Introduction:
Several scores have been proposed to predict hemorrhagic transformation (HT) after IV r-tPA therapy. However, their utility in predicting adverse outcome has been limited to assessment of 30- or 90-day Modified Rankin Scale (mRS) scores, which are not readily available at many stroke centers. Studies evaluating discharge mRS scores as a predictor of long-term adverse patient outcome are lacking.
Hypothesis:
We tested the hypothesis that HT scores predicting long-term adverse outcome after r-tPA therapy correlate with readily available discharge mRS scores.
Methods:
Clinical data was analyzed from consecutive patients (n = 187) receiving IV r-tPA therapy from January 2009 until May 2013 at the Yale-New Haven Hospital. 8 HT scores were calculated for each patient: DRAGON, ASTRAL, Stroke-TPI, HAT Score, MSS, SITS-ICH, SEDAN and SPAN-100. Univariate logistic regression was performed using each HT score as an independent variable and discharge mRS≥5 or discharge mRS≥4 as the dependent variable. Receiver Operating Characteristic (ROC) analysis was used to test goodness of fit by the DeLong method.
Results:
All metrics except the HAT Score and SPAN-100 showed good agreement with discharge mRS≥5 (ROC area > 0.7). The 3 HT scores showing the best agreement with discharge mRS≥5 were Stroke-TPI (ROC 0.85 ± 0.07), ASTRAL (ROC 0.85 ± 0.06) and DRAGON (0.83 ± 0.07), with odds ratios: Stroke-TPI (1.09 ± 0.03), ASTRAL (1.17 ± 0.06) and DRAGON (2.23 ± 0.58). These 3 HT scores still performed best when the discharge mRS cutoff was relaxed to mRS ≥ 4, with similar odds ratios but modestly reduced ROC areas: ASTRAL (ROC 0.82 ± 0.06), DRAGON (ROC 0.79 ± 0.07) and Stroke-TPI (ROC 0.76 ± 0.07).
Conclusions:
Most established HT scores in our study showed good agreement with the discharge mRS score using two separate cutoff values. Therefore, readily available discharge mRS scores may be a useful predictor of adverse outcome following IV r-tPA therapy. These results warrant larger prospective studies to establish associations between discharge mRS scores and 90-day patient outcomes.