Is there a dose-response effect of exercise on inflammation, fatigue and activity in cancer survivors?Systematic review with meta-regression analysis of randomised trials.Adults diagnosed with cancer, regardless of specific diagnosis or treatment.Exercise interventions including aerobic and/or resistance as a key component.The primary outcome measures were markers of inflammation (including C-reactive protein and interleukins) and various measures of fatigue. The secondary outcomes were: measures of activity, as defined by the World Health Organization's International Classification of Functioning, Disability and Health, including activities of daily living and measures of functional mobility (eg, 6-minute walk test, timed sit-to-stand and stair-climb tests). Risk of bias was evaluated using the PEDro scale, and overall quality of evidence was assessed using the Grades of Research, Assessment, Development and Evaluation (GRADE) approach.Forty-two trials involving 3816 participants were included. There was very low-quality to moderate-quality evidence that exercise results in significant reductions in fatigue (SMD 0.32, 95% CI 0.13 to 0.52) and increased walking endurance (SMD 0.77, 95% CI 0.26 to 1.28). A significant negative association was found between aerobic exercise intensity and fatigue reduction. A peak effect was found for moderate-intensity aerobic exercise for improving walking endurance. No dose-response relationship was found between exercise and markers of inflammation or exercise duration and outcomes. Rates of adherence were typically high and few adverse events were reported.Exercise is safe, reduces fatigue and increases endurance in cancer survivors. The results support the recommendation of prescribing moderate-intensity aerobic exercise to reduce fatigue and improve activity in people with cancer.PROSPERO CRD42015019164.