Interlaboratory variation of antibiograms of methicillin-resistant and methicillin-susceptible Staphylococcus aureus strains with conventional and commercial testing systems

Abstract

Laboratory-prepared (conventional) and commercial susceptibility testing systems were compared by using a group of methicillin-resistant (MR) and methicillin-susceptible (MS) strains of Staphylococcus aureus. A group of 25 MR and 15 MS S. aureus strains were coded and tested blindly by disk diffusion, agar dilution, broth microdilution, Sensititre, Micro-Media, Sceptor, API 3600S, MicroScan, Autobac I, and MS-2 systems. All systems were incubated at 35 degrees C and read with either a manual or automated reader at the recommended times. Where applicable, systems were also read at 48 h. Among the conventional assays, the broth and agar dilution methods were comparable, both detecting 88% of the MR strains at 24 h and detecting 92 and 96%, respectively, at 48 h. The disk diffusion method was less efficient, detecting only 36 and 72% at 24 and 48 h, respectively. Detection of cephalothin resistance was low for all systems at both time periods, with agar dilution and disk diffusion being the most and least efficient, respectively. Some variability was also seen with detection of resistance to clindamycin and gentamicin. Among the MS strains, variability among the conventional systems occurred with methicillin, gentamicin, ampicillin, and penicillin. Comparison of the commercial systems with manual readers with the broth microdilution method (reference method) showed that for MR strains, the Sceptor system gave identical results at 24 and 48 h. Sensititre detected 68 and 88% of the MR strains, whereas Micro-Media was least effective detecting 12 and 80% at 24 and 48 h, respectively. None of the commercial systems detected cephalothin resistance well, with only one strain being indicated by the Sceptor and Sensititre systems at 48 h. Slight differences were also seen among the systems with clindamycin and gentamicin. With regard to the MS strains, variability among the systems was seen with methicillin, penicillin, ampicillin, clindamycin, and gentamicin. Among commercial systems with automated readers, the API system detected a greater number of MR strains than did the reference method at 24 and 48 h, 96 and 100%, respectively. The MicroScan method was comparable to the reference method detecting 80 and 88% of the MR strains at both time periods, respectively. Both Autobac I and MS-2 were much less effective in detecting MR strains, noting only 32 and 16%, respectively, at the 3- to 6-h readings. Poor detection of cephalothin resistance among MR strains was evident in all systems. Variability also occurred among the systems with clindamycin, gentamicin, and ampicillin. A single strain of the MR group was reported to be vancomycin resistant by the API system. Among the MS group, the greatest variability was seen with methicillin. Less variability occurred with penicillin, ampicillin, gentamicin, and vancomycin.