Association between stress hyperglycemia ratio and delirium in older hospitalized patients: a cohort study

Abstract

Abstract Background It remains unclear whether stress hyperglycemia is associated with delirium. We performed this cohort study to determine the association between stress hyperglycemia and delirium. Methods We consecutively enrolled patients aged ≥70 years who were admitted to the Geriatric Department of West China Hospital between March 2016 and July 2017. Stress hyperglycemia ratio (SHR) was calculated as fasting blood glucose divided by estimated average glucose derived from glycosylated hemoglobin (HbA1c) and was classified into three tertiles. Delirium was screened within 24 h of admission and three times daily thereafter, using the confusion assessment method. The Cox proportional hazards models were used to assess the association of SHR with delirium. Results Among 487 included patients (mean age 83.0 years, 72.0% male), 50 (10.3%) patients experienced delirium during hospitalization. Compared to the second tertile, both the lowest and the highest SHR tertiles were independently associated with delirium (hazard ratio [HR] 3.71, 95% confidence interval [CI] 1.45-9.51; and HR 2.97, 95% CI 1.29-6.81, respectively). Similar results were found after further adjusting for statin comedication. Multiple-adjusted restricted cubic splines revealed a nonlinear relationship between SHR and delirium (Pnonlinearity=0.04). Adding SHR to conventional risk factors improved the risk prediction of delirium (net reclassification index 0.39, P=0.01; integrated discrimination improvement 0.07, P=0.03). Subgroup analyses indicated that the relationship between SHR and delirium was more apparent in patients with HbA1c <6.5%, with significantly higher HR in the first (3.65, 95% CI 1.11-11.97) and third (3.13, 95% CI 1.13-8.72) SHR tertiles compared to the second tertile, while there was no significant association between SHR and delirium in those with HbA1c ≥6.5%. Conclusions Both lower and higher SHR were associated with increased risk of delirium but only in patients with HbA1c <6.5%. Admission SHR may serve as a promising predictor of delirium, and incorporating this biomarker into prediction algorithms might have potential clinical utility in aiding delirium risk stratification, especially in those with HbA1c <6.5%.