Frailty, clinical outcomes, and treatment effects of a care bundle in acute intracerebral haemorrhage: a post-hoc analysis of the INTERACT3 trial

Background: Frailty has been widely associated with poor outcomes in mixed stroke cohorts, but evidence specific to intracerebral haemorrhage, particularly from settings in low-income and middle-income countries and regarding treatment-effect modification, remains scarce. We aimed to evaluate the association between frailty and several clinical outcomes and to ascertain whether frailty modifies the treatment effect of a care bundle in individuals with acute intracerebral haemorrhage. Methods: We conducted a post-hoc analysis of the third Intensive Care Bundle with Blood Pressure Reduction in Acute Cerebral haemorrhage (INTERACT3) international, stepped-wedge, cluster-randomised controlled trial (NCT03209258; Chinese Clinical Trial Registry ChiCTR-IOC-17011787), which enrolled individuals with acute spontaneous intracerebral haemorrhage at 121 hospitals in nine low-income and middle-income countries and one high-income country. Frailty was quantified using a 30-item cumulative-deficit frailty index; individuals were categorised as non-frail (frailty index ≤0·10), pre-frail (frailty index >0·10 to <0·21), or frail (frailty index ≥0·21). The primary outcome was 6-month all-cause mortality. Mixed-effects regression models, adjusted for key prognostic variables and clustering by hospital site, were used to estimate associations and treatment interactions across frailty strata. Findings: Of the 7036 individuals enrolled in INTERACT3, 7035 with available frailty data were included in this analysis (mean age 62·0 years; 2533 [36·0%] women and 4502 [64·0%] men). 3185 (45·3%) were non-frail, 3000 (42·6%) were pre-frail, and 850 (12·1%) were frail. Increasing frailty was associated with higher 6-month mortality and poorer functional outcomes. These associations were attenuated but persisted after multivariable adjustment (mortality: adjusted hazard ratio [HR] 1·48 [95% CI 1·20–1·83]; major disability among survivors: adjusted odds ratio 1·73 [1·40–2·14] for frail vs non-frail individuals). The care bundle was associated with lower mortality point estimates across frailty groups (non-frail adjusted HR 0·66 [95% CI 0·46–0·94], pre-frail adjusted HR 0·80 [0·61–1·05], and frail adjusted HR 0·65 [0·43–0·98]; p for interaction=0·30). Absolute mortality rate reductions were larger among individuals with frailty (116·9 events per 1000 person-years) than among individuals without frailty (44·1 events per 1000 person-years). No clear evidence of treatment-by-frailty interaction was observed for death or major disability at 6 months (p for interaction=0·27). Interpretation: Frailty showed a graded association with poor prognosis in acute intracerebral haemorrhage, and these associations were attenuated but generally persisted after adjustment for age, baseline neurological severity, and other prognostic factors. We found no clear evidence that frailty modified the effect of the INTERACT3 care bundle on mortality. Mortality estimates favoured the care bundle across frailty strata, with larger absolute mortality reductions observed among individuals with frailty. These findings support offering protocolised acute care to individuals with imaging-confirmed spontaneous intracerebral haemorrhage who present within 6 h of symptom onset, irrespective of frailty status, while also maintaining individualised clinical judgement regarding prognosis and functional recovery. Funding: None.