Abstract W P64: DRAGON Score Outperforms 7 Other Hemorrhagic Transformation Scores in Ischemic Stroke Patients Treated With IV r-tPA
Introduction: Hemorrhagic transformation (HT) is a major complication of IV r-tPA therapy following ischemic stroke. At least 8 scores have been proposed to predict HT in patients undergoing IV thrombolytic therapy; however, studies comparing these 8 scores in an independent dataset are missing.
Hypothesis: We tested the hypothesis that stroke metrics can predict adverse outcome after r-tPA therapy.
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 Stroke Center. 8 stroke metrics were calculated for each patient: DRAGON, ASTRAL, Stroke-TPI, HAT score, MSS, SITS-ICH, SEDAN and SPAN-100. Adverse outcome was assessed by presence of symptomatic intracerebral hemorrhage (sICH). The NINDS trial defined a hemorrhage as symptomatic if it was not seen on a previous CT scan and there had subsequently been either a suspicion of hemorrhage or any decline in neurologic status. Univariate logistic regression was performed using each stroke metric as an independent variable and sICH as the dependent variable. Receiver Operating Characteristic (ROC) analysis was used to test goodness of fit by the DeLong method.
Results: 18 patients (9.6%) had a symptomatic intracerebral hemorrhage after IV rt-PA treatment. Only 4 scores showed good agreement with sICH (ROC area > 0.7): DRAGON (0.76 ± 0.12), ASTRAL (0.73 ± 0.11), Stroke-TPI (0.72 ± 0.11) and HAT score (0.71 ± 0.13) with odds ratios as follows: DRAGON (1.73 ± 0.47), ASTRAL (1.10 ±0.06), Stroke-TPI (1.04 ± 0.03) and HAT score (2.0 ± 0.67).
Conclusions: The DRAGON score yielded the highest combination of ROC area and odds ratios, and thus performed best at predicting sICH in our independent dataset. While the HAT score yielded a higher odds ratio, it yielded a lower ROC area and was therefore less robust at predicting sICH. Our results demonstrate the potential utility of the DRAGON score as a predictor of adverse clinical outcome in patients undergoing IV r-tPA. Further studies are warranted to assess its discriminatory ability in a large prospective cohort.
Author Disclosures: D.T. Asuzu: None. K.V. Nystrom: None. N. Chi: None. D.M. Greer: None. J.R. Halliday: None. C. Wira: None. J.L. Sschindler: None. K.N. Sheth: None.
- © 2014 by American Heart Association, Inc.