We developed a modification of the Fisher computed tomographic rating scale and compared it with the original Fisher scale to determine which scale best predicts symptomatic vasospasm after subarachnoid hemorrhage.We analyzed data from 1355 subarachnoid hemorrhage patients in the placebo arm of four randomized, double-blind, placebo-controlled studies of tirilazad. Modified Fisher computed tomographic grades were calculated on the basis of the presence of cisternal blood and intraventricular hemorrhage. Crude odds ratios (OR) reflecting the risk of developing symptomatic vasospasm were calculated for each scale level, and adjusted ORs expressing the incremental risk were calculated after controlling for known predictors of vasospasm.Of 1355 patients, 451 (33%) developed symptomatic vasospasm. For the modified Fisher scale, compared with Grade 0 to 1 patients, the crude OR for vasospasm was 1.6 (95% confidence interval [CI], 1.0-2.5) for Grade 2, 1.6 (95% CI, 1.1-2.2) for Grade 3, and 2.2 (95% CI, 1.6-3.1) for Grade 4. For the original Fisher scale, referenced to Grade 1, the OR for vasospasm was 1.3 (95% CI, 0.7-2.2) for Grade 2, 2.2 (95% CI, 1.4-3.5) for Grade 3, and 1.7 (95% CI, 1.0-3.0) for Grade 4. Early angiographic vasospasm, history of hypertension, neurological grade, and elevated admission mean arterial pressure were identified as risk factors for symptomatic vasospasm. After adjusting for these variables, the modified Fisher scale remained a significant predictor of vasospasm (adjusted OR, 1.28; 95% CI, 1.06-1.54), whereas the original Fisher scale was not.The modified Fisher scale, which accounts for thick cisternal and ventricular blood, predicts symptomatic vasospasm after subarachnoid hemorrhage more accurately than original Fisher scale.