Serum selenium has been implicated as a risk factor for lung cancer, but the issue remains unsettled. In a cohort of 3,333 males aged 53-74 yrs, we tested the hypothesis that a low serum selenium concentration would be associated with an increased risk of lung cancer mortality. Over 16 yrs, 167 (5.1%) subjects died of lung cancer: 48 (5.0%) out of 965 with low serum selenium (0.4-1.0 μmol · L(-1)), 57 (5.1%) out of 1,141 with medium serum selenium (1.1-1.2 μmol · L(-1)) and 62 (5.1%) out of 1,227 with high serum selenium (1.3-3.0 μmol · L(-1)). After adjustment for age, referencing the lowest level of serum selenium, hazard ratios (HRs) for medium and high levels of serum selenium were 0.97 (95% CI 0.66-1.43) and 0.99 (95% CI 0.68-1.45), respectively. Taking into account pack-years of smoking, spirits intake, dietary markers (salt and fat preferences) and health measures (chronic bronchitis and peak flow), referencing the lowest level of serum selenium, HRs were 1.17 (95% CI 0.79-1.75) and 1.43 (95% CI 0.96-2.14), for medium and high levels respectively. Among heavy smokers, a high serum selenium concentration was associated with a significantly increased risk of lung cancer mortality after taking into account all potential confounders. The hypothesis that low serum selenium is an independent risk factor for lung cancer was not supported.