Abstract
Background
Gestational diabetes mellitus (GDM) is a common metabolic disorder during pregnancy and is closely associated with adverse pregnancy outcomes (APO), such as preeclampsia, preterm birth, and macrosomia. The triglyceride-glucose (TyG) index and growth differentiation factor 15 (GDF-15) have emerged as novel biomarkers in cardiovascular and metabolic diseases. However, their potential relevance to APO in late pregnancy among women with GDM remains unclear.
Methods
A total of 275 pregnant women diagnosed with GDM between January 2024 and March 2025 were enrolled in this study. Participants were categorized into two groups based on the presence or absence of APO: GDM patients with APO and GDM patients without APO. Serum levels of GDF-15 were measured using enzyme-linked immunosorbent assay (ELISA), and the TyG index was calculated using fasting triglyceride and glucose levels.
Results
Compared with the GDM without APO group, patients in the GDM with APO group exhibited significantly higher serum levels of GDF-15 and TyG index (both p < 0.05). Multivariate logistic regression analysis revealed that both the TyG index (OR = 1.672, 95%CI:1.074–2.602, p < 0.05) and GDF-15 (OR = 1.002, 95%CI:1.001–1.003, p < 0.05) were independent predictors of APO. After adjusting for all potential confounders, restricted cubic spline analysis indicated a positive association between both biomarkers and the risk of APO. Furthermore, receiver operating characteristic (ROC) analysis demonstrated that the combined model incorporating TyG and GDF-15 yielded an AUC of 0.722 (95%CI:0.661–0.784), suggesting good predictive performance.
Conclusion
Elevated TyG index and GDF-15 levels in late pregnancy are potential predictive biomarkers for APO in women with GDM. Combined assessment of these markers may enhance clinical risk stratification and inform targeted management strategies.
Keywords: TyG index, GDF-15, GDM, Adverse pregnancy outcomes, Late pregnancy
Introduction
Gestational diabetes mellitus (GDM) is one of the most common metabolic disorders during pregnancy. With the ongoing changes in lifestyle and dietary habits, its global prevalence has shown a steady upward trend. Epidemiological data indicate that the incidence of GDM ranges from 2% to 25% worldwide [1, 2]. GDM not only disrupts maternal metabolic and endocrine homeostasis but also increases the risk of pregnancy-related complications, such as gestational hypertension, preeclampsia, and preterm birth. Moreover, it is associated with adverse neonatal outcomes, including macrosomia, fetal growth restriction, and neonatal hypoglycemia [3–5]. Therefore, early identification and prediction of adverse pregnancy outcomes (APO) in women with GDM are essential for improving maternal and neonatal health.
The triglyceride-glucose (TyG) index, derived from fasting triglyceride and glucose levels, is a widely used surrogate marker for insulin resistance and metabolic syndrome [6, 7]. For instance, in patients with type 2 diabetes mellitus (T2DM), an elevated TyG index has been linked to an increased risk of cardiovascular events and all-cause mortality [8]. Other studies have demonstrated its predictive value for metabolic dysfunction-associated fatty liver disease (MAFLD) and other metabolism-related disorders [9]. In pregnant populations, Guo et al. reported that elevated TyG levels in early pregnancy were significantly associated with an increased risk of developing GDM and gestational hypertension, although no significant association was observed with preeclampsia, placental abruption, fetal distress, or premature rupture of membranes [10]. However, it remains unclear how TyG levels change in late pregnancy in women with GDM, and whether these changes are associated with the development of adverse pregnancy outcomes.
Growth differentiation factor 15 (GDF-15), also known as macrophage inhibitory cytokine-1 (MIC-1), is a member of the transforming growth factor-β (TGF-β) superfamily. It is secreted by various cell types, including macrophages and endothelial cells, and plays a critical role in inflammation, metabolic dysregulation, and cellular stress responses [11]. In recent years, the role of GDF-15 during pregnancy has garnered increasing attention. A 2024 study published in Nature provided compelling evidence that GDF-15 functions as a neuroactive hormone involved in pregnancy-related nausea, vomiting, and even hyperemesis gravidarum [12]. As early as 2003, it was reported to be highly expressed in the placenta and implicated in the regulation of labor and the pathophysiology of preeclampsia [13].Further investigations have revealed that GDF-15 is not only associated with increased risk of T2DM and obesity, but is also closely linked to metabolic disturbances in late pregnancy among women with GDM, potentially through modulation of insulin signaling pathways and inflammatory responses [11, 14]. Given the clinical relevance of both the TyG index and GDF-15 in metabolic disorders, this study aimed to investigate the changes in their serum levels among GDM patients and explore their associations with adverse pregnancy outcomes in late gestation. The findings may offer novel insights for risk stratification and clinical management to improve maternal and neonatal outcomes in this high-risk population.
Patients and methods
Patients selection
Between September 2023 and March 2025, pregnant women aged 18 to 45 years who were diagnosed with GDM at the Renmin Hospital of Wuhan University were consecutively enrolled. Demographic and clinical information was obtained from attending obstetricians, including age, body mass index (BMI), gestational weeks, smoking, history of polycystic ovary syndrome (PCOS), and maternal family history of diabetes/hypertension/PCOS. Exclusion criteria included incomplete medical history or data missing, severe cardiac disease, hepatic or renal insufficiency, malignancy, active severe infections or autoimmune diseases. Women who required insulin therapy for glycemic control after unsuccessful dietary and lifestyle intervention were not excluded, provided they met al.l inclusion criteria and had no severe comorbidities. Insulin use was documented and later adjusted for in multivariable analyses to minimize potential confounding. Participants were categorized into two groups based on the presence or absence of APO, which included stillbirth, spontaneous preterm birth, small for gestational age (SGA), large for gestational age (LGA), macrosomia, and postpartum hemorrhage [15]. A total of 152 patients were classified into the GDM without APO group, and 121 into the GDM with APO group (Fig. 1).This study was approved by the Ethics Committees of the Renmin Hospital of Wuhan University, and conducted in accordance with the principles of the Declaration of Helsinki.Written informed consent was obtained from all participants prior to enrollment.
Fig. 1.
Diagram of patient selection
The diagnosis of GDM and definition of late pregnancy
An oral glucose tolerance test (OGTT) was conducted between 24 and 28 gestation weeks to screen for GDM. The diagnosis of GDM was based on the criteria established by the International Association of the Diabetes and Pregnancy Study Groups (IADPSG) [16]. A diagnosis was made if any one of the following plasma glucose thresholds was met or exceeded: (1) fasting plasma glucose (FPG) ≥ 5.1 mmol/L; (2) 1-hour plasma glucose ≥ 10.0 mmol/L; (3) 2-hour plasma glucose (2 h PG) ≥ 8.5 mmol/L. Late pregnancy, also defined as the third trimester of gestation, refers to the period from 28 gestational weeks until childbirth.
Blood sampling and measurement
Upon hospital admission, all GDM patients underwent venipuncture of the median cubital vein after fasting for at least 8 h. Blood samples were collected into vacuum tubes containing clot activators. Following a 30-minute incubation at room temperature, the samples were centrifuged at 3500 rpm for 5 min, and the supernatant serum was carefully separated for subsequent analyses. Fasting plasma glucose (FPG) was measured in serum using the hexokinase enzymatic method recommended by the International Federation of Clinical Chemistry (IFCC). Serum levels of triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea, and creatinine (Cr) were measured using an automated biochemical analyzer (ADVIA 2400, Siemens).
The TyG index was calculated using the following formula: ln[(TG (mmol/L)*88.545)×(FPG(mmol/L)*18.02)/2] [6, 8, 19]. Serum GDF-15 concentrations were quantified using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, USA), with a validated linear detection range of 0 to 500 pg/mL. To determine the optimal dilution ratio for accurate quantification, a pilot study was conducted using serial dilutions at 1×, 2×, 5×, 10×, and 20×. Based on the titration curve, a 5-fold dilution was selected for all samples. To minimize inter-assay variability and enhance comparability, all serum samples were assayed in a single batch under identical experimental conditions.
Statistical analysis
All analyses were performed using SPSS 26.0 (IBM, USA), R 4.2.3, and SPSSAU software. To reduce baseline bias, a propensity score matching (PSM) analysis was conducted before group comparisons, using logistic regression based on maternal age, gestational weeks, and BMI, with nearest-neighbor matching without replacement. Post-hoc power analysis indicated sufficient sample size (power ≥ 0.8 at α = 0.05). Normality was tested before analysis. Variables with normal distribution (Cr, TyG, LDL-C) were expressed as mean ± SD and compared using the independent-sample t-test; non-normally distributed variables (e.g., Age, Gestational weeks, BMI, FPG, TG, GDF-15) were expressed as median (IQR) and compared by Mann–Whitney U test. Categorical variables were presented as counts and percentages and analyzed using the chi-square test. Logistic regression was applied to identify independent predictors of APO in GDM women. To avoid omission of potential predictors and keep model stability and interpretability, variables with p < 0.1 in univariate analysis were included in the multivariate model. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs). For quartile-based analyses, TyG and GDF-15 were ranked in ascending order and divided into four quartiles with roughly equal sample sizes. Quartiles were coded in ascending order for logistic regression to represent increasing exposure levels. The median value of each quartile was calculated and used as the representative exposure level in the trend logistic regression models. Trend p-values were obtained by treating these median values as continuous variables to evaluate linear trends across quartiles. Restricted cubic spline regression was used to explore nonlinear dose–response associations and receiver operating characteristic (ROC) analysis evaluated the diagnostic performance of key markers. A two-sided P value < 0.05 was considered statistically significant.
Results
Baseline characteristics
A total of 273 women with GDM were included in the analysis, with 121 assigned into the GDM with APO group and 152 to the GDM without APO group. There were no significant differences between the groups in terms of age, gestational weeks, family history, or insulin use. However, the APO group exhibited significantly higher levels of BMI, Cr, TG, TC, LDL-C, TyG index, and serum GDF-15 levels (all p < 0.05). Details are shown in Table 1.
Table 1.
Baseline characteristics between GDM women with and without APO
| Variables | GDM without APO (n = 152) | GDM with APO (n = 121) | p value |
|---|---|---|---|
| Clinical Variables | |||
| Age (years) | 34(24,40) | 34(28,41) | 0.459 |
| Gestational weeks | 36(33,37) | 35(32,37) | 0.257 |
| BMI | 24.17(23.19,25.64) | 24.64(23.61,25.99) | 0.047 |
| Smoking (n,%) | 17.76 | 17.36 | 0.930 |
| Maternal family history of hypertension (n,%) | 13.82 | 14.88 | 0.804 |
| Maternal family history of diabetes (n,%) | 14.47 | 18.18 | 0.408 |
| Maternal family history of PCOS (n,%) | 21.05 | 27.27 | 0.231 |
| PCOS history (n,%) | 19.08 | 26.44 | 0.147 |
| Insulin use (n,%) | 57.89 | 57.85 | 0.994 |
| Laboratory Variables | |||
| FPG (mmol/L) | 5.62(4.69,8.05) | 6.01(5.10,7.60) | 0.300 |
| 2 h PG (mmol/L) | 9.88(8.02,12.31) | 10.47(8.58,11.76) | 0.205 |
| ALT(U/L) | 19(14,26) | 20(15,26) | 0.525 |
| AST(U/L) | 17(14,23) | 18(14,22) | 0.667 |
| Urea(mmol/L) | 5.54(4.68,6.46) | 5.81(4.86,6.61) | 0.057 |
| Cr(µmol/L) | 64.47 ± 0.89 | 67.26 ± 1.06 | 0.042 |
| TG (mmol/L) | 1.57(1.06,2.15) | 1.89(1.36,2.94) | < 0.001 |
| TyG (mmol/L) | 8.96 ± 0.06 | 9.27 ± 0.06 | < 0.001 |
| TC (mmol/L) | 4.47(3.75,5.21) | 4.75(4.20,5.40) | 0.009 |
| HDL-C (mmol/L) | 0.93(0.82,1.09) | 0.97(0.82,1.20) | 0.173 |
| LDL-C (mmol/L) | 2.55 ± 0.09 | 2.85 ± 0.08 | 0.017 |
| GDF-15 (pg/mL) | 779.48(632.67,986.85) | 1103.97(747.97,1467.36) | < 0.001 |
Data were presented as percentages (%) for categorical variables and mean ± SD or median (interquartile range) for continuous variables
Abbreviations: FPG Fasting plasma glucose, 2h PG 2h plasma glucose, ALT Alanine aminotransferase, AST Aspartate aminotransferase, Cr Creatinine, TG Triglyceride, TyG TG-Glucose index, TC Total cholesterol, HDL-C High-density lipoprotein cholesterol, LDL-C Low-density lipoprotein cholesterol, GDF-15 Growth differentiation factor 15
Risk factors associated with APO
Univariate logistic regression identified several variables associated with APO risk, including BMI (OR = 1.139, 95%CI:0.995–1.304, p = 0.058), Cr (OR = 1.022, 95%CI:1.001–1.045, p = 0.043), LDL-C (OR = 1.336, 95%CI:1.050–1.701, p = 0.018), TyG index (OR = 1.922, 95%CI:1.333–2.771, p < 0.001), and GDF-15 (OR = 1.002, 95%CI:1.002–1.003, p < 0.001). These variables were then entered into a multivariate logistic regression model. After adjusting for confounding factors including BMI, Urea, Cr, TC and LDL-C, TyG index (adjusted OR = 1.672, 95%CI:1.074–2.602, p = 0.023) and GDF-15 (adjusted OR = 1.002, 95%CI: 1.001–1.003, p < 0.001) remained independent predictors of APO, whereas other variables such as Cr and LDL-C lost statistical significance (Table 2).
Table 2.
Univariate and multivariate analysis for the risk factors of APO in GDM women
| variables | Univariate Logistic Analysis | Multivariate Logistic Analysis | ||
|---|---|---|---|---|
| 95%CI | p value | 95%CI | p value | |
| Age | 1.016(0.987,1.045) | 0.285 | ||
| Gestational weeks | 0.963(0.893,1.038) | 0.322 | ||
| BMI | 1.139(0.995,1.304) | 0.058 | 1.083(0.933,1.258) | 0.294 |
| Smoking | 0.972(0.519,1.822) | 0.930 | ||
| Maternal family history of hypertension | 1.090(0.552,2.153) | 0.804 | ||
| Maternal family history of diabetes | 1.313(0.688,2.506) | 0.409 | ||
| Maternal family history of PCOS | 1.406(0.804,2.459) | 0.232 | ||
| PCOS history | 1.525(0.861,2.701) | 0.148 | ||
| Insulin use | 0.998(0.665,1.619) | 0.994 | ||
| 2 h PG | 1.069(0.968,1.179) | 0.187 | ||
| ALT | 0.996(0.975,1.016) | 0.676 | ||
| AST | 1.001(0.971,1.031) | 0.995 | ||
| Urea | 1.193(1.001,1.423) | 0.049 | 1.214(0.991,1.488) | 0.061 |
| Cr | 1.022(1.001,1.045) | 0.043 | 1.019(0.994,1.044) | 0.139 |
| TyG | 1.922(1.333,2.771) | < 0.001 | 1.672(1.074,2.602) | 0.023 |
| TC | 1.228(0.994,1.516) | 0.057 | 0.969(0.723,1.298) | 0.883 |
| HDL-C | 1.974(0.757,5.146) | 0.164 | ||
| LDL-C | 1.336(1.050,1.701) | 0.018 | 1.325(0.976,1.800) | 0.071 |
| GDF-15 | 1.002(1.002,1.003) | < 0.001 | 1.002(1.001,1.003) | < 0.001 |
Abbreviations: 2h PG 2h plasma glucose, ALT Alanine aminotransferase, AST Aspartate aminotransferase, Cr Creatinine, TyG TG-Glucose index, TC Total cholesterol, HDL-C High-density lipoprotein cholesterol, LDL-C Low-density lipoprotein cholesterol, GDF-15 Growth differentiation factor 15
Restricted cubic spline analysis of GDF-15 and TyG with APO
To further evaluate the association between GDF-15 and TyG levels and the risk of APO, patients were stratified into quartiles based on their serum concentrations. As shown in Table 3, compared with the lowest GDF-15 quartile (Q1: <688.36 pg/mL) in Model 3, the adjusted odds ratios (OR) for APO in quartiles 2, 3, and 4 were 0.941 (95%CI:0.387–2.291), 2.717 (95%CI:1.197–6.170), and 7.585 (95%CI:3.151–18.262), respectively. A clear upward trend in APO risk was observed across increasing quartiles (p < 0.001). Similarly, for TyG index quartiles (Table 4) in Model 3, using Q1 (< 8.582) as the reference, the adjusted ORs for quartiles 2, 3, and 4 were 2.922(95%CI:1.202–7.104), 4.354(95%CI:1.734–10.931), and 4.978(95%CI:1.905–13.006), respectively. The association between increasing TyG levels and higher APO risk remained statistically significant after full adjustment (p < 0.05).
Table 3.
Association of the occurrence of APO with GDF-15 in GDM population
| GDF-15 quartiles | n | Range | Median (pg/ml) | OR(95%CI) | ||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | ||||
| Quartile 1(low) | 68 | < 688.36 | 539.33 | Reference | Reference | Reference |
| Quartile 2 | 68 | 688.36–896.70 | 756.07 | 0.738(0.343,1.588) | 0.839(0.380,1.856) | 0.941(0.387,2.291) |
| Quartile 3 | 68 | 896.71–1192.83.71.83 | 1011.70 | 2.400(1.185,4.861) | 2.506(1.195,5.253) | 2.717(1.197,6.170) |
| Quartile 4(high) | 69 | > 1192.84 | 1488.79 | 6.800(3.216,0.14.380) | 7.420(3.370,16.339) | 7.585(3.151,18.262) |
| p value for trend | < 0.001 | < 0.001 | < 0.001 | |||
Model 1: no adjustment
Model 2: adjusted for Age, Gestational weeks, BMI, Smoking, Maternal family history of hypertension, Maternal family history of diabetes, Maternal family history of PCOS, PCOS history, Insulin use
Model 3: adjusted for Age, Gestational weeks, BMI, Smoking, Maternal family history of hypertension, Maternal family history of diabetes, Maternal family history of PCOS, PCOS history, Insulin use, FPG, 2h PG, ALT, AST, Urea, Cr, TG, TyG, TC, HDL-C, LDL-C
Table 4.
Association of the occurrence of APO with TyG index in GDM population
| TyG quartiles | n | Range | Median | OR(95%CI) | ||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | ||||
| Quartile 1(low) | 68 | < 8.582 | 8.314 | Reference | Reference | Reference |
| Quartile 2 | 68 | 8.582–9.055 | 8.838 | 3.178(1.511,6.685) | 3.580(1.638,7.827) | 2.922(1.202,7.104) |
| Quartile 3 | 68 | 9.056–9.526 | 9.253 | 3.269(1.551,6.890) | 3.845(1.758,8.410) | 4.354(1.734,10.931) |
| Quartile 4(high) | 69 | > 9.526 | 9.825 | 4.782(2.265,10.095) | 5.627(2.549,12.424) | 4.978(1.905,13.006) |
| p value for trend | < 0.001 | < 0.001 | 0.001 | |||
Model 1: no adjustment
Model 2: adjusted for Age, Gestational weeks, BMI, Smoking, Maternal family history of hypertension, Maternal family history of diabetes, Maternal family history of PCOS, PCOS history, Insulin use
Model 3: adjusted for Age, Gestational weeks, BMI, Smoking, Maternal family history of hypertension, Maternal family history of diabetes, Maternal family history of PCOS, PCOS history, Insulin use, 2h PG, ALT, AST, Urea, Cr, GDF-15, TC, HDL-C, LDL-C
Restricted cubic spline regression was employed to further explore the dose–response relationship. As illustrated in Figs. 2 and 3, both TyG index and GDF-15 levels exhibited a nonlinear positive association with APO risk, suggesting that the risk increased more steeply beyond specific thresholds of these markers.
Fig. 2.
Nonlinear associations between GDF-15 and the risk of APO
Fig. 3.
Nonlinear associations between TyG and the risk of APO
ROC analysis of predictive performance of GDF-15 and TyG index
The diagnostic utility of TyG and GDF-15 in predicting APO was evaluated using ROC curve analysis. The AUC for GDF-15 was 0.708 (95%CI:0.645–0.772), with a sensitivity of 59.5% and specificity of 78.9% at an optimal cut-off value of 1002.83 pg/mL. The TyG index alone showed lower predictive ability, with an AUC of 0.629 (95%CI:0.563–0.695). Notably, combining TyG and GDF-15 improved predictive performance, yielding an AUC of 0.722 (95%CI:0.661–0.784), as shown in Table 5 and Fig. 4.
Table 5.
ROC analysis of predictive performance of GDF-15 and TyG index
| Variables | AUC | SD | p value | 95%CI | Threshold | Sensitivity | Specificity | Youden index |
|---|---|---|---|---|---|---|---|---|
| GDF-15 | 0.708 | 0.032 | < 0.001 | 0.645–0.772 | 1002.83 | 0.595 | 0.789 | 0.385 |
| TyG | 0.629 | 0.034 | < 0.001 | 0.563–0.695 | 8.5880 | 0.876 | 0.349 | 0.225 |
| combination | 0.722 | 0.031 | < 0.001 | 0.661–0.784 | - | 0.686 | 0.684 | 0.370 |
Abbreviations: GDF-15 Growth differentiation factor 15, TyG TG-glucose index, Combination GDF-15 + TyG
Fig. 4.
ROC curve of GDF-15 and TyG in predicting the occurrence of APO
Discussion
This study investigated the association between serum TyG index and GDF-15 levels and APO in women with GDM during late pregnancy. The results demonstrated that both biomarkers were significantly elevated in GDM patients with APO and positively correlated with outcomes including preterm birth, macrosomia, fetal growth restriction, and preeclampsia. These findings suggest that TyG and GDF-15 may serve as promising biomarkers for early risk stratification in late pregnancy among women with GDM, providing valuable insights for future mechanistic studies and personalized clinical interventions.
The TyG index, as a key surrogate marker for insulin resistance and metabolic syndrome, has attracted increasing attention across various metabolic disease fields. For instance, previous studies have demonstrated a strong association between TyG levels and hepatic fat accumulation as well as disease severity in patients with non-alcoholic fatty liver disease (NAFLD), underscoring its value in liver-related metabolic disorders [17, 18]. Similarly, in cardiovascular research, elevated TyG levels have been linked to an increased risk of coronary artery disease and atherosclerosis [19], aligning with the current study’s findings on TyG as a predictor of APO in women with GDM, and further emphasizing its clinical relevance in assessing metabolic stress and systemic dysfunction. In the obstetric context, several studies have explored the relationship between TyG levels at different gestational stages and pregnancy-related complications. Most of these investigations have focused on early pregnancy, particularly the association between elevated TyG and the subsequent development of GDM [20]. In contrast, this study emphasizes the importance of TyG in late pregnancy, demonstrating its positive correlation with APO such as preterm birth and macrosomia. As pregnancy progresses, maternal metabolic burden increases, and insulin resistance is further exacerbated, which potentially directly impairing placental function, fetal growth, and maternal vascular homeostasis. This may explain the stronger association observed in late gestation. Previous studies in early-pregnancy have provided limited evidence linking TyG to outcomes such as preeclampsia or placental abruption [21, 22]. The discrepancy may be due to dynamic changes in maternal metabolic state across trimesters, as well as variations in study design, sample characteristics, and outcome definitions.
GDF-15 plays a crucial role in both physiological and pathological processes during pregnancy. Recent studies have reported significantly elevated serum GDF-15 levels in patients with preeclampsia, indicating its involvement in the pathogenesis of hypertensive disorders of pregnancy [23]. This observation is consistent with our finding that GDF-15 levels are markedly increased in GDM patients who developed APO. Moreover, GDF-15 has been implicated in regulating trophoblast function, placental angiogenesis, and fetal development, possibly through its effects on cellular migration, vascular remodeling, and the immune microenvironment [24, 25]. These biological fuctions provide plausible mechanistic insights into the association between GDF-15 and APO. In contrast, the present study broadens the scope of GDF-15 research by investigating its metabolic and inflammatory roles in late pregnancy, especially in women with GDM. Our findings suggest that GDF-15 may contribute to the pathophysiology of APO through multiple mechanisms, including interference with insulin signaling pathways, amplification of systemic inflammation, and disruption of endothelial function. Collectively, these effects may contribute to adverse maternal and fetal outcomes in GDM.
Although this study identified significant associations between elevated TyG index and GDF-15 levels and APO in women with GDM, the underlying mechanisms remain incompletely understood. Insulin resistance and chronic low-grade inflammation are currently considered key mediators. In insulin resistance, maternal glucose metabolism is impaired due to reduced insulin sensitivity, leading to insufficient glucose uptake and utilization. This promotes lipolysis and increases serum triglyceride levels, consequently elevating the TyG index [26]. Insulin resistance may also activate endocrine and metabolic cascades such as the renin–angiotensin–aldosterone system (RAAS), resulting in vasoconstriction and elevated blood pressure, thereby increasing the risk of preeclampsia. Concurrently, hyperglycemia can impair endothelial cell function in the placenta, disrupt angiogenesis and nutrient transfer, and ultimately contribute to fetal growth restriction or macrosomia [27]. GDF-15, a stress-induced cytokine, plays roles in inflammation, metabolic dysregulation, and cellular stress responses. Its elevated levels in GDM patients may reflect an adaptive response to underlying metabolic disturbances. On one hand, GDF-15 may interfere with insulin signaling by inhibiting the phosphorylation of insulin receptor substrate (IRS), thereby exacerbating insulin resistance. On the other hand, it can induce the release of inflammatory cytokines such as tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6), promoting endothelial dysfunction and reducing placental perfusion, which increases the risk of APO [28]. Furthermore, abnormal overexpression of GDF-15 in placental tissue may directly impair placental development and function, hinder maternal-fetal exchange, and negatively affect fetal growth [29–31]. While these proposed mechanisms are biologically plausible and supported by emerging evidence, the causal relationship between TyG/GDF-15 and APO cannot yet be established based on this current observational data. It is currently unclear whether these biomarkers are direct pathogenic contributors or merely reflective indicators of underlying metabolic and placental dysfunction. Further mechanistic investigations and prospective multicenter studies are warranted to validate these hypotheses and explore potential intervention strategies.
Several limitations should be acknowledged. First, this single-center cross-sectional study had a relatively modest sample size, which limits the generalizability of the findings. Future studies with larger, more diverse populations are warranted to validate these results and enhance external validity. Second, dynamic changes of TyG and GDF-15 across gestation were not assessed, precluding causal inference. Third, although key confounders were adjusted for, residual confounding cannot be entirely excluded. Fourth, the predictive ability of individual markers was moderate, as reflected by a Youden index below 0.5 in ROC analysis. This limitation likely results from the multifactorial nature and clinical heterogeneity of adverse pregnancy outcomes rather than methodological bias, and the ROC findings should therefore be interpreted as exploratory. Finally, APO result from complex, multifactorial etiologies; thus, integrating placental biomarkers and functional studies would provide deeper mechanistic insights.
Conclusion
In conclusion, this is the first study to comprehensively evaluate the association between TyG index, GDF-15, and APO in late pregnancy among women with GDM. These biomarkers hold promise for improving risk prediction and guiding clinical decision-making. Their combined assessment may facilitate the identification of high-risk individuals for early intervention and ultimately enhancing maternal and neonatal health outcomes. Future multicenter prospective studies and mechanistic investigations are essential to validate and expand upon our findings.
Authors’ contributions
Hang Liu: design the methodology, drafted the manuscripts, and funding. Yufeng Mei: conceived the study, collected data, and performed statistical analyses. Jingru Cheng: performed statistical analyses. Shulin Zeng: reviewed and revised the manuscript. Hua Liang: reviewed and revised the manuscript, funding, and approved its final version for publication.
Funding
This study was supported by the Open Project of Hubei Key Laboratory(2023KFZZ010), the Natural Science Foundation of Hubei Province, China(2024AFB111) and the Development Center for Medical Science & Technology, National Health Commission of the People’s Republic of China(WKZX2024DN0169).
Data availability
The datasets generated during and analyzed during the current study are not publicly available due to privacy or ethical restrictions but are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committees of the Renmin Hospital of Wuhan University (No.WDRY2024-K249), and conducted in accordance with the principles of the Declaration of Helsinki.Written informed consent was obtained from all participants prior to enrollment.
Consent for publication
Not Applicable.
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.Omori Y, Jovanovic L. Proposal for the reconsideration of the definition of gestational diabetes. Diabetes Care. 2005;28(10):2592–3. 10.2337/diacare.28.10.2592. [DOI] [PubMed] [Google Scholar]
- 2.Sacks DA, Hadden DR, Maresh M, et al. Frequency of gestational diabetes mellitus at collaborating centers based on IADPSG consensus panel-recommended criteria: the hyperglycemia and adverse pregnancy outcome (HAPO) study. Diabetes Care. 2012;35(3):526–8. 10.2337/dc11-1641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Jiang L, Tang K, Magee LA, et al. A global view of hypertensive disorders and diabetes mellitus during pregnancy. Nat Rev Endocrinol. 2022;18(12):760–75. 10.1038/s41574-022-00734-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hay WW Jr. Care of the infant of the diabetic mother. Curr Diab Rep. 2012;12(1):4–15. 10.1007/s11892-011-0243-6. [DOI] [PubMed] [Google Scholar]
- 5.Li H, Gao L, Yang X, Chen L. Development and validation of a risk prediction model for preterm birth in women with gestational diabetes mellitus. Clin Endocrinol (Oxf). 2024;101(3):206–15. 10.1111/cen.15044. [DOI] [PubMed] [Google Scholar]
- 6.Tao LC, Xu JN, Wang TT, Hua F, Li JJ. Triglyceride-glucose index as a marker in cardiovascular diseases: landscape and limitations. Cardiovasc Diabetol. 2022;21(1):68. 10.1186/s12933-022-01511-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lopez-Jaramillo P, Gomez-Arbelaez D, Martinez-Bello D, et al. Association of the triglyceride glucose index as a measure of insulin resistance with mortality and cardiovascular disease in populations from five continents (PURE study): a prospective cohort study. Lancet Healthy Longev. 2023;4(1):e23–33. 10.1016/S2666-7568(22)00247-1. [DOI] [PubMed] [Google Scholar]
- 8.Yao Y, Wang B, Geng T, et al. The association between TyG and all-cause/non-cardiovascular mortality in general patients with type 2 diabetes mellitus is modified by age: results from the cohort study of NHANES 1999–2018. Cardiovasc Diabetol. 2024;23(1):43. 10.1186/s12933-024-02120-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wang J, Yan S, Cui Y, et al. The diagnostic and prognostic value of the triglyceride-glucose index in metabolic dysfunction-associated fatty liver disease (MAFLD): a systematic review and meta-analysis. Nutrients. 2022;14(23):4969. 10.3390/nu14234969. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Guo Y, Lu J, Bahani M, et al. Triglyceride-glucose index in early pregnancy predicts the risk of gestational diabetes: a prospective cohort study. Lipids Health Dis. 2024;23(1):87. 10.1186/s12944-024-02076-2 Published 2024 Mar 25. [published correction appears in Lipids Health Dis. 2024;23(1):132. https://doi.org/10.1186/s12944-024-02110-3.] . [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wang D, Day EA, Townsend LK, Djordjevic D, Jørgensen SB, Steinberg GR. GDF15: emerging biology and therapeutic applications for obesity and cardiometabolic disease. Nat Rev Endocrinol. 2021;17(10):592–607. 10.1038/s41574-021-00529-7. [DOI] [PubMed] [Google Scholar]
- 12.Fejzo M, Rocha N, Cimino I, et al. GDF15 linked to maternal risk of nausea and vomiting during pregnancy. Nature. 2024;625(7996):760–7. 10.1038/s41586-023-06921-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Marjono AB, Brown DA, Horton KE, Wallace EM, Breit SN, Manuelpillai U. Macrophage inhibitory cytokine-1 in gestational tissues and maternal serum in normal and pre-eclamptic pregnancy. Placenta. 2003;24(1):100–6. 10.1053/plac.2002.0881. [DOI] [PubMed] [Google Scholar]
- 14.Li E, Chen P, Lu J, et al. Serum growth differentiation factor 15 is closely associated with metabolic abnormalities in Chinese pregnant women. J Diabetes Invest. 2021;12(8):1501–7. 10.1111/jdi.13488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tao J, Huang Y, Li Y, Dai W. Platelet-to-lymphocyte ratio and serum hsCRP levels in third trimester and adverse pregnancy outcomes in women with gestational diabetes mellitus. Sci Rep. 2023;13(1):20963. 10.1038/s41598-023-48371-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Weinert LS. International Association of Diabetes and Pregnancy Study Groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy: comment to the International Association of Diabetes and Pregnancy Study Groups Consensus Panel. Diabetes Care. 2010;33(7):e97–8. 10.2337/dc10-0544. [DOI] [PubMed] [Google Scholar]
- 17.Ling Q, Chen J, Liu X, et al. The triglyceride and glucose index and risk of nonalcoholic fatty liver disease: a dose-response meta-analysis. Front Endocrinol (Lausanne). 2023;13:1043169. 10.3389/fendo.2022.1043169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Liu SJ, Duan JH, Chen YY, et al. Unraveling the triglyceride-glucose index: a key predictor of liver fat content and the amplifying role of BMI: evidence from a large physical examination data. Front Endocrinol (Lausanne). 2025;16:1555300. 10.3389/fendo.2025.1555300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Liang S, Wang C, Zhang J, et al. Triglyceride-glucose index and coronary artery disease: a systematic review and meta-analysis of risk, severity, and prognosis. Cardiovasc Diabetol. 2023;22(1):170. 10.1186/s12933-023-01906-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zhang J, Fang X, Song Z, et al. Positive association of triglyceride glucose index and gestational diabetes mellitus: a retrospective cohort study. Front Endocrinol (Lausanne). 2025;15:1475212. 10.3389/fendo.2024.1475212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Li Q, Zhao C, Liu M, Li M, Zhang Y, Yue C. Association between triglyceride-glucose index in early pregnancy and risk of preeclampsia: a multicenter retrospective cohort study. Lipids Health Dis. 2025;24(1):152. 10.1186/s12944-025-02556-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pan Y, Zou S, Xu Y, et al. Is there any association between early trimester triglyceride-glucose index and incidence of hypertensive disorder of pregnancy and adverse pregnancy outcomes? Front Endocrinol (Lausanne). 2023;14:1093991. 10.3389/fendo.2023.1093991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sugulle M, Dechend R, Herse F, et al. Circulating and placental growth-differentiation factor 15 in preeclampsia and in pregnancy complicated by diabetes mellitus. Hypertension. 2009;54(1):106–12. 10.1161/HYPERTENSIONAHA.109.130583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zeng YT, Liu WF, Zheng PS, Li S. GDF15 deficiency hinders human trophoblast invasion to mediate pregnancy loss through downregulating Smad1/5 phosphorylation. iScience. 2023;26(10):107902. 10.1016/j.isci.2023.107902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Al-Mudares F, Reddick S, Ren J, Venkatesh A, Zhao C, Lingappan K. Role of growth differentiation factor 15 in lung disease and senescence: potential role across the lifespan. Front Med (Lausanne). 2020;7:594137. 10.3389/fmed.2020.594137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sánchez-García A, Rodríguez-Gutiérrez R, Mancillas-Adame L, et al. Diagnostic accuracy of the triglyceride and glucose index for insulin resistance: a systematic review. Int J Endocrinol. 2020;2020:4678526. 10.1155/2020/4678526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.McElwain CJ, Tuboly E, McCarthy FP, McCarthy CM. Mechanisms of endothelial dysfunction in pre-eclampsia and gestational diabetes mellitus: windows into future cardiometabolic health? Front Endocrinol (Lausanne). 2020;11:655. 10.3389/fendo.2020.00655. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Asrih M, Wei S, Nguyen TT, Yi HS, Ryu D, Gariani K. Overview of growth differentiation factor 15 in metabolic syndrome. J Cell Mol Med. 2023;27(9):1157–67. 10.1111/jcmm.17725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kim J, Kim SH, Kang H, et al. TFEB-GDF15 axis protects against obesity and insulin resistance as a lysosomal stress response. Nat Metab. 2021;3(3):410–27. 10.1038/s42255-021-00368-w. [DOI] [PubMed] [Google Scholar]
- 30.Adela R, Banerjee SK. GDF-15 as a target and biomarker for diabetes and cardiovascular diseases: a translational prospective. J Diabetes Res. 2015;2015:490842. 10.1155/2015/490842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ruszała M, Niebrzydowska M, Pilszyk A, Kimber-Trojnar Ż, Trojnar M, Leszczyńska-Gorzelak B. Novel biomolecules in the pathogenesis of gestational diabetes mellitus. Int J Mol Sci. 2021;22(21):11578. 10.3390/ijms222111578. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets generated during and analyzed during the current study are not publicly available due to privacy or ethical restrictions but are available from the corresponding author on reasonable request.