Vinceti M
Oral presentation at the FESTEM 2019 International Symposium on Trace Elements and Minerals, Potsdam 2-5 April 2019
Abstract
Risk assessment for human selenium exposure is difficult, complex, and fascinating, like for other essential minerals with both nutritional and toxicological relevance, such as manganese, iodine, copper and iron. Selenium is ubiquitous in the environment, and primary sources of exposure include diet (generally the most important one), drinking water, air pollution from motorized traffic, coal combustion and industrial facilities, hair care products, and selected occupations. Selenium exists in inorganic and organic chemical forms, each having distinctive nutritional and toxicological properties.
Any risk assessment of this trace element, including effects of deficiency and overexposure, should take into consideration the rapidly evolving evidence in both epidemiology and laboratory sciences. Interestingly, the agencies addressing selenium risk assessment have yielded markedly different recommendations for the safe range of intake, depending on their interpretation of the literature. In addition, such agencies did not generally use epidemiologic human data for assessment, mainly relying on biochemical evidence based on proteomic endpoints (selenoprotein upregulation). Results of the most recent experimental and nonexperimental epidemiologic studies have yielded concerning results about selenium safety, contrary to results based on biochemical endpoints, challenging the reliability of current assessments of the safe range of selenium exposure. These results have not yet been incorporated into risk assessment of the health effects of selenium exposure by regulatory and scientific agencies.
This new evidence generated by the epidemiology must therefore be carefully assessed, weighting the evidence yielded by both experimental and nonexperimental studies, and bridging such evidence with that yielded by recent nutritional and toxicologic investigations. Such a reassessment should be mainly driven by the evidence from the low-bias controlled randomized trials, and by the dose-response relations from meta-analyses of epidemiologic studies.