This is an ongoing debate: Is soy good or bad for prostate cancer survivors? Here’s a comprehensive update, courtesy MedPage TODAY:
The Two Sides of Soy and Prostate Cancer
JULY 15, 2013 – By Charles Bankhead, Staff Writer, MedPage Today
Last week’s MedPage Today news coverage included a report on a study showing that soy supplementation did not reduce the risk of biochemical recurrence in prostate cancer. In an accompanying discussant video, Derek Raghavan, MD, of Carolinas Healthcare System in Charlotte, N.C., cited a number of limitations in the study, which he summed up as disappointing.
The first author of the journal article reporting the study results, Maarten Bosland, DVSc, PhD, of the University of Illinois at Chicago, responded to the criticism by email, leading to correspondence between Bosland and Raghavan, with MedPage Today as the conduit.
The following is a summary of that correspondence, presented in a “point-counterpoint” format to give both parties a chance to express their views and to allow MedPage Today (MPT) readers to draw their own conclusions.
Rahgavan Comment in original MPT article: This is a really disappointing study.
Bosland response: The outcome of the study was indeed disappointing because based on the hypothesis, a reduction in the development of recurrence of prostate cancer after a radical prostatectomy was expected. On the other hand, the results were statistically very solid and the lack of a treatment effect convincing. In addition, the soy supplementation was found to be safe.
Raghavan reply: I don’t think the safety of soy supplementation was in question. I still believe that the study was NOT statistically solid — the study was under-powered, and it is not possible to tell with certainty that the control group and the test group truly were well matched.
My reason for saying the study is disappointing is that it took them such a long time to address a relatively simple question, and they still ended up with an under-powered study that didn’t finally resolve the question. As noted before, p-values are less helpful as predictive of statistical significance in an under-powered study, and this has been known since the writings of Maurice Staquet more than 30 years ago.
Rahgavan Comment in original MPT article: I don’t think it tells us anything new.
Bosland response: This is the first study ever to examine the effect of soy supplementation on the risk of recurrence of prostate cancer after a radical prostatectomy. The only other studies examining the effects of soy have been small short-duration clinical trials with changes [in] PSA as [the] endpoint. The results of these studies have been mixed, and short-term changes in PSA caused by diet supplements may not necessarily reflect clinical benefit.
Rahgavan Comment in original MPT article: Half of the patients got what appears to be a poorly characterized soy product, and the other half got a calcium caseinate product, which in other contexts is used as a source of energy in diet supplementation. It was meant to be a placebo, yet it was actually a mechanism for giving calories.
Bosland response: The soy and placebo supplements were nutritionally well characterized, and details are presented in the paper. Both supplements provided equal amounts of vitamins and minerals as well as an equal small amount of energy (175 kcal per daily serving). Dietary protein supplements are not typically used to provide additional calories but additional protein.
Raghavan reply: I agree that dietary protein supplements are not typically used to provide additional calories as their primary purpose; my point is that calcium caseinate is not a pure placebo.
Rahgavan Comment in original MPT article: The trial included men with various levels of risk, required 13 years to conduct, and enrolled relatively few men despite screening thousands of patients.
Bosland response: Slightly more than 3,000 men were screened to identify 952 men who were potentially at elevated risk to develop recurrence. After reviewing their medical records and pathology, 590 were confirmed to be eligible, 104 of whom could not be reached to ask whether they would participate. Of the remaining 486 men that could be interviewed, one third (167) declined to participate and one third (151) was excluded because of various medical and other reasons. One third (168) of these 486 men were enrolled as well as 9 other men who were referred to the study by urologists elsewhere. Enrolling one-third of all eligible men who could be reached is actually a very good record for a prevention-type clinical trial. That this took 13 years was due to the fact that only 19% of men undergoing a radical prostatectomy were confirmed to be at elevated risk for recurrence in this study (which is positive because more than 80% were not at high risk and potentially cured).
Raghavan reply: No disagreement with any of the above response, but it doesn’t alter the fact that they had a rather small number of cases in their trial after 13 years. I didn’t imply that they did not TRY to do the study well; but the reality is that the study is under-powered. In the field of cancer epidemiology, prevention studies are usually conducted with much larger numbers of cases randomized; I respect the fact that the investigators TRIED to do the right study with high risk patients, but they didn’t succeed in getting adequate numbers — in my book, they failed to achieve their goals.
Rahgavan Comment in original MPT article: It’s really impossible to figure out whether those confounding factors would have made the study null and void … . This is a really disappointing study, because it was poorly designed and executed.
Bosland response: The soy and placebo supplement groups were carefully randomized and did not significantly differ in distribution of risk factors for recurrence or in any other respect for that matter. Thus, the notion that “confounding factors would have made the study null and void” seems not supported by the data presented in the paper and was not shared by the peer reviewers of the paper for the Journal of the American Medical Association (JAMA) who also found the study to be well designed and well executed.
Raghavan reply: This is a hard one to address. I think the reviewers from JAMA made a mistake. It’s a little unusual for authors to make a fuss in this type of educational activity, and I can only reiterate that my points are statistically valid, and the authors simply cannot claim the robustness of their data that they imply. The two populations, with small numbers, could easily have been quite different, and there were apparent differences between the two populations that would not be identified statistically in an under-powered study. As also noted in my earlier email, the investigators are well respected and have done good work; this just doesn’t happen to be their best published data.
Well, glad we cleared that up! My opinion is simple. Eating soy several times a week is good for just about anyone; beats the red meat alternative.
Feel good and keep smiling! Pat