Intermittent administration of ampicillin alone has resulted in high failure rates in previously described animal models of enterococcal endocarditis. We developed a rat model of enterococcal endocarditis which permits comparison of continuous intravenous infusion of ampicillin with intramuscular therapy. Continuous low-dose ampicillin infusion (450 mg/kg [body weight] per day) was compared with the same dose given intramuscularly in three divided doses and with high-dose infusion (4.5 g/kg per day) of the drug. For the infecting strain of Streptococcus faecalis, the MIC and MBC were 1 microgram/ml. Mean ampicillin levels in serum were 53.9 +/- 4.8 (peak) and less than 1 (trough), 8.7 +/- 1.4, and 244 +/- 29 micrograms/ml for intramuscular, low-dose, and high-dose regimens, respectively. Ampicillin infusion therapy significantly increased the survival rate and sterilization of blood cultures. Continuous infusions were superior to intermittent therapy in eradicating bacteremia. After 5 days of treatment, low-dose ampicillin infusion was more effective than intermittent therapy in sterilizing cardiac vegetations (P less than 0.01). Continuous-infusion therapy at either dose was significantly more effective than intramuscular injection in reducing bacterial titers in cardiac vegetations (5.4 +/- 1.0 log10 CFU/g [low dose], 4.8 +/- 0.3 log10 CFU/g [high dose], and 7.7 +/- 0.3 log10 CFU/g [intramuscular]). However, no statistically significant advantage was found for high-dose compared with low-dose ampicillin infusion in lowering bacterial titers in vegetations (P greater than 0.3).