Non-linear association of the uric acid to high-density lipoprotein cholesterol ratio with non-alcoholic fatty liver disease and diabetes effect modification in Chinese adults: a secondary analysis.

Background: Global non-alcoholic fatty liver disease (NAFLD) prevalence has risen sharply, highlighting an urgent need for improved non-invasive biomarkers. The uric acid to high-density lipoprotein cholesterol ratio (UHR) integrates key metabolic pathways in NAFLD, but its precise relationship and effect modifiers remain unclear. Herein, we explored the correlation between UHR and NAFLD. Methods: This secondary analysis used data from 1,592 Chinese adults (40-70 years) undergoing health checks. NAFLD was diagnosed via ultrasound. UHR was calculated as serum uric acid (μmol/L) divided by high-density lipoprotein cholesterol ratio (HDL-C) (mmol/L). Association between UHR and NAFLD was assessed using multivariable logistic regression, restricted cubic splines (RCS) for non-linearity, and threshold analysis. Receiver Operating Characteristic analysis was utilized to assess the ability of UHR to predict the NAFLD. Subgroup analyses explored effect modification. Results: NAFLD prevalence was 61.1%. After full adjustment, each one standard deviation (SD) increase in UHR was associated with 22% higher NAFLD odds [odds ratio (OR) = 1.22, 95% confidence interval (CI): 1.03-1.46]. RCS revealed significant non-linearity (p < 0.001). A threshold was identified at UHR Z-score = -0.75: below this, each 1-SD increase conferred a 4.6-fold higher NAFLD risk (OR = 4.64, 95%CI: 1.57-13.68); above it, no association was found. Diabetes significantly modified the association (P-interaction = 0.028): UHR predicted NAFLD in non-diabetics (OR = 1.27, 95% CI: 1.04-1.56) but not diabetics. UHR's predictive ability (area under the curve = 0.670) exceeded that of uric acid or HDL-C alone. Conclusion: UHR is an independent, non-linear predictor of NAFLD in Chinese adults, with a distinct risk threshold. Its association with NAFLD is significantly modified by diabetes status. UHR represents a simple, readily available biomarker potentially enhancing NAFLD risk assessment, particularly in non-diabetic individuals.