Neonatal Outcomes in Twin Pregnancies with Birthweight Discordance: A five-year retrospective cohort study

Summary

Objectives:

This study aimed to evaluate neonatal outcomes associated with birthweight discordance in twin pregnancies managed at a regional tertiary care centre in Oman.

Methods:

This five-year retrospective cohort study (2020–2024) was conducted at Nizwa Hospital, Nizwa, Oman, and included twin pregnancies of ≥24 weeks’ gestation. Birthweight discordance was defined as ≥20%. Multivariable logistic regression analysis was performed, adjusting for gestational age at delivery, chorionicity, preeclampsia and gestational diabetes mellitus.

Results:

A total of 226 twin pregnancies were included, of which 45 (19.9%) had birthweight discordance ≥20%. Discordant pregnancies were delivered at an earlier gestational age than non-discordant pregnancies (34.2 ± 2.9 versus 35.3 ± 2.2 weeks; P = 0.017). NICU admission was more frequent in the discordant group (60.0% versus 31.5%; P <0.001), as was the composite adverse neonatal outcome (62.2% versus 33.1%; P <0.001). Birthweight discordance ≥20% remained independently associated with the composite adverse neonatal outcome (adjusted odds ratio [aOR] = 2.74, 95% CI: 1.21–6.25; P = 0.016), while increasing gestational age at delivery was protective (aOR = 0.47 per additional week, 95% CI: 0.37–0.59; P <0.001).

Conclusion:

Birthweight discordance ≥20% was associated with adverse neonatal outcomes in twin pregnancies, with this association largely mediated by prematurity.

Advances in Knowledge

• This study found that birthweight discordance ≥20% independently predicted adverse neonatal outcomes.

• Gestational age at delivery was the primary determinant of neonatal morbidity.

• The findings support integrating growth discordance, chorionicity and Doppler parameters for improved risk stratification.

• To the best of the authors’ knowledge, this study is the first Omani investigation evaluating the impact of birthweight discordance on neonatal outcomes in twin pregnancies.

Application to Patient Care

• Regular ultrasound and Doppler surveillance in twin pregnancies can help identify high-risk cases.

• Avoiding unnecessary preterm delivery based solely on growth discordance may be associated with improved neonatal outcomes.

• The findings support individualised management strategies that integrate chorionicity, fetal well-being and gestational maturity.

1. Introduction

The incidence of twin pregnancies has increased over recent decades, partly due to delayed childbearing, particularly among women aged 35–39 years, and the increasing use of assisted reproductive technologies (ART).^1,2,3 This trend has been accompanied by increased recognition of twin-specific complications that contribute substantially to neonatal morbidity and mortality. Twin pregnancies are consistently associated with higher rates of preterm birth, fetal growth restriction, neonatal morbidity and perinatal death compared with singleton gestations.^1,2 These adverse outcomes arise from both biological factors intrinsic to multifetal gestation and complications related to placentation, chorionicity and uteroplacental insufficiency.

Among the unique challenges of twin gestation, fetal growth discordance represents an important clinical manifestation of asymmetric intrauterine growth and unequal placental sharing.^5,6 Birthweight discordance, defined as the percentage difference between the larger and smaller twin, is commonly used as a postnatal surrogate for antenatal fetal growth discordance.^7,8 Although the cut-off varies between 15% and 25% across studies, a threshold of ≥20% is most frequently applied in both research and clinical practice, as it identifies twin pairs at a clinically significant increased risk of adverse neonatal outcomes.^1,2,9

In monochorionic twin pregnancies, selective fetal growth restriction (sFGR) is additionally characterised by an estimated fetal weight below the 10th percentile, often reflecting unequal placental sharing and intertwin vascular anastomoses.^8,10,11,12 Numerous cohort studies and meta-analyses have demonstrated that increasing degrees of growth discordance are associated with higher rates of neonatal morbidity, including respiratory distress syndrome, necrotising enterocolitis, intraventricular haemorrhage and perinatal death.^1,2,13 However, interpretation of these associations remains complex, as discordant twins are more likely to be delivered preterm, raising the possibility that prematurity and low absolute birthweight—rather than discordance per se—are the principal drivers of adverse outcomes.^3,14

In Oman, where twin pregnancy rates are rising due to increasing maternal age and the use of ART, understanding the clinical implications of birthweight discordance is critical for improving perinatal care.¹⁵ Therefore, this study aimed to evaluate neonatal outcomes among twin pregnancies managed at a regional tertiary care centre in Oman and to determine whether clinically significant birthweight discordance (≥20%) is independently associated with adverse neonatal outcomes after adjusting for gestational age at delivery.

2. Methods

This retrospective cohort study was conducted at the Department of Obstetrics and Gynecology, Nizwa Hospital, Ministry of Health, Nizwa, Oman, over a five-year period from January 2020 to December 2024. Nizwa Hospital is a regional referral centre providing specialised obstetric care for high-risk pregnancies, including complicated twin gestations, and serves both urban and rural populations.

All women with twin pregnancies with a gestational age of ≥24 weeks at diagnosis or delivery who received antenatal care and/or delivered at Nizwa Hospital during the study period were eligible for inclusion. The reference population consisted of twin pregnancies with documented antenatal ultrasound examinations, complete delivery records and neonatal outcome data available for analysis. Cases with missing key variables, including birthweight or neonatal outcome data, were excluded to ensure completeness of follow-up. Both monochorionic diamniotic and dichorionic diamniotic twin pregnancies were included, provided that antenatal ultrasound data and delivery outcome records were available. Pregnancies complicated by major structural or chromosomal fetal anomalies were excluded. Monoamniotic twin pregnancies, twin—twin transfusion syndrome, intrauterine fetal demise of one or both twins before 24 weeks’ gestation, major congenital or chromosomal abnormalities and cases with incomplete follow-up or missing essential data (e.g., birthweight or neonatal outcome information) were excluded from the analysis. Women with twin pregnancies were managed according to standard institutional protocols, including regular antenatal clinic visits and ultrasound surveillance; however, detailed information regarding home-based care and visit frequency was not consistently available in the retrospective records.

Fetal growth discordance was calculated as the percentage difference in estimated fetal weight or birthweight between the two twins using the following standard formula:

$$Discordance(\%)=\frac{Weightoflargertwin-Weightofsmallertwin}{Weightoflargertwin}\times 100$$

The larger twin was defined as the twin with the higher birthweight at delivery, and the smaller twin as the twin with the lower birthweight. Clinically significant fetal growth discordance was defined as a discordance of ≥20%, based on published evidence demonstrating an increased risk of adverse neonatal outcomes beyond this threshold. Pregnancies with discordance ≥20% were classified as discordant (exposed group), while those with discordance <20% were classified as non-discordant (comparison group).

sFGR was not formally classified in this study, as the exposure variable was based on postnatal birthweight discordance rather than antenatal diagnostic criteria.

Chorionicity was determined during the first trimester using standard ultrasound criteria, including identification of the lambda (λ) sign or T-sign. Twin pregnancies were classified as monochorionic diamniotic or dichorionic diamniotic. Chorionicity was included as a stratification variable and as an adjustment factor in multivariable analyses due to its established influence on neonatal outcomes.

Fetal growth was monitored through serial ultrasound examinations. Estimated fetal weight was calculated using standard biometric parameters, including biparietal diameter, head circumference, abdominal circumference and femur length. Placental insufficiency was defined based on antenatal Doppler abnormalities (elevated umbilical artery pulsatility index or absent or reversed end-diastolic flow) and/or a clinical diagnosis of fetal growth restriction. Birthweight was measured immediately after delivery. Postnatal birthweight discordance was used as the primary exposure variable in the final analysis to ensure objective classification and to minimise the potential error associated with antenatal ultrasound estimation.

The primary outcome used in multivariable logistic regression was a composite adverse neonatal outcome, defined as the occurrence of either neonatal intensive care unit (NICU) admission or at least one major neonatal morbidity. Major neonatal morbidity included respiratory distress syndrome, necrotising enterocolitis, intraventricular haemorrhage, neonatal sepsis or neonatal death. NICU admission followed standard institutional clinical practice; however, criteria may vary and could influence outcome rates.

Continuous variables are presented as mean ± standard deviation and were compared using Welch's t-test. Categorical variables are expressed as frequencies and percentages and were compared using Chi-square tests. Multivariable logistic regression analysis was performed to identify independent predictors of the composite adverse neonatal outcome, adjusting for gestational age at delivery, chorionicity, preeclampsia and gestational diabetes mellitus. Adjusted odds ratios (aOR) with 95% confidence intervals (CI) were reported. A two-sided P value <0.05 was considered statistically significant. The unit of analysis was the pregnancy. Outcomes were defined as present if occurring in at least one twin (‘any twin’), reflecting a pregnancy-level analytical approach.

Nizwa Hospital functions as a regional referral centre serving a diverse population. The inclusion of both monochorionic and dichorionic twin pregnancies, referral and non-referral cases, and the use of standardised diagnostic criteria support the applicability of the findings to similar tertiary care settings, although extrapolation to other populations should be made with caution.

3. Results

A total of 226 twin pregnancies were included in the analysis, of which 45 (19.9%) demonstrated birthweight discordance of ≥20% [Fig. 1]. Pregnancies with clinically significant discordance were delivered at an earlier gestational age compared with non-discordant twin pregnancies (34.2 ± 2.9 versus 35.3 ± 2.2 weeks; P = 0.017). The mean birthweight discordance was significantly higher in the discordant group compared with the non-discordant group (29.4 ± 10.6% versus 7.8 ± 5.7%; P <0.001), confirming clear separation between exposure groups. The mean birthweight of Twin 1 (the larger twin) was comparable between the discordant and non-discordant groups (2.0 ± 0.6 kg versus 2.1 ± 0.4 kg; P = 0.077), whereas Twin 2 (the smaller twin) had a significantly lower mean birthweight in the discordant group (1.7 ± 0.6 kg versus 2.1 ± 0.5 kg; P <0.001). Maternal demographic and obstetric characteristics, including maternal age, parity and mode of delivery, were comparable between groups [Table 1]. Placental insufficiency was rare (1.7%) and was not observed in the discordant group, limiting meaningful comparison. Body mass index was not analysed due to incomplete data in the discordant group.

Fig. 1.

STROBE flow diagram of study population selection.

Flow diagram illustrating the identification, eligibility assessment, exclusions, and final inclusion of twin pregnancies in the study, as well as classification into discordant (≥20%) and non-discordant (<20%) groups.

Table 1.

Baseline maternal and pregnancy characteristics of twin pregnancies by discordance group (N = 226).

	Mean ± SD
Characteristic*	Non-discordant group (n = 181)	Discordant group (n = 45)	P value
Maternal age in years	33.0 ± 5.2	34.0 ± 6.6	0.345
Parity	2.5 ± 2.0	2.4 ± 2.2	0.663
Gravidity	3.6 ± 2.2	3.5 ± 2.2	0.642
Gestational age at delivery in weeks	35.3 ± 2.2	34.2 ± 2.9	0.017
Birthweight discordance in %	7.8 ± 5.7	29.4 ± 10.6	<0.001
Birthweight Twin 1 in kg	2.1 ± 0.4	2.0 ± 0.6	0.077
Birthweight Twin 2 in kg	2.1 ± 0.5	1.7 ± 0.6	<0.001
Monochorionic, n (%)	52 (28.7)	14 (31.1)	0.896
Preeclampsia, n (%)	7 (3.9)	3 (6.7)	0.680
Gestational diabetes, n (%)	56 (30.9)	11 (24.4)	0.502
Caesarean delivery, n (%)	151 (83.4)	42 (93.3)	0.147
Antenatal hospitalisation, n (%)	32 (17.7)	7 (15.6)	0.907

SD = standard deviation.

^*Twin 1 = larger twin (higher birthweight at delivery); Twin 2 = smaller twin (lower birthweight at delivery).

Rates of NICU admission and major neonatal morbidity were significantly higher among twin pregnancies complicated by growth discordance; NICU admission occurred in 60.0% of discordant twins compared with 31.5% of concordant twins (P <0.001). Major neonatal morbidities were also more frequent in the discordant group, including respiratory distress syndrome (48.9% versus 27.1%; P = 0.008), necrotising enterocolitis (15.6% versus 5.5%; P = 0.049), intraventricular haemorrhage (8.9% versus 1.1%; P = 0.017), neonatal sepsis (17.8% versus 5.5%; P = 0.016) and neonatal death (8.9% versus 1.1%; P = 0.017). The composite adverse outcome (major morbidity or NICU admission) was observed in 62.2% of discordant versus 33.1% of concordant twin pregnancies (P <0.001), indicating a substantially higher burden of neonatal morbidity associated with birthweight discordance [Table 2].

Table 2.

Neonatal outcomes of twin pregnancies by discordance group (N = 226).

	n (%)
Outcome	Non-discordant group (n = 181)	Discordant group (n = 45)	P value
NICU admission (any twin)	57 (31.5)	27 (60.0)	<0.001
Respiratory distress syndrome (any)	49 (27.1)	22 (48.9)	0.008
Necrotising enterocolitis (any)	10 (5.5)	7 (15.6)	0.049
Intraventricular haemorrhage (any)	2 (1.1)	4 (8.9)	0.017
Neonatal sepsis (any)	10 (5.5)	8 (17.8)	0.016
Neonatal death (any)	2 (1.1)	4 (8.9)	0.017
Major morbidity composite	54 (29.8)	24 (53.3)	0.005
Composite adverse outcome (major morbidity or NICU)	60 (33.1)	28 (62.2)	<0.001

NICU = neonatal intensive care unit.

In multivariable logistic regression analysis, birth-weight discordance ≥20% remained independently associated with the composite adverse neonatal outcome (aOR = 2.74, 95% CI: 1.21–6.25; P = 0.016). However, gestational age at delivery was the strongest protective factor (aOR = 0.47 per additional week, 95% CI: 0.37–0.59; P <0.001), indicating that prematurity is the principal determinant of neonatal morbidity rather than discordance alone [Table 3].

Table 3.

Multivariable logistic regression analysis for the composite adverse neonatal outcome of twin pregnancies.

Predictor	aOR (95% CI)	P value
Birthweight discordance ≥20%	2.74 (1.21–6.25)	0.016
Gestational age at delivery (per week)	0.47 (0.37–0.59)	<0.001
Monochorionic	0.69 (0.34–1.44)	0.325
Preeclampsia	2.98 (0.48–18.41)	0.240
Gestational diabetes	1.48 (0.72–3.03)	0.283

aOR = adjusted odds ratio; CI = confidence interval.

4. Discussion

Birthweight discordance in twin pregnancies has long been recognised as a clinically relevant marker associated with adverse neonatal outcomes. The current study evaluated the relationship between birthweight discordance and neonatal morbidity in a cohort of twin pregnancies managed at a regional tertiary care centre in Oman.

Antenatal ultrasound is central to fetal growth surveillance in twin pregnancy, although estimated fetal weight remains an indirect measure of fetal size. Postnatal birthweight discordance therefore provides an objective measure for classifying intertwin size discordance in retrospective outcome studies.^1,2,16,17

The threshold of ≥20% for defining clinically significant discordance is supported by multiple large cohort studies and meta-analyses demonstrating a stepwise increase in neonatal morbidity and mortality beyond this cut-off.^1,2,9 However, variation in definitions, study designs and gestational age distributions across cohorts has contributed to ongoing debate regarding the independent role of discordance in adverse neonatal outcomes.^1,2,9,16

In the current cohort, twin pregnancies complicated by clinically significant growth discordance were delivered at an earlier gestational age, consistent with prior reports linking discordance to both spontaneous and iatrogenic preterm birth.^3,5,8,9 Discordance ≥20% remained independently associated with adverse neonatal outcomes after adjustment for confounders; however, gestational age at delivery emerged as the strongest protective factor, suggesting that prematurity plays a central role in the observed association rather than discordance alone. These findings suggest that birthweight discordance primarily reflects underlying placental dysfunction and fetal vulnerability, with its impact on neonatal outcomes largely associated with earlier delivery and prematurity rather than representing a direct causal effect.^7,16,18 Clinical responses to discordance, including intensified surveillance or earlier intervention, may further contribute to this pathway by increasing the likelihood of preterm birth and its associated complications.^14,17,19 This interpretation is consistent with previous studies demonstrating that adjustment for gestational age attenuates the association between discordance and neonatal outcomes, highlighting the central role of prematurity.^3,5,6,9

Furthermore, the current study's findings support evidence that absolute birthweight and gestational maturity are stronger predictors of neonatal outcomes than intertwin weight difference alone.^2,5,14 Accordingly, birthweight discordance should be interpreted as a marker of underlying placental dysfunction and fetal vulnerability rather than as a direct causal factor, and should be considered within the broader clinical context of chorionicity, Doppler velocimetry and placental function.^11,13,16

This study's results reinforce international and regional recommendations advocating individualised surveillance and caution against iatrogenic preterm delivery based solely on growth discordance in the absence of clear evidence of fetal compromise.^14,17,19 Early and accurate determination of chorionicity, combined with structured ultrasound surveillance and Doppler-based monitoring—particularly in monochorionic twin pregnancies—remains essential for optimal clinical management.^7,17,18 In addition, preventive strategies such as timely antenatal corticosteroid administration, optimisation of preterm birth prevention measures and preparedness for NICU are critical to reducing neonatal morbidity and mortality.^20,21

The current study was subject to certain limitations. This study does not distinguish between constitutionally small twins and pathological fetal growth restriction, which represents an inherent limitation of using postnatal birthweight discordance as a surrogate marker.^1,2 The retrospective design limits causal inference and is subject to documentation variability. The unit of analysis was the pregnancy, with outcomes defined as present if occurring in at least one twin (‘any twin’), which partially accounts for within-pair correlation. However, clustering between twins within the same pregnancy was not explicitly modelled using multilevel or generalised estimating equation approaches, which may result in underestimated variance.^5,13 Formal model diagnostics, including goodness-of-fit assessment and evaluation for multicollinearity, were not systematically performed. In addition, although chorionicity was included as an adjustment variable, stratified or interaction analyses by chorionicity were not performed due to sample size limitations. Given the differing pathophysiology between monochorionic and dichorionic twins, this may limit the ability to detect subgroup-specific effects. NICU admission was included in the composite outcome but may be influenced by institutional practices, potentially inflating event rates.^1,3 Separate multivariable logistic regression analyses were not performed for individual neonatal outcomes because several events were infrequent, increasing the risk of model instability and overfitting. Important variables such as mode of conception, antenatal corticosteroid administration and indication for delivery were not consistently available in the dataset and could, therefore, not be included in the analysis. Finally, detailed antenatal Doppler parameters, postnatal follow-up and long-term neurodevelopmental outcomes were not consistently available, restricting the ability to assess the full spectrum of neonatal sequelae. Similar methodological constraints have been noted in previous retrospective twin studies.^11,12

Future prospective, multicentre studies across Oman and the Gulf region are warranted. Such studies should incorporate standardised definitions, advanced statistical approaches to account for intrapair correlation and stratified analyses by chorionicity.

In addition, inclusion of socioeconomic and environmental factors, detailed placental assessment, antenatal Doppler evaluation and long-term neurodevelopmental follow-up would provide a more comprehensive understanding of the determinants of neonatal outcomes and help optimise clinical management strategies.^17,21

5. Conclusion

Birthweight discordance of ≥20% was associated with increased neonatal morbidity and higher rates of NICU admission in twin pregnancies in this five-year study from a regional tertiary care centre in Oman. However, this association appears to be largely explained by prematurity, confirming gestational age at delivery as the principal determinant of neonatal outcomes. Clinical management of twin pregnancies should therefore integrate assessment of birthweight discordance with chorionicity, Doppler findings and serial ultrasound surveillance, while prioritising strategies to prevent and optimally manage preterm birth. Avoiding unnecessary iatrogenic preterm delivery in the absence of clear fetal compromise is likely to provide the greatest benefit in reducing neonatal morbidity.

Generative AI Declaration

Artificial intelligence tools (ChatGPT, OpenAI) were used in a limited manner for language editing, grammar refinement, sentence restructuring and formatting assistance only. No AI tools were used for data generation, statistical analysis, results interpretation or scientific decision-making. All analyses, interpretations and conclusions are the original work of the authors, who take full responsibility for the content of the manuscript.

Authors’ Contribution

Marjan Mahboubi: Conceptualization, Methodology, Formal analysis, Writing — original draft, Writing — review & editing. Shadya Mohammed Salim Al Saeghi: Conceptualization, Data curation, Investigation, Validation, Resources. Hansa Dhar: Data curation, Writing — review & editing. Dina Mohamed Abd El Fatah: Data curation. Rowaida Elraih Omran Mensour: Data curation. Qamariya Khalfan Saud Ambu-Saeidi: Supervision.

Ethics Statement

The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Ministry of Health Research and Ethics Committee (Unique Identification Code: 29335). As this was a retrospective review of anonymised medical records, the requirement for individual informed consent was waived.

Conflict of Interest

The authors declare no conflicts of interest.

Funding

No funding was received for this study.

Data Availability

Data are available upon reasonable request from the corresponding author.