RESUMO
Objective:To investigate the predictive value of stress hyperglycemia for the functional outcomes in patients with acute ischemic stroke (AIS).Methods:Patients with AIS admitted to the Affiliated Hospital of Lianyungang, Xuzhou Medical University from September 1, 2019 to December 31, 2020 were enrolled prospectively. The glucose to glycated hemoglobin ratio (GAR) was used to express stress hyperglycemia. The functional outcome was evaluated by the modified Rankin Scale at 3 month after discharge, 0-2 was defined as a good outcome and >2 as a poor outcome. Multivariate logistic regression analysis was used to determine the independent related factors of functional outcome. Receiver operating characteristic (ROC) curve was used to evaluate the predictive value of GAR for poor functional outcome. Results:A total of 1 286 patients with AIS were included. Their median age was 67 years old, and 762 were males (59.3%). The median baseline National Institutes of Health Stroke Scale (NIHSS) score was 2. The median fasting blood glucose level was 5.48 mmol/L, the median glycosylated hemoglobin was 6.00%, and the median GAR was 0.92. Eight hundred and twelve patients (63.1%) had good outcomes and 474 (36.9%) had poor outcomes. The GAR in the good outcome group was significantly lower than that in the poor outcome group (0.86 vs. 1.03; P<0.001). All patients were divided into 4 groups (GAR1, GAR2, GAR3, and GAR4) according to the GAR quartile from low to high. Multivariate logistic regression analysis showed that after adjusting for relevant confounding factors, GAR4 (taking the GAR1 as a reference, odds ratio [ OR] 8.896, 95% confidence interval [ CI] 5.775-13.702; P<0.001), age ( OR 1.041, 95% CI 1.027-1.055; P<0.001) and baseline NIHSS score classification ( OR 25.898, 95% CI 14.221-47.163; P<0.001) were closely associated with the poor outcomes at 3 months. Further subgroup analysis showed that the higher level of GAR was significantly and independently correlated with the poor functional outcome, regardless of whether the patients had diabetes or not. The ROC curve analysis showed that the area under the curve of GAR predicting poor outcome at 3 months was 0.705 (95% CI 0.675-0.735; P<0.001), and the predictive value was significantly higher than that of glycosylated hemoglobin and fasting blood glucose. When the cut-off value of GAR was 0.97, the Yoden index was the largest, which was 0.370. The sensitivity and specificity of its predicting the poor outcome at 3 months were 61.6% and 75.4%. Conclusion:Whether or not diabetes is present, GAR is an independent predictor of the poor outcomes in patients with AIS.
