In the context of recent social and economic transitions in China, lack of physical activity among adolescents is an emerging health risk, particularly so in rapidly expanding urban areas. Evidence from Western countries suggests that built environment attributes can influence the physical activity participation of young people, but whether or not this is the case for China is unknown. We recruited high school students from ten urban districts in Nanjing, Mainland China (n=2,375; mean age=13.9± 1.0 years old; 46% boys; survey response rate=89%). The outcome variable was selfreported recreational physical activity time; the primary explanatory variable was the residential density of the urban districts. Analysis was conducted using mixed-effects logistic regression models. After adjustment for potential confounding variables, including sedentary behavior and green space, there was a consistent and graduated association between residential density and physical activity; residential density was significantly negatively associated with recreational physical activity time for students from the higher tertile of residential density (OR; 95% CI=0.64; 0.42 to 0.97) compared to those from the lower tertile. The direction and magnitude of the negative association between residential density and physical activity was similar for boys and girls. It should be a public health priority to identify the particular urban environment attributes that can encourage and support young people's participation in physical activity.

Residential density was positively associated with overweight among urban Chinese adolescents. Our findings warrant further research examining attributes of urban environments associated with adolescents' obesity in China and potential mechanisms between them.