Abstract
Background
Fetal growth restriction (FGR) is a common obstetric complication where a fetus fails to reach its genetically determined growth potential. Current diagnostic methods rely on population-based fetal biometric percentiles, which struggle to distinguish pathological FGR from healthy small-for-gestational-age (SGA) infants and may miss cases of growth restriction in fetuses with a high genetic growth potential who fall above conventional SGA thresholds. This study aimed to develop a novel diagnostic framework for FGR by incorporating individualized genetic potential modeling to improve precision in identification and risk stratification.
Methods
Using data from the Shanghai Birth Cohort (1,806 mother-infant pairs), fetal growth potential was calculated using a polygenic risk score for birth weight derived from fetal genome-wide association studies. A new metric, “FGR_degree,” was developed as the difference between standardized genetic growth potential and actual birth weight Z-score. Fetuses were stratified into risk categories using a dual-threshold approach: the top 10% of FGR_degree values combined with SGA criteria (birth weight < 10th percentile). The high-risk “trueFGR” group included fetuses meeting both criteria. Maternal characteristics, pregnancy complications, neonatal outcomes, and neurodevelopmental outcomes at six months were analyzed using regression models.
Results
Among 1,806 mother-child pairs, 181 newborns were classified into the top 10% FGR_degree group, including 46 trueFGR cases. Compared to SGA infants, trueFGR infants exhibited consistent growth impairments: lower birth weight (2,599.7 g vs. 2,720.9 g, P = 0.02), reduced GROW centiles (4.8% vs. 5.7%, P < 0.001), and disproportionate body proportionality. NICU admission rates were higher in trueFGR (17.4% vs. 12.5%, P = 0.59). Neurodevelopmental outcomes showed a dose-response relationship between FGR_degree and 6-month deficits, with trueFGR conferring the highest adjusted risk for gross motor delay (OR 5.04, 95% CI [2.17, 11.69] vs. 3.51, 95% CI [1.85, 6.67] in SGA). Sensitivity analyses confirmed enhanced risk prediction by combining Top10_FGR_degree with SGA.
Conclusions
Integrating fetal genomic data with birth weight enabled the development of a novel FGR severity metric. Combining FGR_degree with SGA classifications improved risk stratification, identifying high-risk fetuses more precisely. TrueFGR was associated with maternal complications, adverse outcomes, and neurodevelopmental delays at six months. Future research should explore non-invasive fetal DNA profiling and multi-ethnic validation.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12884-026-08785-z.
Keywords: Birth weight, Fetal growth restriction, Genetic potential, Polygenic risk score, Small-for-gestational-age.
Introduction
Fetal growth restriction (FGR) affects 5–10% of uncomplicated pregnancies and up to 25% of high-risk pregnancies, leading to increased perinatal morbidity, mortality, and long-term metabolic and developmental risks [1–3].
Diagnosing FGR is challenging due to the lack of precise criteria. Current definitions rely on fetal size (e.g., estimated fetal weight [EFW] or abdominal circumference [AC]) below the 3rd, 5th, or 10th percentile, often combined with abnormal fetoplacental Doppler parameters [4–7]. Infants with birth weight (BW) below the 10th percentile are termed “small for gestational age” (SGA), but these criteria fail to distinguish true FGR from healthy SGA infants who are constitutionally small. Additionally, growth-restricted infants may have BWs above the SGA threshold, complicating diagnosis [8].
The evolution of fetal growth standards has shifted from population-based to individualized approaches. While widely used charts like INTERGROWTH-21st [9] and WHO fetal growth standards [10] assume uniform growth potential, they overlook intrinsic variability [11]. The Gestation Related Optimal Weight (GROW) charts represent an advancement by incorporating maternal height, weight, parity, and ethnicity to create customized growth curves [12, 13]. However, GROW charts do not account for genetic determinants of fetal growth, limiting their ability to fully quantify genetically programmed growth potential [11, 14, 15].
Fetal growth is influenced by both genetic and environmental factors. Genome-wide association studies (GWAS) have identified numerous genetic loci associated with BW, with polymorphic single-nucleotide polymorphisms (SNPs) explaining 15.1% of BW variance [16–18]. To date, a large number of genetic loci have been identified as susceptibility markers for BW by GWAS [16, 19, 20]. The polygenic risk score (PRS) captures the cumulative effects of these genetic variants, providing a tool to assess genetic growth potential [21, 22].
This study aims to develop an FGR severity index by integrating fetal PRS for BW (PRSBW) with actual BW measurements, proposing a dual-parameter diagnostic framework validated through pregnancy outcomes and child neurodevelopmental assessments.
Materials and methods
Study population
The Shanghai Birth Cohort (SBC) recruited women who were trying to conceive or in early pregnancy from six participating hospitals in Shanghai, China, between 2013 and 2016. Detailed enrollment criteria and recruitment process have been described previously [23]. In the present study, mother-child pairs were eligible for inclusion if they met the following criteria: (1) maternal age ranged from 20 to 40 years and the pregnancies resulted in singleton, live births; (2) fetal genomic data were available; (3) BW, sex, and gestational age (GA) were available. Ultimately, 1,806 mother-child pairs were included in our study.
Genotyping and polygenic risk score
Offspring DNA was obtained from umbilical cord blood samples. Details of genotyping and quality control are described in Supplementary materials.
A total of six PRSs were replicated. We searched the PGS Catalog, GWAS Catalog, and the Early Growth Genetics Consortium (EGG) database for relevant polygenic scores associated with BW. From the PGS Catalog, we identified five PRSs, all derived from European populations. Four of these had been evaluated across multi-ancestry populations, including Asian populations, with accession IDs PGS001226, PGS001892, PGS002105, and PGS003494. From the GWAS Catalog, we identified 16 full summary statistics; however, only one had been developed and replicated across multi-ancestry populations, including the Asian population (GWAS Catalog accession ID: GCST005146). Additionally, we included one polygenic score derived from a meta-analysis in the EGG database [24].
PRS calculation was conducted using the weighted-allele count approach, which assumes an independent effect across the genetic risk loci, drawn from the PGS catalog (https://www.pgscatalog.org/, accessed on 16 November 2024). Full summary statistics were drawn from the GWAS catalog and the EGG database. (https://www.ebi.ac.uk/gwas/studies/GCST005146, accessed on 16 November 2024; http://egg-consortium.org/birth-weight-2019.html, accessed on 16 November 2024).
The formula used to calculate PRS is
 |
where PGSi, is the PGS for each individual, βj is the effect size associated with each SNP, Gij is the genotype value (0, 1, 2) for each SNP and for each individual in the study sample. PRS was conducted using PLINK (v1.90b7.2).
Validation and selection of PRS for BW
We systematically replicated and evaluated the explanatory capacity of six PRS models (PRSEGG, PRSGGST005146, PRSPGS001226, PRSPGS001892, PRSPGS002105, and PRSPGS003494) for BW and BWzscore. Linear regression analysis revealed that the PRSPGS003494 model performed best, explaining 4.6% of the variance in BWzscore (R² = 0.046; Fig. 1). This PRS integrates 881,151 significant genetic variants, was used for subsequent analyses. Moreover, the combined model, which integrates clinical predictors from the GROW model and PRSPGS003494, demonstrated robust generalizability across populations. PRS significantly enhanced predictive accuracy when combined with clinical factors in all subgroups, with absolute ΔR² values ranging from 0.005 to 0.043 (Fig. S1). Thus, PRSPGS003494 was selected as the core indicator of fetal genetic growth potential in this study.
Fig. 1.
Explained variance in BWzscore by different PRS models
Birth outcome measurements
Trained nurses measured BW according to a standardized protocol. The GA represents the best obstetric estimate, confirmed by first-trimester ultrasound. SGA was defined by the individually customized GROW standard, with customization coefficients derived from BW data to adjust the expected BW, accounting for key physiological determinants (maternal height, weight, parity, ethnicity and fetal sex) while excluding pathological influences, using population-specific coefficients developed for over 120 ethnic groups through the global GROW standards software (http://www.gestation.net) [13]. BWzscore was calculated using Generalized Additive Models for Location, Scale, and Shape (GAMLSS) [25]. Additionally, birth length (BL) and weight-for-length (WFL) were measured, with WFL z-scores calculated using WHO Anthro software (version 3.2.2.1, https://www.who.int/tools/child-growth-standards/software, accessed on November 1, 2024).
Neurodevelopment assessment
Neurodevelopment in children was evaluated using the Ages and Stages Questionnaire (ASQ) at 6 months. It consists of five domains: communication, gross motor function, fine motor function, problem-solving ability, and personal social skills [26–28]. A score in each domain was categorized based on the standard deviation (SD) from the mean, using a cutoff of 1 SD below the mean to indicate suspected developmental delay, as validated in Chinese populations [28, 29].
Statistical analysis
As shown in Figure S2, covariates were selected based on the literature and guided by the directed acyclic graphs. The details of covariates are described in the Supplementary materials.
See study diagram in Fig. 2, we developed a novel measure to quantify the difference between genetically predicted BW and observed BW. Fetal growth potential was estimated using the fetal PRS for BW (PRSBW). Observed BW was standardized as z-scores (BWzscore) using GAMLSS [25] adjusted for GA and sex. The GAMLSS method, based on the Box-Cox t distribution, models kurtosis and asymmetry while generating smooth BW percentile curves.
Fig. 2.
Study Diagram
To enable direct comparison, both indicators were scaled to the same dimension. We then defined the FGR_degree as FGR_degree = scale(PRSBW) - scale(BWzscore), to quantify the disparity between genetic growth potential (PRSBW) and observed BW (BWzscore). FGR_degree > 0 indicates fetal growth below the growth potential. Fetuses in the top 10% of FGR_degree values (Top10_FGR_degree) and classified as SGA are labeled “trueFGR,” indicating a high risk of not reaching genetic growth potential and requiring close monitoring. Infants with both BWzscore and FGR_degree values between the 25th and 75th percentiles were defined as the control group.
Maternal baseline characteristics, pregnancy complications, neonatal outcomes [anthropometrics, neonatal intensive care unit (NICU) admission], and child neurodevelopmental outcomes were then analyzed across different diagnostic criteria. Linear and logistic regression models were employed to analyze the association between FGR indicators (FGR_degree and trueFGR) and ASQ results.
Continuous variables were summarized as means ± SD, and categorical variables as frequencies and proportions. Group differences were evaluated using the t-test or Wilcoxon rank-sum test for continuous variables and Fisher’s exact test for categorical variables.
The GROW centiles of BW were calculated using the Individual centile calculator (version 2.1.6.1), provided by the Perinatal Institute through the Gestation Network (www.gestation.net). All statistical analyses were performed using R software (R version 4.4.0).
Results
Baseline characteristics of study participants
A total of 1,806 mother-child pairs were included in this study. Among them, there were 104 (5.8%) cases of SGA. 181 (10.0%) newborns were classified as Top10_FGR_degree, with 46 (2.5%) cases confirmed as trueFGR (meeting both Top10_FGR_degree and SGA criteria). The mean, median, and range of FGR_degree were 0.00 ± 1.25, 0.03, and − 4.51 to 4.60, respectively. Also included was a control group of 532 (29.5%) newborns.
As shown in Table 1, while pre-pregnancy BMI was similar across all groups, the trueFGR group had significantly lower gestational weight gain than the controls (11.0 ± 3.5 kg vs. 12.5 ± 4.0 kg; P = 0.01), with a trend toward higher rates of inadequate weight gain (< 10 kg: 30.4% vs. 24.6%). Compared to the SGA cases, trueFGR mothers exhibited marginally elevated rates of inadequate weight gain (30.4% vs. 28.8%). No significant differences were found in demographic characteristics (age, ethnicity, parity), lifestyle factors (smoking, alcohol consumption, physical activity), or socioeconomic status among the groups (P > 0.05).
Table 1.
Baseline characteristics and pregnancy outcomes of study participants
| Variable | Levels | All (N = 1,806) |
Control (N = 532) |
SGA (N = 104) |
Top10_FGR degree (N = 181) |
trueFGR (N = 46) |
P* | P # |
|---|---|---|---|---|---|---|---|---|
| Age, years | Mean ± SD | 28.7 ± 3.5 | 28.9 ± 3.5 | 28.9 ± 3.5 | 28.4 ± 3.3 | 28.8 ± 3.3 | 0.83 | 0.92 |
| Age group, years | 20 ~ 30 | 1129 (62.5%) | 322 (60.5%) | 61 (58.7%) | 119 (65.7%) | 26 (56.5%) | 0.07 | 0.27 |
| 30 ~ 35 | 556 (30.8%) | 170 (32.0%) | 38 (36.5%) | 56 (30.9%) | 20 (43.5%) | |||
| ≥ 35 | 121 (6.7%) | 40 (7.5%) | 5 (4.8%) | 6 (3.3%) | 0 (0.0%) | |||
| Ethnic group | Han | 1785 (98.8%) | 526 (98.9%) | 104 (100.0%) | 180 (99.4%) | 46 (100.0%) | 1.00 | 1.00 |
| Parity | 0 | 1568 (86.8%) | 460 (86.5%) | 94 (90.4%) | 160 (88.4%) | 44 (95.7%) | 1.00 | 1.00 |
| Maternal height, cm | Mean ± SD | 162.2 ± 4.8 | 162.1 ± 4.6 | 162.7 ± 4.7 | 161.9 ± 4.6 | 161.7 ± 4.2 | 0.59 | 0.22 |
| Pre-pregnancy Weight, kg | Mean ± SD | 56.0 ± 8.1 | 55.5 ± 7.7 | 55.7 ± 8.0 | 53.6 ± 7.3 | 54.4 ± 7.7 | 0.36 | 0.35 |
| Pre-pregnancy BMI(kg/m2) | Mean ± SD | 21.2 ± 2.8 | 21.1 ± 2.6 | 21.0 ± 2.8 | 20.5 ± 2.7 | 20.8 ± 2.9 | 0.48 | 0.66 |
| < 18.5 | 279 (15.4%) | 81 (15.2%) | 22 (21.2%) | 49 (27.1%) | 13 (28.3%) | 0.13 | 0.81 | |
| 18.5 ~ 24 | 1250 (69.2%) | 371 (69.7%) | 64 (61.5%) | 112 (61.9%) | 26 (56.5%) | |||
| 24 ~ 28 | 237 (13.1%) | 73 (13.7%) | 16 (15.4%) | 17 (9.4%) | 6 (13.0%) | |||
| ≥ 28 | 40 (2.2%) | 7 (1.3%) | 2 (1.9%) | 3 (1.7%) | 1 (2.2%) | |||
| Gestational weight gain, kg | Mean ± SD | 12.6 ± 4.1 | 12.5 ± 4.0 | 11.7 ± 3.5 | 11.7 ± 3.8 | 11.0 ± 3.5 | 0.01 | 0.27 |
| Gestational weight gain group, kg | Less than 10 kg | 441 (24.4%) | 131 (24.6%) | 30 (28.8%) | 54 (29.8%) | 14 (30.4%) | 0.08 | 0.46 |
| 10 ~ 15 kg | 922 (51.1%) | 279 (52.4%) | 57 (54.8%) | 98 (54.1%) | 28 (60.9%) | |||
| Greater than 15 kg | 443 (24.5%) | 122 (22.9%) | 17 (16.3%) | 29 (16.0%) | 4 (8.7%) | |||
| Maternal education | Primary education | 24 (1.3%) | 5 (0.9%) | 0 (0.0%) | 2 (1.1%) | 0 (0.0%) | 0.61 | 1.00 |
| High school/Technical school | 122 (6.8%) | 30 (5.6%) | 10 (9.6%) | 12 (6.6%) | 2 (4.3%) | |||
| College/University | 1406 (77.9%) | 408 (76.7%) | 80 (76.9%) | 137 (75.7%) | 39 (84.8%) | |||
| Master and above | 254 (14.1%) | 89 (16.7%) | 14 (13.5%) | 30 (16.6%) | 5 (10.9%) | |||
| Maternal economic status | ||||||||
| Affluent | 360 (19.9%) | 112 (21.1%) | 22 (21.2%) | 31 (17.1%) | 8 (17.4%) | 0.29 | 1.00 | |
| Comfortable | 1266 (70.1%) | 368 (69.2%) | 75 (72.1%) | 138 (76.2%) | 37 (80.4%) | |||
| Struggling | 165 (9.1%) | 51 (9.6%) | 7 (6.7%) | 9 (5.0%) | 1 (2.2%) | |||
| Impoverished | 15 (0.8%) | 1 (0.2%) | 0 (0.0%) | 3 (1.7%) | 0 (0.0%) | |||
| Maternal life factors | ||||||||
| smoke_active | 45 (2.5%) | 14 (2.6%) | 3 (2.9%) | 0 (0%) | 0 (0%) | 0.74 | 1.00 | |
| smoke_passive | 860 (47.6%) | 248 (46.6%) | 47 (45.2%) | 70 (38.7%) | 19 (41.3%) | 0.59 | 0.79 | |
| smoke_all | 873 (48.3%) | 254 (47.7%) | 47 (45.2%) | 70 (38.7%) | 19 (41.3%) | 0.49 | 0.79 | |
| Alcohol_consumption | 249 (13.8%) | 88 (16.5%) | 14 (13.5%) | 18 (9.9%) | 7 (15.2%) | 0.98 | 0.98 | |
| Physical exercise (MET-minutes/week) | Mean ± SD | 786.7 ± 900.9 | 788.5 ± 820.1 | 660.5 ± 813.3 | 773.7 ± 954.3 | 705.3 ± 905.6 | 0.51 | 0.76 |
| Physical exercise level | Low | 443 (24.5%) | 123 (23.1%) | 28 (26.9%) | 46 (25.4%) | 10 (21.7%) | 0.30 | 0.76 |
| Medium | 936 (51.8%) | 267 (50.2%) | 61 (58.7%) | 95 (52.5%) | 28 (60.9%) | |||
| High | 427 (23.6%) | 142 (26.7%) | 15 (14.4%) | 40 (22.1%) | 8 (17.4%) | |||
| Pregnancy complications | ||||||||
| GDM | 203 (11.2%) | 64 (12%) | 13 (12.5%) | 15 (8.3%) | 2 (4.3%) | 0.18 | 0.22 | |
| HDP | 47 (2.6%) | 14 (2.6%) | 4 (3.8%) | 7 (3.9%) | 2 (4.3%) | 0.83 | 1.00 | |
| PE | 32 (1.8%) | 5 (0.9%) | 7 (6.7%) | 6 (3.3%) | 3 (6.5%) | 0.01 | 1.00 | |
| Neonatal characteristics | ||||||||
| IVF-ET | 48 (2.7%) | 19 (3.6%) | 2 (1.9%) | 4 (2.2%) | 1 (2.2%) | 0.94 | 1.00 | |
| GA | Mean ± SD | 39.0 ± 1.4 | 39.1 ± 1.3 | 39.2 ± 1.4 | 38.9 ± 1.6 | 38.9 ± 1.5 | 1.00 | 0.76 |
| Preterm birth | 83 (4.6%) | 20 (3.8%) | 2 (1.9%) | 12 (6.6%) | 2 (4.3%) | 0.38 | 0.20 | |
| Fetal sex | male | 909 (50.3%) | 256 (48.1%) | 53 (51%) | 96 (53%) | 24 (52.2%) | 0.71 | 1.00 |
| BW, g | Mean ± SD | 3396.3 ± 443.5 | 3385.4 ± 258.7 | 2720.9 ± 279.3 | 2975.0 ± 415.4 | 2599.7 ± 307.3 | < 0.001 | 0.02 |
| BL, cm | Mean ± SD | 49.9 ± 1.3 | 49.9 ± 1.0 | 48.8 ± 1.8 | 49.1 ± 1.7 | 48.4 ± 2.3 | < 0.001 | 0.30 |
| BWzscore | Mean ± SD | 0.0 ± 1.0 | 0.0 ± 0.4 | -1.9 ± 0.5 | -1.1 ± 0.9 | -2.1 ± 0.5 | < 0.001 | 0.01 |
| WFLzscore | Mean ± SD | 0.1 ± 1.1 | 0.2 ± 0.8 | -1.5 ± 1.1 | -0.8 ± 1.3 | -1.9 ± 1.2 | < 0.001 | 0.13 |
| GROW centile, % | Mean ± SD | 56.9 ± 28.4 | 58.4 ± 15.8 | 5.7 ± 2.8 | 27.6 ± 22.5 | 4.8 ± 3.1 | < 0.001 | 0.08 |
| FGR degree | Mean ± SD | 0.0 ± 1.3 | -0.0 ± 0.5 | 1.5 ± 1.0 | 2.2 ± 0.5 | 2.4 ± 0.7 | < 0.001 | < 0.001 |
| NICU admission | 187 (10.4%) | 50 (9.4%) | 13 (12.5%) | 22 (12.2%) | 8 (17.4%) | 0.14 | 0.59 | |
SD standard deviation, BMI body mass index, GDM gestational diabetes mellitus, HDP hypertensive disorders of pregnancy, PE preeclampsia, IVF-ET in vitro fertilization and embryo transfer, GA gestational age at delivery, BW birthweight, BL birth length, WFL weight-for-length, NICU neonatal intensive care unit
GROW BW centiles are calculated based on the using the Individual Centile Calculator
P*: trueFGR vs. Control; P#: trueFGR vs. SGA; P values less than 0.05, indicating statistical significance, are presented in italic
Pregnancy outcomes of study participants
As shown in Table 1, the trueFGR group exhibited a significantly higher incidence of preeclampsia (PE) compared to the controls (6.5% vs. 0.9%, P = 0.01). Rates of gestational diabetes (GDM) and hypertensive disorders (HDP) showed no significant differences across groups (P > 0.05). The trueFGR newborns displayed consistent growth impairments: BW was significantly lower than the controls (2600 ± 307 g vs. 3385 ± 259 g, P < 0.001) and significantly lower than SGA (2600 g vs. 2721 g, P = 0.02). Growth restriction severity was further evidenced by critically diminished weight centiles (4.8 ± 3.1% vs. 58.4 ± 15.8% in controls and 5.7 ± 2.8% in SGA, P < 0.001), elevated FGR_degree (2.4 ± 0.7 vs. −0.0 ± 0.5 in controls and 1.5 ± 1.0 in SGA, P < 0.001), and abnormal body proportionality reflected in weight-for-length z-scores (WFLzscore: −1.9 ± 1.2 vs. 0.2 ± 0.8 in controls, P < 0.001; vs. −1.5 ± 1.1 in SGA, P = 0.13). Although NICU admission rates trended to be higher in trueFGR (17.4% vs. 9.4% in controls and 12.5% in SGA), these differences did not reach statistical significance (P = 0.14–0.59). Collectively, these findings underscore trueFGR as a distinct clinical entity characterized by PE-associated placental dysfunction and severe, multifaceted FGR that exceeds conventional SGA diagnostic thresholds.
Association between FGR and adverse neurodevelopmental outcomes
FGR_degree exhibited a significant inverse association with 6-month ASQ scores, particularly in the personal social (unadjusted β = -0.63 ± 0.30, P = 0.034) and gross motor domains (unadjusted β = -0.68 ± 0.26, P = 0.011). After adjustment for covariates, these associations intensified: personal social scores decreased by -0.94 ± 0.40 (P = 0.018), and gross motor scores declined by -0.92 ± 0.36 (P = 0.010). These findings underscore a dose-dependent relationship between the FGR_degree and early developmental deficits. The influence of FGR_degree on neurodevelopment is greater than that of fetal size, with its effect decreasing as fetal size increases (Tables S1 and S2).
We assessed neurodevelopmental delay risks across multiple domains in infants classified as trueFGR, Top10_FGR_degree, and SGA compared to controls. Figure 3 demonstrates that in unadjusted analyses, trueFGR infants showed the greatest risk elevation for gross motor delay (OR = 4.17, 95% CI: 1.96–8.87, P < 0.001), followed by SGA (OR = 3.21, 95% CI: 1.81–5.70, P < 0.001) and Top10_FGR_degree (OR = 1.87, 95% CI: 1.17–2.99, P < 0.01). Following covariate adjustment, the risk magnitude increased substantially for trueFGR (OR = 5.04, 95% CI: 2.17–11.69, P < 0.001), maintained significance for SGA (OR = 3.51, 95% CI: 1.85–6.67, P < 0.001), and remained clinically relevant for Top10_FGR_degree (OR = 2.73, 95% CI: 1.45–5.15, P < 0.01). These results highlight trueFGR as the highest-risk phenotype for gross motor delay, while other domains showed no significant associations across groups.
Fig. 3.
Neurodevelopmental delay risk across groups at 6 months. ASQ: Ages and Stages Questionnaires; The cases with both birth weight z-scores and FGR_degree within the 25th to 75th percentiles were considered the control group; Model adjusted for the following covariates: maternal age, race, parity, IVF, pre-pregnancy BMI level, fetal sex, weight gain level during pregnancy, GDM, HDP, PE, smoke, alcohol consumption, physical exercise level, financial status, education level and PRS
Sensitivity analysis
We conducted sensitivity analyses by removing preterm birth participants (n = 83). Fig. S3 shows that consistent with the primary results, Top10_FGR_degree had a higher predictability for adverse neurodevelopment, and combining Top10_FGR_degree with SGA improves risk prediction for SGA.
Discussion
This study is the first in FGR research to integrate genetic predisposition with clinical phenotyping, thereby advancing our understanding of FGR. By incorporating fetal genetic data with BW, we developed a novel index—FGR_degree. This index quantifies genetic potential, defines FGR severity independently of small-for-gestational-age (SGA) criteria, and captures the interaction between genetic potential and actual fetal size at birth, particularly in relation to adverse developmental outcomes. Our findings demonstrate that incorporating genetic growth potential into FGR definitions significantly improves the identification of high-risk fetuses affected by pathological growth restriction.
The FGR_degree, a polygenic risk-informed metric, represents a paradigm shift in quantifying FGR severity by capturing the genetic-environmental mismatch underlying pathological growth restriction. This aligns with evidence that PRS enhances fetal growth prediction beyond traditional markers [30]. Our dual-axis classification, integrating FGR_degree and SGA criteria, improves predictive precision and aligns with the DOHaD framework, addressing diagnostic ambiguity highlighted in recent guidelines [7, 31].
The trueFGR group emerges as a distinct clinical entity linked to placental dysfunction, with elevated PE risk reinforcing its association with maternal vascular malperfusion [32]. TrueFGR infants exhibit severe growth compromise and abnormal body proportionality, reflecting chronic fetal malnutrition and systemic metabolic consequences, termed “placental starvation syndrome” [33]. Higher NICU admission rates in trueFGR infants suggest a severe clinical spectrum requiring intensified neonatal surveillance, consistent with studies linking growth restriction severity to morbidity risk [34]. Epidemiological data show that antenatally undiagnosed FGR contributes to about 10% of perinatal mortality and 52% of stillbirths, with neonates facing risks of birth asphyxia, hypoglycemia, and metabolic acidosis [35, 36]. Chronic hypoxia-ischemia impairs organ development and metabolic programming, increasing vulnerability [37–40]. Furthermore, longitudinal associations with adverse neurodevelopmental outcomes further support the biological plausibility of trueFGR, linking intrauterine nutrient deprivation to impaired neurogenesis [41–43].
The findings highlight the clinical potential of integrating genetic risk stratification into FGR diagnosis and management. The FGR_degree metric enhances the identification of high-risk fetuses, addressing the limitations of conventional anthropometric criteria. This approach could improve neonatal surveillance and intervention strategies, particularly for trueFGR infants who exhibit a more severe clinical spectrum. Additionally, the study underscores the importance of considering genetic potential in multidimensional FGR evaluation, which could refine risk stratification and inform personalized prenatal care. However, the clinical translation of this approach requires further validation in diverse healthcare settings and longitudinal investigations to elucidate long-term neurodevelopmental and metabolic outcomes.
Accurately measuring fetal growth potential has long been a clinical quest. The individually customized SGA definition (e.g., GROW) has, in part, addressed this issue by incorporating maternal weight, height, race, parity, and infant sex into the definition [13]. However, our study demonstrates that this approach still has room to improve. When we combined growth potential with the aforementioned clinical factors, the PRS significantly enhanced predictive accuracy.
It should be noted that among the six PRS evaluated, PRSPGS003494 was the top-performing model, yet its explanatory power for BW remained limited. This PGS003494 model, developed from European GWAS data using LDpred2-auto (optimizing 979,739 variants), has been validated in East Asian cohorts with a partial R² = 0.03. Its modest predictive capacity for BW likely reflects ancestry-specific genetic architecture (e.g., Eurocentric variant-weighting bias) and environmental confounders (e.g., socioeconomic deprivation, which attenuates the expression of genetic potential) [20]. Notably, in our study, PRSPGS003494 outperformed its East Asian validation benchmarks. Its independent contributions remained stable in the general population and term-born subgroups (R² = 0.046–0.052), although its diminished explanatory power in moderate-sized fetuses (R² = 0.009) suggests a preferential applicability for identifying genetic susceptibility to extreme BW phenotypes (e.g., macrosomia or low birth weight). Despite inherent limitations in PRSBW interpretability, the PRSPGS003494-derived FGR metric demonstrates enhanced risk stratification capability when synergistically integrated with SGA criteria. Its emerging discriminatory power in identifying high-risk fetuses highlights its clinical translation potential and underscores the importance of considering individual genetic potential for multidimensional evaluation when diagnosing FGR.
Future research should focus on multi-ancestry genome-wide association studies (GWAS) and gene-environment interaction modeling to refine genetic potential assessment, addressing Eurocentric biases and socioeconomic confounders [20]. The integration of non-invasive prenatal testing (NIPT)-based prenatal PRS with ultrasound-estimated fetal weight (EFW) should be explored to establish dynamic trajectories of FGR_degree and investigate their association with perinatal mortality [44, 45]. Additionally, longitudinal studies are needed to evaluate the long-term neurodevelopmental and metabolic consequences of FGR_degree-defined phenotypes, bridging the gap between research insights and actionable clinical interventions.
This study leverages the SBC, a large prospective birth cohort with longitudinal clinical phenotyping, serial covariate assessments, and robust genetic data, providing a unique platform. The consistency of sensitivity analyses across models reinforces the internal validity of our findings.
However, this study has several limitations. First, the small sample size of trueFGR infants may limit statistical power, necessitating validation in larger cohorts. Second, the small number of early-FGR cases limited our ability to explore genetic risk differences between early- and late-onset FGR. In addition, the ethnic homogeneity of our cohort (predominantly Han Chinese) also requires external validation in multi-ethnic populations to assess generalizability. Future studies should also explicitly compare FGR_degree–defined risk groups with infants classified as severely growth-restricted by size alone (e.g., birth weight below the 3rd percentile), to determine whether genetic information provides incremental risk stratification beyond extreme size-based thresholds. Finally, the PRS in this study relies on umbilical cord blood genotyping, limiting its application in prenatal management.
However, as FGR_degree is derived using observed birthweight and the PRS was derived from postnatal cord blood DNA, its current application remains retrospective rather than predictive. Therefore, the present work should be interpreted as a proof-of-concept demonstrating biological feasibility rather than an immediately deployable clinical screening tool. Future implementation of FGR_degree as a prenatal decision-support tool will require prospective evaluation using antenatal data—such as ultrasound-estimated fetal weight, Doppler findings or non-invasive prenatal testing (NIPT)-based PRS—to determine sensitivity, specificity and overall predictive value. Importantly, future studies with larger sample sizes and prospective antenatal data will also be required to evaluate the incremental predictive and clinical value of FGR_degree specifically within non-SGA populations, including formal assessments of sensitivity, specificity, and net reclassification beyond conventional size-based criteria. With advances in NIPT, fetal DNA can be collected from maternal peripheral blood as cell-free DNA (cfDNA). Studies have shown that fetal SNP genotypes obtained through targeted sequencing of maternal cfDNA exhibit up to 100% concordance with those derived from fetal genomic DNA in amniotic fluid, laying the foundation for prenatal PRS implementation [44, 45]. Future research should focus on multi-ancestry GWAS, gene-environment interaction modeling, and integrating cfDNA-based prenatal PRS with ultrasound-EFW to establish dynamic FGR_degree trajectories and investigate associations with perinatal mortality. Further validation is needed to translate this approach into clinical practice and explore its long-term neurodevelopmental and metabolic consequences.
Conclusions
This study integrates fetal genetic information and birth weight to develop a novel measure for assessing FGR severity, offering a comprehensive approach independent of arbitrary birth weight cutoffs. By combining FGR_degree with SGA criteria, our framework enhances risk prediction and enables more precise identification of high-risk FGR cases, underscoring the value of integrating genetic and growth data to improve screening and personalized interventions.
Supplementary Information
Acknowledgements
Not applicable.
Abbreviations
- AC
Abdominal circumference
- ASQ
Ages and Stages Questionnaire
- BL
Birth length
- BMI
Body mass index
- BW
Birth weight
- CI
Confidence interval
- EFW
Estimated fetal weight
- EGG
Early Growth Genetics Consortium
- FGR
Fetal growth restriction
- GA
Gestational age
- GAMLSS
Generalized additive models for location, scale, and shape
- GDM
Gestational diabetes
- GROW
Gestation Related Optimal Weight
- GWAS
Genome-wide association studies
- HDP
Hypertensive disorders
- NICU
Neonatal intensive care unit
- OR
Odds ratio
- PE
Preeclampsia
- PRS
Polygenic risk score
- SBC
Shanghai Birth Cohort
- SD
Standard deviation
- SGA
Small for gestational age
- SNP
Single-nucleotide polymorphism
- WFL
Weight-for-length
Authors’ contributions
Xiaoqing He: Data Curation, Formal analysis, Visualization, Software, Writing- Original draft preparation; Wen Jiang: Methodology, Software, Data Curation, Formal analysis, Writing - review & editing; Lei Wang: Data Curation, Writing - review & editing; Xiaojing Zeng: Methodology, Writing - review & editing; Fang Fang: Data Curation, Writing - review & editing; Yan Chen: Data Curation, Writing - review & editing; Luming Sun: Methodology, Writing - review & editing; Yun Huang: Conceptualization, Methodology, Investigation, Supervision, Writing - review & editing. Jun Zhang: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing - review & editing.
Funding
This work was supported by the Collaborative Innovation Program of the Shanghai Municipal Health Commission (2020CXJQ01), the National Natural Science Foundation of China (82404277), the Natural Science Foundation of Shanghai (24ZR1449200), the National Key Research and Development Program of China (2022YFC2704700), and the Xinhua Hospital, affiliated with Shanghai Jiao Tong University School of Medicine (XKPF2024A100).
Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the ethical principles of the Declaration of Helsinki.The study obtained ethical approval from the Institutional Review Boards of the Xinhua Hospital, Shanghai Jiao Tong University School of Medicine (XHEC-C-2013-001-3). Written informed consent was obtained from the participants in this study.
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.
Xiaoqing He and Wen Jiang contributed equally to this work and should be considered as first authors.
Contributor Information
Yun Huang, Email: huangyun@shsmu.edu.cn.
Jun Zhang, Email: junjimzhang@sjtu.edn.cn.
References
- 1.Subramanian S, Remadevi S, Subramaniam K. Prevalence, characteristics and maternal risk factors of small for gestational age fetuses in a tertiary care center from Kerala. New Indian J OBGYN. 2020;7(1):76–81. [Google Scholar]
- 2.Gaccioli F, Aye I, Sovio U, Charnock-Jones DS, Smith GCS. Screening for fetal growth restriction using fetal biometry combined with maternal biomarkers. Am J Obstet Gynecol. 2018;218(2s):S725–37. [DOI] [PubMed] [Google Scholar]
- 3.D’Agostin M, Di Morgia S, Vento C, Nobile G. Long-term implications of fetal growth restriction. World J Clin Cases. 2023;11(13):2855–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lees CC, Stampalija T, Baschat A, da Silva Costa F, Ferrazzi E, Figueras F, et al. ISUOG practice guidelines: diagnosis and management of small-for-gestational-age fetus and fetal growth restriction. Ultrasound Obstet Gynecol. 2020;56(2):298–312. [DOI] [PubMed] [Google Scholar]
- 5.Lausman A, Kingdom J. Intrauterine growth restriction: screening, diagnosis, and management. J Obstet Gynaecol Can. 2013;35(8):741–8. [DOI] [PubMed] [Google Scholar]
- 6.ACOG Practice bulletin no. 134: fetal growth restriction. Obstet Gynecol. 2013;121(5):1122–33. [DOI] [PubMed] [Google Scholar]
- 7.Fetal Growth Restriction. ACOG practice Bulletin, number 227. Obstet Gynecol. 2021;137(2):e16–28. [DOI] [PubMed] [Google Scholar]
- 8.ACOG Practice Bulletin No. 204: fetal growth restriction. Obstet Gynecol. 2019;133(2):e97–109. [DOI] [PubMed] [Google Scholar]
- 9.Villar J, Cheikh Ismail L, Victora CG, Ohuma EO, Bertino E, Altman DG, et al. International standards for newborn weight, length, and head circumference by gestational age and sex: the newborn Cross-Sectional study of the INTERGROWTH-21st project. Lancet. 2014;384(9946):857–68. [DOI] [PubMed] [Google Scholar]
- 10.Organization WH, Multicentre Growth Reference Study Group. WHO child growth standards based on length/height, weight and age. Acta Paediatr Suppl. 2006;450:76–85. [DOI] [PubMed] [Google Scholar]
- 11.Anderson NH, Sadler LC, McKinlay CJD, McCowan LME. INTERGROWTH-21st vs customized birthweight standards for identification of perinatal mortality and morbidity. Am J Obstet Gynecol. 2016;214(4):509. e1-.e7. [DOI] [PubMed] [Google Scholar]
- 12.Gardosi J, Mongelli M, Wilcox M, Chang A. An adjustable fetal weight standard. Ultrasound Obstet Gynecol. 1995;6(3):168–74. [DOI] [PubMed] [Google Scholar]
- 13.Gardosi J, Francis A, Turner S, Williams M. Customized growth charts: rationale, validation and clinical benefits. Am J Obstet Gynecol. 2018;218(2s):S609–18. [DOI] [PubMed] [Google Scholar]
- 14.Morris RK, Johnstone E, Lees C, Morton V, Smith G, Obstetricians, RCo, et al. Investigation and care of a Small-for‐Gestational‐Age fetus and a growth restricted fetus (Green‐top guideline No. 31). BJOG: An International Journal of Obstetrics & Gynaecology; 2024. [DOI] [PubMed] [Google Scholar]
- 15.Wilcox AJ, Snowden JM, Ferguson K, Hutcheon J, Basso O. On the study of fetal growth restriction: time to abandon SGA. Eur J Epidemiol. 2024;39(3):233–9. [DOI] [PubMed] [Google Scholar]
- 16.Horikoshi M, Beaumont RN, Day FR, Warrington NM, Kooijman MN, Fernandez-Tajes J, et al. Genome-wide associations for birth weight and correlations with adult disease. Nature. 2016;538(7624):248–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Magnus P, Gjessing HK, Skrondal A, Skjaerven R. Paternal contribution to birth weight. J Epidemiol Community Health. 2001;55(12):873–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lunde A, Melve KK, Gjessing HK, Skjaerven R, Irgens LM. Genetic and environmental influences on birth weight, birth length, head circumference, and gestational age by use of population-based parent-offspring data. Am J Epidemiol. 2007;165(7):734–41. [DOI] [PubMed] [Google Scholar]
- 19.Juliusdottir T, Steinthorsdottir V, Stefansdottir L, Sveinbjornsson G, Ivarsdottir EV, Thorolfsdottir RB, et al. Distinction between the effects of parental and fetal genomes on fetal growth. Nat Genet. 2021;53(8):1135–42. [DOI] [PubMed] [Google Scholar]
- 20.Tanigawa Y, Qian J, Venkataraman G, Justesen JM, Li R, Tibshirani R, et al. Significant sparse polygenic risk scores across 813 traits in UK biobank. PLoS Genet. 2022;18(3):e1010105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Choi SW, Mak TS, O’Reilly PF. Tutorial: a guide to performing polygenic risk score analyses. Nat Protoc. 2020;15(9):2759–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Khera AV, Chaffin M, Aragam KG, Haas ME, Roselli C, Choi SH, et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to Monogenic mutations. Nat Genet. 2018;50(9):1219–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhang J, Tian Y, Wang W, Ouyang F, Xu J, Yu X, et al. Cohort profile: the Shanghai birth cohort. Int J Epidemiol. 2019;48(1):21–g. [DOI] [PubMed] [Google Scholar]
- 24.Warrington NM, Beaumont RN, Horikoshi M, Day FR, Helgeland Ø, Laurin C, et al. Maternal and fetal genetic effects on birth weight and their relevance to cardio-metabolic risk factors. Nat Genet. 2019;51(5):804–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Cole TJ. Sample size and sample composition for constructing growth reference centiles. Stat Methods Med Res. 2021;30(2):488–507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Squires J, Bricker DD, Twombly E. Ages & stages questionnaires: Paul H. Brookes Baltimore; 2009.
- 27.Sheldrick RC, Marakovitz S, Garfinkel D, Carter AS, Perrin EC. Comparative accuracy of developmental screening questionnaires. JAMA Pediatr. 2020;174(4):366–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wei M, Bian X, Squires J, Yao G, Wang X, Xie H, et al. Studies of the norm and psychometrical properties of the ages and stages questionnaires, with a Chinese National sample. Zhonghua Er Ke Za zhi=. Chin J Pediatr. 2015;53(12):913–8. [PubMed] [Google Scholar]
- 29.Yue A, Jiang Q, Wang B, Abbey C, Medina A, Shi Y, et al. Concurrent validity of the ages and stages questionnaire and the Bayley scales of infant development III in China. PLoS ONE. 2019;14(9):e0221675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Beaumont RN, Warrington NM, Cavadino A, Tyrrell J, Nodzenski M, Horikoshi M, et al. Genome-wide association study of offspring birth weight in 86 577 women identifies five novel loci and highlights maternal genetic effects that are independent of fetal genetics. Hum Mol Genet. 2018;27(4):742–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Barker DJ. The origins of the developmental origins theory. J Intern Med. 2007;261(5):412–7. [DOI] [PubMed] [Google Scholar]
- 32.Redman CW, Sargent IL. Latest advances in Understanding preeclampsia. Science. 2005;308(5728):1592–4. [DOI] [PubMed] [Google Scholar]
- 33.Thornburg KL, O’Tierney PF, Louey S, Review. The placenta is a programming agent for cardiovascular disease. Placenta. 2010;31(SupplSuppl):S54–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Malhotra A, Allison BJ, Castillo-Melendez M, Jenkin G, Polglase GR, Miller SL. Neonatal morbidities of fetal growth restriction: pathophysiology and impact. Front Endocrinol (Lausanne). 2019;10:55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Frøen JF, Gardosi JO, Thurmann A, Francis A, Stray-Pedersen B. Restricted fetal growth in sudden intrauterine unexplained death. Acta Obstet Gynecol Scand. 2004;83(9):801–7. [DOI] [PubMed] [Google Scholar]
- 36.Sharma D, Shastri S, Sharma P. Intrauterine growth restriction: antenatal and postnatal aspects. Clin Med Insights. 2016;10:CMPed.S40070. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Miller SL, Huppi PS, Mallard C. The consequences of fetal growth restriction on brain structure and neurodevelopmental outcome. J Physiol. 2016;594(4):807–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Crispi F, Figueras F, Cruz-Lemini M, Bartrons J, Bijnens B, Gratacos E. Cardiovascular programming in children born small for gestational age and relationship with prenatal signs of severity. Am J Obstet Gynecol. 2012;207(2):e1211–9. [DOI] [PubMed] [Google Scholar]
- 39.Cruz-Lemini M, Crispi F, Valenzuela-Alcaraz B, Figueras F, Sitges M, Bijnens B, et al. Fetal cardiovascular remodeling persists at 6 months in infants with intrauterine growth restriction. Ultrasound Obstet Gynecol. 2016;48(3):349–56. [DOI] [PubMed] [Google Scholar]
- 40.Miranda J, Simões RV, Paules C, Cañueto D, Pardo-Cea MA, García-Martín ML, et al. Metabolic profiling and targeted lipidomics reveals a disturbed lipid profile in mothers and fetuses with intrauterine growth restriction. Sci Rep. 2018;8(1):13614. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Adanikin A, Lawlor DA, Pell JP, Nelson SM, Smith GCS, Iliodromiti S. Association of birthweight centiles and early childhood development of Singleton infants born from 37 weeks of gestation in scotland: A population-based cohort study. PLoS Med. 2022;19(10):e1004108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Gutbrod T, Wolke D, Soehne B, Ohrt B, Riegel K. Effects of gestation and birth weight on the growth and development of very low birthweight small for gestational age infants: a matched group comparison. Arch Dis Child Fetal Neonatal Ed. 2000;82(3):F208–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Arcangeli T, Thilaganathan B, Hooper R, Khan KS, Bhide A. Neurodevelopmental delay in small babies at term: a systematic review. Ultrasound Obstet Gynecol. 2012;40(3):267–75. [DOI] [PubMed] [Google Scholar]
- 44.Ge Q, Bai Y, Liu Z, Liu Q, Yan L, Lu Z. Detection of fetal DNA in maternal plasma by microarray coupled with emulsions PCR. Clin Chim Acta. 2006;369(1):82–8. [DOI] [PubMed] [Google Scholar]
- 45.Tam JCW, Chan YM, Tsang SY, Yau CI, Yeung SY, Au KK, et al. Noninvasive prenatal paternity testing by means of SNP-based targeted sequencing. Prenat Diagn. 2020;40(4):497–506. [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 datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.