A grading scale to predict outcomes after intra-arterial thrombolysis for stroke complicated by contrast extravasation

OBJECTIVE: Contrast extravasation after intra-arterial thrombolysis for str oke occurs frequently and is identifiable on a computed tomographic (CT) sc an, but it is often unrecognized or misdiagnosed. Few articles describing t his phenomenon have been published. The clinical outcomes of patients after contrast extravasation are poorly understood. We designed a grading system to predict outcomes after contrast extravasation and tested the grading sc ale prospectively. METHODS: We studied 27 patients who had contrast extravasation exhibited on a CT scan immediately after intra-arterial thrombolysis. The National Inst itutes of Health Stroke Scale was used to quantify neurological examination s preoperatively, postoperatively, and at follow-up an average of 3 months later. A grading scale from 0 to 10 was developed from a retrospective anal ysis of the first 18 patients using odds ratios and Fisher's exact test. Th e grading system was then applied prospectively to the next 9 consecutive p atients. RESULTS: Six components of the grading system were weighted approximately p roportional to corresponding odds ratios: 1) incomplete recanalization (3 p oints), 2) prolonged angiographic blush (2 points), 3) hyperdensity greater than 150 Hounsfield units (2 points), 4) lesion volume greater than 50 cc exhibited on a CT scan (1 point), 5) lesion in eloquent parenchyma (1 point ), and 6) hypodensity demonstrated on an immediate postoperative CT scan (1 point). The contrast extravasation grades for each outcome category (excel lent, fair, poor, died) increased in stepwise fashion, There was a direct l inear correlation between the assigned grade and National Institutes of Hea lth Stroke Scale score improvement at follow-up. CONCLUSION: This grading system should prove useful as a preliminary guide for predicting outcomes of patients with contrast extravasation after intra -arterial thrombolysis for stroke. Further analysis in a large cohort of pr ospective patients is necessary to ensure extensibility.