Unlike some animal, ecologic, case-control, and prospective studies that have associated higher selenium levels with reduced cancer risk, a large cohort study recently reported in the Journal of the National Cancer Institute found that selenium intake (as reflected by toenail selenium levels) within the range typically consumed by most women in the United States was not associated with a reduced risk of several common cancers.

Miriam Garland, SCD, Harvard School of Public Health, led the research team that analyzed toenail clippings (which may reflect dietary selenium intake over the previous year or longer) obtained from 62,641 women with no cancer history. All of the women were participants in the Nurses' Health Study, and were aged 30 to 55 years at the study's inception in 1976.

The toenail clippings were requested from the cancer-free women in 1982, and during 41 months of follow-up, 503 cases of cancer other than breast cancer or nonmelanoma skin cancer were diagnosed. For each cancer patient, a control subject was chosen from among the women who remained free of diagnosed cancer, matched by age and by date of nail return. In addition, toenail selenium data from a previous study of 434 breast cancer patients and their controls were included in some of the analyses.

For cancers at all sites combined (excluding breast cancer), case and control subjects had similar mean toenail selenium levels, and a nonsignificant positive association between selenium levels and cancer risk (age- and smoking-adjusted) was found. Potentially confounding factors were considered (eg, state of residence, weight relative to height, aspirin use, intake of various dietary constituents, reproductive variables, and variables related to hormone use and melanoma risk), but none materially altered the results. In a combined analysis, including the breast cancer study data, the results were more weakly positive.

In the analyses by cancer site, toenail selenium level was not associated with reduced cancer risk at any major site, including uterine cancer, colorectal cancer, melanoma, ovarian cancer, or lung cancer (after adjusting for smoking). In fact, nonsignificant positive associations were observed for colorectal cancer, melanoma, and for lung cancer. Relevant potential confounding factors were considered for each of the cancer sites; inclusion of these factors did not substantially alter the results, except that results for colorectal and uterine cancer became more strongly positive, though the trends were not statistically significant.

The authors conclude that the data provide evidence against the hypothesis of an overall protective effect of selenium within the range of intakes among women in the U.S. but acknowledge that the implications of the observed positive associations are unclear. In addition, they note that knowledge of selenium metabolism is incomplete, and suggest further research in this area that may help elucidate the biological basis for any relationship between selenium and cancer risk.

In an accompanying editorial, Larry C. Clark, MPH, PhD, and David S. Alberts, MD, Arizona Cancer Center, suggest that despite the magnitude of this cohort study, interpretation of the results is complicated by possible selenium status differences in population subgroups with particular cancer risks from age, exposure or nutrient status, changes in participants' toenail selenium levels over time, and other factors. They believe, in light of conflicting results from epidemiologic studies, that the findings from this study should not be the sole basis upon which individual or public health decisions are made. They emphasize the importance of prospective clinical trials of selenium to address the issue of cancer prevention directly.