Development and external validation of a prognostic model for occult atrial fibrillation in patients with ischemic stroke

Abstract

ObjectiveCurrently, the risk of occult atrial fibrillation (AF) could not be predicted in patients with acute ischemic stroke (AIS) using a simple scoring system. Therefore, in this study, we developed and externally validated a nomogram to predict occult AF in patients with AIS.MethodsIn this study, we prospectively conducted a development cohort study with data collected at our stroke center from July 2017 to February 2018, and an external validation cohort from March 2019 to December 2019.ResultsFollow-up data were collected from 177 participants (56.5% older than 65 years, 29.4% female) for generating the nomogram model. Multivariate logistic regression analysis was performed with AF as the dependent variable indicated that age >65 years, heart rate >100, C-reactive protein (CRP), N-terminal pro-B-type natriuretic peptide (NT-proBNP) >270, hemorrhagic transformation (HT) as independent variables for predicting the development of AF, and a nomogram was generated based on these factors. The area under the receiver operating characteristic curve (AUC-ROC) for the model was 0.937, the C-index was 0.926, and the AUC-ROC for the validation cohort was 0.913.ConclusionTo our knowledge, this is the first nomogram developed and externally validated in a stroke center cohort for individualized prediction of risk of developing AIS in patients with occult AF. This nomogram could provide valuable information for the screening of occult AF after a stroke.