beta-carotene has been hypothesised to reduce lung cancer risk. We studied the effect of 14 weeks of beta-carotene supplementation (20 mg d-1) on the frequency of micronuclei in sputum in 114 heavy smokers in a double-blind trial. Micronuclei reflect DNA damage in exfoliated cells and may thus provide a marker of early-stage carcinogenesis. Pre-treatment blood levels of cotinine, beta-carotene, retinol and vitamins C and E were similar in the placebo group (n = 61) and the treatment group (n = 53). Plasma beta-carotene levels increased 13-fold in the treatment group during intervention. Initial micronuclei counts (per 3,000 cells) were higher in the treatment group than in the placebo group (5.0 vs 4.0, P < 0.05). During intervention, the treatment group showed a 47% decrease, whereas the placebo group showed a non-significant decrease (16%). After adjustment for the initial levels, the treatment group had 27% lower micronuclei counts than the placebo group at the end of the trial (95% CI: 9-41%). These results indicate that beta-carotene may reduce lung cancer risk in man by preventing DNA damage in early-stage carcinogenesis.