Abstract WP62: Intracerebral Hemorrhage Score and Glasgow Coma Scale Are Not Accurate Predictors of Subsequent Mechanical Ventilation in Initially Non-Intubated Intracerebral Hemorrhage Patients

Introduction: Intracerebral hemorrhage (ICH) often requires mechanical ventilation (MV), increasing patient risks and hospital costs. Two standard tools were evaluated for their ability to predict patients needing subsequent MV. A predictive tool could reduce costs and complications from unnecessary MV and reduce risks for patients requiring non-elective intubation. Methods: Spontaneous ICH patients were evaluated at a comprehensive stroke center, focusing on those not intubated upon arrival or in the Emergency Department. Retrospective chart review included ICH score (ICHS), Glasgow Coma Scale (GCS), MV rates, and timing of intubation. To assess the ability of ICHS and GCS to predict subsequent MV within two weeks, we used the area under the receiver operating characteristic curve (AUC) and Youden's J statistic (range 0-1), which combines sensitivity and specificity. Higher values indicate better test performance. Results: Over 4 years, 700 consecutive ICH patients were evaluated, with 231 meeting analysis criteria. Of these, 20/231 (8.66%) received subsequent MV within the following two weeks. Neither the ICHS nor GCS accurately predicted subsequent MV, with AUC/Standard Error/Youden's J of 0.630/0.066/0.254 for ICHS and 0.643/0.070/0.235 for GCS. Specificity was >97% (95% Confidence Interval: 94-99%) for both an ICHS > 4 and a GCS < 8; however, corresponding sensitivities were low (5% and 10%, respectively). Conclusions: The ICHS and GCS are not accurate predictors of subsequent MV, as shown by the low AUC and Youden's J. Both scales are >97% specific when the ICHS is > 4 or GCS is < 8, which may identify high-risk patients but would overlook others. Development of an intubation-specific tool, further exploration of alternative existing ICH scores, and evaluation of a larger sample size are needed.