Abstract: Objective To establish the prediction model for ischemic stroke recurrence. Methods Consecutive patients with ischemic stroke from January 1st, 2008 to December 31st, 2009 were followed up until June 30th, 2010. The rate of recurrence were described with Kaplan-Meier curve, and the factors associated with recurrence were analyzed with monovariant and multivariate Cox's proportional hazard regression. Results There were 79 cases relapsed during the follow-up. The independent risk factors associated with recurrence were age(X),history of hypertension (X2), family stroke history (X3), total cholesterol (X4), disease progression (X5), total scores of Essen stroke risk score (X6). The personal prognosis index (PI) for predicting recurrence was as: PI=0.025X1+0.681X2+0.973X3+0.395X4+0.636X5+0.283X6. As the receiver operating characteristic (ROC) curve showed, when the cut-off point of PI was 2.289, the sensitivity of the model was 0.731 and specificity was 0.795. Conclusion The model for predicting recurrence of ischemic stroke was established.

Key words: ischemic stroke, recurrence, Cox's regression model, prognosis index