Relative efficacies of broad-spectrum cephalosporins for treatment of methicillin-susceptible Staphylococcus aureus experimental infective endocarditis

The effects of treatment with broad-spectrum parenterally administered cephalosporins and cefuroxime, cefazolin, or nafcillin were compared in an experimental model of Staphylococcus aureus infective endocarditis, and the results in vivo were compared with the activities of the study drugs in vitro. After 3 days of treatment, all antimicrobial agents tested were more effective than no treatment in reducing the number of surviving bacteria in cardiac valve vegetations. Nafcillin was the most effective agent studied and was significantly more active than was ceftizoxime, ceftriaxone, cefotaxime, cefoperazone, cefuroxime, or cefazolin (P < or = 0.05). Cefpirome and ceftazidime were the most effective broad-spectrum cephalosporins. The outcome of treatment with cefpirome or ceftazidime was similar to that of treatment with nafcillin and significantly better than that of treatment with ceftizoxime or cefotaxime (P < or = 0.05). Treatment outcome correlated closely with the MICs of the antimicrobial agents for the study strain with the exception of ceftazidime, which was significantly more active in vivo in comparison with other agents than predicted by its MIC (P < or = 0.0003). When ceftazidime was excluded as an outlier, treatment outcome correlated with the MICs of the remaining study drugs (Spearman's correlation coefficient, 0.95; P < or = 0.0004), as well as with the estimated percentage of time during which the concentration of total drug (correlation coefficient, -0.85; P < or = 0.007) or free drug (correlation coefficient, -0.90; P < or = 0.003) exceeded the MIC. A consideration of total or free drug concentrations in relation to MICs did not significantly improve the correlation with outcome observed with the MICs alone.