JNCI Journal of the National Cancer Institute 2006 98(20):1501; doi:10.1093/jnci/djj407
© The Author 2006. Published by Oxford University Press.
Erratum
"Phytoestrogen Intake and Endometrial Cancer Risk" by Horn-Ross et al. [J Natl Cancer Inst 2003;95:1158–64]. Recently, during additional analyses of the dietary data from this case–control study, we identified the need to revise our initial coding of portion size for a number of foods. The initial coding had inadvertently resulted in the calculation of caloric intake in a manner that was often different for case and control subjects. Thus, we reanalyzed the data reported in the above-referenced article. Although estimates of phytoestrogen intake changed only minimally (corrected Table 2, available at http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue20/), caloric intake changed substantially for some women. When caloric intake was included as a covariate in the exposure–disease models, all the compounds that were associated with reduced risk of endometrial cancer in the original report yielded odds ratios that were essentially equivalent to unity (corrected Table 3, available at http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue20/). Thus, for example, the odds ratio of endometrial cancer for women in the highest versus lowest quartile of total isoflavone intake should be odds ratio (OR) = 1.1 (95% confidence interval [CI] = 0.66 to 1.7, Ptrend = .84) and that for women in the highest versus lowest quartile of total lignan intake should be OR = 1.1 (95% CI = 0.69 to 1.7, Ptrend = .94). Similarly, the statistically significantly reduced risks in postmenopausal women specified in the original report also became essentially equivalent to unity when corrections were made (corrected Table 4, available at http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue20/). These corrected results suggest that there is no association between phytoestrogen consumption and endometrial cancer risk at levels consumed in this non-Asian population, contrary to our original report.
Interestingly, however, we still obtained evidence confirming our a priori hypothesis that body mass index modifies the phytoestrogen–endometrial cancer association (corrected Table 5, available at http://jncicancerspectrum.oxfordjournals.org/jnci/content/vol98/issue20/). Among postmenopausal women, those who were obese and consumed low levels of isoflavones (<1.5 mg/d) had a higher risk of endometrial cancer than normal-weight women consuming high levels (
1.5 mg/d) of isoflavones (OR = 4.7, 95% CI = 2.3 to 9.5). In addition, this risk increase was greater than that for obese women consuming high levels of isoflavones when compared with normal-weight women consuming high levels of isoflavones (OR = 2.5, 95% CI = 1.4 to 4.6). However, as in our original article, the interaction between body mass index and isoflavone intake was not statistically significant. A similar pattern was seen for consumption of total lignans. Thus, our conclusion regarding effect modification—i.e., that the effects of phytoestrogens in this high-risk subgroup of obese women is worthy of further investigation—remains valid. The authors regret the error.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||