Evaluation of current methods for detection of staphylococci with reduced susceptibility to glycopeptides

The sensitivity and specificity of seven methods (agar dilution, broth micr odilution, Etest at 0.5 and 2.0 McFarland (McF) inocula, two agar screening methods, and population studies [PS]) were evaluated in a double-blind stu dy involving 284 methicillin-resistant Staphylococcus aureus (MRSA) strains and 45 Staphylococcus strains with reduced susceptibilities to vancomycin (SRSV). The results were compared to the population analysis profile-area u nder the curve ratio method (PAP-AUC ratio compared to that of Mu3) as desc ribed by Wootton et al, The agar screening method using brain heart infusio n agar (6 mug of vancomycin per mi) gave a sensitivity of 22% and a specifi city of 97%. A similar method using Mueller-Hinton agar (5 mug of vancomyci n per mi) gave a sensitivity of 20% and a specificity of 99%, The PS method detected 34 false positives (12%) and gave a sensitivity of 71% and a spec ificity of 88%. Etest using 0.5 and 2.0 McF inocula gave sensitivities and specificities of 82 and 93% and of 96 and 97%, respectively. The best Etest interpretative criteria for the 2.0 McF inoculum was greater than or equal to8 mg of vancomycin per liter and greater than or equal to8 mug teicoplan in per mi or greater than or equal to 12 mug of teicoplanin per mi. The dir ect colony suspension inoculum for this method was found to be equally accu rate in detecting (hetero-) glycopeptide-intermediate S. aureus compared to the overnight broth inoculum preparation method. Agar dilution and broth m icrodilution using the NCCLS breakpoint criteria for vancomycin gave sensit ivities and specificities of 20 and 100% and of 11 and 100%, respectively, Using the Etest with a 2.0 McF inoculum, six different media were assessed against a selection of SRSV (n = 48) and MRSA (n = 12). Brain heart infusio n agar yielded tile highest sensitivity and specificity values: 88 and 88%, respectively.