Although considerable evidence suggests that leisure-time physical activity is associated with a reduced risk of endometrial cancer (EC), the shape of dose-response relationship has not been investigated and previous meta-analyses have not accounted for differences in measures of physical activity. To address such issues, we conducted linear and nonlinear dose-response meta-analyses by metabolic equivalent of task (MET)-hour/week and hour/week, respectively, based on observational studies published up to September 2013 identified from PubMed and Embase databases. Summary relative risks (RRs) and 95% confidence intervals (CIs) were calculated using a random-effects model. In the linear dose-response analysis, an increase in leisure-time physical activity by 3 MET-hour/week was associated with an ∼2% reduced risk of EC (summary RR = 0.98, p = 0.02, 95% CI = 0.95-1.00, I(2) = 53%, p(heterogeneity) = 0.06, three case-control studies and three cohort studies, 3,460 cases, range of activity = 0-50 MET-hour/week) and an increase by an hour/week was associated with an ∼5% reduced risk of EC (summary RR = 0.95, p < 0.001, 95% CI = 0.93-0.98, I(2) = 31%, p(heterogeneity) = 0.20, four case-control studies and two cohort studies, 3,314 cases, range of activity = 0-12 hour/week). Nonlinear dose-response meta-analysis suggested that the curve may plateau at 10 MET-hour/week (p(change) in slope = 0.04) but this statistical significance was sensitive to one study. No evidence of a nonlinear association was indicated by hour/week (p(change) in slope > 0.69). In conclusion, an increase in leisure-time physical activity may continue to decrease EC risk, within the range of 0-50 MET-hour/week or 0-15 hour/week. Future studies should evaluate possible independent role of intensity of physical activity and effect modification by obesity.
Keywords: dose-response; endometrial cancer; meta-analysis; physical activity.
© 2013 UICC.