Predicting Prognosis of Intracerebral Hemorrhage (ICH): Performance of ICH Score Is Not Improved by Adding Oral Anticoagulant Use

Rik Houben, Floris H. B. M. Schreuder, Kim J. Bekelaar, Danny Claessens, Robert J. van Oostenbrugge, Julie Staals

The intracerebral hemorrhage (ICH) score is a commonly used prognostic model for 30-day mortality in ICH, based on five independent predictors (ICH volume, location, Glasgow Coma Scale, age, and intraventricular extension). Use of oral anticoagulants (OAC) is also associated with mortality but was not considered in the ICH score. We investigated (a) whether the predictive performance of ICH score is similar in OAC-ICH and non-OAC-ICH and (b) whether addition of OAC use to the ICH score increases the prognostic performance of the score. We retrospectively selected all consecutive adult non-traumatic ICH cases (three hospitals, region South-Limburg, the Netherlands 2004-2009). Mortality at 30 days was recorded. Using univariable and multivariable logistic regression, association between OAC use and 30-day mortality was tested. Then (a) we computed receiver operating characteristic (ROC) curves for ICH score and determined the area under the curve (AUC) in OAC-ICH and non-OAC-ICH. Then (b) we created a New ICH score by adding OAC use to the ICH score. We calculated correlation between 30-day mortality and ICH score, respectively, New ICH score using Spearman correlation test. We computed ROC curves and calculated the AUC. We analyzed 1,232 cases, 282 (22.9%) were OAC related ICH. Overall, 30-day mortality was 39.3%. OAC use was independently associated with 30-day mortality (OR 2.09, 95% CI, 1.48-2.95; p < 0.001), corrected for the five predictors of the ICH score. The ICH score performed slightly better in non-OAC-ICH (AUC 0.840) than in OAC-ICH (AUC 0.816), but this difference was not significant (p = 0.39). The ICH score and the New ICH score were both significantly correlated with 30-day mortality (rho 0.58, p < 0.001 and 0.59, p < 0.001, respectively). The AUC for the ICH score was 0.837, for New ICH score 0.840. This difference was not significant. The ICH score is a useful tool for predicting 30-day mortality both in patient who use and patients who do not use OAC. Although OAC use is an independent predictor of 30-day mortality, addition of OAC use to the existing ICH score does not increase the prognostic performance of this score.