Re: Physical activity and risks of proximal and distal colon cancers: a systematic review and meta-analysis.

We read with great interest the systematic review and meta-analysis by Boyle et al. (1) on physical activity and risks of proximal colon cancers (PCCs) and distal colon cancers (DCCs), under the hypothesis that PCCs and DCCs may be two distinct cancer types with different genetic and environmental risk factors. The results suggested that physical activity is associated with a reduced risk of both PCCs and DCCs. This is a very important investigation because colorectal cancer is the third most common cancer in both men and women (2). However, as a limitation of their study, the authors stated that future research should focus on whether the intensity of physical activity influences the association between physical activity and the risk of colon cancer, which was also stressed in the recent article by Barton (3). Recently, we published a systematic review and metaanalysis (4) on physical activity and risk of breast cancer and found that intensity of physical activity influences its association with the risk of breast cancer. Thus, in this letter, we explore whether the intensity of physical activity influences its association with the risk of PCCs and DCCs. A two-stage, random-effects, dose– response meta-analysis taking into account the between-study heterogeneity was performed, as has been well described by Orsini et al. (5). Briefly, three knots at the 25th, 50th, and 75th percentiles of the levels of physical activity were first estimated in a restricted cubic spline model using generalized least square regression taking into account the correlation within each set of published relative risk. Then a multivariable random-effects meta-analysis was conducted to combine the study-specific estimates using the restricted maximum likelihood method. Metabolic equivalent (MET) is the most commonly used measurement of physical activity intensity (6). Thus studies were included if they provided the following information (7): relative risk with 95% confidence interval (CI) for three or more quantitative categories (METhours) of physical activity (we presented all results with relative risk for simplicity), as well as the number of cases and participants (or person-years) for each category of physical activity. The median or mean level of physical activity for each category was assigned to the corresponding relative risk for every study. For PCC, data from seven studies [references (9,19,20,27,32,33,39) in the study by Boyle et al. (1)], which included 1534 PCC case subjects, were used. A linear relationship was found, and the risk of PCC was 0.97 (95% CI = 0.90 to 1.04), 0.94 (95% CI = 0.83 to 1.07), 0.90 (95% CI = 0.77 to 1.03), 0.85 (95% CI = 0.73 to 0.98), 0.79 (95% CI = 0.67 to 0.93), and 0.74 (95% CI = 0.59 to 0.92) for 10, 20, 30, 40, 50, and 60 MET-hours/week, respectively (Figure 1). For DCC, data from the above-mentioned seven studies, which included 1583 DCC case subjects, were used. Some evidence of a nonlinear relationship was found, and the risk of DCC was 0.90 (95% CI = 0.83t o 0.96), 0.80 (95% CI = 0.69 to 0.92), 0.77 (95% CI = 0.67 to 0.91), 0.81 (95% CI = 0.70 to 0.93), 0.84 (95% CI = 0.71 to 0.99), and 0.86 (95% CI = 0.70 to 1.06) for 10, 20, 30, 40, 50, and 60 METhours/week, respectively.