Preliminary Results of External Validation of the RAISE Score for Prediction of Prolonged Mechanical Ventilation of Subarachnoid Hemorrhage Patients

Introduction: The RAISE score was designed in Austria to predict whether patients with non-traumatic subarachnoid hemorrhage (SAH) would require prolonged mechanical ventilation, defined as >7 days, with higher scores predicting a higher probability of prolonged ventilation. As prolonged mechanical ventilation leads to worse outcomes, we sought externally validate the RAISE score.

Methods: In this United States-based single-institution retrospective cohort study of patients admitted to the intensive care unit from June 2017-March 2022 with spontaneous non-cortical SAH, we collected patient demographics, radiographic parameters, SAH severity based on Hunt-Hess grade, and other variables of the RAISE score to determine whether the model was externally valid in our patient population. The RAISE score was used to determine binary outcomes (prolonged mechanical ventilation or not) using different thresholds. We calculated the sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC).

Results: A total of 189 patients met study criteria. In bivariate analysis, Fisher score, Hunt-Hess grade, SAH early brain edema score, aneurysmal etiology, hydrocephalus requiring treatment, and endovascular coiling were associated with mechanical ventilation >7 days. Although a RAISE score of 12 predicted >50% chance of prolonged mechanical ventilation in the original study, it corresponded to an 30% chance in our cohort. However, higher RAISE scores did generally correlate with a greater probability of prolonged ventilation. In testing different thresholds of the RAISE score for predicting whether a patient would require prolonged mechanical ventilation or not as a binary outcome, the AUROC was 0.89. Utilizing a score of 12, sensitivity was 82.9%, and specificity was 69.6%.

Conclusion : Higher RAISE scores do predict prolonged mechanical ventilation in SAH patients. However, the threshold score for predicting prolonged mechanical ventilation may vary between institutions. It is unclear whether use of the RAISE score can help improve outcomes or is simply a metric to compare patients across institutions.