RESUMEN
OBJECTIVE: To evaluate the utility of predicting nonalcoholic fatty liver disease (NAFLD) and obesity by retinal vascular changes (RVC) found in fundoscopy and to determine whether this is related to a low-grade inflammatory process. METHODS: We carried out a cross-sectional analysis that included 152 (ages 18-45 years) patients divided into four groups: NAFLD and BMI less than 25, absence of NAFLD and BMI less than 25, NAFLD and BMI more than 30, and absence of NAFLD and BMI more than 30. Retinal fundoscopy, hepatic ultrasound, metabolic profile, and cytokine measurement were performed. We calculated odds ratio [95% confidence interval (CI)], performed diagnostic utility tests, and carried out a 2 factorial analysis. RESULTS: Obesity was associated with RVC (odds ratio: 21.25; 95% CI: 8.79-51.4) and NAFLD [25 (9.07-72.6)]. NAFLD was associated with RVC [11.24 (4.98-26.48)], and the prediction of NAFLD showed a sensitivity of 75% (95% CI: 68-82) and a specificity of 81% (75-86); when RVC-obesity were combined, sensitivity increased to 90% (88-91.7), with a specificity of 85% (84-85.8). C-reactive protein was associated with the three factors, suggesting an independent contribution. Thin patients with RVC and NAFLD had higher concentrations of interleukin-2, interleukin-6, tumor necrosis factor-α, and interferon-γ. CONCLUSION: NAFLD in patients with obesity without diabetes or hypertension can be predicted by RVC, a noninvasive technique carried out by eye fundoscopy. NAFLD alone can drive inflammatory conditions in the absence of obesity that manifests as RVC.