Diet and lifestyle factors and risk of subtypes of esophageal and gastric cancers: classification tree analysis

Abstract

Purpose

Although risk factors for squamous cell carcinoma of the esophagus and adenocarcinomas of the esophagus (EA), gastric cardia (GC), and other (noncardia) gastric (OG) sites have been identified, little is known about interactions among risk factors. We sought to examine interactions of diet, other lifestyle, and medical factors with risks of subtypes of esophageal and gastric cancers.

Methods

We used classification tree analysis to analyze data from a population-based case-control study (1095 cases, 687 controls) conducted in Connecticut, New Jersey, and western Washington State.

Results

Frequency of reported gastroesophageal reflux disease symptoms was the most important risk stratification factor for EA, GC, and OG, with dietary factors (EA, OG), smoking (EA, GC), wine intake (GC, OG), age (OG), and income (OG) appearing to modify the risk of these cancer sites. For esophageal squamous cell carcinoma, smoking was the most important risk stratification factor, with gastroesophageal reflux disease, income, race, noncitrus fruit, and energy intakes further modifying risk.

Conclusion

Various combinations of risk factors appear to interact to affect risk of each cancer subtype. Replication of these data mining analyses are required before suggesting causal pathways; however, the classification tree results are useful in partitioning risk and mapping multilevel interactions among risk variables.

Introduction

Esophageal adenocarcinoma (EA) and, to a lesser extent, gastric cardia (GC) adenocarcinoma have been increasing in incidence [1], [2], [3]. It has been reported that the annual incidence of EA increased 350% between 1976 and 1994 [2], and Holmes and Vaughan [4] reported a 6.75-fold increase among white men between 1973 and 2002. Increases have been found in the United States [1], [2], [3], [4], [5], Great Britain, Australia, The Netherlands, Denmark, and other western nations [6]. Etiologic studies in the United States and elsewhere have identified several important risk factors, including cigarette smoking [7], [8], [9], [10], obesity [11], [12], [13], [14], and gastroesophageal reflux disease (GERD) [15], [16], [17]. It has also been shown that Helicobacter pylori colonization may be protective for EA and GC adenocarcinoma, particularly so for CagA-positive strains [18], [19], [20].

Epidemiologic studies have reported that fruit and vegetable consumption may be inversely associated with risks of esophageal and gastric cancer without regard to subsite or histologic type [21]. A review of the literature conducted by Thrift et al. [22] indicated that a moderate to substantially decreased risk of EA is associated with regular fruit and vegetable intake. There is limited evidence, however, examining the role of dietary factors on subtypes of these cancers in combination with other factors. We have previously reported significant inverse associations between intake of nutrients found primarily in plant-based foods and the risk of EA and GC adenocarcinoma [23], [24]. In addition, Steevens et al. [25] noted a statistically significant reduced risk of EA associated with raw vegetable consumption and a significant inverse association between brassica vegetables and GC adenocarcinoma among a cohort of Dutch men and women. However, in the AARP cohort, Freedman et al. [26] found a significant inverse association between fruit intake and risk of esophageal squamous cell carcinoma but not of EA. Although we observed a significant positive association between intake of meat and animal protein and risk of EA and GC adenocarcinoma [23], [24], Keszei et al. [27], in an analysis of data from The Netherlands Cohort Study, did not find any association between red or processed meat and esophageal or gastric adenocarcinomas. They did, however, find a significantly elevated risk of esophageal squamous cell carcinoma associated with both red and processed meats among men [27]. According to the World Cancer Research Fund expert panel report, the available evidence suggests a positive association between noncardia gastric adenocarcinoma and nitrite-related foods in western countries and salted or preserved foods in Asian countries [21]. There is also evidence of an inverse association between dietary fiber intake and risks of EA [11], [23] and GC adenocarcinoma [23], [28].

Dietary behaviors are complex. For example, consumption of fruit and vegetables is associated both positively and negatively with consumption of other food groups [29]. In addition, dietary behaviors correlate with other health behaviors and demographic factors [30], [31], [32]. Recursive partitioning techniques, including classification trees, have been used as a means of examining the complex interactions or patterns of risk factors in a variety of diseases [33], [34], including colon [35] and lung [36] cancers. Classification tree analysis is agnostic in evaluating interactions that do not need to be prespecified, in contrast to standard regression models, in which interactions are generally prespecified [37]. Given that most cancers are multifactorial in nature, often involving combinations of both host and genetic factors in determining risk, classification tree models can give clues to important interactions by sorting through the complex, multilevel nature of risk factors associated with these cancers. Thus, the purpose of this analysis was to explore a variety of dietary and lifestyle variables as predictors of risk of subtypes of esophageal and gastric cancer, better understand which of these correlated variables appears to be most important for risk stratification, and examine multilevel interactions involving these same variables.