© The Author 2006. Published by Oxford University Press.
EDITORIAL |
(Nutritional) Chemoprevention of Cancer: What's Up?
Correspondence to: John A. Baron, MD, Section of Biostatistics and Epidemiology, Dartmouth Medical School, Evergreen Center, 46 Centerra Pkwy., Ste. 300, Lebanon, NH 03766 (e-mail: john.a.baron{at}dartmouth.edu).
Some of the initial motivation for cancer chemoprevention was derived from observations regarding the effects of nutrients on carcinogenesis, and so it is fitting that the Journal now publishes a progress report of sorts for nutritional interventions. Davies et al. (1) have assumed the difficult task of summarizing the effects of all nutritional interventions across all the studied populations and endpoints. With the heterogeneity inherent in each of these, there is clearly a risk of mixing apples and oranges into an uninterpretable stew. However, even if the summary estimates may be difficult to interpret in some cases, it is valuable to see the studies arrayed together in one place. Moreover, it certainly is sensible to examine mortality across studies of a common treatment, even if different cancers are involved.
It seems disappointing that Davies et al. found the quality of the trials reviewed to be low. Their quality rating used an assessment of three facets of the trials: generation of the (randomized) allocation sequence, concealment of allocation, and masking of outcome assessors and participants. Most of the negative assessments actually reflected a lack of reporting, so arguably a better characterization of the trial quality might be "unknown." In any case, there are several other important aspects of clinical trials that were not judged but are commonly reported. Certainly compliance with study treatments is one such factor; low compliance erodes the exposure differences between study groups and so biases a trial toward a null result. There is a similar potential issue with use of the study agents outside the study. Many chemoprevention trials use interventions that are readily available to everyone; examples include dietary change, nutritional supplements, and aspirin. If—even in a blinded study—subjects access the study agents on their own, this activity will act like noncompliance to decrease differences between the study groups. Therefore, low drop-in rates are also important markers of trial quality. Similarly, losses to follow-up may compromise the ability of a trial to study its endpoints, and so another fundamental marker of a well-done clinical trial is a high rate of follow-up. In judging the quality of the trials included in the overview, it would have been informative to include these dimensions of trial performance.
The overview came to fairly negative conclusions, namely that most interventions cannot be reliably assessed because of limited data. Nonetheless, there were some tantalizing findings in the report. The summary estimates suggested non–statistically significant reductions in risk of all-cause mortality and cancer mortality with "healthy diet" interventions in cancer (largely breast cancer) patients. Although the reviewed studies were small, the larger Women's Nutrition Intervention Study among breast cancer patients was larger and has reported positive results in abstract (2).
Another article in this issue of the Journal was a trial of three treatments (two of them nutritional) to reduce the prevalence of precancerous gastric lesions (3). The study might have been a candidate for inclusion in the overview had it been published earlier, because the study population in China had already been shown to have an almost universal prevalence of preinvasive or preneoplastic lesions (4). However, in this study it was the nonnutrient intervention—antibiotic treatment for Helicobacter pylori—that had the positive result.
In many ways, this population, chosen because of its high prevalence of antecedents of gastric cancer, was an ideal one in which to conduct a chemoprevention study. The high prevalence of precursor lesions permitted assessment of effects on the occurrence of dysplasia or cancer and the related issues of progression or regression. However, studies in such populations, with marginal or frank nutritional deficiencies, may differ from those in Western populations. In Linxian, antioxidants conferred a reduced risk of gastric cancer (4), but beneficial effects have been notably absent in studies of other cancers in Western populations (5,6), with the possible exception of selenium (7).
The study population also brings another, more unexpected complication. The high prevalence of some of the endpoints actually introduces complexity in the interpretation of relative risk measures, in particular the odds ratios presented in their Table 3 (3). Odds ratios diverge progressively from risk ratios as event rates increase above 10% or so. As a result, the interpretation of the odds ratios in this study is not obvious. For example, in their Table 3, the odds ratio for the combined endpoint of severe chronic atrophic gastritis, intestinal metaplasia, dysplasia, or gastric cancer in 1999 was 0.77. This statistic, derived from prevalences of 73.4% in the treated group and 77.9% in the active group, translates to a risk ratio of 0.94 and a risk difference of 4.5%. Thus, the odds ratio does not mean a 23% reduction in risk as might be thought. The progression-versus-regression analysis arguably provides more user-friendly measures of effect, though of a different sort of endpoint.
In contrast, another methodologic point might suggest a greater efficacy than indicated in the report. This trial was not a direct test of H. pylori eradication but rather an investigation of one particular regimen, amoxicillin and omeprazole, with one retreatment if seronegativity was not obtained. A high proportion, but far from all, of the treated subjects became H. pylori seronegative. Therefore, a different eradication regimen that resulted in a more complete and sustained eradication of the bacterium might well show even stronger results.
Together the two articles in this issue of the Journal well illustrate the contemporary status of chemoprevention: hard to summarize, many negative findings, but some hopeful nuggets of progress.
REFERENCES
(1) Davies AA, Davey Smith G, Harbord R, Bekkering GE, Sterne JAC, Beynon R, Thomas S. Nutritional interventions and outcome in patients with cancer or preinvasive lesions: systematic review. J Natl Cancer Inst 2006;98:961–73.
(2) Chlebowski, Blackburn, Elashoff. Proc ASCO 2005;23:10.
(3) You W-c, Brown LM, Zhang L, Li J-y, Jin M-l, Chang Y-s, et al. Randomized double-blind factorial trial of three treatments to reduce the prevalence of precancerous gastric lesions. J Natl Cancer Inst 2006;98:974–83.
(4) Blot WJ, Li JY, Taylor PR, Guo W, Dawsey S, Wang GQ, et al. Nutrition intervention trials in Linxian, China: supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst 1993;85:1483–92.
(5) Lee IM, Cook NR, Gaziano JM, Gordon D, Ridker PM, Manson JE, et al. Vitamin E in the primary prevention of cardiovascular disease and cancer: the Women's Health Study: a randomized controlled trial. JAMA 2005;294:56–65.
(6) Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Antioxidant supplements for prevention of gastrointestinal cancers: a systematic review and meta-analysis. Lancet 2004;364:1219–28.[CrossRef][ISI][Medline]
(7) Clark LC, Combs GF Jr, Turnbull BW, Slate EH, Chalker DK, Chow J, et al. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 1996;276:1957–63. Erratum in: JAMA 1997;277:1520.[Abstract]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||