This review evaluated the strength of the evidence for a causal relationship between physical activity (PA) and colorectal cancer (CRC). A systematic review of databases through February 2008 was conducted to identify studies that assessed the association between total or recreational PA and incidence or mortality of CRC (including CRC, rectal cancer, colon cancer, and proximal or distal colon cancer). Studies were evaluated for significant associations between PA and risk of CRC endpoints and for evidence of dose-response relationships in the highest quality studies. Twenty cohort studies were evaluated; 11 were high-quality. Fifty percent of all studies and 64% of highest quality studies reported at least one significant association between PA and risk of a CRC endpoint (P<0.05). However, only 28% of all analyses (31% of analyses of highest quality studies) were significant (P<0.05). Only 40% of analyses of highest quality studies resulted in a significant P for trend (P<0.05); however, a non-significant inverse linear association between PA and colon cancer risk was apparent. Heterogeneity in the evidence from all studies and from the highest quality studies was evident. Evidence from cohort studies is not sufficient to claim a convincing relationship exists between PA and CRC risk.