Serum selenium levels and all-cause, cancer, and cardiovascular mortality among US adults.

BACKGROUND Selenium, an essential trace element involved in defense against oxidative stress, may prevent cancer and cardiovascular disease. We evaluated the association between selenium levels and all-cause and cause-specific mortality in a representative sample of US adults. METHODS Serum selenium levels were measured in 13,887 adult participants in the Third National Health and Nutrition Examination Survey. Study participants were recruited from 1988 to 1994 and followed up for mortality for up to 12 years. RESULTS The mean serum selenium level was 125.6 ng/mL. The multivariate adjusted hazard ratios comparing the highest (> or = 130.39 ng/mL) with the lowest (< 117.31 ng/mL) serum selenium level tertile were 0.83 (95% confidence interval [CI], 0.72-0.96) for all-cause mortality, 0.69 (95% CI, 0.53-0.90) for cancer mortality, and 0.94 (95% CI, 0.77-1.16) for cardiovascular mortality. However, based on spline regression models, the association between serum selenium levels and all-cause and cancer mortality was nonlinear, with an inverse association at low selenium levels (< 130 ng/mL) and a modest increase in mortality at high selenium levels (> 150 ng/mL). There was no association between serum selenium levels and cardiovascular mortality. CONCLUSIONS In a representative sample of the US population, we found a nonlinear association between serum selenium levels and all-cause and cancer mortality. Increasing serum selenium levels were associated with decreased mortality up to 130 ng/mL. Our study, however, raises the concern that higher serum selenium levels may be associated with increased mortality.