Objective: We wanted to explore if FTO genotype interacts with fat intake, or leisure-time physical activity, on fat mass, lean mass and mortality. Subjects and methods: Among 22 799 individuals (44-74 years) in the population-based Malmö diet and cancer cohort that were genotyped for rs9939609 in FTO and had information on dietary intake (from a modified diet history method) and no history of diabetes, cancer or cardiovascular disease, 2255 deaths (including 1100 cancer and 674 cardiovascular deaths) occurred during 12.0 years of follow-up. Leisure-time physical activity was determined from a list of 17 different physical activities in a questionnaire. Body composition was measured using bioelectric impedance method. Results: FTO genotype associated strongly with both fat mass and lean mass (P trend o1 Â 10 À16 for both) but we found only significant interactions with fat intake, or physical activity, on fat mass (P interaction ¼ 0.01 and 0.004). No significant interaction between FTO genotype and fat intake (P interaction ¼ 0.72), or leisure-time physical activity (P interaction ¼ 0.07), on total mortality were observed. However, we observed a significant interaction between leisure-time physical activity and FTO genotype on cardiovascular mortality (P interaction ¼ 0.03). The highest vs lowest quintile of physical activity was associated with 46% (95% confidence interval, 17-64%) reduced cardiovascular mortality among TT-carriers (P trend ¼ 0.004), and 11% reduced cardiovascular mortality among A-allele carriers (P trend ¼ 0.68). Conclusion: Our results indicate that FTO genotype associates with both fat mass and lean mass, but the level of fat intake and physical activity only modify the association with fat mass. In addition, FTO genotype may modify the association between physical activity and cardiovascular mortality.