Skip to main content
Download PDF
Download ePub

Non-linear relationship between triglyceride glucose index and new-onset diabetes among individuals with non-alcoholic fatty liver disease: a cohort study

Lipids in Health and Disease volume 24, Article number: 94 (2025) Cite this article

Abstract

Background

The relationship between the triglyceride glucose (TyG) values and the development of diabetes in non-alcoholic fatty liver disease (NAFLD) patients is not yet well researched. This study aims to examine how the baseline TyG levels correlate with the incidence of new-onset diabetes in this specific cohort.

Methods

This cohort included 2,506 normoglycemic Japanese adults with NAFLD who underwent routine health check-ups at Murakami Memorial Hospital between 2004 and 2015. Several statistical approaches, including restricted cubic splines and two-piecewise linear regression, were utilized to assess the relation between the TyG levels and diabetes risk.

Results

Among the 2,506 participants (mean age: 44.78 ± 8.32 years; 81.09% male), 203 individuals (8.10%) developed diabetes over the course of the 11-year follow-up period. A U-shaped relationship was observed between the levels of TyG and the onset of diabetes, with an inflection point identified at a TyG value of 7.82 (95% CI: 7.72-8.00). Below this threshold, each one-unit elevation in TyG values reduced the probability of diabetes by 93% (HR = 0.07, 95% CI: 0.01–0.32, P = 0.001). Conversely, above this threshold, each one-unit elevation increased the probability of diabetes by 70% (HR = 1.70, 95% CI: 1.19–2.44, P = 0.004).

Conclusions

The findings validate a U-shaped association between TyG levels and new-onset diabetes in adults with NAFLD. Both low and high TyG levels increase diabetes probability in such a group.

Introduction

Diabetes, a condition of hyperglycemia, most commonly results from insulin resistance (IR) or an insufficiency of insulin secretion [1]. Global estimates predict that 629 million individuals will be affected by diabetes in 2045 [2]. Diabetic complications and their influence have a profound negative impact on individual health, impose a huge financial burden, and propel rising healthcare costs globally [3]. A vast amount of information is available on the etiology and risk factors of diabetes.

Non-alcoholic fatty liver disease (NAFLD) represents the most common worldwide chronic liver disease, with an estimated 25% of adults affected by it [4]. NAFLD is distinguished by excessive hepatic fat accumulation, which can range from simple steatosis to more virulent forms such as nonalcoholic steatohepatitis [5, 6]. The progression of NAFLD can lead to serious liver complications, including cirrhosis and even hepatocellular carcinoma [5, 6]. In addition to its hepatic symptoms, NAFLD is a strong relative factor for IR and a variety of types of metabolic disorder [7,8,9,10].

The triglyceride glucose (TyG) index, calculated using triglyceride (TG) levels and fasting plasma glucose (FPG) levels, is universally considered to be associated with IR [11]. The index is particularly valuable as it integrates both glucose and lipid metabolism, thereby providing a more holistic view of metabolic dysfunction [12, 13]. Previous studies have consistently demonstrated a strong positive relationship between levels of TyG and the development of diabetes in a range of populations [14,15,16,17,18]. In NAFLD, however, a direct relationship between the levels of TyG and the probability of diabetes is not well characterized, and whether a non-linear relationship between the TyG levels and NAFLD subjects’ probability of diabetes exists is not yet established.

Therefore, the current study aimed to elucidate the relationship between baseline TyG index levels and new-onset diabetes probability in normoglycemic Japanese adults with NAFLD, with a specific intention to detect any threshold effects using nonlinear statistical analysis.

Methods

Study design and data source

This cohort analysis made use of the NAGALA database, which is accessible through the DRYAD repository and contains information collected at Murakami Memorial Hospital (2004–2015) [19]. The initial group consisted of 20,944 willing participants who underwent comprehensive medical exams. The original study was authorized by the Murakami Memorial Hospital Ethics Committee, and all subjects were required to provide written informed consent before recruitment.

Study participants

A well-characterized study population was established by excluding individuals based on specific criteria from an original cohort of 20,944 Japanese participants. Inclusion in the study was not possible for the following reasons: a positive hepatitis B or C virus serology, a diagnosis of alcoholic fatty liver disease, diabetes, a FPG level of 6.1 mmol/L or above, prior drug use, and lack of covariate data. Additional criteria used to exclude individuals with NAFLD included excessive alcohol consumption [20, 21], the absence of ultrasonographic evidence of fatty liver, and an extreme TyG index. Certified sonographers used standardized ultrasonographic examinations to diagnose NAFLD [19]. The final analysis included 2,506 participants after all exclusion criteria were applied (Fig. 1).

Fig. 1
figure 1

Study population

Full size image

Definition of NAFLD

Fatty liver diagnosis was performed through abdominal ultrasonography by certified technicians using a calibrated ultrasound system [19]. Gastroenterologists systematically evaluated ultrasound images based on predefined diagnostic criteria, focusing on liver brightness, deep attenuation, vascular blurring, and hepatorenal echo contrast [19, 22]. Assessments were made without access to participants’ additional clinical data to reduce interpretation bias [19].

Consistent with recent international diagnostic criteria, NAFLD was defined as an individual’s alcohol consumption falling below the cutoff levels and the presence of fatty liver on ultrasonography [20, 21].

Covariates

Demographic information (gender and age) and lifestyle characteristics (exercise frequency, alcohol consumption, smoking status) were collected using standardized questionnaires [19]. Anthropometric measurements (WC, diastolic blood pressure (DBP), systolic blood pressure (SBP), body mass index (BMI)), and laboratory assessments were performed by certified healthcare professionals following standardized protocols in consistent laboratory conditions. All biochemical parameters, including aspartate aminotransferase (AST), TG, high density lipoprotein cholesterol (HDL-c), FPG, glycosylated hemoglobin A1c (HbA1c), gamma-glutamyl transferase (GGT), total cholesterol (TC), and alanine aminotransferase (ALT), were analyzed using validated analytical methods in an accredited clinical laboratory [19]. Engaging in physical activity at least once a week was considered regular exercise [19]. The definition of visceral fat obesity was WC ≥ 80 cm for women and ≥ 90 cm for men [19].

Exposure and outcome

The exposure variable was the levels of TyG, calculated:\(\:\:\:Ln\:\left[TG\:\right(mg/dL)\times\:FPG(mg/dL)/2]\) [23, 24]. The outcome was diabetes. Diabetes diagnosis was confirmed when meeting any of the following criteria during follow-up: self-reported physician diagnosis, HbA1c ≥ 6.5%, or FPG ≥ 7.0 mmol/L [1].

Statistical analyses

Differences across TyG tertile levels were evaluated using one-way ANOVA for continuous variables and chi-square tests for categorical variables. Missing data for HDL-c were imputed using mean substitution [25].

Kaplan-Meier survival curves, accompanied by log-rank tests, were employed to compare diabetes-free survival rates across TyG index tertiles. Hazard ratios (HRs) with 95% confidence intervals (CIs) were generated using Cox proportional hazards models. Models were adjusted for potential confounders identified based on previous literature [14,15,16,17,18, 26, 27] and statistical criteria (change-in-estimate > 10% [28, 29] or P < 0.05 in univariate analysis): HDL-c, WC, AST, age, GGT, SBP [30, 31], ALT, sex [32, 33], TC, DBP [30, 31], regular exerciser, BMI, HbA1c [34], and smoking status [35,36,37].

Dose-response relationships were evaluated using restricted cubic splines (RCS). The optimal inflection point was identified through a recursive algorithm based on maximum likelihood estimation [38]. Model fitness was compared between linear and two-piece-wise models using log-likelihood ratio tests [38]. Bootstrap resampling with 1000 replications generated 95% CI for the inflection point [38].

To test the robustness of the results, sensitivity analyses were conducted, including imputation of missing data and simplified covariate adjustments. E-values were computed to assess the potential influence of unmeasured confounders [39]. Subgroup analyses were also performed to investigate possible effect modifications.

All statistical analyses were carried out using EmpowerStats (version 4.2) and R software (version 4.2.0), with statistical significance set at P < 0.05.

Results

Baseline characteristics

Table 1 Baseline characteristics of participants
Full size table

As shown in Table 1, of the 2,506 participants, men comprised 2,032 (81.09%), while women accounted for 474 (18.91%). The participants’ average age was 44.78 ± 8.32. A significant increase was observed in DBP, FPG, SBP, HbA1c, TG, TC, GGT, AST, ALT, WC, BMI, the proportion of current smokers, and the male ratio as the levels of TyG increased. HDL-c showed a significant decrease with an increasing TyG index (Table 1). New-onset diabetes was identified in 203 participants (8.10%), with its prevalence increasing from T1 to T3 (4.91%, 6.83%, and 12.56%, respectively).

Fig. 2
figure 2

Kaplan–Meier diabetes-free survival curve across TyG tertile (log-rank, P < 0.001)

Full size image

Differences in diabetes-free probability across the TyG tertile levels were evident in Kaplan-Meier curves (P < 0.001). Higher TyG values were linked to lower diabetes-free probability over the 4,000-day follow-up period (Fig. 2).

Univariate analyses

Table 2 Univariate Cox proportional hazards regression
Full size table

Table 2 demonstrated a significant relationship between the outcome and variables such as FPG, WC, HbA1c, TG, TC, GGT, AST, ALT, BMI, age, current smoking status, TyG, and HDL-c. Specifically, increases in these variables increased the probability of diabetes, with HDL-c being the exception.

Multivariate analyses

Table 3 Influence of TyG on new-onset diabetes under various models
Full size table

As provided in Table 3, in Model II, each one-unit increase in TyG levels corresponded to a 43% higher probability of diabetes (HR = 1.43, 95% CI: 1.01–2.02, P = 0.044). In tertile analysis (T1: 6.96–8.35, T2: 8.35–8.82, T3: 8.82–10.18), participants in the T3 group showed significantly increased diabetes risk (HR = 1.46, 95% CI: 1.03–2.08, P = 0.034) compared to the T2 group, whereas no significant difference was observed in the T1 group (HR = 0.94, 95% CI: 0.62–1.44, P = 0.791).

Nonlinear analyses

Fig. 3
figure 3

Nonlinear association between TyG index and new-onset diabetes. RCS analysis revealed a threshold, nonlinear relationship. The red solid line represents the adjusted log relative risk, while the blue dashed curves indicate the 95% CI. Models adjusted for HDL-c, WC, AST, age, GGT, SBP, ALT, sex, TC, DBP, regular exerciser, BMI, HbA1c, and smoking status

Full size image
Table 4 Threshold effect analysis of TyG and diabetes using two-piece-wise regression (N = 2,506)
Full size table

A nonlinear relationship between baseline TyG levels and diabetes incidence was observed (Fig. 3; Table 4). The inflection point occurred at a TyG value of 7.82 (95% CI: 7.72-8.00). For TyG values at or below 7.82, each one-unit elevation in TyG values decreased the probability of diabetes by 93% (HR = 0.07, 95% CI: 0.01–0.32, P = 0.001). In contrast, for TyG values above 7.82, each one-unit elevation increased the probability of diabetes by 70% (HR = 1.70, 95% CI: 1.19–2.44, P = 0.004) (Table 4).

Subgroup analyses

Table 5 Stratified analyses of the association between TyG levels and diabetes risk across different subgroups
Full size table

Stratified analyses showed the TyG-diabetes relationship remained consistent across various subgroups, with no significant interactions (all P for interaction > 0.05) (Table 5).

Sensitivity analysis

Table 6 Threshold effect analysis of TyG and diabetes with missing data (N = 2,502)
Full size table

The results remained consistent between analyses with missing data (Table 6) and after imputation (Table 4), with identical threshold effects at the TyG index of 7.82. Similarly, simplified adjusted models (Table S1), including only variables significant in univariate analysis, showed comparable threshold points and risk estimates to Table 4, supporting the reliability of the identified associations and threshold effect.

To examine the potential impact of unmeasured confounding variables, an E-value was calculated. The results remained robust, unless an unobserved confounder had an HR exceeding 2.79.

Discussion

This cohort study, which involved 2,506 normoglycemic NAFLD patients, sought to investigate the TyG-diabetes relationship over an 11-year follow-up period. A U-shaped relationship was identified, with a threshold effect at a TyG index of 7.82 (95% CI: 7.72-8.00). Below this threshold, the risk of diabetes decreased (HR = 0.07, 95% CI: 0.01–0.32), but above it, the risk increased substantially (HR = 1.70, 95% CI: 1.19–2.44).

This study bears methodological similarities to prior cohort studies, yet reveals unique findings. All studies utilized cohort designs to explore the TyG index-diabetes relationships in diverse populations. Zhang et al. [14] studied 5,706 normal-weight participants from rural China, finding a significantly increased diabetes risk in the highest quartile levels compared to the lowest of the TyG (HR = 3.91, 95% CI: 2.22–6.87). Similarly, Li et al. [15] observed that TyG level elevation increased diabetes risk (HR = 3.34, 95% CI: 3.11–3.60) with a significant nonlinear relationship. In Japanese populations, Xuan et al. [16] detected a threshold effect at 7.97, with a decreased risk below (HR = 0.21, 95% CI: 0.08–0.57) and increased risk above (HR = 2.42, 95% CI: 1.66–3.53). The current study, focusing on 2,506 NAFLD patients, revealed a comparable threshold effect at 7.82, with decreased risk below (HR = 0.07, 95% CI: 0.01–0.32) and increased risk above (HR = 1.70, 95% CI: 1.19–2.44). These variations in risk patterns and magnitudes highlight the distinct metabolic characteristics of NAFLD patients, especially concerning IR and glucose homeostasis. These findings highlight the complex, nonlinear relationship between the baseline levels of TyG and the probability of diabetes in NAFLD individuals.

In the stratified analyses, the TyG-diabetes relationship remained consistent across different subgroups, including sex, smoking status, exercise habits, WC, BMI, and age, with no significant interaction effects (all P for interaction > 0.05). Notably, never smokers showed a significantly stronger association between the baseline levels of TyG and the probability of diabetes (HR = 1.80, 95%CI: 1.00-3.23) compared with past or current smokers (HR = 1.26, 95%CI: 0.81–1.94), consistent with previous findings that smoking status influences IR and glucose metabolism [35,36,37]. Furthermore, the TyG-diabetes relationship was more pronounced in women (HR = 1.74, 95%CI: 0.70–4.35) versus men (HR = 1.38, 95%CI: 0.94–2.02), supporting evidence of sex-specific differences in the TyG-diabetes relationship [32, 33].

Multivariate analyses indicated a positive linear association between the baseline levels of TyG, treated as a continuous variable, and the probability of diabetes. However, tertile analysis revealed a more complex pattern: the highest tertile showed an increased risk, while the lowest tertile showed a reduced risk when compared to the middle tertile. This divergent pattern suggested the presence of a nonlinear relationship, prompting further analyses that confirmed a U-shaped association with a threshold at a baseline TyG value of 7.82. The robustness of these findings was validated through extensive sensitivity analyses. Consistent results were obtained when comparing analyses with missing data (N = 2,502) and after imputation (N = 2,506), revealing identical threshold effects at a TyG index of 7.82. Both simplified and fully adjusted models demonstrated comparable inflection points and risk estimates, substantiating the reliability of the identified U-shaped association and threshold effect.

In adults with NAFLD, TyG-diabetes relationship demonstrates a distinctive U-shaped pattern, where both elevated and reduced TyG levels contribute to diabetes pathogenesis through different mechanisms. In NAFLD patients with elevated TyG, excessive hepatic lipid accumulation triggers lipotoxicity through overwhelming free fatty acid levels [40, 41]. This lipotoxicity promotes diabetes development through multiple pathways: IR induced by saturated fatty acid-induced inflammation and GP130-STAT3 signaling [42, 43]; β-cell dysfunction through reactive oxygen species generation [44,45,46,47]; and disrupted glucose homeostasis via altered hepatokine secretion and PPARγ-mediated impairment of insulin signaling [48,49,50]. Hepatic steatosis in NAFLD exacerbates systemic IR, consequently increasing TyG levels [42, 43]. NAFLD progression is further promoted with elevated TyG via lipotoxicity and oxidative stress, forming a vicious cycle [40, 41]. These processes are all interconnected and constitute a vicious cycle promoting diabetes progression.

In contrast, profoundly low TyG levels signify a critical metabolic anomaly precipitating diabetes development [51]. The underlying pathophysiology includes impaired glucose utilization and metabolic rigidity, disrupting cellular energy balance via defective fuel substrate transition [52]. Such disruptions initiate compensatory mechanisms: diminished glucose absorption enhances IR [12], and metabolic rigidity compromises cellular energy efficiency [52]. Moreover, malnutrition-related pancreatic impairments and modified glucose-TG interactions [53, 54] establish a continuous cycle of metabolic decline culminating in diabetes [55].

Current findings demonstrate a clear U-shaped correlation between the baseline levels of TyG and probability of diabetes in NAFLD populations. The analysis pinpointed a critical value around 7.82, providing potential clinical implications for varying groups based on their TyG levels. For individuals with TyG values above 7.82, evidence suggests that each incremental increase correlates with a heightened risk of diabetes (HR = 1.70), underscoring the importance of vigilant metabolic surveillance, lifestyle adjustments, or therapeutic interventions. Conversely, for those with TyG values below this threshold, the risk appears reduced; however, very low TyG levels might signal essential malnutrition and metabolic issues that warrant careful consideration. Thus, it is advisable to avoid excessive dietary limitations that may lead to further reductions in TyG and subsequent metabolic issues. A moderate approach to maintaining metabolic equilibrium is advisable for this demographic. Expanding this research to multicenter studies could further authenticate the inflection point’s applicability across varied demographics.

Study strengths and limitations

This investigation is distinguished by several strengths. Primarily, it serves as an inaugural study linking the TyG index with the incidence of diabetes in NAFLD cohorts, unveiling a novel U-shaped association (inflection point: 7.82, 95% CI: 7.72-8.00). Secondly, it employs standardized diagnostic protocols for NAFLD through ultrasound assessments conducted by accredited technicians and evaluated by unbiased gastroenterologists. Thirdly, it includes exhaustive statistical evaluations adjusted for multiple confounders, including metabolic, anthropometric, and lifestyle variables. Fourthly, advanced statistical methods utilizing RCS and two-piece-wise regression models facilitate a detailed exploration of non-linear associations. Additionally, the integrity of the results is reinforced by multiple sensitivity tests and E-value computations.

It is important to acknowledge certain limitations. Primarily, the diagnosis of NAFLD depends solely on ultrasound outcomes without comprehensive hepatic steatosis grading, which could impact the accuracy of results since the severity of steatosis may affect the TyG index and diabetes risk [20, 21, 56]. The absence of liver biopsy data restricts the quantitative analysis of hepatic lipid content. Secondly, the NAGALA database (2004–2015) does not include Fibroscan data, which could have offered a more precise evaluation of liver fibrosis [20, 21, 57]. Prospective studies incorporating Fibroscan and detailed assessments of liver steatosis are recommended to further investigate the interactions between liver fibrosis, steatosis, baseline TyG levels, and diabetes probability. Thirdly, the absence of random glucose measurements and oral glucose tolerance tests in the criteria for diabetes diagnosis likely led to an underestimation of diabetes prevalence in this cohort. Furthermore, given that the study exclusively involved Japanese adults, additional research is necessary to confirm the applicability of these findings to other ethnic groups. Additionally, whereas the study was based on NAFLD criteria, a recent consensus (2023) has proposed MASLD (metabolic dysfunction-associated steatotic liver disease) to emphasize metabolic dysfunction [58]. Future research might consider incorporating MASLD diagnostic parameters.

Conclusion

In a cohort of 2,506 Japanese adults diagnosed with NAFLD, using data from the NAGALA database, this study delineated a U-shaped correlation between the baseline levels of TyG and the probability of diabetes onset, identifying a critical inflection point at a TyG index of 7.82 (95% CI: 7.72-8.00). Clinically, stabilizing the TyG index near this value could prove advantageous, given that deviations either above or below this point are linked to an elevated risk of diabetes. Distinct approaches are advisable depending on the TyG values: those with elevated indices may benefit from lifestyle modifications, pharmacological treatment, and rigorous metabolic surveillance, whereas for individuals with lower indices, it is crucial to monitor for signs of malnutrition and metabolic imbalances without further reducing the TyG levels. Additional multicenter research might help further enhance the robustness of these findings and assess the applicability of this inflection point across varied demographic groups.

Data availability

Data can be accessed at: https://doiorg.publicaciones.saludcastillayleon.es/10.5061/dryad.8q0p192.

Abbreviations

TyG:

Triglyceride glucose index

WC:

Waist circumference

BMI:

Body mass index

DBP:

Diastolic blood pressure

SBP:

Systolic blood pressure

ALT:

Alanine aminotransferase

AST:

Aspartate aminotransferase

GGT:

Gamma-glutamyl transferase

HDL-c:

High-density lipoprotein cholesterol

TC:

Total cholesterol

TG:

Triglycerides

FPG:

Fasting plasma glucose

HbA1c:

Hemoglobin A1c

HR:

Hazard ratios

CI:

Confidence intervals

IR:

Insulin resistance

NAFLD:

Non-alcoholic fatty liver disease

References

  1. Basevi V, Di Mario S, Morciano C, Nonino F, Magrini N. Comment on: American Diabetes Association. Standards of Medical Care in Diabetes—2011. Diabetes Care. 2011;34(Suppl. 1):S11–S61. Diabetes Care. 2011;34:e53.

  2. Cho NH, Shaw JE, Karuranga S, Huang Y, Fernandes JD da, Ohlrogge R et al. AW,. IDF Diabetes Atlas: Global estimates of diabetes prevalence for 2017 and projections for 2045. Diabetes Research and Clinical Practice. 2018;138:271–81.

  3. American Diabetes Association. Economic costs of diabetes In the U.S. In 2017. Diabetes Care. 2018;41:917–28.

    Article  PubMed Central  Google Scholar 

  4. Younossi ZM, Koenig AB, Abdelatif D, Fazel Y, Henry L, Wymer M. Global epidemiology of nonalcoholic fatty liver disease—Meta-analytic assessment of prevalence, incidence, and outcomes. Hepatology. 2016;64:73.

    Article  PubMed  Google Scholar 

  5. Lazarus JV, Mark HE, Anstee QM, Arab JP, Batterham RL, Castera L, et al. Advancing the global public health agenda for NAFLD: a consensus statement. Nat Rev Gastroenterol Hepatol. 2022;19:60–78.

    Article  CAS  PubMed  Google Scholar 

  6. Gordon C, Fraysse S, Li J, Ozbay S, Wong AB. Disease severity is associated with higher healthcare utilization in nonalcoholic steatohepatitis medicare patients. Official J Am Coll Gastroenterol| ACG. 2020;115:562.

    Article  Google Scholar 

  7. Angulo P. Nonalcoholic fatty liver disease. N Engl J Med. 2002;346:1221–31.

    Article  CAS  PubMed  Google Scholar 

  8. Sung K-C, Wild SH, Byrne CD. Resolution of fatty liver and risk of incident diabetes. J Clin Endocrinol Metabolism. 2013;98:3637–43.

    Article  CAS  Google Scholar 

  9. Morrison AE, Zaccardi F, Khunti K, Davies MJ. Causality between non-alcoholic fatty liver disease and risk of cardiovascular disease and type 2 diabetes: A meta-analysis with bias analysis. Liver Int. 2019;39:557–67.

    Article  CAS  PubMed  Google Scholar 

  10. Tarantino G, Crocetto F, Di Vito C, Creta M, Martino R, Pandolfo SD, et al. Association of NAFLD and insulin resistance with Non metastatic bladder cancer patients: A Cross-Sectional retrospective study. J Clin Med. 2021;10:346.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Vasques ACJ, Novaes FS, de Oliveira Mda, Souza S, Yamanaka JRM, Pareja A. TyG index performs better than HOMA in a Brazilian population: a hyperglycemic clamp validated study. Diabetes Res Clin Pract. 2011;93:e98–100.

    Article  CAS  PubMed  Google Scholar 

  12. Khan SH, Sobia F, Niazi NK, Manzoor SM, Fazal N, Ahmad F. Metabolic clustering of risk factors: evaluation of Triglyceride-glucose index (TyG index) for evaluation of insulin resistance. Diabetol Metab Syndr. 2018;10:74.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Wang H, Yan F, Cui Y, Chen F, Wang G, Cui W. Association between triglyceride glucose index and risk of cancer: A meta-analysis. Front Endocrinol (Lausanne). 2022;13:1098492.

    Article  PubMed  Google Scholar 

  14. Zhang M, Wang B, Liu Y, Sun X, Luo X, Wang C, et al. Cumulative increased risk of incident type 2 diabetes mellitus with increasing triglyceride glucose index in normal-weight people: the rural Chinese cohort study. Cardiovasc Diabetol. 2017;16:30.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Li X, Li G, Cheng T, Liu J, Song G, Ma H. Association between triglyceride-glucose index and risk of incident diabetes: a secondary analysis based on a Chinese cohort study. Lipids Health Dis. 2020;19:236.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Xuan X, Hamaguchi M, Cao Q, Okamura T, Hashimoto Y, Obora A, et al. U-shaped association between the triglyceride-glucose index and the risk of incident diabetes in people with normal glycemic level: A population-base longitudinal cohort study. Clin Nutr. 2021;40:1555–61.

    Article  CAS  PubMed  Google Scholar 

  17. Chen C-L, Liu L, Lo K, Huang J-Y, Yu Y-L, Huang Y-Q, et al. Association between triglyceride glucose index and risk of New-Onset diabetes among Chinese adults: findings from the China health and retirement longitudinal study. Front Cardiovasc Med. 2020;7:610322.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Park B, Lee HS, Lee Y-J. Triglyceride glucose (TyG) index as a predictor of incident type 2 diabetes among Nonobese adults: a 12-year longitudinal study of the Korean genome and epidemiology study cohort. Translational Res. 2021;228:42–51.

    Article  CAS  Google Scholar 

  19. Okamura T, Hashimoto Y, Hamaguchi M, Obora A, Kojima T, Fukui M. Ectopic fat obesity presents the greatest risk for incident type 2 diabetes: a population-based longitudinal study. Int J Obes. 2019;43:139–48.

    Article  Google Scholar 

  20. Chitturi S, Farrell GC, Hashimoto E, Saibara T, Lau GK, Sollano JD. Non-alcoholic fatty liver disease in the Asia–Pacific region: definitions and overview of proposed guidelines. J Gastroenterol Hepatol. 2007;22:778–87.

    Article  PubMed  Google Scholar 

  21. Powell EE, Wong VW-S, Rinella M. Non-alcoholic fatty liver disease. Lancet. 2021;397:2212–24.

    Article  CAS  PubMed  Google Scholar 

  22. Okamura T, Hashimoto Y, Hamaguchi M, Obora A, Kojima T, Fukui M. Creatinine-to‐bodyweight ratio is a predictor of incident non‐alcoholic fatty liver disease: A population‐based longitudinal study. Hepatol Res. 2020;50:57–66.

    Article  CAS  PubMed  Google Scholar 

  23. Yassin A, Haider A, Haider KS, Caliber M, Doros G, Saad F, et al. Testosterone therapy in men with hypogonadism prevents progression from prediabetes to type 2 diabetes: Eight-Year data from a registry study. Diabetes Care. 2019;42:1104–11.

    Article  CAS  PubMed  Google Scholar 

  24. Park K, Ahn CW, Lee SB, Kang S, Nam JS, Lee BK, et al. Elevated TyG index predicts progression of coronary artery calcification. Diabetes Care. 2019;42:1569–73.

    Article  CAS  PubMed  Google Scholar 

  25. Groenwold RHH, White IR, Donders ART, Carpenter JR, Altman DG, Moons KGM. Missing covariate data in clinical research: when and when not to use the missing-indicator method for analysis. CMAJ. 2012;184:1265–9.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Liang X, Xing Z, Lai K, Li X, Gui S, Li Y. Sex differences in the association between metabolic score for insulin resistance and the reversion to normoglycemia in adults with prediabetes: a cohort study. Diabetol Metab Syndr. 2024;16:183.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Liang X, Xing Z, Li Y, Gui S, Hu H. Non-linear dose-response relationship between the visceral adiposity index and diabetes in adults with normoglycemia: a cohort study. Front Endocrinol (Lausanne). 2024;15:1441878.

    Article  PubMed  Google Scholar 

  28. Jaddoe VWV, de Jonge LL, Hofman A, Franco OH, Steegers EAP, Gaillard R. First trimester fetal growth restriction and cardiovascular risk factors in school age children: population based cohort study. BMJ. 2014;348:g14.

    Article  PubMed  PubMed Central  Google Scholar 

  29. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Int J Surg. 2014;12:1495–9.

    Article  Google Scholar 

  30. Nazarzadeh M, Bidel Z, Canoy D, Copland E, Wamil M, Majert J, et al. Blood pressure Lowering and risk of new-onset type 2 diabetes: an individual participant data meta-analysis. Lancet. 2021;398:1803–10.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Wikanendra GB, Suhadi R, Setiawan CH, Virginia DM, Hendra P, Fenty F. Correlation among sleep duration, blood pressure, and blood glucose level of Morangan people, Sindumartani, Ngemplak, Sleman. J Pharm Sci Community. 2020;17.

  32. Tian X, Zuo Y, Chen S, Liu Q, Tao B, Wu S, et al. Triglyceride–glucose index is associated with the risk of myocardial infarction: an 11-year prospective study in the Kailuan cohort. Cardiovasc Diabetol. 2021;20:19.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Millstein RJ, Pyle LL, Bergman BC, Eckel RH, Maahs DM, Rewers MJ, et al. Sex-specific differences in insulin resistance in type 1 diabetes: the CACTI cohort. J Diabetes Complicat. 2018;32:418–23.

    Article  Google Scholar 

  34. Choi SH, Kim TH, Lim S, Park KS, Jang HC, Cho NH. Hemoglobin A1c as a diagnostic tool for diabetes screening and new-onset diabetes prediction: a 6-year community-based prospective study. Diabetes Care. 2011;34:944–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Rönnemaa T, Rönnemaa EM, Puukka P, Pyörälä K, Laakso M. Smoking is independently associated with high plasma insulin levels in nondiabetic men. Diabetes Care. 1996;19:1229–32.

    Article  PubMed  Google Scholar 

  36. Baek W, Lee J-W, Lee HS, Han D, Choi S-Y, Chun EJ, et al. Concurrent smoking and alcohol consumers had higher triglyceride glucose indices than either only smokers or alcohol consumers: a cross-sectional study in Korea. Lipids Health Dis. 2021;20:49.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Sun K, Liu J, Ning G. Active smoking and risk of metabolic syndrome: A Meta-Analysis of prospective studies. PLoS ONE. 2012;7:e47791.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Lin L, Chen C, Yu X. [The analysis of threshold effect using empower stats software]. Zhonghua Liu Xing Bing Xue Za Zhi. 2013;34:1139–41.

    PubMed  Google Scholar 

  39. Haneuse S, VanderWeele TJ, Arterburn D. Using the E-Value to assess the potential effect of unmeasured confounding in observational studies. JAMA. 2019;321:602–3.

    Article  PubMed  Google Scholar 

  40. Su S, Liu X, Zhu M, Liu W, Liu J, Yuan Y, et al. Trehalose ameliorates nonalcoholic fatty liver disease by regulating IRE1α–TFEB signaling pathway. J Agric Food Chem. 2025;73:521–40.

    Article  CAS  PubMed  Google Scholar 

  41. Pei K, Gui T, Kan D, Feng H, Jin Y, Yang Y, et al. An overview of lipid metabolism and nonalcoholic fatty liver disease. Biomed Res Int. 2020;2020:4020249.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Min H-K, Mirshahi F, Verdianelli A, Pacana T, Patel V, Park C-G, et al. Activation of the GP130-STAT3 axis and its potential implications in nonalcoholic fatty liver disease. Am J Physiology-Gastrointestinal Liver Physiol. 2015;308:G794–803.

    Article  CAS  Google Scholar 

  43. Cui N, Li H, Dun Y, Ripley-Gonzalez JW, You B, Li D et al. Exercise inhibits JNK pathway activation and lipotoxicity via macrophage migration inhibitory factor in nonalcoholic fatty liver disease. Front Endocrinol. 2022;13.

  44. Gehrmann W, Elsner M, Lenzen S. Role of metabolically generated reactive oxygen species for lipotoxicity in pancreatic β-cells. Diabetes Obes Metabolism. 2010;12:149–58.

    Article  CAS  Google Scholar 

  45. Eguchi N, Vaziri ND, Dafoe DC, Ichii H. The role of oxidative stress in pancreatic Β cell dysfunction in diabetes. Int J Mol Sci. 2021;22:1509.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Drews G, Krippeit-Drews P, Düfer M. Oxidative stress and beta-cell dysfunction. Pflugers Arch - Eur J Physiol. 2010;460:703–18.

    Article  CAS  Google Scholar 

  47. Fang Y, Zhang Q, Tan J, Li L, An X, Lei P. Intermittent hypoxia-induced rat pancreatic β-cell apoptosis and protective effects of antioxidant intervention. Nutr Diabetes. 2014;4:e131–131.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Zhao J, Fan H, Wang T, Yu B, Mao S, Wang X, et al. TyG index is positively associated with risk of CHD and coronary atherosclerosis severity among NAFLD patients. Cardiovasc Diabetol. 2022;21:123.

    Article  PubMed  PubMed Central  Google Scholar 

  49. Antunes GC, de Lima RD, Vieira RFL, Macêdo APA, Muñoz VR, Zambalde EP, et al. Resistance exercise attenuates IKKε phosphorylation and hepatic fat accumulation of obese mice. Clin Exp Pharmacol Physiol. 2022;49:1072–81.

    Article  CAS  PubMed  Google Scholar 

  50. Lutkewitte AJ, Singer JM, Shew TM, Martino MR, Hall AM, He M, et al. Multiple antisense oligonucleotides targeted against monoacylglycerol acyltransferase 1 (Mogat1) improve glucose metabolism independently of Mogat1. Mol Metabolism. 2021;49:101204.

    Article  CAS  Google Scholar 

  51. Liu T, Xuan H, Yin J, Wang L, Wang C, Xu X, et al. Triglyceride Glucose Index Increases Significantly Risk of Hypertension Development in Chinese Individuals Aged ≥45 Years Old: Analysis from the China Health and Retirement Longitudinal Study. JMDH. 2023;16:63–73.

  52. Lewis GF, Carpentier A, Adeli K, Giacca A. Disordered fat storage and mobilization in the pathogenesis of insulin resistance and type 2 diabetes. Endocr Rev. 2002;23:201–29.

    Article  CAS  PubMed  Google Scholar 

  53. Hameed EK, Al-Ameri LT, Hasan HS, Abdulqahar ZH. The Cut-off values of Triglycerides - Glucose index for metabolic syndrome associated with type 2 diabetes mellitus. Baghdad Sci J. 2022;19:0340–0340.

    Article  Google Scholar 

  54. Zeng Y, Yin L, Yin X, Zhao D. Association of triglyceride-glucose index levels with gestational diabetes mellitus in the US pregnant women: a cross-sectional study. Front Endocrinol. 2023;14.

  55. Dang K, Wang X, Hu J, Zhang Y, Cheng L, Qi X, et al. The association between triglyceride-glucose index and its combination with obesity indicators and cardiovascular disease: NHANES 2003–2018. Cardiovasc Diabetol. 2024;23:8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Guo W, Lu J, Qin P, Li X, Zhu W, Wu J, et al. The triglyceride-glucose index is associated with the severity of hepatic steatosis and the presence of liver fibrosis in non-alcoholic fatty liver disease: a cross-sectional study in Chinese adults. Lipids Health Dis. 2020;19:218.

    Article  PubMed  PubMed Central  Google Scholar 

  57. Amernia B, Moosavy SH, Banookh F, Zoghi G. FIB-4, APRI, and AST/ALT ratio compared to fibroscan for the assessment of hepatic fibrosis in patients with non-alcoholic fatty liver disease in Bandar Abbas, Iran. BMC Gastroenterol. 2021;21:453.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Rinella ME, Lazarus JV, Ratziu V, Francque SM, Sanyal AJ, Kanwal F, et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. J Hepatol. 2023;79:1542–56.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

We give our sincere gratitude to Professor Takuro Okamura and their team for primary research data provision.

Funding

This work was supported by Sanming Project of Medicine in Shenzhen (No. SZSM202211016), Shenzhen Fund for Guangdong Provincial High-level Clinical Key Specialties (No. SZGSP006), Shenzhen Second People’s Hospital Clinical Research Fund of Guangdong Province High-level Hospital Construction Project (Grant No. 20223357008, 2023xgyj3357003).

Author information

Author notes

  1. Ying Li, Zemao Xing and Shuiqing Gui contributed equally to this work.

Authors and Affiliations

  1. Department of Critical Care Medicine, Shenzhen Second People’s Hospital, Shenzhen, China

    Xiaomin Liang, Kai Lai, Xiaohong Li, Ying Li, Zemao Xing & Shuiqing Gui

  2. Department of Critical Care Medicine, The First Affiliated Hospital of Shenzhen University, Shenzhen, China

    Xiaomin Liang, Kai Lai, Xiaohong Li, Ying Li, Zemao Xing & Shuiqing Gui

Authors
  1. Xiaomin Liang
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Kai Lai
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Xiaohong Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Ying Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Zemao Xing
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Shuiqing Gui
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

X.M.L: Research design, data analysis and manuscript drafting. K.L and X.H.L: Data collection and cleansing. Y.L, Z.M.X, and S.Q. G: Research designs and manuscript revisions. All authors collaborated on article and cleared final version.

Corresponding authors

Correspondence to Ying Li, Zemao Xing or Shuiqing Gui.

Ethics declarations

Ethics approval and consent to participate

The study was authorized by the Murakami Memorial Hospital Ethics Committee, and all subjects were required to provide written informed consent before recruitment.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary Material 2

Supplementary Material 3

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Liang, X., Lai, K., Li, X. et al. Non-linear relationship between triglyceride glucose index and new-onset diabetes among individuals with non-alcoholic fatty liver disease: a cohort study. Lipids Health Dis 24, 94 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-025-02518-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12944-025-02518-5

Keywords

Lipids in Health and Disease

ISSN: 1476-511X