Introduction Frailty is known as a factor associated with risk of disability in elderly.

Objectives To examine the association between frailty status and incidence of activities of daily living (ADL) disability in elderly.

Methods Data comes from a longitudinal survey—SABE (Health, Well-being and Ageing), which began in 2000 with a multistage clustered sample of 2143 people aged ≥60 years-old in Sao Paulo-Brazil. Frailty components (Fried's model) were included in 2006, during the second wave (n=1115 re-interviewed). In 2006, used here as baseline, a subsample of 688 persons aged ≥75 years was followed-up, in 2 rounds (2008–2009). Measures included: frailty status in baseline; incidence of ADL disability (baseline-2009). We tested two adjusted logistic multiple regression models. Inferences were weighted to account for sample design.

Results In 2006, 389 elderly had no ADL limitation. In 2009, after excluding those who scored <19 in Mini-Mental State Exam, ADL were analysed in a sample of 242 elderly. The disability incidence was 91.5/1000 person-years in robust, 118.6/1000 person-years in pre-frail, and 223.2/1000 person-years in frail elderly. Frailty is associated to the risk of developing disability (OR=3.32; p=0.015). In model 1, adjusted for age, sex and education, frailty remains significantly associated (OR=2.54; p=0.045). In model 2, after adding depression, falls, BMI, and mobility limitation, frailty lost significance, but still shows risk (OR=2.61; p=0.066).

Conclusions Incidence rate of ADL limitation was greater in frail elderly, but other factors, as depression and falls, are also important and should be considered.