Evaluation of mupirocin E-test for determination of isolate susceptibility: comparison with standard agar dilution techniques.
Mupirocin E-test strips have been evaluated for their ease of use and accuracy in determining the susceptibilities of 171 strains of Staphylococcus spp., Streptococcus spp., Haemophilus influenzae, and Moraxella catarrhalis. The susceptibility of each strain was determined on two occasions, using parallel E-test and agar dilution methodologies each time. To ensure similar precisions for statistical analyses, E-test MICs were rounded up to a standard twofold agar dilution scale. Clear, elliptical zones were obtained against Staphylococcus spp. M. catarrhalis also gave clear zones, but the scale intercept was often difficult to interpret because of the irregular shape of the inhibition zone. Poor growth sometimes resulted in less-distinct zones of inhibition against Streptococcus spp. and H. influenzae. Excellent correlation was observed between the the E-test and agar dilution against Staphylococcus spp. and H. influenzae, with > 95% of the E-test values falling within one log2 dilution of the corresponding agar MIC. The correlation was lower for Streptococcus spp. and M. catarrhalis, with 86 and 83%, respectively, of E-test results falling within one log2 dilution of the agar MIC. When E-test MICs did not agree exactly with the corresponding agar MIC against Staphylococcus spp. or Streptococcus spp., there was a tendency for the E-test to give a lower MIC. This bias has little effect upon individual MICs in staphylococci or in the generation of susceptibility interpretation errors ( < 1.5% overall), but it could reduce population geometric mean MICs by factors of 0.78 to 0.83. This effect was more marked for Streptococcus spp., reducing the population mean by a factor of 0.73 and resulting in 0.7% major and 8% very major errors. In contrast, the E-test tended to give higher MICs against M. catarrhalis, resulting in 7.3% major errors and increasing the population geometric mean MIC by a factor of 1.60.