Fruit and vegetable intake and breast cancer risk: a case for subtype-specific risk?

In this issue of the Journal, Jung et al. (1) report no overall effect of fruit and vegetable consumption on breast cancer risk among women in the large, long-term Pooling Project of Prospective Studies of Diet and Cancer (Pooling Project). When the authors considered hormone-responsive, estrogen receptor (ER)–positive patients separately from the nonhormonal, ER-negative patients under a hypothesis of separate etiology, a statistically significant protective effect of fruit and vegetable consumption was observed for risk of ER-negative, but not ER-positive breast cancer. Most striking, the authors provide evidence that the protective effect of fruit and vegetable consumption for ER-negative breast cancer, in terms of magnitude and direction, were largely consistent across the 20 pooled studies. With these findings, Jung et al. (1) add to the growing number of studies reporting the differential effect of risk factors, including the classic breast cancer reproductive risk factors such as age of first pregnancy, when breast cancer is considered as separate subgroups such as ER-positive, ER-negative and triple negative breast cancer (2). These studies indicate that considering ER-positive and ER-negative breast cancer as a single disease in diet association studies has likely resulted in an underestimation of risk and a possible failure to detect even modest effects that may modify ER-negative breast cancer risk. These findings support the concept of distinct natural histories of hormone-dependent and hormone-independent tumors, as has been suggested by molecular profiling studies (3) and as heralds back to the findings from the Iowa Women's Study, which described differences in reproductive factors and risk for ER-positive, progesterone receptor–positive, and ER-negative breast cancers (4). For more than four decades, the role of fruit and vegetable consumption as dietary modifiers of breast cancer risk has been studied (5). Despite several mechanism-based hypotheses favoring a role for a protective effect of higher intake of plant-rich diets (6,7), results from population-based studies of fruit and vegetable intake and breast cancer risk have been inconsistent and suggestive of protection only for premenopausal disease (5). Given the evidence for etiologic heterogeneity among breast cancers, questions emerge. Most immediate is whether or not we know enough about the tumor subtypes to begin to separate breast cancer into distinct diseases for testing epidemiological associations. And, importantly, why do so? Support for the separation of breast cancer on ER status as a first approximation of distinct etiological histories exists from hierarchical clustering from gene expression profiling that shows a clear …