Frailty in Chronic Subdural Hematoma: Leveraging a Combined RAI-GCS Model Demonstrates Superior Discrimination for Predicting Non-Routine Discharge and In-Hospital Mortality

Introduction: The impact of frailty, or decreased baseline physiologic reserve, on chronic subdural hematoma (cSDH) outcomes has been explored primarily through comorbidity-based frailty indices, which inherently neglect multiple other frailty domains besides comorbidities. This study analyzes the impact of frailty, as measured by the accumulation-of-deficit-based Risk Analysis Index (RAI), on clinical outcomes for cSDH patients, and its predictive ability when paired with Glasgow Coma Scale (GCS) scores.

Methods: Hospitalizations for chronic subdural hematoma (cSDH) were identified in the National Inpatient Sample using diagnostic and procedural codes from 2019-2020. Frailty’s association with clinical outcomes, measured per RAI, was analyzed via multivariate logistic regression models. The discrimination of the RAI for clinical outcomes, stratified by cSDH severity (as measured by GCS score), was tested by computation of concordance statistics with area under the receiver operating characteristic curve analysis.

Results: A total of 7,070 cSDH patients (mean age = 74) were identified, among which 2,055 (29.1%) were robust (R) (RAI 0-26), 2,390 (33.8%), normal (N) (RAI 27-35), 1,885 (26.7%), frail (F) (RAI 36-45), and 740 (10.5%) very frail (VF) (RAI > 46). Non-routine discharge (R 30.7%, N 51.7%, F 71.1%, VF 79.7%; p< 0.001) and in-hospital mortality (R 1.9%, N 5.2%, F 7.7%, VF 18.9%; p< 0.001) significantly increased with increasing frailty thresholds. The addition of the pseudo-continuous RAI (0-81 scale) to GCS alone significantly improved model discrimination for non-routine discharge (AUC RAI-GCS 0.76 vs. AUC GCS alone 0.66, De Long p< 0.001) and in-hospital mortality (AUC RAI-GCS 0.82 vs. AUC GCS alone 0.76, De Long p< 0.001).

Conclusion : Increasing frailty, as measured by the RAI, has a stepwise, robust association with non-routine discharge, in-hospital mortality, and worse GCS scores, independent of established risk factors and treatment course. The RAI may be deployed by-the-bedside of cSDH patients, in conjunction with GCS scores, for reliable, personalized patient assessment and outcome prognostication.