Abstract
Background
Low birth weight (LBW) remains a critical global health challenge, yet the differential contributions of risk factors in pregnancies with and without gestational diabetes mellitus (GDM) are poorly characterized. This study aimed to identify modifiable first-trimester risk factors for LBW in GDM and non-GDM subpopulations within a Chinese context, and quantify the preventable LBW burden by population attributable fraction (PAF) estimation adjusted for risk factor interdependence.
Methods
This prospective cohort study enrolled 34,031 pregnant women from Central China (2013–2019), stratified by GDM (n = 5414) and non-GDM (n = 28,617) groups. Early-pregnancy exposures included advanced maternal age, passive smoking, alcohol intake, no folic acid used, low education, and pre-pregnancy underweight. Multivariable logistic regression and principal component analysis-adjusted PAFs were used to quantify preventable LBW burdens while accounting for risk factor overlap.
Results
Five risk factors were identified in both GDM (advanced maternal age, passive smoking, underweight, no folic acid used, and low education) and non-GDM (advanced maternal age, alcohol intake, underweight, no folic acid used, and low education) pregnancies. It was found that 18.4% and 33.4% of LBW cases could be theoretically prevented through these five identified risk factors in both GDM and non-GDM cohorts, respectively. Among four shared contributors, low education emerged as the primary modifiable risk factor (adjusted PAF: 9.3% in GDM; 20.1% in non-GDM), followed by underweight (adjusted PAF: 3.7% in GDM; 5.8% in non-GDM), advanced maternal age (adjusted PAF: 3.3% in GDM; 4.8% in non-GDM), no folic acid used (adjusted PAF: 0.9% in GDM; 2.2% in non-GDM). Passive smoking and alcohol intake are specific risk factors for pregnant women with GDM and non-GDM, respectively, with adjusted PAF of 1.1% and 0.5%.
Conclusion
This study significantly contributes to the advancement of precision public health approaches by elucidating modifiable risk profiles associated with LBW in both GDM and non-GDM populations. These findings highlight population-specific heterogeneity in preventable disease burdens, revealing that a significant proportion of LBW cases could be reduced through targeted antenatal interventions.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12884-026-09097-y.
Keywords: Risk factors, Gestational diabetes mellitus, Low birth weight, Population attributable fraction, Epidemiology
Introduction
Low birth weight (LBW), defined as a birth weight less than 2500 g, persists as a critical global public health challenge with profound implications for infant mortality, childhood morbidity, and lifelong health trajectories [1]. Current estimates indicate that 14.6% of live births worldwide are classified as LBW, with low- and middle-income countries disproportionately affected [2]. LBW infants face elevated risks of neonatal complications, impaired neurodevelopment, and increased susceptibility to adult-onset chronic conditions [3–6]. The associated economic burden is staggering, encompassing both immediate neonatal intensive care costs and long-term healthcare expenditures [7]. These multifaceted consequences underscore LBW as a multifactorial condition requiring urgent, targeted interventions to mitigate its enduring societal impact.
The first-trimester represents a critical window for modifying LBW risk through addressable maternal factors [8]. Established determinants including pre-pregnancy body mass index (BMI), suboptimal gestational weight gain, smoking, psychological stress, and micronutrient deficiencies during early gestation exert profound effects on placental development and fetal growth trajectories [9–12]. Emerging evidence links these early exposures to downstream metabolic dysregulation, particularly gestational diabetes mellitus (GDM). Crucially, GDM fundamentally alters maternal metabolic physiology through insulin resistance and β-cell dysfunction, leading to distinct pathways of fetal growth disturbance compared to non-GDM pregnancies [13, 14]. This pathophysiological divergence establishes GDM as a distinct clinical entity that reshapes the etiological basis of adverse pregnancy outcomes. Risk factors for LBW may operate through biological pathways that differ by GDM status, such as being exacerbated by the insulin resistance and chronic inflammation characteristic of GDM. Alternatively, their effect size may vary substantially between these populations [15, 16]. We therefore hypothesized that a priori stratification by GDM status is essential to uncover these distinct risk profiles. This approach, justified by established biological and clinical heterogeneity, aims to enable precision prevention strategies, which are currently hindered by a lack of GDM-stratified evidence [17]. Recent advances in predictive modeling, exemplified by the GDM-specific risk stratification system developed by Cooray et al. for adverse pregnancy outcomes, highlight the potential of tailored interventions for high-risk subgroups [18]. Nevertheless, there is currently insufficient evidence to justify LBW prevention strategies specifically tailored to GDM status.
In China, where LBW affects 800,000–900,000 newborns annually [2], addressing this gap holds significant public health potential. Quantifying the modifiable risk burden through population attributable fraction (PAF) analysis is central to effective intervention [19]. However, conventional PAF estimates often assume unrealistic independence among correlated risk factors (e.g., physical inactivity, obesity, and socioeconomic status), potentially inflating combined risk estimates and distorting preventive prioritization [20]. Advanced methodologies that accounting for these interdependencies are thus essential for accurate risk prioritization.
This prospective cohort study addresses these critical gaps through two primary aims: (1) identifying distinct first-trimester modifiable risk factors for LBW in GDM and non-GDM subpopulations within a Chinese context, and (2) applying counterfactual-based PAF estimation adjusting for risk factor interdependencies to quantify the preventable LBW burden. By elucidating etiology-specific risk pathways, our findings aim to inform risk-stratified prevention guidelines and optimize resource allocation within precision maternal care frameworks, with implications for national LBW reduction strategies and global maternal health equity.
Materials and methods
Study design and participant enrollment
This prospective cohort study was conducted at Hunan Provincial Maternal and Child Health Care Hospital in Central China between 13 March 2013 and 31 December 2019. We enrolled pregnant women aged ≥ 18 years initiating prenatal care before 14 gestational weeks who intended to receive ongoing care and deliver at the study hospital. Participants were recruited from outpatient departments (reproductive medicine, obstetrics, and ultrasonography) following written informed consent. The study protocol received approval from the Ethics Committee of Xiangya School of Public Health, Central South University (Approval No. XYGW-2019–01).
At 24–28 gestational weeks, participants underwent a 75 g oral glucose tolerance test (OGTT) following International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria. Plasma glucose levels (fasting, 1-h, and 2-h post-load) were measured using an automated analyzer (Toshiba TBA-120FR, Tokyo, Japan). GDM was diagnosed if any value met the thresholds: fasting ≥ 5.1 mmol/L, 1-h ≥ 10.0 mmol/L, or 2-h ≥ 8.5 mmol/L. Pregnant women who met any of the following criteria were excluded: 1) multiple births; 2) pre-pregnancy diabetes; 3) pregnancy termination due to miscarriage, stillbirth, or induced labor; 4) conception via assisted reproductive technologies (ART); or 5) loss to follow-up. It is worth noting that ART pregnancies may introduce confounding by obscuring early pregnancy modifiable factors, such as the relationship between lifestyle exposure and LBW [21]. After applying these exclusion criteria, the final analytical cohort comprised 34,031 non-duplicated pregnant women (Fig. 1).
Fig. 1.
Flow chart showing study design and inclusion of participants
Data collection
Upon enrollment, trained interviewers administered a structured questionnaire through face-to-face interviews and telephone follow-ups to collect data on sociodemographic characteristics, behavioral habits, and lifestyle factors during early pregnancy (< 14 weeks). The questionnaire used in this study has previously been published elsewhere [22]. Variables included parity, maternal age, pre-pregnancy BMI, education level, smoking, alcohol intake, and folic acid supplementation. All self-reported data were cross-verified with electronic medical records (EMRs) to ensure accuracy. Pregnancy complications, clinical measurements, and delivery outcomes were systematically extracted from the hospital’s EMR system. Quality control measures included duplicate data entry and periodic audits by an independent team to resolve discrepancies.
Exposures and covariates
Early-pregnancy (< 14 weeks) modifiable risk factors included: 1) advanced maternal age: defined as maternal age ≥ 35 years at conception, calculated using birthdates from national identification cards [22]; 2) smoking: categorized as active (self-reported inhalation of lit tobacco products) and passive (exposure to secondhand smoke at least one day a week) during the first three months of pregnancy [23]; 3) alcohol intake: consumption of any alcoholic beverage (beer, wine, or spirits) during early pregnancy, regardless of quantity [23]; 4) underweight: body mass index (BMI) < 18.5 kg/m2, calculated from self-reported values of weight and height before conception by participants at the first prenatal visit [24]; 5) no folic acid used: self-reported whether folic acid was used before or during pregnancy; 6) low education: defined as completion of senior middle school or below [25]. Additionally, residence (urban/rural), parity (primiparous/multiparous), per capita monthly family income (≤ CNY 2500, CNY 2500 to CNY 5000, > CNY 5000), ethnicity (Han/minority), gestational weight gain, and infant sex (male/female) were included as covariates.
Outcomes
The primary outcome was LBW, defined as birth weight < 2500 g. Birth weight and neonatal outcomes were extracted from EMRs and cross-checked by obstetricians blinded to exposure status. To address fetal growth restriction on a gestational age- and sex-standardized scale, we conducted a sensitivity analysis using small for gestational age (SGA) defined by the Chinese national birthweight for gestational age standard WS/T 800–2022 (percentiles for 24–42 gestational weeks) [26]. Infants with birthweight < 10th percentile for sex and gestational age were classified as SGA. Dating followed routine obstetric practice; when last menstrual period and early ultrasound differed by > 1 week, ultrasound dating was prioritized per standard clinical convention.
Statistical analysis
We used EpiData 3.1 for data management with double-entry verification procedures. Participants were stratified into two independent groups based on GDM status. Continuous variables with normal distributions are reported as mean (SD); categorical variables as counts (percentages). We employed Student's t-tests for continuous variables and χ2 or Fisher's exact tests for categorical variables, as appropriate, for group comparisons. We developed multivariable logistic regression models to estimate adjusted odds ratios with 95% confidence intervals (CIs) for LBW risk factors, stratified by GDM status. Prior to this, the interaction between potential risk factors and GDM status was also tested (Table S1). Two adjustment levels were implemented. Model 1 was adjusted for demographic covariates, including residence, parity, per capita monthly family income, ethnicity, gestational weight gain, and infant sex. Model 2 was further adjusted for all early-pregnancy risk factors. The adjusted ORs from Model 2 were used to calculate PAF for each risk factor using Levin’s formula:
 |
where P represents the population prevalence of each factor and OR represents the corresponding odds ratio. The combined estimate of the PAF assumed independence of risk factors using this formula:
 |
Traditional Levin's formula assumes independence among risk factors, potentially overestimating combined PAFs due to exposure overlap. To address this limitation, we implemented a principal component analysis (PCA)-adjusted approach accounting for shared variance (communality) among risk factors. First, we constructed a tetrachoric correlation matrix to quantify pairwise associations between binary risk factors. Second, PCA was applied to this matrix to extract eigenvalues and factor loadings, retaining components with eigenvalues > 1 (Kaiser criterion). Third, communality for each risk factor, representing the proportion of variance explained by shared latent factors, was calculated as the sum of squared factor loadings across retained components. Combined PAFs were then adjusted using the weighting formula:
 |
where the 
was 1 minus the communality of each risk factor. This downweighs PAFs estimates for factors with high shared variance, reflecting their reduced unique contribution to LBW risk. Final adjusted individual PAFs were calculated based on their proportional contributions to the weighted combined PAF, thereby mitigating overestimation due to exposure collinearity. Methodological details of this PCA-adjusted approach are described in our previous work [27, 28].
Furthermore, we repeated the modeling strategy used for LBW with SGA as the dependent variable, fitting multivariable logistic regression models stratified by GDM status and adjusting for the same covariates. We then estimated PAFs for SGA (individual and combined) using the same framework as for LBW, including the PCA-adjusted approach to account for interdependence among risk factors.
All statistical analyses and graphics were produced with the R statistical program (version 4.2.0) in this study. Two-sided P values less than 0.05 were considered to represent statistical significance.
Results
The study included 34031participants from the first-trimester pregnancy cohort, comprising 5414 women diagnosed with GDM and 28,617 non-GDM pregnancies (Fig. 1). LBW occurred in 2990 neonates (8.8%) overall, with cases distributed across both cohorts. Baseline characteristics revealed significant differences between the LBW cases of GDM and non-GDM pregnancies (Table S2). Notably, women with GDM and LBW presented with a higher mean age at delivery (32.85 years vs. 30.93 years), an increased pre-pregnancy BMI (22.05 kg/m2 vs. 21.12 kg/m2), and a greater prevalence of advanced maternal age (40.4% vs. 27.1%) as well as passive smoking (9.3% vs. 6.7%). Conversely, non-GDM pregnancies with LBW showed higher mean infant birth weight (1931.29 g vs. 1843.68 g), greater gestational weight gain (mean: 11.80 kg vs. 10.22 kg), and higher prevalence of maternal underweight (19.3% vs. 12.4%) compared to GDM pregnancies.
GDM pregnancies complicated by LBW exhibited higher prevalence of advanced maternal age (40.4% vs. 30.9%), underweight (12.4% vs. 7.9%), passive smoking (9.3% vs. 6.3%), no folic acid used (7.8% vs. 4.0%), and low educational attainment (50.3% vs. 32.9%), along with lower gestational age at delivery (34.11 vs. 38.88 weeks) and gestational weight gain (10.22 kg vs. 12.53 kg) than GDM pregnancies with normal birth weight (Table 1). In non-GDM pregnancies, LBW cases showed higher proportions of advanced maternal age (27.1% vs. 20.1%), multiparity (75.5% vs. 67.9%), alcohol consumption (2.5% vs. 1.3%), no folic acid used (8.1% vs. 4.2%), low educational attainment (54.5% vs. 34.4%), and underweight status (19.3% vs. 15.2%). Additionally, LBW cases had lower gestational age (34.23 vs. 38.83 weeks) and gestational weight gain (11.88 kg vs. 14.00 kg) compared to non-LBW pregnancies (Table 1).
Table 1.
Characteristics of the participants with or without low birth weight in two cohorts
| Characteristics | GDM | Non-GDM | ||||||
|---|---|---|---|---|---|---|---|---|
| All (N = 5414) | LBW (N = 485) | Non-LBW (N = 4929) | P-value | All (N = 28,617) | LBW (N = 2505) | Non-LBW (N = 26,112) | P-value | |
| Age at delivery (years) | 32.43 ± 4.52 | 32.85 ± 4.42 | 32.39 ± 4.53 | 0.031 | 30.87 ± 4.47 | 30.93 ± 5.29 | 30.87 ± 4.38 | 0.517 |
| Advanced maternal age (≥ 35 years) | < 0.001 | < 0.001 | ||||||
| Yes | 1719 (31.8) | 196 (40.4) | 1523 (30.9) | 5930 (20.7) | 678 (27.1) | 5252 (20.1) | ||
| No | 3695 (68.2) | 289 (59.6) | 3406 (69.1) | 22,687 (79.3) | 1827 (72.9) | 20,860 (79.9) | ||
| Residence | < 0.001 | < 0.001 | ||||||
| Urban | 3429 (63.3) | 232 (47.8) | 3197 (64.9) | 17,598 (61.5) | 1131 (45.1) | 16,467 (63.1) | ||
| Rural | 1985 (36.7) | 253 (52.2) | 1732 (35.1) | 11,019 (38.5) | 1374 (54.9) | 9645 (36.9) | ||
| Pre-pregnancy BMI (kg/m2) | 22.20 ± 3.16 | 22.05 ± 3.02 | 21.22 ± 3.17 | 0.270 | 21.15 ± 2.85 | 21.12 ± 3.08 | 21.15 ± 2.82 | 0.670 |
| Pre-pregnancy BMI category | 0.001 | < 0.001 | ||||||
| Underweight | 448 (8.3) | 60 (12.4) | 388 (7.9) | 4459 (15.6) | 483 (19.3) | 3976 (15.2) | ||
| Normal weight | 4090 (75.5) | 363 (74.8) | 3727 (75.6) | 21,764 (76.1) | 1774 (70.8) | 19,990 (76.6) | ||
| Overweight or obese | 876 (16.2) | 62 (12.8) | 814 (16.5) | 2394 (8.4) | 248 (9.9) | 2146 (8.2) | ||
| Parity | < 0.001 | < 0.001 | ||||||
| Primiparous | 1477 (27.3) | 74 (15.3) | 1403 (28.5) | 9000 (31.4) | 613 (24.5) | 8387 (32.1) | ||
| Multiparous | 3937 (72.7) | 411 (84.7) | 3526 (71.5) | 19,617 (68.6) | 1892 (75.5) | 17,725 (67.9) | ||
| Ethnicity | 0.484 | 0.242 | ||||||
| Han | 5339 (98.6) | 480 (99.0) | 4859 (98.6) | 28,244 (98.7) | 2466 (98.4) | 25,778 (98.7) | ||
| Minority | 75 (1.4) | 5 (1.0) | 70 (1.4) | 373 (1.3) | 39 (1.6) | 334 (1.3) | ||
| Education level | < 0.001 | < 0.001 | ||||||
| Senior middle school or below | 1865 (34.4) | 244 (50.3) | 1621 (32.9) | 10,343 (36.1) | 1366 (54.5) | 8977 (34.4) | ||
| College or above | 3549 (65.6) | 241 (40.7) | 3308 (67.1) | 18,274 (63.9) | 1139 (45.5) | 17,135 (65.6) | ||
| Per-capita monthly family income (CNY) | 0.001 | < 0.001 | ||||||
| ≤ 2500 | 944 (17.4) | 81 (16.7) | 863 (17.5) | 4936 (17.2) | 368 (14.7) | 4568 (17.5) | ||
| 2500 to 5000 | 2860 (52.8) | 293 (60.4) | 2567 (52.1) | 15,300 (53.5) | 1549 (61.8) | 13,751 (52.7) | ||
| > 5000 | 1610 (29.7) | 111 (22.9) | 1499 (30.4) | 8381 (29.3) | 588 (23.5) | 7793 (29.8) | ||
| Alcohol intake | 0.154 | < 0.001 | ||||||
| Yes | 91 (1.7) | 12 (2.5) | 79 (1.6) | 407 (1.4) | 62 (2.5) | 345 (1.3) | ||
| No | 5323 (98.3) | 473 (97.5) | 4850 (98.4) | 28,210 (98.6) | 2443 (97.5) | 25,767 (98.7) | ||
| Active smoking | 0.547 | 0.157 | ||||||
| Yes | 73 (1.3) | 8 (1.6) | 65 (1.3) | 380 (1.3) | 41 (1.6) | 339 (1.3) | ||
| No | 5341 (98.7) | 477 (98.4) | 4864 (98.7) | 28,237 (98.7) | 2464 (98.4) | 25,773 (98.7) | ||
| Passive smoking | 0.011 | 0.559 | ||||||
| Yes | 355 (6.6) | 45 (9.3) | 310 (6.3) | 1989 (7.0) | 167 (6.7) | 1822 (7.0) | ||
| No | 5059 (93.4) | 440 (90.7) | 4619 (93.7) | 26,628 (93.0) | 2338 (93.3) | 24,290 (93.0) | ||
| Folic acid supplementation | < 0.001 | < 0.001 | ||||||
| Yes | 5178 (95.6) | 447 (92.2) | 4731 (96.0) | 27,309 (95.4) | 2303 (91.9) | 25,006 (95.8) | ||
| No | 236 (4.4) | 38 (7.8) | 198 (4.0) | 1308 (4.6) | 202 (8.1) | 1106 (4.2) | ||
| Infant sex | 0.264 | < 0.001 | ||||||
| Male | 2732 (50.5) | 233 (48.0) | 2499 (50.7) | 15,180 (53.0) | 1236 (49.3) | 13,944 (53.4) | ||
| Female | 2682 (49.5) | 252 (52.0) | 2430 (49.3) | 13,437 (47.0) | 1269 (50.7) | 12,168 (46.6) | ||
| Gestational age at delivery (weeks) | 38.45 ± 2.14 | 34.11 ± 3.17 | 38.88 ± 1.42 | < 0.001 | 38.43 ± 2.12) | 34.23 ± 3.16 | 38.83 ± 1.45 | < 0.001 |
| Gestational weight gain (kg) | 12.32 ± 4.69 | 10.22 ± 4.39 | 12.53 ± 4.67 | < 0.001 | 13.81 ± 4.72 | 11.80 ± 5.34 | 14.00 ± 4.61 | < 0.001 |
| Birth weight (g) | 3160.32 ± 594.28 | 1843.68 ± 468.47 | 3289.88 ± 423.05 | < 0.001 | 3142.46 ± 544.24 | 1931.29 ± 472.54 | 3258.66 ± 385.96 | < 0.001 |
Continuous data are presented as mean ± SD, and categorical data are presented as n (%)
Two-sample t tests were used to compare mean values, and Pearson’s Chi-square or Fisher’s exact test was used to test the difference between proportions
LBW Low birth weight, BMI Body mass index
A multivariable logistic regression analysis was conducted, adjusting for residence, parity, per capita monthly family income, ethnicity, gestational weight gain, and infant sex, which revealed distinct risk profiles between populations with GDM and non-GDM (Table 2). Among GDM pregnancies, passive smoking (OR 1.49 [95% CI 1.06, 2.09]) emerged as a significant independent risk factor for LBW. Non-GDM pregnancies exhibited a unique association with alcohol intake (OR 1.52 [95% CI 1.14, 2.04]). Shared modifiable risk factors across both groups included advanced maternal age (GDM: OR 1.33 [95% CI 1.08, 1.64]; non-GDM: OR 1.38 [95% CI 1.25, 1.53]), underweight (GDM: OR 2.46 [95% CI 1.79, 3.37]; non-GDM: OR 1.62 [95% CI 1.45, 1.81]), no folic acid used (GDM: OR 1.61 [95% CI 1.10, 2.37]; non-GDM: OR 1.74 [95% CI 1.47, 2.05]), and low education (GDM: OR 2.07 [95% CI 1.70, 2.52]; non-GDM: OR 2.21 [95% CI 2.03, 2.41]). Additionally, overweight or obesity was associated with a significantly lower risk of LBW in the GDM group (OR 0.55 [95% CI 0.41, 0.74]), but not in the non-GDM group (OR 0.92 [95% CI 0.80, 1.07]). This protective association was excluded from subsequent PAF calculations, as PAF is conventionally reserved for adverse risk factors and our analysis focused on identifying modifiable risks for population-level intervention. Sensitivity analysis using Chinese birthweight standards identified two independent risk factors for SGA in the GDM group versus four in the non-GDM group (Table S3). Underweight and low education were common to both, whereas advanced maternal age and alcohol intake were specific to the non-GDM group.
Table 2.
Related factors of low birth weight in patients with and without gestational diabetes mellitus based on multivariable logistic regression
| Characteristics | Model 1 | Model 2 | ||
|---|---|---|---|---|
| OR (95% CI) | P-value | OR (95% CI) | P-value | |
| GDM | ||||
| Advanced maternal age | 1.22 (1.00, 1.50) | 0.049 | 1.33 (1.08, 1.64) | 0.007 |
| Alcohol intake | 1.52 (0.81, 2.86) | 0.197 | 1.35 (0.70, 2.63) | 0.372 |
| Passive smoking | 1.47 (1.05, 2.06) | 0.024 | 1.49 (1.06, 2.09) | 0.023 |
| Underweight* | 2.38 (1.75, 3.24) | < 0.001 | 2.46 (1.79, 3.37) | < 0.001 |
| Overweight or obese* | 0.59 (0.44, 0.78) | < 0.001 | 0.55 (0.41, 0.74) | < 0.001 |
| No folic acid used | 1.72 (1.19, 2.50) | < 0.001 | 1.61 (1.10, 2.37) | 0.014 |
| Low education | 1.99 (1.64, 2.41) | < 0.001 | 2.07 (1.70, 2.52) | < 0.001 |
| Non-GDM | ||||
| Advanced maternal age | 1.31 (1.18, 1.44) | < 0.001 | 1.38 (1.25, 1.53) | < 0.001 |
| Alcohol intake | 1.78 (1.34, 2.36) | < 0.001 | 1.52 (1.14, 2.04) | 0.005 |
| Passive smoking | 0.99 (0.84, 1.17) | 0.872 | 0.94 (0.79, 1.11) | 0.437 |
| Underweight* | 1.57 (1.40, 1.75) | < 0.001 | 1.62 (1.45, 1.81) | < 0.001 |
| Overweight or obese* | 1.01 (0.87, 1.17) | 0.899 | 0.92 (0.80, 1.07) | 0.277 |
| No folic acid used | 1.82 (1.55, 2.14) | < 0.001 | 1.74 (1.47, 2.05) | < 0.001 |
| Low education | 2.20 (2.02, 2.40) | < 0.001 | 2.21 (2.03, 2.41) | < 0.001 |
Model 1 was adjusted for residence, parity, per capita monthly family income, ethnicity, gestational weight gain, and infant sex
Model 2 was adjusted for all potential exposures in addition to covariates in model 1
Values are presented as adjusted odds ratio with 95% confidence interval
GDM Gestational diabetes mellitus
*vs. Normal weight
The PAFs of modifiable risk factors associated with LBW were comprehensively assessed in both GDM and non-GDM cohorts, as depicted in Fig. 2. In the GDM population, it was found that 44.3% (95% CI 26.9, 59.9) of LBW cases could potentially be prevented by five identified risk factors (Fig. 2A). After adjusting for the overlap among these risk factors, the combined PAF was 18.4% (95% CI 10.9, 26.1) (Fig. 2A). On the other hand, among participants without GDM, the PAFs attributable to five risk factors were 43.5% (95% CI 36.7, 50.2), with an adjusted combined PAF of 33.4% (95% CI 29.0, 37.9) (Fig. 2B). Within the GDM group, the most significant contributor to the incidence of LBW was low educational attainment, with a PAF of 26.9% (95% CI 19.4, 34.3). Additional factors included underweight (10.8% [95% CI 6.2, 16.4]), advanced maternal age (9.5% [95% CI 2.5, 16.9]), passive smoking (3.1% [95% CI 0.4, 6.7]), and no folic acid used (2.6% [95% CI 0.4, 5.7]) (Fig. 2A). Following adjustment for the communality of risk factors, low education retained the highest adjusted PAFs (9.3%), and no folic acid used demonstrated the lowest adjusted contribution (0.9%) among the five evaluated factors (Fig. 2C). In the non-GDM population, low educational attainment emerged as the primary modifiable risk factor, with a PAF of 30.4% (95% CI 27.1, 33.7). This was followed by underweight (8.8% [95% CI 6.6, 11.2]), advanced maternal age (7.3% [95% CI 4.9, 9.9]), no folic acid used (3.3% [95% CI 2.1, 4.6]), and alcohol intake (0.7% [95% CI 0.2, 1.4]) (Fig. 2B). The adjusted PAFs for these factors varied widely, ranging from 0.5% for alcohol intake to 20.1% for low education, indicating a considerable reduction in attributable risk after considering overlapping exposures (Fig. 2D). Figure S1 presents the PAFs and PCA-adjusted PAFs for SGA. In both cohorts, low education and underweight accounted for the largest attributable fractions, and their ranking was largely consistent with the patterns observed for LBW.
Fig. 2.
Estimates for population attribution fractions and adjusted population attribution fractions of risk factors for low birth weight in participants with gestational diabetes mellitus (A, C) and non-gestational diabetes mellitus (B, D)
Discussion
This stratified analysis of maternally modifiable first-trimester risk factors for LBW in pregnancies with and without GDM reveals population-specific heterogeneity in preventable disease burdens. We discovered divergent risk profiles, with five risk factors contributing to LBW in both GDM and non-GDM pregnancies, underscoring differential etiological pathways. Using an advanced PAF methodology adjusted for risk factor interdependencies, we further estimated that 18.4% and 33.4% of LBW cases could be theoretically prevented through these five identified risk factors in both GDM and non-GDM cohorts, respectively. These findings substantiate precision public health approaches, emphasizing risk-stratified antenatal care pathways to maximize resource efficiency while achieving etiology-specific prevention targets.
Unlike recent PAF reports quantifying LBW risk factors (typically assessed during mid-to-late pregnancy) across entire pregnant populations [20, 29, 30], this study is designed to demonstrate the prevention potential of common modifiable behavioral factors during the first-trimester, a clinically significant window, separately in GDM and non-GDM subpopulations. Among the shared risk factors, low education emerges as the leading modifiable risk factor in both groups, with adjusted PAFs of 9.3% (GDM) and 20.1% (non-GDM). While individual educational attainment itself may not be modifiable during pregnancy, its impact operates through downstream mediators such as health literacy, healthcare access, health-seeking behaviors, and maternal metabolic status [31–33]. For pregnant women with low education, targeted antenatal health education and community-based outreach programs can reduce LBW risks. In China, integrating such interventions into primary maternal care, especially in rural areas (rural residency > 50% in LBW groups, Table 1), may bridge the education gap. The prominence of underweight in both cohorts demands targeted interventions. For GDM pregnancies, nutritional supplementation protocols should prioritize rapid correction of pre-pregnancy undernutrition, given its potent effect on LBW risk. However, broader nutritional support programs may yield greater population-level impact due to higher underweight prevalence in non-GDM populations. Advanced maternal age contributed modestly to LBW risk (adjusted PAF: 3.3% for GDM and 4.8% for non-GDM), likely reflecting age-related declines in placental function and comorbid conditions [34]. In the context of China’s three-child policy [35], childbearing delay may be unavoidable. Effective risk mitigation requires a multifaceted public health response, including providing professional pre-pregnancy optimization to prevent reproductive aging, enhanced fetal surveillance protocols for advanced-age pregnancies, and policy-level initiatives promoting workplace flexibility and childcare support to reduce delayed childbearing pressures [36, 37]. Folic acid deficiency, known to disrupt DNA synthesis and repair, leading to developmental abnormalities and restricted fetal growth [38–40], was identified as a significant risk factor in both GDM and non-GDM pregnancies. Our findings highlight the importance of universal prenatal folic acid supplementation programs. Fortunately, the relatively low PAFs for non-use of folic acid supplementation suggest that such programs are already well-implemented in China [41, 42], though continued vigilance is necessary.
Our analysis also revealed distinct independent risk factors in GDM and non-GDM populations. Alcohol intake was identified as an independent risk factor for LBW, predominantly observed in non-GDM pregnancies. Despite its well-established risks to fetal development [43], the potential for prevention through the modification of these factors remains limited. This may be attributed to the widespread adoption of comprehensive preconception planning and lifestyle modifications in China [41]. Passive smoking exposes fetuses to toxic chemicals, including nicotine, carbon monoxide, and DNA adducts, which cause vasoconstriction and reduce fetal oxygen supply, impairing growth [44]. Notably, our findings indicated that passive smoking emerged as a unique risk factor for LBW in GDM pregnancies, with an estimated 1.1% of cases potentially preventable through avoidance of second-hand smoke exposure. This highlights the need to integrate smoking cessation initiatives targeting close contacts of pregnant women into prenatal care protocols.
In this study, our decision to stratify analyses by GDM status was predicated on the a priori understanding of distinct pathophysiological milieus, rather than being contingent solely upon the statistical significance of interaction terms in our cohort. While only passive smoking exhibited a formal statistical interaction (Table S1), the stratified analyses revealed meaningfully divergent risk profiles, including the specificity of passive smoking to GDM and alcohol intake to non-GDM, and substantially different magnitudes of attributable burden for shared factors such as low education. A pooled analysis would have obscured these differences, yielding homogenized risk estimates ill-suited for precision public health. The divergent risk profiles and PAFs themselves are the primary findings that validate the stratification approach, demonstrating that GDM status is a key determinant of the preventable LBW landscape. Additionally, the supplementary analysis using the Chinese-specific SGA standard served as a robust validation of our primary LBW findings. The high consistency in the core risk profiles, particularly the persistent and dominant roles of low education and underweight, across both outcome definitions underscores the robustness of these associations for fetal growth impairment in Central China. Minor discrepancies, however, offer additional etiological insights. For instance, the fact that passive smoking and no folic acid use were significant for LBW but not consistently for SGA suggests that their impact on birth weight may be partly mediated through pathways influencing gestational length (i.e., preterm birth) rather than purely restricting growth velocity. The reaffirmation of alcohol intake as a risk factor for SGA in the non-GDM population further solidifies the evidence for its direct growth-restricting effects. The use of the SGA standard mitigates concerns about regional variation in fetal size, confirming that the identified risk profiles are pertinent to growth restriction in the Central Chinese population.
This study employs a prospective design along with standardized follow-up procedures, thereby strengthening the ability to draw causal inferences. Our findings support the implementation of stratified strategies for the prevention of LBW, highlighting the necessity of incorporating targeted interventions within China’s maternal health framework, particularly in light of the increasing prevalence of GDM associated with the three-child policy [45]. To address the temporal challenge of GDM diagnosis at 24–28 weeks, we propose a pragmatic, two-stage preventive framework guided by our findings. First, during the first-trimester, universal and precautionary interventions should target factors with high attributable burden across both populations, including the correction of underweight and initiatives to mitigate the risks associated with low educational attainment. Avoidance of exposures with universal detriment, for instance alcohol consumption and secondhand smoke, should be promoted. Subsequently, after the formal GDM diagnosis, clinical management can be intensively refined based on the GDM-specific risk profile we identified, with particular emphasis on protecting patients from passive smoke exposure. This framework leverages early-actionable factors while respecting diagnostic timelines, thereby maximizing the window of opportunity for prevention. Our PCA-adjusted model supports this approach by quantifying unique risk contributions beyond GDM status itself. On the other hand, traditional approaches that focus on single risk factors, while beneficial for assessing individual contributions, rely on the assumption of independence among these factors when calculating combined population attributable fractions. A significant advantage of our study is the provision of adjusted combined PAFs that account for the interrelationships among risk factors. Notably, the total attributable disease proportion associated with the various causal mechanisms of LBW may exceed 100%, representing a statistical paradox due to the breach of the mutual exclusivity assumption. In our analysis, the adjusted combined PAFs yielded lower estimates than their unadjusted counterparts. This outcome is expected and aligns with biological plausibility [46], with stable estimates across all individual risk factors. It is crucial to recognize that adjusted and unadjusted PAFs do not offer conflicting views; rather, they provide complementary perspectives regarding the influence of overlapping risk factors [47]. While unadjusted PAFs may more accurately reflect the potential impact of entirely removing a single risk factor on disease burden, adjusted PAFs are likely to present a more realistic assessment of the effects resulting from simultaneous modifications of multiple risk factors.
Despite the strengths of our study, several limitations must be acknowledged. First, while comprehensive phenotyping enhances internal validity, recall bias in self-reported data and residual confounding from unmeasured variables, such as dietary patterns, psychosocial stress, and environmental pollutants remain, and the latter may cause underestimation of the preventable burden of LBW. Second, although we have proposed several plausible biological mechanisms in this analysis, the causal pathways involved are not yet fully understood. The collection of maternal feces and serum from this first-trimester cohort for the integrated biospecimen repository presents an opportunity for future multi-omics research to address these gaps. Third, PAF estimates are context-specific, influenced by region-specific exposure prevalence and effect sizes, limiting generalizability. Moreover, the theoretical "complete elimination" of risk factors is often unrealistic; PAFs therefore estimate the maximum preventable burden under counterfactual conditions. Real-world interventions may achieve smaller but still meaningful reductions.
Conclusion
This study enhances precision public health by delineating distinct, modifiable risk profiles for LBW in GDM and non-GDM populations. Furthermore, our findings support a two-stage prevention strategy, initiating universal first-trimester measures followed by GDM-specific interventions post-diagnosis. This approach underscores that the significant preventable burden of LBW necessitates targeted, timely, and stratified prenatal planning. Future research should aim to validate these results across a variety of cohorts and assess the effectiveness of cost-effective interventions that prioritize risk factors during the first-trimester.
Supplementary Information
Acknowledgements
We would like to thank the researchers who participated in this survey.
Disclosures
All authors report no disclosures relevant to the manuscript.
Authors’ contributions
Yuhang Wu: Conceptualization (lead); formal analysis (lead); software data (lead), methodology (lead); curation (equal); writing-original draft (lead); formal analysis (lead); writing-review and editing (equal). Hanyu Xiao: Data curation (equal); investigation (equal); project administration(equal). Lizhang Chen: Investigation (supporting); data curation (equal); project administration(equal). Tingting Wang: Investigation (equal); project administration(equal); resource (equal); conceptualization (supporting); supervision (equal). Jiabi Qin: Investigation (equal); project administration(equal); writing-review and editing (equal). All authors gave their final approval and agree to be accountable for all aspects of the work.
Funding
This work was supported by the National Natural Science Foundation of China (82173608, 82473644, and 82404362) and Excellent Youth Project of Hunan Province Office of Education (23B0001).
Data availability
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.
Declarations
Ethics approval and consent to participate
This study was conducted under the guidance of the Declaration of Helsinki, and all participants in this study filled in informed consent forms at the time of recruitment. The study was approved by the ethics committees of Xiangya School of Public Health, Central South University (XYGW-2019–01).
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.
Contributor Information
Jiabi Qin, Email: qinjiabi123@163.com.
Tingting Wang, Email: wangting91123@126.com.
References
- 1.Katz J, Lee AC, Kozuki N, Lawn JE, Cousens S, Blencowe H, et al. Mortality risk in preterm and small-for-gestational-age infants in low-income and middle-income countries: a pooled country analysis. Lancet. 2013;382:417–25. 10.1016/S0140-6736(13)60993-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Blencowe H, Krasevec J, De Onis M, Black RE, An X, Stevens GA, et al. National, regional, and worldwide estimates of low birthweight in 2015, with trends from 2000: a systematic analysis. Lancet Glob Health. 2019;7:e849. 10.1016/S2214-109X(18)30565-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Blencowe H, Lee AC, Cousens S, Bahalim A, Narwal R, Zhong N, et al. Preterm birth–associated neurodevelopmental impairment estimates at regional and global levels for 2010. Pediatr Res. 2013;74:17–34. 10.1038/pr.2013.204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Christian P, Lee SE, Donahue Angel M, Adair LS, Arifeen SE, Ashorn P, et al. Risk of childhood undernutrition related to small-for-gestational age and preterm birth in low- and middle-income countries. Int J Epidemiol. 2013;42:1340–55. 10.1093/ije/dyt109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ediriweera DS, Dilina N, Perera U, Flores F, Samita S. Risk of low birth weight on adulthood hypertension - evidence from a tertiary care hospital in a South Asian country, Sri Lanka: a retrospective cohort study. BMC Public Health. 2017;17:358. 10.1186/s12889-017-4268-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hansen AL, Brøns C, Engelhard LM, Andersen MK, Hansen T, Nielsen JS, et al. Low birthweight in patients with type 2 diabetes is associated with elevated risk of cardiovascular events and mortality. Diabetologia. 2024;67:1616–29. 10.1007/s00125-024-06170-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Beam AL, Fried I, Palmer N, Agniel D, Brat G, Fox K, et al. Estimates of healthcare spending for preterm and low-birthweight infants in a commercially insured population: 2008–2016. J Perinatol. 2020;40:1091–9. 10.1038/s41372-020-0635-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Poon LC, McIntyre HD, Hyett JA, Da Fonseca EB, Hod M. The first-trimester of pregnancy – a window of opportunity for prediction and prevention of pregnancy complications and future life. Diabetes Res Clin Pract. 2018;145:20–30. 10.1016/j.diabres.2018.05.002. [DOI] [PubMed] [Google Scholar]
- 9.Lyu Y, Cui M, Zhang L, Zheng G, Zuo H, Xiu Q, et al. Pre-pregnancy body mass index, gestational diabetes mellitus, and gestational weight gain: individual and combined effects on fetal growth. Front Public Health. 2024;12:1354355. 10.3389/fpubh.2024.1354355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Brand JS, Gaillard R, West J, McEachan RRC, Wright J, Voerman E, et al. Associations of maternal quitting, reducing, and continuing smoking during pregnancy with longitudinal fetal growth: findings from Mendelian randomization and parental negative control studies. PLoS Med. 2019;16:e1002972. 10.1371/journal.pmed.1002972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wu J, Zhou F, Wang Y, Niu Y, Zhang C, Meng Y, et al. Associations between maternal early pregnancy depression and longitudinal fetal growth. J Affect Disord. 2024;362:808–15. 10.1016/j.jad.2024.07.068. [DOI] [PubMed] [Google Scholar]
- 12.Christian P, Mullany LC, Hurley KM, Katz J, Black RE. Nutrition and maternal, neonatal, and child health. Semin Perinatol. 2015;39:361–72. 10.1053/j.semperi.2015.06.009. [DOI] [PubMed] [Google Scholar]
- 13.Seghieri G, Anichini R, De Bellis A, Alviggi L, Franconi F, Breschi MC. Relationship between gestational diabetes mellitus and low maternal birth weight. Diabetes Care. 2002;25:1761–5. 10.2337/diacare.25.10.1761. [DOI] [PubMed] [Google Scholar]
- 14.He L-R, Yu L, Guo Y. Birth weight and large for gestational age trends in offspring of pregnant women with gestational diabetes mellitus in southern China, 2012-2021. Front Endocrinol. 2023;14:1166533. 10.3389/fendo.2023.1166533. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.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;37:2316731. 10.1080/14767058.2024.2316732. [DOI] [PubMed] [Google Scholar]
- 16.Wu Y, Xiao H, Chen L, Qin J, Wang T. Pre-pregnancy body mass index and risk of macrosomia: glycemic status-specific thresholds and subgroup interactions in a prospective cohort. Front Nutr. 2025;12:1633088. 10.3389/fnut.2025.1633088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sina M, Cade TJ, Flack J, Nolan CJ, Rajagopal R, Wong V, et al. Antenatal models of care for women with gestational diabetes mellitus: vignettes from an international meeting. Aust N Z J Obstet Gynaecol. 2020;60:720–8. 10.1111/ajo.13144. [DOI] [PubMed] [Google Scholar]
- 18.Cooray SD, Boyle JA, Soldatos G, Allotey J, Wang H, Fernandez-Felix BM, et al. Development, validation and clinical utility of a risk prediction model for adverse pregnancy outcomes in women with gestational diabetes: The PeRSonal GDM model. EClinicalMedicine. 2022;52:101637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mansournia MA, Altman DG. Population attributable fraction. BMJ. 2018;360:k757. 10.1136/bmj.k757. [DOI] [PubMed] [Google Scholar]
- 20.Nishihama Y, Nakayama SF, Tabuchi T. Population attributable fraction of risk factors for low birth weight in the Japan Environment and Children’s Study. Environ Int. 2022;170:107560. 10.1016/j.envint.2022.107560. [DOI] [PubMed] [Google Scholar]
- 21.Eckler R, Liu YD, Pooni A, Yong PJ, Lisonkova S, Bedaiwy MA. Maternal and perinatal outcomes in primiparous singleton pregnancies conceived with assisted reproductive technology in British Columbia. Eur J Obstet Gynecol Reprod Biol. 2023;291:106–11. 10.1016/j.ejogrb.2023.10.010. [DOI] [PubMed] [Google Scholar]
- 22.Chen Y, Zhong T, Song X, Zhang S, Sun M, Liu 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. 10.1136/bmjpo-2023-001931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tunçalp Ӧ, Pena‐Rosas J, Lawrie T, Bucagu M, Oladapo O, Portela A, et al. WHO recommendations on antenatal care for a positive pregnancy experience—going beyond survival. BJOG: An International Journal of Obstetrics & Gynaecology. 2017;124:860–2. 10.1111/1471-0528.14599. [DOI] [PubMed] [Google Scholar]
- 24.Sun M, Luo M, Wang T, Wei J, Zhang S, Shu J, et al. Effect of the interaction between advanced maternal age and pre-pregnancy BMI on pre-eclampsia and GDM in Central China. BMJ Open Diabetes Res Care. 2023;11:e003324. 10.1136/bmjdrc-2023-003324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Li R, Zhang J, Gao Y, Zhang Y, Lan X, Dong H, et al. Duration and quality of sleep during pregnancy are associated with preterm birth and small for gestational age: a prospective study. Int J Gynaecol Obstet. 2021;155:505–11. 10.1002/ijgo.13584. [DOI] [PubMed] [Google Scholar]
- 26.National Health Commission of the People’s Republic of China. WS/T 800–2022 Growth evaluation criteria for newborns of different gestational ages at birth. 2022. https://www.nhc.gov.cn/wjw/c100311/202208/07787ef64ba34fe1bc8bbdae9fd0d4e5.shtml.
- 27.Wu Y, Zheng H, Xu F, Guo J, Liu Z, Wang S, et al. Population attributable fractions for risk factors and disability burden of dementia in Jiangxi Province, China: a cross-sectional study. BMC Geriatr. 2022;22:811. 10.1186/s12877-022-03507-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mukadam N, Sommerlad A, Huntley J, Livingston G. Population attributable fractions for risk factors for dementia in low-income and middle-income countries: an analysis using cross-sectional survey data. Lancet Glob Health. 2019;7:e596-603. 10.1016/S2214-109X(19)30074-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Letson GW, French J, Ricketts S, Trierweiler K, Juhl A, Gujral I, et al. Utility of population attributable fraction assessment in guiding interventions to reduce low birthweight in the high-altitude state of Colorado. Matern Child Health J. 2016;20:2457–64. 10.1007/s10995-016-2037-6. [DOI] [PubMed] [Google Scholar]
- 30.Siza JE. Risk factors associated with low birth weight of neonates among pregnant women attending a referral hospital in northern Tanzania. Tanzan J Health Res. 2008;10:1–8. 10.4314/thrb.v10i1.14334. [DOI] [PubMed] [Google Scholar]
- 31.Rogne T, Gill D, Liew Z, Shi X, Stensrud VH, Nilsen TIL, et al. Mediating factors in the association of maternal educational level with pregnancy outcomes: a Mendelian randomization study. JAMA Netw Open. 2024;7:e2351166. 10.1001/jamanetworkopen.2023.51166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Zhu Y, Zhang H, Qi J, Liu Y, Yan Y, Wang T, et al. Evaluating causal influence of maternal educational attainment on offspring birthweight via observational study and Mendelian randomization analyses. SSM - Population Health. 2024;25:101587. 10.1016/j.ssmph.2023.101587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Berete F, Gisle L, Demarest S, Charafeddine R, Bruyère O, Van Den Broucke S, et al. Does health literacy mediate the relationship between socioeconomic status and health related outcomes in the Belgian adult population? BMC Public Health. 2024;24:1182. 10.1186/s12889-024-18676-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Aradhya S, Tegunimataka A, Kravdal Ø, Martikainen P, Myrskylä M, Barclay K, et al. Maternal age and the risk of low birthweight and pre-term delivery: a pan-Nordic comparison. Int J Epidemiol. 2023;52:156–64. 10.1093/ije/dyac211. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Li H, Nawsherwan, Fan C, Mubarik S, Nabi G, Ping YX. The trend in delayed childbearing and its potential consequences on pregnancy outcomes: a single center 9-years retrospective cohort study in Hubei. China BMC Pregnancy Childbirth. 2022;22:514. 10.1186/s12884-022-04807-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ye X, Baker PN, Tong C. The updated understanding of advanced maternal age. Fundam Res. 2024;4:1719–28. 10.1016/j.fmre.2023.09.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cao J, Xu W, Liu Y, Zhang B, Zhang Y, Yu T, et al. Trends in maternal age and the relationship between advanced age and adverse pregnancy outcomes: a population-based register study in Wuhan, China, 2010–2017. Public Health. 2022;206:8–14. 10.1016/j.puhe.2022.02.015. [DOI] [PubMed] [Google Scholar]
- 38.Tamura T, Picciano MF. Folate and human reproduction. Am J Clin Nutr. 2006;83:993–1016. 10.1093/ajcn/83.5.993. [DOI] [PubMed] [Google Scholar]
- 39.Bailey LB, Gregory JF. Folate metabolism and requirements. J Nutr. 1999;129:779–82. 10.1093/jn/129.4.779. [DOI] [PubMed] [Google Scholar]
- 40.Bleker OP, Buimer M, Van Der Post JAM, Van Der Veen F. Ted (G.J.) Kloosterman: on intrauterine growth. The significance of prenatal care. Studies on birth weight, placental weight and placental index. Placenta. 2006;27:1052–4. 10.1016/j.placenta.2006.01.001. [DOI] [PubMed] [Google Scholar]
- 41.Zhou Q, Acharya G, Zhang S, Wang Q, Shen H, Li X. A new perspective on universal preconception care in China. Acta Obstet Gynecol Scand. 2016;95:377–81. 10.1111/aogs.12865. [DOI] [PubMed] [Google Scholar]
- 42.Zhou Q, Zhang S, Wang Q, Shen H, Tian W, Chen J, et al. China’s community-based strategy of universal preconception care in rural areas at a population level using a novel risk classification system for stratifying couples´ preconception health status. BMC Health Serv Res. 2016;16:689. 10.1186/s12913-016-1930-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Pielage M, El Marroun H, Odendaal HJ, Willemsen SP, Hillegers MHJ, Steegers EAP, et al. Alcohol exposure before and during pregnancy is associated with reduced fetal growth: the Safe Passage Study. BMC Med. 2023;21:318. 10.1186/s12916-023-03020-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Cui H, Gong T-T, Liu C-X, Wu Q-J. Associations between passive maternal smoking during pregnancy and preterm birth: evidence from a meta-analysis of observational studies. PLoS ONE. 2016;11:e0147848. 10.1371/journal.pone.0147848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Tatum M. China’s three-child policy. Lancet. 2021;397:2238. 10.1016/S0140-6736(21)01295-2. [DOI] [PubMed] [Google Scholar]
- 46.Norton S, Matthews FE, Barnes DE, Yaffe K, Brayne C. Potential for primary prevention of Alzheimer’s disease: an analysis of population-based data. Lancet Neurol. 2014;13:788–94. 10.1016/S1474-4422(14)70136-X. [DOI] [PubMed] [Google Scholar]
- 47.Wu Y, Chen X, Hu S, Zheng H, Chen Y, Liu J, et al. The impact of potentially modifiable risk factors for stroke in a middle-income area of China: a case-control study. Front Public Health. 2022;10:815579. 10.3389/fpubh.2022.815579. [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
Data Availability Statement
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy.