Abstract WMP78: Investigating the Utility of Previously Developed Prediction Scores in AIS patients in the Stroke Belt

Abstract

Background: A number of scoring systems have been developed in different geographic and treatment populations, however, it remains unknown how these scores perform in the Stroke Belt population. We sought to validate and assess the utility of the SEDAN, THRIVE, HIAT and HIAT2 scoring systems among patients receiving systemic (IV tPA) and endovascular (IAT) reperfusion. Methods: We retrospectively reviewed all IV tPA and IAT patients presenting to our tertiary care center from 2009-2011. The scores were assessed in IV tPA only patients, IAT only patients and in patients who received both therapies (IV-IA). We tested THRIVE for predicting mRS 3-6, HIAT and HIAT2 for mRS 4-6, and SEDAN for symptomatic intracerebral hemorrhage (sICH). sICH was defined as a Type 2 parenchymal hemorrhage with deterioration in NIHSS score of 4 points or death. ROC curves were used to evaluate each score within the three groups. Result: Of the 366 patients who were included in this study, 89 had IAT only, 243 had IV tPA only and 34 had IV-IA. Figure 1a shows their demographic and baseline characteristics. Figure 1b shows the performance of SEDAN, THRIVE, HIAT and HIAT2 scores in our population (ROC range from 0.512-0.818). Conclusion: The two scores developed in the Stroke Belt, HIAT and HIAT2, performed well in the patient group for which they were developed (IAT). Additionally, they are also good predictors in other groups (IV and IV-IA). THRIVE performed well for predicting mRS 3-6 in all three groups. SEDAN was only moderately useful in predicting sICH after IV tPA. SEDAN had poor predictive value in IAT and IV-IA. These results highlight the need for validating clinical scores in different patient populations to determine their generalizability to all stroke patients.