Serum selenium contents and the risk of cancer

To investigate the relationship between serum Se contents and the risk of cancer, 4857 serum samples were obtained from cancerous (2730) and non-cancerous patients (2127). In this cohort, the incidence of cancer in the non-cancerous patients was followed for the subsequent 3 years. The serum Se level of non-cancerous patients who later developed cancer during the 3 years was determined and compared with that of the non-cancerous patients. A high incidence of cancer was observed in the lower serum Se patients of the non-cancerous group. The serum Se levels of cancerous patients were significantly lower than in non-cancerous patients. These results suggest that the low serum Se level in cancerous patients may not be induced by the tumor but it was more likely already present before the tumor. To calculate the odds ratio, cancerous and non-cancerous patients were divided into group of less than 110 ppb and more than 111 ppb group in serum Se. The odds ratio was estimated to be 1.95, suggesting the high risk of cancer in the low serum Se group.