BACKGROUND: Existing evidence about the associations of non-alcoholic fatty liver disease (NAFLD) and serum uric acid (SUA) with subclinical atherosclerosis is controversial. The aim of the present study was to examine the associations of NAFLD and SUA with subclinical atherosclerosis. METHODS: In the present cross-sectional study, 1354 obese adults underwent hepatic ultrasonography and arteriosclerosis detection. Indices of subclinical atherosclerosis were brachial-ankle pulse wave velocity (ba-PWV) and the ankle-brachial index (ABI). Linear regression using multivariable fractional polynomial (MFP) modeling was used to examine independent associations of NAFLD and SUA with a-PWV and ABI. RESULTS: Compared with controls, mean (± SD) ba-PWV was significantly higher in subjects with NAFLD (1534 ± 292 vs 1433 ± 259 cm/s; P < 0.001) and hyperuricemia (HUA; 1519 ± 275 vs 1476 ± 287 cm/s; P = 0.007). After adjustment for sociodemographic and lifestyle factors, NAFLD and SUA were both positively related to ba-PWV (β = 0.120 and 0.064, respectively; P < 0.05 for both). With further adjustment for insulin resistance and components of metabolic syndrome (MetS), the positive correlations were no longer significant (β = 0.017 and 0.006; P > 0.05 for both). In addition, NAFLD, but not SUA, was negatively correlated with ABI (β = -0.073; P = 0.015). Using MFP modeling, the best fractional polynomial (FP) transformation model showed that non-linear transformations were appropriate for two variables in their relationship with ba-PWV, namely age and fasting insulin as first-degree FP transformations (age3 and 1/insulin0.5 , respectively). CONCLUSIONS: Neither NAFLD nor SUA was related to ba-PWV with increases in insulin resistance and MetS, but NAFLD was independently and negatively correlated with ABI.
BACKGROUND: Existing evidence about the associations of non-alcoholic fatty liver disease (NAFLD) and serum uric acid (SUA) with subclinical atherosclerosis is controversial. The aim of the present study was to examine the associations of NAFLD and SUA with subclinical atherosclerosis. METHODS: In the present cross-sectional study, 1354 obese adults underwent hepatic ultrasonography and arteriosclerosis detection. Indices of subclinical atherosclerosis were brachial-ankle pulse wave velocity (ba-PWV) and the ankle-brachial index (ABI). Linear regression using multivariable fractional polynomial (MFP) modeling was used to examine independent associations of NAFLD and SUA with a-PWV and ABI. RESULTS: Compared with controls, mean (± SD) ba-PWV was significantly higher in subjects with NAFLD (1534 ± 292 vs 1433 ± 259 cm/s; P < 0.001) and hyperuricemia (HUA; 1519 ± 275 vs 1476 ± 287 cm/s; P = 0.007). After adjustment for sociodemographic and lifestyle factors, NAFLD and SUA were both positively related to ba-PWV (β = 0.120 and 0.064, respectively; P < 0.05 for both). With further adjustment for insulin resistance and components of metabolic syndrome (MetS), the positive correlations were no longer significant (β = 0.017 and 0.006; P > 0.05 for both). In addition, NAFLD, but not SUA, was negatively correlated with ABI (β = -0.073; P = 0.015). Using MFP modeling, the best fractional polynomial (FP) transformation model showed that non-linear transformations were appropriate for two variables in their relationship with ba-PWV, namely age and fasting insulin as first-degree FP transformations (age3 and 1/insulin0.5 , respectively). CONCLUSIONS: Neither NAFLD nor SUA was related to ba-PWV with increases in insulin resistance and MetS, but NAFLD was independently and negatively correlated with ABI.