Use of a predictor panel to evaluate susceptibility testing methods for ampicillin-sulbactam.

A predictor panel of clinical isolates that produce a variety of types and amounts of beta-lactamases was used to assess the accuracies of a variety of susceptibility tests for ampicillin-sulbactam. Combinations of ampicillin-sulbactam in ratios of 1:1 and 2:1 and with sulbactam held constant at concentrations of 4 and 8 micrograms/ml were examined in dilution tests performed in agar and broth. In addition, disks containing 10/10, 20/10, 20/20, and 20/30 micrograms of ampicillin-sulbactam were examined in diffusion tests. The results indicated that the MICs obtained in broth microdilution tests performed with each of the four combinations differed, on average, less than twofold. Of the disks tested, the 20/10-micrograms ampicillin-sulbactam disk provided the best separation between susceptible and resistant strains when interpretive criteria for resistance was a zone size of < or = 16 mm and that for susceptibility was a zone size of > or = 21 mm. This disk also gave the highest overall agreement with MICs, regardless of the combination used in the broth microdilution test. Discrepancies between agar and broth microdilution MICs were greater than twofold, on average, and this necessitated recommendation of separate criteria for the two methods. Thus, a predictor panel was very useful in identifying the parameters of susceptibility tests that were most accurate in identifying strains that were susceptible and resistant to ampicillin-sulbactam.