OBJECTIVE To determine the efficacy and safety of short-contact administration of topical tretinoin on foot ulcers in patients with diabetes. DESIGN Randomized, double-blind, placebo-controlled trial. SETTING Outpatient clinic at a Veterans Affairs medical center. PATIENTS Twenty-four volunteers with diabetic foot ulcers but without evidence of peripheral arterial disease or infection. INTERVENTIONS Patients were randomized to 4 weeks of daily treatment with either topical 0.05% tretinoin solution (Retin-A) or placebo saline solution. Photographs and assessment of wound size and appearance were assessed every 2 weeks for a total of 16 weeks. MAIN OUTCOME MEASURES The proportion of ulcers that healed in each group and the degree of change in ulcer size. RESULTS Twenty-two patients, with a total of 24 foot ulcers, completed the study. At the end of 16 weeks, 2 (18%) of 11 ulcers in the control group and 6 (46%) of 13 ulcers in the tretinoin treatment group healed completely. Topical tretinoin therapy significantly decreased ulcer area and depth compared with placebo treatment over the 16 weeks of the study (P < .01 for surface area; P = .02 for depth). Adverse effects mainly consisted of mild pain at the ulcer site. CONCLUSIONS Short-contact application of topical tretinoin improved the healing of foot ulcers in patients with diabetes. The tretinoin therapy was generally well tolerated, without serious local or systemic adverse effects.