Selenium supplements don't prevent type-2 diabetes, and may increase risk

Secondary analysis of an existing study has found that taking selenium supplements does not appear to prevent type-2 diabetes, and may even increase risk for the disease. There is epidemiological and experimental evidence to suggest that dietary intake of antioxidants, including selenium, may protect against type-2 diabetes, however there have been no long-term randomised controlled trials on this.

The authors of this paper used data from an existing long-term controlled trial of selenium supplementation on cancer incidence to investigate whether any effect on diabetes risk could be shown. The study was an investigation into whether such supplements reduced the risk of non-melanoma skin cancer in people living in areas of the US with low selenium consumption. Participants were recruited from dermatology clinics between 1983 and 1991, and were randomised to receive identical tablets containing a high-selenium yeast (equivalent to 200microgm daily) or an ordinary yeast control. For this analysis, the authors assessed the incidence of type-2 diabetes throughout the blinded phase of the trial (1983-1996) in participants who did not have it at baseline. Diabetes was assessed by self-report, corroborated by medical records.

The study involved 1,312 participants, of whom 1,202 did not have type-2 diabetes at baseline: there were no statistically significant baseline differences between the selenium and control groups. Average follow-up was 7.7 years, and over this time 97 new cases of type-2 diabetes were identified. This gives a rate similar to that found in other studies of largely white populations, however there was a statistically significant imbalance between the selenium and control groups: of the 97 cases, 58 developed in the selenium group and 39 in the placebo group. These give incidences of 12.6 vs. 8.4 cases per 1000 patient-years respectively and a hazard ration of 1.55 (95% CI 1.03 to 2.33).

The authors conclude that selenium supplementation in this group of people from low-selenium areas of the US did not appear to reduce risk for type-2 diabetes. In contrast, those in the supplement group had a statistically significant increase in risk that persisted across subgroups. When risk was analysed by baseline selenium level, the risk increased with greater baseline levels. They caution that their study has limitations: for example, incidence of diabetes was a secondary outcome of the study, it was self-reported, and the numbers were relatively small so a few more cases in the placebo group would attenuate the association. Nevertheless, it was a placebo-controlled study with good adherence to treatment.

An accompanying editorial discusses the study and its potential implications. The authors note that selenium-containing supplements are widely used in the US, and that while it is an essential micronutrient, it has a narrow therapeutic range. Further trials are need to investigate this possible link; meanwhile, those whose diet provides adequate daily supplies should not take supplements except as part of a controlled trial.

Ann Intern Med 2007; 147: 217-23 (published online in advance of print, link to abstract); Ann Intern Med 2007; 147(4) (Editorial, published online in advance of print; link to full text, may be restricted to subscribers)