RESUMO
BACKGROUND: Frailty phenotype has been extensively modified. Among the five criteria, the low physical activity (PA) is often changed, however, it is still uncertain how this modification might impact frailty classification. AIMS: To examine the variance in the prevalence of frailty by modifying PA criterion using different cut-points of both subjective and objective measures, and to determine the agreement between these on classifying individuals with low PA. Finally, a surrogate PA criterion of frailty phenotype was proposed using objectively measured moderate-to-vigorous physical activity (MVPA). METHODS: This cross-sectional study comprised a convenience sample of 135 community-dwelling older adults. Frailty was evaluated using a modified frailty phenotype. PA was assessed using International Physical Activity Questionnaire-short form (IPAQ-SF) and objectively measured using a uniaxial accelerometer for 7 days. Four different low PA criteria were created and compared (population dependent and independent cut-points) using subjective and objective measures. RESULTS: Different measures and cut-points resulted in an overall variation of 12.5% on frailty prevalence. The agreement in the categorization of participants with low PA between population dependent cut-points of both IPAQ-SF and accelerometer was none to slight (%Overallagreement = 43.70%; Kappa = 0.082, p = 0.114). Results from ROC curve analysis showed an optimal threshold of 15.13 min/day of MVPA to discriminate between non-frail and pre-frail individuals. CONCLUSION: Modifications of the low PA criterion of frailty phenotype can greatly impact frailty classification. MVPA measured through an accelerometer may present a possible solution to standardize this criterion, and improve frailty screening and between-studies comparability.