Abstract
This study aimed to investigate the relationship between the triglyceride glucose-body mass index (TyG-BMI) and metabolic dysfunction-associated steatotic liver disease (MASLD) in Korean pregnant women. A secondary analysis was conducted on 585 singleton pregnancies using publicly available data from South Korea. The TyG-BMI was examined as a continuous variable and by quartiles. The association between the TyG-BMI and MASLD was evaluated using logistic regression, restricted cubic spline (RCS) regression, sensitivity analyses, and subgroup analyses. The diagnostic performance of the TyG-BMI for MASLD was further assessed using receiver operating characteristic (ROC) curve analysis. The mean maternal age was 32.1 ± 3.8 years, and MASLD was present in 110 (18.8%) women. After adjusting for potential confounders, TyG-BMI was significantly and positively associated with MASLD (adjusted odds ratio per 10-unit increase = 1.31, 95% confidence interval (CI): 1.21–1.43, P < 0.001). Sensitivity and subgroup analyses supported the robustness of this association. RCS analysis indicated a linear relationship between TyG-BMI and MASLD risk. The area under the ROC curve was 0.7483 (95% CI 0.6953–0.8012), suggesting a good diagnostic accuracy. An optimal TyG-BMI cut-off value of 190.99 was identified, with a sensitivity of 63.64% and a specificity of 74.32%.
Subject terms: Diseases, Gastrointestinal diseases
Introduction
Metabolic dysfunction-associated steatotic liver disease (MASLD) is a prevalent health issue on a global scale, affecting approximately one-quarter of the world’s population1. Its prevalence is higher among individuals with sedentary lifestyles and unhealthy diets2. If left untreated, MASLD can potentially advance to more severe conditions, such as cirrhosis and liver cancer, which may result in significant health risks and socioeconomic burdens3–5. Recent studies have highlighted the correlation between MASLD and unfavorable pregnancy outcomes, including gestational diabetes, preterm labor, and fetal overgrowth6,7. This highlights the necessity for early detection in pregnant women8.
However, liver biopsy, the gold standard for diagnosing MASLD, is invasive and considered inappropriate for pregnant women because of the potential risks involved9. Therefore, it is imperative to develop non-invasive, easily accessible, and highly diagnostic serological markers for MASLD screening in this population8.
The triglyceride glucose (TyG) index, which serves as a marker of insulin resistance (IR)10–12, has been extensively employed for the assessment of metabolic health. The triglyceride glucose-body mass index (TyG-BMI), which integrates the TyG index with body mass index (BMI), offers a more comprehensive indicator of metabolic status and has been associated with IR and MASLD13,14. Despite its potential clinical utility, the relationship between TyG-BMI and MASLD in pregnant women has not been explored in any published research thus far. To the best of our knowledge, this study is the first to demonstrate an association between TyG-BMI and MASLD during pregnancy, highlighting its potential as a readily available biomarker for risk stratification and longitudinal monitoring in this vulnerable population.
Methods
Study design and data source
This study was a cross-sectional analysis based on publicly available data. The original dataset, developed by Lee et al. in Korea, is accessible via PLOS ONE (https://journals.plos.org/plosone)15. The initial study was approved by the Institutional Review Board of the Seoul National University Medical Center and the Public Institutional Review Board of the Ministry of Health and Welfare of South Korea. As this study involved a secondary analysis of anonymized data, no additional ethical approval was required. All methods were performed in accordance with the relevant guidelines and regulations.
Participants
Lee et al. conducted an original study (NCT 02276144)16 to investigate whether MASLD is a risk factor for adverse pregnancy outcomes. Singleton pregnant women were recruited before 14 weeks of gestation. The exclusion criteria were as follows: (1) history of alcohol consumption, (2) history of diabetes mellitus, (3) history of chronic liver disease (CLD), (4) loss to follow-up during pregnancies, and (5) preterm delivery before 34 weeks of gestation. A total of 623 singleton pregnancies were enrolled in the study. For the present analysis, 38 women were excluded due to missing exposure, outcome, or covariate variables, resulting in a final sample of 585 participants (Fig. 1). All participants provided written informed consent.
Fig. 1.
Flow chart of study participants. TG triglyceride, FPG fasting plasma glucose, BMI body mass index, GDM gestational diabetes mellitus, AST aspartate aminotransferase, ALT alanine aminotransferase, MASLD metabolic dysfunction-associated steatotic liver disease.
Data collection
The participating pregnant women provided blood samples between 10 and 14 weeks of gestation after fasting for more than 8 h. Laboratory assessments included alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma-glutamyl transferase (GGT), total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL), low-density lipoprotein cholesterol (LDL), fasting plasma glucose (FPG), and insulin levels. Trained healthcare professionals collected data on pre-pregnancy BMI, age, reproductive history, alcohol consumption, and the medical history of the participants. All participants underwent ultrasound examination to assess for MASLD. Gestational diabetes mellitus (GDM) was evaluated between 24 and 28 weeks of gestation using an oral glucose tolerance test (OGTT) and diagnosed according to the criteria of the International Association of Diabetes and Pregnancy Study Groups (IADPSG). GDM was confirmed if any of the following thresholds were met: fasting plasma glucose ≥ 5.1 mmol/L, 1-h post-load plasma glucose ≥ 10.0 mmol/L, or 2-h post-load plasma glucose ≥ 8.5 mmol/L17,18.
Definitions
Definition of triglyceride glucose-body mass index: TyG = Ln [TG (mg/dL) * FPG (mg/dL)/2], TyG -BMI = TyG * BMI19.
Definition of MASLD: The presence of MASLD was determined by ultrasound and diagnosed by the presence of two or more of the following: (1) diffuse echogenic enhancement; (2) liver echoes that are stronger than those of the renal parenchyma; (3) blurred and narrowed hepatic venous lumen20,21.
Statistical analysis
Participating pregnant women were stratified into quartiles according to the TyG-BMI index: Q1 (< 161.2, n = 146), Q2 (161.2–177.6, n = 146), Q3 (177.7–199.2, n = 146), and Q4 (> 199.2, n = 147). Categorical variables were presented as numbers and percentages and analyzed using the chi-square test. Continuous variables with an approximately normal distribution were expressed as means and standard deviations and were compared using analysis of variance. Skewed continuous variables were reported as medians with interquartile ranges and were compared using the Kruskal–Wallis test. Four models were developed using univariate and multivariate logistic regression analyses. Covariates (parity, age, ALT, AST, GGT, HDL, LDL, insulin, and GDM) were retained if they were associated with MASLD at P < 0.05 in the univariate analysis, if their inclusion changed the exposure–outcome association by > 10%, or if they were supported by prior evidence and clinical relevance. Model 1 was unadjusted. Model 2 was adjusted for sociodemographic characteristics, including parity and age. Model 3 was further adjusted for ALT, AST, GGT, HDL, and LDL. Model 4 was additionally adjusted for insulin and GDM. The risk of MASLD was estimated using adjusted odds ratios (ORs) with 95% confidence intervals (CIs). TyG-BMI was analyzed as both a continuous and categorical variable based on quartiles, and P-values for trends were calculated. Restricted cubic spline (RCS) curve fitting with full adjustment was applied to examine the linear relationship between TyG-BMI and MASLD. To assess the robustness of the findings, a sensitivity analysis was conducted after excluding participants with GDM. Stratified logistic regression analyses were then performed according to age, parity, GDM status, and homeostasis model assessment of insulin resistance (HOMA-IR). Lastly, receiver operating characteristic (ROC) curve analysis was used to evaluate the ability of TyG-BMI to discriminate MASLD and to determine the optimal cut-off value.
We used the Free Statistical Analysis Platform (version 2.0, http://www.clinicalscientists.cn/freestatistics, Beijing, China) and the R statistical program (version 4.2.2, http://www.R-project.org, The R Foundation) for statistical analyses. Statistical significance was set at P < 0.05, and the results are expressed as ORs and 95% CIs.
Results
Baseline characteristics of participating pregnant women
A total of 38 participants were excluded due to missing data, leaving 585 pregnant women for the analysis. Table 1 presents the baseline characteristics of the participants, stratified into four groups according to the TyG-BMI. The mean age was 32.1 ± 3.8 years, and 110 (18.8%) women were diagnosed with MASLD. Across the TyG-BMI quartiles, MASLD was present in 11 women (7.5%) in Q1, 15 (10.3%) in Q2, 26 (17.8%) in Q3, and 58 (39.5%) in Q4, with a significantly higher incidence in Q4 than in the other quartiles. Analysis of Table 1 further indicates that the Q4 group exhibited higher BMI, ALT, GGT, TC, TG, LDL, FPG, insulin, and insulin resistance values, a higher GDM prevalence, and lower HDL levels.
Table 1.
Baseline characteristics of participants.
| Characteristics | Total | TyG-BMI | P value | |||
|---|---|---|---|---|---|---|
| Q1 (< 161.2) n = 146 |
Q2 (161.2-177.6) n = 146 |
Q3 (177.7-199.2) n = 146 |
Q4(> 199.2) n = 147 |
|||
| Age (year) | 32.1 ± 3.8 | 31.7 ± 3.7 | 31.9 ± 3.7 | 32.1 ± 3.3 | 32.6 ± 4.2 | 0.188 |
| Parity | 0.039 | |||||
| No | 307 (52.5) | 82 (56.2) | 82 (56.2) | 81 (55.5) | 62 (42.2) | |
| Yes | 278 (47.5) | 64 (43.8) | 64 (43.8) | 65 (44.5) | 85 (57.8) | |
| BMI (kg/m2) | 22.0 ± 3.5 | 18.6 ± 1.3 | 20.4 ± 1.0 | 22.5 ± 1.1 | 26.5 ± 3.2 | < 0.001 |
| BMI ≥ 25 (kg/m2) | 95 (16.2) | 0 (0) | 0 (0) | 2 (1.4) | 93 (63.3) | < 0.001 |
| AST(IU/L) | 16.0 (14.0, 20.0) | 16.0 (14.0, 18.0) | 16.0 (14.0, 19.0) | 17.0 (14.0, 20.0) | 16.0 (14.0, 20.5) | 0.632 |
| ALT(IU/L) | 11.0 (8.0, 15.0) | 10.0 (8.0, 13.0) | 10.0 (8.0, 14.0) | 12.0 (8.2, 16.0) | 13.0 (9.0, 18.0) | < 0.001 |
| GGT (IU/L) | 12.0 (10.0, 15.0) | 11.0 (10.0, 14.0) | 11.0 (9.0, 13.0) | 12.0 (10.0, 14.8) | 14.0 (11.0, 21.0) | < 0.001 |
| TC (mg/dL) | 172.8 ± 27.1 | 163.8 ± 24.0 | 166.9 ± 23.7 | 176.7 ± 27.2 | 183.9 ± 28.6 | < 0.001 |
| TG (mg/dL) | 110.0 (87.0, 139.0) | 87.5 (73.0, 103.0) | 105.0 (85.0, 128.0) | 114.0 (90.0, 146.0) | 142.0 (113.0, 182.0) | < 0.001 |
| HDL (mg/dL) | 65.0 ± 13.5 | 67.5 ± 13.3 | 64.7 ± 12.9 | 64.9 ± 14.2 | 62.8 ± 13.4 | 0.033 |
| LDL (mg/dL) | 84.0 ± 21.7 | 78.4 ± 20.0 | 79.9 ± 18.7 | 87.2 ± 20.6 | 90.4 ± 24.9 | < 0.001 |
| FPG (mg/dL) | 77.0 ± 9.7 | 74.3 ± 9.1 | 76.1 ± 7.9 | 77.1 ± 8.1 | 80.4 ± 12.2 | < 0.001 |
| Insulin (µIU/mL) | 8.4 (5.4, 11.5) | 6.2 (3.7, 8.8) | 7.5 (5.0, 10.4) | 9.0 (5.6, 11.9) | 11.4 (8.1, 17.9) | < 0.001 |
| HOMA-IR | 1.5 (1.0, 2.3) | 1.1 (0.7, 1.7) | 1.4 (0.9, 2.1) | 1.7 (1.0, 2.4) | 2.3 (1.5, 3.7) | < 0.001 |
| GDM | < 0.001 | |||||
| No | 549 (93.8) | 144 (98.6) | 142 (97.3) | 140 (95.9) | 123 (83.7) | |
| Yes | 36 (6.2) | 2 (1.4) | 4 (2.7) | 6 (4.1) | 24 (16.3) | |
| MASLD | < 0.001 | |||||
| No | 475 (81.2) | 135 (92.5) | 131 (89.7) | 120 (82.2) | 89 (60.5) | |
| Yes | 110 (18.8) | 11 (7.5) | 15 (10.3) | 26 (17.8) | 58 (39.5) | |
| TyG | 8.3 (8.1, 8.6) | 8.1 (7.9, 8.3) | 8.3 (8.0, 8.5) | 8.4 (8.2, 8.6) | 8.6 (8.4, 8.9) | < 0.001 |
| HSI | 29.4 (27.0, 33.1) | 26.2 (25.0, 27.7) | 27.8 (26.6, 29.5) | 30.5 (29.0, 32.9) | 34.9 (32.1, 37.7) | < 0.001 |
| FLI | 11.5 (6.0, 22.7) | 4.5 (3.4, 6.3) | 8.4 (6.1, 11.5) | 15.5 (11.4, 20.4) | 35.2 (24.2, 53.6) | < 0.001 |
| AIP | 0.2 (0.1, 0.4) | 0.1 (0.0, 0.2) | 0.2 (0.1, 0.3) | 0.2 (0.1, 0.4) | 0.4 (0.2, 0.5) | < 0.001 |
Values were expressed as mean (standard deviation) or median (interquartile range) or n (%). TyG-BMI triglyceride glucose-body mass index, BMI body mass index, AST aspartate aminotransferase, ALT alanine aminotransferase, GGT gamma-glutamyl transferase, TC total cholesterol, TG triglyceride, HDL high-density lipoprotein cholesterol, LDL low-density lipoprotein cholesterol, FPG fasting plasma glucose, HOMA-IR homeostasis model assessment-insulin resistance, GDM gestational diabetes mellitus, MASLD metabolic dysfunction-associated steatotic liver disease, TyG triglyceride glucose, HSI hepatic steatosis index, FLI fatty liver index, AIP Atherogenic index of plasma.
Relationship between TyG-BMI and MASLD
Figure 2 illustrates that the restricted cubic spline model maintained a linear relationship between TyG-BMI and MASLD after adjusting for all covariates. Univariate analysis results (Table 2) indicated that the risk of MASLD was positively associated with BMI, ALT, GGT, TG, FPG, insulin, HOMA-IR, GDM, and TyG-BMI, whereas HDL demonstrated a negative association.
Fig. 2.
Association between TyG-BMI and MASLD risk in RCS. Adjusted for age, parity, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, insulin and gestational diabetes mellitus.
Table 2.
The unadjusted association between baseline variables and MASLD (n = 585). OR odds ratio, CI confidence interval, other abbreviations as in Table 1.
| Variable | OR (95%CI) | P value |
|---|---|---|
| Age (year) | 0.96 (0.91 ~ 1.01) | 0.141 |
| Parity | 0.99 (0.65 ~ 1.5) | 0.954 |
| BMI (kg/m2) | 1.3 (1.22 ~ 1.39) | < 0.001 |
| AST(IU/L) | 1.02 (1 ~ 1.04) | 0.111 |
| ALT(IU/L) | 1.04 (1.02 ~ 1.06) | < 0.001 |
| GGT(IU/L) | 1.03 (1.01 ~ 1.05) | 0.008 |
| TC (mg/dL) | 1.01 (1 ~ 1.01) | 0.175 |
| TG (mg/dL) | 1.01 (1.01 ~ 1.01) | < 0.001 |
| HDL (mg/dL) | 0.97 (0.96 ~ 0.99) | 0.001 |
| LDL (mg/dL) | 1.01 (1 ~ 1.02) | 0.107 |
| FPG (mg/dL) | 1.03 (1.01 ~ 1.05) | 0.01 |
| Insulin (µIU/mL) | 1.09 (1.06 ~ 1.13) | < 0.001 |
| HOMA-IR | 1.38 (1.19 ~ 1.6) | < 0.001 |
| GDM | 6.37 (3.18 ~ 12.78) | < 0.001 |
| TyG-BMI | 1.03 (1.02 ~ 1.04) | < 0.001 |
Multivariate logistic regression analyses (Table 3) confirmed a positive association between TyG-BMI and MASLD risk after adjusting for covariates. Specifically, Model 1 presented an OR of 1.34 (95% CI: 1.25–1.43), Model 2 an OR of 1.35 (95% CI: 1.26–1.45), and Model 3 reported an OR of 1.35 (95% CI: 1.25–1.46). Each ten-unit increase in TyG-BMI was associated with a 31% increased risk of MASLD, accounting for various influencing factors. The correlation remained significant when TyG-BMI was assessed as a categorical variable (trend P < 0.001).
Table 3.
Relationship between TyG-BMI and MASLD in in Korean Pregnant Women.
| Variable | Model 1 | Model 2 | Model 3 | Model 4 | ||||
|---|---|---|---|---|---|---|---|---|
| OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | |
| TyG-BMI*0.1 | 1.34 (1.25 ~ 1.4) | < 0.001 | 1.35 (1.26 ~ 1.45) | < 0.001 | 1.35 (1.25 ~ 1.46) | < 0.001 | 1.31 (1.21 ~ 1.43) | < 0.001 |
| TyG-BMI | ||||||||
| Q1 (< 161.2) | 1(Ref) | 1(Ref) | 1(Ref) | 1(Ref) | ||||
| Q2 (161.2-177.6) | 1.41 (0.62 ~ 3.17) | 0.413 | 1.43 (0.63 ~ 3.23) | 0.393 | 1.28 (0.56 ~ 2.93) | 0.551 | 1.23 (0.54 ~ 2.82) | 0.625 |
| Q3 (177.7-199.2) | 2.66 (1.26 ~ 5.61) | 0.01 | 2.75 (1.3 ~ 5.82) | 0.008 | 2.37 (1.1 ~ 5.11) | 0.028 | 2.2 (1.01 ~ 4.77) | 0.047 |
| Q4 (> 199.2) | 8 (3.98 ~ 16.07) | < 0.001 | 8.75 (4.31 ~ 17.76) | < 0.001 | 7.23 (3.44 ~ 15.2) | < 0.001 | 5.43 (2.48 ~ 11.86) | < 0.001 |
| P for Trend | < 0.001 | < 0.001 | < 0.001 | < 0.001 | ||||
Model 1: unadjusted; Model 2: adjusted for age, parity. Model 3: adjusted for age, parity, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol. Model 4: adjusted for age, parity, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, insulin and gestational diabetes mellitus. Ref reference.
Sensitivity analyses and subgroup analyses
Given the established association between MASLD and insulin resistance, two sensitivity analyses were performed. Initially, after excluding pregnant women with HOMA-IR ≥ 2, TyG-BMI continued to demonstrate a positive association with MASLD risk after adjusting for all covariates (OR: 1.36; 95% CI 1.20–1.55). In the second analysis, excluding women with GDM also yielded a positive association (OR: 1.29; 95% CI: 1.18–1.41) after controlling for confounding factors (Table 4). The relationship between TyG-BMI and MASLD was consistent across all subgroups, remaining robust regardless of age, parity, HOMA-IR, GDM, and other factors (Fig. 3). The stability of these findings reinforces the reliability of our results, suggesting that the association between TyG-BMI and MASLD is not influenced by these potential confounders.
Table 4.
Relationship between TyG-BMI and MASLD in different sensitivity analyses.
| Variable | Model I (OR, 95% CI, P) | Model II (OR, 95% CI, P) |
|---|---|---|
| TyG-BMI*0.1 | 1.36 (1.2 ~ 1.55) < 0.001 | 1.29 (1.18 ~ 1.41) < 0.001 |
| TyG-BMI | ||
| Q1 (< 161.2) | 1 (Ref) | 1 (Ref) |
| Q2 (161.2-177.6) | 1.1 (0.41 ~ 2.94) 0.844 | 1.18 (0.51 ~ 2.74) 0.695 |
| Q3 (177.7-199.2) | 2.47 (0.99 ~ 6.13) 0.052 | 2.25 (1.03 ~ 4.9) 0.042 |
| Q4 (> 199.2) | 5.66 (2.21 ~ 14.49) < 0.001 | 4.74 (2.12 ~ 10.58) < 0.001 |
| P for Trend | < 0.001 | < 0.001 |
Model I was sensitivity analysis after excluding those with HOME-IR ≥ 2, n = 387. Model II was a sensitivity analysis after excluding those with GDM, n = 550.
Fig. 3.
The forest plot for the effect size of TyG-BMI on MASLD in prespecified and exploratory subgroups. Each stratification was adjusted for age, parity, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, insulin, and gestational diabetes mellitus, except the stratification factor itself.
ROC curve analysis
As displayed in Table 5; Fig. 4, TyG-BMI demonstrated good discriminatory ability for MASLD in pregnant women, yielding an area under the curve (AUC) of 0.748 (95% CI: 69.53%– 80.12%). The optimal cut-off value for TyG-BMI was determined to be 190.99, with a sensitivity of 63.64% and a specificity of 74.32%. Comparatively, TyG-BMI exhibited the highest AUC among other indices, including TyG, hepatic steatosis index (HSI), fatty liver index (FLI), and atherogenic index of plasma (AIP), confirming its superior efficacy as a screening test for MASLD.
Table 5.
Areas under the receiver operating characteristic curves of each evaluated parameter in identifying MASLD.
| Variable | AUC (95% CI) | Best threshold | Specificity | Sensitivity | Youden index |
|---|---|---|---|---|---|
| TyG-BMI | 0.7483 (69.53% ~ 80.12%) | 190.99 | 0.7432 | 0.6364 | 0.3796 |
| TyG | 0.6395(57.8% ~ 70.09%) | 8.56 | 0.7684 | 0.5 | 0.2684 |
| HSI | 0.7355 (67.99% ~ 79.1%) | 31.91 | 0.7479 | 0.6545 | 0.4024 |
| FLI | 0.7248 (66.73% ~ 78.23%) | 24.67 | 0.8517 | 0.5182 | 0.3699 |
| AIP | 0.6501 (59.03% ~ 70.99%) | 0.3634 | 0.7881 | 0.4818 | 0.2699 |
AUC Area under the curve.
Fig. 4.
ROC curves for TyG-BMI and MASLD in pregnant women. The results showed that the AUC of the TyG-BMI was 0.748. Compared to TyG, HSI, FLI, and AIP, the AUC of the TyG-BMI for screening MASLD was the highest. TyG triglyceride glucose, HSI hepatic steatosis index, FLI fatty liver index, AIP Atherogenic index of plasma.
Discussion
This study analyzed data from 585 pregnant Korean women and demonstrated that each 10-unit increment in TyG-BMI was associated with a significantly increased risk of MASLD (adjusted OR = 1.31, 95% CI: 1.21–1.43, P < 0.001). Sensitivity and stratified analyses confirmed the robustness of these findings. ROC analysis demonstrated an AUC of 0.748, suggesting a favorable diagnostic performance. To the best of our knowledge, this is the first study to report an association between TyG-BMI and MASLD in this specific population of pregnant women.
A large-scale cross-sectional study involving 14,280 participants undergoing comprehensive health screenings demonstrated that the TyG-BMI serves as a simple, non-invasive surrogate marker capable of accurately identifying MASLD within the general population22. Zhang et al. found that TyG-BMI was significantly associated with MASLD and its hepatic steatosis and fibrosis status14. A comprehensive analysis of data from the National Health and Nutrition Examination Survey (NHANES III) revealed a significant nonlinear dose-response relationship between increasing TyG-BMI levels and all-cause, cardiovascular, and diabetes-related mortality among individuals with MASLD19. Although previous studies have suggested an association between TyG-BMI and MASLD, this relationship has not been clearly characterized in pregnant women, which provides the rationale for the present study.
A decade-long retrospective cohort study spanning 2014–2023 demonstrated that among 51,708 pregnant individuals, the prevalence of CLD increased significantly from 1.6% to 2.3%, corresponding to a 2.5-fold increase during the study period. Among cases of chronic liver disease during pregnancy, the proportion of MASLD rose sharply from 14.1% in 2014 to 57.9% in 2023, becoming the primary cause of chronic liver disease during pregnancy23. According to a recent study, the overall prevalence of MASLD in Korea is approximately 30.3%, with a prevalence of approximately 20.3% among women24. In this study, the prevalence of MASLD among pregnant participants was 18.8%, which is comparatively lower than that reported in the existing literature. This difference is likely due to the characteristics of the study population. First, stringent exclusion criteria were applied, and individuals with pre-existing diabetes mellitus were systematically excluded from the study. In contrast, contemporary evidence indicates that the prevalence of MASLD among individuals with diabetes mellitus ranges from 65% to 68.8%1. Second, the cohort in this study exhibited a relatively low adiposity burden, with only 16.2% of participants exhibiting a BMI ≥ 25 kg/m². In contrast, published data suggest that the prevalence of MASLD among overweight individuals is close to 70%25. Consequently, the leaner risk profile of the study population and exclusion of patients with diabetes mellitus contributed to the comparatively lower prevalence of MASLD observed in this study.
The present study addresses an important knowledge gap by evaluating the diagnostic utility of the TyG-BMI for screening MASLD during pregnancy. Indices such as HSI, FLI, and AIP are used to screen for MASLD in the general population, and FLI and AIP have been examined specifically in pregnant women26,27. Accordingly, ROC curve analyses were conducted to compare the diagnostic performance of TyG-BMI with TyG, HSI, FLI, and AIP in detecting gestational MASLD. TyG-BMI demonstrated a strong discriminative capacity (AUC = 0.74; 95% CI: 69.53%–80.12%), outperforming all comparator indices. These findings suggest that TyG-BMI exhibits greater clinical utility than conventional surrogate markers of insulin resistance. The optimal cut-off, determined using the Youden index, was 190.99, with a sensitivity of 63.6% and a specificity of 74.3%. This corresponds to a false-negative rate of approximately 36%, which exceeds the acceptable threshold for its use as a sole screening tool. Therefore, TyG-BMI alone cannot be recommended for the routine screening of gestational MASLD. This limitation may be overcome by integrating TyG-BMI with additional validated biomarkers and clinical variables. For instance, combined prediction models incorporating TyG-BMI with FLI or HSI, thereby capturing both hepatic enzymatic and lipid-related abnormalities, may enhance the overall diagnostic accuracy in pregnant populations.
MASLD pathogenesis is highly complex and involves multiple, interrelated mechanisms. Key contributors include insulin resistance, unhealthy dietary patterns, oxidative stress due to excess free radical production, and alterations in the intestinal microenvironment28,29. Hepatic triglyceride accumulation plays a central role; reduced expression of diacylglycerol acyltransferase 2 (DGAT2) can promote the generation of hepatotoxic intermediates that contribute to liver injury30. Under insulin resistance conditions, altered signaling via insulin receptor substrate 2 enhances hepatic de novo lipogenesis, while serine kinase–mediated disruption of the insulin signaling pathway promotes an increase in hepatocellular free fatty acids and their efflux31. In parallel, activation of inflammatory pathways, mitochondrial dysfunction, and endoplasmic reticulum stress further drive the progression of MASLD32. Collectively, these processes culminate in hepatic inflammation and fibrosis. During pregnancy, hormonal changes physiologically induce a state of insulin resistance, which may predispose pregnant women to a cascade of metabolic disturbances that facilitate the development or exacerbation of MASLD.
Recent epidemiological data indicate an increasing prevalence of MASLD during pregnancy7. Several studies have demonstrated that MASLD in pregnant women is strongly associated with adverse pregnancy outcomes, including preterm birth, miscarriage, GDM, preeclampsia, and delivery of a macrosomic foetus33,34. Therefore, careful monitoring of MASLD during pregnancy is warranted. Although liver biopsy remains the gold standard for diagnosing MASLD, its invasive nature, cost, and limited feasibility for repeated use render it unsuitable for pregnant women. In contrast, TyG-BMI, a composite index of the TyG index with BMI, provides a more comprehensive indication of metabolic status than TyG alone. It is non-invasive, readily obtainable from routine clinical measurements, and therefore well-suited for evaluation in pregnant populations.
This study has several limitations. First, its cross-sectional design precludes the establishment of a causal relationship between TyG-BMI and MASLD. Second, MASLD was diagnosed using ultrasonography rather than liver biopsy, which is the gold standard but is unsuitable for pregnant women35. Ultrasonography has reduced diagnostic accuracy when hepatic steatosis involves less than 15% of the liver parenchyma9. Third, pregnant women with a history of diabetes mellitus before pregnancy were excluded. Because individuals with diabetes are at a higher risk of MASLD36, this criterion likely led to an underestimation of the prevalence of MASLD and limited the generalizability of the findings to pregnant women with pre-existing diabetes. Fourth, as this was a secondary analysis, BMI in the original dataset referred only to the pre-pregnancy BMI. In future studies, we plan to establish a dedicated cohort and collect both pre-pregnancy and gestational BMI to enable more detailed comparative analyses. Lastly, the study population did not include pregnant women from other racial or ethnic groups. Consequently, future cohort studies should include more diverse populations to assess the external validity and generalizability of these findings.
Conclusion
This study set out, for the first time, to examine the association between TyG-BMI during pregnancy and MASLD. An independent, positive dose–response relationship was observed between TyG-BMI and gestational MASLD, indicating that maintaining TyG-BMI below the identified optimal cut-off value (190.99) may help reduce the risk of MASLD during pregnancy. Accordingly, incorporating TyG-BMI into routine prenatal surveillance could facilitate the early identification of women at an elevated risk. Given the inherent limitations of the cross-sectional design, causal inferences cannot be made. Accordingly, longitudinal multicenter investigations are needed to clarify causality and evaluate the generalizability of these findings across diverse ethnic and geographic populations.
Acknowledgements
In this secondary analysis, the primary sources of data and methodology are derived from the research undertaken by Lee et al.15. We sincerely thank the original investigators for granting permission to reuse the dataset and acknowledge their contribution and openness in making these data available.
Author contributions
Study design: Yuqin Shen and Rong Shuai. Data acquisition: Rong Shuai. Statistical analysis: Yuqin Shen, Dongqian Yang, and Li Zhang. Discussion: Yuqin Shen and Rong Shuai. Manuscript drafting: Yuqin Shen. All the authors provided final approval of the manuscript submitted for publication.
Data availability
Data were fully available at https://journals.plos.org/plosone.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Younossi, Z. M., Kalligeros, M. & Henry, L. Epidemiology of metabolic dysfunction-associated steatotic liver disease. Clin. Mol. Hepatol.31, S32–s50. 10.3350/cmh.2024.0431 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Vancells Lujan, P., Viñas Esmel, E. & Sacanella Meseguer, E. Overview of non-alcoholic fatty liver disease (NAFLD) and the role of sugary food consumption and other dietary components in its development. Nutrients1310.3390/nu13051442 (2021). [DOI] [PMC free article] [PubMed]
- 3.Eslam, M. et al. A new definition for metabolic dysfunction-associated fatty liver disease: An international expert consensus statement. J. Hepatol.73, 202–209. 10.1016/j.jhep.2020.03.039 (2020). [DOI] [PubMed] [Google Scholar]
- 4.Eslam, M., Sanyal, A. J. & George, J. M. A. F. L. D. A consensus-driven proposed nomenclature for metabolic associated fatty liver disease. Gastroenterology158, 1999–2014.e1991. 10.1053/j.gastro.2019.11.312 (2020). [DOI] [PubMed]
- 5.Sarin, S. K. et al. Liver diseases in the Asia-Pacific region: a Lancet Gastroenterology & Hepatology Commission. Lancet Gastroenterol. Hepatol.5, 167–228. 10.1016/s2468-1253(19)30342-5 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lee, S. M. et al. Metabolic Dysfunction-Associated Fatty Liver Disease and Subsequent Development of Adverse Pregnancy Outcomes. Clin. Gastroenterol. Hepatol.20, 2542–2550e2548. 10.1016/j.cgh.2021.11.007 (2022). [DOI] [PubMed] [Google Scholar]
- 7.Sarkar, M. et al. Non-alcoholic fatty liver disease in pregnancy is associated with adverse maternal and perinatal outcomes. J. Hepatol.73, 516–522. 10.1016/j.jhep.2020.03.049 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Vali, Y. et al. Biomarkers for staging fibrosis and non-alcoholic steatohepatitis in non-alcoholic fatty liver disease (the LITMUS project): a comparative diagnostic accuracy study. lancet Gastroenterol. Hepatol.8, 714–725. 10.1016/s2468-1253(23)00017-1 (2023). [DOI] [PubMed] [Google Scholar]
- 9.Ajmera, V. & Loomba, R. Imaging biomarkers of NAFLD, NASH, and fibrosis. Mol. Metab.50, 101167. 10.1016/j.molmet.2021.101167 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Alizargar, J., Bai, C. H., Hsieh, N. C. & Wu, S. V. Use of the triglyceride-glucose index (TyG) in cardiovascular disease patients. Cardiovasc. Diabetol.19, 8. 10.1186/s12933-019-0982-2 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Tahapary, D. L. et al. Challenges in the diagnosis of insulin resistance: Focusing on the role of HOMA-IR and Tryglyceride/glucose index. Diabetes metabolic syndrome. 16, 102581. 10.1016/j.dsx.2022.102581 (2022). [DOI] [PubMed] [Google Scholar]
- 12.Wang, Y. et al. Baseline level and change trajectory of the triglyceride-glucose index in relation to the development of NAFLD: a large population-based cohort study. Front. Endocrinol.14, 1137098. 10.3389/fendo.2023.1137098 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Dang, K. et al. The association between triglyceride-glucose index and its combination with obesity indicators and cardiovascular disease: NHANES 2003–2018. Cardiovasc. Diabetol.23, 8. 10.1186/s12933-023-02115-9 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Zhang, F. et al. Association between triglyceride glucose-body mass index and the staging of non-alcoholic steatohepatitis and fibrosis in patients with non-alcoholic fatty liver disease. Ann. Med.56, 2409342. 10.1080/07853890.2024.2409342 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lee, S. M. et al. Nonalcoholic fatty liver disease is a risk factor for large-for-gestational-age birthweight. PloS one. 14, e0221400. 10.1371/journal.pone.0221400 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lee, S. M. et al. Non-alcoholic fatty liver disease in the first trimester and subsequent development of gestational diabetes mellitus. Diabetologia62, 238–248. 10.1007/s00125-018-4779-8 (2019). [DOI] [PubMed] [Google Scholar]
- 17.Zhang, C. & Catalano, P. Screening for Gestational Diabetes. Jama326, 487–489. 10.1001/jama.2021.12190 (2021). [DOI] [PubMed] [Google Scholar]
- 18.Hillier, T. A. et al. A Pragmatic, Randomized Clinical Trial of Gestational Diabetes Screening. N. Engl. J. Med.384, 895–904. 10.1056/NEJMoa2026028 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chen, Q. et al. Association between triglyceride-glucose related indices and mortality among individuals with non-alcoholic fatty liver disease or metabolic dysfunction-associated steatotic liver disease. Cardiovasc. Diabetol.23, 232. 10.1186/s12933-024-02343-7 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Saadeh, S. et al. The utility of radiological imaging in nonalcoholic fatty liver disease. Gastroenterology123, 745–750. 10.1053/gast.2002.35354 (2002). [DOI] [PubMed] [Google Scholar]
- 21.Taylor, K. J. et al. Quantitative US attenuation in normal liver and in patients with diffuse liver disease: importance of fat. Radiology160, 65–71. 10.1148/radiology.160.1.3520657 (1986). [DOI] [PubMed] [Google Scholar]
- 22.Hu, H., Han, Y., Cao, C. & He, Y. The triglyceride glucose-body mass index: a non-invasive index that identifies non-alcoholic fatty liver disease in the general Japanese population. J. Transl Med.20, 398. 10.1186/s12967-022-03611-4 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Jutras, G. et al. Epidemiologic Trends in Acute, Chronic, and Pregnancy-Specific Liver Diseases in Pregnancy. Am. J. Gastroenterol.10.14309/ajg.0000000000003687 (2025). [DOI] [PubMed] [Google Scholar]
- 24.Im, H. J., Ahn, Y. C., Wang, J. H., Lee, M. M. & Son, C. G. Systematic review on the prevalence of nonalcoholic fatty liver disease in South Korea. Clin. Res. Hepatol. Gastroenterol.45, 101526. 10.1016/j.clinre.2020.06.022 (2021). [DOI] [PubMed] [Google Scholar]
- 25.Quek, J. et al. Global prevalence of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in the overweight and obese population: a systematic review and meta-analysis. lancet Gastroenterol. Hepatol.8, 20–30. 10.1016/s2468-1253(22)00317-x (2023). [DOI] [PubMed] [Google Scholar]
- 26.Koralegedara, I. S., Warnasekara, J. N., Rathnayake, A., Dayaratne, K. G. & Agampodi, S. B. Fatty Liver Index is a valid predictor of non-alcoholic fatty liver disease (NAFLD) in pregnancy. BMJ Open Gastroenterol.910.1136/bmjgast-2022-000913 (2022). [DOI] [PMC free article] [PubMed]
- 27.Shuai, R., He, Y., Yang, D., Zhang, Y. & Zhang, L. Association between the atherogenic index of plasma and non-alcoholic fatty liver disease in Korean pregnant women: secondary analysis of a prospective cohort study. Front. Nutr.12, 1511952. 10.3389/fnut.2025.1511952 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Guilherme, A., Virbasius, J. V., Puri, V. & Czech, M. P. Adipocyte dysfunctions linking obesity to insulin resistance and type 2 diabetes. Nat. Rev. Mol. Cell. Biol.9, 367–377. 10.1038/nrm2391 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kirpich, I. A., Marsano, L. S. & McClain, C. J. Gut-liver axis, nutrition, and non-alcoholic fatty liver disease. Clin. Biochem.48, 923–930. 10.1016/j.clinbiochem.2015.06.023 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yamaguchi, K. et al. Inhibiting triglyceride synthesis improves hepatic steatosis but exacerbates liver damage and fibrosis in obese mice with nonalcoholic steatohepatitis. Hepatology45, 1366–1374. 10.1002/hep.21655 (2007). [DOI] [PubMed] [Google Scholar]
- 31.Hotamisligil, G. S. Role of endoplasmic reticulum stress and c-Jun NH2-terminal kinase pathways in inflammation and origin of obesity and diabetes. Diabetes54 (Suppl 2), 73–78. 10.2337/diabetes.54.suppl_2.s73 (2005). [DOI] [PubMed] [Google Scholar]
- 32.Wang, M. & Kaufman, R. J. The impact of the endoplasmic reticulum protein-folding environment on cancer development. Nat. Rev. Cancer. 14, 581–597. 10.1038/nrc3800 (2014). [DOI] [PubMed] [Google Scholar]
- 33.Hagström, H. et al. Adverse outcomes of pregnancy in women with non-alcoholic fatty liver disease. Liver Int.36, 268–274. 10.1111/liv.12902 (2016). [DOI] [PubMed] [Google Scholar]
- 34.Qian, Y. et al. Nonalcoholic Fatty Liver Disease and Adverse Pregnancy Outcomes in Women With Normal Prepregnant Weight. J. Clin. Endocrinol. Metab.108, 463–471. 10.1210/clinem/dgac567 (2023). [DOI] [PubMed] [Google Scholar]
- 35.Castera, L., Friedrich-Rust, M. & Loomba, R. Noninvasive Assessment of Liver Disease in Patients With Nonalcoholic Fatty Liver Disease. Gastroenterology156, 1264–1281e1264. 10.1053/j.gastro.2018.12.036 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Younossi, Z. M. et al. The global epidemiology of nonalcoholic fatty liver disease and nonalcoholic steatohepatitis among patients with type 2 diabetes. Clin. Gastroenterol. Hepatol. Off. Clin. Pract. J. Am. Gastroenterol. Assoc.22, 1999–2010.e1998. 10.1016/j.cgh.2024.03.006 (2024). [DOI] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data were fully available at https://journals.plos.org/plosone.