Establishing population-specific gestational weight gain ranges for twin pregnancies and comparing their performance with IOM recommendations: a retrospective cohort study from central China

Abstract

Objectives

To establish optimal pre-pregnancy body mass index (BMI)-specific gestational weight gain (GWG) ranges for twin pregnancies and compare their association with maternal and neonatal adverse outcomes against Institute of Medicine (IOM) recommendations.

Design

Retrospective cohort study. Retrospective cohort study. Adjusted ORs (aORs) with 95% CI were used to quantify associations; average marginal effects (AME) with 95% CI (in percentage points) were used to compare absolute risk differences.

Setting

Perinatal data from >70 obstetric institutions in Wuhan, China, collected via the Wuhan Maternal and Child Health Service Management Information System.

Participants

10 636 women with twin deliveries at ≥28 weeks (2011–2023). Pre-pregnancy BMI categorised using Chinese cut-offs: underweight <18.5, normal 18.5–23.9, overweight 24.0–27.9, obesity ≥28.0 kg/m².

Primary and secondary outcome measures

Hypertensive disorders of pregnancy (HDP), gestational diabetes mellitus (GDM), premature rupture of membranes (PROM), small for gestational age (SGA) and large for gestational age.

Results

Optimal GWG ranges were: underweight 18.0–25.0 kg, normal 14.0–24.0 kg, overweight 12.2–24.0 kg, obesity 13.3–20.0 kg. Compared with IOM guidelines, study-derived ranges showed more favourable risk identification. In normal weight women, excessive GWG increased HDP risk (aOR 1.79, 95% CI 1.49 to 2.14); 13.57% versus 8.79%, AME 5.90 pp (95% CI 3.88 to 7.91 pp). In underweight women, inadequate GWG increased PROM (aOR 1.64, 95% CI 1.05 to 2.57); 14.48% versus 7.51%, AME 4.18 pp (95% CI 0.31 to 8.06 pp) and SGA (aOR 1.72, 95% CI 1.29 to 2.31); 45.58% versus 41.40%, AME 11.74 pp (95% CI 5.55 to 17.94 pp). In overweight women, excessive GWG increased HDP (aOR 1.81, 95% CI 1.21 to 2.70); 24.39% versus 16.32%, AME 9.68 pp (95% CI 2.49 to 16.88 pp) and inadequate GWG increased SGA (aOR 1.60, 95% CI 1.20 to 2.14); 35.15% versus 27.82%, AME 9.85 pp (95% CI CI 3.72 to 15.97 pp), which IOM failed to detect. In obese women, inadequate GWG increased SGA (aOR 2.76, 95% CI 1.37 to 5.53); 27.18% versus 17.16%, AME 17.95 pp (95% CI 5.74 to 30.17 pp), which was missed by IOM.

Conclusions

Our findings support adopting region-specific GWG standards for twin pregnancies in Asian populations.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• A retrospective cohort of over 10 000 twin pregnancies from Central China using a centralised electronic health information system.

• The combination of restricted cubic spline and IQR methods producing consistent and robust estimation of optimal gestational weight gain intervals for twin pregnancies across all Chinese-specific body mass index groups at pre-pregnancy.

• Lack of control for chorionicity.

• Possible under-reporting of overweight based on the self-reported pre-pregnancy weight.

Introduction

Globally, twin birth rates have risen by nearly 10% in recent decades, driven by increasing use of assisted reproductive technology (ART) and advanced maternal age; and this upward trend is predicted to continue.^{1 2} Compared with singleton pregnancy, twin pregnancies are not only associated with higher risks of maternal complications and fetal adverse outcomes,^{3 4} but also impose greater metabolic demands and postpartum health risks on mothers.⁵ Another public health challenge amplified by the trend of advanced maternal age is the growing incidence of overweight and obesity, and the rise of pregnant women with associated complications, including hypertension and diabetes mellitus.^{6 7} This phenomenon seems to be more common in pregnant women with twins.

Proactive intervention of controllable risk factors can reduce adverse pregnancy outcomes efficiently. Both preconception overweight/obesity and gestational weight gain (GWG) have been established as two controllable determinants of adverse pregnancy outcomes,⁸,¹¹ which can be modified through lifestyle changes.¹² Optimising preconception weight faces barriers due to limited pre-pregnancy care access, as few women seek preemptive medical care for overweight/obesity before pregnancy; however, GWG management possesses greater clinical feasibility through routine antenatal care.¹³ Substantial evidence has confirmed that optimal GWG could reduce adverse pregnancy outcomes, including halving the incidence of term pre-eclampsia, which seriously affects perinatal mortality and intrauterine growth restriction.¹⁴ Thus, it is so important to determine the appropriate range of GWG. In 2009, the Institute of Medicine (IOM) published the provisional GWG recommendations for twins.¹⁵ Recently, Lin et al developed equations for assessing optimal GWG ranges based on US birth weight data.^{16 17} However, well-documented differences in body composition between Asian and Western populations (higher body fat percentage at a given body mass index (BMI)) lead to discrepancies in BMI classification standards. The IOM uses WHO international cut-offs (normal 18.5–24.9, overweight 25.0–29.9 and obesity ≥30.0 kg/m²), whereas Chinese guidelines define overweight as ≥24.0 and obesity as ≥28.0 kg/m². Given these differences in BMI classification, as well as differences in birth weight distributions, both the IOM recommendations and existing equations need further validation in Asian populations.¹⁸,²⁰ Our prior study established BMI-stratified GWG ranges with the sole objective of optimising twin birth weights.²¹ However, suboptimal GWG is associated with significantly increased risks of maternal complications, including hypertensive disorders of pregnancy (HDP) and gestational diabetes mellitus (GDM), as confirmed by previous studies.²²,²⁴ Therefore, developing gestational management standards that simultaneously address the dual outcomes of preventing maternal pregnancy complications and ensuring appropriate neonatal weight has become imperative. Current existing evidence from only two single-centre Chinese studies^{25 26} is insufficient to establish nationwide standards. This gap necessitates population-based studies with adequate sample sizes to establish Chinese-specific GWG ranges for twin pregnancies.

Using data on twin mother-infant deliveries from over 70 obstetric institutions in Wuhan, China, between 1 January 2011 and 31 December 2023, this study aimed to establish pre-pregnancy BMI-stratified optimal GWG ranges that are associated with reduced adverse maternal and neonatal outcomes, including HDP, GDM, premature rupture of membranes (PROM), small for gestational age (SGA) or large for gestational age (LGA). Located in central China with high population mobility, Wuhan has 10%–20% of pregnant women originating from across the nation annually. This diversity enhances the generalisability of our twin cohort for the broader Chinese pregnant population.

Methods

This study extracted maternal and neonatal health data from the Wuhan Maternal and Child Health Service Management Information System. Developed in 2003, this urban electronic networked platform collects all clinical perinatal information, including antepartum examinations, intrapartum conditions and postpartum visit data for mothers and children in Wuhan City.

Study population

This study included the following mothers: (1) delivered twins between 1 January 2011 and 31 December 2023; (2) received complete perinatal care from pregnancy to delivery at obstetric institutions in Wuhan, China; (3) gestational age ≥28 weeks; (4) twin live-born twins without major congenital anomalies. The exclusion criteria were: (1) resided outside Wuhan for >1 month during pregnancy; (2) having pre-pregnancy complications (cardiac, renal, endocrine or other critical systemic disorders); (3) missing pre-pregnancy maternal height/weight or neonatal anthropometric data. Among 21 888 women with twin deliveries during this study period, 10 636 mothers and their twins met eligibility criteria (figure 1).

Figure 1. The selection process of this study.

Pre-pregnancy BMI categorisation and GWG calculation

According to the Chinese adult standards, pregnant women were categorised into four groups based on pre-pregnancy BMI: underweight (BMI<18.5 kg/m²), normal weight (18.5≤BMI<24.0 kg/m²), overweight (24.0≤BMI<28.0 kg/m²) and obesity (BMI≥28.0 kg/m²). GWG was calculated as maternal weight at delivery minus pre-pregnancy weight. Pre-pregnancy weight was self-reported by the mothers during early pregnancy registration (≤14 gestational weeks), while delivery weight was measured by professionally trained nurses at the obstetric facility at admission.

Outcomes of interest and confounders

Based on clinical practices and published literature, this study focused on common adverse maternal and neonatal pregnancy outcomes, including HDP, GDM, PROM, SGA or LGA. HDP encompassed gestational hypertension and pre-eclampsia/eclampsia. Neonates were classified as SGA (defined as birth weight below the 10th percentile for gestational age and sex), LGA (above the 90th percentile) or the appropriate-for-age (AGA). The neonatal outcome for the pregnancy was classified as SGA if at least one twin met SGA criteria; LGA if at least one twin met LGA criteria and neither twin was SGA; AGA only when both twins were AGA. Confounding variables included maternal age, gravidity, parity (number of deliveries ≥28 weeks) and conception method.

Statistical analysis

Descriptive statistics summarised maternal characteristics, GWG and pregnancy outcomes stratified by pre-pregnancy BMI. Restricted cubic spline (RCS) models evaluated non-linear relationships between GWG and adverse outcomes. To improve statistical power and simplify interpretation, GDM, HDP and PROM were combined into a composite outcome (CO) for covariate-adjusted models. Low-risk GWG intervals were defined as ranges where adjusted ORs (aORs) for adverse outcomes remained below 1.0 or showed declining trends. Concurrently, the IQR of GWG among healthy pregnancies (term deliveries without complications and birth weight 2500–4000 g) served as an alternative benchmark. Cohen’s kappa assessed agreement between RCS- and IQR-derived GWG ranges. The final recommended GWG intervals prioritised broader ranges within each pre-pregnancy BMI category to balance clinical feasibility and risk reduction. Multivariable logistic regression models compared adverse outcome associations between the study-derived optimal GWG ranges and the IOM recommendations, quantified using aORs with 95% CIs; average marginal effects (AME) with 95% CIs expressed in percentage points (pp) were used to compare absolute risk differences of outcomes among groups. All analyses were performed with R software (V.4.2.1); statistical significance was set at p<0.05. For AME, a 95% CI that does not cross zero was also considered statistically significant at the α=0.05 level.

Results

Baseline characteristics of the study population

Among 10 636 mothers, maternal age ranged from 18 to 44 years, with 14.2% (1507) aged ≥35 years. Pre-pregnancy BMI categories included underweight (11.3%, 1204), normal weight (73.1%, 7772), overweight (12.8%, 1364) and obesity (2.8%, 296). As maternal age increased, the prevalence of underweight declined, whereas the prevalence of overweight and obesity rose significantly. Compared with primiparous women, multiparous women (those who had given birth ≥3 times) exhibited a significantly higher prevalence of overweight/obesity. The prevalence of overweight (19.9%) and obesity (4.8%) among pregnant women conceived through ART was significantly higher than that in the naturally conceived pregnant women (12.2% and 2.6%, respectively, p<0.001).

GWG varied across pre-pregnancy BMI groups. The underweight group had a mean GWG of 20.7±5.9 kg (range: 6.0–40.0 kg; IQR: 16.5–24.0 kg). Followed by the normal weight group with 19.2±6.6 kg (range: 5.0–40.0 kg; IQR: 14.8–23.0 kg), lower mean GWG was observed in the overweight group (16.7±6.1 kg; range: 5.0–38.5 kg; IQR: 12.5–20.0 kg) and the obese group (15.8±5.9 kg; range: 5.0–35.0 kg; IQR: 11.0–20.0 kg). The differences in GWG among pre-pregnancy BMI groups were statistically significant (p<0.01).

Among 10 636 pregnant women, 1028 (9.7%) were diagnosed with GDM, 1113 (10.3%) with HDP (including 576 cases of pre-eclampsia and 17 cases of eclampsia) and 999 with PROM. A total of 2794 women had at least one of the three common pregnancy complications studied. The incidence of pregnancy complications in the pre-pregnancy overweight/obese groups was significantly higher than that in the other two groups. Regarding neonatal outcomes, 3627 mothers delivered SGA neonates and 534 delivered LGA neonates, corresponding to the rates of 34.1% and 5.0%, respectively. χ² tests indicated statistically significant differences (p<0.01) in the incidence of CO, SGA or LGA among pre-pregnancy BMI groups. Baseline characteristics are summarised in table 1.

Table 1. Comparison of maternal characteristics, pregnancy outcomes stratified by pre-pregnancy body mass index.

	Total (n=10 636)	Underweight (n=1204, 11.3%)	Normal weight (n=7772, 73.1%)	Overweight (n=1364, 12.8%)	Obesity (n=296, 2.8%)	P value
Age						<0.001
<25	902 (8.5)	173 (14.4)	666 (8.6)	53 (3.9)	10 (3.4)
25–34	8227 (77.3)	929 (77.1)	5997 (77.2)	1064 (78.0)	237 (80.1)
≥35	1507 (14.2)	102 (8.5)	1109 (14.2)	247 (18.1)	49 (16.5)
Parity						0.154
1	7434 (69.9)	862 (71.6)	5406 (69.6)	953 (69.9)	213 (72.0)
2	2825 (26.6)	311 (25.8)	2093 (26.9)	351 (25.7)	70 (23.6)
≥3	377 (3.5)	31 (2.6)	273 (3.5)	60 (4.4)	13 (4.4)
Gestational weeks						<0.001
<37	5897 (55.4)	664 (55.1)	4212 (54.2)	824 (60.4)	197 (66.6)
≥37	4739 (44.6)	540 (44.9)	3560 (45.8)	540 (39.6)	99 (33.4)
Mode of conception						<0.001
Natural	9742 (91.6)	1119 (92.9)	7184 (92.4)	1186 (86.9)	253 (85.5)
ART	894 (8.4)	85 (7.1)	588 (7.6)	178 (13.1)	43 (14.5)
CO	2794 (26.3)	249 (20.7)	1924 (24.8)	492 (36.1)	129 (43.6)	<0.001
HDP	1113 (10.5)	80 (6.6)	739 (9.5)	224 (16.4)	70 (23.6)	<0.001
GDM	1028 (9.7)	75 (6.2)	676 (8.7)	228 (16.7)	49 (16.5)	<0.001
PROM	999 (9.4)	116 (9.6)	711 (9.1)	137 (10.0)	35 (11.8)	0.339
SGA	3627 (34.1)	494 (41.0)	2677 (34.4)	395 (29.0)	61 (20.6)	<0.001
LGA	534 (5.0)	34 (2.8)	372 (4.8)	94 (6.9)	34 (11.5)	<0.001

ART, assisted reproductive technology; CO, composite outcome; GDM, gestational diabetes mellitus; HDP, hypertensive disorders of pregnancy; LGA, large for gestational age; PROM, premature rupture of membranes; SGA, small for gestational age.

Using RCS to assess the interval for suitable GWG

We used RCS to model the relationship between GWG and CO, LGA and SGA, adjusting for confounding factors such as maternal age, parity, method of conception and gestational week at delivery.

In the underweight group, the risk of SGA continued to decline with increasing GWG, whereas the risk of LGA gradually increased. When GWG ranged from 18.5 kg to 25.0 kg, the SGA risk reached a plateau. For the GWG level below 25.0 kg, the aOR for LGA consistently remained below 1.0. Within this range, the risk of CO initially showed an upward then downward trend, after the risk value fell to 0.72; it rebounded with an upward trajectory. However, when GWG was within 17.5–25.0 kg, the aOR for CO remained <1.0. We therefore concluded that the appropriate GWG range was 18.5–25.0 kg because this interval was associated with the lowest combined risk of CO, SGA and LGA.

As shown in figure 2, we tentatively determined the suitable GWG range as follows: 14.0–24.0 kg for the normal weight group, 12.2–24.0 kg for the overweight group, 13.3–19.8 kg for the obesity group, based on similar risk-benefit analyses.

Figure 2. Restricted cubic spline (RCS) analysis of gestational weight gain and adverse outcomes by pre-pregnancy body mass index (BMI). This figure illustrates the non-linear relationship between GWG and adverse pregnancy outcomes across pre-pregnancy BMI categories, modelled using RCS. Adjusted ORs (aORs) for CO (presence of hypertensive disorders of pregnancy, gestational diabetes mellitus, premature rupture of membranes), LGA and SGA are plotted against GWG, with shaded intervals indicating low-risk ranges where aORs remained ≤1.0 or exhibited declining trends. Key findings: optimal GWG ranges were 18.5–25.0 kg for underweight, 14.0–24.0 kg for normal-weight, 12.2–24.0 kg for overweight and 13.3–19.8 kg for obesity. Models adjusted for maternal age, parity, conception method and gestational week at delivery. Chinese adults BMI categories applied. CO, composite outcome; GWG, gestational weight gain; LGA, large for gestational age; SGA, small for gestational age.

Using IQR to assess the interval for suitable GWG

We identified 3436 women with healthy pregnancies from 10 636 study participants. Reference intervals for GWG were established using the IQR method, stratified by pre-pregnancy BMI categories. For the underweight group, the observed GWG ranged from 6.0 kg to 40.0 kg, with an IQR of 18.0–25.0 kg. In the normal weight group, GWG spanned from 5.0 kg to 40.0 kg, with an IQR of 15.0–24.0 kg, while the overweight group exhibited a GWG range of 5.0 kg–37.5 kg, with an IQR of 12.5–21.0 kg. The obesity group demonstrated a GWG range of 6.0 kg–35.0 kg, with an IQR of 13.3–20.0 kg.

Determination of optimal GWG

The Kappa values for agreement between the RCS and IQR methods exceeded 0.7 for all BMI groups, indicating good consistency and robustness between the two methods. We selected the broader interval from the two methods as clinically recommended GWG intervals for Chinese twin pregnancies. This study ultimately determined the optimal GWG ranges for twin pregnancies stratified by pre-pregnancy BMI as follows: 18.0–25.0 kg for underweight women, 14.0–24.0 kg for normal weight women, 12.2–24.0 kg for overweight women and 13.3–20.0 kg for women with obesity. Table 2 shows these GWG ranges stratified by pre-pregnancy BMI categories, and a comparison between the determined GWG optimal ranges and the IOM recommendations, which lack GWG guidance for underweight women with twin pregnancies.

Table 2. Comparison of optimal GWG in this study and IOM recommendation.

Group	Estimated GWG		Kappa value	Determined GWG
	N (IQR)	RCS		This study	IOM
Underweight	395 (18.0–25.0)	18.5–25.0	0.92	18.0–25.0	–
Normal weight	2577 (15.0–24.0)	14.0–24.0	0.91	14.0–24.0	17.0–25.0
Overweight	383 (12.5–21.0)	12.2–24.0	0.75	12.2–24.0	14.0–23.0
Obesity	81 (13.3–20.0)	13.3–19.8	0.89	13.3–20.0	11.0–19.0

GWG, gestational weight gain; IOM, Institute of Medicine; RCS, restricted cubic spline.

Validation of the optimal GWG ranges: comparison with IOM recommendations

Based on the optimal GWG ranges identified in this study and the IOM recommendations, we categorised GWG as inadequate, adequate or excessive. Using these two sets of GWG standards as references, we performed multivariable logistic regression and absolute risk difference analyses; statistically significant findings are presented in figure 3.

Figure 3. Comparison of the appropriate gestational weight gain’s role with the IOM recommendations in guiding clinical practice. AMEs and aORs were calculated using multivariable models adjusted for maternal age, gravidity, parity and conception method. AME, average marginal effect; aORs, adjusted ORs; IOM, Institute of Medicine; HDP, hypertensive disorders of pregnancy; GDM, gestational diabetes mellitus; LGA, large for gestational age; PROM, premature rupture of membranes; SGA, small for gestational age.

Among underweight women, using the optimal GWG ranges established in this study as the reference, inadequate GWG was found to be associated with increased risks of PROM (aOR 1.64, 95% CI 1.05 to 2.57) and SGA (aOR 1.72, 95% CI 1.29 to 2.31), whereas excessive GWG was associated with a decreased risk of SGA (aOR 0.67, 95% CI 0.48 to 0.95). Specifically, among women with inadequate GWG, the incidence of PROM was 14.48% compared with 7.51% in those with adequate GWG (AME 4.18 pp, 95% CI 0.31 to 8.06 pp), and 45.58% delivered SGA neonate versus 41.40% in those with adequate and 45.58% GWG (AME 11.74 pp, 95% CI 5.55 to 17.94 pp). In contrast, among women with excessive GWG, the incidence of SGA was lower than that in the adequate GWG group (32.76% vs 41.40%; AME, −8.08 pp, 95% CI −14.87 to -.29 pp). The IOM recommendations provide no optimal GWG range for this population.

Among normal-weight women, we used the optimal GWG ranges developed in this study as the reference. Those women with excessive GWG had a 79% increased risk of HDP (aOR 1.79, 95% CI 1.49 to 2.14) and an 81% increased risk of LGA (aOR 1.81, 95% CI 1.41 to 2.33), while those with inadequate GWG had a 44% increased risk of SGA (aOR 1.44, 95% CI 1.26 to 1.64). Specifically, among women with excessive GWG, the incidences of HDP and LGA were 13.57% and 6.95%, respectively, compared with 8.79% and 4.38% in those with adequate GWG; the corresponding AMEs were 5.90 pp (95% CI 3.88 to 7.91 pp) and 4.75 pp (95% CI 2.53 to 6.98 pp). Among women with inadequate GWG, 37.79% delivered an SGA neonate versus 35.13% in those with adequate GWG, with an AME of 7.92 pp (95% CI 5.07 to 10.78 pp). With reference to the IOM recommendations, only inadequate GWG was associated with an increased risk of GDM, with a 36% increased risk (aOR 1.36, 95% CI 1.15 to 1.61). Among women with inadequate GWG, 11.51% were diagnosed with GDM compared with 7.89% in those with adequate GWG, with an AME of 2.53 pp (95% CI 1.15 to 3.92 pp).

Using the optimal GWG ranges established from this study as the reference for overweight women, those with excessive GWG had an 81% increased risk of HDP (aOR 1.81, 95% CI 1.21 to 2.70), while those with inadequate GWG had a 60% increased risk of SGA (aOR 1.60, 95% CI 1.20 to 2.14). Specifically, among women with excessive GWG, the incidence of HDP was 24.39% compared with 16.32% in those with adequate GWG; the AME was 9.68 pp (95% CI 2.49 to 16.88 pp). Among women with inadequate GWG, 35.15% delivered an SGA neonate versus 27.82% in those with adequate GWG, with an AME of 9.85 pp (95% CI 3.72 to 15.97 pp). With reference to the IOM recommendations, inadequate GWG was associated with increased risks of GDM and PROM, with a 46% increased risk of GDM (aOR 1.46, 95% CI 1.01 to 2.12) and a 103% increased risk of PROM (aOR 2.03,95% CI 1.25 to 3.30). Among women with inadequate GWG, 22.34% were diagnosed with GDM and 16.22% had PROM, compared with 16.20% and 7.34%, respectively, in those with adequate GWG; the corresponding AMEs were 5.68 pp (95% CI 0.09 to 11.26 pp) for GDM and 6.64 pp (95% CI 2.20 to 11.08 pp) for PROM.

Using the optimal GWG ranges established from this study as the reference for women with obesity, those with excessive GWG had an increased risk of HDP (aOR 2.08, 95% CI 1.01 to 4.28), although the absolute difference in HDP incidence between the two groups was not statistically significant. Inadequate GWG was associated with an increased risk of SGA (aOR 2.76, 95% CI 1.37 to 5.53). Among women with inadequate GWG, 27.18% delivered an SGA neonate versus 17.16% in those with adequate GWG, with an AME of 17.95 pp (95% CI 5.74 to 30.17 pp). With reference to the IOM recommendations, neither excessive nor inadequate GWG was associated with any adverse maternal or neonatal outcomes.

Discussion

This study established optimal pre-pregnancy BMI-specific GWG ranges for twin pregnancies that were doubly associated with reduced pregnancy complications and optimised neonatal birth weight, through analysis of twin birth data from Wuhan, a major city in central China. The prevalence of underweight (11.3%), overweight (12.8%) and obesity (2.8%) varied with maternal age and parity. Overall, with advancing maternal age and a higher number of parities, the prevalence of underweight decreased while that of overweight/obesity increased. Furthermore, we found that ART-conceived pregnancies posed higher overweight/obesity rates, which were consistent with previous reports.⁷ In addition, the proportion of SGA newborns among pre-pregnant underweight women was significantly higher than that in the other three groups. The association between higher parity and overweight/obesity suggests a need for heightened postpartum weight management, as postpartum weight retention may trigger cycles of visceral adiposity leading to metabolic complications.^{27 28} Consequently, pre-pregnancy underweight women and multiparas should be considered high-risk groups for perinatal weight management, alongside individuals with pre-pregnancy overweight/obese.

Inadequate GWG usually reflects nutritional deficiencies, insufficient plasma volume expansion and an underlying metabolic state that inhibits insulin metabolism signal, leading to insulin resistance,²⁹ which could partially explain GDM risk elevation in specific subpopulations. However, the association between insufficient GWG and GDM risk remains inconclusive, as several studies have reported that excessive GWG was a risk factor for GDM.^{30 31} In this study, our analysis demonstrated that IOM-defined inadequate GWG increased GDM risk in pre-pregnancy normal weight or overweight women; under our population-optimised ranges, this association was confined solely to normal weight women. This divergence in risk associations suggested that an appropriate GWG range established according to population characteristics, specific pregnancy complications such as GDM and pre-pregnancy BMI status can more accurately guide gestational weight management.

Previous studies have conclusively demonstrated the positive correlation between excessive GWG and HDP. Deihl et al suggested that in normotensive women, first-trimester GWG was positively associated with blood pressure changes in the second trimester,³² and other researchers reported that cumulative GWG correlated with blood pressure fluctuations throughout the pregnancy period.³³ Recent evidence suggested that maintaining GWG within appropriate ranges could reduce full-term preeclampsia incidence by 60% in dichorionic twin pregnancies.¹⁴ In this study, regardless of pre-pregnancy BMI status, excessive GWG according to our ranges was associated with significantly elevated HDP risk, which was consistent with prior findings. When applying IOM recommendations, this association was only observed in pre-pregnancy normal weight women, with a smaller AME (5.69 vs 5.90). These results indicated that IOM standards exhibited lower sensitivity than our GWG ranges in predicting HDP risk, particularly among women with pre-pregnancy overweight/obesity.

In clinical practice, a diagnosis of GDM based on blood glucose testing at a specific gestational week triggers immediate nutritional and exercise interventions to prevent disease progression.^{34 35} However, HDP may be less promptly addressed, as some pregnant women tolerate elevated blood pressure discomfort and delay seeking medical care, risking progression to pre-eclampsia, which is a leading cause of maternal and perinatal mortality/morbidity worldwide. Thus, from a clinical safety perspective, HDP carries more severe potential consequences than GDM. The GWG optimal ranges derived from this study were associated with greater risk reduction for HDP, LGA and SGA across all BMI categories. Whereas the IOM recommendations offered better detection of GDM risk, particularly in overweight women. Given the greater clinical severity of HDP, preventing HDP should take priority over preventing GDM when the two standards conflict. Therefore, we recommend that weight management for Chinese twin pregnancies should primarily use our population-specific GWG ranges.

This study used data collected prospectively through a standardised real-time entry system over a decade, which helped to reduce potential information bias. Despite the fact that the study sample was geographically limited to a single city, it is notable that nearly 20% of pregnant women in Wuhan come from other regions nationwide annually. As a transportation hub in central China, Wuhan has regional characteristics that enhance the representativeness of the cohort for the Chinese population by incorporating individuals with diverse genetic, socioeconomic and lifestyle backgrounds. Furthermore, this study introduced methodological integration of RCS and IQR to define BMI-stratified optimal GWG ranges and validated the derived ranges against the IOM recommendations using real-world data. The study confirms that establishing population-specific GWG standards offers superior clinical guidance, while providing actionable references for Asian populations where IOM recommendations demonstrate limited clinical concordance.

There are several limitations that need to be considered. First, the small size of women with pre-pregnancy obesity precludes further stratification according to obesity level. Second, pre-pregnancy weight was mainly self-reported by pregnant women during early pregnancy, which can lead to recall bias; however, the high concordance between the self-reported weight and actual weight suggested this bias was likely negligible.³⁶ Third, chorionicity data were unavailable in our study. Monochorionic and dichorionic twins are associated with different risks of gestational complications and birth weight discordance,^{37 38} and our inability to stratify or adjust for chorionicity may introduce significant unmeasured confounding. Therefore, our findings should be interpreted with caution, and future studies with chorionicity information are needed to validate whether the proposed GWG ranges apply equally to monochorionic and dichorionic twins. Fourth, we excluded 8217 women (37.5% of initial twin deliveries) who left Wuhan for more than 1 month during pregnancy because their delivery outcomes were not captured by the information system. These excluded women may differ systematically from those included (eg, migrant status, prenatal care access), potentially introducing selection bias.³⁹ Due to data system constraints, we could not retrieve their baseline characteristics for comparison.

Conclusions

This study established BMI-stratified optimal GWG ranges for twin pregnancies in a central Chinese cohort. Our ranges were associated with more favourable maternal and neonatal outcomes than the IOM guidelines, particularly for HDP, SGA/LGA prevention and provided an optimal range for underweight women, a group not covered by IOM recommendations. Our findings highlight the need for region-specific standards. We recommend adopting these ranges as the primary clinical reference for weight management in Chinese twin pregnancies. This study further confirms the necessity of establishing region-specific GWG standards and provides a reference for developing GWG guidelines in Asian populations with similar BMI distribution patterns.

Data availability statement

No data are available.