Obesity is usually associated with increased mortality in young and middle-aged people. However, in older people, higher body mass index (BMI) (calculated as weight in kilograms divided by the square of height in meters) is apparently protective, which casts doubt on whether obesity is a significant public health concern. A possible explanation for the observed relationship between BMI and mortality in older people is reverse causality, because BMI in this age group is the result, not the cause, of underlying illness. In this case, BMI in older people would be a predictor of mortality as a marker of aging and health status, and the association of BMI with mortality would be expected to differ with health status. In a prospective study of 54 088 older people, Schooling et al showed that the association between BMI and mortality varied with health status. In the small minority (9%) of healthy older people who have never smoked, higher BMI (>25) was associated with higher mortality compared with normal BMI; however, in unhealthy older people, higher BMI was associated with lower mortality.