The relative importance of pharmacokinetic and pharmacodynamic parameters for the feasibility of a single daily dose (SDD) of antibiotics remains to be established. Therefore, we studied the relationship between in vitro bacteriological parameters (MIC, MBC, and killing rate [KR], defined as the reduction in the inoculum within 3 h), pharmacokinetic parameters (t1/2 and protein binding [PB], and in vivo antibacterial effect of a single antibiotic dose in an experimental rabbit model of Escherichia coli endocarditis. Nine antibiotics were investigated: two aminoglycosides, two quinolones, and five beta-lactams. For each drug, the minimal effective dose (MED) (in milligrams per kilogram) was defined as the lowest dose able to achieve a significant difference (P less than 0.05) of CFU in the vegetations in comparison with controls 24 h after a single intravenous injection. Aminoglycosides and quinolones had the lowest MEDs, followed by beta-lactams. Univariate regression analysis showed that KR was the major determinant of MED. A stepwise regression analysis showed that t1/2 significantly improved the predictive value of KR, while PB, MIC, and MBC did not. The final equation was MED = 1,586-238 KR-297 t1/2 (r = 0.90, P = 0.01). We concluded that the pharmacodynamic parameters (especially the high KR) of aminoglycosides and quinolones explained their low MEDs and might allow SDD. In contrast, the low KR of beta-lactams emphasized the critical importance of a long t1/2, as for ceftriaxone, allowing the use of this beta-lactam alone in SDD.