Frailty status, functional recovery and survival trajectories following neurocritical illness in older adults: a retrospective cohort study

Frailty is an important prognostic marker in critically ill older adults, but its association with recovery after neurocritical illness remains incompletely characterised. We evaluated the association between pre-admission Clinical Frailty Scale (CFS) score and six-month neurological outcome and explored mortality, frailty, and residential trajectories. We conducted a retrospective single-centre cohort study of consecutive adults aged ≥ 65 years admitted to a tertiary neurocritical care unit between 2020 and 2025. Pre-admission CFS was categorised as 1–3 (non-frail), 4–5 (very mild/mild frailty), and ≥ 6 (moderate-to-severe frailty). The primary outcome was favourable neurological recovery at six months (modified Rankin Scale [mRS] ≤ 3). For this outcome, adjusted ORs were oriented towards favourable recovery and reported per one-point decrease in CFS. Models adjusted for age, admission Glasgow Coma Scale, pupillary status, and diagnosis; a sensitivity model used SAPS II instead of age and GCS. Secondary analyses were exploratory. Among 404 patients, 277 (68.6%) had CFS 1–3, 98 (24.3%) had CFS 4–5, and 29 (7.2%) had CFS ≥ 6. Six-month mRS was available for 386 patients; favourable outcome occurred in 70/262 (26.7%), 9/95 (9.5%), and 1/29 (3.4%), respectively (global p < 0.001). Each one-point decrease in CFS was associated with higher adjusted odds of favourable outcome (OR 1.92, 95% CI 1.47–2.56; p < 0.001; n = 383), whereas each one-point increase was associated with higher six-month mortality (HR 1.22, 95% CI 1.10–1.34; p < 0.001). The CFS-plus-GCS-plus-pupillary-reactivity model yielded an AUC of 0.81 (95% CI 0.76–0.86). Higher pre-admission CFS scores were associated with a lower probability of favourable neurological recovery and with higher six-month mortality in this single-centre cohort. These findings describe group-level associations and require prospective multicentre validation, particularly because the CFS ≥ 6 subgroup was small and neurocritical diagnoses were heterogeneous.