Background The WHO has concluded that physical activity reduces the risk of numerous diseases. However, few systemic reviews have been performed to assess the role of leisure time physical activity (LTPA) in lowering the risk of cancer in a dose-dependent manner and furthermore the suitability of recommendation of physical activity by the WHO.

Methods A systematic review and meta-analysis was designed to estimate cancer risk by LTPA in binary comparison and in a dose-dependent manner. MEDLINE and Web of Science were searched up to 30 December 2014 without language restrictions. Reference lists were reviewed for potential articles.

Results A total of 126 studies were recruited into the meta-analysis. Overall, the total cancer risk was reduced by 10% in people who undertook the most LTPA as compared with those who did the least. Dose–response meta-analysis indicated that the current WHO recommendation (equal to an average of 10 metabolic equivalents of energy hours per week) induced a 7% (95% CI 5% to 9%) cancer reduction. Moreover, the protective role of LTPA against cancer becomes saturated at 20 metabolic equivalents of energy hours per week, with a relative risk of 0.91 (95% CI 0.88 to 0.93). Subanalyses results based on cancer types showed that LTPA only exhibited significant protection against breast cancer and colorectal cancer.

Conclusions Our meta-analysis indicates that the current WHO recommendation of physical activity can result in a 7% reduction in cancer risk, which is mainly attributed to its protective role against breast cancer and colorectal cancer. Furthermore, two-fold of current recommendation level is considered to give its saturated protection against cancer.