Abstract
BACKGROUND: Diabetic retinopathy (DR), a leading cause of vision loss in working-age adults, is strongly influenced by glycemic control. The stress hyperglycemia ratio (SHR), derived from admission glucose and HbA1c, has emerged as a potential predictor of adverse outcomes, yet its link to DR remains unclear. This study aimed to explore the association between SHR and DR using the NHANES database, as well as its impact on recurrent vitreous hemorrhage (RVH) and neovascular glaucoma (NVG) following pars plana vitrectomy (PPV) in patients with proliferative diabetic retinopathy (PDR). METHODS: First, the association between SHR and DR was evaluated using multivariable logistic regression analysis based on NHANES database (2005-2018), including 4539 eligible diabetic patients, of whom 968 had DR. Second, 250 eyes from 201 PDR patients undergoing PPV were retrospectively analyzed and divided into two groups by median preoperative SHR. Risk factors for complications were identified using multivariable logistic regression, and paired analyses compared SHR levels at complication onset versus initial PPV surgery. RESULTS: In the cross-sectional study based on the NHANES database, multivariate logistic regression revealed a significant positive association between SHR and DR risk after adjusting for all covariates (corrected OR = 1.65, 95% CI: 1.12-2.44; p = 0.014). In the retrospective cohort study, multivariable regression indicated that older age and higher fibrinogen levels were linked to increased postoperative complication risk (p < 0.01). Paired analyses demonstrated significantly higher SHR levels at complication onset compared to initial PPV surgery (p < 0.001). CONCLUSION: SHR was significantly and positively correlated with the incidence of DR. Furthermore, in individuals with PDR, elevation of SHR was closely associated with the occurrence of postoperative complications, suggesting that SHR may be a valuable marker for postoperative adverse events. These results indicate that SHR provides a new clinical basis for the early warning of DR and risk assessment of postoperative complications. TRIAL REGISTRATION: Clinicaltrials.gov NCT05631054 Registered November 30, 2022 https//clinicaltrials.gov/study/NCT05631054.