Vitamin E and β-carotene affect the immune function and might influence the predisposition of man to infections. To examine whether vitamin E or β-carotene supplementation affects tuberculosis risk, we analysed data of the Alpha-Tocopherol Beta-Carotene Cancer Prevention (ATBC) Study, a randomised controlled trial which examined the effects of vitamin E (50 mg/d) and β-carotene (20 mg/d) on lung cancer. The trial was conducted in the general community in Finland in 1985–93; the intervention lasted for 6·1 years (median). The ATBC Study cohort consists of 29 023 males aged 50–69 years, smoking at baseline, with no tuberculosis diagnosis prior to randomisation. Vitamin E supplementation had no overall effect on the incidence of tuberculosis (risk ratio (RR) = 1·18; 95 % CI 0·87, 1·59) nor had β-carotene (RR = 1·07; 95 % CI 0·80, 1·45). Nevertheless, dietary vitamin C intake significantly modified the vitamin E effect. Among participants who obtained 90 mg/d or more of vitamin C in foods (n 13 502), vitamin E supplementation increased tuberculosis risk by 72 (95 % CI 4, 185)%. This effect was restricted to participants who smoked heavily. Finally, in participants not supplemented with vitamin E, dietary vitamin C had a negative association with tuberculosis risk so that the adjusted risk was 60 (95 % CI 16, 81) % lower in the highest intake quartile compared with the lowest. Our finding that vitamin E seemed to transiently increase the risk of tuberculosis in those who smoked heavily and had high dietary vitamin C intake should increase caution towards vitamin E supplementation for improving the immune system.