This nationwide cohort study was conducted using the National Health Information Database (NHID). The public healthcare system in South Korea operates under the National Health Insurance Service (NHIS), a universal and compulsory single-insurer system. The NHIS forms the public health database of NHID and includes a comprehensive range of information, such as demographic data, diagnoses (coded using the 10th International Classification of Diseases [ICD-10]), insured medical services, and laboratory data of the general population (13). In addition, all insured individuals are required to undergo annual or biannual National Health Screening programs. Healthcare utilization data were integrated with health screening records to constitute a study cohort. This study was approved by the Institutional Review Board (KHUH 2021-11-056), and the use of the NHIS database (NHIS-2023-1-380) was authorized.

Patients with CKD who participated in the National Health Screening between 2012 and 2015 were screened for inclusion. As the screening program is performed annually or biennially for all adults in South Korea, CKD was defined when either presence of proteinuria more than 1+ or eGFR less than 60mL/min/1.73m² was consistently observed across at least two consecutive examinations. eGFR was calculated using the 2009 Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI) (14). Exclusion criteria included individuals younger than 20 years, those with a history of ischemic heart disease (I20–25 in ICD-10 codes) or stroke (I60–69 in ICD-10 codes) at enrollment, those with eGFR less than 15mL/min/1.73m², those receiving hemodialysis or peritoneal dialysis, and those with missing medical information. A total of 47,592 participants were included in the final analysis (Figure 1).

Baseline demographic and clinical variables collected included age, sex, body mass index (BMI), height, weight, waist circumference, Charlson comorbidity index (CCI), CKD-EPI, systolic and diastolic blood pressure, fasting blood glucose, hemoglobin, total cholesterol, triglyceride, high density lipoprotein (HDL) and low density lipoprotein (LDL) cholesterol levels. The CCI was determined by calculating the score allocated to each comorbid condition using the ICD-10 diagnoses. Metabolic syndrome was defined as the presence of three or more of the following (13, 15): elevated blood pressure (≥130 mmHg systolic and/or ≥85 mmHg diastolic and/or use of antihypertensive drugs), hypertriglyceridemia (≥150 mg/dL or treatment for elevated triglycerides and/or use of lipid-lowering drugs), low high-density lipoprotein cholesterol (<40 mg/dL in men,<50 mg/dL in women), elevated fasting glucose (≥110 mg/dL and/or use of antidiabetic agents) or increased waist circumference (>90 cm for men, >80 cm in women). Self-reported lifestyle variables included smoking, alcohol consumption, and physical activity. Smoking status was classified as non-smoker, former smoker, or current smoker, and alcohol consumption was classified as 0 or 1 drinking per week, or drinking ≥ 2 times per week. Moderate physical activity was defined as exercise ≥ 3 times per week. All baseline data were obtained at the time when participants first fulfilled the CKD criteria and follow-up began thereafter.

Participants were categorized into four groups based on their TyG index levels. The TyG index was calculated using the following formula (16);

The four quartiles based on TyG index levels were classified as follows; quartile 1,< 8.46; quartile 2, ≥ 8.46 and< 8.80; quartile 3, ≥ 8.80 and< 9.18; and quartile 4, ≥ 9.18.

The primary end points of this study were the occurrence of CV events, progression to ESKD, and all-cause mortality. CV events included myocardial infarction and ischemic stroke. Myocardial infarction and ischemic stroke were identified using ICD codes I20–25 and I60-69, respectively, which were newly prescribed during hospitalization or on at least two outpatient visits during the follow-up period. Progression to ESKD was defined as sustained requirement for hemodialysis, peritoneal dialysis, or renal transplantation. Participants who progressed to ESKD were identified by (i) specific insurance codes for renal replacement therapy to receive additional insurance coverage, (ii) procedure codes for renal replacement therapy prescribed for more than 3 months, and/or (iii) procedure or diagnostic codes for kidney transplantation. Subgroup analyses were performed based on age, sex, BMI, presence of metabolic syndrome, urine protein, and diabetes to account for diverse patterns of the TyG index associated with older age, obesity, kidney injury, glucose dysregulation, and metabolic dysfunction, all of which may affect clinical outcomes. Patients were followed up from January 1, 2016, until the date of death, last recorded checkup, or December 31, 2020, whichever occurred first.

Quantitative data were expressed as mean ± standard deviation, and categorical data were presented as numbers and percentages. Baseline characteristics and differences between groups were evaluated using chi-square tests and analysis of variance, as appropriate. Incident rates were calculated as the number of events per 1,000 person-years. The association between the TyG index and CV events, progression to ESKD, and all-cause mortality was analyzed using univariable and multivariable Cox regression models. Covariates in the multivariate Cox regression model included age, sex, BMI, smoking, drinking, physical activity, CCI, systolic blood pressure, eGFR, hemoglobin, LDL-cholesterol. Kaplan-Meier plots were generated to illustrate cumulative event rates for each clinical outcome. In the analysis treating exposures as a continuous variable, the multivariable-adjusted restricted cubic spline curves were constructed with three knots. In addition to the main effects in the fully adjusted models, interactions between the TyG index and subgroup characteristics or eGFR were evaluated. Statistical significance was defined as P value< 0.05. All statistical analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC) and R (version 3.4.1; The R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org).

Baseline demographics and clinical parameters across to TyG index quartiles are presented in Table 1. The mean TyG index was 8.10 ± 0.27 in quartile 1 (n = 11,900), 8.64 ± 0.12 in quartile 2 (n = 11,890), 9.04 ± 0.13 in quartile 3 (n = 11,904), and 9.75 ± 0.45 in quartile 4 (n = 11,898), with an overall mean of 8.88 ± 0.67 for the study population. Participants in quartile 4 were younger and included a higher proportion of males compared with those in the other quartiles. Higher TyG index quartiles were associated with increased BMI, greater waist circumference and greater prevalence of unfavorable lifestyle factors, including current smoking, frequent alcohol consumption, and physical inactivity. Blood pressure, fasting plasma glucose, and triglyceride level increased progressively with the rise of TyG index level.

During a median follow-up of 90.8 ± 18.5 months, CV events and death were observed in 5,514 patients (11.6%), progression to ESKD in 3,475 patients (7.3%), and all-cause mortality in 7,502 patients (15.8%). The cumulative event rate of CV events varied significantly across TyG index quartiles, with the highest rate observed in quartile 4 (Figure 2). The cumulative event rate of progression to ESKD increased with higher TyG index quartiles.

Table 2 shows the number of events, incidence rate in person-years, and hazard ratios (HRs) for all-cause mortality, CV events, and progression to ESKD. In univariate analyses, individuals in higher TyG index quartiles exhibited significantly greater risk of CV events and progression to ESKD. The association between the TyG index and risk of CV events remained significant after adjusting for multiple factors. Compared to TyG index quartile 1, the adjusted HRs (95% CI) for CV events were 1.08 (1.00 - 1.17), 1.13 (1.04 - 1.22), and 1.39 (1.28 - 1.51) for quartiles 2, 3, and 4, respectively. In a similar manner, a significant association was observed between the TyG index and the risk of progression to ESKD across all TyG index quartiles, that the risk increasing in higher quartiles as follows: adjusted HRs with 95% CI as 1.22 (1.08 - 1.38) in quartile 2, 1.35 (1.18 - 1.54) in quartile 3, 1.86 (1.65 - 2.10) in quartile 4. The risk of all-cause mortality was also significantly elevated in TyG index quartile 3 (adjusted HR 1.08, 95% CI 1.02 - 1.16) and 4 (adjusted HR 1.31, 95% CI 1.23 - 1.41). Additional adjustment models are provided in the Supplementary Table 1.

Restricted cubic spline curves indicated a continuous association between the TyG index and HRs for CV events, progression to ESKD, and all-cause mortality (Figure 3). The HRs for CV events and progression to ESKD increased with higher TyG index values, with risk accelerating at elevated levels. Meanwhile, the risk of all-cause mortality increased markedly only at higher TyG index values, with no significant changes observed at lower values.

Pre-specified subgroup analysis based on age, sex, BMI, urine protein, diabetes, and metabolic syndrome are presented in Table 3. Risks of CV events, progression to ESKD, and all-cause mortality generally increased with each TyG index quartile increment across all subgroups. A more pronounced increase in HRs for CV events was observed in patients aged< 65 years (P for interaction<0.001), those with proteinuria (P for interaction<0.001), and those with diabetes (P for interaction = 0.003) compared to patients aged ≥ 65 years, those without proteinuria and those without diabetes. The risk of progression to ESKD was greater in participants with proteinuria than in those without proteinuria (P for interaction = 0.001). Similarly, the increase in all-cause mortality risk was more prominent in participants aged< 65 years and those with diabetes or metabolic syndrome. Significant interactions were observed between the TyG index and predefined subgroup characteristics.

To evaluate whether the relationship between kidney function and adverse clinical outcomes differed by TyG index, the risks of CV events, progression to ESKD, and all-cause mortality were evaluated across eGFR levels stratified by TyG index quartiles using spline curves (Figure 4). A J-shaped relationship was observed between eGFR and CV event risk, while progression to ESKD followed an L-shaped pattern and all-cause mortality exhibited a U-shaped association. The shape of the curve for CV events remained consistent across TyG index quartiles, suggesting a relatively uniform association (P for interaction = 0.616). However, the increase in ESKD risk with declining eGFR was more pronounced in CKD patients with higher TyG index quartiles (P for interaction< 0.001). Similarly, the interaction between TyG index and eGFR in relation to all-cause mortality was statistically significant, indicating a stronger association in higher TyG index quartiles (P for interaction< 0.001).