Objectives Patients suffering from spontaneous subarachnoid hemorrhage frequently require mechanical ventilation. Here, we aimed to identify factors associated with prolonged mechanical ventilation in subarachnoid hemorrhage patients and to create a new predictive score for prolonged mechanical ventilation. Design Prospective cohort study with retrospective data analysis. Setting Neurocritical care unit at a tertiary academic medical center. Patients Two hundred ninety-seven consecutive nontraumatic adult subarachnoid hemorrhage patients. Methods In patients with mechanical ventilation, we identified factors associated with mechanical ventilation greater than 48 hours, greater than 7 days, and greater than 14 days compared with mechanical ventilation less than or equal to 48 hours, less than or equal to 7 days, or less than or equal to 14 days in multivariable generalized linear models. Ventilated patients who died before 48 hours, 7 days, or 14 days and those never ventilated were excluded from the respective analysis. We incorporated those factors into a new prognostic score (the RAISE score) to predict prolonged mechanical ventilation greater than 7 days. The calculation was based on a random dataset of 60% of subarachnoid hemorrhage patients and was internally validated. Interventions None. Measurements and main results Patients were 57 years old (interquartile range, 47-68 yr) and presented with a median Hunt and Hess grade of 3 (1-5). Two hundred forty-two patients (82%) required mechanical ventilation for 9 days (2-20 d). In multivariable analysis, a higher Acute Physiology Score was associated with mechanical ventilation greater than 48 hours, greater than 7 days, and greater than 14 days, a higher Hunt and Hess grade with greater than 7 days and greater than 14 days. Early neuroimaging findings were associated with mechanical ventilation greater than 48 hours (hydrocephalus; high-grade Subarachnoid Hemorrhage Early Brain Edema Score), greater than 7 days (high-grade Subarachnoid Hemorrhage Early Brain Edema Score, co-occurrence of intracerebral bleeding) but not with prolonged mechanical ventilation greater than 14 days. The RAISE score, including age, Acute Physiology Score, Hunt and Hess grade, Subarachnoid Hemorrhage Early Brain Edema Score, and the co-occurrence of intracerebral hemorrhage accurately stratified patients by prolonged mechanical ventilation greater than 7 days (C-statistic 0.932). A RAISE score of 12 predicted 60% likelihood of mechanical ventilation greater than 7 days. Conclusions Initial disease severity and neuroimaging findings detected within 24 hours after ICU admission were associated with the need for prolonged mechanical ventilation in patients with subarachnoid hemorrhage. These results may be helpful for patient families and caregivers to better anticipate the course of therapy.
