Abstract
BACKGROUND: The significant association between serum uric acid (SUA) and nonalcoholic fatty liver disease (NAFLD) is well documented. In this report, we tested the hypothesis that SUA might improve the widely studied fatty liver index (FLI) to predict NAFLD. METHODS: A cross-sectional study was performed in a community of Nanjing, China. The population data on sociodemographics, physical examinations, and biochemical tests were collected from July to September 2018. The associations of SUA and FLI with NAFLD were analyzed using linear correlation, multiple linear regressions, binary logistic analyses, and area under receiver-operating characteristic curve (AUROC). RESULTS: A total of 3,499 people were included in this study, of which 36.9% had NAFLD. The prevalence of NAFLD increased with the increase of SUA levels (all P <.05). Logistic regression analyses revealed that SUA was significantly associated with an increased risk of NAFLD (all P <.001). The NAFLD predictive value after combining SUA with FLI was superior to FLI, especially in females (AUROC(SUA + FLI) = 0.911 vs. AUROC(FLI) = 0.903, P <.05). The reclassification of NAFLD also significantly improved, based on net reclassification improvement (0.053, 95% confidence interval [CI]: 0.022-0.085, P <.001) and integrated discrimination improvement (0.096, 95% CI: 0.090-0.102, P <.001). A regression formula of this combined algorithm was proposed as: The novel formula = 0.032* waist circumference + 0.303* body mass index + 1.301* natural logarithm of triglyceride + 0.478* natural logarithm of glutamyl transpeptidase + 0.002* SUA- 18.823. At the cutoff value of 13.3, the sensitivity and specificity of this model were 89.2% and 78.4%, respectively. CONCLUSIONS: SUA level was positively associated with NAFLD prevalence. A new formula combining SUA with FLI may serve as a better indicator to predict NAFLD compared to FLI, especially in females.