LAUSR

dose-response relationship between moderate-intensity and vigorous-intensity physical activity and all-cause mortality risk in “insufficiently active weekend warriors” who reported 1 or 2 sessions per week but did not meet physical activity guidelines of at least 150 minutes per week of moderate-intensity aerobic activity or at least 75 minutes per week of vigorousintensity aerobic activity. We did observe a linear trend when investigating total physical activity of any intensity. We concluded that some of the health benefits might be explained by nonexercise activity, such as light-intensity walking. More than 40% of the weekend warriors were in desk-bound occupations, and we would suggest that participation in sport and exercise at the weekend is enough to increase cardiorespiratory fitness and to reduce the mortality risk associated with the sedentary lifestyle of Western societies. We thank Lam and colleagues for their letter too. They suggested that air pollution was subject to geographical variation and that air pollution was a relevant confounding variable. We did not adjust for air pollution; however, the available evidence suggests that air pollution is only related to lung cancer mortality.2 Lam and colleagues also suggested that treatment availability was subject to geographical variation and that treatment availability was also a pertinent confounding variable. There is some evidence of a North-South divide in health care in the United Kingdom; however, socioeconomic factors may explain differences in physical activity3 and other exposures and outcomes.4,5 Compared with the inactive participants in our study, the hazard ratio for cancer mortality was 0.79 (95% CI, 0.66-0.94) in the regularly active and 0.82 (95% CI, 0.63-1.06) in the weekend warriors after adjustment for age, sex, smoking habit, longstanding illness, and socioeconomic status (the regularly active reported ≥150 minutes/wk in moderate-intensity aerobic activity or ≥75 minutes/wk in vigorousintensity aerobic activity from ≥3 sessions; the weekend warriors reported the same amounts of activity per week from 1 or 2 sessions). Lam and colleagues mentioned a clustering effect in our subsample. The core sample is weighted so that it might be representative of the population living in private households.6 When we have weighted the subsample, it has had little bearing on the association between physical activity and mortality.