Physical activity as a predictor of absence of frailty in the elderly

Rev Assoc Med Bras (1992). May-Jun 2012;58(3):341-7.
[Article in English, Portuguese]

Abstract

Objective: To analyze physical activity in different domains (work, transportation, housework, and leisure) as a predictor of the absence of frailty.

Methods: An epidemiological cross-sectional study with a random sample of 622 individuals aged > 60 years was carried out in Uberaba, MG, Brazil. Receiver operating characteristic (ROC) curves were generated and compared with physical activity in different domains and the absence of frailty. Cutoffs points of physical activity (minutes/week) were established to predict the absence of frailty. A confidence interval of 95% was established in order to find the largest areas under the ROC curves for work, transportation, household, and leisure. The total physical activity and its domains were assessed by the international physical activity questionnaire. The index of frailty (unintentional weight loss, functional limitation to chair rise, handgrip strength, physical activity, and exhaustion) was based on the study by Fried; the participants were dichotomously classified as not frail and frail.

Results: The prevalence of frailty was 19.7% (men) and 20% (women). Physical activity of moderate or vigorous intensity accumulated in different domains for 145 minutes/week for women and 140 minutes/week for men or 85 minutes/week for women and 112.5 minutes/week for men for activities in the leisure domain showed the best cutoff to predict the absence of frailty.

Conclusion: The practice of physical activity, especially in leisure time or accumulated in different domains, contributes to the prevention of frailty in the elderly.

MeSH terms

  • Activities of Daily Living*
  • Aged
  • Aged, 80 and over
  • Brazil / epidemiology
  • Cross-Sectional Studies
  • Female
  • Frail Elderly / statistics & numerical data*
  • Humans
  • Leisure Activities*
  • Life Style*
  • Male
  • Middle Aged
  • Motor Activity*
  • Prevalence
  • ROC Curve
  • Socioeconomic Factors
  • Work / statistics & numerical data*