Triglyceride-glucose index associated with the risk of future renal function decline: a 13-year prospective cohort study Authors: Liying Tian, Xinyu Wu, Qi Qi, Xinyi Li, Bocheng Yue, Bo Yang, Nan Wang, Quanle Han, Shouling Wu, Kangbo Li Published: 2026 ABSTRACT We aimed to characterize the role of triglyceride–glucose index (TyG index), a surrogate marker of insulin resistance (IR), and renal function decline (RFD) in a large community-based prospective cohort study. A sum of 85,569 participants were included in the current study, all participants were categorized based on quartiles of the TyG index levels. The primary outcome of this study was future RFD, which was defined as a reduction of ≥ 25% in the estimated glomerular filtration rate from the baseline to below 60 mL/min/1.73 m2. Incidence rate was expressed as cases per 1,000 person-years. The association between the TyG index and the risk of RFD was investigated using Cox proportional hazards models. During a median follow-up period of 14.01 years (interquartile range: 13.63–14.21 years), a total of 3668 cases RFD were recorded. The incidence rate of incident RFD increased substantially with the quartiles of TyG index, presenting 2.529, 3.110, 3.431, and 4.050 per 1,000-person year in quartile 1 to quartile 4. Multivariate Cox regression analysis revealed that the hazard ratio of RFD risk was 1.656 (95% confidence interval: 1.479–1.854) in quartile 4 compared with the quartile 1. The findings of this study indicate that an elevated TyG index correlates with an increased risk of future RFD, suggesting that IR could play a pivotal role in the onset of RFD. Introduction The Global Burden of Disease studies indicate that chronic kidney disease (CKD) has become a prominent cause of mortality on a global scale1. CKD presents a substantial challenge to public health and represents a significant burden across the world2. It affects more than 10% of the global population, which amounts to over 800 million individuals3. CKD manifests when renal function is impaired for a duration exceeding 3 months, defined by either a glomerular filtration rate (GFR) falling below 60 mL/min/1.73 m2 or the detection of kidney damage indicators, such as albuminuria, abnormalities in urine sediment, or other disturbances in electrolytes4. In the context of CKD, physiological adaptations take place over time to sustain GFR and facilitate the balance of creatinine and various electrolytes5. End-stage renal disease (ESRD) represents the terminal phase, stage 5, and the most critical level of CKD. It pertains to individuals whose estimated glomerular filtration rate (eGFR) is below 15 mL per minute per 1.73 m2 of body surface area, or those who require dialysis regardless of their GFR6. Given that CKD is characterized by a gradual deterioration and loss of kidney function, it underscores the necessity of recognizing and comprehending renal function decline (RFD) as a foundation for formulating treatment approaches aimed at preventing the onset and advancement of CKD. On the other hand, insulin resistance (IR) is one of cardiometabolic syndromes that affects many organs7, it has been identified as a risk factor of CKD8. Research shows that IR even occurs in the initial stages of CKD9. However, the association between the IR and future RFD has not been fully explored. Therefore, in the current study, by analyzing data of 85,569 participants from the Kailuan study, we aimed to characterize the role of triglyceride–glucose index (TyG index), a surrogate marker of IR, and RFD in a large community-based prospective cohort study. Methods Study population The Kailuan study, conducted in the Kailuan community in Tangshan, China, is a prospective cohort study. From June 2006 to October 2007, a total of 101,510 participants were included in the initial baseline survey and underwent a thorough biennial health examination. We excluded 14,486 participants with a baseline eGFR < 60 ml/min/1.73 m2, and 1455 participants with incomplete data at the beginning of the study. In the end, a sum of 85,569 participants were included in the current study (Fig. 1).Fig. 1Study flow diagram. TyG index: triglyceride–glucose index, eGFR: estimated glomerular filtration rate. Study flow diagram. TyG index: triglyceride–glucose index, eGFR: estimated glomerular filtration rate. Data collection and TyG index calculation Demographic characteristics and lifestyle factors (smoking habits and alcohol intake) were obtained through a standardized questionnaire administered by trained personnel. The participants’ educational level was categorized as illiterate or primary school, middle school, and high school or above. All blood samples were analyzed on the same day using an auto-analyzer (Hitachi 747, Hitachi, Tokyo, Japan). The biochemical indicators tested included fasting blood glucose (FPG), total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), high-sensitivity C-reactive protein (hs-CRP). The TyG index was determined as the formula Ln [TG (mg/dl) × FBG (mg/dl)/2]10. All participants were categorized based on quartiles of the TyG index levels: quartile 1 (3.61 ≤ TyG index < 8.15), quartile 2 (8.15 ≤ TyG index < 8.55), quartile 3 (8.55 ≤ TyG index < 9.02), quartile 4 (9.02 ≤ TyG index < 12.51). Future RFD and follow-up The primary outcome of this study was future RFD. According to the previously described criteria, RFD was defined as a reduction of ≥ 25% in the eGFR from the baseline to below 60 mL/min/1.73 m11–13. eGFR is calculated according to the CKD-Epidemiology Collaboration equation14. Follow-up started after the completion of the initial examination and ended until their death or December 31, 2020. All participants were visited biennially to measure or record the above clinical and laboratorial parameters mentioned in initial examination. Statistical analysis SAS 9.4 software was used to analyze the data. P-value < 0.05 was considered statistically significant (two-sided). It was assumed that the missing covariate data were random, and complete conditional random imputation was employed to address the missing covariate values. The imputation process utilized discriminant analysis for categorical variables and regression modeling for continuous variables. Baseline characteristics were presented as means ± standard deviation (SD) or median with interquartile range. The Chi-Square test was utilized for categorical variables, while one-way analysis of variance or Kruskal–Wallis test was used for continuous variables to compare baseline characteristics among quartiles. A log-rank test is used to compare Kaplan–Meier cumulative incidence curves. Incidence rate was expressed as cases per 1,000 person-years. The association between the TyG index and the risk of RFD was investigated using Cox proportional hazards models (the proportional hazard assumption was satisfied). Model 1 was a univariate Cox regression analysis model. Model 2 was adjusted for age and sex. Model 3 was adjusted for age, sex, body mass index (BMI), systolic blood pressure (SBP), TC, LDL-C, hs-CRP. Model 4 was adjusted for age, sex, BMI, SBP, TC, LDL-C, hs-CRP, eGFR, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. The restricted cubic spline function analysis was used to analyze whether the TyG index was nonlinearly correlated with the risk of RFD. In addition, to ensure the robustness of the results, the following sensitivity analyses were conducted based on Model 4: (1) Excluding participants experienced RFD during the first 2 years of follow-up. (2) Excluding participants experienced myocardial infarction and stroke during follow-up. (3) Excluding all-cause death during follow-up. Furthermore, subgroup analyses were conducted according to baseline age (≥ 60/< 60 years), sex (men/women), and BMI (≥ 28 kg/m2/< 28 kg/m2)15. Results Population characteristics A total of 85,569 eligible participants were included in the current study, with an average age of 50.8 ± 12.2 years, and 80.38% of them were male. The mean value of the TyG index was 8.62. In comparison to those in quartile 1, individuals in the other quartiles tended to be older, more male, with elevated BMI, blood pressure, FBG, TC, TG, and hs-CRP levels, more likely to smoke, had a higher prevalence of hypertension and diabetes (Table 1).Table 1Baseline characteristics.VariablesTotal (n = 85,569)Quartile 1 (n = 21,392)Quartile 2 (n = 21,395)Quartile 3 (n = 21,390)Quartile 4 (n = 21,392)PAge, years50.8 ± 12.249.5 ± 13.450.8 ± 12.351.5 ± 11.851.3 ± 10.9 < 0.001Male, n (%)68,780 (80.38)15,745 (73.60)17,128 (80.06)17,653 (82.53)18,254 (85.33) < 0.001BMI, kg/m224.96 ± 3.4923.34 ± 3.2224.53 ± 3.2925.56 ± 3.3126.40 ± 3.35 < 0.001SBP, mmHg129.6 ± 20.5123.8 ± 19.5128.6 ± 20.0131.4 ± 20.3134.8 ± 20.4 < 0.001DBP, mmHg83.0 ± 11.679.4 ± 10.982.4 ± 11.384.1 ± 11.486.1 ± 11.8 < 0.001TyG index8.62 ± 0.687.83 ± 0.278.36 ± 0.118.77 ± 0.139.54 ± 0.45 < 0.001FBG, mmol/l5.43 ± 1.594.79 ± 0.675.09 ± 0.765.42 ± 1.096.43 ± 2.52 < 0.001TC, mmol/l4.95 ± 1.094.64 ± 0.924.91 ± 0.955.11 ± 0.995.15 ± 1.37 < 0.001TG, mmol/l1.25 (0.87–1.89)0.69 (0.57–0.80)1.07 (0.95–1.20)1.52 (1.34–1.75)2.66 (2.13–3.72) < 0.001HDL-C, mmol/l1.54 ± 0.401.57 ± 0.411.55 ± 0.391.52 ± 0.391.51 ± 0.41 < 0.001LDL-C, mmol/l2.31 ± 0.912.15 ± 0.932.35 ± 0.872.42 ± 0.882.33 ± 0.95 < 0.001hs-CRP, mg/l0.80 (0.30–2.30)0.60 (0.21–2.01)0.73 (0.29–2.10)0.90 (0.33–2.30)1.03 (0.40–2.75) < 0.001Hypertension, n (%)35,460 (41.44)5977 (27.94)8261 (38.61)9820 (45.91)11,402 (53.30) < 0.001Diabetes, n (%)7323 (8.56)231 (1.08)552 (2.58)1541 (7.20)4999 (23.37) < 0.001Higher income, n (%)12,580 (14.70)3363 (15.72)3007 (14.05)3092 (14.46)3118 (14.58) < 0.001Smoking, n (%)30,819 (36.02)7251 (33.90)7469 (34.91)7776 (36.35)8323 (38.91) < 0.001Drinking, n (%)33,614 (39.28)8128 (38.00)8041 (37.58)8427 (39.40)9018 (42.16) < 0.001Physical exercise, n (%)77,927 (91.07)19,577 (91.52)19,525 (91.26)19,399 (90.69)19,426 (90.81)0.020Higher educational background, n (%)17,963 (20.99)5347 (25.00)4377 (20.46)4270 (19.96)3969 (18.55) < 0.001BMI: body mass index, SBP: systolic blood pressure, DBP: diastolic blood pressure, TyG index: triglyceride–glucose index, FBG: fasting blood glucose, TC: total cholesterol, TG: triglycerides, HDL-C: high-density lipoprotein cholesterol, LDL-C: low-density lipoprotein cholesterol, hs-CRP: high-sensitivity C-reactive protein. Table 1Baseline characteristics.VariablesTotal (n = 85,569)Quartile 1 (n = 21,392)Quartile 2 (n = 21,395)Quartile 3 (n = 21,390)Quartile 4 (n = 21,392)PAge, years50.8 ± 12.249.5 ± 13.450.8 ± 12.351.5 ± 11.851.3 ± 10.9 < 0.001Male, n (%)68,780 (80.38)15,745 (73.60)17,128 (80.06)17,653 (82.53)18,254 (85.33) < 0.001BMI, kg/m224.96 ± 3.4923.34 ± 3.2224.53 ± 3.2925.56 ± 3.3126.40 ± 3.35 < 0.001SBP, mmHg129.6 ± 20.5123.8 ± 19.5128.6 ± 20.0131.4 ± 20.3134.8 ± 20.4 < 0.001DBP, mmHg83.0 ± 11.679.4 ± 10.982.4 ± 11.384.1 ± 11.486.1 ± 11.8 < 0.001TyG index8.62 ± 0.687.83 ± 0.278.36 ± 0.118.77 ± 0.139.54 ± 0.45 < 0.001FBG, mmol/l5.43 ± 1.594.79 ± 0.675.09 ± 0.765.42 ± 1.096.43 ± 2.52 < 0.001TC, mmol/l4.95 ± 1.094.64 ± 0.924.91 ± 0.955.11 ± 0.995.15 ± 1.37 < 0.001TG, mmol/l1.25 (0.87–1.89)0.69 (0.57–0.80)1.07 (0.95–1.20)1.52 (1.34–1.75)2.66 (2.13–3.72) < 0.001HDL-C, mmol/l1.54 ± 0.401.57 ± 0.411.55 ± 0.391.52 ± 0.391.51 ± 0.41 < 0.001LDL-C, mmol/l2.31 ± 0.912.15 ± 0.932.35 ± 0.872.42 ± 0.882.33 ± 0.95 < 0.001hs-CRP, mg/l0.80 (0.30–2.30)0.60 (0.21–2.01)0.73 (0.29–2.10)0.90 (0.33–2.30)1.03 (0.40–2.75) < 0.001Hypertension, n (%)35,460 (41.44)5977 (27.94)8261 (38.61)9820 (45.91)11,402 (53.30) < 0.001Diabetes, n (%)7323 (8.56)231 (1.08)552 (2.58)1541 (7.20)4999 (23.37) < 0.001Higher income, n (%)12,580 (14.70)3363 (15.72)3007 (14.05)3092 (14.46)3118 (14.58) < 0.001Smoking, n (%)30,819 (36.02)7251 (33.90)7469 (34.91)7776 (36.35)8323 (38.91) < 0.001Drinking, n (%)33,614 (39.28)8128 (38.00)8041 (37.58)8427 (39.40)9018 (42.16) < 0.001Physical exercise, n (%)77,927 (91.07)19,577 (91.52)19,525 (91.26)19,399 (90.69)19,426 (90.81)0.020Higher educational background, n (%)17,963 (20.99)5347 (25.00)4377 (20.46)4270 (19.96)3969 (18.55) < 0.001BMI: body mass index, SBP: systolic blood pressure, DBP: diastolic blood pressure, TyG index: triglyceride–glucose index, FBG: fasting blood glucose, TC: total cholesterol, TG: triglycerides, HDL-C: high-density lipoprotein cholesterol, LDL-C: low-density lipoprotein cholesterol, hs-CRP: high-sensitivity C-reactive protein. Baseline characteristics. BMI: body mass index, SBP: systolic blood pressure, DBP: diastolic blood pressure, TyG index: triglyceride–glucose index, FBG: fasting blood glucose, TC: total cholesterol, TG: triglycerides, HDL-C: high-density lipoprotein cholesterol, LDL-C: low-density lipoprotein cholesterol, hs-CRP: high-sensitivity C-reactive protein. Association of TyG index and risk of future RFD During a median follow-up period of 14.01 years (interquartile range: 13.63–14.21 years), a total of 3668 cases of RFD were recorded. The incidence rate of incident RFD increased substantially with the quartiles of TyG index, presenting 2.529, 3.110, 3.431, and 4.050 per 1,000-person year in quartile 1 to quartile 4. There was also an increasing trend in the cumulative incidence of incident RFD as the TyG index quartiles increased (Fig. 2).Fig. 2Cumulative incidence of future RFD according to the quartiles of TyG index. RFD: renal function decline. Cumulative incidence of future RFD according to the quartiles of TyG index. RFD: renal function decline. Multivariate Cox regression analysis reveals that the hazard ratio (HR) of RFD risk is 1.656 (95% confidence interval [CI]: 1.479–1.854) in quartile 4 compared with the quartile 1 (Table 2). In addition, the results of RCS showed that overall association is significant, but there is no evidence of significant non-linearity between the TyG index and the risk of future RFD (P for overall < 0.001, P for non-linearity = 0.168) (Fig. 3). The significant associations between the TyG index and the risk of RFD remained similar in the sensitivity analyses (Table 3). Furthermore, subgroup analysis showed that the association between the TyG index and incident RFD is stronger in participants aged less than 60 years (HR: 1.749, 95% CI: 1.521–2.010) than in participants aged above 60 years (HR: 1.358, 95% CI: 1.110–1.662), in women (HR: 1.461, 95% CI: 1.263–1.695) than in men (HR: 1.227, 95% CI: 1.026–1.467) (Table 4).Table 2Hazard ratios for risk of future RFD according to TyG index quartiles.βSEχ2PHR95% CIModel 1 Quartile 20.2060.05016.700 < 0.0011.2291.113–1.357 Quartile 30.3040.04937.781 < 0.0011.3551.230–1.493 Quartile 40.4690.04795.883 < 0.0011.5981.455–1.755Model 2 Quartile 20.2360.05121.824 < 0.0011.2671.147–1.399 Quartile 30.3260.05043.030 < 0.0011.3861.257–1.528 Quartile 40.5530.048131.467 < 0.0011.7401.583–1.912Model 3 Quartile 20.220.05118.522 < 0.0011.2461.127–1.378 Quartile 30.3010.05234.052 < 0.0011.3511.221–1.495 Quartile 40.5180.051103.228 < 0.0011.6791.520–1.856Model 4 Quartile 20.2270.05119.535 < 0.0011.2551.135–1.388 Quartile 30.3090.05334.240 < 0.0011.3621.228–1.511 Quartile 40.5040.05876.597 < 0.0011.6561.479–1.854RFD: renal function decline. Model 1 was a univariate Cox regression analysis model. Model 2 was adjusted for age and sex. Model 3 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein. Model 4 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity.Fig. 3Restricted cubic splines for the relative risk of future RFD.Multivariable-adjusted hazard ratios for future RFD based on restricted cubic spines with 5 knots at 5th, 25th, 50th, 75th, and 95th percentiles of the baseline TyG index levels. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity.Table 3Sensitivity analysis of the association between baseline TyG index and risk of future RFD.Total numberCase numberP ValueHR95% CIP for trendAnalysis 1 < 0.001 Quartile 121,194516-1.000- Quartile 221,156633 < 0.0011.2791.136–1.440 Quartile 321,139708 < 0.0011.4021.241–1.583 Quartile 421,083814 < 0.0011.6341.431–1.866Analysis 2 < 0.001 Quartile 120,862694-1.000- Quartile 220,761850 < 0.0011.2701.146–1.406 Quartile 320,572912 < 0.0011.3531.217–1.503 Quartile 420,3891070 < 0.0011.6731.491–1.877Analysis 3 < 0.001 Quartile 118,222589-1.000- Quartile 217,891710 < 0.0011.2771.142–1.427 Quartile 317,696761 < 0.0011.3761.226–1.544 Quartile 417,158853 < 0.0011.6971.494–1.928RFD: renal function decline. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Analysis 1: Excluding participants experienced RFD during the first 2 years of follow-up (n = 997). Analysis 2: Excluding participants experienced myocardial infarction and stroke during follow-up (n = 2985). Analysis 3: Excluding all-cause death during follow-up (n = 14,602).Table 4Subgroup analysis of the association between baseline TyG index and risk of future RFD.SubgroupTotal numberCase numberP ValueHR95% CIP for interactionAge < 0.001 ≥ 60 Quartile 14270203-1.000- Quartile 243272840.2091.1250.936–1.351 Quartile 345203800.0111.2671.056–1.520 Quartile 440093760.0031.3581.110–1.662 < 60 Quartile 117,122511-1.000- Quartile 217,068588 < 0.0011.3421.189–1.514 Quartile 316,870579 < 0.0011.3831.217–1.572 Quartile 417,383747 < 0.0011.7491.521–2.010Sex < 0.001 Men Quartile 115,745290-1.000- Quartile 217,1283320.6451.0380.885–1.218 Quartile 317,6533420.7201.0310.874–1.215 Quartile 418,2544340.0251.2271.026–1.467 Women Quartile 15647424-1.000- Quartile 242675400.0061.2021.054–1.370 Quartile 33737617 < 0.0011.2911.129–1.476 Quartile 43138689 < 0.0011.4611.263–1.695BMI0.303 ≥ 28 Quartile 1162653-1.000- Quartile 229501390.0501.3741.000–1.890 Quartile 345542110.0231.4311.051–1.948 Quartile 462453100.0041.6081.165–2.218 < 28 Quartile 119,766661-1.000- Quartile 218,445733 < 0.0011.2261.102–1.364 Quartile 316,836748 < 0.0011.3431.203–1.500 Quartile 415,147813 < 0.0011.6501.462–1.863Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Table 2Hazard ratios for risk of future RFD according to TyG index quartiles.βSEχ2PHR95% CIModel 1 Quartile 20.2060.05016.700 < 0.0011.2291.113–1.357 Quartile 30.3040.04937.781 < 0.0011.3551.230–1.493 Quartile 40.4690.04795.883 < 0.0011.5981.455–1.755Model 2 Quartile 20.2360.05121.824 < 0.0011.2671.147–1.399 Quartile 30.3260.05043.030 < 0.0011.3861.257–1.528 Quartile 40.5530.048131.467 < 0.0011.7401.583–1.912Model 3 Quartile 20.220.05118.522 < 0.0011.2461.127–1.378 Quartile 30.3010.05234.052 < 0.0011.3511.221–1.495 Quartile 40.5180.051103.228 < 0.0011.6791.520–1.856Model 4 Quartile 20.2270.05119.535 < 0.0011.2551.135–1.388 Quartile 30.3090.05334.240 < 0.0011.3621.228–1.511 Quartile 40.5040.05876.597 < 0.0011.6561.479–1.854RFD: renal function decline. Model 1 was a univariate Cox regression analysis model. Model 2 was adjusted for age and sex. Model 3 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein. Model 4 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Hazard ratios for risk of future RFD according to TyG index quartiles. RFD: renal function decline. Model 1 was a univariate Cox regression analysis model. Model 2 was adjusted for age and sex. Model 3 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein. Model 4 was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Restricted cubic splines for the relative risk of future RFD.Multivariable-adjusted hazard ratios for future RFD based on restricted cubic spines with 5 knots at 5th, 25th, 50th, 75th, and 95th percentiles of the baseline TyG index levels. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Table 3Sensitivity analysis of the association between baseline TyG index and risk of future RFD.Total numberCase numberP ValueHR95% CIP for trendAnalysis 1 < 0.001 Quartile 121,194516-1.000- Quartile 221,156633 < 0.0011.2791.136–1.440 Quartile 321,139708 < 0.0011.4021.241–1.583 Quartile 421,083814 < 0.0011.6341.431–1.866Analysis 2 < 0.001 Quartile 120,862694-1.000- Quartile 220,761850 < 0.0011.2701.146–1.406 Quartile 320,572912 < 0.0011.3531.217–1.503 Quartile 420,3891070 < 0.0011.6731.491–1.877Analysis 3 < 0.001 Quartile 118,222589-1.000- Quartile 217,891710 < 0.0011.2771.142–1.427 Quartile 317,696761 < 0.0011.3761.226–1.544 Quartile 417,158853 < 0.0011.6971.494–1.928RFD: renal function decline. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Analysis 1: Excluding participants experienced RFD during the first 2 years of follow-up (n = 997). Analysis 2: Excluding participants experienced myocardial infarction and stroke during follow-up (n = 2985). Analysis 3: Excluding all-cause death during follow-up (n = 14,602). Sensitivity analysis of the association between baseline TyG index and risk of future RFD. RFD: renal function decline. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Analysis 1: Excluding participants experienced RFD during the first 2 years of follow-up (n = 997). Analysis 2: Excluding participants experienced myocardial infarction and stroke during follow-up (n = 2985). Analysis 3: Excluding all-cause death during follow-up (n = 14,602). Table 4Subgroup analysis of the association between baseline TyG index and risk of future RFD.SubgroupTotal numberCase numberP ValueHR95% CIP for interactionAge < 0.001 ≥ 60 Quartile 14270203-1.000- Quartile 243272840.2091.1250.936–1.351 Quartile 345203800.0111.2671.056–1.520 Quartile 440093760.0031.3581.110–1.662 < 60 Quartile 117,122511-1.000- Quartile 217,068588 < 0.0011.3421.189–1.514 Quartile 316,870579 < 0.0011.3831.217–1.572 Quartile 417,383747 < 0.0011.7491.521–2.010Sex < 0.001 Men Quartile 115,745290-1.000- Quartile 217,1283320.6451.0380.885–1.218 Quartile 317,6533420.7201.0310.874–1.215 Quartile 418,2544340.0251.2271.026–1.467 Women Quartile 15647424-1.000- Quartile 242675400.0061.2021.054–1.370 Quartile 33737617 < 0.0011.2911.129–1.476 Quartile 43138689 < 0.0011.4611.263–1.695BMI0.303 ≥ 28 Quartile 1162653-1.000- Quartile 229501390.0501.3741.000–1.890 Quartile 345542110.0231.4311.051–1.948 Quartile 462453100.0041.6081.165–2.218 < 28 Quartile 119,766661-1.000- Quartile 218,445733 < 0.0011.2261.102–1.364 Quartile 316,836748 < 0.0011.3431.203–1.500 Quartile 415,147813 < 0.0011.6501.462–1.863Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Subgroup analysis of the association between baseline TyG index and risk of future RFD. Model was adjusted for age, sex, body mass index, systolic blood pressure, total cholesterol, low-density lipoprotein cholesterol, high-sensitivity C-reactive protein, estimated glomerular filtration rate, history of hypertension, history of diabetes, income level, smoking status, drinking status, physical activity. Discussion In this study, we confirmed a notable correlation between the TyG index and the risk of RFD. The results indicate that individuals exhibiting a high TyG index are at an increased risk of experiencing future RFD, in contrast to those with a low TyG index. Additionally, the relationship between the TyG index and renal events suggests that IR could play a pivotal role in the onset of RFD. In line with our result, Yoshida et al. conducted an analysis involving data from 10,758 Japanese individuals. Future RFD was characterized as a decrease of ≥ 25% in eGFR from the baseline measurement. Throughout the follow-up duration (37.8 ± 23.6 months), the incidence rates of RFD were recorded at 0.31 and 0.69 per 100 person-years for the low and high TyG index groups, respectively. Furthermore, a high TyG index was found to be significantly correlated with future RFD, with a HR of 2.25 and a 95% CI of 1.40–3.6013. Chen et al. examined data from 548 Taiwanese participants with hypertension. RFD was defined as a decline of > 25% in eGFR. Over an average follow-up period of 4.7 ± 3.1 years, 97 patients experienced a decline of > 25% in eGFR. The findings indicated that the TyG index (HR = 1.490; 95% CI = 1.016–2.185, P = 0.041), office SBP (HR = 1.013; 95% CI = 1.000–1.026, P = 0.047), diabetes mellitus (HR = 1.797, 95% CI = 1.026–3.147, P = 0.040), and baseline eGFR (HR = 1.015; 95% CI = 1.002–1.028, P = 0.025) were all significantly associated with RFD. In addition, patients with eGFR decline (> 25%) exhibited remarked elevated levels of the TyG index compared with patients who did not experience eGFR decline (P = 0.014)12. Lei et al. explored the association between the TyG index and eGFR in a cohort study including 7,822 elderly adults from South China. Their results showed that the TyG index is significantly associated with 30% decline of eGFR (HR per SD increase:1.38; 95% CI: 1.26–1.50), or 40% decline of eGFR (HR per SD increase: 1.42; 95% CI: 1.24–1.63)16. Liu et al. conducted an analysis involving data from 3,899 participants in the China Health and Retirement Longitudinal Study. The main outcome measured was rapid kidney function decline (RKFD), which is characterized as an annualized decrease in eGFRcr-cys of 5 ml/min per 1.73 m2 or greater. Over a follow-up period of 3.99 years, 191 instances of RKFD were documented. The TyG index and its related two indices (triglyceride-glucose × BMI (TyGBMI), and triglyceride-glucose × waist circumference (TyGWC)) showed significant associations with the occurrence of RKFD, with TyGWC demonstrating the highest predictive capability for RKFD17. CKD is the primary cause of sustained, pathological eGFR decline. A sustained decrease (especially below 60 mL/min/1.73 m2 for more than 3 months) is the defining feature of CKD18. Numerous studies have shown that a high TyG index is associated with the risk of CKD. Regarding this, Ren et al. investigated the relationship between the TyG index and CKD in a cohort study including 10,498 subjects, their results showed that the HR (95% CI) for the highest quartile compared to the lowest quartile of TyG index was 1.30 (1.08–1.57). Furthermore, each SD increase in the TyG index was linked to an 11% higher risk of new-onset CKD (HR: 1.11, 95% CI: 1.03–1.19)19. Okamura et al. studied the association between the TyG index and incident CKD in a cohort study including 11,712 participants in Japan, they demonstrated that a higher TyG index was significantly associated with risk for incident CKD in both men (HR: 1.32, 95% CI 1.02–1.70, P = 0.036) and women (HR: 1.50, 95% CI: 1.05–2.13, P = 0.024)20. Liu et al. conducted an analysis involving data from 4361 participants who were part of the National Health and Nutrition Examination Survey conducted between 2015 and 2018. The average TyG index recorded was 8.60 ± 0.68, while the prevalence of CKD was found to be 13.35%. It was observed that participants exhibiting a higher TyG index were linked to an increased risk of CKD (OR = 1.34, 95% CI: 1.13, 1.59, P = 0.0006)21. Furthermore, our results also identified that a stronger association between the TyG index and the risk of RFD in individuals aged less than 60 years compared to those above, in women compared to men. Considering age as a significant risk factor for RFD22, the influence of TyG index appeared to diminish in the ≥ 60 years cohort after stratification and adjustment for age-related covariates. This reduction may be due to the prevalence of comorbidities, including cardiovascular diseases. An increasing amount of evidence suggests that IR plays a role in RFD23. IR is highly prevalent in CKD due to chronic inflammation, where increased proinflammatory cytokines impair insulin signalling24. In cases of IR, a low-grade inflammatory response is triggered, leading to heightened macrophage infiltration in both adipose tissue and the kidney. This macrophage infiltration results in the generation of various pro-inflammatory cytokines and subsequent elements. Elevated pro-inflammatory cytokines, particularly interleukin-6, and activation of the NFκB pathway drive renal injury, metabolic dysfunction, and cardiovascular complications25. Moreover, recent findings have provided substantial evidence to support the significant involvement of insulin in the function of the glomerular filtration barrier26. Additionally, insulin could play a role in the progression of microalbuminuria and CKD by impacting the production of transforming growth factor β in mesangial and proximal tubule cells27. Studies have indicated that transforming growth factor β production leads to heightened fibrosis in the interstitium and podocyte remodeling, ultimately leading to increased extracellular matrix production and damage to the filtration barrier, which are factors potentially involved in the development of microalbuminuria28. Therefore, it is crucial to aggressively manage metabolic abnormalities including IR/hyperinsulinemia to prevent or delay the progression of renal function loss and the development of CKD7. Our study has some limitations. First, the study endpoint was a future RFD, which was defined as a 25% decrease in eGFR from baseline to below 60 ml/min/1.73 m2. CKD is the primary cause of sustained, pathological eGFR decline18. However, with a biennial examination schedule, it is impossible to distinguish between CKD and acute kidney injury (AKI). CKD is defined as the existence of kidney impairment or an eGFR below 60 ml/min/1.73 m2, enduring for a period exceeding 3 months. In CKD, the body undergoes gradual adaptations to sustain GFR and achieve a balance of creatinine and various electrolytes. Conversely, AKI represents a sudden deterioration in kidney function, typically occurring within a matter of hours or days, characterized by increased serum creatinine levels and diminished urine output. During AKI, these adaptive mechanisms frequently fail to take place, leading to a swift advancement towards renal failure5. In addition, the existence of albuminuria or proteinuria indicates kidney damage and, when combined with the assessment of GFR, forms the basis for evaluating CKD. Unfortunately, proteinuria/albuminuria data were not incorporated in the current study. While the supplementary sensitivity analysis, involved excluding all-cause death during the follow-up period, was performed, it does not entirely alleviate the concern that death could serve as a competing event in the analysis of long-term renal outcomes. The primary time-to-event analyses consider death as a form of censoring, which may consequently lead to an overestimation of the cumulative incidence of RFD. Lastly, given that our study population is 80.38% male, which substantially differs from typical community-based populations, the results could not be generalized to the general population of North China. Conclusions The findings of this study indicate that an elevated TyG index correlates with an increased risk of future RFD. Assessing the TyG index could prove beneficial in recognizing individuals who are at significant risk for future RFD.