Association between triglyceride glycemic index and ejection fraction preserved heart failure in hypertensive patients

Study design and participants

This is a single-center retrospective study of patients diagnosed with essential hypertension and treated at the Department of Cardiology, First Affiliated Hospital of Xinjiang Medical University from December 2018 to September 2020. This is a sub-study of a larger study aimed at elucidating the pathogenesis of cardiovascular disease. We retrospectively analyzed data from that study to determine whether there is an association between insulin resistance and the development of heart failure with preserved ejection fraction in hypertensive patients. Written informed consent was obtained from all participants. This project is in line with the Declaration of Helsinki, and the study protocol was approved by the Human Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University (approval number: S220722-25). Participants’ information is recorded in the JiaHe eCase system.

Inclusion criteria

Participants were included if they met the standard definition of the International Society of Hypertension Global Hypertension Practice Guidelines 2020 (Unger et al., 2020). The criteria were: (1) Essential hypertension was diagnosed as repeated blood pressure measurements with office systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg or taken anti-hypertensive treatment. (2) Complete clinical data collection.

Exclusion criteria

All participants with one of the following conditions were excluded. (1) White coat hypertension is defined as an elevated office blood pressure in the presence of a normal out-of-office blood pressure, typically with an office blood pressure of at least 140/90 mmHg and a 24-h average blood pressure of less than 130/80 mmHg (Nuredini et al., 2020). (2) Secondary hypertension; (3) other structural heart disease including congenital, valvular and pericardial diseases; (4) hypertrophic cardiomyopathy; (5) HF due to other cardiac disease rather than hypertension (6) abnormal thyroid function; (7) Mental illness or psychological disorders. (8) Severe liver and renal dysfunction; (9) unwilling to participate in the study or refused to sign the informed consent form.

General clinical data collection and laboratory test

Upon admission, detailed patient demographics, medical history, and records were meticulously documented by a physician specializing in cardiovascular medicine. This comprehensive data collection included gender, age, height, weight, history of smoking and alcohol consumption, systolic and diastolic blood pressure (SBP and DBP), past medical history detailing diabetes mellitus or coronary artery disease (CAD), and current medication usage. Eight milliliters of fasting venous blood was drawn from each patient early in the morning of the day following hospitalization (after at least 8 h of fasting) for assessment of fasting blood glucose (FBG), triglycerides, total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), uric acid (UA), creatinine, and N-terminal pro-B-type natriuretic peptide (NT-proBNP). The TyG index was calculated using the formula: TyG = ln [fasting triglycerides (mg/dL) × fasting glucose (mg/dL)/2] (Tao et al., 2022). Body Mass Index (BMI) was calculated as weight (kg) divided by height (m) squared. Patients who smoked regularly within the past 6 months were classified as current smokers. Alcohol consumption was defined as drinking at least 100 g of alcoholic beverages per week in the past month. Diabetes mellitus was defined as having a history of or suffering from diabetes mellitus, and/or having at least two fasting plasma glucose levels >7.0 mmol/L, or at least one random blood glucose value >11.1 mmol/L, or a glycosylated hemoglobin (HbA1c) >6.5%, or a 2-h OGTT blood glucose ≥11.1 mmol/L before the current admission, or being on antidiabetic medication.

Echocardiographic evaluation

Transthoracic echocardiography was performed on all patients within 24 h of admission utilizing the Vivid 7 Ultrasound System (GE Medical Systems, Chicago, IL, USA). This examination, conducted by a specialized sonographer, meticulously recorded a series of echocardiographic parameters. These parameters included: (1) Left atrial diameter (LAD): indicative of atrial size and potentially atrial pressure. (2) LV end-diastolic diameter (LVEDD): reflecting LV size and volume status at the end of diastole. (3) Interventricular septum thickness (IVS): measuring the thickness of the septum separating the left and right ventricles. (4) LV posterior wall thickness (LVPWT): assessing the thickness of the LV’s posterior wall. (5) E/A ratio: the ratio of peak velocity flow from early LV filling (E wave) to late filling (A wave) via the mitral valve, offering insight into diastolic function. (6) E/e′ ratio: the ratio of peak E wave velocity at the mitral valve orifice to peak velocity of early diastolic mitral annulus motion (e′), providing an estimation of LV filling pressure. (7) LV mass index (LVMI): a quantification of LV mass relative to body surface area (Ren et al., 2020). (8) LV ejection fraction (LVEF): indicating the percentage of blood volume ejected from the LV with each heartbeat, calculated using the biplane method of disks (modified Simpson’s rule) (Brooks et al., 2016).

Diagnosis of heart failure with preserved ejection fraction

The diagnosis of heart failure with preserved ejection fraction (HFpEF) adheres to the criteria set forth by the 2016 European Society of Cardiology guidelines, encompassing a comprehensive approach that mandates the fulfillment of all the following points: (1) clinical signs or symptoms of heart failure. (2) elevated N-terminal pro-B-type natriuretic peptide (NT-proBNP) > 125 pg/ml) (3) evidence of structural changes in the heart (left atrial volume index ≥ 34 ml/m² or LVMI ≥ 115 g/m² for men and (≥95 g/m² for women); (4) normal or mildly abnormal LVEF ≥ 50% (Ponikowski et al., 2016).

Statistical analyses

All statistical analyses in this study were conducted using R software (version 4.2.3). Initially, participants were stratified into three groups according to the tertiles of the triglyceride glucose (TyG) index to compare baseline characteristics. Continuous variables with normal distribution were presented as mean ± standard deviation (SD), while those not normally distributed were expressed as the median along with the 25th and 75th percentiles. For the comparison between groups, one-way Analysis of Variance (ANOVA) was employed for normally distributed variables, whereas the Kruskal-Wallis H test was utilized for variables without a normal distribution. Count data were denoted as frequencies (percentages), with inter-group comparisons facilitated by the chi-square test. The relationship between the TyG index and various functional and structural cardiac parameters was assessed using Spearman’s correlation analysis. Furthermore, logistic regression models were implemented to explore the association between HFpEF and different variables among patients with hypertension. To investigate the potential non-linear relationships between the TyG index and HFpEF, restricted cubic spline (RCS) plots were utilized. The diagnostic efficiency of various indices for HFpEF in the hypertensive patient cohort was evaluated through receiver operating characteristic (ROC) curves. Statistical significance was determined at a two-sided p-value of less than 0.05 across all statistical tests conducted.

Baseline characteristics of hypertensive patients grouped by TyG index tertiles

Initially, 1,602 patients with essential hypertension were enrolled. Of these, 422 patients were excluded because they did not fulfill the inclusion criteria, and 188 patients were excluded because they were missing relevant clinical data. Finally, 992 patients (634 males and 358 females) who completed the study were included in the data analysis (Fig. 1).

Patients were categorized into three tertiles based on the percentile of the TyG index: Tertile 1 (TyG ≤ 8.483), Tertile 2 (8.483 < TyG ≤ 9.058), and Tertile 3 (TyG > 9.058), and their baseline characteristics were compared (Table 1). In Tertile 3, the median age of patients was 61 years, and there were higher proportions of BMI, SBP, male gender, prevalence of diabetes, as well as the use of ACEI/ARB and diuretics. Additionally, this group exhibited lower levels of HDL-C, and higher levels of serum creatinine, uric acid, and NT-pro BNP. In terms of echocardiographic data, Tertile 3 also had higher LVMI and MVE/e′. Regarding the incidence of HFpEF, there were nine cases (2.7%) in Tertile 1, 41 cases (12.4%) in Tertile 2, and 152 cases (45.9%) in Tertile 3. Additionally, we further compared TyG index levels between hypertensive patients with and without HFpEF, finding that TyG index levels were significantly higher in patients with both hypertension and HFpEF (9.446 [9.064, 9.747] vs. 8.599 [8.250, 8.993], p < 0.001) (Fig. 2).

Correlation analysis of TyG index with important cardiac indices in hypertensive patients

We analyzed the association of the TyG index with several key ultrasound indices and NT-pro BNP using Spearman correlation analysis and plotted scatter plots (Fig. 3). The analysis revealed that the TyG index had a positive correlation with LVMI (rs = 0.200, p < 0.001), MVE/e′ (rs = 0.163, p < 0.001), and NT-pro BNP (rs = 0.163, p < 0.001), and a negative correlation with LVEF (rs = −0.087, p < 0.018).

Association between TyG index and occurrence of HFpEF in patients with essential hypertension

Logistic regression analysis revealed an association between the TyG index and the occurrence of HFpEF in hypertensive patients (Table 2). The base model showed an OR of 4.739 (95% CI [3.738–6.082], p < 0.001). In Model 1, adjusted for demographic details such as age and gender, the adjusted OR was 4.842 (95% CI [3.747–6.344], p < 0.001). Model 2, further adjusted for laboratory indices in addition to the adjustments in Model 1, confirmed that the TyG index remained an independent predictor with an OR of 5.127 (95% CI [3.894–6.856], p < 0.001). After standardization and categorization of the TyG index data, it was found that the standardized TyG index still independently associated with HFpEF in the fully adjusted Model 2, with an OR of 3.267 (95% CI [2.677–4.031], p < 0.001). Moreover, compared to Tertile 1, the OR was 5.355 (95% CI [2.587–12.271], p < 0.001) for Tertile 2 and 34.214 (95% CI [17.025–77.350], p < 0.001) for Tertile 3, indicating a significant increase in risk across tertiles. We conducted an in-depth analysis of the nonlinear relationship between the TyG index and HFpEF incidence using a restricted cubic spline plot (RCS). This analysis revealed a significant nonlinear association between the TyG index and HFpEF occurrence in hypertensive patients, with a nonlinear p value of less than 0.001. Notably, as the TyG index increased, the incidence of HFpEF also rose significantly, delineating a roughly S-shaped curve (Fig. 4).

Predictive power of TyG index for HFpEF

In evaluating the predictive performance for HFpEF among hypertensive patients, our ROC analysis revealed distinct efficacies across various markers and their combinations (Table 3). The TyG Index alone showed a promising AUC of 0.824 (95% CI [0.795–0.854]), indicating its substantial role in risk identification. Similarly, NT-proBNP demonstrated excellent sensitivity, with an AUC of 0.840 (95% CI [0.816–0.864]), while LVMI also presented strong predictive power, marked by an AUC of 0.847 (95% CI [0.820–0.875]). On the other hand, MVE/e′ and LVEF displayed moderate and relatively low predictive values, with AUCs of 0.688 (95% CI [0.648–0.729]) and 0.538 (95% CI [0.490–0.586]), respectively. The analysis of combined markers significantly enhanced predictive accuracy; notably, the TyG+LVMI combination emerged as the most potent predictor, achieving the highest AUC of 0.907 (95% CI [0.887–0.927]) (Fig. 5).