Value of the Triglyceride Glucose index Combined with Body Mass index in Predicting non-Alcoholic Fatty Liver Disease in Patients with Type 2 Diabetes



Objective: To explore the predictive value of the triglyceride glucose-body mass index (TyG-BMI) in relation to the occurrence of non-alcoholic fatty liver disease (NAFLD) in the Chinese population with type 2 diabetes (T2D).

Methods: We selected 826 patients with T2D who were hospitalized at the Department of Endocrinology and Metabolism of Karamay People's Hospital from September 2016 to October 2018 for this research. The height, weight, fasting blood glucose, serum insulin, and lipid profiles of the subjects were collected. Stepwise logistic regression analysis was performed to calculate the triglyceride glucose (TyG) index, TyG-BMI, and the ratio of the insulin resistance index (HOMA-IR) to triglyceride high-density lipoprotein-cholesterol at risk for non-alcoholic fatty liver disease. The curve method was used to analyze its predictive value for non-alcoholic fatty liver.

Results Results of the logistic regression analysis showed that the risk ratios of NAFLD were 6.54 (3.70-11.53) and 4.868 (2.576-9.200) for the TyG-BMI before and after correction. The area under the curve (AUC) for TyG-BMI was 0.727 (0.691–0.764), which was the highest among all the other parameters studied.

Conclusion Compared with the TyG index, the TG/HDL-C, HOMA-IR, and the TyG-BMI are more effective predictors of NAFLD in T2D.

1. Background

Non-alcoholic fatty liver disease (NAFLD) refers to a clinic pathological syndrome characterized by liver steatosis, which is caused by the exclusion of alcohol and other clear liver damage factors. If the disease progresses, it can evolve into liver fibrosis and even liver cancer. Several studies have shown that NAFLD is related to diseases such as type 2 diabetes, abdominal obesity, dyslipidemia, hypertension, and cardiovascular diseases, and is closely related to insulin resistance (IR) and genetic susceptibility to metabolic stress-induced liver injury. [1–5]

In recent years, with the global epidemic of obesity and metabolic syndrome, the incidence of NAFLD has gradually increased worldwide [6–7]. NAFLD is one of the important global public health issues of the 21st century. NAFLD affects 30% of adults and 10% of children in the United States [8–9].In the past decade, NAFLD has grown rapidly and has been presenting a trend of younger onset in China. In 2004, the prevalence of NAFLD among adults in Shanghai was approximately 15.4%.and in 2011, the prevalence of NAFLD in adult communities in Beijing was up to 35.1% [10-11] .NAFLD is a chronic liver condition that is gaining more and more importance in China.

In a retrospective study assessing the clinical characteristics and initial disease severity of patients with NAFLD and the incidence and risk factors of NAFLD progression[12], 12.3% of NAFLD-free patients showed progression, 24.7% of patients with NAFLD combined with T2DM progressed. The risk of T2DM and disease progression is about twice the risk of T2DM-free disease progression, and the mortality risk increases as the disease progresses. In this study, this association between T2DM and NAFLD was found in nearly 40% of NAFLD patients, further illustrating the importance of diabetes management to reduce the risk and adverse consequences of liver-related death. Therefore, it is particularly important to screen for NAFLD patients in the T2DM population as well as monitor disease progression of NAFLD.

 T2DM and NAFLD are both related to IR. IR refers to a state in which the body exhibits reduced sensitivity and reactivity to insulin metabolism [13]. The current “gold standard” for evaluating IR is the high insulin-positive glucose clamp test. However, as it is a complicated, time-consuming, and labor-intensive test, its wide application in clinical work remains limited.Recent studies have reported that the triglyceride glucose (TyG) index, which is calculated on the basis of triglyceride and fasting blood glucose levels, and the steady-state model can be used to identify the insulin resistance index (HOMA-IR). There is a significant correlation between the glucose metabolism rate M value obtained from the glucose clamp test [14–16], and it has become a reliable proxy for evaluating IR. To date, a few studies have evaluated the relationship between the TyG index and the incidence of NAFLD in non-diabetic populations [17–19], but no study has investigated the relationship between the TyG-BMI and incidence of NAFLD in diabetic populations. In this study, we tried to clarify the relationship between NAFLD and the TyG with BMI in the T2DM population and explore its predictive value for the occurrence of NAFLD in the T2DM population in China.

2. Materials And Methods

2.1 Study participants

A total of 826 in-patients who were treated at the Department of Endocrinology and Metabolism, Karamay People’s Hospital from September 1, 2016, to October 31, 2018 were screened for participation in this cross-sectional observational study referring to the practice guidelines of the American Gastroenterology Association and the American Liver Disease Research Association [20]. Patients with type 2 diabetes (T2D) were included in the study. Of the 826 patients, 552 had NAFLD, 274 did not have NAFLD. The following inclusion criteria were applied for patients with T2D and NAFLD: no history of drinking or an alcohol consumption of <140 g/week (women < 70 g/week), and T2D patients had diffuse fatty liver detected by ultrasound imaging. Patients were screened according to the following exclusion criteria: previous long-term heavy drinking or combined with viral hepatitis, drug-induced liver disease, total parenteral nutrition, hepatolenticular degeneration, autoimmune liver disease and other specific diseases that can lead to fatty liver; inflammatory bowel disease; hypothyroidism; Cushing's syndrome; β-lipoproteinemia; and insulin resistance-related conditions such as lipoatrophic diabetes and Mauriac syndrome; and other patients with T2D that could cause fatty liver.

2.2 Medical data collection and physical examination

This research project follows the Helsinki Declaration and China’s clinical research management norms and regulations. The research plan was approved by the Medical Ethics Committee of Karamay People’s Hospital. Inpatients had to voluntarily sign informed consent forms before they could be used as research objects for research data collection. General demographic information and anthropometric measurement data were collected for use in the research. For the demographic information, age, ethnicity, gender, occupation, education level, previous medical history, and personal lifestyle(e.g. alcohol abuse, smoking). For anthropometric measurements, the subjects were asked to fast, take off their shoes and wear light clothing, and height and weight measurements were obtained. The body mass index (BMI) = weight (kg)/height (m)2. Blood pressure was measured according to the recommendations of the American Heart Association, and the blood pressure in the right arm was recorded by a qualified investigator (nurse) with a mercury sphygmomanometer. Blood pressure (BP) was measured three times, and the average of the systolic and diastolic blood pressure values was used for analysis.

2.3 Laboratory measurement and index calculation

Venous blood was collected early in the morning after the patient had fasted for at least 8 hours. Determination of biochemical indicators: fasting blood glucose, triglyceride, total cholesterol, and high-density lipoprotein cholesterol (HDL-C) levels were measured using the COBAS 8000 chemical analyzer (Roche, Germany). Plasma insulin levels were measured using an E601 automatic chemiluminescence system (Roche, Germany). On the day of blood collection, blood biochemical indicators were determined at the Medical Test Center of Karamay People's Hospital, Xinjiang, China. Quantitative analysis of insulin resistance uses the insulin resistance index (HOMA-IR) obtained by the steady-state model to identify insulin resistance. The HOMA-IR is calculated as follows: HOMA-IR = fasting insulin (μU/dL) × fasting blood glucose (mg/dL)/22.5. TyG index [21]: Ln [TG (mg/dL) × fasting blood glucose (mg/dL)/2]. The TyG-BMI represents the TyG index × BMI [22].

2.4 Ultrasound analysis

A color Doppler ultrasound system (IU22, Philips Healthcare, Andover, MA) with a 1.0–5.0 MHz sensor was used to perform an abdominal ultrasound scan to diagnose fatty liver. Abdominal ultrasonography was performed on the subjects by two professionally trained and experienced ultrasound diagnostic physicians in a blinded manner. The examiner was blinded to the clinical information of the subject, and used the echogenicity of the liver tissue, the difference between the liver and the right kidney, and the visibility of the vascular structure to arrive at a diagnosis [23]. In this study, signs of hepatic steatosis were considered to be NAFLD if the liver ultrasound showed any degree of echogenic enhancement of liver tissue and if the liver appeared brighter than the renal cortex on ultrasound. Subjects with NAFLD were classified according to the presence and severity of this disease.

2.5 Statistical analysis

Excel 2007 was used to generate the database, and errors were corrected after double data entry. All statistical analyses were performed using the SPSS 22.0 statistical software package (IBM, Armonk, New York). Continuous data for skewed distributions were expressed as medians and interquartile ranges (IQR) and compared using Kruskal-Wallis H test or the Mann-Whitney U test. Categorical variables were compared using the chi-square test. Before and after adjusting for any confounding factors (age, gender, weight, systolic BP, diastolic BP, smoking, drinking, diabetes duration, and exercise), compare with quartile 1 (Below 25% with quartile) for reference, and stepwise logistic regression analysis was applied to calculate the TyG index, TyG-BMI, and TG/HDL-C quartiles 2–4. The hazard ratio of NAFLD and the 95% confidence interval (CI) for HOMA-IR were compared with the reference value below the 75th percentile. The hazard ratio of NAFLD was calculated above the 95th percentile and 95% CIs. Then, we determined the receiver operating characteristic curve (ROC) of each parameter and calculated the area under the curve (AUC) and compared the AUC between different groups. We then studied the ability of these parameters to predict the occurrence of NAFLD, derived the coordinate values of the curve, and calculated their sensitivity and specificity. The point with the highest sensitivity as well as specificity was considered the cutoff point. The difference was statistically significant with a P value of <0.05 (two-tailed).

3. Results

3.1 Patient clinical characteristics

A total of 826 subjects were included in the study. Of these, 274 had no NAFLD and had an average age of 59 (49–67) years, while 552 patients had NAFLD, and had an average age of 55 (47–64) years. Among the patients with NAFLD, There were 375 men, 177 women, and among the patients without NAFLD, there were 178 men and 96 women. There were no statistically significant differences in the gender ratio, SBP, and HbA1C between the two groups (P > 0.05); the age, BMI, DBP, and duration of diabetes, fasting blood glucose (FBG), postprandial blood glucose (PBG), HOMA-IR, triglyceride (TG), total cholesterol (TC), high-density lipoprotein-cholesterol (HDL-C), and TyG were significantly different among the groups (P < 0.05). The BMI, DBP, FBG, PBG, HOMA-IR, TG, TC, and TyG were significantly higher in patients with NAFLD than in those without NAFLD, while the age, duration of diabetes, and HDL-C were significantly lower in patients with NAFLD than in those without NAFLD (see Table 1).

3.2 Relevant indicators and risk assessment of the incidence of NAFLDA

The results of the logistic regression analysis showed that compared with participants in the minimum quartile (Q1), the TyG index, TyG-BMI, and Q2–Q4 of the TG/HDL-C ratio had a higher risk ratio (OR) of NAFLD (Table 2). The TyG-BMI had the highest hazard ratio of NAFLD. Before and after adjustment, the hazard ratio (OR) of Q4 reached 6.54 (95% CI 3.70-11.53) and 4.868 (95% CI 2.576–9.200, P < 0.001). The second highest hazard ratio was for the TyG index, with OR values of 3.455 (2.060–5.795) and 3.405 (1.900–6.102), and the hazard ratios of HOMA-IR before and after adjusting for confounding factors were 2.451 (1.539–3.904) and 2.951 (1.732–5.026), respectively.

3.3 Comparison of the parameters to the predictive power of NAFLD

The results of the ROC curve analysis of the TyG index, TyG-BMI, HOMA-IR, and the TG/HDL-C ratio corresponding to 95% CI are shown in Table 3 and Figure 1. The AUC of NAFLD was the highest for TyG-BMI at 0.727 (95% CI 0.691–0.764), followed by TG/HDL-C (0.657, 95% CI 0.617–0.696), and HOMA-IR (0.655, 95% CI 0.616–0.694), and TyG (0.651, 95% CI 0.611–0.691). In the gender subgroup analysis, the AUCs for the TyG-BMI of men and women with NAFLD were 0.739 (0.695–0.783) and 0.702(0.636–0.768), respectively. When the cutoff value of the ROC curve drawn by the TyG-BMI for NAFLD is 169.92, the sensitivity and specificity of the predicted NAFLD were 62.2% and 73.8%, respectively. Compared with the other three parameters, the sensitivity + specificity value corresponding to the cutoff value of TyG-BMI was the largest. (see Table 3).

Among men, the most sensitive parameter for predicting NAFLD was TyG, followed by TG/HDL-C and HOMA-IR, and the most specific parameter for predicting NAFLD was TyG-BMI, followed by HOMA-IR and TG/HDL-C. Among women, the most sensitive parameter for predicting NAFLD is TyG-BMI (61.7%), followed by the TyG index and the TG/HDL-C. Furthermore, the most specific parameter for predicting NAFLD was HOMA-IR (84%), followed by the TyG index and the TG/HDL-C ratio.

4. Discussion

Currently, regarding the pathogenesis of NAFLD, it is believed that the combination of genetic susceptibility to this condition and the presence of multiple factors such as IR, inflammatory factors secreted by adipose tissue, gut microbiota, and specific genetic and epigenetic factors trigger the onset of NAFLD. Of these, IR plays a key role in the pathogenesis of fatty liver, which can cause excessive lipid deposition in liver cells, which is closely related to the occurrence of NAFLD [24–26]. The pathophysiology of NAFLD is IR, which is clinically manifested as metabolic syndrome, i.e., hypertension, hyperlipidemia, central obesity, hyperglycemia, and NAFL.

The results revealed that the TyG index can effectively identify the risk of IR in Chinese individuals [27]. In the San Antonio metabolism (SAM) study, Gastaldelli et al. [28] proposed that because TyG is closely related to the liver fat mass, it is not a good method to measure peripheral IR, but it is a good method to measure liver IR. In fact, hypertriglyceridemia can increase the transport of free fatty acids to the liver, cause liver fat accumulation, hepatic IR, cause fatty liver, and increase glucose output in the liver. Studies have found that the TyG index calculated on the basis of the TGs and FBG levels can diagnose steatosis, is associated with IR and can predict IR. However, this measurement is confounded by the presence of fibrosis and inflammation, as a result of which steatosis is not accurately quantified [29].

Studies have shown that increasing TG and decreasing HDL-C levels can lead to IR. When the circulating TG levels are high, heparin activates lipoprotein lipase to increase intravascular lipolysis of TG, thereby increasing the risk of tissue exposure to free fatty acids (FFAs). High FFAs can cause IR through oxidative stress pathways [30]. Clinical studies of Caucasian populations have proven that the TG/HDL-C ratio can predict IR, and several studies conducted in China have also shown that TG/HDL-C can predict IR [31–32].

The relationship between obesity and IR has also been well established, and excess adipose tissue has been shown to promote insulin resistance [33]. Studies have shown that obesity is closely related to liver steatosis. BMI is related to the occurrence of NAFLD in the general population or in specific disease groups such as among patients with hypertension. Furthermore, it has been reported that 65–92.3% of patients with a BMI of >40 kg/m2 have NAFLD, and the higher the BMI in NAFLD patients, the more severe is the case of liver steatosis [34]. The BMI may affect the predicted TyG value for NAFLD.

Therefore, combining the TyG and obesity indices can help better predict the occurrence of IR and NAFLD compared to the TyG index alone. Zhang et al. revealed that after adjusting for potential confounding factors, there is a strong positive correlation between the TyG-BMI and NAFLD risk. The TyG-BMI can accurately identify NAFLD, as the AUC of TyG-BMI was 0.835 (0.824–0.845), which is higher than that of TyG, BMI, TG, FPG, and other components. Thus, TyG-BMI is an effective indicator for identifying non-obese NAFLD patients. In this study, we compared the efficacy of TyG, TyG-BMI, and four other parameters to predict the occurrence of NAFLD in patients with T2DM. The results revealed that TyG-BMI has an AUC of 0.727 (95% CI, 0.691–0.764) in the accuracy of predicting NAFLD in T2D. The optimal cutoff point for the diagnosis of NAFLD is 169.92. At this time, the sensitivity and specificity of this factor were 62.2% and 73.8%, respectively. The results suggest that compared with TyG, TG/HDL-C ratio, and HOMA-IR, the combination of TyG index and BMI can better predict the occurrence of T2D and NAFLD in both men and women, and the accuracy of TyG-BMI in predicting NAFLD with type 2 diabetes was also the highest in both men and women.

Abdominal obesity includes subcutaneous adipose tissue and visceral adipose tissue. Visceral adipose tissue has a greater effect on the IR [33]. Studies have shown that visceral fat produces more FFAs than subcutaneous fat, thereby increasing the risk of IR and diabetes [34]. In addition, visceral fat secretes a variety of inflammatory cytokines and adipokines, which may also promote the occurrence of IR and diabetes [34,35]. In this study, TyG-BMI was a more accurate predictor of NAFLD in men than in women, which may be related to the fact that male obesity is mostly abdominal obesity.

Currently, liver biopsies are the best diagnostic and staging methods for nonalcoholic steatohepatitis (NASH) and NAFLD. However, it is invasive, and its associated complications and irregular liver biopsy sampling limit its use. Noninvasive tools for detecting NAFLD include ultrasound, computed tomography, and magnetic resonance spectroscopy. The latter two are expensive and time-consuming tools, and ultrasound is currently recommended as the first-line imaging technique for the clinical screening of NAFLD patients. In addition, researching a simple and effective diagnostic tool that can identify the risk of NAFLD at an early stage will help the early detection and management of such patients, which is very important for public health. The results of this study suggest that the combination of the triglyceride glucose index and body mass index (TyG-BMI) is a good indicator for identifying IR and predicting NAFLD in patients with T2D.

The present study has several limitations. First of all, due to its cross-sectional design, the identified relationship is not forward-looking, and causality cannot be determined. Further prospective cohort studies are needed to determine whether TyG-BMI can predict the future occurrence of NAFLD. Second, due to the lack of waist circumference information, TyG and abdominal obesity indicators could not be combined for analysis and comparison. In addition, the research subjects are from inpatients and the number of cases is relatively small. If there is a large sample of natural populations derived from outpatient examinations and participating in health examinations, the research results may be better.

This study shows that TyG-BMI is a strong predictor of NAFLD in T2DM patients. This result also suggests that reducing blood TG levels, weight loss, and increased physical activity are important measures that will help prevent NAFLD in T2DM patients. This is also the main management measure to prevent the occurrence of NAFLD in patients with T2D.

5. Conclusion

A very high proportion of T2D patients have NAFLD. In our study, nearly 67% of patients with T2D had NAFLD. Therefore, it is worthwhile to screen for effective NAFLD markers in patients with T2D. TyG-BMI is a valuable index for screening NAFLD, and it is an effective noninvasive method to identify NAFLD. To improve the prediction performance of NAFLD in patients with T2D, it can be predicted at a low cost using values obtained from routine laboratory tests. Therefore, we recommend applying the TyG-BMI value to the risk assessment of NAFLD in people with T2D in clinical practice and in future epidemiological studies.


NAFLD non-alcoholic fatty liver disease

ROC receiver operating characteristics

BMI body mass index

AUC Areas under the curve

CI Confidence intervals

TyG index, a product of triglyceride and fasting glucose

HOMA-IR, the homeostasis model assessment for insulin resistance


Ethics approval and consent to participate

This study was approved by the Medical Ethics Committee of People’s Hospital of Karamay. Informed consent was obtained from each patient.

Consent for publication

Not applicable.

Availability of data and materials

All data generated or analyzed during this study are included in this article.

Competing interests

All authors declare that they have no competing interests.


Not applicable.


We thank all the doctors and nurses at the Department of Endocrinology and Metabolism and Department of Ultrasound Imaging (People's Hospital of Karamay) who participated in this study. We thank the Karamay Science and Technology Bureau (XinJiang, Karamay city, China) for its financial support.

Contributorship statement

Nong Li has substantially contributed to the conception, acquisition, analysis, and interpretation of data, and has drafted the paper, revised it critically, and approved of the final version. Huiwen Tan has substantially contributed to the interpretation of data, and drafted the paper or revised it critically. Xue fang Huang has substantially contributed to the paper, drafted it, revised it critically, and approved of the final version Aixia Xie has substantially contributed to the acquisition of data, drafted the paper. Cheng Li has substantially contributed to the interpretation of data. Xuan Fu has substantially contributed to the acquisition of data.Weiting Xang has substantially contributed to the acquisition of data.Amina Kirim has substantially contributed to the acquisition of data.


  1. Yki-Jarvinen H. Non-alcoholic fatty liver disease as a cause and a consequence of metabolic syndrome. Lancet Diabetes Endocrinol. 2014;2:901–910.
  2. Anstee QM, Targher G, Day CP. Progression of NAFLD to diabetes mellitus, cardiovascular disease or cirrhosis. Nat Rev Gastroenterol Hepatol. 2013;10:330–344.
  3. Vanni E, Bugianesi E, Kotronen A, De Minicis S, Yki-Jarvinen H, Svegliati-Baroni G. From the metabolic syndrome to NAFLD or vice versa? Dig Liver Dis. 2010;42:320–330.
  4. van den Berg EH, Amini M, Schreuder TC, Dullaart RP, Faber KN, Alizadeh BZ, et al. Prevalence and determinants of non-alcoholic fatty liver disease in lifelines: a large Dutch population cohort. PLoS One. 2017;12:e0171502.
  5. Lee JI, Kim MC, Moon BS, Song YS, Han EN, Lee HS, et al. The relationship between 10-year cardiovascular risk calculated using the pooled cohort equation and the severity of non-alcoholic fatty liver disease. Endocrinol Metab (Seoul) 2016;31:86–92.
  6. Browning JD, Szczepaniak LS, Dobbins R, Nuremberg P, Horton JD, Cohen JC, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology. 2004;40:1387–1395.
  7. Smits MM, Ioannou GN, Boyko EJ, Utzschneider KM. Non-alcoholic fatty liver disease as an independent manifestation of the metabolic syndrome: results of a US national survey in three ethnic groups. J Gastroenterol Hepatol. 2013;28:664–670.
  8. Eguchi A, Povero D, Alkhouri N, Feldstein AE.Novel therapeutic targets for nonalcoholic fatty liver disease. Expert Opin Ther Targets. 2013;17(7):773-9. doi: 10.1517/14728222.2013.789502.
  9. Cohen JC, Horton JD, Hobbs HH. Human fatty liver disease: old questions and new insights. Science. 2011;332:1519–1523.
  10. Fan JG, Zhu J, LiXJ, et al. Fatty liver and the metabolic syndromeamong Shanghai adults. J Gastroenterol Hepatol, 2005, 20:1825-1832.
  11. Yan J, Xie W, Ou WN, et al. Epidemiological survey and risk factoranalysis of fatty liver disease of adult residents, Beijing, China [J]. J Gastroenterol Hepatol, 2013, 28:1654-1659.
  12. Simeone JCBae JPHoogwerf BJLi Q1, Haupt AAli AKBoardman MK, Nordstrom BL.Clinical course of non-alcoholic fatty liver disease: an assessment of severity, progression, and outcomes.Articles from Clinical Epidemiology 2017 Dec 14;9:679-688. doi: 10.2147/CLEP.S144368. eCollection 2017.
  13. MuniyappaR.Lee S, a1.Current approachesforassessing insulin sensitivity and resistance in vivo:advantages,limitations,and appropriateusage.Am J Physiel Endoednol Melab,2008.294:E15-E26.
  14. Simental-Mendia LE, Rodriguez-Moran M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat Disord. 2008;6:299–304.
  15. Guerrero-Romero F, Simental-Mendia LE, Gonzalez-Ortiz M, Martinez-Abundis E, Ramos-Zavala MG, Hernandez-Gonzalez SO, et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab. 2010;95:3347–3351.
  16. Vasques AC, Novaes FS, de Oliveira Mda S, Souza JR, Yamanaka A, Pareja JC, et al. TyG index performs better than HOMA in a Brazilian population: a hyperglycemic clamp validated study. Diabetes Res Clin Pract. 2011;93:e98–e100.
  17. Simental-Mendia LE, Simental-Mendia E, Rodriguez-Hernandez H, Rodriguez-Moran M, Guerrero-Romero F. The product of triglycerides and glucose as biomarker for screening simple steatosis and NASH in asymptomatic women. Ann Hepatol. 2016;15:715–720.
  18. Zhang S, Du T, Zhang J, Lu H, Lin X, Xie J, et al. The triglyceride and glucose index (TyG) is an effective biomarker to identify non-alcoholic fatty liver disease. Lipids Health Dis. 2017;16:15.
  19. Zhang S, Du T, Li M, Jia J, Lu H, Lin X, et al. Triglyceride glucose-body mass index is effective in identifying non-alcoholic fatty liver disease in nonobese subjects. Medicine (Baltimore) 2017;96:e7041.
  20. Chalasani N, Younossi Z, Lavine JE, et al. The diagnosis and management of non-alcoholic fatty liver disease:practice guideline by the American Gastroenterological Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology. Gastroenterology, 2012,142(7):1592-1609.
  21. Simental-Mendía LE, Rodríguez-Morán M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat Disord. 2008;6: 299–304. 10.1089/met.2008.0034
  22. Zheng S, Shi S, Ren X, Han T, Li Y, Chen Y, et al. Triglyceride glucose-waist circumference, a novel and effective predictor of diabetes in first-degree relatives of type 2 diabetes patients: cross-sectional and prospective cohort study. J Transl Med. 2016;14: 260 10.1186/s12967-016-1020-8
  23. Saadeh S, Younossi ZM, Remer EM, Gramlich T, Ong JP, Hurley M, et al. The utility of radiological imaging in non-alcoholic fatty liver disease. Gastroenterology. 2002;123:745–750.
  24. Birkenfeld AL, Shulman GI. Non-alcoholic fatty liver disease, hepatic insulin resistance, and type 2 diabetes. Hepatology. 2014;59:713–723.
  25. Machado M, Cortez-Pinto H. Non-alcoholic fatty liver disease and insulin resistance. Eur J Gastroenterol Hepatol. 2005;17:823–826.
  26. Buzzetti E, Pinzani M, Tsochatzis EA.The multiple-hit pathogenesis of non-alcoholic fatty liver disease (NAFLD).Metabolism. 2016;65(8):1038-48.
  27. Du T, Yuan G, Zhang M, Zhou X, Sun X, Yu X. Clinical usefulness of lipid ratios, visceral adiposity indicators, and the triglycerides and glucose index as risk markers of insulin resistance. Cardiovasc Diabetol. 2014;13:146. doi: 10.1186/s12933-014-0146-3.
  28. Gastaldelli A, Ferrannini E, Miyazaki Y, Matsuda M, DeFronzo RA; San Antonio metabolism study.Beta-cell dysfunction and glucose intolerance: results from the San Antonio metabolism (SAM) study.Diabetologia. 2004 Jan;47(1):31-9. doi: 10.1007/s00125-003-1263-9
  29. Fedchuk L, Nascimbeni F, Pais R, Charlotte F, Housset C, Ratziu V, et al. Performance and limitations of steatosis biomarkers in patients with non-alcoholic fatty liver disease. Aliment Pharmacol Ther. 2014;40:1209–1222. doi: 10.1111/apt.12963.
  30. Lam TKT, Carpentier A, Lewis GF, et al. Mechanisms of the free fatty acid-induced increase in hepatic glucose production. Am J Physiol Endocrinol and Metab. 2003;284(5):E863–73.
  31. Chiang JK, Lai NS, Chang JK, Koo M. Predicting insulin resistance using the triglyceride-to-high-density lipoprotein cholesterol ratio in Taiwanese adults. Cardiovasc Diabetol. 2011; 17;10: 93.doi: 10.1186/1475-2840-10-93.
  32. Zhang L, Chen S, Deng A, et al. Association between lipid ratios and insulin resistance in a Chinese population. PLoS One. 2015;10(1) doi: 10.1371/journal.pone.0116110.
  33. Kahn BB, Flier JS. Obesity and insulin resistance. J Clin Invest. 2000;106: 473–481. 10.1172/JCI10842
  34. Gelli C, Tarocchi M, Abenavoli L, Di Renzo L, Galli A, De Lorenzo A.Effect of a counseling-supported treatment with the Mediterranean diet and physical activity on the severity of the non-alcoholic fatty liver disease.World J Gastroenterol. 2017;23(17):3150-3162.
  35. Al-Saeed A. Comments on: Serum nesfatin-1 and galanin concentrations in the adult with metabolic syndrome. Relationships to insulin resistance and obesity.Saudi Med J. 2019;40(6):630-631.


Table 1 Comparison of baseline characteristics between the non-alcoholic fatty liver disease (NAFLD) and non-NAFLD groups





P -value


























Diabetic course(year)





HbA1C (%)













































BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting blood glucose; PBG, blood glucose 2 hours after meal; HOMA-IR, homeostatic model assessment for insulin resistance; TG, triglycerides; TC, Total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TYG, a product of triglyceride and fasting glucose

Table 2 Odds ratios and adjusted odds ratios for NAFLD in quartiles of each parameter


Unadjusted OR (95% CI)


Adjusted OR (95% CI)


TyG index








1.738 (1.071–2.822)


1.992 (1.142–3.475)



3.08 (1.859–5.108)


2.942 (1.677–5.163)

















1.843 (1.096–3.098)



4.754 (2.802–8.067)





6.535 (3.704–11.529)


4.868 (2.576–9.200)












2.951 (1.732-5.026)










1.566 (0.972–2.522)


1.356 (0.788–2.333)



2.288 (1.393–3.760)


1.882 (1.079–3.281)



4.171 (2.435–7.147)




TyG index, a product of triglyceride and fasting glucose; TyG-BMI, TyG index × BMI; HOMA-IR, homeostatic model assessment for insulin resistance; (μU/dL) × fasting blood glucose (mg/dL)/22.5;

Table 3 Areas under the receiver-operating characteristic curves for each parameter for predicting non-alcoholic fatty liver disease


Area under the curve

95% CI



Cutoff value



All subjects








TyG index








































TyG index








































TyG index
































CI = confidence interval SE = standard error.