Longitudinal study of the long-term relation between physical activity and obesity in adults

Int J Obes Relat Metab Disord. 2004 Jan;28(1):105-12. doi: 10.1038/sj.ijo.0802548.


Background: Earlier observational studies of the relation between physical activity and obesity are inconsistent and ambiguous, showing a clear cross-sectional inverse relation, and a prospective association only when physical activity at the time of follow-up is included.

Objective: To examine the long-term effect of leisure time physical activity (LTPA) on subsequent development of obesity and the effect of body weight on later physical inactivity in a population-based longitudinal setting taking into account the effects of historical changes on future changes as well as pertinent confounders.

Design: The study included 3653 women and 2626 men aged 20-78 y selected at random within sex-age strata from the general population of Copenhagen. At two surveys, 5 y apart, LTPA, body mass index (BMI) (weight/height2, kg/m2), several possible confounders and modifying factors were assessed. Obesity (defined as BMI > or =30 kg/m2) and LTPA was assessed at the 3rd survey 10 y later. Odds ratios (with 95% confidence limits) for developing obesity between the last two surveys were estimated by logistic regression analysis, taking into account baseline and preceding changes in BMI and LTPA. A similar analysis of odds ratios for physical inactivity as outcome at the 3rd survey was conducted.

Results: Compared to physical inactivity, the odds ratios of development of obesity among women with medium and high level of activity were 0.81 (0.53, 1.25) and 1.16 (0.73, 1.84), respectively, and among men, the odds ratios were 1.28 (0.71, 2.33) and 1.65 (0.91, 2.99), respectively. Compared to median BMI, the odds ratio of later physical inactivity among women with high BMI was 1.91 (1.39, 2.61), and among men the odds ratio was 1.50 (1.01, 2.22). The associations were not confounded or modified by age, pre-existing diseases, smoking, alcohol intake, educational level, occupational physical activity or by familial predisposition to obesity.

Conclusion: This study did not support that physical inactivity as reported in the freely living adult population in the long term is associated with the development of obesity, but the study indicates that obesity may lead to physical inactivity.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Age Factors
  • Aged
  • Aged, 80 and over
  • Body Mass Index
  • Cross-Sectional Studies
  • Exercise / physiology*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Obesity / etiology*
  • Prospective Studies
  • Regression Analysis
  • Sex Factors