Using a rat model of aortic valve infective endocarditis, we previously found that oxacillin was equally effective against an oxacillin-susceptible strain of Staphylococcus aureus and a beta-lactamase-hyperproducing borderline oxacillin-susceptible strain of S. aureus; also, ampicillin-sulbactam was less effective than oxacillin against both isolates and at low doses was less effective against the borderline-susceptible strain than against the fully oxacillin-susceptible strain (C. Thauvin-Eliopoulos, L. B. Rice, G. M. Eliopoulos, and R. C. Moellering, Jr., Antimicrob. Agents Chemother. 34:728-732, 1990). In the present study, we extended this work, using alternative treatment schedules and additional bacterial strains. Extending treatment with low doses of ampicillin-sulbactam (500 and 250 mg/kg of body weight per day, respectively) to 6.5 days resulted in equalization of effectiveness against the previously studied strains BOSSA-1 and OSSA-1 (3.75 +/- 1.61 log10 and 4.71 +/- 1.79 log10 CFU of residual viable bacteria per g, respectively). Against the borderline oxacillin-susceptible strain BOSSA-1, increasing the sulbactam dosage from 500 to 2,000 mg/kg/day while maintaining a fixed dose of ampicillin (1,000 mg/kg/day) by continuous infusion resulted in lower bacterial counts (4.93 +/- 1.84 log10 versus 3.65 +/- 1.26 log10 CFU of residual viable bacteria per g, respectively), but this difference was of only borderline significance; differences in efficacy between the low-dose and high-dose sulbactam regimens were exaggerated when intermittent intravenous administration was used (6.19 +/- 1.90 log10 versus 3.37 +/- 1.41 log10 CFU/g, respectively; P < 0.001). However, for any individual sulbactam dosage, the model of administration (continuous versus intermittent infusion) did not affect the activity of the regimen. When additional strains were used in the model, oxacillin and ampicillin-sulbactam (1,000 plus 2,000 mg/kg/day) were equally effective against both oxacillin-susceptible and borderline oxacillin-resistant strains of S. aureus. These results support the predictions that oxacillin would be clinically effective in the treatment of infections caused by borderline oxacillin-susceptible strains of S. aureus and that, except at very low doses, ampicillin-sulbactam would also be as effective against borderline-susceptible strains as against fully oxacillin-susceptible strains of S. aureus.