The use of intravenous antibiotics as prophylaxis in penetrating eye injuries is strictly empiric and not based on scientific data supporting their use. To determine the efficacy of prophylactic intravenous cefazolin in penetrating eye injuries, a rabbit model of posttraumatic endophthalmitis was developed. Forty rabbits received penetrating eye injuries followed immediately by an intravitreal inoculum of live Staphylococcus epidermidis. The rabbits then were randomly divided into four groups: group 1 received three doses of intravenous cefazolin; group 2 received six doses, and group 3 received nine doses; group 4 received no treatment and served as controls. All control rabbits developed 4+ vitreitis; rabbits receiving three doses of the antibiotic developed 2+ vitreitis, and those receiving six or nine doses of cefazolin showed no vitreous inflammation (P < or = 0.0001). Histologic examination of control eyes showed an exuberant reaction with formation of retrolental membranes, vitreous abscess, and retinal detachment. Eyes treated with nine doses of cefazolin were devoid of inflammatory cells. These findings suggest that intravenous cefazolin is effective in preventing the development of posttraumatic endophthalmitis in a rabbit model.