OBJECTIVE We studied the effects of inflammation, repeated antibiotic doses, and the surgical status of the eye on penetration of vancomycin hydrochloride into the rabbit vitreous cavity after intravenous administration. METHODS We studied three anatomic states (phakic, aphakic, and aphakic, vitrectomy-treated eyes) subdividing each into inflamed and noninflamed groups. Intravenous vancomycin hydrochloride (15 mg/kg of body weight) was administered every 12 hours for 48 hours. Eyes were harvested for the assay of vitreous cavity antibiotic levels at various intervals from one to 49 hours. We determined concentrations and calculated mean values and S.E.M. RESULTS Therapeutic levels were not established in the vitreous cavity at any time period in the two phakic groups. At 25 hours, the inflamed aphakic eyes had concentrations of 5.05 +/- 1.9 micrograms/ml and the control noninflamed aphakic eyes 4.5 +/- 1.23 micrograms/ml; slight increases were found by 49 hours. Concentrations tested in the aphakic, vitrectomy-treated eyes at two, 13, 25, and 49 hours demonstrated progressive increases both in the inflamed eyes (5.4 +/- 2.4 micrograms/ml, 9.64 +/- 4.25 micrograms/ml, 9.2 +/- 3.96 micrograms/ml, 10.34 +/- 4.49 micrograms/ml) and noninflamed eyes (3.52 +/- 2.1 micrograms/ml, 5.4 +/- 1.96 micrograms/ml, 6.8 +/- 2.53 micrograms/ml, 8.7 +/- 5.44 micrograms/ml). CONCLUSIONS Vitreous vancomycin concentrations in aphakic and aphakic, vitrectomy-treated eyes after intravenous administration exceed the minimal inhibitory concentrations for the usual gram-positive pathogens that create endophthalmitis, suggesting a role for intravenous vancomycin in the treatment of bacterial endophthalmitis.