Current Diabetes Reviews · 7 authors, 3 centres
This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.
This cross-sectional study of 251 adults with type 2 diabetes found that diabetic retinopathy (DR) was associated with a 2.5-fold increased odds of worse cognitive performance, even after adjusting for age, education, hypertension, and depression. The findings suggest that routine eye exams for DR may help identify patients at higher risk for cognitive decline. This supports the use of retinal evaluation as a potential screening tool for cognitive dysfunction in diabetes care.
**Background:** Type 2 diabetes mellitus (T2DM) is associated with a twofold increased risk of dementia, suggesting that cognitive impairment may be a chronic diabetes complication. Diabetic retinopathy (DR), a microvascular complication of diabetes, shares embryological and pathophysiological similarities with the brain, making the retina a potential window into cerebral health. However, the association between DR and cognitive performance in T2DM populations, particularly in upper-middle-income countries, remains underexplored. This study aimed to evaluate the association of DR and macular edema with cognitive performance in a T2DM cohort.
**Methods:** A cross-sectional study was conducted at a tertiary hospital in Southern Brazil from September 2017 to December 2020. Among 400 T2DM adults recruited during routine appointments, 251 were included after exclusions (e.g., use of cognition-altering medications, illiteracy, prior dementia, stroke, or inability to perform tests). Data collected included sociodemographics, clinical history, lifestyle factors, and laboratory tests (HbA1c, lipids, creatinine, etc.). Retinal examination was performed via indirect binocular ophthalmoscopy and slit-lamp biomicroscopy under mydriasis, classifying DR as none, non-proliferative, proliferative, or macular edema. Cognitive function was assessed using validated Portuguese versions of the Mini-Mental State Examination (MMSE), Semantic Verbal Fluency Test, Trail Making Test A and B, and CERAD Word Memory Test. Depression symptoms were screened using the Patient Health Questionnaire-9 (PHQ-9). All cognitive test scores were converted into a Global Cognition z-Score (GCS(z)), with GCS(z) < 0 indicating worse performance. Multivariate binary logistic regression (backward conditional) was used to estimate the association between GCS(z) < 0 and DR, adjusting for variables with p < 0.25 in univariate analysis (age ≥ 65 years, school years ≤ 6, diabetes duration ≥ 10 years, physical activity, PHQ-9 > 9, arterial hypertension, cardiovascular disease, and macular edema).
**Key Results:** The final sample of 251 patients had a mean age of 61.1 (±9.8) years, 56.6% female, mean diabetes duration of 12.6 (±8.9) years, and mean schooling of 7.6 (±4.2) years. DR prevalence was 46.5%, and 54% had at least one microvascular complication. Hypertension was present in 82.4%, dyslipidemia in 89%, and cardiovascular disease in 35.2%. Only 30% had HbA1c < 7%. The prevalence of GCS(z) < 0 was 46.6%. In the multivariate logistic regression model, DR was significantly associated with GCS(z) < 0, with an odds ratio (OR) of 2.50 (95% CI: 1.18–5.34), after adjustment for age ≥ 65 years (OR 5.46, 95% CI: 2.42–12.34), school education ≤ 6 years (OR 12.19, 95% CI: 5.62–26.46), arterial hypertension (OR 2.55, 95% CI: 0.88–7.39), and depression symptoms (PHQ-9 > 9) (OR 3.53, 95% CI: 1.55–8.07). Macular edema was not independently associated in the final model.
**Clinical Implications:** This study demonstrates that DR is independently associated with worse cognitive performance in T2DM patients, even after controlling for major confounders such as age, education, hypertension, and depression. The findings align with prior meta-analyses reporting a pooled OR of 2.45 for cognitive impairment with DR. Given that retinal examination is already a standard annual screening in diabetes care, it may serve as a low-cost, non-invasive tool to identify patients at higher risk for cognitive decline. The strong effect of low education (OR 12.19) underscores the need to adjust cognitive assessments for educational level. Limitations include the cross-sectional design (precluding causality), lack of brain imaging or retinal neurodegeneration biomarkers, and potential selection bias from a single tertiary center. Nonetheless, the results support integrating cognitive risk assessment into routine diabetes management, particularly in resource-limited settings. Future longitudinal studies should explore whether DR severity and retinal neurodegeneration markers can predict cognitive trajectory over time.