n engl j med 360;14 nejm.org april 2, 2009 1463 daily for up to 4 years. Novelose was used because it was relatively easy to mix with a variety of foods. The dose used was as much as we thought we could ask participants to add to their diet. There were 97 withdrawals on the basis of “minor disorders,” as compared with 76 among participants receiving placebo starch. The difference was attributable to symptoms of bloating, suggesting that we had delivered a clinically significant dose. Analysis of crypt-cell proliferation and dissected crypt dimensions in mucosal-biopsy specimens (unpublished data) revealed no difference between the placebo group and either of the treatment groups. We agree that the literature supports the case for an antineoplastic effect of resistant starch. The lack of support for this hypothesis in our trial may simply reflect the major challenge of testing nutritional chemoprevention with the use of the traditional approach of a randomly assigned supplement. D. Timothy Bishop, M.D.