Commentary: Luke and Cooper are wrong: physical activity has a crucial role in weight management and determinants of obesity

Introduction

Luke and Cooper claim that evidence does not support a relationship between physical activity (PA) and obesity, and that policy recommendations to control the obesity epidemic should be limited to ‘altering the modern food environment’.¹ We disagree and contend that to support their position Luke and Cooper misrepresent and/or ignore an extensive evidence base of observational and experimental studies that clearly support an effect of PA on obesity, as follows.

Secular trends in physical activity

In direct contradiction to claims that secular trends in PA are not consistent with trends in obesity prevalence and that the ‘labour-saving culture was fully in place by the 1960s–70s’,¹ Church et al. examined US occupational PA from more than 140 000 businesses and government agencies, and >440 000 individual worksites. They report a significant decreasing trend in occupational energy expenditure over the past 50 years. Mathematical modelling showed that decrements in occupational PA were 140 and 120 kcal/day in men and in women, respectively. Their model accurately predicted changes in body mass index (BMI) (National Health and Nutrition Examination Surveys) for men and women from 1971 to 2006.² We examined household energy expenditure over the past 45 years and found decrements of >1800 kcal/week for all women and >2500 kcal/week in non-employed women.³ These results combined with decreases in active transport⁴ suggest that the decrements in PA energy expenditure in multiple domains surely must be an important cause of population increases in weight. It is implausible that the ‘labour-saving culture’¹ has not changed since the 1960s–70s.

Observational cohort studies

There are numerous longitudinal studies showing that PA prevents weight gain. One excellent example by Hankinson et al. examined 3554 men and women with seven follow-up measurements over 20 years and concluded that maintaining high activity levels lessens gains in weight and visceral adiposity.⁵ Women and men who maintained PA at recommended levels had smaller increases in both BMI and waist circumference when compared with inactive individuals.

Ecological studies

Perhaps the gravest misrepresentation by Luke and Cooper is citation of data from the Cuban economic crisis of 1989–2000.⁶ The abstract from this report states that ‘during the crisis period, the proportion of physically active adults increased from 30% to 67%, and a [downward] 1.5-unit shift in the body mass index distribution was observed’.⁶ We agree that this is a ‘rare contemporary instance’,¹ but the data presented clearly show an increase in PA and a substantial weight loss during this period. We do not assert that this increase in PA is the sole reason for the decline in BMI, and think that the decline in food availability also had an effect. However, it is nonsensical to claim that the BMI change was due entirely to declines in per capita food availability and that the substantial increases in PA had no role.

Experimental evidence

In direct contrast to Luke and Cooper’s¹ argument that experimental evidence does not support physical activity as an effective intervention for weight loss, there is a significant body of experimental evidence indicating that physical activity affects weight and fat mass losses and weight management efforts. This evidence arises from multiple controlled, randomized trials consisting of exercise applied independently or in conjunction with dietary interventions.

The effectiveness of exercise as an intervention for weight loss can be illustrated no better than in Luke and Cooper’s references^7,8 which they use to support their contention that increasing physical activity does not lead to weight loss. These studies demonstrate significant weight loss, decreased waist circumference and/or improvements in body composition with exercise. Perhaps the strongest evidence is from the 16-month intervention in the Midwest Exercise Trial.⁸ Contrary to the Luke and Cooper position that no weight loss was reported, Donnelly and colleagues show that exercising men, when compared with controls, reduced their body weight (−4.8 kg, P = 0.01), BMI and fat mass with no reductions in fat-free mass. Whereas the exercising women remained weight stable, women in the control group gained weight, with a significant difference between groups (−5.2 kg, P = 0.03). Exercising women also had decreased fat mass in contrast to the increased fat mass in the control group. Of importance, there were no reported changes in total energy or macronutrient intake by either exercising group. So the weight loss in men and prevention of weight gain in women, and sparing of lean body tissues in both sexes, were likely due to the increased physical activity.

Finally, Luke and Cooper propose that ‘exercise plus calorie restriction achieves virtually the same amount of weight loss as calorie restriction alone.’ We agree that physical activity does not add substantial weight loss to that resulting from severe calorie restriction of ≥500 kcal/day. However, recent literature reviews and meta-analyses conclude that at equivalent energy expenditure the combination of exercise and modest calorie restriction is more effective for weight loss than calorie restriction alone.^9,10

To summarize, evidence from controlled experimental studies shows that increased energy expenditure through physical activity is effective for improving body composition via reductions in body weight and fat mass, while sparing lean tissue mass. Further, the combination of increased energy expenditure and moderate calorie restriction is more effective for weight management than calorie restriction alone, whereas the additive effect may be overshadowed in interventions utilizing severe and unsustainable caloric restriction.

Conclusion

We agree that the causes of the obesity epidemic are multiple and that weight regulation is complex and not well understood. Interrelationships among dietary factors and PA habits require further study. However, there is a large body of evidence supporting the hypotheses that declines in PA from multiple domains contribute to the obesity epidemic^2,3 and that moderately active persons gain less weight over time than those who are sedentary,⁵ and there is strong, unequivocal evidence from multiple well-controlled randomized trials showing that PA has a role in weight management, especially in preserving lean mass and reducing fat mass.⁸

Funding

Drs. Blair, Archer, and Hand are funded via unrestricted research grants from The Coca-Cola Company for analyses of dietary trends and for an energy balance study. Additional funding for Dr. Blair: NHLBI (U01AG022376-S01); and for Drs. Blair and Hand: NIH/NINR (1R21NR011281)

Conflict of interest: None declared.