Physical and brain frailty in stroke: shared occurrence and outcomes. A cohort study

Abstract

Background:There is increasing interest in the concept of frailty in stroke, including both physical frailty and imaging evidence of brain frailty. We aimed to 1 establish concurrent validity of a brain frailty measurement against traditional measures of physical and global frailty 2) establish prevalence of brain frailty in stroke survivors with and without pre-existing frailty 3) establish the predictive validity of measures of physical, global, and brain frailty against long-term cognitive outcomes.<break><break>Methods:We included consecutively admitted stroke or transient ischaemic attack (TIA) survivors from participating stroke centres. Baseline CT scans were used to identify leukoaraiosis, atrophy, or old infarcts; these were then combined to generate an overall brain frailty score for each participant on a scale ranging from 0-3 (3=higher brain frailty). Global frailty was measured using Rockwood frailty index, and physical frailty using a Fried frailty screening tool. Presence of major or minor neurocognitive disorder at 18-months following stroke was established via a multicomponent assessment. We assessed the concurrent validity of brain frailty and frailty scales via Spearmans rank correlation. Prevalence of brain frailty was established based upon observed percentages within groups defined by frailty status (robust, pre-frail, frail). We conducted multivariable logistic regression analyses to evaluate association between brain, global, and physical frailty with 18-month cognitive impairment.<break><break>Results:Three-hundred-forty-one stroke survivors participated. Brain frailty was weakly correlated with Rockwood frailty(Rho:0.336;p<0.001) and with Fried frailty(Rho:0.230;p<0.001). Three-quarters of people who were frail had moderate-severe brain frailty and prevalence increased according to frailty status. Brain frailty(OR:1.64,95%CI=1.17- 2.32), Rockwood frailty (OR:1.05,95%CI=1.02-1.08) and Fried frailty(OR:1.93,95%CI=1.39-2.67) were each independently predictive of cognitive impairment at 18 months following stroke. Fried frailty was predictive independent of age, sex, stroke severity, education, baseline cognitive test performance, and brain frailty status (OR:1.49,95%CI=1.03-2.18)<break><break>Conclusions:Physical and brain frailty are separate concepts, although they frequently co-exist. Both are associated with adverse cognitive outcomes and physical frailty remains important when assessing cognitive outcomes.