A nomogram by adding atrial fibrillation to multicenter stroke survey score to predict intracranial hemorrhage in stroke patients undergoing thrombolysis

Abstract Background: Identification of stroke patients at risk of postthrombolysis intracranial hemorrhage (ICH) in the clinical setting is essential. However, studies in this area spare.We aimed to develop a nomogram by adding atrial fibrillation to the Multicenter Stroke Survey (MSS) score to predict the probability of ICH in acute ischemic stroke patients undergoing thrombolysis. Methods: A retrospective observational study was conducted with 287 participants from a single center (67.2% males, median age 65 years). Head computed tomography scan was performed after 24 hour to evaluate ICH occurrence, and a computed tomography scan was done immediately in case of clinical worsening. The risk factors associated with ICH were analysed. Based on multivariate logistic model, a nomogram was generated for ICH on the basis of atrial fibrillation and the MSS score. We assessed the discriminative performance by using the area under curve (AUC) of receiver-operating characteristic (ROC) and calibration of risk prediction model by using calibration plot. Results: A total of 41(14.3%) ICH events occurred.The MSS score and atrial fibrillation were independent predictors of ICH in multivariate logistic regression analysis. Discrimination of the nomogram was superior to the MSS score alone (0.794 vs 0.741; P=0.034). The model was internally validated by using bootstrap (1000 samples) with AUC-ROC of 0.795. The calibration plot showed good agreement. Conclusion: We developed and internally validated a new nomogram using the MSS score and atrial fibrillation as predictors.The nomogram is a simple and accurate tool for predicting ICH in acute ischemic stroke patients undergoing thrombolysis. Further studies are warranted to validate our findings.