Forty-four RCTs (3,254 participants) were included in the review. The mean methodological quality score was 6.8 (range 3 to 10) out of a maximum of 11 points.
Exercise interventions reduced cancer-related fatigue levels in cancer patients more than in usual care controls (d=0.31, 95% CI 0.22 to 0.40; 44 RCTs, I2=50%). Results remained significant for breast cancer (25 RCTs, I2=42%) and prostate cancer (four RCTs), but not for lymphoma (four RCTs), colorectal cancer (one RCT) one leukaemia (one RCT). There was evidence of heterogeneity for the overall analysis and for breast cancer.
Cancer-related fatigue levels reduced to a greater extent when interventions were theoretically driven (p<0.001) than when they were not and when cancer survivors were older rather than younger (p=0.04). Cancer-related fatigue levels reduced more than controls when moderate-intensity resistance exercise was used compared with lower intensity or aerobic exercises of any level of physical exertion, particularly for higher quality interventions (p<0.05). Time since diagnosis, aerobic exercise, flexibility exercise or supervision of exercise sessions did not moderate the results.
There was no evidence of publication bias.