Background: Many overweight and obese individuals use exercise when attempting to lose weight. However, the improvements in weight and body composition are often far less than expected. Levels of physical activity outside of the structured exercise program are believed to change and may be responsible for the unsuccessful weight loss.

Objective: The purpose of this meta-analysis was to provide a quantitative estimate of the change in non-exercise physical activity (NEPA) during exercise interventions.

Methods: All studies included in the meta-analysis were peer-reviewed and published in English. Participants were randomized to a non-exercise comparison group or exercise training group with an intervention lasting ≥2 weeks. NEPA was measured at baseline and at various times during the study. Hedges' d effect size (ES) was used to adjust for small sample bias, and random-effects models were used to calculate the mean ES and explore potential moderators.

Results: The cumulative results of 44 effects gathered from ten studies published between 1997 and 2015 indicated that NEPA did not change significantly during exercise training (ES = 0.02, 95% confidence interval [CI] -0.09 to 0.13; p = 0.723). Duration of the exercise session (β = -0.0039), intervention length (β = 0.0543), and an age × sex (β = -0.0005) interaction indicated that the increase in NEPA may be attenuated in older women during exercise training and during shorter exercise interventions with longer sessions (all p < 0.005).

Conclusion: On average, no statistically or clinically significant mean change in NEPA occurs during exercise training. However, session duration and intervention length, age, and sex should be accounted for when designing exercise programs to improve long-term sustainability and improve the likelihood of weight loss success, as the initial decrease in NEPA appears to dissipate with continued training.