Development and validation of nomograms for predicting stroke recurrence after firstepisode ischemic stroke

Abstract

To explore the risk factors for recurrence in first-episode ischemic stroke survivors and establish a model for predicting stroke recurrence using a nomogram. We collected the data from a total of 821 first-episode ischemic stroke survivors admitted in the Department of Neurology, West China Hospital, Sichuan University from January, 2010 to December, 2018. R software was used for random sampling of the patients, and 70% of the patients were included in the training set to establish the prediction model and 30% were included in the validation set. Cox proportional risk regression model was used to analyze the factors affecting stroke recurrence, and R software rms package was used to construct the histogram and establish the visual prediction model. C-index and calibration curve were used to evaluate the performance of the model for predicting stroke occurrence. Among the 821 survivors, the recurrence rate was 16.81% at 3 years and 19.98% at 5 years. Multivariate analysis of the training set by Cox regression model showed that an age over 65 years (HR= 2.596, P=0.024), an age of 45-64 years (HR=2.510, P=0.006), a mRS score beyond 3 (HR=2.284, P=0.004) and a history of coronary heart disease (HR=1.353, P=0.034) were all risk factors for stroke recurrence. The C-indexes of the nomogram for the 3-and 5-year relapse prediction model were 0.640 and 0.671, respectively. Age, mRS score and peripheral vascular disease are the factors affecting stroke recurrence in first-episode ischemic stroke survivors, and the nomogram has a high discrimination and predictive power for predicting ischemic stroke recurrence.