Abstract

BACKGROUND:

The relationship between exercise and long-term weight loss is definitive. However, in deconditioned individuals, the basis of that relationship beyond minimal energy expenditures is unclear. Effects emanating from exercise's proposed association with changes in emotional eating and its psychosocial correlates were tested.

METHODS:

Women with class 1 (n = 67) and class 2 (n = 53) obesity, participating in a new community-based weight-loss treatment based on tenets of social cognitive theory, were assessed at baseline and Months 3, 6, 12, and 24 on weight and measures of exercise outputs, emotional eating, eating self-regulation, and negative mood. An exercise-support component was provided for 2 months prior to group nutrition-change sessions.

RESULTS:

Improvements from baseline were significant on all measures but did not significantly differ by group. Although not for the initial 3 months, changes in exercise significantly predicted weight changes over 6, 12, and 24 months from baseline. However, only 26%-31% of the lost weight was directly attributable to exercise-related energy expenditures. Changes in emotional eating significantly mediated the exercise-weight-loss relationship, and changes in self-regulation and mood significantly mediated the exercise-emotional eating relationship.

CONCLUSIONS:

Findings supported exercise's role in long-term weight loss primarily through psychosocial factors.