Objective  To investigate the predictive value of serum matrix metalloproteinase-9 (MMP-9) and neurofilament light chain (NFL) for post-stroke cognitive impairment (PSCI) in patients with first-onset acute ischemic stroke (AIS).

Methods  Eighty patients with AIS were selected. According to Montreal Cognitive Assessment (MoCA) score after 6 months of stroke, they were divided into PSCI group (MoCA score < 26) and non-PSCI group (MoCA score ≥26). The serum levels of MMP-9 and NFL were compared between the two groups on admission; the correlations of MMP-9 and NFL level with MoCA score were analyzed; the Logistic regression model was used to determine the risk factors of PSCI; the receiver operating characteristic (ROC) curve was used to evaluate the predictive efficacy of MMP-9 and NFL levels on PSCI.

Results  The serum levels of MMP-9 and NFL in the PSCI group were significantly higher than those in the non-PSCI group (P < 0.05). Pearson correlation analysis showed that MMP-9 and NFL levels at admission were negatively correlated with MoCA scores 6 months after discharge (r=-0.457, -0.512; P < 0.05). Logistic regression analysis showed that age (OR=1.468), MMP-9 level (OR=1.694) and NFL level (OR=1.802) were independent influencing factors for PSCI in patients with first-ever AIS (P < 0.05). ROC curve analysis showed that the area under the curve (AUC) of MMP-9 and NFL level predicting the occurrence of PSCI at admission were 0.805 and 0.833, respectively; the AUC predicted by the combination of the two indexes was 0.905, which was higher than that of a single index, and the difference was statistically significant (P < 0.05).

Conclusion  Both MMP-9 and NFL can be used as predictors of PSCI occurrence in patients with first-onset AIS, and their combined evaluation can further improve the predictive efficacy.