Natural Regression of Frailty Among Community-Dwelling Older Adults: A Systematic Review and Meta-Analysis

Background and Objectives Frailty is a dynamic process with potential transitions over time. However, there is limited understanding of the patterns of frailty improvement. We conducted a systematic review and meta-analysis to estimate the natural rate of frailty regression among community-dwelling older adults aged at least 60 years. Research Design and Methods Systematic searches for studies reporting frailty improvement were performed in 5 databases (Medline, Embase, CINAHL plus, Web of Science, and PsycINFO) from inception until January 2019. Results Twenty-five studies from 26 countries were included. Among a baseline population of more than 50,000 individuals, the pooled prevalence of pre-frailty and frailty was 50.5% (95% confidence interval [CI] 47.8-53.3) and 12.8% (95% CI 9.1-17.0), respectively. During a median follow-up of 3.0 (range 1-10.0) years, 23.3% of surviving pre-frail individuals regressed to a robust state and 35.2% of surviving frail individuals reversed to a pre-frail or robust state. The pooled remission rates among people with pre-frailty and frailty were 80.4 (95% CI 61.7-104.6) and 135.3 (95% CI 98.1-186.5) per 1,000 person-years, respectively. Frailty and pre-frailty improvement rates varied by sex, diagnostic criteria, study region, and follow-up duration. The remission rates were significantly reduced when accounting for progressions to death. The heterogeneity of included studies was high which reflected considerable differences in methodological approach. Discussion and Implications Although frailty is highly prevalent in older people, natural remission is possible and common. Improved understanding of the factors that confer increased likelihood of frailty regression may support the design of interventions to reduce the burden of frailty.