Association between plasma-glycosylated hemoglobin A1c/high-density lipoprotein cholesterol ratio and urinary albumin-creatinine ratio in Chinese adults

Abstract

Objective: To explore the relationship between glycosylated hemoglobin A1c/high-density lipoprotein cholesterol ratio (HbA1c/HDL-C) and urinary albumin-creatinine ratio (UACR) in Chinese adults. Methods: In this cross-sectional study, the clinical data of 43 820 community residents (age>40 years) from the Risk Evaluation of Cancers in Chinese Diabetic Individuals (REACTION study; March-December 2012) across eight centers (Liaoning, Guangdong, Shanghai, Gansu, Guangxi, Henan, Hubei, and Sichuan) in China were collected and analyzed. Participants were divided into three groups based on UACR levels:<10 mg/g, 10-30 mg/g, and >30 mg/g. The HbA1c/HDL-C ratio was divided into four groups according to quartile division of the subjects: 1st quartile (Q1<3.79), 2nd quartile (3.79≤Q2<4.59), 3rd quartile (4.59≤Q3≤5.66), and 4th quartile (Q4>5.66). Multivariate ordinal logistic regression model was used to analyze the relationship between HbA1c/HDL-C and UACR. Receiver operating characteristic (ROC) analysis was used to explore the predictive value of HbA1c/HDL-C to UACR. Results: The 43 820 subjects included 13 452 (30.70%) male and 30 378 (69.30%) female patients, with an average age of (58.00±0.05) years. According to results of one-way analysis of variance analysis, the HbA1c/HDL-C ratio was significantly associated with the risk of increased UACR (F=495.73, P<0.001). After adjusting for clinically relevant confounding variables in logistic regression model, compared with participants with the lowest HbA1c/HDL-C ratio (Q1), women with the highest HbA1c/HDL-C ratio (Q4) had a 1.483-fold (95%CI 1.376-1.598, P<0.001) and men had a 1.161-fold (95%CI 1.019-1.323, P<0.001) increased risk of UACR. The ROC curve analysis showed that the area under the curve of HbA1c/HDL-C for predicting increased UACR was 0.623 (95%CI 0.597-0.606), with a sensitivity of 60.18% and a specificity of 54.91%. The HbA1c/HDL-C ratio showed the highest predictive value of all glycemic and lipidemic parameters. In individuals with well-controlled blood glucose (HbA1c<6.5%) or lipid levels (HDL-C≥1.0 mmol/L), the HbA1c/HDL-C ratio was still independently associated with the risk of increased UACR after adjusting for confounding variables [OR(95%CI) of quartile 4: 1.563 (1.210-2.019, P=0.001) in participants with HbA1c<6.5% and 1.822 (1.687-1.968, P<0.001) in participants with HDL-C≥1.0 mmol/L]. Conclusion: As a novel compound indicator for evaluating glucose homeostasis and dyslipidemia, the HbA1c/HDL-C ratio was independently associated with increased UACR in the general population aged>40 years in China, which was superior to both glycemic and lipid parameters alone.