Retrospective Cohort Study of Foetal Growth Restricted Births After 40 Weeks at an Australian Tertiary Maternity Hospital Between 2018 and 2021

ABSTRACT

Background

Foetal growth restriction (FGR) increases risks of neonatal morbidity and mortality. Identifying and preventing FGR is key to improving neonatal outcomes. FGR risk assessment tools are designed to recognise women at high risk of FGR.

Aims

This study aims to compare the theoretical effectiveness of four different FGR risk factor assessment pathways in identifying women at risk for FGR: New South Wales (NSW) Foetal Safety Risk Assessment, Safer Baby Bundle FGR pathway, NHS Saving Babies' Lives risk assessment tool and RCOG Screening for Small Gestational Age foetus pathway. These pathways recommend management plans for women identified to be at risk for FGR.

Materials and Methods

This retrospective study at Westmead Hospital compared singleton births between 2018 and 2021 that were ≥ 40 weeks' gestation with birthweight < 3rd percentile, against the four FGR pathways. Prior to 2022, NSW had no recommended algorithm to identify patients at risk for FGR.

Results

116 women and 116 growth‐restricted babies were included for analysis. The NSW Foetal Safety Risk Assessment would have identified 32 (27.6%) women to be at risk for FGR, the Safer Baby Bundle 39 (33.6%), the NHS tool 44 (37.9%) and the RCOG pathway 36 (30%). 32 (23%) babies required neonatal intensive care or special care nursery admission with no neonatal deaths or stillbirths.

Conclusion

Foetal risk assessment tools increase identification of women at risk of FGR, although none are perfect. Ongoing improvements to risk assessment algorithms will help identify larger cohorts of patients at risk of FGR and reduce poor neonatal outcomes.

1. Introduction

Foetal Growth Restriction (FGR) is when a foetus fails to reach its biologically determined growth potential. In clinical practice, a surrogate for measuring foetal growth restriction is foetal weight measuring very small($<$3rd percentile) [1]. This may be diagnosed early (< 32 weeks' gestation) or late (≥ 32 weeks' gestation). Identification of FGR can prevent stillbirth as these babies are susceptible to higher morbidity and mortality. This includes complications from prematurity, impaired thermoregulation and hypoglycaemia as well as perinatal asphyxia, stillbirth or neonatal death [2].

Traditionally assessment of FGR is based on examination findings using symphysial fundal height (SFH). However, the margin of error in SFH measurement may lead to undetected cases. To decrease stillbirth rates, several risk factor assessment tools have been developed by various organisations. The Safer Baby Bundle (SBB) was created by the Stillbirth Centre of Research Excellence in conjunction with the Perinatal Society of Australia and New Zealand, and the New South Wales (NSW) Ministry of Health mandated the implementation of a modified version of the Safety Baby Bundle risk assessment in 2022 in all public maternity institutions in NSW. Internationally, the United Kingdom uses guidelines from the Royal College of Obstetricians (RCOG), and this was simplified into the National Health Service (NHS) Saving Babies' Lives Bundle. All programmes aim to identify women with risk factors, recommend an ultrasound surveillance programme, and some also recommend low‐dose aspirin. The programmes are intended to be used in conjunction with clinical assessment for FGR to direct patient care, with the goal of reducing the number of babies born with growth restriction, and ultimately stillbirths. However, it is recognised that no system is completely fail‐proof.

Knowing that there are undetected FGR births that occur, the aim of this study is to compare the aforementioned four different FGR risk factor assessments pathways to see if they would have made a hypothetical difference in reducing the number of babies born with undetected growth restriction. This will be done by assessing the ability of these pathways in identifying known undetected FGR births at our institution prior to the state‐wide implementation of the NSW Foetal Safety Risk Assessment Pathway in 2022.

2. Materials and Methods

This retrospective cohort study was conducted at the tertiary‐level maternity and neonatal hospital in Sydney, Australia. Ethics approval was granted by Western Sydney Local Health District Human Research Ethics Committee (2023/ETHO1546). This study included singleton births between 2018 and 2021, who delivered ≥ 40 weeks' gestation with birthweight ($<$3rd percentile). Live births, stillbirths, as well as vaginal, instrumental and emergency caesarean section (CS) deliveries were included. Elective CS births ≥ 40 weeks were also included as the hospital supports vaginal births after CS, where women can opt for an elective CS if they do not spontaneously deliver by their choice of 40 or 41 weeks' gestation. The birth weight percentile was determined by the Fenton growth chart. This cohort was chosen as live births of growth‐restricted babies at ≥ 40 weeks are considered undetected FGR and near‐misses for stillbirth. Women with suspected or ultrasound‐confirmed FGR are usually offered an induction of labour or CS prior to this gestation. Exclusion criteria included multiple gestation pregnancies, inductions or elective CS for suspected or ultrasound‐confirmed FGR, and known foetal congenital anomalies.

Data was extracted from e‐Maternity, an NSW state‐wide electronic medical record system that collects data on obstetric history, antenatal and intrapartum care, and neonatal outcomes. This includes maternal age, mode of conception, parity, body mass index, smoking status, illicit drug use, medical conditions in the index pregnancy and ultrasound results. Maternal data was compared against four FGR pathways—the NSW Foetal Safety Risk Assessment Pathway [3], SBB FGR Care Pathway [4], NHS Saving Babies' Lives Algorithm and Risk Assessment Tool [5], and the RCOG Screening for Small‐for‐Gestational‐Age Foetus pathway [6]. Statistical analysis was completed using descriptive and quantitative statistics in Excel.

3. Results

6154 babies were born at ≥ 40 weeks' gestation between 2018 and 2021 139 (2.25%) babies born to 139 women were growth restriction < 3rd percentile. Among these women, there had been clinical suspicion for FGR in 20 of whom 10 had growth ultrasounds that confirmed FGR late in their pregnancy and one had a history of previous stillbirth (unknown cause). 3 babies were known to have congenital anomalies. These 23 cases were removed from further analysis. 116 babies and 116 women were included in this study. A comparison of the women's risk factors against the different FGR algorithms is reflected in Figure 1.

FIGURE 1.

Comparison of risk factors for foetal growth restriction (FGR) used in the different FGR pathways, and the number of women who delivered growth restricted babies at ≥ 40 weeks' gestation Westmead Hospital between 2018 and 2021 with those risk factors (N = 139). ³NSW, NSW Foetal Safety Risk Assessment Pathway; ⁴SBB, Safer Baby Bundle FGR Care Pathway (1); ⁵NHS, NHS Saving Babies' Lives Algorithm and Risk Assessment Tool; ⁶RCOG, RCOG Screening for Small‐for‐Gestational‐Age Foetus pathway; SGA, small for gestational age.

Of the 116 babies born with undetected FGR, there were no deaths or stillbirths. 27 babies (23.3%) needed special care nursery (SCN) or neonatal intensive care unit (NICU) admission. 16 neonates were admitted to the NICU/Special Care Nursery immediately after birth for respiratory distress, hypoglycaemia or metabolic acidosis, and 11 were admitted from the postnatal ward to the SCN for hypoglycaemia or jaundice. 5 (4.3%) babies were born with APGAR scores ≤ 7 at 5 min. 61 (52.6%) babies were delivered vaginally, 22 (19.0%) via instrumental delivery and 33 (28.4%) via caesarean section. Among the CS births, 23 (69.7%) were performed for suspected foetal compromise in labour. 35 (30.2%) babies had meconium liquor in labour, of whom 10 (28.6%) went on to have vaginal births.

3.1. NSW Foetal Safety Risk Assessment Pathway

The NSW Foetal Safety Risk Assessment would have identified 32 (27.6%) women as higher risk for FGR. 26 (22.4%) women would have been considered Level 3 risk and offered low‐dose aspirin from prior to 16 weeks plus serial growth ultrasounds every 2–4 weeks from 24 weeks' gestation until birth, while six (5.2%) women would have been identified as having two risk factors for FGR and offered low‐dose aspirin only from before 16 weeks with no additional ultrasounds unless clinical concern arose (Table 1). 84 (72.4%) women would have been identified as having only one risk factor for FGR and been considered suitable for routine antenatal care. No women had zero risk factors. According to this pathway, the 32 women missed out on interventions that might have detected foetuses with FGR. When comparing the real‐life obstetric and neonatal outcomes for these women individually against the pathway, up to 13 instrumental or CS deliveries that happened in actuality may not have occurred after accounting for elective CS births, as well as eight admissions to the SCN/NICU, seven births with meconium liquor, and two births with APGARS ≤ 7 (Table 2). The usual practice at our unit is for women with ultrasound‐confirmed FGR to be offered delivery earlier than 40 weeks, with potentially improved obstetric and neonatal outcomes as the babies would be less compromised. For the remaining 84 cases, applying this pathway may not have made any difference in their management and outcome.

TABLE 1.

Application of the different foetal growth restriction (FGR) risk assessment pathways^a to the women who delivered growth restricted babies at ≥ 40 weeks' gestation at Westmead Hospital between 2018 and 2021, N = 116.

NSW	Level 1 (0 RF)	cLevel 2 (≥ 2 minor RFb)	dLevel 2 (≥ 3 minor RFb)	Level 3 (≥ 1 major RF)
	84 (72.4%)	6 (5.2%)	0	26 (22.4%)
SBB	Level 1 (0 RF)	Level 2 (≥ 1 minor RF)		Level 3 (≥ 1 major RF)
	77 (66.4%)	39 (33.6%)		0
NHS	Low Risk (0 RF)	—		High Risk (≥ 1 RF)
	72 (72.1%)	—		44 (37.9%)
RCOG	Low Risk (0 RF)	Moderate Risk (≥ 1 mod RF)		High Risk (≥ 1 high RF)
	80 (30%)	15 (12.9%)		21 (18.1%)

Note: ^aFoetal Growth Restriction pathways; ^bRF, risk factor. Minor/Major/Moderate/High RF defined in Figure 1; ^cLevel 2 with ≥ 2 RF will be recommended aspirin; ^dLevel 2 with ≥ 3 RF will be recommended aspirin and growth scans.

TABLE 2.

Potential changes in neonatal outcomes after application of the different FGR risk assessment pathways^a, N = 116.

Neonatal outcome	Actual	NSW	SBB	NHS	RCOG
Births ≥ 40 weeks' gestation/No difference	116	84	77	72	80
Births < 40 weeks' gestation	0	32 (27.6%)	39 (33.6%)	44 (37.9%)	36 (31.0%)
Meconium liquor	35 (30.2%)	7 (6.0%)	8 (6.9%)	14 (12.1%)	9 (7.8%)
Assisted deliveryb	55 (47.4%)	13 (11.2%)	14 (12.1%)	18 (15.5%)	13 (11.2%)
NICUc/SCNd admission	27 (23.3%)	8 (6.9%)	10 (8.6%)	12 (10.3%)	9 (7.8%)
APGAR score ≤ 7 at 5 min	5 (4.3%)	2 (1.7%)	2 (1.7%)	2 (1.7%)	2 (1.7%)

Note: ^aFetal Growth Restriction pathways; ^bIncludes instrumental deliveries, elective and emergency caesarean sections; ^cNICU, Neonatal intensive care unit; ^dSpecial Care Nursery.

3.2. SBB FGR Care Pathway

The SBB FGR Care Pathway would have identified 39 (33.6%) women as Level 2 risk for FGR, of whom five patients had two risk factors (Table 1). No women were identified as Level 3 risk. These 39 women would have missed out on interventions to detect or prevent FGR as they would have been offered growth ultrasounds at 28 and 36 weeks' gestation, with additional growth ultrasounds as clinically indicated. 1 woman would have also been identified as high risk for preterm pre‐eclampsia and been offered low‐dose aspirin in the pregnancy. When following their real‐life obstetric and neonatal outcomes, up to 14 instrumental or CS deliveries may not have occurred after accounting for elective CS births, as well as ten admissions to SCN/NICU. There were eight births with meconium liquor, and two births with APGARS ≤ 7 (Table 2). For 77 (66.4%) women and babies, applying this pathway may not have made any difference in their management and outcome.

3.3. NHS Saving Babies' Lives Algorithm and Risk Assessment Tool

The NHS Saving Babies' Lives Algorithm would have identified 44 (37.9%) women as high risk for FGR, where nine patients had two high risk factors and one patient had three high risk factors. If the pathway was applied, 44 (37.9%) women would have been offered serial ultrasounds for foetal weight and umbilical artery dopplers, which may have identified FGR earlier in the pregnancy (Table 1). Up to 18 instrumental or CS deliveries that actually eventuated may not have occurred after accounting for elective CS births, as well as 12 admissions to SCN/NICU, 14 births with meconium liquor and two births with APGARS ≤ 7 (Table 2). For 72 (62.1%) women and babies, applying this pathway would not have made any difference in their management and outcome.

3.4. RCOG Screening for Small‐For‐Gestational‐Age (SGA) Foetus Pathway

The RCOG Screening for SGA Foetus pathway would have assessed 36 (30%) women as having risk factors for FGR, of whom 21 (18.1%) would have been considered high risk (Table 1). If the pathway was applied, 15 (12.9%) women with moderate risk factors would have been offered 4‐weekly growth scans. 12 of the 21 women considered at high risk would have been offered aspirin from before 16 weeks, and uterine artery dopplers at morphology scan. Subsequent growth scans would have been offered based on the morphology growth and doppler findings. These interventions might have detected foetuses with FGR. Tracking the obstetric and neonatal outcomes, 13 instrumental or CS deliveries may not have occurred after accounting for the elective CS births, as well as nine admissions to SCN/NICU. There were nine births with meconium liquor, and two births with APGARS ≤ 7 (Table 2). For 80 (70%) women and babies, applying this pathway would not have made any difference in their management and outcome.

4. Discussion

This retrospective cohort study highlights the effectiveness of four FGR risk assessment pathways in identifying at‐risk women. The average neonatal admission rate in Australia is 17% [7], whereas 23.3% of this cohort required SCN/NICU admission. This finding supports current evidence that low birth weight predisposes neonates to further morbidity and longer hospital admissions. It highlights the importance of having an effective risk assessment to screen and prevent low birth weights antenatally.

While all pathways would have detected some undiagnosed FGR cases, potentially improving neonatal outcomes through earlier delivery, most cases would not have changed in management or outcome. The pathways identified similar proportions of women at higher risk for FGR (27.6%–37.9%) with the NSW Foetal Safety Risk Assessment classifying the fewest, while the NHS Saving Babies' Lives Algorithm identified the most. The NSW algorithm's identification of 22.4% as Level 3 risk, coupled with the detection of an additional 5.2% patients at Level 2 risk, indicates the pathway's ability to differentiate levels of risks. While the Safer Baby Bundle algorithm identified a similar number of women with Level 2 risk factors for FGR, it did not classify any patients as Level 3, suggesting potential limitations in its ability to differentiate between risk levels. The NHS and RCOG assessment tools identified a comparable number of higher risk patients. RCOG's emphasis on major risk factors likely resulted in fewer patients being identified as at‐risk for FGR overall.

It must be noted that although the NSW pathway classified the fewest women as higher risk for FGR, it stands out as the algorithm with the potential to make the greatest difference. Most FGR risk assessment tools rely on serial growth ultrasounds; however, the NSW algorithm was unique in recommending low‐dose aspirin for FGR prevention. By contrast, the SBB recommended aspirin only for women at high risk of pre‐eclampsia and the RCOG if there was a history of FGR or hypertensive disease. That being said, while FGR and pre‐eclampsia share many risk factors and underlying placental pathology, the evidence supporting aspirin for the prevention of FGR is limited. A systematic review and metanalysis [8] (2021) demonstrated that there was no high‐quality evidence that aspirin use in pregnancy prevents late onset FGR. Additionally, low‐dose aspirin is not without side effects. As such, the NSW pathway's approach sits at odds with other risk assessment tools, highlighting the need for further research to determine whether aspirin use in this context provides meaningful benefit.

The NSW pathway was also unique in including patient ethnicity as a risk factor for FGR, namely being of South Asian, African or Aboriginal descent. There is international evidence that women from these ethnic backgrounds are at increased risk for FGR, stillbirth and neonatal mortality [9]. There is also strong evidence that women from poorer socio‐economic backgrounds have higher risks of poor perinatal outcomes [10]. Addressing ethnicity and socio‐economic status is particularly relevant to our institution as it services a metropolitan region that has a high migrant population and a high proportion of socio‐economic disadvantage [11]. The inclusion of ethnicity in multicultural communities highlights the importance of validating and optimising risk assessment tools within the populations they are intended to serve. Future research is needed to determine how pathways can be customised to multi‐ethnic populations.

The study underscores the need for ongoing evaluation and refinement of FGR risk assessment pathways. According to the pathways, almost one third of undetected FGR cases potentially could have been identified during serial growth ultrasounds, emphasising the pivotal role of implementing any type of FGR risk assessment algorithm with the aim of reducing FGR morbidity and mortality. The variability in the performance of different pathways suggests that a customised or combined approach could enhance accuracy. However, while these pathways improve detection of true FGR, they could increase the risk of false positives. This can lead to anxiety and higher rates of unnecessary inductions of labour, caesarean sections and/or preterm deliveries. Another consideration alongside risk‐factor–based pathways is the role of universal third‐trimester growth scans for detection of FGR. While universal ultrasound can improve the detection of FGR compared with selective screening, it has not been shown to reduce adverse perinatal outcomes [12]. Moreover, universal scanning also increases false positives, leading to more interventions that increase demands on service capacity. It must be recognised, however, that no algorithm is completely fool‐proof and that there will still be a proportion of women who will not be picked up by any screening pathway. Ultimately, the evaluation of any pathway must balance the potential benefits of improved detection against the risks of over‐intervention and resource utilisation.

A limitation of this study is the small sample size, as it only includes women with undetected FGR who delivered ≥ 40 weeks, excluding deliveries, stillbirths and neonatal deaths related to FGR at earlier gestations. A larger sample might have enabled statistical analysis to compare pathway effectiveness. The retrospective nature of the study and reliance on electronic medical records also poses limitations. Prospective studies with real‐time monitoring and a larger sample size could provide more robust insights. Additionally, investigating the impact of interventions and adherence to recommended guidelines on perinatal outcomes would be valuable. It is intended that a similar audit of growth restricted births ≥ 40 weeks' gestation will be performed following the implementation of NSW Foetal Safety Risk Assessment, and a comparison made of the two epochs, to determine if the assessment tool is effective in reducing the number of undetected FGR births.

In conclusion, this study contributes valuable insights into the challenges of FGR identification and prevention, emphasising the importance of continuous improvement in risk assessment algorithms to enhance maternal, foetal and neonatal outcomes.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.