We conducted a double-blind placebo-controlled trial to evaluate the chemopreventive potential of either vitamin A alone or beta carotene alone in subjects with oral leukoplakia in Kerala, India. We randomised 160 fishermen and women with oral precancerous lesions to receive oral vitamin A (retinyl acetate 300,000 IU/week x 12 months, n = 50), or beta carotene (360 mg/week x 12 months, n = 55), or placebo (n = 55). Blood, saliva and urine samples were collected at baseline and at exit to study serum micronutrients and mutagenicity assays. Biopsies of the mucosal lesions at entry were performed for histopathological exclusion of malignancy. The subjects were examined once every 2 months to establish clinical response of lesions and toxicity, if any. The results are based on 43 complaint subjects on placebo, 42 on vitamin A and 46 on beta carotene. The complete regression rates were: 10% in the placebo arm, 52% with vitamin A and 33% with beta carotene (P < 0.0001). Homogeneous leukoplakias and smaller lesions responded better than non-homogeneous and larger lesions. No major toxicities were observed. Half of the responders with beta carotene and two thirds with vitamin A relapsed after stopping supplementation. Serum beta carotene concentration increased substantially with beta carotene administration while with vitamin A supplementation there was no change in serum retinol levels. In the vitamin A treated group there was a significant decrease in serum alpha tocopherol. Vitamin A administration resulted in a significant remission of oral leukoplakia without any side effects of prolonged vitamin A supplementation. The results of this study, as well as those from previous studies, appear to provide strong supporting evidence to justify long term trials with vitamin A in subjects with high-risk leukoplakias with oral cancer as an endpoint.