Early predictors of outcome in patients receiving hypervolemic and hypertensive therapy for symptomatic vasospasm after subarachnoid hemorrhage

Objective: Symptomatic vasospasm after subarachnoid hemorrhage (SAH) is ass ociated with a high incidence of permanent disability and death. For early identification of patients who are at risk for poor outcome, we determined the predictors of outcome in patients with symptomatic vasospasm after SAH. Design: We retrospectively determined the prognostic value of clinical char acteristics and computed tomographic scan both at admission and at the time of initiation of hypervolemic and hypertensive therapy. Settings: Neurosciences critical care unit at a University hospital. Patients: A total of 70 consecutive patients who developed symptomatic vaso spasm after SAH. Intervention: Treatment with oral nimodipine, hypervolemic therapy, and hyp ertensive therapy. Angioplasty and intra-arterial papaverine were used in p atients with vasospasm resistant to standard treatment. Measurements and Main Results: Poor outcome, defined as Glasgow Outcome Sca le Score of 3-5 at 2 months or discharge, was observed in 32 (46%) patients . In the logistic regression analysis, a Glasgow Coma Scare (GCS) score of less than or equal to 11 (odds ratio, 11.0; 95% confidence interval, 3.6-39 .3) and hydrocephalus (odds ratio, 4.3; 95% confidence Interval, 1.2-18.2) at the time of initiation of hypervolemic and hypertensive therapy were sig nificantly associated with poor outcome. Poor outcome was observed in 91% o f the patients who had both a GCS score of less than or equal to 11 and hyd rocephalus compared with 15% of patients with a GCS score of >11 and no hyd rocephalus at the time of initiation of hypervolemic and hypertensive thera py, a GCS score of less than or equal to 11 was also independently associat ed with length of intensive care unit stay (F ratio = 18.0; p = .0011) and hospital stay (F ratio = 9.2; p = .0034) after initiation of hypervolemic a nd hypertensive therapy. Conclusions: The results of this study suggest that outcome in patients wit h symptomatic vasospasm can be effectively predicted by routinely available information, including GCS scare at the time of initiation of hypervolemic and hypertensive therapy. This information can be used for selection and s tratification of patients in future treatment studies of patients with symp tomatic vasospasm.