Abstract
BACKGROUND: Unfavorable prognoses are often accompanied for hyperglycemic stroke patients. This study aimed to construct a hyperglycemia/diabetes-derived polygenic risk score (PRS) to improve the predictive performance for poor outcome risks after a stroke and to evaluate its potential clinical application. METHODS: A hospital-based cohort study was conducted including 1320 first-ever acute ischemic stroke (AIS) patients and 1210 patients who completed the follow-up at 3 months. PRSs were calculated for hyperglycemia/diabetes mellitus using results from genome-wide association studies in Asians. An unfavorable functional outcome was defined as a modified Rankin Scale score of ≥3 at 3, 6, and 12 months of follow-up. The prediction of a poor prognosis was evaluated using measures of model discrimination, calibration, and net reclassification improvement (NRI). RESULTS: The second to fourth PRS quartiles (≥Q2) were significantly associated with higher risks of unfavorable outcomes at 3 months compared with the first quartile as the reference group after adjusting for age, baseline stroke severity, hypertension, diabetes, dyslipidemia, smoking, heart disease, and ischemic stroke subtype (p for trend <0.0001). The addition of the PRS to traditional risk predictors of poor outcomes after an AIS significantly improved the model fit (likelihood ratio test p < 0.0001) and enhanced measures of reclassification (NRI, 0.245; 95% confidence interval [CI], 0.195-0.596). The corrected C-index for the PRS combining traditional risk factors at 3 months after a stroke was 0.899 (95% CI, 0.878-0.980). Among hyperglycemic AIS patients, those who did not take an antidiabetic drug and whose PRS was ≥Q2 had higher risks of an unfavorable outcome at 3 months compared with patients who took the medicine. CONCLUSION: The hyperglycemia/diabetes-derived PRS was associated with poor outcomes after an AIS, but further studies are needed to validate its use for clinical applications.