Introduction:
A risk stratification model is pretty important to prevent the recurrent stroke for ischemic stroke patients. The present study aimed to meta-analysis the ability of Essen Stroke Risk Score (ESRS) to accurately predict recurrence of ischemic stroke.

Material and methods:
Studies on the diagnostic performance of Essen Stroke Risk Score in predicting recurrent stroke were searched by electronic and manual methods. Quality pooled C-statistics, and 95% confidence intervals (95% CI) were evaluated.

Results:
Fifteen studies with a total of 94,052 patients were included in the meta-analysis. The pooled C-statistics of ESER for patients without atrial fibrillation (AF) experiencing recurring strokes at 90-day or one-year were 0.65 (95% CI: 0.58-0.73) and 0.57 (95% CI: 0.53-0.60), and the heterogeneity was weak. The average ratio of one-year recurrent stroke in the low-risk and high-risk groups classified according to ESRS is 5.6%(range 1.4 to 12.1%) and 9.2%( range 3.2 to 20.1%), respectively. And the calibration analysis showed the pooled RR in the low-risk group is 0.88 (95%CI: 0.24-3.19) and 0.88 (0.24-3.31) with wide confidence intervals and high levels of heterogeneity, indicating the calibration ability was low.

Conclusions:
ESRS had low to moderate ability to predict recurrence of stroke in patients with ischemic stroke and low calibration ability, which need to be further improved.