We studied the antipneumococcal efficacy of cefotaxime and vancomycin alone and a combination of cefotaxime with various dosages of vancomycin in the treatment of prolonged (48 h) experimental fibrin clot infections in rabbits. A clinical pneumococcal strain for which MICs were 2, 0.5 and 0.5 mg/L of penicillin, cefotaxime and vancomycin respectively, was used in this study. Cefotaxime was given iv at a fixed dose of 50 mg/kg and vancomycin iv at 1, 2.5, 5, 10 or 20 mg/kg. Maximal concentrations in clots were (mean +/- S.D.): 2.1 +/- 0.9, 1.1 +/- 0.4, 1.9 +/- 1, 2.3 +/- 1.5, 3.6 +/- 0.4 and 4 +/- 0.3 mg/g, respectively. The mean half-lives of elimination from clots were 2.2 h for cefotaxime and 7 h for vancomycin. We observed the highest bacterial reductions for the highest doses of vancomycin with or without cefotaxime. The combination of intermediate doses of vancomycin with cefotaxime led to higher antibacterial effects than either monotherapy. The low dose of vancomycin gave no significant additional effect compared with cefotaxime alone. The times of regrowth were similar for cefotaxime and cefotaxime-vancomycin 1, and also for vancomycin 10 and vancomycin 20 with or without cefotaxime but were significantly delayed for the combination cefotaxime-vancomycin 2.5 and cefotaxime-vancomycin 5 as compared with vancomycin 2.5 and vancomycin 5. By using a multivariate analysis, we demonstrated that the most important parameters were Cmax (r = 0.43) and AUC (r = 0.58) for cefotaxime alone and Cmax (r = 0.70) for vancomycin alone; none of the tested parameters was found to be significantly correlated with the efficacy of the combinations of cefotaxime and vancomycin. From these findings, and under the experimental conditions used (i.e., relatively low concentrations of cefotaxime), we demonstrated that the in-vivo antibacterial efficacy of the combination of cefotaxime and vancomycin was higher than each monotherapy when the local concentrations of vancomycin were at least 1.9 mg/L.