Serum C-reactive protein adds predictive information for post-stroke delirium: The PROPOLIS study.

OBJECTIVE

Delirium is common and serious complication after stroke. Accurate prediction of delirium is important for prevention and monitoring of high-risk patients. Our study aimed to determine if addition of C-reactive protein (CRP) to a model based on easy-to-access clinical predictors improves accuracy of delirium prediction in acute stroke patients.

METHODS

We analyzed data of patients participating in the Prospective Observational Polish Study on post-stroke delirium. We included patients admitted within 24 h after stroke or transient ischemic attack (TIA) in whom serum CRP was measured on admission. We examined core features of delirium during first 7 days of hospitalization. We assessed if addition of CRP to two clinical models improved metrics of discrimination and reclassification. Model A included age and stroke severity and Model B included stroke severity, atrial fibrillation, diabetes mellitus, pre-stroke dependency, and hemorrhagic stroke.

RESULTS

We included 459 patients. We diagnosed delirium in 29.2% of them. Patients who developed delirium had higher CRP level than those without delirium (median: 13.2 vs. 4.4 mg/L, p < 0.001). CRP >7.09 mg/L was associated with an increased risk of delirium (adjusted OR: 2.98, 95%CI: 1.71-5.19, p < 0.001). After adding CRP to clinical models, an area under receiver operator curve increased from 0.77 to 0.80 (p = 0.038) for Model A and from 0.81 to 0.84 (p = 0.016) for Model B. There was also improvement in reclassification.

CONCLUSIONS

Addition of CRP to clinical predictors moderately improved prediction of post-stroke delirium. CRP could be considered as a potential biomarker to stratify risk of delirium after stroke.