The effect of a low-protein diet for diabetic nephropathy is underestimated.

We appreciate the opportunity to respond to the letter from Drs Bekaert and Rayman regarding our recently published article (1) on the association between plasma selenium concentrations and prostate cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC). Bekaert and Rayman point out that the mean plasma selenium concentration of EPIC participants is lower than that reported in North American populations, and we acknowledge that the range of selenium concentrations may have to extend beyond that required to saturate plasma glutathione peroxidase and selenoprotein P to observe an association with prostate cancer risk. Nonetheless, our results are consistent with some other studies conducted in men with relatively high selenium concentrations (2–4), suggesting that our null finding may not be solely due to low baseline concentrations. Evidence from trials is also inconsistent; although the Nutritional Prevention of Cancer trial showed a significantly reduced risk of prostate cancer among participants receiving supplementation who had a low prostate specificantigen concentration and a low baseline selenium concentration (5), the Selenium and Vitamin E Cancer Prevention Trial (SELECT) was prematurely discontinued because there was evidence that selenium supplementation did not prevent prostate cancer and some concern that it may increase the risk of diabetes (6). Only if results of detailed subgroup analyses from the SELECT can confirm the apparent protective effect of selenium supplementation among men with low baseline selenium concentrations do we think that further research on selenium supplementation and prostate cancer risk is warranted.