Abstract
Objective
The differential impact of pre‐pregnancy body mass index (BMI) on large for gestational age (LGA) risk by gestational diabetes mellitus (GDM) status remains unquantified. We aimed to assess whether GDM modifies the association between pre‐pregnancy BMI and LGA to inform precision prevention strategies.
Methods
In this prospective cohort study, 34 031 pregnant women enrolled at Hunan Provincial Maternal and Child Health Care Hospital (2013–2019) were stratified by GDM status. Multivariable logistic regression and restricted cubic spline (RCS) models adjusted for sociodemographic, behavioral, and clinical covariates evaluated the association between pre‐pregnancy BMI and LGA.
Results
LGA incidence was significantly higher in GDM than non‐GDM pregnancies (14.8% vs. 11.9%, P < 0.001). Adjusted models demonstrated a steeper BMI‐LGA dose–response gradient in GDM: each 1‐unit BMI increase conferred 20% higher odds (adjusted odds ratio [aOR]: 1.20, [95% confidence interval [CI]: 1.17, 1.23]) versus 11% in non‐GDM pregnancies (aOR: 1.11, [1.09, 1.12]). Obesity amplified LGA risk more substantially in GDM (aOR: 6.66, [4.27, 10.39]) than non‐GDM pregnancies (aOR: 2.78, [2.00, 3.87]). RCS models revealed nonlinear BMI‐LGA trajectories in both cohorts (reference level: 21.05 for GDM and 26.18 for non‐GDM). Associations remained consistent across demographic subgroups. Notably, significant interaction effects occurred exclusively in non‐GDM pregnancies, with multigravida (aOR: 1.13) exhibiting a stronger BMI‐LGA association than primigravida (aOR: 1.08).
Conclusion
GDM status significantly modifies the relationship between pre‐pregnancy BMI and LGA risk, identifying a threshold that substantially increases risk in women with GDM and revealing subgroup‐specific vulnerabilities. These findings suggest that incorporating pre‐pregnancy BMI and estimated risk of developing GDM into early risk stratification can help identify high‐risk pregnancies, thereby guiding targeted antenatal monitoring and individualized interventions of varying stringency to prevent LGA.
Keywords: gestational diabetes mellitus, large for gestational age, maternal metabolic health, pre‐pregnancy body mass index, prospective cohort, risk stratification
1. INTRODUCTION
Large for gestational age (LGA), defined as birth weight exceeding the 90th percentile for gestational age and sex, poses significant clinical and public health challenges. 1 Globally, LGA prevalence is increasing in parallel with rising maternal obesity and gestational diabetes mellitus (GDM) rates, with estimates ranging between 4% and 22% depending on population characteristics. 2 , 3 , 4 LGA births are strongly associated with immediate perinatal complications, including shoulder dystocia, birth trauma, and neonatal hypoglycemia. 5 Furthermore, LGA status has been linked to long‐term adverse health outcomes, increasing the risk of childhood obesity, metabolic syndrome, and cardiometabolic diseases later in life. 6 This imposes a substantial burden on healthcare systems, families, and individuals. Consequently, developing preventive strategies targeting modifiable risk factors to reduce LGA occurrence represents a critical public health priority.
Maternal pre‐pregnancy body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters) is a well‐established, potent predictor of fetal growth and birth weight outcomes. A prospective study involving 3172 pregnant women demonstrated that pre‐pregnancy overweight (BMI 25.0–29.9) and obesity (BMI ≥30.0) were associated with 1.7‐fold and 1.6‐fold increased risks of LGA infants, respectively, compared to a normal pre‐pregnancy BMI (18.5–24.9). 7 Conversely, pre‐pregnancy underweight status (BMI <18.5) was linked to a nearly 30% reduction in LGA risk. 8 The underlying mechanisms linking elevated pre‐pregnancy BMI to fetal overgrowth are complex but may involve dysregulation of glucose metabolism. Elevated maternal adiposity prior to pregnancy has been shown to exacerbate pregnancy‐induced insulin resistance, resulting in increased nutrient availability (particularly glucose and free fatty acids) to the fetus, thereby stimulating excessive fetal growth and adipogenesis. 9 However, the influence of maternal hyperglycemia itself, particularly GDM introduces significant heterogeneity. GDM represents a distinct pathophysiological state characterized by pronounced insulin resistance and β‐cell dysfunction, potentially amplifying the impact of pre‐pregnancy BMI on fetal growth. Conversely, emerging evidence suggests that a significant proportion of LGA cases occur in non‐GDM pregnancies, 10 indicating that the metabolic environment in normoglycemic pregnancies may differentially modulate the BMI‐LGA relationship. Furthermore, current clinical guidelines and research employ uniform BMI‐based management strategies to reduce adverse pregnancy outcomes, potentially overlooking the modifying effect of GDM status. This standardized approach raises critical questions regarding whether GDM status modifies the association between pre‐pregnancy BMI and LGA risk. Advances in precision medicine offer new opportunities for risk‐stratified pregnancy management. 7 A recent small‐scale retrospective cohort study suggested that among women with GDM, those with pre‐pregnancy overweight or obesity exhibited a markedly higher LGA risk compared to their normal‐weight counterparts. 11 Nevertheless, robust evidence supporting the clinical necessity and potential benefit of distinct, BMI‐stratified management protocols for LGA prevention, specifically within GDM versus non‐GDM populations, remains scarce. Rigorously comparing BMI‐associated LGA risks across glycemic strata can provide important evidence to support the development of precision‐based clinical guidelines.
Using a prospective first‐trimester pregnancy cohort, the aim of the present study was to characterize associations between pre‐pregnancy BMI and LGA, stratified by GDM diagnosis. The magnitude and direction of these BMI‐LGA relationships were assessed for differences between GDM and non‐GDM populations, including within key demographic subgroups. Findings may provide evidence supporting differentiated pre‐pregnancy BMI‐based management thresholds that could inform early risk stratification before GDM diagnosis, potentially contributing to the development of precision‐based pregnancy management approaches in the future. Given the rising global burden of obesity and GDM, addressing these critical knowledge gaps has important implications for maternal‐child health policy programs and clinical practice.
2. MATERIALS AND METHODS
2.1. Study design and participant enrollment
Conducted at Hunan Provincial Maternal and Child Health Care Hospital, China (March 13, 2013 to December 31, 2019), this prospective cohort study enrolled pregnant women aged ≥18 years. Eligible participants initiated prenatal care prior to 14 weeks' gestation and intended to receive ongoing care and deliver at the study hospital, and were limited to singleton pregnancies. Recruitment followed informed consent in relevant outpatient clinics (reproductive medicine, obstetrics, ultrasonography). Ethical approval was granted by the Ethics Committee of Xiangya School of Public Health, Central South University (approval no. XYGW‐2018‐36). Exclusion criteria included conception via assisted reproductive technologies (ART) to minimize confounding. Participants underwent a standardized 75 g oral glucose tolerance test (OGTT) at 24–28 weeks' gestation, adhering to International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria. Plasma glucose (fasting, 1‐h, 2‐h post‐load) was measured using an automated analyzer (Toshiba TBA‐120FR, Tokyo, Japan). GDM diagnosis required any value meeting thresholds: fasting ≥5.1 mmol/L, 1‐h ≥10.0 mmol/L, or 2‐h ≥8.5 mmol/L. Following exclusions (Figure 1), 34 031 women constituted the final analytical cohort. This study followed the STROBE reporting guidelines.
FIGURE 1.
Flow chart showing the study design and the inclusion of participants. GDM, gestational diabetes mellitus; LGA, large for gestational age.
2.2. Data collection
Early‐pregnancy (<14 weeks) sociodemographic, behavioral, and lifestyle data were collected using a validated structured questionnaire, administered by trained interviewers through in‐person interviews and telephone follow‐ups. The questionnaire, which has been applied in other studies, 12 included variables such as maternal age, residence (urban/rural), ethnicity (Han Chinese/minority groups), parity, gravidity, pre‐pregnancy BMI, education level, monthly household income per capita, smoking status, alcohol intake, and folic acid supplementation. Self‐reported information was cross‐referenced with electronic medical records (EMRs) to ensure accuracy. The hospital's EMR system served as the source for pregnancy complications, clinical measurements, and delivery outcomes. Independent quality control teams implemented duplicate data entry and conducted periodic audits to address inconsistencies.
2.3. Exposures and covariates
Pre‐pregnancy BMI served as the primary exposure, calculated from self‐reported values of weight and height before conception by participants at the first prenatal visit. Maternal pre‐pregnancy BMI was categorized according to WHO criteria 1 : underweight (<18.5), normal (18.5–24.9), overweight (25.0–29.9), and obese (≥30.0). Covariates included maternal age (calculated from national identification card birthdates), ethnicity, education level (junior high school or below, high school/technical secondary school, college degree, bachelor's degree or above), residence, gravidity (1 vs. ≥2 times), parity (primiparous/multiparous), monthly household income per capita (≤CNY 2500, CNY 2501 to 5000, >CNY 5000), smoking status (active [self‐reported tobacco use] or passive [≥1 day/week secondhand smoke exposure]), 13 alcohol intake (any consumption during early pregnancy), 13 folic acid supplementation (self‐reported whether folic acid was used before or during pregnancy), 13 gestational weight gain, gestational age at delivery, and infant sex (male/female).
2.4. Outcomes
Based on the weight distribution curves of newborns of different gestational ages obtained from the national birth defect monitoring system as a reference standard, LGA was defined as birth weight exceeding the 90th percentile for gestational age and sex. 14 Birth weight and neonatal outcomes were extracted from EMRs and cross‐checked by obstetricians blinded to exposure status.
2.5. Statistical analysis
Data management was performed using EpiData 3.1 with double‐entry verification. Participants were stratified according to gestational diabetes mellitus (GDM) status. Continuous normally distributed variables are presented as mean (standard deviation [SD]), while categorical variables are expressed as frequency (percentage). Between‐group comparisons across pre‐pregnancy BMI categories were performed using analysis of variance for continuous variables and χ 2 or Fisher exact tests for categorical variables. Trends in LGA incidence by pre‐pregnancy BMI categories were assessed using the Mantel–Haenszel χ 2 test. Multivariable logistic regression models were used to estimate associations between pre‐pregnancy BMI (per 1‐unit increase and categorically) and LGA, stratified by GDM status, with results reported as adjusted odds ratios (aORs) and 95% confidence intervals (CIs). Restricted cubic spline (RCS) regression with three knots was employed to evaluate potential nonlinear relationships between pre‐pregnancy BMI and LGA, the reference BMI for each group was defined as the pre‐pregnancy BMI value at which the model‐estimated odds ratio (OR) for LGA equalled 1, where an OR = 1 indicates no association between pre‐pregnancy BMI and LGA risk, an OR >1 indicates an increased risk of LGA associated with pre‐pregnancy BMI, and an OR <1 indicates a decreased risk of LGA associated with pre‐pregnancy BMI. Subgroup analyses stratified by maternal age, gravidity, parity, and infant sex were conducted to assess the consistency of associations. Interaction terms between pre‐pregnancy BMI and stratification variables were formally tested.
3. RESULTS
This first‐trimester pregnancy cohort comprised 34 031 participants. Maternal age averaged 32.43 years (± 4.52 SD), and mean pre‐pregnancy BMI was 22.20 (± 3.16 SD). GDM was diagnosed in 5414 women (15.9%), with the remaining 28 617 (84.1%) serving as the non‐GDM comparator group. Large‐for‐gestational‐age (LGA) neonates occurred in 4214 infants (12.4% overall). LGA incidence was significantly higher in pregnancies with GDM (802 cases, 14.8%) than in those without GDM (3412 cases, 11.9%; P < 0.001). A significant gradient in LGA incidence was observed across pre‐pregnancy BMI categories (Figure 2). Rates increased progressively from 7.8% (GDM) and 8.6% (non‐GDM) among underweight women to 39.0% (GDM) and 21.5% (non‐GDM) in obese women (P for trend <0.001).
FIGURE 2.
Incidence of large for gestational age according to the pre‐pregnancy BMI categories. BMI, body mass index; GDM, gestational diabetes mellitus; LGA, large for gestational age.
3.1. Baseline characteristics stratified by pre‐pregnancy BMI and GDM status
Table 1 details the baseline characteristics of GDM and non‐GDM participants, stratified by pre‐pregnancy BMI categories. Pre‐pregnancy BMI differed significantly across categories within both the GDM and non‐GDM groups (P < 0.001). Among GDM pregnancies, maternal age (P < 0.001), advanced maternal age (P < 0.001), gravidity (P < 0.001), parity (P < 0.001), ethnicity (P < 0.001), education level (P = 0.002), and gestational weight gain (P < 0.001) varied significantly across BMI categories. In non‐GDM pregnancies, significant variations occurred across BMI strata for maternal age (P < 0.001), advanced maternal age (P < 0.001), gravidity, parity (P < 0.001), education level (P < 0.001), gestational weight gain (P < 0.001), smoking status (P = 0.042), infant sex (P < 0.001), and gestational age at delivery (P = 0.006). Both cohorts exhibited progressively higher infant birth weights with increasing pre‐pregnancy BMI (P < 0.001).
TABLE 1.
Baseline characteristics of GDM and non‐GDM patients grouped according to pre‐pregnancy BMI categories.
| Characteristics | GDM | Non‐GDM | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Overall (N = 5414) | Underweight (N = 448) | Normal weight (N = 4090) | Overweight (N = 771) | Obesity (N = 105) | P value | Overall (N = 28 617) | Underweight (N = 4459) | Normal weight (N = 21 764) | Overweight (N = 2152) | Obesity (N = 242) | P value | |
| Age at delivery (years), mean (SD) | 32.43 (4.52) | 30.24 (3.97) | 32.63 (4.49) | 32.77 (4.50) | 31.74 (5.41) | < 0.001 | 30.87 (4.47) | 29.23 (4.09) | 31.11 (4.45) | 31.62 (4.73) | 32.63 (3.88) | < 0.001 |
| Advanced maternal age (≥35 years), n (%) | < 0.001 | < 0.001 | ||||||||||
| No | 3695 (68.2) | 385 (85.9) | 2720 (66.5) | 512 (66.4) | 78 (74.3) | 22 687 (79.3) | 3980 (89.3) | 16 979 (78.0) | 1568 (72.9) | 160 (66.1) | ||
| Yes | 1719 (31.8) | 63 (14.1) | 1370 (33.5) | 259 (33.6) | 27 (25.7) | 5930 (20.7) | 479 (10.7) | 4785 (22.0) | 584 (27.1) | 82 (33.9) | ||
| Residence, n (%) | 0.270 | 0.295 | ||||||||||
| Urban | 3429 (63.3) | 299 (66.7) | 2592 (63.4) | 471 (61.1) | 67 (63.8) | 17 598 (61.5) | 2768 (62.1) | 13 339 (61.3) | 1330 (61.8) | 161 (66.5) | ||
| Rural | 1985 (36.7) | 149 (33.3) | 1498 (36.6) | 300 (38.9) | 38 (36.2) | 11 019 (38.5) | 1691 (37.9) | 8425 (38.7) | 822 (38.2) | 81 (33.5) | ||
| Gravidity (times), n (%) | < 0.001 | < 0.001 | ||||||||||
| 1 | 1477 (27.3) | 191 (42.6) | 1056 (25.8) | 196 (25.4) | 34 (32.4) | 9000 (31.4) | 1795 (40.3) | 6674 (30.7) | 493 (22.9) | 38 (15.7) | ||
| ≥2 | 3937 (72.7) | 257 (57.4) | 3034 (74.2) | 575 (74.6) | 71 (67.6) | 19 617 (68.6) | 2664 (59.7) | 15 090 (69.3) | 1659 (77.1) | 204 (84.3) | ||
| Parity, n (%) | < 0.001 | < 0.001 | ||||||||||
| Primiparous | 2481 (45.8) | 316 (70.5) | 1773 (43.3) | 342 (44.4) | 50 (47.6) | 13 926 (48.7) | 2660 (59.7) | 10 357 (47.6) | 817 (38.0) | 92 (38.0) | ||
| Multiparous | 2933 (54.2) | 132 (29.5) | 2317 (56.7) | 429 (55.6) | 55 (52.4) | 14 691 (51.3) | 1799 (40.3) | 11 407 (52.4) | 1335 (62.0) | 150 (62.0) | ||
| Ethnicity, n (%) | < 0.001 | 0.391 | ||||||||||
| Han | 5339 (98.6) | 448 (100) | 4032 (98.6) | 765 (99.2) | 94 (89.5) | 28 244 (98.7) | 4410 (98.9) | 21 471 (98.7) | 2126 (98.8) | 237 (97.9) | ||
| Minority | 75 (1.4) | 0 (0) | 58 (1.4) | 6 (0.8) | 11 (10.5) | 373 (1.3) | 49 (1.1) | 293 (1.3) | 26 (1.2) | 5 (2.1) | ||
| Education level, n (%) | 0.002 | < 0.001 | ||||||||||
| Junior high and below | 417 (7.7) | 21 (4.7) | 318 (7.8) | 63 (8.2) | 15 (14.3) | 2116 (7.4) | 346 (7.8) | 1475 (6.8) | 242 (11.2) | 53 (21.9) | ||
| High school or technical secondary school | 1448 (26.7) | 123 (27.5) | 1065 (26.0) | 220 (28.5) | 40 (38.1) | 8227 (28.7) | 1215 (27.2) | 6133 (28.2) | 785 (36.5) | 94 (38.8) | ||
| College degree | 2597 (48) | 216 (48.2) | 1985 (48.5) | 362 (47) | 34 (32.4) | 12 963 (45.3) | 2170 (48.7) | 9903 (45.5) | 827 (38.4) | 63 (26) | ||
| Bachelor's degree or above | 952 (17.6) | 88 (19.6) | 722 (17.7) | 126 (16.3) | 16 (15.2) | 5311 (18.6) | 728 (16.3) | 4253 (19.5) | 298 (13.8) | 32 (13.2) | ||
| Per‐caput monthly family income (CNY), n (%) | 0.541 | 0.416 | ||||||||||
| ≤2500 | 944 (17.4) | 75 (16.7) | 716 (17.5) | 136 (17.6) | 17 (16.2) | 4936 (17.2) | 748 (16.8) | 3765 (17.3) | 383 (17.8) | 40 (16.5) | ||
| 2500 to 5000 | 2860 (52.8) | 242 (54.0) | 2170 (53.1) | 386 (50.1) | 62 (59.0) | 15 300 (53.5) | 2421 (54.3) | 11 603 (53.3) | 1158 (53.8) | 118 (48.8) | ||
| >5000 | 1610 (29.7) | 131 (29.2) | 1204 (29.4) | 249 (32.3) | 26 (24.8) | 8381 (29.3) | 1290 (28.9) | 6396 (29.4) | 611 (28.4) | 84 (34.7) | ||
| Alcohol intake, n (%) | 0.054 | 0.091 | ||||||||||
| No | 5323 (98.3) | 437 (97.5) | 4016 (98.2) | 765 (99.2) | 105 (100) | 28 210 (98.6) | 4413 (99) | 21 438 (98.5) | 2122 (98.6) | 237 (97.9) | ||
| Yes | 91 (1.7) | 11 (2.5) | 74 (1.8) | 6 (0.8) | 0 (0) | 407 (1.4) | 46 (1) | 326 (1.5) | 30 (1.4) | 5 (2.1) | ||
| Smoking, n (%) | 0.654 | 0.042 | ||||||||||
| No | 4991 (92.2) | 409 (91.3) | 3770 (92.2) | 717 (93.0) | 95 (90.5) | 26 267 (91.8) | 4050 (90.8) | 20 000 (91.9) | 1995 (92.7) | 222 (91.7) | ||
| Yes | 423 (7.8) | 39 (8.7) | 320 (7.8) | 54 (7.0) | 10 (9.5) | 2350 (8.2) | 409 (9.2) | 1764 (8.1) | 157 (7.3) | 20 (8.3) | ||
| Folic acid supplementation, n (%) | 0.588 | 0.333 | ||||||||||
| No | 236 (4.4) | 21 (4.7) | 182 (4.4) | 31 (4.0) | 2 (1.9) | 1308 (4.6) | 191 (4.3) | 1018 (4.7) | 92 (4.3) | 7 (2.9) | ||
| Yes | 5178 (95.6) | 427 (95.3) | 3908 (95.6) | 740 (96.0) | 103 (98.1) | 27 309 (95.4) | 4268 (95.7) | 20 746 (95.3) | 2060 (95.7) | 235 (97.1) | ||
| Infant sex, n (%) | 0.313 | < 0.001 | ||||||||||
| Male | 2732 (50.5) | 211 (47.1) | 2090 (51.1) | 382 (49.5) | 49 (46.7) | 15 180 (53.0) | 2406 (54.0) | 11 407 (52.4) | 1193 (55.4) | 174 (71.9) | ||
| Female | 2682 (49.5) | 237 (52.9) | 2000 (48.9) | 389 (50.5) | 56 (53.3) | 13 437 (47.0) | 2053 (46.0) | 10 357 (47.6) | 959 (44.6) | 68 (28.1) | ||
| Gestational age at delivery (weeks), mean (SD) | 38.45 (2.14) | 38.24 (2.30) | 38.45 (2.11) | 38.56 (2.15) | 38.58 (2.25) | 0.078 | 38.43 (2.12) | 38.34 (2.21) | 38.45 (2.09) | 38.36 (2.18) | 38.55 (1.98) | 0.006 |
| Gestational weight gain (g), mean (SD) | 12.32 (4.69) | 13.86 (4.52) | 12.49 (4.59) | 10.91 (4.76) | 9.68 (5.40) | < 0.001 | 13.81 (4.72) | 14.62 (4.69) | 13.87 (4.61) | 11.81 (5.01) | 11.43 (6.23) | < 0.001 |
| Birth weight (g), mean (SD) | 3160.32 (594.28) | 2974.71 (572.86) | 3157.33 (577.48) | 3238.60 (661.75) | 3493.86 (557.46) | < 0.001 | 3142.46 (544.24) | 3041.11 (517.02) | 3159.2 (535.17) | 3169.17 (641.18) | 3267.64 (674.39) | < 0.001 |
Note: BMI, calculated as weight in kilograms divided by the square of height in meters.
Abbreviations: BMI, body mass index; GDM, gestational diabetes mellitus; SD, standard deviation.
3.2. Association between pre‐pregnancy BMI and LGA
A significant positive association between pre‐pregnancy BMI and LGA was confirmed in both cohorts using logistic regression (Table 2). Unadjusted models indicated that each 1‐unit BMI increase elevated LGA odds by 15% in GDM pregnancies (OR 1.15; 95% CI: 1.12, 1.17; P < 0.001) and by 7% in non‐GDM pregnancies (OR 1.07; 95% CI: 1.06, 1.08; P < 0.001). Following comprehensive adjustment for covariates, these associations persisted and intensified, particularly among GDM participants (aOR 1.20; 95% CI: 1.17, 1.23; P < 0.001) compared to non‐GDM participants (aOR 1.11; 95% CI: 1.09, 1.12; P < 0.001). Categorical analysis demonstrated distinct risk profiles. Among GDM pregnancies with full adjustment, pre‐pregnancy obesity conferred a 6.66‐fold higher LGA risk versus normal BMI (aOR 6.66; 95% CI: 4.27, 10.39; P < 0.001), while overweight status yielded a 2.82‐fold increase (aOR 2.82; 95% CI: 2.31, 3.45; P < 0.001). In non‐GDM pregnancies, obesity and overweight were associated with 2.78‐fold (aOR 2.78; 95% CI: 2.00, 3.87; P < 0.001) and 1.59‐fold (aOR 1.59; 95% CI: 1.40, 1.82; P < 0.001) elevated risks, respectively. Conversely, underweight women exhibited significantly reduced LGA risk in both groups (GDM: aOR 0.43; 95% CI: 0.29, 0.62; P < 0.001; non‐GDM: aOR 0.57; 95% CI: 0.51, 0.65; P < 0.001).
TABLE 2.
Odds ratio for large for gestational age associated with severity of pre‐pregnancy BMI among GDM and non‐GDM pregnancies.
| Categories | Events (%) | Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P value | P for trend | OR (95% CI) | P value | P for trend | OR (95% CI) | P‐value | P for trend | ||
| GDM | ||||||||||
| Continuous variable per 1 unit | 1.15 (1.12, 1.17) | < 0.001 | 1.15 (1.12, 1.17) | < 0.001 | 1.20 (1.17, 1.23) | < 0.001 | ||||
| Category | 802 (14.8) | < 0.001 | < 0.001 | < 0.001 | ||||||
| Underweight (N = 448) | 35 (7.8) | 0.56 (0.39, 0.8) | 0.002 | 0.57 (0.39, 0.81) | 0.002 | 0.43 (0.29, 0.62) | < 0.001 | |||
| Normal weight (N = 4090) | 536 (13.1) | Ref. | Ref. | Ref. | ||||||
| Overweight (N = 771) | 190 (21.6) | 2.17 (1.8, 2.62) | < 0.001 | 2.19 (1.81, 2.64) | < 0.001 | 2.82 (2.31, 3.45) | < 0.001 | |||
| Obesity (N = 105) | 41 (39.0) | 4.25 (2.84, 6.35) | < 0.001 | 4.10 (2.73, 6.17) | < 0.001 | 6.66 (4.27, 10.39) | < 0.001 | |||
| Non‐GDM | ||||||||||
| Continuous variable per 1 unit | 1.07 (1.06, 1.08) | < 0.001 | 1.08 (1.06, 1.09) | < 0.001 | 1.11 (1.09, 1.12) | < 0.001 | ||||
| Category | 3412 (11.9) | < 0.001 | < 0.001 | < 0.001 | ||||||
| Underweight (N = 4459) | 383 (8.6) | 0.68 (0.61, 0.76) | < 0.001 | 0.65 (0.58, 0.73) | 0.001 | 0.57 (0.51, 0.65) | < 0.001 | |||
| Normal weight (N = 21 764) | 2651 (12.2) | Ref. | Ref. | Ref. | ||||||
| Overweight (N = 2152) | 326 (15.1) | 1.29 (1.14, 1.46) | < 0.001 | 1.33 (1.18, 1.51) | < 0.001 | 1.59 (1.40, 1.82) | < 0.001 | |||
| Obesity (N = 242) | 52 (21.5) | 1.97 (1.45, 2.69) | < 0.001 | 2.14 (1.57, 2.93) | < 0.001 | 2.78 (2.00, 3.87) | < 0.001 | |||
Note: BMI, calculated as weight in kilograms divided by the square of height in meters. Model 1: unadjusted. Model 2: adjusted for maternal age, residence location, gravidity, parity, per caput monthly family income, ethnicity, and education level. Model 3: adjusted for maternal age, residence location, gravidity, parity, per caput monthly family income, ethnicity, education level, smoking, alcohol intake, folic acid supplementation, gestational weight gain, delivery gestational age in the subsequent pregnancy, and infant sex.
Abbreviations: BMI, body mass index; CI, confidence interval; GDM, gestational diabetes mellitus; OR, odds ratio.
RCS models revealed non‐linear relationships between pre‐pregnancy BMI and LGA risk (Figure 3). In GDM pregnancies, the curve exhibited minimal variation within the lower BMI range and rose more steeply from the overweight range onward. In non‐GDM pregnancies, the risk curve was less pronounced but remained significant (P for nonlinearity <0.001), indicating a non‐linear association between pre‐pregnancy BMI and LGA. Specifically, risk elevation occurred when pre‐pregnancy BMI exceeds 26.18, and even modest further elevations increase the risk of fetal overgrowth in euglycemic pregnancies.
FIGURE 3.
Restricted cubic spline of the association between pre‐pregnancy BMI and large for gestational age among GDM and non‐GDM pregnancies. Heavy central lines represent the estimated adjusted odd ratios, with shaded ribbons denoting 95% CIs. Pre‐pregnancy BMI 21.05 and 26.18 were selected as the reference level for GDM and non‐GDM pregnancies, respectively. Pre‐pregnancy BMI 18.5, 25.0, and 30.0 were categorized based on the WHO criteria represented by the vertical dotted lines. The horizontal dotted lines represent the odds ratio of 1.0. The model was adjusted for maternal age, residence location, gravidity, parity, per caput monthly family income, ethnicity, education level, smoking, alcohol intake, folic acid supplementation, gestational weight gain, delivery gestational age in the subsequent pregnancy, and infant sex. BMI, body mass index; CI, confidence interval; GDM, gestational diabetes mellitus; OR, odds ratio.
3.3. Subgroup and interaction analyses
Stratified analyses evaluated the relationship between pre‐pregnancy BMI and LGA across maternal age, gravidity, parity, and infant sex subgroups (Figure 4). Elevated pre‐pregnancy BMI significantly increased LGA risk across all subgroups in both GDM and non‐GDM cohorts (P < 0.05). Significant effect modification emerged exclusively in non‐GDM pregnancies, with a stronger association observed in multigravida women (OR 1.13 vs. primigravida: 1.08; P for interaction = 0.001). These results collectively demonstrate context‐dependent risk patterns, supporting personalized preventive strategies for high‐risk subgroups.
FIGURE 4.
Forest plots of odds ratios for large for gestational age in different subgroups among GDM and non‐GDM pregnancies. CI, confidence interval; GDM, gestational diabetes mellitus; OR, odds ratio.
4. DISCUSSION
This large prospective cohort study provides compelling evidence that the relationship between pre‐pregnancy BMI and LGA risk is significantly modified by maternal glycemic status. Statistically significant differences in pre‐pregnancy BMI and gestational weight gain across BMI categories were expected in both GDM and non‐GDM groups, and each cohort exhibited a pronounced positive dose–response relationship between elevated pre‐pregnancy BMI and LGA. Women with overweight (aOR: 2.82 for GDM; 1.59 for non‐GDM) or obesity (aOR: 6.66 for GDM; 2.78 for non‐GDM) pre‐pregnancy demonstrated significantly elevated risks of delivering LGA infants compared to those with normal BMI. Lower pre‐pregnancy BMI categories were associated with a protective effect against LGA (aOR: 0.43 for GDM; 0.57 for non‐GDM). Restricted cubic spline (RCS) modeling revealed nonlinear risk trajectories, and both GDM and non‐GDM pregnancies exhibited a significant association between BMI and LGA risk. By comparing odds ratios and the BMI thresholds (ref BMI: 21.05 for GDM; 26.18 for non‐GDM) at which risk elevation occurred, the analysis showed a stronger effect in GDM pregnancies. Collectively, these results support establishing distinct glycemic status‐specific pre‐pregnancy BMI thresholds for targeted LGA prevention, thereby facilitating early risk stratification of high‐risk individuals before GDM is identified. Based on this stratification, individualized interventions can be implemented with varying degrees of stringency, including weight management, blood glucose monitoring, and pregnancy follow‐up, among others.
The differential risk gradients observed between GDM and non‐GDM populations challenge the current paradigm of uniform BMI‐based prenatal management. Although the universal association of pre‐pregnancy overweight/obesity with fetal overgrowth is well‐established, 15 the substantially steeper risk escalation quantified in GDM pregnancies underscores how maternal hyperglycemia and insulin resistance potentiate adiposity‐driven pathways. While the underlying mechanisms have not been fully elucidated, this association appears biologically plausible. Evidence suggests that inflammatory cytokines derived from adipose tissue, such as leptin and TNF‐α, may exacerbate placental nutrient transport dysregulation and fetal hyperinsulinemia, 16 , 17 representing an inherent amplification process within GDM. Crucially, the present findings extend prior evidence by demonstrating that elevated pre‐pregnancy BMI remains an independent and significant driver of LGA risk even in non‐GDM pregnancies (aOR 1.11 per unit BMI increase). This suggests contributions from subclinical insulin resistance, inflammatory pathways, or other adiposity‐related mechanisms beyond overt hyperglycemia. 18 Early in pregnancy (prior to GDM diagnosis), fetuses of overweight/obese women have consistently exhibited above‐average values across all ultrasound biometric parameters, indicating that maternal overweight/obesity is directly associated with fetal growth acceleration independently of GDM status. 19 This observation challenges the notion that the BMI‐LGA association is primarily mediated by hyperglycemia and highlights underrecognized risks in pregnancies classified as “healthy.” The nonlinear risk curves quantified by RCS modeling further distinguish this study. The identification of a specific BMI threshold (≥21.05) for accelerated LGA risk exclusively in GDM pregnancies represents a key novel contribution. This implies that the current WHO BMI categories, optimized for general pregnant populations, may underestimate risk thresholds in metabolically compromised women. Consequently, integrating continuous BMI metrics with glycemic parameters for precision risk prediction is supported. Moreover, the persistent risk elevation observed in non‐GDM women with overweight/obesity emphasizes that universal weight management remains a public health imperative. Consistent with prior reports, 8 , 20 a strong negative association between low pre‐pregnancy BMI and LGA was also demonstrated, suggesting that reduced substrate availability (glucose, lipids) may limit fetal overgrowth. 21 However, while confirming the protective effect of low pre‐pregnancy BMI against LGA, these findings necessitate cautious interpretation. Despite adjustment for potential confounders, the observed protection does not preclude an increased susceptibility to intrauterine growth restriction or micronutrient deficiencies among severely underweight women, particularly in resource‐limited settings. 21 , 22 Future guidelines should, therefore, advocate for achieving an evidence‐based optimal pre‐pregnancy BMI range rather than merely promoting weight reduction below standard thresholds.
Subgroup analyses reinforced the context‐dependency of the BMI‐LGA relationship and identified potential modifiers. Significant effect modification was observed exclusively within the non‐GDM cohort, where multigravidity amplified the association between pre‐pregnancy BMI and LGA risk (multigravida aOR: 1.13 per unit BMI vs. primigravida aOR: 1.08). This suggests that cumulative metabolic stressors from prior pregnancies, or associated sociodemographic factors influencing weight retention, 23 , 24 may exacerbate the impact of pre‐pregnancy adiposity on fetal growth in the absence of GDM. Conversely, within the GDM cohort, elevated pre‐pregnancy BMI consistently increased LGA risk across all examined demographic subgroups (age, gravidity, parity, infant sex), without statistically significant interactions. This uniformity underscores the dominant role of the dysmetabolic state in GDM, wherein hyperglycemia and insulin resistance may override the modifying influence of other factors. In non‐GDM women, glycemic values tend to vary with BMI, meaning that stratification alone cannot fully account for the influence of glycaemia on the BMI–LGA relationship. The results showed that only certain variables exhibited significant effect modification in this group, suggesting that the nuanced contributions of subclinical hyperglycemia and insulin resistance to LGA risk in this group remain incompletely understood and warrant further investigation. While the lack of significant interaction within GDM subgroups simplifies risk messaging for this high‐risk population, it also highlights the pervasive impact of the underlying pathophysiology. It should be emphasized that these findings, derived from observational data, should not be interpreted as causal pathways; further research is warranted to explore potential biological mechanisms driving the observed effect modification in non‐GDM pregnancies.
Several limitations warrant consideration. First, despite extensive covariate adjustment, residual confounding from unmeasured factors, such as detailed dietary habits, physical activity, precise gestational weight gain patterns, and psychosocial stressors, cannot be excluded. Second, pre‐pregnancy anthropometrics relied on self‐report. While self‐reported data show reasonable validity, 25 systematic weight underestimation in pregnancy is well‐documented. Third, generalizability is limited by the predominantly Han Chinese cohort; external validation across diverse ethnicities is essential given known ethnic differences in adiposity‐BMI correlations and GDM mechanisms. Fourth, excluding ART pregnancies may not fully account for confounding by subfertility or endocrine conditions. Future research requires multiethnic cohorts with objectively measured pre‐pregnancy anthropometrics, longitudinal lifestyle monitoring, and biomarker integration. Intervention studies should also assess whether stratified BMI management in GDM pregnancies lowers LGA risk more effectively than conventional approaches. Fifth, given sample size constraints and to minimize confounding from multiple concurrent analyses, we focused our study on LGA alone. However, WHO BMI categories are associated with multiple maternal and neonatal outcomes beyond LGA risk, and other relevant outcomes merit further exploration in future research.
5. CONCLUSION
In conclusion, this study contributes to evolving paradigms in precision perinatal care by demonstrating the critical modifying role of GDM status in pre‐pregnancy BMI‐associated LGA risks. These findings support the implementation of stratified management strategies to optimize maternal and fetal outcomes, advancing precision prevention efforts.
AUTHOR CONTRIBUTIONS
Yuhang Wu designed the study, managed the acquisition, the analysis and the interpretation of the data. Lizhang Chen, Jiabi Qin, and Tingting Wang were involved in designing the study. Jiye Zhang participated in acquiring and analyzing the data. Xiaochan Wang participated in data acquisition.All authors drafted the manuscript and approved the final version for publication, and take responsibility for its accuracy and integrity.
FUNDING INFORMATION
This work was supported by the National Natural Science Foundation of China (82 173 608, 82 473 644, and 82 404 362) and Excellent Youth Project of Hunan Province Office of Education (23B0001).
CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest.
DATA AVAILABILITY STATEMENT
Data available on request from the authors.
REFERENCES
- 1. American College of Obstetricians and Gynecologists . Committee opinion No. 549: obesity in pregnancy. Obstet Gynecol. 2013;121:213‐217. [DOI] [PubMed] [Google Scholar]
- 2. Surkan PJ, Hsieh C‐C, Johansson ALV, Dickman PW, Cnattingius S. Reasons for increasing trends in large for gestational age births. Obstet Gynecol. 2004;104:720‐726. [DOI] [PubMed] [Google Scholar]
- 3. Koyanagi A, Zhang J, Dagvadorj A, et al. Macrosomia in 23 developing countries: an analysis of a multicountry, facility‐based, cross‐sectional survey. Lancet. 2013;381:476‐483. [DOI] [PubMed] [Google Scholar]
- 4. Harvey L, Van Elburg R, Van Der Beek EM. Macrosomia and large for gestational age in Asia: one size does not fit all. J Obstet Gynaecol Res. 2021;47:1929‐1945. [DOI] [PubMed] [Google Scholar]
- 5. Weissmann‐Brenner A, Simchen MJ, Zilberberg E, et al. Maternal and neonatal outcomes of large for gestational age pregnancies. Acta Obstet Gynecol Scand. 2012;91:844‐849. [DOI] [PubMed] [Google Scholar]
- 6. Freeman DJ. Effects of maternal obesity on fetal growth and body composition: implications for programming and future health. Semin Fetal Neonatal Med. 2010;15:113‐118. [DOI] [PubMed] [Google Scholar]
- 7. Sun Y, Shen Z, Zhan Y, et al. Effects of pre‐pregnancy body mass index and gestational weight gain on maternal and infant complications. BMC Pregnancy Childbirth. 2020;20:390. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Song X, Chen L, Zhang S, et al. Gestational diabetes mellitus and high triglyceride levels mediate the association between pre‐pregnancy overweight/obesity and macrosomia: a prospective cohort study in Central China. Nutrients. 2022;14:3347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Lin L, Lin J, Yan J. Association of maternal body mass index change with risk of large for gestational age among pregnant women with and without gestational diabetes mellitus: a retrospective cohort study. J Matern Fetal Neonatal Med. 2024;37:2316732. [DOI] [PubMed] [Google Scholar]
- 10. Brenner B, Papadakis E, Greer IA, Gris J. Assessment‐based management of placenta‐mediated pregnancy complications: pragmatism until a precision medicine approach evolves. Br J Haematol. 2023;202:18‐30. [DOI] [PubMed] [Google Scholar]
- 11. Liao Q, Yu T, Chen J, Zheng X, Zheng L, Yan J. Relationship between maternal pre‐pregnancy BMI and neonatal birth weight in pregnancies with gestational diabetes mellitus: a retrospective cohort study. Front Med. 2025;11:1478907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Chen Y, Zhong T, Song X, et al. Maternal anaemia during early pregnancy and the risk of neonatal outcomes: a prospective cohort study in Central China. BMJ Paediatr Open. 2024;8:e001931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Tunçalp Ö, Pena‐Rosas J, Lawrie T, et al. WHO recommendations on antenatal care for a positive pregnancy experience‐going beyond survival. BJOG. 2017;124:860‐862. [DOI] [PubMed] [Google Scholar]
- 14. Dai L, Deng C, Li Y, et al. Birth weight reference percentiles for Chinese. PLoS One. 2014;9:e104779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Poston L, Caleyachetty R, Cnattingius S, et al. Preconceptional and maternal obesity: epidemiology and health consequences. Lancet Diabetes Endocrinol. 2016;4:1025‐1036. [DOI] [PubMed] [Google Scholar]
- 16. Ramsay JE, Ferrell WR, Crawford L, Wallace AM, Greer IA, Sattar N. Maternal obesity is associated with dysregulation of metabolic, vascular, and inflammatory pathways. J Clin Endocrinol Metab. 2002;87:4231‐4237. [DOI] [PubMed] [Google Scholar]
- 17. Spradley F, Palei A, Granger J. Immune mechanisms linking obesity and preeclampsia. Biomolecules. 2015;5:3142‐3176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Goossens GH. The metabolic phenotype in obesity: fat mass, body fat distribution, and adipose tissue function. Obes Facts. 2017;10:207‐215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Grivell R, Yelland L, Deussen A, Crowther C, Dodd J. Antenatal dietary and lifestyle advice for women who are overweight or obese and the effect on fetal growth and adiposity: the LIMIT randomised trial. BJOG. 2016;123:233‐243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Song X, Shu J, Zhang S, et al. Pre‐pregnancy body mass index and risk of macrosomia and large for gestational age births with gestational diabetes mellitus as a mediator: a prospective cohort study in Central China. Nutrients. 2022;14:1072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Ronnenberg AG, Chen C, Xu X, et al. Low preconception body mass index is associated with birth outcome in a prospective cohort of Chinese women. J Nutr. 2003;133:3449‐3455. [DOI] [PubMed] [Google Scholar]
- 22. Han Z, Lutsiv O, Mulla S, et al. Low gestational weight gain and the risk of preterm birth and low birthweight: a systematic review and meta‐analyses. Acta Obstet Gynecol Scand. 2011;90:935‐954. [DOI] [PubMed] [Google Scholar]
- 23. Juan J, Yang H, Wei Y, et al. Prevalence and characteristics of macrosomia in the first and subsequent pregnancy: a multi‐center retrospective study. Chin Med J. 2022;135:1492‐1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Rademaker D, Hukkelhoven CWPM, Van Pampus MG. Adverse maternal and perinatal pregnancy outcomes related to very advanced maternal age in primigravida and multigravida in The Netherlands: a population‐based cohort. Acta Obstet Gynecol Scand. 2021;100:941‐948. [DOI] [PubMed] [Google Scholar]
- 25. Stommel M, Schoenborn CA. Accuracy and usefulness of BMI measures based on self‐reported weight and height: findings from the NHANES & NHIS 2001‐2006. BMC Public Health. 2009;9:421. [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
Data available on request from the authors.