Abstract
Background
Existing diagnostic criteria for gestational diabetes mellitus (GDM) rely solely on glucose thresholds, which are insufficient to capture metabolic heterogeneity. We aimed to evaluate the role of C-peptide measured during the oral glucose tolerance test (OGTT) in assisting the development of stratified treatment strategies and predicting the risk of adverse pregnancy outcomes.
Methods
This nested case–control study conducted within the Xi’an Longitudinal Mother–Child Cohort included 1014 pregnant women with GDM and 1014 without GDM (non-GDM) who delivered singleton live-born infants between January 1, 2017, and December 31, 2018. C-peptide levels were measured at three intervals during the OGTT. Latent class trajectory modeling was used to identify distinct C-peptide trajectories, and logistic regression was used to assess their associations with adverse fetal and maternal outcomes.
Results
Two principal C-peptide trajectories were identified in GDM despite similar glucose profiles. GDM Class 1 (771, 76.04%) presented a delayed 120-min C-peptide peak and poorer beta-cell secretion, whereas GDM Class 2 (243, 23.96%) presented a sharp 60-min peak followed by a decline and significantly increased insulin resistance, with greater risks of delivering large for gestational age (LGA) (adjusted odds ratio (aOR), 1.52; 95% confidence interval (CI), 1.07–2.15) and macrosomia (aOR, 1.83; 95% CI, 1.13–2.97). Surprisingly, 21.7% (220) of the non-GDM group had a high C-peptide response associated with elevated preeclampsia risk (aOR, 2.91; 95% CI, 1.25–6.74).
Conclusions
Dynamic OGTT-derived C-peptide trajectories revealed clinically significant metabolic subgroups of GDM that were obscured by glucose-only diagnostics, with the predominantly insulin-resistant Class being at higher risk of fetal overgrowth.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12916-025-04281-x.
Keywords: Gestational diabetes mellitus, C-peptide, Latent class trajectory analysis, Beta-cell function, Insulin resistance
Background
Gestational diabetes mellitus (GDM) is characterized by glucose intolerance first identified during pregnancy [1]. Affecting 14.7% (14.7–14.8%) of pregnancies in the Western Pacific region [2], GDM is strongly associated with adverse maternal and neonatal outcomes, including cesarean delivery, macrosomia, preeclampsia, and long-term metabolic dysfunction [3, 4]. Despite standardized diagnostic criteria, GDM represents a heterogeneous disorder [5] with variable pathophysiology [6]. Current diagnostic approach through oral glucose tolerance test (OGTT) relies on static glucose thresholds and fails to capture dynamic metabolic dysregulation. These emphasize the need for biomarkers that reflect underlying insulin secretion and resistance.
C-peptide, a cleavage product of proinsulin, is produced in equimolar amounts with insulin. Unlike insulin, C-peptide is not significantly cleared by the liver and can be used to assess endogenous insulin secretion capacity. In contrast to insulin, C-peptide exhibits minimal hepatic extraction (approximately 5–10%), with a half-life of 20–30 min, making it a more reliable indicator for assessing endogenous insulin secretion capacity [7]. Previous studies showed that fasting serum C-peptide in the first and second trimesters was associated with the risk of developing GDM [8], and the accuracy of fasting C-peptide to predict GDM with pharmacotherapy was superior to other traditional risk factors, such as prepregnancy BMI [9]. Likewise, postpartum cord blood C-peptide levels were elevated in women with GDM and women delivering macrosomia [10], and the risk of developing type 2 diabetes in women with previous GDM was associated with elevated postpartum fasting C-peptide but not with insulin levels [11]. Additionally, metabolic indices incorporating C-peptide have revealed distinct GDM subtypes with unique physiological profiles and pregnancy outcomes [12, 13]. These findings highlight the potential of C-peptide as a biomarker for improving pregnancy risk stratification and elucidating the heterogeneity of GDM pathophysiology.
However, most existing evidence has relied on traditional summary metrics or static comparisons, which can only provide overall secretion profiles or a static snapshot of beta-cell function. Dynamic fluctuations in insulin secretion during glucose challenge in GDM have not been fully characterized. The OGTT, which is routinely employed for diagnosing GDM [14], provides a unique yet underexploited opportunity to measure C-peptide at fasting, 1-h, and 2-h intervals. Based on this, it is possible not only to evaluate the different dynamic responses to insulin secretion but also not to increase unnecessary blood draws in patients. Recently, latent class trajectory modeling (LCTM) has been recognized as a valuable method for providing pathophysiological insights to effectively stratify glucose metabolism characteristics in non-diabetic and obese populations [15, 16] but not currently in women with GDM. Therefore, with LCTM, we aimed to identify distinct C-peptide trajectories of women with GDM during OGTT and evaluate the associations with maternal and neonatal adverse outcomes to assist in precision clinical management.
Methods
Study design and population
This nested case–control study was based on the Xi’an Longitudinal Mother–Infant Cohort (XAMC), a prospective cohort established in 2013 in Northwest China to investigate pregnancy outcomes in women enrolled before 20 weeks of gestation [17]. We included participants with complete longitudinal electronic health records from preconception to delivery from January 2017 to December 2018. Exclusion criteria included artificial fertilization, multiple pregnancies, abortion, stillbirths, delivery before 28 weeks of gestation, pre-existing type 1 or type 2 diabetes mellitus, and any severe systemic diseases such as cancer, chronic renal failure, severe anemia, immune disorders, or other endocrine disorders. From this cohort, 1014 women diagnosed with GDM according to the International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria were included as cases. Controls (n = 1014) with normal glucose tolerance were selected through 1:1 matching by delivery time (Additional file 1: Fig. S1). The diagnosis of GDM was confirmed via a standardized 75-g 2-h OGTT between 24 and 28th weeks of gestation [14], with a fasting plasma glucose (FPG) level of ≥ 5.1 mmol/L, a 1-h blood glucose level ≥ 10.0 mmol/L, or a 2-h blood glucose level ≥ 8.5 mmol/L [18].
This study was conducted in accordance with the Declaration of Helsinki and received ethical approval from the Ethics Committee of Xi’an Jiaotong University (XJTU 2016-053) and the Ethics Committee of Northwest Women and Children’s Hospital (NWCH 2012-013). All participants provided written informed consent upon recruitment.
Laboratory tests and the calculation of glucometabolic parameters
The OGTT was conducted in the morning after 8–12 h of fasting. Venous blood samples were collected at three time points: fasting (0 h), 1 h, and 2 h after glucose loading, and measured for glucose, insulin, and C-peptide levels. Glucose levels were measured using the hexokinase method on an AU5800 clinical chemistry analyzer (Beckman Coulter Inc., USA) with intra- and inter-assay coefficients of variation (CVs) of 0.8% and 1.2%, respectively. Insulin and C-peptide levels were measured using electrochemiluminescent immunoassay kits (insulin, range 0.2–1000 µIU/mL; C-peptide, range 0.010–40.0 ng/mL) on a Cobas e 801 analyzer (Roche Diagnostics, Mannheim, Germany). The intra- and inter-assay CVs were 1.8% and 3.2% for insulin, and 1.8% and 3.5% for C-peptide, respectively. All laboratory measurements were performed in the hospital’s accredited central laboratory following standard manufacturers’ protocols, with data extracted from the Hospital Information System (HIS).
To characterize glucose metabolism dynamics, we calculated glucose composite indices. Insulin sensitivity and resistance were assessed by the homeostasis model assessment of insulin resistance (HOMA-IR) [19], quantitative sensitivity check indices for insulin (QUICKIi) and C-peptide (QUICKIc) and Matsuda index [20, 21], whereas beta-cell function and insulin secretion were measured via homeostasis models assessment of beta-cell function (HOMA-B), insulin secretion-sensitivity index-2 (ISSI-2), and modified insulinogenic indices for insulin (IGIi) and C-peptide (IGIc) [22, 23], the calculations of which are shown in the Additional file 1 (pp. 2).
Covariates
Maternal demographics and reproductive and medical histories were prospectively collected by physicians during enrollment interviews (< 6 weeks of gestation) and established in records. Clinical test data, gestational and delivery records, and discharge diagnoses were extracted from HIS records. Prepregnancy BMI was calculated as weight (kg)/height2 (m2), with overweight/obesity defined as ≥ 24 kg/m2 per the Chinese criteria [24]. In accordance with the Chinese Standard of Recommendation for Weight Gain During Pregnancy Period [25], gestational weight gain (GWG) was classified as insufficient, adequate, or excessive (recommended GWG, 11–16 kg for prepregnancy BMI < 18.5 kg/m2; 8–14 kg for 18.5 ≤ BMI < 24 kg/m2; 7–11 kg for 24 ≤ BMI < 28 kg/m2; and 5–9 kg for BMI > 28 kg/m2). Parity was dichotomized as primiparous or multiparous. History of pregnancy loss was defined as any of the following events in previous pregnancies: spontaneous abortion, induced abortion, stillbirth, or fetal demise.
Pregnancy outcomes
The primary outcome was a composite of adverse fetal outcomes, defined as having at least one of the following outcomes: prematurity, macrosomia, low birth weight (LBW), small for gestational age (SGA), large for gestational age (LGA), fetal distress, and acute chorioamnionitis. Prematurity referred to infants born before 37 weeks of gestation. Neonates with a birth weight exceeding 4000 g or less than 2500 g were classified as having macrosomia or LBW, respectively. In accordance with the Chinese growth standard for newborns [26], infants with birth weights below the 10th percentile for gestational age were considered SGA, whereas those with birth weights above the 90th percentile for gestational age were considered LGA.
The secondary outcomes were gestational metabolic diseases, including pregnancy-induced hypertension, preeclampsia, and thyroid dysfunctions, and the individual outcomes included in the primary outcome. Pregnancy-induced hypertension was defined as elevated blood pressure (≥ 140/90 mmHg) de novo after 20 gestational weeks without proteinuria [27]. Preeclampsia was defined as hypertension after 20 weeks accompanied by proteinuria or organ dysfunction [27]. Thyroid dysfunctions included hyperthyroidism, hypothyroidism, subclinical hypothyroidism, and Hashimoto’s thyroiditis, diagnosed based on serum TSH and free T4 levels, and/or thyroid autoantibodies [28].
Statistical analysis
Categorical variables were expressed as frequencies (%). Continuous variables were tested for normality using the Shapiro–Wilk test and were presented as mean ± standard deviation (SD) if normally distributed or as median with interquartile range (IQR) otherwise. Continuous variables were compared between the GDM and non-GDM groups via the Wilcoxon test or Student’s t test and among groups stratified by age, prepregnancy BMI, and family history via the Kruskal‒Wallis test or analysis of variance. Categorical variables were analyzed by the chi-square test or Fisher’s exact test. Spearman’s correlation analysis was employed to assess correlations among glucose, C-peptide, and insulin levels given their non-normal distributions across all three time points.
Latent class trajectory modeling was used to identify distinct C-peptide response trajectories in the GDM and non-GDM populations respectively, using fixed (group) and random (individual) effects. Potential categories of 1–7 were modeled using a quadratic time term for 2 h C-peptide dynamics. Model selection relied on the Akaike information criterion (AIC), Bayesian information criterion (BIC), sample-size adjusted BIC (SABIC), and entropy values, with lower AIC/BIC/SABIC and higher entropy (closer to 1) indicating superior fit. Class membership probabilities were calculated for each participant to assign individuals to the most likely latent C-peptide response class. The optimal number of categories was determined on the basis of model fit metrics, class size and clinical interpretability, requiring an average posterior probability > 0.7 and a category headcount ratio > 0.2.
After identifying the best-fitting model, we analyzed glucose metabolism profiles across the latent classes via the Kruskal–Wallis and chi‒square tests, and predicted its risk factors by logistic regression. We also analyzed the association between latent classes and study outcomes by logistic regression and adjusted for covariates to calculate adjusted odds ratios (aORs) and 95% confidence intervals (CIs). Three adjustment models were applied in the following order: Model 1, adjusted for prepregnancy BMI and family history of diabetes; Model 2, adjusted for maternal age, history of previous cesarean delivery, and history of pregnancy loss; and Model 3, adjusted for all the covariates adjusted for in Model 1 and Model 2. The LCTM was constructed in R version 4.0.2 (package ‘lcmm’), and the remaining analyses were performed with IBM SPSS Statistics 26. P values less than 0.05 (two-tailed) were considered statistically significant.
Results
The study included 2028 women with a median maternal age of 30 (28, 34) years, gestational age at delivery of 39 (38, 40) weeks, and prepregnancy BMI of 21.4 (19.5, 23.6) kg/m2. Compared with the non-GDM group, the GDM group presented significantly higher maternal age, prepregnancy BMI, GWG, and family history of diabetes (all P < 0.05). Among women with GDM, 84.7% (859) controlled their blood glucose through diet and/or exercise, and 4.5% (46) required insulin therapy. The primary outcome (any adverse neonatal outcome) occurred in 56.5% overall (GDM: 58.6%, non-GDM: 54.3%), and the prevalence of gestational metabolic diseases was 16.2% in both groups (Additional file 1: Table S1).
Glucose metabolism indicator profiles
Compared with the non-GDM group, the GDM group exhibited significantly higher fasting, 60-min, and 120-min glucose, insulin, and C-peptide levels during the OGTT, along with higher areas under the curve (AUCs) for these indicators (all P < 0.001; Table 1). The GDM group also presented lower QUICKIi, QUICKIc, Matsuda, IGIc, and ISSI-2 indices and higher HOMA-IR indices (all P < 0.001). The correlations between C-peptide and insulin were stronger in the GDM group (Additional file 1: Table S2). Age-stratified analyses revealed that GDM women aged ≥ 35 years had lower fasting C-peptide and insulin levels and higher blood glucose levels. Moreover, insulin and C-peptide levels were all higher in the prepregnancy BMI ≥ 24 kg/m2 groups, and those with a family history of diabetes had elevated postprandial glucose levels (all P < 0.05) (Additional file 1: Tables S2–S4).
Table 1.
Glucose metabolism indicator levels in the GDM and non-GDM groups
| All participants | Non-GDM group | GDM group | P | |
|---|---|---|---|---|
| N | 2028 | 1014 | 1014 | |
| Glucose level (mmol/L) | ||||
| Fasting | 4.79 (4.49, 5.20) | 4.54 (4.34, 4.75) | 5.20 (4.87, 5.39) | <0.001 |
| 60 min | 8.26 (6.95, 9.75) | 7.24 (6.32, 8.22) | 9.66 (8.32, 10.48) | <0.001 |
| 120 min | 6.92 (6.10, 8.00) | 6.36 (5.72, 7.00) | 7.86 (6.80, 8.22) | <0.001 |
| Insulin level (µIU/mL) | ||||
| Fasting | 7.8 (5.30, 11.10) | 5.95 (4.55, 8.50) | 9.80 (7.30, 13.03) | <0.001 |
| 60 min | 47.4 (29.20, 69.38) | 37.25 (18.18, 56.25) | 56.90 (39.80, 81.48) | <0.001 |
| 120 min | 41.3 (25.50, 61.60) | 31.05 (15.30, 45.63) | 52.80 (36.28, 75.20) | <0.001 |
| C-peptide level (ng/mL) | ||||
| Fasting | 1.60 (1.27, 2.01) | 1.42 (1.15, 1.74) | 1.82 (1.45, 2.26) | <0.001 |
| 60 min | 7.11 (5.71, 8.78) | 6.72 (5.40, 8.24) | 7.53 (6.05, 9.19) | <0.001 |
| 120 min | 7.25 (5.79, 8.98) | 6.57 (5.29, 7.94) | 8.11 (6.48, 9.92) | <0.001 |
| Area under the curve | ||||
| Glucose (mmol/L*h) | 14.21 (12.47, 16.09) | 12.74 (11.53, 13.90) | 16.05 (14.64, 17.10) | <0.001 |
| Insulin (µIU/mL*h) | 73.68 (48.56, 103.46) | 58.43 (32.48, 83.59) | 89.83 (65.09, 122.58) | <0.001 |
| C-peptide (ng/mL *h) | 11.52 (9.50, 14.24) | 10.74 (8.96, 13.00) | 12.59 (10.31, 17.57) | <0.001 |
| Glucose Composite Indicator | ||||
| HOMA-IR | 1.67 (1.10, 2.49) | 1.20 (0.91, 1.74) | 2.28 (1.64, 3.08) | <0.001 |
| QUICKIi × 102 | 35.33 (33.30, 37.76) | 37.23 (35.11, 38.98) | 33.71 (32.31, 35.43) | <0.001 |
| QUICKIc × 102 | 46.75 (44.17, 49.40) | 48.50 (46.31, 50.92) | 44.86 (42.89, 47.31) | <0.001 |
| Matsuda index | 6.01 (4.23, 9.39) | 8.78 (5.99, 14.15) | 4.62 (3.36, 6.04) | <0.001 |
| HOMA-B | 122.73 (91.99, 165.00) | 121.45 (90.07, 169.58) | 123.16 (93.41, 160.72) | 0.497 |
| IGIi | 11.58 (7.31, 17.81) | 11.78 (5.70, 17.98) | 11.48 (7.69, 17.76) | 0.063 |
| IGIc | 1.68 (1.20, 2.38) | 1.99 (1.53, 2.73) | 1.37 (0.99,1.92) | <0.001 |
| ISSI-2 | 533.10 (422.80, 679.07) | 634.75 (499.50, 781.80) | 464.88 (398.87, 546.30) | <0.001 |
Bolded P values indicate statistical significance (P＜0.05)
The area under the curve (AUC) was calculated via the time–blood glucose curve and on the basis of the calculus principle for the AUC at the 3 time points in hours
Abbreviations HOMA-IR Homeostasis models of assessment of insulin resistance, HOMA-B Homeostasis models of assessment of beta-cell function, QUICKIi Quantitative sensitivity check indices for insulin, QUICKIc Quantitative sensitivity check indices for C-peptide, IGIi (µIU/mg) Insulinogenic indices for insulin, IGIc (ng/mg) Insulinogenic indices for C-peptide, ISSI-2 Insulin secretion-sensitivity index-2
Latent classes of C-peptide patterns and corresponding glucose and insulin responses
Latent class trajectory modeling identified two distinct C-peptide response patterns in both GDM group and non-GDM group based on model fitting statistics and class membership probabilities (Additional file 1: Tables S5–S7). GDM Class 1 (76.04%) presented a slower increase in C-peptide, peaking at 8.06 (7.90, 8.26) ng/mL at 120 min, whereas GDM Class 2 (23.96%) exhibited a sharp C-peptide peak at 60 min (10.64 (10.31, 10.96) ng/mL), followed by a decline (Fig. 1A). Notably, the non-GDM Class 2 had a similar C-peptide trajectory to GDM Class 2, peaking at 60 min (9.97 (9.68, 10.26) ng/mL), and additionally its insulin trajectory was second only to that of the GDM Class 2 (Fig. 1B).
Fig. 1.
C-peptide trajectories for each class (A) identified by the LCTM approach and corresponding insulin (B) and glucose (C) responses. Curves are the estimated mean trajectories (solid line) and 95% CIs (shaded)
Despite similar glucose curves between the two GDM classes (Fig. 1C), GDM Class 2 had the highest peak and AUC values of C-peptide and insulin (Fig. 2), the greatest insulin resistance (highest HOMA-IR, lowest QUICKIi, QUICKIc and Matsuda indices; all P < 0.001), and predominantly impaired fasting glucose (48.1%) (Table 2). Conversely, GDM Class 1 had the lowest beta-cell secretion indicators (HOMA-B, IGIi, IGIc, and ISSI-2; all P < 0.001) and the highest proportion of impaired glucose tolerance (37.9%). Notably, non-GDM Class 2 presented peak and AUC values of C-peptide and insulin second only to GDM Class 2 (Fig. 2B–C, E–F) and had significantly higher HOMA-IR and lower QUICKIi, QUICKIc, and Matsuda index values compared to non-GDM Class 1 (Table 2), suggesting subclinical dysmetabolism. High correlation existed between blood glucose levels at 120 min with C-peptide and insulin levels at 120 min in the GDM Class 2 group (Additional file 1: Fig. S3).
Fig. 2.
Area under the curve for glucose (A), C-peptide (B), and insulin (C) levels, and the peak values for glucose (D), C-peptide (E), and insulin (F) levels. The classes were identified via latent class trajectory analysis
Table 2.
Glucose metabolism characteristics of the latent classes
| Non-GDM group | GDM group | ||||||
|---|---|---|---|---|---|---|---|
| Class 1 | Class 2 | P | Class 1 | Class 2 | P | P′ | |
| N | 794 (78.30) | 220 (21.70) | 771 (76.04) | 243 (23.96) | |||
| HOMA-IR | 1.15 (0.90, 1.59) | 1.49 (1.01, 2.18)ab | <0.001 | 2.07 (1.52, 2.86) | 2.87 (2.15, 3.83) | <0.001 | <0.001 |
| QUICKIi × 102 | 37.49 (35.60, 39.07) | 35.96 (33.95, 38.27)ab | <0.001 | 34.20 (32.65, 35.86) | 32.63 (31.34, 34.01) | <0.001 | <0.001 |
| QUICKIc × 102 | 48.99 (46.89, 51.32) | 46.46 (44.62, 48.66)ab | <0.001 | 45.44 (43.44, 47.87) | 43.07 (41.28, 45.47) | <0.001 | <0.001 |
| Matsuda index | 9.44 (6.67, 14.87) | 5.90 (4.43, 10.27)ab | <0.001 | 5.02 (3.81, 6.47) | 3.43 (2.54, 4.57) | <0.001 | <0.001 |
| HOMA-B | 118.32 (89.31, 161.77) | 143.46 (94.32, 220.04)a | <0.001 | 114.56 (88.34, 150.54) | 148.57 (117.37, 191.81) | <0.001 | <0.001 |
| IGIi | 10.74 (2.34, 15.50) | 18.71 (11.21, 27.17)ab | <0.001 | 9.66 (6.89, 13.64) | 20.52 (14.71, 28.28) | <0.001 | <0.001 |
| IGIc | 1.87 (1.42, 2.54) | 2.49 (1.89, 3.34)ab | <0.001 | 1.21 (0.92, 1.66) | 1.91 (1.53, 2.48) | <0.001 | <0.001 |
| ISSI-2 | 758.84 (619.54, 884.24) | 915.05 (716.65, 1096.17)ab | <0.001 | 521.58 (444.09, 596.98) | 629.16 (531.33, 712.69) | <0.001 | <0.001 |
| Types of blood glucose abnormalities (%) | |||||||
| Impaired fasting glucose (IFG) | - | - | - | 227 (35.9) | 117 (48.1) | 0.003 | - |
| Impaired glucose tolerance (IGT) | - | - | - | 292 (37.9) | 71 (29.2) | ||
| IFG+IGT | - | - | - | 202 (26.2) | 55 (22.6) | ||
Bolded P values indicate statistical significance (P＜0.05)
aStatistically significant difference between non-GDM Class 2 and GDM Class 1, P < 0.05
bStatistically significant difference between non-GDM Class 2 and GDM Class 2, P < 0.05. P′ represents the difference in outcome between the four subgroups
Cross-sectional associations with C-peptide pattern classes
Multivariate logistic regression revealed prepregnancy BMI as a common risk factor for non-GDM Class 2 (20.9%; OR = 1.75, 95% CI: 1.18–2.58) and both GDM groups, with the strongest association observed in GDM Class 2 (34.6%; OR = 3.42, 95% CI: 2.42–4.83). Additional risk factors included family history of diabetes (GDM Class 2: OR = 2.57; GDM Class 1: OR = 2.07), advanced maternal age for GDM Class 1 (30.1%; OR = 1.85), and history of previous cesarean delivery for GDM Class 2 (18.5%; OR = 1.99) (Table 3).
Table 3.
Demographic characteristics and predictors of potential categories
| Non-GDM Class 1 | Non-GDM Class 2 | GDM Class 1 | GDM Class 2 | ||||
|---|---|---|---|---|---|---|---|
| M(P25, P75)/N (%) | M(P25, P75)/N (%) | OR (95% CI) | M(P25, P75)/N (%) | OR (95% CI) | M(P25, P75)/N (%) | OR (95% CI) | |
| Maternal age (years, %)a | 29 (28, 32) | 29 (27, 32) | – | 32 (29, 35)e | – | 31 (28, 34) | – |
| <35 | 677 (85.3) | 188 (85.5) | Ref. | 539 (69.9) | Ref. | 199 (81.9) | Ref. |
| ≥35 | 117 (14.7)b | 32 (14.5)c | 0.89 (0.54, 1.45) | 232 (30.1)e | 1.85 (1.37, 2.50) | 44 (18.1) | 1.11 (0.71, 1.74) |
| Prepregnancy BMI (kg/m2)a | 20.6 (19.0, 22.5) | 21.1 (19.2, 23.6)d | – | 22.0 (20.0, 24.0) | – | 22.7 (21.1, 24.6) | – |
| BMI <24 | 693 (87.3) | 174 (79.1) | Ref. | 576 (74.8) | Ref. | 159 (65.4) | Ref. |
| BMI ≥24 | 101 (12.7)b | 46 (20.9)d | 1.75 (1.18, 2.58) | 195 (25.3)e | 2.13 (1.62, 2.80) | 84 (34.6) | 3.42 (2.42, 4.83) |
| Gestational weight gain (kg)a | 14.0 (11.5, 17.0) | 15.0 (12.0, 18.0)d | – | 12.0 (10.0, 15.0)e | – | 13.0 (10.0, 16.0) | – |
| Adequate | 376 (47.4) | 91 (41.4) | Ref. | 385 (49.9) | Ref. | 108 (44.4) | Ref. |
| Insufficient | 34 (4.3) | 4 (1.8) | 0.46 (0.16, 1.33) | 104 (13.5) | 2.70 (1.77, 4.13) | 20 (8.2) | 1.80 (0.98, 3.29) |
| Excessive | 384 (48.4)b | 125 (56.8)cd | 1.27 (0.93, 1.73) | 282 (36.6)e | 0.66 (0.53, 0.82) | 115 (47.3) | 0.90 (0.66, 1.22) |
| Graviditya | 2 (1, 2)b | 2 (1, 2) | 1.17 (0.87, 1.58) | 2 (1, 3) | 1.08 (0.89, 1.31) | 2 (1, 3) | 1.15 (0.88, 1.50) |
| Paritya | |||||||
| Primiparous | 546 (68.8) | 148 (67.3) | Ref. | 442 (57.3) | Ref. | 170 (70.0) | Ref. |
| Multiparous | 248 (31.2)b | 72 (32.7)c | 0.90 (0.53, 1.53) | 329 (42.7)e | 0.82 (0.58, 1.18) | 73 (30.0) | 0.43 (0.25, 0.75) |
| Previous cesarean deliverya | 103 (13.0)b | 31 (14.1)c | 1.04 (0.61, 1.77) | 170 (22.0) | 1.34 (0.95, 1.89) | 45 (18.5) | 1.99 (1.16, 3.39) |
| History of pregnancy lossa | 307 (38.7)b | 83 (37.7)cd | 0.73 (0.43, 1.24) | 413 (53.6) | 1.40 (0.99, 1.99) | 122 (50.2) | 1.30 (0.80, 2.12) |
| Family history of diabetesa | 33 (4.2)b | 16 (7.3) | 1.72 (0.92, 3.22) | 71 (9.2) | 2.07 (1.33, 3.23) | 26 (10.7) | 2.57 (1.48, 4.46) |
| Family history of hypertension | 87 (11.0) | 30 (13.6) | 1.19 (0.76, 1.87) | 113 (14.7) | 1.23 (0.89, 1.68) | 37 (15.2) | 1.28 (0.83, 1.97) |
The non-GDM Class 1 served as the reference group for the OR calculation
aThe difference among the four groups was statistically significant
bstatistically significant difference between non-GDM Class 1 and non-GDM Class 2
cstatistically significant difference between non-GDM Class 2 and GDM Class 1
dstatistically significant difference between non-GDM Class 2 and GDM Class 2
estatistically significant difference between GDM Class 1 and GDM Class 2, P < 0.05
C-peptide pattern classes and perinatal outcomes
According to the fully adjusted models, GDM Class 2 had a higher risk of delivering LGA (aOR = 1.51, 95% CI: 1.07–2.14) and macrosomia (aOR = 1.82, 95% CI: 1.13–2.95). Women in GDM Class 1 also presented an increased risk of delivering LGA in the crude model (OR = 1.30; 95% CI, 1.04–1.72), which attenuated to non-significance after adjustment (Table 4). Among maternal outcomes, non-GDM Class 2 had an increased risk of preeclampsia (OR = 2.85; 95% CI, 1.25–6.50) and GDM Class 2 had an increased risk of hyperthyroidism (OR = 12.13; 95% CI, 2.26–65.05) (Additional file 1: Table S8).
Table 4.
Multivariate analysis of latent C-peptide classes and important perinatal outcomes (ORs and 95% CIs)
| Crude model | Model 1 | Model 2 | Model 3 | |
|---|---|---|---|---|
| Adverse fetal outcome | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 0.91 (0.65, 1.26) | 0.89 (0.64, 1.24) | 0.90 (0.65, 1.26) | 0.99 (0.71, 1.39) |
| GDM, Class 1 | 1.15 (0.93, 1.42) | 1.11 (0.90, 1.38) | 1.07 (0.86, 1.33) | 1.19 (0.94, 1.51) |
| GDM, Class 2 | 1.36 (1.01, 1.84) | 1.29 (0.95, 1.75) | 1.31 (0.96, 1.77) | 1.22 (0.90, 1.65) |
| Prematurity | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 0.90 (0.63, 1.27) | 0.91 (0.64, 1.30) | 0.90 (0.64, 1.28) | 0.91 (0.64, 1.30) |
| GDM, Class 1 | 0.96 (0.76, 1.21) | 0.98 (0.78, 1.24) | 1.05 (0.83, 1.32) | 1.06 (0.84, 1.34) |
| GDM, Class 2 | 1.01 (0.73, 1.41) | 1.05 (0.75, 1.47) | 1.06 (0.76, 1.47) | 1.08 (0.77, 1.51) |
| LGA | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 1.05 (0.72, 1.55) | 1.02 (0.69, 1.51) | 1.05 (0.71, 1.55) | 1.02 (0.69, 1.51) |
| GDM, Class 1 | 1.34 (1.04, 1.72) | 1.28 (1.00, 1.65) | 1.18 (0.91, 1.52) | 1.14 (0.88, 1.47) |
| GDM, Class 2 | 1.75 (1.25, 2.44) | 1.61 (1.14, 2.27) | 1.62 (1.15, 2.27) | 1.51 (1.07, 2.14) |
| Macrosomia | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 1.17 (0.66, 2.07) | 1.14 (0.65, 2.02) | 1.18 (0.67, 2.08) | 1.14 (0.64, 2.02) |
| GDM, Class 1 | 1.10 (0.74, 1.62) | 1.06 (0.71, 1.57) | 1.04 (0.70, 1.54) | 1.01 (0.67, 1.49) |
| GDM, Class 2 | 2.04 (1.28, 3.27) | 1.91 (1.18, 3.09) | 1.95 (1.22, 3.13) | 1.82 (1.13, 2.95) |
| Fetal distress | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 0.33 (0.04, 2.53) | 0.32 (0.04, 2.48) | 0.33 (0.04, 2.57) | 0.32 (0.04, 2.50) |
| GDM, Class 1 | 1.41 (0.65, 3.10) | 1.37 (0.62, 3.04) | 1.59 (0.72, 3.53) | 1.54 (0.69, 3.44) |
| GDM, Class 2 | 0.59 (0.13, 2.68) | 0.56 (0.12, 2.60) | 0.61 (0.13, 2.79) | 0.57 (0.12, 2.64) |
| Gestational metabolic diseases | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 1.35 (0.92, 1.99) | 1.31 (0.89, 1.94) | 1.39 (0.92, 2.00) | 1.32 (0.90,1.94) |
| GDM, Class 1 | 1.12 (0.85, 1.47) | 1.07 (0.81, 1.41) | 1.13 (0.86, 1.48) | 1.08 (0.82,1.43) |
| GDM, Class 2 | 0.94 (0.62, 1.41) | 0.88 (0.58, 1.32) | 0.95 (0.63, 1.43) | 0.88 (0.58,1.34) |
| Hypertension | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 1.83 (0.73, 4.59) | 1.60 (0.63, 4.04) | 1.82 (0.73, 4.58) | 1.61 (0.64, 4.08) |
| GDM, Class 1 | 1.56 (0.79, 3.09) | 1.27 (0.63, 2.55) | 1.65 (0.83, 3.31) | 1.38 (0.68, 2.79) |
| GDM, Class 2 | 1.17 (0.42, 3.28) | 0.84 (0.29, 2.41) | 1.22 (0.43, 3.42) | 0.88 (0.31, 2.54) |
| Preeclampsia | ||||
| Non-GDM, Class 1 | 1 | 1 | 1 | 1 |
| Non-GDM, Class 2 | 3.16 (1.40, 7.16) | 2.71 (1.18, 6.23) | 3.18 (1.40, 7.21) | 2.85 (1.25, 6.50) |
| GDM, Class 1 | 1.68 (0.84, 3.38) | 1.23 (0.60, 2.55) | 1.54 (0.76, 3.15) | 1.30 (0.63, 2.69) |
| GDM, Class 2 | 1.26 (0.45, 3.58) | 0.83 (0.28, 2.43) | 1.27 (0.45, 3.60) | 0.94 (0.33, 2.73) |
Bolded OR (95% CI) values indicate statistical significance, defined as 95% CI not including 1
Model 1 was adjusted for prepregnancy BMI and family history of diabetes; Model 2 was adjusted for maternal age, history of previous cesarean delivery, and history of pregnancy loss; and Model 3 was adjusted for all the covariates adjusted for in Model 1 and Model 2
Abbreviations LGA Large for gestational age, SGA Small for gestational age, LBW Low birth weight
Discussion
Our study demonstrated a dynamic C-peptide profiling during OGTT, which can capture the heterogeneous pathophysiology of GDM. The latent class trajectory modeling revealed two major C-peptide trajectory patterns within GDM, despite similar blood glucose levels: a slow 120-min peak (Class 1) and a sharp 60-min peak with decline (Class 2). Class 1 was characterized by decreased insulin secretion, whereas Class 2 was characterized by pronounced insulin resistance with increased risks of delivering large for LGA and macrosomia. Interestingly, a subset of women without GDM also exhibited an early high-response C-peptide pattern similar to GDM with an increased risk of preeclampsia. Clinically, these findings have significant implications for personalized management.
Studies on the heterogeneity of GDM are emerging. However, these methods ignore measurement error, omit certain individual classifications, and fail to capture subtle group differences [12, 29]. Using a data-driven method, we observed different C-peptide trajectories, which may reflect differing balances between insulin secretion capacity and resistance. GDM Class 1, characterized by a slow-rise C-peptide trajectory, had high glucose levels throughout the OGTT, with relatively mild insulin and C-peptide responses. Despite only mild insulin resistance, women in this Class had the worst levels of insulin secretion indices. This pattern may indicate a suboptimal beta-cell compensatory response to the glucose load, reflecting a tendency toward insufficient insulin secretion. Such a profile may be related to limited beta-cell reserve or dysfunction in early-phase insulin release [30], which has been implicated in the development of postprandial hyperglycemia [31]. In contrast, the sharp and early C-peptide peak in GDM Class 2 aligned with compensatory hyperinsulinemia driven by profound insulin resistance, analogous to the physiology of type 2 diabetes [32]. Elevated HOMA-IR and reduced Matsuda indices in this group further support impaired peripheral glucose uptake [33, 34]. However, nearly half of the women had elevated fasting glucose levels, which may suggest that insulin resistance was severe enough to expose the fetus to higher ambient glucose.
Our findings indicated that GDM Class 2 had the strongest association with adverse fetal outcomes, especially LGA and macrosomia. These insights are consistent with previous studies linking maternal insulin resistance and hyperglycemia to fetal overgrowth [35, 36], supporting the biological plausibility of our findings. Maternal insulin resistance and hyperinsulinemia may upregulate placental glucose and amino acid transporters, leading to excess nutrient delivery to the fetus [37]. GDM Class1 demonstrated more modest risk associations after adjusting for key confounding factors, which was consistent with the findings of previous studies. Powe et al. [38] reported that GDM dominated by insulin resistance carried roughly double the risk of adverse outcomes relative to GDM dominated by insulin secretion. Similarly, a large prospective study in Europe stratified GDM by insulin sensitivity and showed that GDM with high insulin resistance had a worse metabolic profile and greater risk of complications than both normoglycemic women and GDM with low insulin resistance [39]. Indeed, a recent Korean study using latent profile analysis of OGTT glucose curves identified clusters but found no independent association with outcomes after adjusting for confounders [40]. The ability to demonstrate differences in adverse outcomes using C-peptide–based classification highlights the additional informational value of assessing insulin secretory responses rather than solely assessing glucose.
The recognition of this metabolic heterogeneity is essential for enhancing clinical management practices. Current clinical guidelines for GDM treatment often adopt a “one-size-fits-all” approach. Nutritional therapy is commonly initiated, and pharmacological intervention is chosen based on the persistence of hyperglycemia, without regard to metabolic heterogeneity [41]. However, our findings suggest that simply adding the testing of C-peptide during routine OGTT measurements could differentiate pathogenic mechanisms that direct therapeutic strategies. Patients with a high-risk, hyperinsulinemic phenotype (Class 2) should receive targeted interventions addressing insulin resistance, including healthy diets, physical activity, and weight loss to reduce glucose excursions [42]. Insulin sensitizers such as metformin may be more effective in pregnancies with poor glycemic control after lifestyle interventions. Given the increased risk of fetal overgrowth, serial ultrasound examinations are also recommended. Conversely, women in GDM Class 1 may respond better to strategies targeting impaired insulin secretion and improving postprandial hyperglycemia. Nutritional advice should focus on smaller, more frequent meals to manage postprandial hyperglycemia. As beta-cell function and insulin reserves may deteriorate as pregnancy progresses, continuous monitoring of blood glucose remains necessary and insulin support may be needed.
An interesting finding of our research was the identification of a subgroup of normoglycemic women exhibiting high C-peptide responses (non-GDM Class 2) with elevated insulin resistance [43]. Moreover, these women were at greater risk of developing preeclampsia, possibly owing to elevated levels of C-peptide and insulin, which increased vascular resistance and caused high blood pressure [44]. Although obesity is a common risk factor for insulin resistance and preeclampsia [45, 46], the association between non-GDM Class 2 and preeclampsia remained significant after adjusting for BMI, supporting prior evidence that insulin resistance independently predicts preeclampsia risk during pregnancy [47]. These findings highlight the limitation of relying solely on glucose thresholds for GDM diagnosis, underscoring a continuum of metabolic abnormalities extending below current diagnostic criteria. Therefore, we propose supplemental testing of C-peptide levels during the OGTT to aid in clinical precision. This option would not add extra pain to blood sampling in pregnant women and would be easy to implement. Even if the test is not available to all pregnant women, it is desirable to test C-peptide in women with the risk factors identified in this study, such as a family history of diabetes, advanced maternal age, multiparous women, and prepregnancy overweight and obesity.
Strengths and limitations
Our study has several strengths. First, this study used a dynamic model to depict individual C peptide trajectories, offering novel insights into the metabolic heterogeneity in GDM and helping identify distinct physiological patterns associated with adverse pregnancy outcomes. Specifically, GDM Class 1, a slow-rise trajectory, was characterized by impaired insulin secretion, whereas GDM class 2, an early-sharp rising trajectory, was characterized by marked insulin resistance with greater risks of adverse fetal outcomes. Additionally, a high-risk normoglycemic subgroup, which was at risk of abnormal insulin resistance and developing preeclampsia, was identified. These findings may provide a foundation for future studies exploring personalized management strategies to improve maternal and infant outcomes. Finally, and crucially determining its feasibility, the C peptide measurements in this study were obtained from the routine OGTT samples, which are a practical and non-invasive method that does not require additional blood draws.
Our study had several limitations. This study was a single ethnicity study conducted at one medical center in China, which may limit generalizability to other populations. Multicenter studies across diverse populations are needed to validate the results. In addition, this study evaluated only short-term follow-up outcomes, and differences between the classes may not have been fully captured. Further follow-up should be performed in the future to refine the long-term consequences of different C-peptide response patterns. Furthermore, although we explored the impact of a series of C-peptide indicators, the predictive value, feasibility, and cost-effectiveness of specific C-peptide indicators in guiding clinical management should be further evaluated in future studies. Finally, this study lacked detailed information on genetic inheritance and mechanisms, necessitating further exploration in future research.
Conclusions
C-peptide levels measured via the 2-h OGTT can disclose the heterogeneity of glucose metabolism between GDM and identify high-risk groups of pregnant women without GDM. This study demonstrated the inadequacy of glucose-centered diagnostic criteria and supported the potential value of incorporating dynamic C-peptide assessment into the OGTT in clinical practice.
Supplementary Information
Additional file 1: Additional Details on Calculation of the glucose composite indicators. Fig. S1. Flow chart. Table S1. Baseline Characteristics of the 2028 Participants. Fig. S2. Correlation analysis between 9 indicators of insulin, C-peptide and glucose in GDM and non-GDM. Table S2. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by maternal age. Table S3. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by pre-pregnancy BMI. Table S4. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by family history. Table S5. Model fit statistics in non-GDM. Table S6. Model fit statistics in GDM. Table S7. Descriptive statistics of class membership probabilities used in hard assignment (highest probability for each individual). Fig. S3. Correlation analysis between 9 indicators of insulin, C-peptide and glucose in latent C peptide pattern classes. Table S8. Multivariate analysis of latent C peptide pattern classes and secondary outcomes (OR and 95%).
Acknowledgements
The authors extend their gratitude to the participants and the research team at Northwest Women’s and Children’s Hospital for their dedicated involvement.
Abbreviations
- GDM
Gestational diabetes mellitus
- OGTT
Oral glucose tolerance test
- LCTM
Latent class trajectory modeling
- FPG
Fasting plasma glucose
- HOMA–IR
Homeostasis model assessment of insulin resistance
- HOMA–B
Homeostasis model assessment of beta–cell function
- QUICKIi
Quantitative insulin sensitivity check index for insulin
- QUICKIc
Quantitative insulin sensitivity check index for C-peptide
- IGIi
Modified insulinogenic indices for insulin
- IGIc
Modified insulinogenic indices for C-peptide
- ISSI–2
Insulin Secretion–Sensitivity Index–2
- GWG
Gestational weight gain
- LBW
Low birth weight
- SGA
Small for gestational age
- LGA
Large for gestational age
Authors’ contributions
XW performed the analysis. XW, ZH1, and SZ conceived the study, interpreted the results, and drafted the original manuscript. JL, JZ, and WY contributed to editing the manuscript. JJ, HY, ZH2 and YM critically reviewed the manuscript for important intellectual content. XL had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.
Funding
This research was funded by the National Natural Science Foundation of China (No. 81874263), the Key Special Project of the Development Center for Medical Science & Technology of the National Health Commission of the People’s Republic of China (W2015CAE060) and Social development in key industry innovation chains (2021ZDLSF02–14).
Data availability
The datasets generated during the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The study was approved by the Research Ethics Committee of Northwest Women and Children’s Hospital (NWCH 2012–013) and Xi’an Jiaotong University (XJTU2016–053). Written informed consent was obtained from all participants.
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.
Xinyue Wang and Zhangya He contributed equally to this work.
References
- 1.Hartling L, Dryden DM, Guthrie A, Muise M, Vandermeer B, Donovan L. Benefits and harms of treating gestational diabetes mellitus: a systematic review and meta-analysis for the U.S. Preventive Services Task Force and the National Institutes of Health Office of Medical Applications of Research. Ann Intern Med. 2013;159(2):123–9. [DOI] [PubMed] [Google Scholar]
- 2.Wang H, Li N, Chivese T, Werfalli M, Sun H, Yuen L, et al. IDF Diabetes Atlas: estimation of global and regional gestational diabetes mellitus prevalence for 2021 by International Association of Diabetes in Pregnancy Study Group’s criteria. Diabetes Res Clin Pract. 2022;183: 109050. [DOI] [PubMed] [Google Scholar]
- 3.Saravanan P. Diabetes in Pregnancy Working Group, Maternal Medicine Clinical Study Group, Royal College of Obstetricians and Gynaecologists UK. Gestational diabetes: opportunities for improving maternal and child health. Lancet Diabetes Endocrinol. 2020;8(9):793–800. [DOI] [PubMed] [Google Scholar]
- 4.Wambua S, Singh M, Okoth K, Snell KIE, Riley RD, Yau C, et al. Association between pregnancy-related complications and development of type 2 diabetes and hypertension in women: an umbrella review. BMC Med. 2024;22(1):66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Plows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH. The pathophysiology of gestational diabetes mellitus. Int J Mol Sci. 2018;19(11):3342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cundy T, Holt RIG. Gestational diabetes: paradigm lost? Diabet Med. 2017;34(1):8–13. [DOI] [PubMed] [Google Scholar]
- 7.Maddaloni E, Bolli GB, Frier BM, Little RR, Leslie RD, Pozzilli P, et al. C-peptide determination in the diagnosis of type of diabetes and its management: a clinical perspective. Diabetes Obes Metab. 2022;24(10):1912–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhao C, Liu H, Deng Y, Wu H, Wang S, Lyu X, et al. Maternal fasting serum C-peptide concentrations in the first and second trimesters and subsequent risk of gestational diabetes mellitus: a nested case-control study among Chinese women. Diabetes Res Clin Pract. 2024;208: 111111. [DOI] [PubMed] [Google Scholar]
- 9.Falcone V, Kotzaeridi G, Breil MH, Rosicky I, Stopp T, Yerlikaya-Schatten G, et al. Early assessment of the risk for gestational diabetes mellitus: can fasting parameters of glucose metabolism contribute to risk prediction? Diabetes Metab J. 2019;43(6):785–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Niknam A, Ramezani Tehrani F, Behboudi-Gandevani S, Rahmati M, Hedayati M, Abedini M, et al. Umbilical cord blood concentration of connecting peptide (C-peptide) and pregnancy outcomes. BMC Pregnancy Childbirth. 2022;22(1):764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lappas M, Jinks D, Ugoni A, Louizos CCJ, Permezel M, Georgiou HM, et al. Post-partum plasma C-peptide and ghrelin concentrations are predictive of type 2 diabetes in women with previous gestational diabetes mellitus. J Diabetes. 2015;7(4):506–11. [DOI] [PubMed] [Google Scholar]
- 12.Gong Y, Wang Q, Chen S, Liu Y, Li C, Kang R, et al. Heterogeneity of gestational diabetes and risk for adverse pregnancy outcome: a cohort study. J Clin Endocrinol Metab. 2025;110(7):e2264–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Shen Y, Zheng Y, Su Y, Jiang S, Ma X, Hu J, et al. Insulin sensitivity, β cell function, and adverse pregnancy outcomes in women with gestational diabetes. Chin Med J (Engl). 2022;135(21):2541–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Obstetrics Subgroup, Chinese Society of Obstetrics and Gynecology, Chinese Medical Association. Guideline of preconception and prenatal care (2018). Chin J Obstet Gynecol. 2018;53(1):7–13. Chinese. [DOI] [PubMed]
- 15.Hulman A, Witte DR, Vistisen D, Balkau B, Dekker JM, Herder C, et al. Pathophysiological characteristics underlying different glucose response curves: a latent class trajectory analysis from the prospective EGIR-RISC study. Diabetes Care. 2018;41(8):1740–8. [DOI] [PubMed] [Google Scholar]
- 16.Tricò D, McCollum S, Samuels S, Santoro N, Galderisi A, Groop L, et al. Mechanistic insights into the heterogeneity of glucose response classes in youths with obesity: a latent class trajectory approach. Diabetes Care. 2022;45(8):1841–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ji J, He Z, Qu P, Gao J, Zhang W, Wu P, et al. The Xi’an longitudinal mother–child cohort study: design, study population and methods. Eur J Epidemiol. 2021;36(2):223–32. [DOI] [PubMed] [Google Scholar]
- 18.International Association of Diabetes and Pregnancy Study Groups Consensus Panel, Metzger BE, Gabbe SG, Persson B, Buchanan TA, Catalano PA, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33(3):676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–9. [DOI] [PubMed] [Google Scholar]
- 20.Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care. 1999;22(9):1462–70. [DOI] [PubMed] [Google Scholar]
- 21.Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab. 2000;85(7):2402–10. [DOI] [PubMed] [Google Scholar]
- 22.Retnakaran R, Qi Y, Goran MI, Hamilton JK. Evaluation of proposed oral disposition index measures in relation to the actual disposition index. Diabet Med. 2009;26(12):1198–203. [DOI] [PubMed] [Google Scholar]
- 23.Mari A, Tura A, Natali A, Anderwald C, Balkau B, Lalic N, et al. Influence of hyperinsulinemia and insulin resistance on in vivo β-cell function: their role in human β-cell dysfunction. Diabetes. 2011;60(12):3141–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Parker SE, Carlson JM, Kebede N, Werler MM, Janulewicz PA. Pre-pregnancy body mass index and parent and teacher-reported behavioral outcomes among offspring in childhood. Neurotoxicol Teratol. 2022;89: 107049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.National Health Commission of the People's Republic of China. Standard of recommendation for weight gain during pregnancy period (WS/T 801—2022). Beijing: National Health Commission of China; 2022. Available from: https://www.nhc.gov.cn/wjw/c100311/202208/deb61e5c2299451ea1b957b0672272b3.shtml. Cited 2025 Jul 5.
- 26.National Health Commission of the People's Republic of China. Growth standard for newborns by gestational age (WS/T 800–2022). Beijing: National Health Commission of China; 2022. Available from: https://www.nhc.gov.cn/wjw/c100311/202208/deb61e5c2299451ea1b957b0672272b3.shtml . Cited 2025 Jul 5.
- 27.Wu P, Green M, Myers JE. Hypertensive disorders of pregnancy. BMJ. 2023;381: e071653. [DOI] [PubMed] [Google Scholar]
- 28.Stagnaro-Green A, Abalovich M, Alexander E, Azizi F, Mestman J, Negro R, et al. Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and postpartum. Thyroid. 2011;21(10):1081–125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Cosson E, Tatulashvili S, Vicaut E, Pinto S, Sal M, Nachtergaele C, et al. Glycemic status during pregnancy according to fasting and post-load glucose values: The association with adverse pregnancy outcomes. An observational study Diabetes Metab. 2023;49(5): 101469. [DOI] [PubMed] [Google Scholar]
- 30.Luppi P, Drain P. C-peptide antioxidant adaptive pathways in β cells and diabetes. J Intern Med. 2017;281(1):7–24. [DOI] [PubMed] [Google Scholar]
- 31.Lee MH, Febriana E, Lim M, Baig S, Shen L, Dalakoti M, et al. Performance of the 1 h oral glucose tolerance test in predicting type 2 diabetes and association with impaired β-cell function in Asians: a national prospective cohort study. Lancet Reg Health West Pac. 2024;54: 101278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Gedebjerg A, Bjerre M, Kjærgaard AD, Nielsen JS, Rungby J, Brandslund I, et al. CRP, C-peptide, and risk of first-time cardiovascular events and mortality in early type 2 diabetes: a Danish cohort study. Diabetes Care. 2023;46(5):1037–45. [DOI] [PubMed] [Google Scholar]
- 33.Claremeau F, Bale G, Lanthier N. Cirrhosis and insulin resistance: current knowledge, pathophysiological mechanisms, complications and potential treatments. Clin Sci (Lond). 2020;134(16):2117–35. [DOI] [PubMed] [Google Scholar]
- 34.Anoop S, Misra A, Bhatt SP, Gulati S, Mahajan H. High fasting C-peptide levels and insulin resistance in non-lean & non-obese (BMI >19 to <25 kg/m2) Asian Indians with type 2 diabetes are independently associated with high intra-abdominal fat and liver span. Diabetes Metab Syndr. 2019;13(1):708–15. [DOI] [PubMed] [Google Scholar]
- 35.Francis EC, Powe CE, Lowe WL Jr, White SL, Scholtens DM, Yang J, et al. Refining the diagnosis of gestational diabetes mellitus: a systematic review and meta-analysis. Commun Med (Lond). 2023;3(1):185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Liang X, Fu Y, Lu S, Shuai M, Miao Z, Gou W, et al. Continuous glucose monitoring-derived glycemic metrics and adverse pregnancy outcomes among women with gestational diabetes: a prospective cohort study. Lancet Reg Health West Pac. 2023;39: 100823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Liong S, Lappas M. Lipopolysaccharide and double stranded viral RNA mediate insulin resistance and increase system a amino acid transport in human trophoblast cells in vitro. Placenta. 2017;51:18–27. [DOI] [PubMed] [Google Scholar]
- 38.Powe CE, Allard C, Battista MC, Doyon M, Bouchard L, Ecker JL, et al. Heterogeneous contribution of insulin sensitivity and secretion defects to gestational diabetes mellitus. Diabetes Care. 2016;39(6):1052–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Benhalima K, Van Crombrugge P, Moyson C, Verhaeghe J, Vandeginste S, Verlaenen H, et al. Characteristics and pregnancy outcomes across gestational diabetes mellitus subtypes based on insulin resistance. Diabetologia. 2019;62(11):2118–28. [DOI] [PubMed] [Google Scholar]
- 40.Kim KJ, Kim NH, Choi J, Kim SG, Lee KJ. How can we adopt the glucose tolerance test to facilitate predicting pregnancy outcome in gestational diabetes mellitus? Endocrinol Metab (Seoul). 2021;36(5):988–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Obstetrics Subgroup, Chinese Society of Obstetrics and Gynecology, Chinese Medical Association, Chinese Society of Perinatal Medicine, Chinese Medical Association, Commitee of Pregnancy with Diabetes Mellitus, China Maternal and Child Health Association. Guideline of diagnosis and treatment of hyperglycemia in pregnancy (2022) [Part one]. Chin J Obstet Gynecol. 2022;57(1):3–12.
- 42.Mastrototaro L, Roden M. Insulin resistance and insulin sensitizing agents. Metabolism. 2021;125: 154892. [DOI] [PubMed] [Google Scholar]
- 43.Manco M, Panunzi S, Macfarlane DP, Golay A, Melander O, Konrad T, et al. One-hour plasma glucose identifies insulin resistance and beta-cell dysfunction in individuals with normal glucose tolerance: cross-sectional data from the Relationship between Insulin Sensitivity and Cardiovascular Risk (RISC) study. Diabetes Care. 2010;33(9):2090–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Wang Y, Wan H, Chen Y, Xia F, Zhang W, Wang C, et al. Association of C-peptide with diabetic vascular complications in type 2 diabetes. Diabetes Metab. 2020;46(1):33–40. [DOI] [PubMed] [Google Scholar]
- 45.Ferrara A, Hedderson MM, Brown SD, Ehrlich SF, Tsai AL, Feng J, et al. A telehealth lifestyle intervention to reduce excess gestational weight gain in pregnant women with overweight or obesity (GLOW): a randomised, parallel-group, controlled trial. Lancet Diabetes Endocrinol. 2020;8(6):490–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Borges MC, Clayton GL, Freathy RM, Felix JF, Fernández-Sanlés A, Soares AG, et al. Integrating multiple lines of evidence to assess the effects of maternal BMI on pregnancy and perinatal outcomes. BMC Med. 2024;22(1):32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.HHauth JC, Clifton RG, Roberts JM, Myatt L, Spong CY, Leveno KJ, et al. Maternal insulin resistance and preeclampsia. Am J Obstet Gynecol. 2011;204(4):327 e1–6. [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.
Supplementary Materials
Additional file 1: Additional Details on Calculation of the glucose composite indicators. Fig. S1. Flow chart. Table S1. Baseline Characteristics of the 2028 Participants. Fig. S2. Correlation analysis between 9 indicators of insulin, C-peptide and glucose in GDM and non-GDM. Table S2. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by maternal age. Table S3. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by pre-pregnancy BMI. Table S4. Comparing glucose metabolism indicator differences in GDM and non-GDM stratified by family history. Table S5. Model fit statistics in non-GDM. Table S6. Model fit statistics in GDM. Table S7. Descriptive statistics of class membership probabilities used in hard assignment (highest probability for each individual). Fig. S3. Correlation analysis between 9 indicators of insulin, C-peptide and glucose in latent C peptide pattern classes. Table S8. Multivariate analysis of latent C peptide pattern classes and secondary outcomes (OR and 95%).
Data Availability Statement
The datasets generated during the current study are available from the corresponding author upon reasonable request.