Re: A prospective investigation of coffee drinking and endometrial cancer incidence

Abstract

authors show statistically significant inverse associations for both caffeinated coffee and decaffeinated coffee, with per-cup hazard ratios of 0.90 (0.86–0.95) and 0.93 (0.87–0.99), respectively, and conclude that the observed reduction in endometrial cancer risk does not differ by caffeine content. In this letter, we attempt to provide an alternate interpretation for the results from this study, especially with regard to caffeinated and decaffeinated coffee. The examination of Table 2 in Ref. 1 reveals a greater magnitude of reduction in endometrial cancer risk with caffeinated coffee than with decaffeinated coffee. Compared to nondrinkers, hazard ratios for caffeinated coffee drinkers are 0.68 (0.53–0.87) and 0.46 (0.34–0.63) for 1 cup and >3 cups per day, respectively. Hazard ratios for decaffeinated coffee drinkers are 0.81 (0.62–1.06) and 0.81 (0.51–1.20) for 1 cup and >3 cups per day, respectively. Even though the p-trend is statistically significant for both coffee types, the 95% confidence intervals are consistently significant only for caffeinated coffee. Proposed biological mechanisms would predict a stronger association for caffeinated coffee than for decaffeinated coffee. Caffeinated coffee/caffeine is associated with higher levels of adiponectin levels, 5 an endogenous insulin sensitizer, and higher levels of sex hormone binding globulin (SHBG), a moiety that reduces bioavailable estrogen. 6,7 If coffee consumption influences endometrial cancer risk by reducing bioavailable estrogen and insulin levels, then there is greater plausibility for caffeinated coffee than for decaffeinated coffee. These mechanistic associations have not been observed for decaffeinated coffee. 5,7 Interestingly, in line with an insulin-mediated mechanism, a recent study suggested that coffee may reduce diabetes risk through SHBG, an association that is apparent for caffeinated coffee and caffeine, but not for decaffeinated coffee. 7 Both types of coffee are associated with lower C-peptide levels 8 and contain other compounds that may reduce endometrial cancer risk such as cholorogenic acid and antioxidants. Although it is possible that both types of coffee may reduce endometrial cancer risk, caffeine would be predicted to provide added reduction in risk. The method of exposure assessment used to ascertain caffeinated or decaffeinated coffee drinkers further provides room for misclassification across coffee types. Participants were asked to report the number of cups of coffee they drank. They were then asked to indicate whether they drank caffeinated or decaffeinated coffee more than half the time. Using only this information, those who drank decaffeinated coffee more than 50% of the time could, by definition, have also consumed some amount of caffeinated coffee. We can hypothesize that the reduction in risk in decaffeinated coffee could in part be due to caffeinated coffee or visa versa. However, given the strength of the proposed underlying biologic rationale for caffeinated coffee, the greater magnitude in reduction observed for caffeinated coffee than for decaffeinated coffee, and the inherent misclassification due to the way coffee type was determined, it is more likely that some of the reduction observed in decaffeinated coffee drinkers could in fact be attributed to caffeine. It is also of note that decaffeinated coffee contains small amounts of caffeine and that this study did not assess information on other sources of caffeine such as tea, soda and chocolate.