Analysis of the correlation between the serum triglyceride glucose index and the risk of death in patients on maintenance hemodialysis: a retrospective cohort study

Participants

This study included patients with MHD who were admitted to the Department of Nephrology of Chaozhou People’s Hospital of Guangdong Province between July 2012 and June 2022. The study adhered to all relevant tenets of the Declaration of Helsinki and was approved by the Medical Ethics Committee of the Chaozhou People’s Hospital (CZSRMYY-20200317032). Patients whose research has been approved by a medical ethics committee agree to the exemption. The clinical data of all patients were derived entirely from the medical records of the hospital. The inclusion criteria for this study were as follows: (1) aged ≥ 18 years, (2) switched from peritoneal dialysis (PD) to MHD; (3) renal transplantation received prior to undergoing MHD treatment; (4) transferred from other hemodialysis centers after receiving MHD treatment; (5) received regular hemodialysis treatment less than 3 months; and (6) absence of baseline fasting plasma glucose and triglyceride values. All patients were followed up until June 30, 2023. Finally, a total of 368 patients with MHD were included in this study (Fig. 1).

Data collection and definitions

All MHD patients were taken as the baseline value measured within one month before the start of the first hemodialysis, and demographic data and clinical data, including gender, age at the first dialysis session, duration of dialysis, comorbidities and clinical testing data, were extracted from the medical information system of Chaozhou People’s Hospital.

All MHD patients were rechecked for dialysis-related indicators every 3 months. After fasting overnight before undergoing hemodialysis, 10 ml of fasting venous blood and 5 ml of venous blood at the end of dialysis were collected on the first dialysis day of the week and then sent to the laboratory department of Chaozhou People’s Hospital. The routine hematological and biochemical parameters were detected via standard laboratory methods, and the corresponding indicators were calculated.

Specifically, the included parameters were as follows: (1) general biochemical index testing: routine indicators of renal function: urea, serum creatinine (SCr), uric acid (UA), fasting plasma glucose (FPG), triglyceride (TG), total cholesterol (TC), hemoglobin (HB), serum albumin (ALB), calcium (Ca), and homocysteine (Hcy); (2) special indicators of renal function: Kt/V (single pool) was calculated via the urea kinetic modelling (UKM) formula: Kt/V = −In (R − 0.008T) + (4 − 3.5R) × UF/W, wherein R = urea after dialysis/urea before dialysis, T = time of dialysis in hours, UF = ultrafiltration capacity (L), W = postdialysis patient weight (kg); corrected serum calcium (corrected Ca, mmol/L) = measured serum calcium (mmol/L) − [0.025 × albumin (g/L)] + 1; TyG index = ln [fasting serum triglyceride (mg/dL) × fasting plasma glucose (mg/dL)/2] (Guerrero-Romero et al., 2010). The above recorded values were the baseline values for 1 month before the start of the patient’s first hemodialysis session.

The criteria for hypertension included the following: (1) resting two non-contemporaneous systolic blood pressures ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg and (2) receiving antihypertensive therapy.

The criteria for diabetes mellitus (DM) include the following factors: (1) patients treated with insulin or oral hypoglycaemic drugs; (2) patients with clinically diagnosed type 1 or type 2 diabetes mellitus; (3) fasting plasma glucose (FPG) ≥ 7.0 mmol/L (126 mg/dL) or 75 g oral glucose tolerance test (OGTT) FPG ≥ 7.0 mmol/L (126 mg/dL) or random blood glucose ≥ 11.1 mmol/L (200 mg/dL) with classic symptoms such as dry mouth, polydipsia, and polyuria; and (4) glycosylated hemoglobin (HbA1c) ≥ 6.5%.

CVD refers to a history of angina, myocardial infarction, arrhythmia, angioplasty, coronary artery bypass grafting, heart failure, or peripheral vascular disease before admission to MHD.

Cardiovascular death was defined as death caused by CVDs such as arrhythmia, cardiac arrest, congestive heart failure, cerebrovascular accidents, ischemic brain injury, hypoxic encephalopathy, peripheral vascular disease, acute myocardial infarction, atherosclerotic heart disease, cardiomyopathy, and aneurysm dissection or rupture.

The criteria for hypertriglyceridemia was defined as triglyceride > 1.70 mmol/L.

The criteria for hypercholesteremia was defined as cholesteremia > 5.17 mmol/L.

Study endpoints

Primary endpoint events: all-cause death and cardiovascular death in patients on MHD.

Patients with MHD who experienced all-cause death were defined as a composite of cardiovascular death, infection, multiorgan failure, and other causes of death.

The follow-up endpoints were as follows: death, transfer to peritoneal dialysis, receipt of kidney transplant therapy, transfer to another blood purification centre for continued treatment, loss to follow-up, or end of the study.

All study endpoints were determined by experts and professors from the Blood Purification Center and hemodialysis nurses.

Statistical analysis

Continuous variables are expressed as the means ± standard deviations (means ± SDs) or medians (P₂₅, P₇₅), and the independent samples t test or Mann‒Whitney U test was used for comparisons between groups. Categorical variables are expressed as numbers (n) and percentages (%), and comparisons between groups were performed via either the chi-square test or Fisher’s exact test. Pearson correlation or Spearman correlation was used to analyse the correlation between the TyG index and the clinical parameters of MHD patients. Kaplan–Meier survival analysis and the log rank test were used to evaluate the survival status between different TyG index groups. Cox regression analysis was used to calculate the risk ratio (HR) and 95% confidence interval (CI) of the risk of death in patients with MHD. Time-dependent area under the curve (Time-AUC) analysis was performed to evaluate the TyG index for predicting death outcome in MHD patients. All the data were statistically analysed via SPSS 25.0 and R version 4.4.2. The above tests were all bilateral, and P < 0.05 was considered to indicate statistical significance.

Baseline characteristics of MHD patients grouped based on the median TyG index

According to the median TyG index, MHD patients were divided into a low TyG group (TyG < 8.86) and a high TyG group (TyG ≥ 8.86), and the clinical baseline characteristics of the total population and groups were compared (Table 1). The average age of the enrolled MHD patients was 58 (49, 65) years, and 196 were male, accounting for 53.26% of the patients. Compared with the patients in the low TyG group, there were more MHD patients with diabetes mellitus in the high TyG group (P = 0.017), and the overall levels of WBC, NE, PLT, FPG, TC, TG and TyG were greater (all P < 0.05), but the dialysis adequacy was worse (P < 0.001), the dialysis duration was shorter (P = 0.001), and there were no obvious differences between the remaining clinical parameters (all P > 0.05).

Correlation analysis between the TyG index and clinical parameters in MHD patients

The TyG index in MHD patients was significantly positively correlated with the WBC, NE, PLT, FPG, TG, TC and Kt/V values (all P ≤ 0.001), negatively correlated with K (r = −0.163, P = 0.002), negatively correlated with ALB (r = −0.134, P = 0.010), and not significantly correlated with other clinical parameters (all P > 0.05) (Table 2).

Follow-up outcomes of MHD patients

In the 11-year follow-up study, the median survival time of MHD patients was 43 months (ranging from 3 to 132 months), during which 167 (45.38%) all-cause deaths were recorded, 6 (1.63%) received kidney transplantation, 2 (0.54%) were transferred to PD treatment, and 51 (13.86%) were transferred to other blood purification centers for continued dialysis treatment. Among the patients who died of MHD, 116 (69.46%) died of CVD, 24 (14.37%) died of infection, 14 (8.38%) died of multiorgan failure, and 13 (7.78%) died of other causes. Among the 116 MHD patients who died due to CVD, 39 (33.62%) had congestive heart failure, 37 (31.90%) each had cerebrovascular accidents and sudden cardiac death, 2 (1.72%) had acute myocardial infarction, and 1 (0.86%) had aortic dissection rupture.

Analysis of mortality risk in patients on TyG and MHD

To show the outcomes of MHD patients with different basic TyG indices, we constructed Kaplan‒Meier survival plots for all-cause mortality and cardiovascular mortality according to the TyG median grouping (Figs. 2A and 2B). As shown in Fig. 2, MHD patients in the high TyG group had a greater risk of all-cause and cardiovascular mortality than did those in the low TyG group, with a log rank of 13.574 (P < 0.001) compared with 15.853 (P < 0.001). Although there was no statistically significant difference in all-cause mortality and cardiovascular mortality between the two TyG groups of MHD patients in DM group, MHD patients with high TyG index in diabetic group had a higher risk of death than the same type of non-diabetic patients as shown in Figs. 2C to 2F.

To determine if the increased risk of death among MHD patients in the high TyG group was due to the high prevalence of metabolic syndrome such as diabetes and lipid parameters in this group, we performed subgroup analyses and interaction tests on MHD patients in the high TyG group with or without diabetes, hypertriglyceridemia and hypercholesteremia. As shown in Table 3, the prediction of high TyG index on the all-cause and cardiovascular mortality of MHD patients was not affected by the prevalence of these conditions (all P for interaction > 0.05).

Prognostic value of the TyG index in MHD patients

Cox regression analysis was used to evaluate the predictive value of the TyG index for the mortality risk of MHD patients when it was used as a continuous variable or a categorical variable. The proportional hazards assumptions of TyG index and mortality in MHD patients based on Schoenfeld residuals were shown in Fig. S1–S4. The results showed that the risk proportional assumptions in Cox proportional risk models were congruent (P > 0.05). In multivariate Cox regression analysis, three models were developed to evaluate the predictive value of the TyG index for the primary endpoint, including variables significantly associated with the TyG index in Spearman correlation analysis (P < 0.05) and/or clinical variables identified as important to death in patients with MHD as confounders. The results of unadjusted Cox regression analysis revealed that the baseline TyG index was a predictor of all-cause and cardiovascular mortality in MHD patients (HR 1.446, 95% CI [1.146–1.824], P = 0.002 vs. HR 1.443, 95% CI [1.072–1.941], P = 0.015). Compared with those in the low TyG group, the risks of all-cause and cardiovascular mortality in MHD patients in the high TyG group were HR 1.771, 95% CI [1.299–2.414], P < 0.001 and HR 1.781, 95% CI [1.207–2.556], P = 0.003, respectively. After adjusting for gender and age at first dialysis in Model 1, the baseline TyG index was also a predictor of all-cause and cardiovascular mortality in MHD patients (HR 1.429, 95% CI [1.122–1.819], P = 0.004 vs. HR 1.433, 95% CI [1.065–1.929], P = 0.017). Compared with those in the low TyG group, the risks of all-cause mortality and cardiovascular mortality in MHD patients in the high TyG group were HR 1.716, 95% CI [1.258–2.340], P = 0.001 and HR 1.773, 95% CI [1.017–2.371], P = 0.004, respectively. After Model 1 and adjusting for complications such as CVD, hypertension and DM in Model 2, the baseline TyG index was still a predictor of all-cause and cardiovascular mortality in MHD patients (HR 1.604, 95% CI [1.151–2.237], P = 0.005 vs. HR 1.582, 95% CI [1.101–2.571], P = 0.049). Compared with those in the low TyG group, the risks of all-cause and cardiovascular mortality in MHD patients in the high TyG group were HR 1.557, 95% CI [1.133–2.141], P = 0.006 and HR 1.734, 95% CI [1.171–2.219], P = 0.007, respectively. Model 3 included Model 2 and WBC, NE, HB, PLT, K, corrected Ca, FPG, ALB, TG, TC and Kt/V, and the baseline TyG index was always an independent predictor of all-cause and cardiovascular mortality in MHD patients (HR 1.550, 95% CI [1.216–1.975], P < 0.001 vs. HR 1.376, 95% CI [1.009–1.875], P = 0.044). Compared with those in the low TyG group, the risks of all-cause and cardiovascular mortality in MHD patients in the high TyG group were HR 1.790, 95% CI [1.177–2.721], P = 0.006 and HR 1.735, 95% CI [1.022–2.612], P = 0.022, respectively (Table 4).

To assess the value of the TyG index on the risk of death in patients with MHD, we generated time-dependent area under the curve (time-AUC) for further analysis. The predictive time-AUC values of the TyG index for all-cause and cardiovascular death in MHD patients were between 0.698 to 0.819. Meanwhile, the overall trends of time-AUC were upward with the extension of follow-up time (Fig. 3 and Table 5).

TyG index, whether as a continuous variable or a nominal variable, can effectively predict the survival time of MHD patients (with a differentiation of 0.752 to 0.806), and improve the prediction ability of the new clinical prediction model for all-cause death and cardiovascular death of MHD patients, with an improvement ratio of at least 10.11% and a maximum of 56.89% (Table 6).