Prediction of piperacillin-tazobactam susceptibility among Enterobacteriaceae, Pseudomonas aeruginosa, and other bacteria using ticarcillin-clavulanic acid, ceftazidime, and other broad-spectrum antimicrobial in vitro test results.

The ability of various in vitro beta-lactam susceptibility test results to predict the susceptibility of piperacillin-tazobactam (a new beta-lactam-beta-lactamase inhibitor combination) was assessed using more than 46,000 recent clinical isolates. The organisms were tested by reference-quality National Committee for Clinical Laboratory Standards (NCCLS) broth microdilution procedures and interpreted by the currently published NCCLS criteria. The recommended antimicrobial tests that would accurately predict the piperacillin-tazobactam in vitro efficacy had an overall very major, false-susceptible rate of only 0.6% (< or = 1.5% is acceptable). The following drug tests can be used to judge piperacillin-tazobactam activity and spectrum (low patient risk) conservatively: for Enterobacteriaceae use ticarcillin-clavulanic acid results (0.6% very major error); for Pseudomonas aeruginosa use piperacillin (0.1%) results; for enterococci use ampicillin or ampicillin-sulbactam (1.8%) results; for Haemophilus influenzae and Moraxella catarrhalis use cefotaxime or cefuroxime or ceftriaxone (1.5%); and for staphylococci use oxacillin by NCCLS recommendations. When the piperacillin-tazobactam testing reagents become available, the direct testing of this combination should be applied to relevant clinical isolates. The piperacillin-tazobactam break points should be reassessed as indicated by the cited minimum inhibitory concentration population analysis to improve predictive accuracy; H. influenzae susceptibility modified to < or = 2/4 micrograms/ml and Enterococcus species susceptibility tested at < or = 16/4 micrograms.