Study Objective. To determine if prophylactic use of cefazolin reduces peristomal wound infection associated with percutaneous endoscopic gastrostomy. Design. Prospective, randomized, double-blind, placebo-controlled clinical trial. Setting. Academic medical center, referral-based, gastroenterology service. Patients. One hundred thirty hospitalized patients, 23 of whom were excluded. Of the remaining 107 patients, 52 (group I) were already using antibiotics at the time of randomization for gastrostomy, whereas 55 (group II) were not. Interventions. Patients received either intravenous saline as a placebo or intravenous cefazolin (1 g) 30 minutes before gastrostomy. Measurements and Main Results. For 1 week after gastrostomy, the peristomal area was evaluated and a score assigned each day for erythema (0 to 4), induration (0 to 3), and exudate (0 to 4). A maximum combined score of 8 or more or the development of pus was a criterion for infection. None of the patients in group I developed a wound infection. Only 2 of 27 group II patients given prophylaxis developed a wound infection, compared with 9 of 28 patients not given prophylaxis, a difference of 25% (95% confidence interval, 4.8 to 44.6%; p less than 0.025). The number of patients who developed a wound infection was 0 of 52 in group I and 2 of 27 in group II patients who received cefazolin, a difference of 7.4% (95% confidence interval, -2.5 to 17.3%; p = 0.07). Conclusion. Cefazolin prophylaxis significantly reduces the risk for peristomal wound infection associated with percutaneous endoscopic gastrostomy. It is needed, however, only for patients not already receiving antibiotic treatment at the time of gastrostomy.