The triglycerides-glucose index and the triglycerides to high-density lipoprotein cholesterol ratio are both effective predictors of in-hospital death in non-diabetic patients with AMI

Introduction

Acute coronary syndrome (ACS) is a severe type of coronary heart disease (Reed, Rossi & Cannon, 2017). The Global Registry of acute coronary events (GRACE) study showed that the 1-year mortality of ACS patients was about 15%, and the cumulative 5-year mortality was about 20% (Fox et al., 2010). The mortality of acute myocardial infarction (AMI) is still on the rise in China. Diabetes and metabolic syndrome are important risk factors for AMI (DeFilippis et al., 2019). Identifying high-risk patients, early diagnosis and timely intervention of patients with AMI to reduce mortality and improve prognosis have been important research directions.

Insulin resistance is a significant risk factor for cardiovascular disease (CVD), which is associated with decreased cardiac autonomic function in the elderly without diabetes (Laakso & Kuusisto, 2014; Ormazabal et al., 2018; Tenenbaum et al., 2007; Howard et al., 1996). Insulin resistance is characterized by the loss of endogenous or exogenous insulin effect, resulting in glucose uptake and utilization disorders (James, Stockli & Birnbaum, 2021). A study showed that nearly 58.4% of non-diabetic Parkinson’s patients have been diagnosed with insulin resistance (Hogg et al., 2018). Tissue-specific insulin resistance, such as abnormalities in the liver, muscle and adipose tissue, can lead to endothelial dysfunction and macrophage accumulation, causing severe CVD events (Poon et al., 2020; Jia et al., 2021; Ryu et al., 2021).

The triglycerides-glucose index (TyG) and the triglycerides to high-density lipoprotein cholesterol (TG/HDL-C) are simple indicators for assessing insulin resistance in epidemiological studies. TyG is the product of fasting triglyceride and fasting blood glucose (Wang et al., 2021a; Wang et al., 2021b). There is a positive correlation between TyG index and insulin resistance (Dikaiakou et al., 2020). Compared with the homeostasis model assessment of insulin resistance (HOMA-IR) index and the euglycemic hyperinsulinemic clamp test, TyG is a new, concise and reliable alternative index for evaluating insulin resistance (Wang et al., 2021a; Wang et al., 2021b). Moreover, studies have shown that a higher TyG index is associated with an increased risk of type 2 diabetes (Xuan et al., 2021; Dikaiakou et al., 2020). In addition, TyG index is positively correlated with coronary artery calcification and is an effective indicator for predicting the progress of CVD (Park et al., 2019).

TG/HDL-C is also a feasible indicator for evaluating insulin resistance (Zhou et al., 2016). A study that enrolled 9764 adults (average age 56 years old) who had participated in the REACTION study in a Guangzhou community found that, compared with other blood lipid markers, TG/HDL-C may be a better indicator for assessing insulin resistance and diabetes among middle-aged and elderly Chinese populations (Lin et al., 2018). Zare et al. (2022) used TG/HDL-C as a surrogate index of insulin resistance in people with pulmonary hypertension without a history of diabetes, and defined TG/HDL-C>3.0 as an indicator of insulin resistance. This study found that insulin resistance was not associated with all-cause mortality in pulmonary hypertension (Zare et al., 2022; Naderi et al., 2014).

Due to the crucial roles of TyG and TG/HDL-C in insulin resistance, this study aimed to use TyG and TG/HDL-C to explore the effects of insulin resistance on cardiac function and in-hospital death in non-diabetic AMI patients, and to provide guidance for the stratification of clinical risk in patients with AMI.

Materials & Methods

Results

TyG and TG/HDL-C, both indicators of insulin resistance, are closely related to death after myocardial infarction

Normality of TyG was tested by the Shapiro–Wilk test and Q-Q plots. The absolute value of the sample kurtosis is less than 10 and the absolute value of the skewness is less than 3, TyG can be described as basically conforming to the normal distribution. Independent-sample t-test was used for comparison of TyG. The TyG index was significantly increased in the death groups compared to the no-death groups in AMI (8.86 ± 0.59 vs 9.22 ± 0.97, P = 0.001), STEMI (8.90 ± 0.60 vs 9.24 ± 0.95, P = 0.024) and NSTEMI (8.81 ±0.58 vs 9.19 ± 1.00, P = 0.022) patients (Fig. 1A). Mann–Whitney U method was used for comparison of TG/HDL-C not conforming to normality.TG/HDL-C was not significantly increased in the death group of AMI patients (P = 0.588) (Fig. 1B). The patients were also divided into different groups according to the quartiles of TyG (Q1 <8.4543, Q2 8.4543−8.8418, Q3 8.8418−9.2294, Q4 ≥9.2294) and TG/HDL-C (T1 <2.0270, T2 2.0270−2.9487, T3 2.9487−4.3456, T4 ≥4.3456). The Chi-square test was used to test for differences in the trend of in-hospital mortality. The results showed that in-hospital mortality in the Q4 group of TyG (6.80%) and T4 group of TG/HDL-C (5.76%) was higher than in other groups (Q1 = 5.10%, Q2 = 3.64%, Q3 = 5.10%; T1 = 4.51%, T2 = 4.75%, T3 = 4.74%), despite the lack of statistically significant differences among these groups.

TyG and TG/HDL-C are important independent predictors of in-hospital death in AMI

Since this study only recorded the death status of enrolled subjects without recording the time of death, logistic regression instead of Cox regression was used to study the predictive ability of TyG and TG/HDL-C on in-hospital death of AMI patients. Due to a large number of variables (about 100) in this study and the sample size of 1,648 patients, the inclusion of excess variables into the logistic model could affect the model’s accuracy. Therefore, we chose to perform univariate logistic regression to screen out the significant variables that were subsequently included in multivariate logistic regression without omitting clinically important variables. Univariate logistic regression analysis showed that age (OR = 1.079, 95% CI [1.056−1.102], P < 0.001), alcohol history (OR =2.234, 95% CI [1.019−4.896], P = 0.045), smoking history (OR = 0.565, 95% CI [0.354−0.902], P = 0.017), ALB (OR = 0.847, 95% CI [0.815−0.881], P < 0.001), ALT (OR = 1.001, 95% CI [1.001−1.002], P = 0.001), BUN (OR =1.135, 95% CI [1.103–1.169], P < 0.001), creatinine (OR =1.003, 95% CI [1.002−1.005], P < 0.001), Hb (OR =0.970, 95% CI [0.961−0.978], P < 0.001), TC (OR =0.991, 95% CI [0.985−0.996], P = 0.001), SBP (OR =0.982, 95% CI [0.972−0.992], P < 0.001), UA (OR =1.003, 95% CI [1.002−1.004], P < 0.001), WBC (OR =1.083, 95% CI [1.043−1.124], P < 0.001) and TyG (OR =2.256, 95% CI [1.646−3.092], P < 0.001) were significant factors predicting in-hospital death in patients with AMI. The significant variables in the univariate analysis were enrolled in the multivariate logistic regression analysis. We used tolerance Tol and variance inflation factor VIF to evaluate whether the collinearity of included variables was met (Supplemental Materials 4). “Enter” method was used for multivariate logistic regression analysis, and the results showed that age (OR =1.069, 95% CI [1.042−1.097], P < 0.001), ALB (OR =0.926, 95% CI [0.879−0.976], P = 0.004), ALT (OR =1.001, 95% CI [1.000−1.002], P = 0.028), BUN (OR =1.063, 95% CI [1.005−1.123], P = 0.031), SBP (OR = 0.982, 95% CI [ 0.971−0.993], P = 0.002), WBC (OR =1.060, 95% CI [1.012−1.111], P = 0.014) and TyG (OR =3.106, 95% CI [2.122−4.547], P < 0.001), were important factors predicting the prognosis of AMI. “Forward” and “Backward” methods were also performed to validate the result of “enter” method (Supplemental Materials 5). The results of TyG by “Forward” (OR =3.021, 95% CI [2.084−4.379], P < 0.001) and “Backward” (OR =3.080, 95% CI [2.123−4.469], P < 0.001) methods were similar with “enter” method. Based on the above results, TyG was an important independent predictor of in-hospital death in AMI after adjusting for multiple clinical confounders (Table 3, Fig. 2A). When TG/HDL-C and significant factors in univariate variables were included in to multivariate analysis, the result of “enter” method showed that TG/HDL-C (OR =1.167, 95% CI [1.062−1.282], P = 0.001) was also an important predictor of in-hospital death (Table 4, Fig. 2B). The results of TG/HDL-C by “Forward” (OR =1.161, 95% CI [1.059−1.273], P = 0.002) and “Backward” (OR =1.163, 95% CI [1.059−1.276], P = 0.001) methods were also similar with “enter” method (Supplemental Materials 5).

Table 3:

Logistic regression of predicting in-hospital death in AMI patients without diabetes.

Parameters	Univariate regression			Variate regression
	OR	95% CI	P value	OR	95% CI	P value
Age (years)	1.079	1.056–1.102	<0.001***	1.069	1.042–1.097	<0.001***
Alcohol	2.234	1.019–4.896	0.045*	1.243	0.488–3.165	0.648
Smoking	0.565	0.354–0.902	0.017*	1.097	0.607–1.981	0.760
ALB (g/L)	0.847	0.815–0.881	<0.001***	0.926	0.879–0.976	0.004**
ALT (U/L)	1.001	1.001–1.002	0.001**	1.001	1.000–1.002	0.028*
BUN (mmol/L)	1.135	1.103–1.169	<0.001***	1.063	1.005–1.123	0.031*
Creatinine (µmol/L)	1.003	1.002–1.005	<0.001***	1.000	0.998–1.003	0.887
Hb (g/L)	0.970	0.961–0.978	<0.001***	0.991	0.979–1.003	0.144
TC (mg/dl)	0.991	0.985–0.996	0.001**	0.999	0.993–1.005	0.698
SBP (mmHg)	0.982	0.972–0.992	<0.001***	0.982	0.971–0.993	0.002**
UA (µmol/L)	1.003	1.002–1.004	<0.001***	1.000	0.998–1.002	0.828
WBC (109/L)	1.083	1.043–1.124	<0.001***	1.060	1.012–1.111	0.014*
TyG	2.256	1.646–3.092	<0.001***	3.106	2.122–4.547	<0.001***

DOI: 10.7717/peerj.14346/table-3

Notes:

*P < 0.05

**P < 0.01

***P < 0.001

ALB

Albumin

ALT

Alanine aminotransferase

BUN

Blood urea nitrogen

SBP

Systolic blood pressure

TC

Total cholesterol

UA

Uric acid

WBC

White blood cells

Table 4:

Logistic regression of predicting in-hospital death in AMI patients without diabetes.

Parameters	Univariate regression			Variate regression
	HR	95% CI	P value	HR	95% CI	P value
Age (years)	1.079	1.056–1.102	<0.001***	1.069	1.041–1.098	<0.001***
Alcohol	2.234	1.019–4.896	0.045*	1.023	0.406–2.574	0.962
Smoking	0.565	0.354–0.902	0.017*	1.101	0.611–1.983	0.748
ALB (g/L)	0.847	0.815–0.881	<0.001***	0.935	0.887–0.986	0.011*
ALT (U/L)	1.001	1.001–1.002	0.001**	1.001	1.000–1.002	0.020*
BUN (mmol/L)	1.135	1.103–1.169	<0.001***	1.066	1.005–1.130	0.033*
Creatinine (µmol/L)	1.003	1.002–1.005	<0.001***	1.001	0.998–1.003	0.638
Hb (g/L)	0.970	0.961–0.978	<0.001***	0.991	0.978–1.003	0.133
TC (mg/dl)	0.991	0.985–0.996	0.001***	1.002	0.996–1.008	0.564
SBP (mmHg)	0.982	0.972–0.992	<0.001***	0.983	0.972–0.995	0.005**
UA (µmol/L)	1.003	1.002–1.004	<0.001***	0.999	0.997–1.001	0.389
WBC (109/L)	1.083	1.043–1.124	<0.001***	1.070	1.022–1.120	0.004**
TG/HDL-C	1.075	1.000–1.155	0.050	1.167	1.062–1.282	0.001**

DOI: 10.7717/peerj.14346/table-4

Notes:

*P < 0.05

**P < 0.01

***P < 0.001

ALB

Albumin

ALT

Alanine aminotransferase

BUN

Blood urea nitrogen

Hb

Hemoglobin

SBP

Systolic blood pressure

TC

Total cholesterol

UA

Uric acid

WBC

White blood cells

Discussion

This study found that a higher TyG index and TG/HDL-C (markers of insulin resistance) indicated poorer prognosis and higher mortality in AMI patients without diabetes. Higher TyG and TG/HDL-C were also associated with higher inflammatory cells count and LDL-C, which meant they were closely related to severe inflammatory states and dyslipidemia.

Insulin resistance is characterized by reduced insulin sensitivity and impaired oxidation and glucose utilization (Wu & Ballantyne, 2020). Studies have shown that insulin resistance was found not only in type 2 diabetes but also in a wide range of diseases with abnormal physiological functions, triggering the proposal of the concept of insulin resistance syndrome (Fletcher & Lamendola, 2004; Willard, Stevenson & Steenkamp, 2016). Insulin resistance is associated with the incidence of CVD and could be used as a potential therapeutic target for CVD (Rutter et al., 2008; Bersch-Ferreira et al., 2018). The gold standard for defining insulin resistance is the euglycemic hyperinsulinemic clamp (Geloneze & Tambascia, 2006). The HOMA-IR index is also an effective marker for diagnosing insulin resistance (Geloneze & Tambascia, 2006). However, both of them are costly, and the identification of HOMA-IR is based on fasting insulin, which is not readily available in a large-scale cross-sectional study or a retrospective design. A study that used TyG and TG/HDL-C index as surrogate markers of insulin resistance to explore the longitudinal association of insulin resistance with the progression of arteriosclerosis in the hypertensive population found that the synergistic effect of hypertension and insulin resistance led to the adverse progression of arteriosclerosis (Wu et al., 2021). Therefore, TyG and TG/HDL-C are expected to perform as reliable and convenient markers, not inferior to the above two indicators.

Based on TyG and TG/HDL-C, we found insulin resistance was associated with inflammatory states and dyslipidemia. Previous studies have demonstrated that the onset age of stroke and hypertension decreased gradually with the increase of TyG index, which clarifies the negative effects of TyG index on the onset of diseases (Zhou et al., 2020; Gu et al., 2020). Insulin resistance is always accompanied by chronic low-grade inflammation (Perry et al., 2021). Increased WBC counts, neutrophil counts and lymphocyte counts are convenient markers for evaluating systemic inflammation (Perry et al., 2021). In our study, the WBC counts and lymphocyte counts were significantly associated with TyG index and TG/HDL-C, indicating that insulin resistance may result from systemic inflammation. A higher TyG index was accompanied by more severe inflammation. Moreover, the relationship between dyslipidemia and insulin resistance was reciprocal (Fassler et al., 2019). On the one hand, lipid accumulation is one of the leading causes of insulin resistance (Wang et al., 2021a; Wang et al., 2021b). On the other hand, insulin resistance is related to changes in lipoprotein and lipid metabolism (van der Kolk et al., 2019). We found that TC and LDL-C levels were positively associated with higher TyG and TG/HDL-C, which confirmed that insulin resistance was accompanied by dyslipidemia and other metabolic disorders.

Insulin resistance is not only important in AMI patients with diabetes, but also in patients without diabetes. Many clinical trials have determined that insulin resistance plays a vital role in the development of CVD and can result in myocardial ischemia (Klarin et al., 2017; Masaki et al., 2019). Furthermore, insulin resistance is currently considered an independent risk factor for cardiovascular death (Zhang et al., 2021). In a large, multi-ethnic and prospective cohort study of more than 15,000 postmenopausal women published in Circulation: Cardiovascular Quality and Outcomes, insulin levels, HOMA-IR, blood glucose, and TG/HDL-C were measured to assess insulin resistance. After adjusting for traditional cardiovascular risk factors such as age, race/ethnicity, smoking, total cholesterol, systolic blood pressure, and treatment with antihypertensive agents, insulin resistance was significantly associated with increased CVD risk (Schmiegelow et al., 2015). Previous studies on TyG and TG/HDL-C and insulin resistance mainly focused on AMI patients with diabetes (Zhang et al., 2021). However, we mainly explored the role of TyG and TG/HDL-C in predicting the prognosis of AMI patients without diagnosis of diabetes due to diabetes being an significant risk factor for AMI. It was interesting and significant for us to find insulin resistance can help predict short-term death in non-diabetes AMI patients.

There are still some limitations of our study. First, the long-term follow-up was not performed, limiting the survival analysis of high TyG and TG/HDL-C patients. Second, the roles of TyG and TG/HDL-C in AMI patients with or without diabetes were not compared. Third, HOMA-IR and serum insulin levels were not measured.

Conclusions

In conclusion, the emerged simple markers of insulin resistance, TyG index and TG/HDL-C are important predictors of in-hospital death in AMI patients without diabetes after adjusting significant variables such as age, SBP, BUN, WBC, ALT and ALB. For the population without diabetes, it is also worth emphasizing the importance of screening for insulin resistance and actively taking early intervention measures.

Supplemental Information