The fitness, obesity, and health equation: is physical activity the common denominator?

THE MODERN LIVING ENVIRONMENT IN DEVELOPED COUNtries is characterized by low daily energy expenditure and an abundant and inexpensive food supply, making positive energy balance common. The result is a rightward skewing of the body mass index (BMI) distribution and an increasing prevalence of obesity. Indisputable evidence links obesity to health problems, including risk of cardiovascular disease, type 2 diabetes, some cancers, and all-cause mortality. These associations are doserelated, temporally consistent, and biologically plausible, which support a causal hypothesis. Physical inactivity also has a dose-related, temporally consistent, and biologically plausible relationship to the same health outcomes as those as for obesity, and both obesity and inactivity have similar patterns of association with clinical risk indicators such as blood pressure, fasting plasma glucose, and inflammatory markers. Furthermore, declines in average daily energy expenditure are a likely underlying cause of the obesity epidemic. However, the majority of studies examining obesity and health have not adequately accounted for physical activity. When physical activity has been considered, investigators have often relied on simple self-report questionnaires in which inaccuracy can increase proportionally with the respondent’s weight. Failure to adequately quantify physical activity when examining the risks of obesity is similar to exploring risk factors for cancer and misclassifying tobacco use. Physical activity and weight are closely linked and each must be measured accurately and considered carefully when examining the other. Previous investigations, including studies from our group, have examined the individual and joint associations of cardiorespiratory fitness (CRF) and body habitus to health outcomes. Use of maximal exercise tests to quantify CRF provides an objective evaluation of an individual’s recent activity patterns and accounts for 70% to 80% of the variance in detailed activity records. Cardiorespiratory fitness is stronger than self-reported physical activity as a predictor of many health outcomes, most likely because fitness measurements are less prone to misclassification and because factors other than activity may influence both fitness and health through related biological factors. For example, the age and examination year–adjusted relative risks for all-cause mortality in 1263 men with type 2 diabetes are 1.8 for inactivity but 2.9 for low fitness. Obese individuals with at least moderate CRF have lower rates of cardiovascular disease (CVD) or all-cause mortality than their normal-weight but unfit peers. In fact, death rates in the former group are about one half those of the latter. These results are similar whether BMI, percent body fat, body fat mass, or waist circumference is used as the body habitus measure and are consistent for women and men and for men with type 2 diabetes. These results also pertain to nonsmokers and after exclusion of individuals with chronic disease or deaths within the first 5 years of followup. Moreover, 150 minutes a week of moderate intensity physical activity is sufficient to avoid the low-fitness category. These results are consistent with those reported in 24 studies identified in a systematic review on physical activity, CRF, obesity, and health. Two reports in this issue of JAMA examine activity and BMI as predictors of health outcomes, with somewhat divergent results. Wessel and colleagues report that women with low self-reported functional ability had higher risks of CVD outcomes than women with higher levels of fitness. In contrast, BMI and fat distribution were not associated with CVD risk. These findings are consistent with studies showing that adequate levels of activity or fitness confer health benefits for women and men in normal-weight, overweight, and obese categories. In contrast, Weinstein and colleagues report that BMI is stronger than physical activity in predicting incident type 2 diabetes and that physical activity has little effect on the relation of BMI to diabetes. They report significant inverse gradients of risk across categories of physical activity for 3 different methods of assessing activity although the associations became nonsignificant after adjustment for BMI in 2 of the 3 analyses. Key questions raised by the 2 articles in this issue of JAMA are why is the association between physical activity and incident diabetes substantially reduced when adjusted for BMI as reported by Weinstein et al and why is there little evi-