ObjectiveTo evaluate the relationship between levels of serum apolipoproteins and the prevalence of cardiovascular autonomic neuropathy (CAN) in type 2 diabetes. MethodsIn total, 3199 individuals with type 2 diabetes were investigated in a cross-sectional study. The diagnosis of CAN was made based on the results of a cardiovascular reflex test. Serum apolipoprotein A-I (apoA-I) and apolipoprotein B (apoB) levels were measured. ResultsSerum apoA-1 levels were significantly low in individuals with CAN, but there was no significant association between serum apoB levels and CAN. According to the degree of cardiovascular autonomic dysfunction, the average apoA-I levels were significantly different after adjusting for other covariates (normal, 1.32 g/l, 95% confidence interval [CI] 1.30–1.35; early, 1.29 g/l, 95% CI 1.27–1.31; definite, 1.27 g/l, 95% CI 1.25–1.30; P for trend = 0.010). In the multivariable analysis, the statistically significant association between apoA-I and CAN remained after adjusting for the risk factors (odds ratio per standard deviation increase in the log-transformed value, 0.65; 95% CI, 0.43–0.97, P = 0.036). Additional adjustments for serum high-sensitivity C-reactive protein (or fibrinogen) concentrations eliminated this relationship. ConclusionsSerum apoA-I levels are inversely associated with the prevalence of CAN in individuals with type 2 diabetes. Our data also suggest that a putatively increased risk of CAN associated with decreased apoA-I levels might be mediated by correlated increases in the levels of inflammatory markers.