Background

Diabetes mellitus (DM) and coronary heart disease (CHD) are two chronic diseases that pose a huge public health burden [1]. CHDs are often accompanied by DM, possibly because both conditions occur with the same risk factors, such as abnormal inflammatory responses or abnormal lipid metabolism [2]. Moreover, as an essential risk factor for CHD, DM can exacerbate the progression of atherosclerosis, resulting in poor clinical outcomes [3, 4]. A 2022 study in China by Junning Fan et al. on 500,000 Chinese has shown that diseases, including CHD and DM, greatly portend the risk of mortality among Chinese adults [5]. The management of glucose metabolism in patients with CHD is therefore particularly important. However, abnormal blood glucose metabolism, including prediabetes (pre-DM) and DM, has become increasingly common; by 2045, more than 600 million people are estimated to develop pre-DM, and the same number will develop DM, according to the 2017 global estimates of DM prevalence and 2045 projections [6]. Besides, Asians are more prone to DM especially type 2 DM (T2DM) and other CHD complications, than Westerners due to various factors [7].

Dyslipidemia often accompanies abnormal glucose metabolism [8]. In addition to traditional lipid parameters, including triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C), non-traditional lipid parameters like TG/HDL-C, LDL-C/HDL-C, non-HDL-C, TC/HDL-C and non-HDL-C/HDL-C are all closely related to the occurrence and development of pre-DM and T2DM. The reason may be that excess cholesterol accumulation leads to β-cell dysfunction, thereby impairing glucose tolerance and affecting insulin secretion. In addition, islet cholesterol deposition may lead to increased islet amyloid polypeptide aggregation and increased islet amyloid formation, further deteriorating β-cell function and affecting glucose homeostasis [9,10,11,12]. More importantly, they are also considered more predictive of CHD than conventional lipid parameters. These non-traditional lipid parameters can provide more information about conventional parameters, are difficult to quantify risk information and can better reflect interactions between lipid components [13]. Of these, the TG/HDL-C has been recognized as a potential predictive marker of insulin resistance (IR), which is a key trigger for the development of T2DM. Increased TG/HDL-C ratios have been shown to indicate a greater risk of new-onset T2DM in some studies [14]. Additionally, past studies have extensively explored sex-specific cutoff points for TG/HDL-C, which classify participants as high or low IR and cardiovascular disease (CVD) risk [15, 16]. However, few studies have compared the strength of the association of TG/HDL-C and other non-traditional lipid parameters with the occurrence of pre-DM and T2DM in the Chinese CHD population.

Therefore, this study aimed to compare the association of TG/HDL-C and other non-traditional lipid parameters with pre-DM and T2DM in the Chinese CHD population, test the association of high TG/HDL-C based on sex-specific cutoff points with pre-DM and T2DM.

Methods

Subjects

We conducted a large, multicenter retrospective cohort study called Retrospective Cohort Study on Adjuvant Treatment of Coronary Heart Disease Angina Pectoris with Chinese Patent Medicine (RCSCD-TCM). During the study, we established a CHD retrospective database, which included 107,301 inpatients with CHD from 6 hospitals in Tianjin, including Tianjin Chest Hospital, Tianjin Hospital of ITCWM Nankai Hospital, Tianjin Academy of Traditional Chinese Medicine Affiliated Hospital, First Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Second Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin Medical University General Hospital from January 1, 2014 to September 30, 2020. The following patients were excluded: (1) those younger than 35 years or older than 75 years; (2) those with oncological, infectious, or serious liver or renal diseases; (3) those who lacked TG, TC, HDL-C, LDL-C, fasting blood glucose (FBG) and hemoglobin A1c (HbA1c) data. Ultimately, 28,476 eligible subjects were enrolled in the final analysis. The flow chart of patient recruitment is shown in Fig. 1. The study was approved by the Ethics Committee of Tianjin University of Traditional Chinese Medicine (approval number TJUTCM-EC20190008) and certified by the Chinese Clinical Trials Registry on July 14, 2019 (registration number ChiCTR-1900024535) and on July 18, 2019, by ClinicalTrials.gov (registration number NCT04026724) [17].

Fig. 1
figure 1

Flow chart of patient recruitment

Full size image

Data collection

The following data for this study analysis were acquired from the CHD retrospective database, where data came from medical records: clinical history, anthropometric data, blood analysis, and medical imaging data. Anthropometric data, including blood pressure, and personal information such as age, sex, smoking status, drinking status, family history of DM, current antihypertensive medication, and current anti-lipid medication were recorded. Fasting venous blood samples were obtained from all subjects on the second day of hospitalization. TG, TC, LDL-C, HDL-C, FBG and HbA1c were measured directly by an automatic hematology analyzer. The laboratory carries out quality control according to standard procedures. Non-HDL-C = TC – HDL-C; Non-HDL-C/HDL-C =  (TC – HDL-C)/HDL-C.

Definitions

Smokers smoke at least 100 cigarettes in their lifetime [18]. Drinkers are defined as consuming alcohol at least 1 time per week [19]. Hypertension was defined as systolic blood pressure (SBP) ≥ 140 mmHg and/or diastolic blood pressure (DBP) ≥ 90 mmHg, or current use of antihypertensive medication [20]. Non-HDL-C = TC-HDL-C; hyperlipidemia was defined as TC ≥ 6.2 mmol/L, TG ≥ 2.3 mmol/L, LDL-C ≥ 4.1 mmol/L, or HDL-C ≤ 1.0 mmol/L [21]. Diabetic status includes normoglycemic (NG) (FBG5.6 < mmol/L or HbA1c < 5.7%), Pre-DM (5.6 ≤ FBG ≤ 6.9 mmol/L or 5.7 ≤ HbA1c ≤ 6.4%), T2DM (FBG ≥ 7.0 mmol/L or HbA1c ≥ 6.5%) [22]. According to the sex-specific cutoff of TG/HDL-C for identifying the risks of IR and CVDs, high TG/HDL-C was defined as TG/HDL-C > 2.5 in females and TG/HDL-C > 3.5 in males [15].

Statistical analysis

The tertiles of TG/HDL-C were T1 (TG/HDL-C < 1.10), T2 (1.10 ≤ TG/HDL-C ≤ 1.89) and T3 (TG/HDL-C > 1.89). The Kolmogorov–Smirnov test was used to test the data for normality. Normally distributed continuous variables were presented as mean ± standard deviation, and non-normally distributed data were presented as median (interquartile). Demographic differences among groups were assessed using the Kruskal–Wallis H test. Categorical variables were expressed as counts and percentages (%), and differences between groups were examined with the chi-square test. Logistic regression models, calculated using odds ratio (OR) and 95% confidence interval (CI), were used to investigate the association of pre-DM and T2DM with various lipid parameters. Two-sided P < 0.05 was considered statistically significant. The collinearity of different models was tested before logistic regression. Missing values were imputed using the multiple imputation method. All statistical analyses were performed using the statistical package for the social sciences version 24.0 (IBM Corp, New York, NY, USA).

Result

Subject characteristics

28,476 participants were included in this study, including 13,321 (46.8%) females, 15,155 (53.2%) males, with a median age of 64 years, and 18,660 (65.5%) were over 60 years old. 30.4% (8465), 25.6% (7281), and 44.1% (12,550) were NG, pre-DM, and T2DM, respectively. TG/HDL-C levels and high TG/HDL-C distribution were higher in pre-DM and T2DM than in NG. Baseline characteristics of participants according to diabetic status are shown in Table 1.

Table 1 Baseline clinical characteristics according to diabetic status
Full size table

Associations between pre-DM and T2DM with univariate

Univariate analysis results showed that age, sex, SBP, hypertension, family history of DM, current antihypertensive medication, TG, TC, HDL-C, LDL-C, non-HDL-C, LDL-C/HDL-C, TC/HDL-C, non-HDL-C/HDL-C, TG/HDL-C, high TG/HDL-C were associated with pre-DM and T2DM. Among non-traditional lipid parameters, TG/HDL-C was the highest risk factor associated with pre-DM and T2DM in CHD patients (OR: 1.19; 95% CI 1.16–1.23), (OR: 1.36; 95% CI 1.33–1.39) (Table 2).

Table 2 Associations between pre-DM and T2DM with univariate
Full size table

Associations between pre-DM and T2DM with traditional lipid parameters and non-traditional lipid parameters

As shown in Table 3, after adjusting for confounding factors, elevated TG (OR: 1.29; 95% CI 1.24–1.33) (OR: 1.50; 95% CI 1.45–1.55), TC (OR: 1.04; 95% CI 1.01–1.07) (OR: 1.08; 95% CI 1.06–1.11), and LDL-C (OR: 1.04; 95% CI 1.01–1.08) (OR: 1.06; 95% CI 1.03–1.10) in traditional lipid parameters were all associated with the risk of pre-DM and T2DM, with TG showing the highest association. Conversely, HDL-C may be a protective factor for pre-DM and T2DM (OR: 0.86; 95% CI 0.78–1.00) (OR: 0.39; 95% CI 0.35–0.43). The non-traditional lipid parameters non-HDL-C (OR: 1.06; 95% CI 1.03–1.09) (OR: 1.16; 95% CI 1.13–1.19), LDL-C/HDL-C (OR: 1.11; 95% CI 1.07–1.14) (OR: 1.27; 95% CI 1.23–1.30), TC/HDL-C (OR: 1.11; 95% CI 1.08–1.13) (OR: 1.28; 95% CI 1.25–1.30) and non-HDL-C/HDL-C (OR: 1.11; 95% CI 1.08–1.13) (OR: 1.28; 95% CI 1.25–1.30) were positively correlated with the risk of pre-DM and T2DM. TG/HDL-C remained the highest risk factor associated with pre-DM or T2DM in patients with CHD (OR: 1.21; 95% CI 1.16–1.25) (OR: 1.35; 95% CI 1.30–1.39).

Table 3 Associations between pre-DM and T2DM with traditional lipid parameters and non-traditional lipid parameters
Full size table

Associations of pre-DM and T2DM with TG/HDL-C

As shown in Table 4, in Model 2, after adjusting for confounders, the results of multiple logistic regression analysis showed that when dividing TG/HDL-C into tertiles, using T1 as a reference, T3 was observed to have the highest association with both pre-DM and T2DM (OR: 1.60; 95% CI 1.48–1.74), (OR: 2.79; 95% CI 2.60–3.00). Compared with pre-DM, the association of TG/HDL-C with the risk of T2DM was stronger. When TG/HDL-C was used as a continuous variable in both unadjusted and adjusted models, the TG/HDL-C (tertiles) was consistent with the P for the trend of the pre-DM and T2DM (Ptrend < 0.001). High TG/HDL-C was significantly associated with pre-DM and T2DM (OR: 1.69; 95% CI 1.52–1.88), (OR: 2.85; 95% CI 2.60–3.12).

Table 4 Associations between pre-DM and T2DM with TG/HDL-C
Full size table

As shown in Table 5, after multivariate adjustment, TG/HDL-C was associated with pre-DM and T2DM in both sexes. But the association between it and pre-DM and T2DM in females (OR: 1.27; 95% CI 1.21–1.33) (OR: 1.49; 95% CI 1.44–1.56) was greater than in males (OR: 1.17; 95% CI 1.13–1.21) (OR: 1.30; 95% CI 1.26–1.34). As shown in Table 6, after adjusting for confounders, it was significantly associated with pre-DM and T2DM at different ages. When TG/HDL-C was used as a continuous variable, this association was greater in patients with CHD over age 60 (OR: 1.23; 95% CI 1.18–1.28) (OR: 1.44; 95% CI 1.39–1.49). Tables 7 and 8 showed that this association was significant across smoking and drinking after multivariate adjustmentc. When TG/HDL-C was used as a continuous variable, the association between it and pre-DM and T2DM in non-smokers (OR: 1.27; 95% CI 1.21–1.32) (OR: 1.46; 95% CI 1.40–1.51) and non-drinkers (OR: 1.24; 95% CI 1.19–1.29) (OR: 1.43; 95% CI 1.38–1.49) was greater. For different sexes, ages, smoking and drinking statuses, using T1 as a reference, T3 levels still presented the highest levels of pre-DM and T2DM risk, and high TG/HDL-C was significantly associated with pre-DM and DM.

Table 5 Associations between pre-DM and T2DM with TG/HDL-C according to sex
Full size table
Table 6 Associations between pre-DM and T2DM with TG/HDL-C according to age
Full size table
Table 7 Associations between pre-DM and T2DM with TG/HDL-C according to smoking status
Full size table
Table 8 Associations between pre-DM and T2DM with TG/HDL-C according to drinking status
Full size table

Discussion

This study investigated the correlation between TG.HDL-C and other unconventional lipid parameters with the risk of pre-DM and T2DM in Chinese patients with CHD. The results showed that non-traditional lipid parameters, especially TG/HDL-C, were related to the risk of pre-DM and T2DM. High TG/HDL-C, defined by the sex-specific TG/HDL-C cutoff point, is a risk factor for pre-DM and T2DM.

T2DM is a chronic metabolic disorder characterized by insufficient insulin production or IR caused by other factors [23]. Pre-DM is the intermediate stage between NG and DM, all people with T2DM pass the pre-DM stage, and about 5% to 10% of pre-DM will progress to T2DM each year [24]. Pre-DM and T2DM have been reported to be associated with an increased risk of CVD, including CHD [23, 25]. Therefore, managing pre-DM and T2DM risk factors is necessary. Glucose metabolism is closely related to lipid metabolism [26, 27]. Previous studies have demonstrated the correlation of lipid parameters including TG, TC, HDL-C, LDL-C, LDL-C/HDL-C, non-HDL-C, TC/HDL-C with pre-DM and DM [28,29,30,31]. Among them, TG and HDL-C have been considered important risk factors for developing CVD in Asians [32, 33]. The potential clinical significance of TG/HDL-C has been widely explored as a product of these two. Recent studies pointed out that TG/HDL-C was associated with IR and cardiometabolic disease risk; its cutoff for identifying risk differs between males and females [16, 34], suggesting that it may be a potential tool for identifying patients with DM. Consequently, we demonstrated its association with the risk of developing pre-DM and T2DM in the CHD population, and sex-specific high TG/HDL-C was a risk factor for pre-DM and T2DM.

The following reasons may explain the association of TG/HDL-C with pre-DM and T2DM: TG elevated results in increased free fatty acids (FFA), reduced insulin sensitivity [35], and continued exposure to FFA due to TG may reduce AMP-activated kinase protein activity and increase TG accumulation, leading to changes in pancreatic α-cell insulin signaling and hypersecretion of glucagon [36], thereby creating a vicious cycle between TG levels and IR. It leads to impaired glucose tolerance and the development of pre-DM and T2DM. At the same time, HDL protected β cells from cytokine- or glucose-induced apoptosis through two components, including ApoA1 (the major protein component of HDL) and S1P. Decreased HDL-C levels affect β-cell function or survival, which has a regulatory role in the pathogenesis of T2DM [37,38,39]. The combination of high TG and low HDL-C, known as atherogenic dyslipidemia, is also a strong risk factor for CHD. Therefore, the TG/HDL-C ratio was considered a potential predictive marker of IR and β-cell dysfunction. It is closely associated with pre-DM and T2DM as well as CVD development [40,41,42]. The study also verified that examined the association of TG/HDL-C with T2DM and pre-DM existed across different sex, age, smoking, and drinking statuses, as IR might changes with these factors [19, 43, 44]. Past studies have generally concluded that females exhibit more favorable metabolic risk profiles than males, including lower TG and higher HDL-C levels, and the association of dyslipidemia with DM appears to be stronger among males [45], middle-aged patients [46], and smokers and drinkers [47]. Conversely, when TG/HDL-C was used as a continuous variable in our study, it was associated with pre-DM and T2DM at different ages, sexes, smoking and drinking status, but stronger in females, people over 60 and those who do not smoke and drink alcohol. These differences might be due to the study population and sample size differences. Our study is aimed at CHD patients in China. Different races, different health conditions, and different sample sizes may affect the results of the study. A previous study in a Chinese population also showed that TG/HDL-C is not a marker of male IR but may be a marker of IR in Chinese non-obese females [48]. John Billimek et al. pointed out that although patients were prescribed similar lipid-lowering drug regimens, females with diabetes had worse lipid control than males [49]. Also, the anabolism is significantly lower in the elderly compared to middle-aged-onset patients. Elderly-onset T2DM patients have relatively preserved β-cell function and higher IR [14]. In addition, the relationship between alcohol consumption and IR T2DM remains controversial [47]. A meta-analysis evaluating the association between alcohol consumption and the risk of metabolic syndrome reported that compared with non-drinkers, very light drinkers were significantly associated with a reduction in the risk of metabolic syndrome. In contrast, heavy drinkers are associated with an increased risk of metabolic syndrome [50]. Further longitudinal studies may be needed for validation.

Strengths and limitations

This present study has some strengths. First, based on our current knowledge, this was the largest population-based study of the association of non-traditional lipid parameters, especially TG/HDL-C, with pre-DM and T2DM in patients with CHD. The association of TG/HDL-C with pre-DM and T2DM was also verified at different ages, sex smoking and drinking statuses to exclude the influence of potential factors on this association. Secondly, possible confounders were also included in the analysis to rule out their interference with the results. Moreover, since a sex-specific cutoff point for TG/HDL-C was included in the core analysis, it may be beneficial to extend the clinical validation of this cutoff point in association with pre-DM and T2DM. Nonetheless, this study still had some limitations. Above all, as an observational study, this study was not suitable for examining the causal relationship between non-traditional lipid parameters and pre-DM and T2DM. Next, the current use of hypoglycemic agents and body mass index (BMI) as important confounders was not included in the regression model due to the missing data. We will conduct prospective cohort studies in the future to investigate causality and collect as comprehensive data as possible. Finally, there may be some unavoidable bias between centers being a multicenter study.

Conclusions

In addition to traditionally determined lipid parameters, non-traditional lipid parameters were significantly correlated with pre-DM and T2DM in CHD patients, among which TG/HDL-C showed a stronger correlation. Early clinical lipid intervention is necessary, especially in CHD patients. Clinicians can take advantage of the potential value of the TG/HDL-C and its sex-specific cutoff points, which may serve as a simple and efficient dyslipidemia management tool for detecting and preventing the risk of DM in patients with CHD.