Information from observational studies and a limited number of clinical trials has raised concerns over the potential adverse effects of long-term selenium supplementation on glucose metabolism and risk of type-2 diabetes in populations with high background selenium status, such as the US. There is limited trial evidence on the effect of selenium supplementation on glucose metabolism in European populations whose selenium status is much lower than that in the US.

We aimed to determine the effect of selenium supplementation at different dose levels on changes in HbA_1c after 6 months and 2 years in a population of relatively low selenium status.

The Denmark PRECISE study was a single-centre, randomized, double-blinded, placebo-controlled, multi-arm, parallel clinical trial with four groups. Four hundred and ninety-one male and female volunteers aged 60–74 years, recruited at Odense University Hospital, Denmark, were randomly assigned to treatment with 100, 200, or 300μg selenium/d as selenium-enriched-yeast or placebo-yeast. HbA_1c measurements were available for 489 participants at baseline (1998–99), 435 at 6 months, and 369 after 2 years of selenium supplementation. Analyses were performed by intention to treat.

The mean (SD) age, plasma selenium concentration, and HbA_1c at baseline were 66.1 (4.1) years, 86.5 (16.3)ng/g and 36.6 (7.0)mmol/mol, respectively. During the initial 6-month intervention period, mean HbA_1c levels (95% CIs) decreased by 1.5 (0.2 to 2.8)mmol/mol for 100μg/d of selenium supplementation and by 0.7 (−0.6 to 2.0)mmol/mol for the 200 and 300μg/d groups compared with placebo (P=0.16 for homogeneity of changes across the four groups). After 2 years of selenium supplementation, HbA_1c decreased significantly in all four treatment groups, with no difference between active treatment groups and placebo. Compared with placebo, mean HbA_1c levels (95% CIs) changed by −0.9 (−2.7 to 0.9)mmol/mol, 0.1 (−1.7 to 2.0)mmol/mol and −1.0 (−2.8 to 0.8)mmol/mol for 100, 200, and 300μg/d of selenium supplementation, respectively (P=0.52 for homogeneity of changes across the four groups).

Selenium supplementation in a European population of elderly individuals of relatively low selenium status did not significantly affect HbA1c levels after two years of supplementation. Our results do not support a beneficial or a harmful effect of selenium supplements on glucose metabolism.