Re: Meat, fish, and colorectal cancer risk: the European Prospective Investigation into Cancer and Nutrition.

In a large cohort comprising 10 populations in the European Prospective Investigation into Cancer and Nutrition, Norat et al. ( 1 ) reported that processed and red meat intake was associated with elevated rates of colorectal cancer and its subtypes. Although the authors considered several study limitations, they may have omitted one that is key: the possibility that confounding by socioeconomic position may be responsible for the diet – disease gradients. In populations drawn from some of the countries featured in the article, markers of socioeconomic position have been shown to be associated with selfreported dietary characteristics, including meat consumption ( 2 ) . Thus, persons who are socioeconomically disadvantaged are more likely to report higher intake than their affl uent counterparts ( 2 ) . A raised risk of colorectal cancer has also been found in persons from deprived social groups, as indexed by lower levels of educational attainment ( 3 ) . In exploring the relationship between meat consumption (indeed, most indicators of food intake) and colon cancer ( indeed, most chronic disease outcomes), surprisingly few investigators adjust for socioeconomic indices, so judging the impact of this covariate on the dietdisease relationship is problematic. However, a suggestion that socioeconomic deprivation may have a role as a confounder in the meat – colon cancer relationship can be found in a study that appears to comprise a socioeconomically homogenous group of women. As cited by the authors ( 1 ) , but not discussed in the present context, an early report from the Nurses’ Health Study ( 4 ) found a positive relationship of both unprocessed meat (beef, pork, or lamb) and processed meat intake with incident colon cancer. This association was essentially lost in a later follow-up study of the same population ( 5 ) containing over four times the number of cases ( n = 670) and therefore greater statistical precision than the earlier report. In a series of articles ( 3 , 6 , 7 ) , the European Union Working Group on Socioeconomic Inequalities in Health has reported that the methodologic issues of comparing the relationship between mortality and socioeconomic indices (i.e., education and occupational social class) across culturally disparate European settings can be surmounted. Assuming that similar data are available in at least some of the cohorts comprising the present report ( 1 ) , as they should be, presumably the potentially confounding role of socioeconomic position in the meat – colon cancer relationship could be explored in a subgroup of study participants and reported by the authors.