Randomized Trial Examining the Effect of a 12-week Exercise Program on Eating Behaviors

Abstract

Overeating and emotional eating can lead to weight gain. While exercise may help protect against weight gain, the mechanisms through which exercise aids in weight control are poorly understood. PURPOSE: This efficacy trial tests the hypothesis that exercise training impacts eating behaviors, specifically through a reduction in overeating and internal disinhibition (the tendency to eat in response to cognitive or emotional cues), in a sample of women who are overweight or obese. METHODS: Participants were inactive at baseline and self-identified as ‘stress eaters’ (eating more than usual when ‘moderately’ or ‘extremely’ stressed). They were randomized to 12 weeks of exercise training (EX) or to a no-exercise control (CON). EX participants were given an exercise goal of 200 min/wk of combined supervised and home-based exercise (30% supervised; home-based exercise was confirmed via accelerometry). No dietary instructions were provided to any participants. Assessments occurred at baseline and 12 weeks. Overeating episodes were measured over 14 days at each assessment using ecological momentary assessment (EMA; 5 surveys/day delivered randomly via smartphone). Internal disinhibition was assessed questionnaire. RESULTS: 39 participants (EX: n=19, CON: n=20) completed the study (age: 40.8±10.3 years BMI: 31.6±3.9 kg/m2). Adherence to the exercise intervention was high (99.4% of all prescribed exercise was confirmed via accelerometry) and 12-week weight change did not differ by condition (EX: -1.1±3.5% vs. 0.4±2.0%, p=0.11). At week 12, the proportion of eating episodes that were characterized as overeating episodes was 18.4% in EX vs. 24.5% in CON (p=.01). The odds of an overeating episode were lower in EX relative to CON and became more pronounced over time (condition*time=-.005, SE=.002, p=.01). Specifically at week 12, the odds of having an overeating episode among EX participants was 0.58 times the odds of having an overeating episode within CON’s. Internal disinhibition decreased in EX (pre: 4.1±2.2, post: 2.8±1.8), but not CON (4.3±2.6 to 4.2±2.3, p=.02). CONCLUSIONS: Exercise training reduced the likelihood of overeating, and eating in response to emotional or cognitive cues, in women who self-identified as stress eaters. Thus, this may be one pathway by which exercise impacts body weight.