S.D. covered the zone range of the susceptible group, and one more S.D. below covered the intermediate/indeterminate group. The new zone breakpoints would place the main population of clinical isolates in the correct susceptibility group, and even make it possible to detect strains with different degrees of reduced susceptibility in the routine test. In a follow-up quality control study in 1988 the interpretive errors for clinical isolates were eliminated for all antibiotics except doxycycline in some laboratories. Laboratory-related zone breakpoints for doxycycline calculated by the single strain regression analysis method led to correct susceptibility interpretations also in these cases.