Transitions in frailty states and associated factors: a multistate analysis of the Italian Longitudinal Study on Aging population-based cohort

Background: Frailty is recognized as a dynamic and potentially reversible process, but comprehensive studies on its progression/regression are rare. Objective: To investigate the frequency and characteristics of frailty transitions over time in a representative sample of older Italians. Design and participants: As secondary analysis of the Italian Longitudinal Study on Aging (ILSA) population-based cohort, we studied all participants (n = 1339; women 47.5 %, age 72.7 ± 5.1) with complete information on changes in frailty status (or death) between consecutive ILSA surveys (T0, T1, T2). Measurements: Frailty was operationalized according to Fried phenotype, analysing transitions between frailty, or death, during T0-T1, T1-T2 (4-, 5-year length). Transition probability at 1, 3, 5 years was estimated through non-hidden continuous-time Markov models, with death as absorbing state. Factors influencing transitions were evaluated with Cox proportional Hazard Ratios (HR). Results: We observed 1931 transitions between frailty states and 241 to death. The estimated probability of: maintaining a stable frailty status (∼80 % within 1 year) halved at 5 years; worsening increased steeply over time and was always greater among women; improvement/remission was twice higher at medium (about 20 % among Frail->preFrail women, preFrail->nonFrail men) than short term. Depressive symptoms were the strongest predictor of worsening [nonFrail->Frail: women HR 3.63 (95 %CI 1.45-9.10), men HR 3.78 (95 %CI 2.0-7.13)]. Not having a spouse/partner was associated with a 30 % reduced probability of pre-frailty remission in both sexes. Conclusions: Our findings confirm the fluctuating nature of frailty with an ample chance of remission/improvement, highlighting the importance of a prompt, multidimensional preventive approach, including psycho-social dimensions.