The SUpplementation en VItamines et Mineraux AntioXydants (SU.VI.MAX) study, a randomised double-blind, primary-prevention trial showed that after 7·5 years, low-dose antioxidant supplementation lowered the total cancer incidence in men, but not in women. To explain this difference in the impact of antioxidant supplementation in men and women, we hypothesised that the effect of supplementation is dependent on initial antioxidant status; 12 741 French adults (7713 females aged 35–60 years; 5028 males aged 45–60 years) received daily antioxidant supplementation (120 mg vitamin C, 30 mg vitamin E, 6 mg β-carotene, 100 μg Se, 20 mg Zn daily) or a matching placebo. Cut-off limits for baseline serum concentrations of the different antioxidant vitamins and minerals were defined as follows for both men and women: 0·3 μmol/l for β-carotene, 11·4 μmol/l for vitamin C, 15 μmol/l for vitamin E, 0·75 μmol/l for Se and 10·7 μmol/l for Zn. The percentage of men with serum concentrations under cut-off limits was higher for vitamins C and E and β-carotene in those who developed a cancer than in those who did not. The risk of cancer was higher in men with baseline concentrations of serum vitamin C or vitamin E under cut-off limits, but not in women. The effect of supplementation was greater in men with baseline serum concentrations of vitamin C, vitamin E and β-carotene below the cut-off limits compared with those above it. This effect was maintained only for vitamin E after adjustment for age, tobacco, and alcohol consumption and BMI. No effect of supplementation could be seen in women. Baseline antioxidant status is related to the risk of cancer in men but not in women and therefore does not entirely explain the differences observed in the effect of antioxidant supplementation on cancer risk between sexes in the SU.VI.MAX study.