Prediction of poststroke cognitive impairment based on the systemic inflammatory response index

Abstract

Background: Poststroke cognitive impairment (PSCI) is a prevalent complication among stroke survivors. Although the systemic inflammatory response index (SIRI) has been shown to be a reliable predictor of a variety of inflammatory diseases, the association between the SIRI and PSCI is still unclear. Therefore, the purpose of this study was to investigate the relationship between SIRI and PSCI, and to design a nomogram to predict the risk of PSCI in acute ischemic stroke (AIS) patients.

Methods: A total of 1342 patients with AIS were included in the study. Using the Mini-Mental State Examination scale, patients were separated into PSCI and non-PSCI groups within 2 weeks of stroke. Clinical data and SIRI values were compared between the groups. We developed the optimal nomogram for predicting PSCI using multivariate logistic regression. Finally, the nomogram was validated using the receiver operating characteristic curve, calibration curve, and decision curve analysis (DCA).

Results: In total, 690 (51.4%) patients were diagnosed with PSCI. After adjusting for potential confounders, the SIRI (OR = 1.226, OR: 1.095-1.373, p < .001) was shown to be an independent risk factor for PSCI in the logistic regression analysis. The nomogram based on patient gender, age, admission National Institutes of Health Stroke Scale scores, education, diabetes mellitus, and SIRI had good discriminative ability with an area under the curve (AUC) of 0.716. The calibration curve and Hosmer-Lemeshow test revealed excellent predictive accuracy for the nomogram. Finally, the DCA showed the good clinical utility of the model.

Conclusion: Increased SIRI on admission is correlated with PSCI, and the nomogram built with SIRI as one of the predictors can help identify PSCI early.

Keywords: acute ischemic stroke; inflammation; nomogram; poststroke cognitive impairment; systemic inflammatory response index.