Due to changes in diet and lifestyle, an aging population, and the increasing prevalence of diabetes and obesity, non-alcoholic fatty liver disease (NAFLD) has become the most common chronic liver disease worldwide (1, 2). Among these, metabolic dysfunction-associated steatotic liver disease (MASLD) has emerged as a new disease subtype, emphasizing the central role of metabolic abnormalities in liver damage (3). The latest epidemiological data show that the prevalence of MASLD in urban populations in China is as high as 28.77%, with approximately 16.87% of patients already exhibiting significant liver fibrosis (4). With the global prevalence of obesity and type 2 diabetes (T2DM), the risk of MASLD patients progressing to liver fibrosis, cirrhosis, and hepatocellular carcinoma has significantly increased. However, the lack of early screening methods means that most patients are diagnosed at an irreversible stage (5). Currently, liver biopsy is the gold standard for diagnosis, but due to its invasiveness, sampling errors, and procedural risks, it is difficult to widely apply in clinical practice (6). Imaging studies (such as Fibrotouch, CT, or MRI) are non-invasive but costly (7). Therefore, there is an urgent need to develop simple, cost-effective serological predictive markers.

Insulin resistance (IR) is considered the core mechanism underlying the development of MASLD (8, 9). Basic research indicates that IR promotes the influx of free fatty acids into the liver, increases lipid synthesis, and inhibits β-oxidation, leading to excessive lipid accumulation in hepatocytes (10, 11). Additionally, the chronic low-grade inflammatory state induced by IR further accelerates the activation of hepatic stellate cells and promotes fibrosis progression (12). The triglyceride-glucose index (TyG index) has been proven to be a surrogate marker for assessing IR and is closely associated with the development of MASLD (13–15). Compared with other IR surrogate markers, the TyG index exhibits high sensitivity and specificity, is easy to calculate, and has extremely low cost constraints (16). However, the traditional TyG index does not incorporate obesity as a key metabolic factor. Obesity not only exacerbates IR to promote hepatic steatosis but can also independently induce inflammation and fibrosis (17). In recent years, the TyG-BMI index (combining TyG with body mass index) has demonstrated superiority in predicting cardiovascular disease, but its value in stratifying the severity of MASLD and liver fibrosis remains controversial (18). Especially in different degrees of hepatic steatosis and fibrosis stages, there is a lack of systematic comparison of the predictive efficacy between the TyG index and the TyG-BMI index.

This study aims to address the following key questions: first, are the TyG index and TyG-BMI index independently associated with the severity of MASLD and liver fibrosis? Second, which of the two indices has superior predictive performance for hepatic steatosis and fibrosis? Additionally, is there a dose-response relationship, meaning that higher TyG/TyG-BMI levels are associated with a higher risk of liver disease?

Through a cross-sectional retrospective analysis of a large clinical dataset, this study comprehensively compared the roles of the TyG index and TyG-BMI index in MASLD risk stratification and liver fibrosis, providing more accurate and cost-effective metabolic indicator options for non-invasive screening. This helps identify high-risk populations, optimize the timing of clinical interventions, and offers new insights into the pathogenesis of MASLD.

Definition of MASLD: (1) Imaging diagnosis of fatty liver: CAP ≥244 dB/m. (2) No other causes of hepatic steatosis or excessive alcohol consumption (≥30 g/d for males, ≥20 g/d for females). (3) At least one metabolic cardiovascular disease risk factor: ① BMI ≥25 kg/m² (23 kg/m² for Asia) or waist circumference (WC) >94 cm for males, >80 cm for females; ② Fasting blood glucose (FBG) ≥5.6 mmol/L (100 mg/dl) or 2-h post-load glucose levels ≥7.8 mmol/L (140 mg/dl) or glycated hemoglobin (HbA1c) ≥5.7% (39 mmol/L) or type 2 diabetes or treatment for type 2 diabetes; ③ Blood pressure ≥130/85 mmHg or specific antihypertensive drug treatment; ④ Plasma triglycerides (TG) ≥1.70 mmol/L (150 mg/dl) or lipid lowering treatment; ⑤ Plasma high-density lipoprotein cholesterol (HDL-C) ≤ 1.0 mmol/L (40 mg/dl) for males or ≤ 1.3 mmol/L (50 mg/dl) for females, or lipid lowering treatment (19). The study participants were divided into 3,450 non-MASLD cases and 3,585 MASLD cases.

The classification of MASLD based on the CAP value of Fibrotouch is as follows: a CAP value ≥ 244 dB/m is defined as hepatic steatosis; a CAP value between 244 and < 269 dB/m is defined as mild MASLD; a CAP value between 269 and < 296 dB/m is defined as moderate MASLD; and a CAP value ≥ 296 dB/m is defined as severe MASLD.

Based on the LSM value, MASLD patients are grouped according to the presence of liver fibrosis: LSM ≥ 7.3 kPa is classified as the liver fibrosis group, and LSM < 7.3 kPa is classified as the non-liver fibrosis group.

Data analysis was performed using statistical software SPSS 26.0 (IBM Corp, Armonk, NY, USA) and R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). Continuous variables are expressed as mean and standard deviation. t-tests or analysis of variance (ANOVA) were used to compare means between two or three groups. Categorical data are presented as counts and percentages, and intergroup comparisons were performed using Pearson's chi-square test. Correlation analyses were conducted between CAP and TyG, TyG-BMI, and their components. The intergroup differences between CAP and the quartiles of TyG and TyG-BMI, as well as between TyG, TyG-BMI, and different degrees of hepatic steatosis, were analyzed by Tukey's multiple comparisons test. Logistic regression analysis was used to estimate the correlation between the TyG index, TyG-BMI index, and MASLD and liver fibrosis. Variables with p < 0.1 in univariate logistic regression were included in multivariate logistic regression to calculate odds ratios (OR) and 95% confidence intervals (CI). To address multicollinearity, variance inflation factors (VIF) were calculated for all variables in the initial full model. VIF > 10 was considered indicative of severe multicollinearity. Due to high multicollinearity between TC and LDL-C (VIF > 15), both TC and LDL-C were excluded. Similarly, since the TyG-BMI index is mathematically derived from TyG and BMI, BMI was excluded from models containing TyG or TyG-BMI to obtain stable estimates. After these exclusions, all variables in the final model had VIF values below 10, indicating acceptable levels of multicollinearity. Three models were established: model I (unadjusted), Model II (adjusted for age, gender, type 2 diabetes, coronary heart disease, and hypertension), and Model III (adjusted for age, gender, type 2 diabetes, coronary heart disease, hypertension, ALT, AST, TBIL, DBIL, IBIL, ALP, CR, UA, UREA, TG, HDL-C, and FBG). The predictive ability of TyG and TyG-BMI indices for MASLD and liver fibrosis was estimated using the area under the receiver operating characteristic curve (AUROC). A two-tailed value of p < 0.05 was considered to indicate statistical significance.

In recent years, with the rapid development of China's economy and changes in lifestyle, the incidence of MASLD has shown a significant upward trend. This disease is characterized by high genetic susceptibility, an increasing proportion of young patients, insufficient public awareness, and limited clinical diagnostic and treatment options (20). Notably, MASLD not only leads to end-stage liver diseases such as decompensated cirrhosis and hepatocellular carcinoma but also increases the risk of type 2 diabetes, cardiovascular disease, chronic kidney disease, and non-hepatic malignant tumors, imposing a significant burden on global public health systems (21–23). Due to the lack of obvious symptoms and slow progression in the early stages of MASLD, patients often underestimate its severity. However, clinical observations have revealed that many patients diagnosed incidentally with hepatic steatosis actually already exhibit significant risks of liver fibrosis (24). This situation underscores the importance and urgency of developing accurate, convenient, and non-invasive diagnostic tools.

Early risk assessment and intervention in the management of MASLD not only prevent liver-related complications but also reduce the risk of metabolic abnormalities such as type 2 diabetes and cardiovascular disease. Among numerous biomarkers, the TyG index has garnered significant attention as a reliable surrogate marker for insulin resistance (13, 25). However, current research on the correlation between the TyG index and the progression of liver fibrosis and the severity of steatosis in MASLD remains insufficient, particularly regarding the predictive value of the TyG index for the risk of liver fibrosis in MASLD patients. Therefore, further investigating the relationship between TyG and its derived indicator (TyG-BMI) with liver fat content and fibrosis severity, and evaluating their predictive efficacy for MASLD, will provide new theoretical basis and clinical strategies for the early screening and intervention of this disease.

This study systematically compared the predictive value of the TyG index and TyG-BMI index for different severities of MASLD and liver fibrosis. The results showed that both TyG and TyG-BMI were independently associated with the severity of MASLD, but TyG-BMI demonstrated superior predictive performance (AUC 0.908 vs. 0.774). After adjusting for multiple confounding factors, only TyG-BMI remained significantly associated with liver fibrosis. Both indices exhibited a clear dose-response relationship, with the highest quartile of TyG-BMI associated with a 239-fold increased risk of hepatic steatosis. Additionally, potential cutoff values for clinical risk stratification were identified. These findings provide new evidence for understanding the metabolic drivers of MASLD and reveal the synergistic role of obesity and insulin resistance in liver disease progression.

This study confirmed that the TyG index is significantly associated with the severity of MASLD, a finding that is highly consistent with the results of Zhang et al.'s (26) large-scale cohort study. Their study showed that for every 1-unit increase in the TyG index, the risk of MASLD increased by 78.4% (95% CI = 1.383–2.302, p < 0.001), while in this study, the risk in the highest quartile of the TyG index increased to 4.48 times. This difference may be attributed to the stricter liver imaging assessment criteria used in this study. Furthermore, our study further refined the dose-response relationship between the TyG index and different degrees of fatty degeneration (mild, moderate, and severe), which has been rarely reported in previous studies. Additionally, the excellent predictive performance of the TyG-BMI index is consistent with recent studies, indicating that composite metabolic indicators are superior to single biomarkers (27–30). Our results are consistent with those of Kim et al. (31), who reported that the TyG-BMI index achieved an AUC of 0.862–0.872 and a sensitivity of 86.3% in predicting MASLD in the general population. From a mechanistic perspective, this advantage may stem from TyG-BMI being a combination of TyG and BMI levels, thereby reflecting both insulin resistance and obesity severity. Obesity not only exacerbates hepatic lipid accumulation through insulin resistance-dependent pathways (increased free fatty acid influx) but also promotes disease progression via insulin resistance-independent mechanisms (e.g., dysregulation of adipokines) (32). Notably, the risk of severe steatosis in the TyG-BMI Q4 group (OR = 239.41) was significantly higher than that in the TyG Q4 group (OR = 4.48), suggesting that BMI may amplify the metabolic damage caused by insulin resistance. This finding supports the “multiple hit” hypothesis for the onset of MASLD (8).

Through restricted cubic spline analysis, we identified a significant “J-shaped” non-linear association between TyG-BMI and MASLD risk. The inflection point occurred at approximately TyG-BMI 213, with risk surging sharply beyond this threshold. This confirms that the ultra-high quartile OR value reflects a genuine dose-response inflection point, suggesting TyG-BMI > 213 serves as a critical warning threshold for metabolic decompensation and steeply increasing risk. Furthermore, subgroup analyses revealed significant population differences. TyG-BMI demonstrated the strongest predictive power for MASLD in younger (< 50 years) and non-diabetic populations, highlighting its exceptional value for early screening of metabolic diseases, particularly in contemporary populations experiencing increasingly younger onset. Furthermore, its association with liver fibrosis was stronger in males. These findings suggest that future clinical applications may require establishing refined risk assessment criteria based on demographic characteristics.

Our study found significant differences between the two indices in predicting liver fibrosis: the TyG index lost its significance after adjusting for multiple confounding factors (OR = 0.903, p = 0.504), while TyG-BMI remained strongly correlated (OR = 1.029, p < 0.001). This difference may reflect that obesity-related pathways (such as chronic inflammation and gut microbiota dysbiosis) may independently drive fibrosis beyond insulin resistance. Visceral fat-released pro-fibrotic cytokines (TGF-β, IL-17) can directly activate hepatic stellate cells (33). The association between TyG and liver fibrosis weakened after adjusting for variables such as ALT and AST, suggesting that the correlation may be mediated by hepatocyte damage rather than insulin resistance alone. Interestingly, these findings contrast with those from cardiovascular studies, which show that TyG alone can robustly predict cardiovascular disease outcomes (34–36), suggesting that different organs have distinct mechanisms of response to metabolic abnormalities.

The TyG index primarily reflects insulin resistance in peripheral tissues, specifically the imbalance in glucose and free fatty acid metabolism. However, the TyG index alone may inadequately assess the dysfunction of adipose tissue itself, which is a key driver of metabolic inflammation and liver injury (37, 38). The TyG-BMI index quantifies overall obesity burden by integrating BMI. This integration is mechanistically crucial. First, enlarged adipose tissue—particularly visceral fat—serves as a primary source of pro-inflammatory cytokines (e.g., TNF-α, IL-6) and lipotoxic substances (e.g., ceramides). These factors directly impact the liver via the portal venous system, inducing and amplifying hepatic inflammation and fibrosis (39). Second, obesity itself induces adipocyte hypoxia, endoplasmic reticulum stress, and macrophage infiltration, establishing a chronic low-grade inflammatory state independent of classical peripheral insulin resistance pathways (40). Therefore, TyG-BMI may simultaneously quantify the dual hepatic impact of metabolic dysfunction (TyG) and inflammatory burden (BMI), potentially representing the core mechanism underlying its superior predictive value for liver fibrosis compared to TyG alone.

To comprehensively evaluate the clinical value of TyG-BMI, we compared it with other commonly used non-invasive indicators. While HOMA-IR, the “gold standard” for insulin resistance, is more direct, it relies on insulin testing, which is costly and exhibits significant inter-laboratory variability, making it unsuitable for large-scale screening (16). In contrast, TyG and TyG-BMI require only routine lipid and glucose data, offering greater universality. Compared to other composite formulas like the Fatty Liver Index (FLI) and Visceral Adipose Index (VAI), TyG-BMI demonstrates distinct advantages. A recent comparative study revealed that in predicting MAFLD in the Chinese population, TyG-BMI (AUC = 0.903) showed significantly superior discriminatory performance compared to FLI (AUC = 0.879) and VAI (AUC = 0.773) (41). This may be because FLI focuses more on hepatic cell injury and central obesity, while VAI primarily estimates visceral fat function, and TyG-BMI more directly integrates the core metabolic axis driving fatty liver—insulin resistance and overall obesity.

TyG-BMI can serve as a first-line screening tool for high-risk populations for MASLD (e.g., central obesity, diabetes), particularly in resource-limited settings such as community hospitals. In this study, TyG-BMI demonstrated predictive ability for fibrosis (AUC = 0.697) comparable to that reported for the liver fibrosis index (FIB-4) by Sun et al. (42), but TyG-BMI is simpler to calculate and thus more suitable for primary care screening. It is important to note that while TyG-BMI has some predictive ability for liver fibrosis in MASLD patients, its AUC remains lower than that of elastography (AUC ≈ 0.85). Therefore, positive cases require further imaging confirmation (43).

This study had a large sample size and employed standardized data collection methods to measure and evaluate participants' various indicators, while controlling for most potential confounding factors (including liver function indicators and multiple metabolic parameters), which enhances the reliability of our findings. However, the limitations of our study should not be overlooked. First, the cross-sectional design cannot establish causal relationships, and prospective studies are needed to validate the predictive value of TyG-BMI for fibrosis progression. Second, the absence of histological confirmation may underestimate the prevalence of liver fibrosis. Furthermore, unaccounted confounding factors (such as diet and exercise) may influence the results. Finally, the study population originated from a single center in China, potentially limiting the generalizability of our findings to other geographic regions and ethnic groups worldwide. Metabolic characteristics, body composition, and the prevalence of MASLD drivers may vary significantly across different regions and ethnic populations. Additionally, we employed FibroTouch as the primary assessment tool for hepatic steatosis and fibrosis. Although this technology is widely used and validated in China, its diagnostic thresholds—particularly for staging liver fibrosis—may be influenced by ethnicity, etiology, and device model. Future longitudinal studies are needed to validate the predictive value of TyG-BMI for liver disease progression. If feasible, comparisons of the predictive efficacy of TyG-BMI vs. emerging biomarkers (such as FIB-4, PRO-C3) across different racial populations should be conducted, and the potential of TyG-BMI-guided interventions (such as bariatric surgery) to improve liver disease outcomes should be explored.

The TyG-BMI index is a strong predictor of the severity of MASLD and liver fibrosis, with predictive performance superior to that of the traditional TyG index, making it an effective tool for clinical screening of high-risk populations for MASLD. As an indicator combining superiority and clinical practicality, TyG-BMI can effectively stratify MASLD risk while reflecting insulin resistance and obesity-related metabolic abnormalities. Its strong correlation with liver fibrosis underscores the importance of addressing obesity in MASLD management. Although further validation is needed, TyG-BMI holds promise for revolutionizing the screening paradigm for metabolic liver diseases.