Association of total cholesterol to high-density lipoprotein cholesterol ratio with diabetes risk: a retrospective study of Chinese individuals

Abstract

A common complication of type 2 diabetes is hypercholesterolemia in many patients. It is still unclear, nevertheless, how high-density lipoprotein cholesterol ratio (TC/HDL-C), total cholesterol, and diabetes are related. The purpose of this study is to look at the prediction ability and causal relationship between TC/HDL-C and diabetes. This study included 117,268 subjects who were undergoing physical examinations. The subjects were grouped into four equal groups according to the TC/HDL-C quartiles; the main outcome was the occurrence of diabetes events. TC/HDL-C is calculated as total cholesterol divided by high-density lipoprotein cholesterol. In 3.1 years (± 0.95) of follow-up, 795 women (0.68%) and 1,894 men (1.62%) received new diabetes diagnoses. TC/HDL-C is an independent predictor of new-onset diabetes, according to multivariable Cox regression analysis (HR 1.27 per SD increase, 95% CI: 1.09-1.48, P for trend < 0.001). It turned out that a cutoff value of 3.55 (area under the curve 0.64, sensitivity 0.66, specificity 0.56), was ideal for TC/HDL-C in predicting new-onset diabetes. A subgroup analysis demonstrated that the younger population had a significantly higher risk of TC/HDL-C-related diabetes than the middle-aged group (interaction P < 0.05). After controlling for confounding variables, this Chinese cohort study reveals a direct correlation between TC/HDL-C and diabetes, with a stronger independent association observed in younger and middle-aged individuals.

Keywords: Cohort study; Diabetes; Risk factor; TC/HDL-C; Total cholesterol: high-density lipoprotein cholesterol ratio.

Fig. 1

Flow diagram of participants included. The exclusion criteria for the study population are shown in this flowchart, resulting in to the enrollment of 117,268 participants in total. BMI body mass index; FPG fasting plasma glucose; TC total cholesterol; HDL-C high-density lipoprotein cholesterol.

Fig. 2

Associations between the TC/HDL-C and new-onset diabetes. Shown is the cumulative incidence curve of the diabetes outcome. The result was estimated with the use of Log rank test (P < 0.001). The inset shows the same data on an expanded y axis.

Fig. 3

Receiver operating characteristic (ROC) curve analyses to predict diabetes. The predictive value of TC/HDL-C for incident diabetes was investigated by the researchers using ROC curve analysis. The AUC for TC/HDL-C was significantly higher than that of either TC or HDL-C alone (both P < 0.001). AUC: area under the curve; DM diabetes mellitus, HDL-C high-density lipid cholesterol, TC total cholesterol.