Background: Little is known about whether initiating physical activity induces change in other health-related behaviors. If other behaviors do change with increasing physical activity, this would complicate interpretation of differences in study outcomes in exercise intervention trials.
Design: Randomized controlled trial.
Setting/participants: 173 sedentary, overweight (body mass index between 24.0 and 25.0 kg/m2 with body fat>33% or BMI>or=25.0 kg/m2), postmenopausal women, ages 50 to 75 years, not using hormone therapy, and living in the Seattle, WA area.
Intervention: Participants were randomly assigned to an exercise intervention (n=87) or a stretching-control group (n=86). The exercise intervention included facility and home-based moderate-intensity exercise.
Main outcome measures: Changes in dietary intake, alcohol consumption, and medication and supplement use were compared from baseline to 3- and 12-month follow-up between exercise and control groups, and by tertiles of exercise adherence. Data were collected between January 1998 and July 2001.
Results: In general, changes in dietary intake between the exercise and control group were not statistically different. The exercise group had a greater increase in the proportion of participants who used multivitamins (+5%) compared to the control group (-10%) at 3 months (p-interaction=0.04), but not at 12 months (p-interaction=0.58). Furthermore, there were few differences when comparing changes in health behaviors across exercise adherence tertiles.
Conclusions: Our results suggest that participation in a year-long exercise intervention trial among post-menopausal women has little effect on other health behaviors. These findings suggest that additional behavior changes in exercise trials are minimal and unlikely to bias primary study results.