Universal late pregnancy ultrasound screening to predict adverse outcomes in nulliparous women: a systematic review and cost-effectiveness analysis

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 15/105/01. The contractual start date was in March 2017. The draft report began editorial review in August 2019 and was accepted for publication in March 2020. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2021 Smith et al. This work was produced by Smith et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2021 The authors

Screening for pregnancy complications

Complications of pregnancy are a major contributor to the global burden of disease as a result of the effects on both the mother and the infant. 1 Identifying and managing the risk of complications is a key element of antenatal care that aims to reduce the number and severity of adverse outcomes. Current clinical guidelines2 describe multiple methods of identifying high-risk women, including (1) identification of maternal risk factors associated with disease (e.g. obesity, being aged > 40 years), (2) assessment of complications in previous pregnancies, (3) identification of pre-existing medical conditions (e.g. diabetes mellitus) and (4) clinical presentation with symptoms that are associated with an increased risk of adverse outcome (e.g. antepartum haemorrhage, reduced fetal movements). In addition, multiple tests are given to pregnant women to assess their risk. Taking the example of screening for Down syndrome, a woman’s risk is first assessed by maternal age; this background risk is then adjusted for the results of ultrasonic imaging (nuchal translucency) and biomarkers (pregnancy-associated plasma protein A and free beta subunit of human chorionic gonadotrophin), and the summative risk is used to inform the use of invasive testing. 3

Use of ultrasound in pregnancy screening

The first trimester ultrasound scan used to screen for Down syndrome is an example of a scan that is offered to all pregnant women as part of their risk assessment. Routine pregnancy care in the UK also involves a second screening ultrasound scan, performed at or after 18 weeks’ gestation but before 21 weeks’ gestation, the primary purpose of which is to identify fetuses with structural abnormalities. 3 A positive result from this scan might inform decisions around termination of pregnancy (e.g. some women may choose to terminate a pregnancy if the fetus has a severe neural tube defect) or it might inform the need for targeted follow-up and changes to the perinatal care of the infant. For example, identifying a congenital diaphragmatic hernia could lead to invasive testing for aneuploidy, prenatal discussions with the paediatric surgery team and modification to neonatal resuscitation (e.g. early intubation to avoid expansion of the stomach with air).

In the UK and the USA, universal ultrasound is not recommended after the mid-pregnancy anomaly scan. 2,4 Instead, it is recommended that ultrasound be offered in a targeted manner and only to women in whom there is a clinical indication. Such indications could include presentation with symptoms (e.g. antepartum haemorrhage), relevant medical history (e.g. antiphospholipid antibody syndrome) and relevant medical history [e.g. previous fetal growth restriction (FGR)], or result from physical examination [e.g. the fetus is small for gestational age (SGA)] on clinical examination.

Use of ultrasound in late pregnancy

When ultrasound scans are performed in late pregnancy, a number of features are commonly reported. Ultrasound allows the estimation of the size (length and circumference) of fetal parts, termed fetal biometry. A variety of methods exist for converting these measurements to an estimated fetal weight (EFW)5 and a number of reference ranges exist for EFW in relation to the exact gestational age. 6,7 The interpretation of EFW and the individual biometric measurements generally focuses on two properties: the position of the value on the distribution for the given gestational age and the change in the value over serial measurements. Taking the first of these, infants in the smallest 10% of measurements for gestational age are referred to as SGA and infants in the largest 10% are referred to as large for gestational age (LGA). The second property examines the growth velocity across the pregnancy. For example, if a fetus is on the 9th percentile at 36 weeks’ gestation and it had also been on the 9th percentile at 20 weeks’ gestation, it would be regarded as SGA but with normal fetal growth velocity. SGA infants with normal growth velocity are often constitutionally small. SGA combined with evidence of reduced fetal growth velocity is regarded as indicating FGR. 8

Another major category of measurement in ultrasound in late pregnancy is Doppler flow velocimetry (referred to as ‘Doppler’). 9 In brief, a blood vessel is imaged and electronic callipers on the screen are placed over the vessel. The machine then plots out the velocity of flow on the y-axis, with time on the x-axis. The resultant plot is termed a flow–velocity waveform. Different blood vessels have different patterns of flow–velocity waveform and the pattern is analysed both qualitatively and quantitatively. One of the key blood vessels for study is the umbilical artery. Flow is characterised qualitatively by the direction of flow in end-diastole (i.e. immediately prior to the rise in flow that occurs with a heartbeat, i.e. systole). The normal state is forward flow, but there can be absent flow or even reversed flow. The waveform can also be analysed mathematically, and a number of indices have been described, such as the pulsatility index (PI) and resistance index (RI). The derivation, calculation and detailed interpretation of these indices are described in detail elsewhere. 9 However, both values correlate positively with the presumed resistance to flow in the vascular bed supplied by the artery. Hence, high values of PI and RI in the umbilical arteries are interpreted as indicating a high resistance to flow in the fetal vascular tree of the placenta. Correlative studies of umbilical artery Doppler and placental microscopy support this interpretation in cases of FGR occurring before 36 weeks’ gestation. 10

The four most common sites for Doppler are the umbilical arteries, the maternal uterine arteries, the fetal middle cerebral arteries (MCAs) and the ductus venosus. 9 In contrast to the other three, it is low resistance in the fetal MCAs that is thought to indicate compromise. The interpretation is that a reduced level of oxygen in the fetal blood leads to cerebral vasodilation and, hence, reduced measures of resistance in the arteries supplying the brain.

Other features that are examined in late pregnancy include the placenta, the amniotic fluid and the fetal presentation. Reporting on the placenta generally focuses on its site in relation to the cervix. Implantation of the placenta over the cervix is called placenta praevia and it can cause massive haemorrhage during labour. Reduced amniotic fluid is called ‘oligohydramnios’ and increased amniotic fluid is called ‘polyhydramnios’. Amniotic fluid volume is quantitatively assessed by measuring the biggest single pool (deepest vertical pool) or by summing the four deepest pools in each quadrant of the uterus (amniotic fluid index) (AFI). One of the simplest findings on scan is the presentation of the fetus. Near term, > 95% of fetuses present head first. Women are examined close to term to assess presentation, but this approach frequently misses infants presenting breech. 11 Ultrasound unambiguously establishes the presentation at the time of a scan.

Coupling interventions to scan results

A limited number of disease-modifying interventions can be coupled with ultrasound performed in late pregnancy to alter the outcome of pregnancy. Most of the interventions involve modifications to either the timing of delivery [e.g. induction of labour (IOL)] or the mode of delivery (e.g. delivery by pre-labour caesarean section). One exception to this is breech presentation. It has been known for many years that vaginal breech delivery, although safe for the majority of women, can be associated with complications that could have severe consequences for the infant. Breech delivery is associated with a number of specific complications, such as increased risk of umbilical cord compression and entrapment of the fetal head after delivery of the fetal body. Vaginal breech birth in the UK has been shown to be associated with an absolute risk of death during labour or in the first 4 weeks of life of 8.3 per 1000. Although the absolute risk is low, it is much higher than the risk associated with a planned caesarean delivery of 0.3 per 1000. 12 The risks associated with vaginal breech birth (an awareness of which has long predated the epidemiological study confirming the higher risk of death) were the basis for the procedure to turn the infant from breech to a cephalic presentation using manual manipulation by a clinician, called external cephalic version (ECV). If this procedure is unsuccessful, generally, delivery by planned caesarean section is recommended. 13 This is based both on the observational data of increased risks associated with vaginal breech birth and on the results of randomised controlled trials (RCTs) of planned caesarean section, which have confirmed the reduced risk of perinatal death with this procedure, compared with planned vaginal breech birth. 14

In the case of most of the other diagnoses that may be made by ultrasound, the primary disease-modifying intervention in the second half of pregnancy is to deliver the infant either by IOL or by planned caesarean section. However, screening may also be used to inform the assessment of fetal well-being to help inform the timing of this intervention. For example, if an infant is found to be SGA and FGR is suspected, there are multiple ways to assess the well-being of the infant. However, these simply constitute another layer of diagnostic and prognostic tests, and ultimately, are used to target the timing of disease-modifying interventions in delivery. The primary reason for expediting delivery is that IOL removes the subsequent risk of stillbirth (i.e. intrauterine fetal death followed by the delivery of an infant showing no signs of life). Most cases of stillbirth are due to complications that can occur to the fetus only in utero (e.g. placental abruption or placental failure); hence, delivering the fetus removes the risk of stillbirth. 15 This is confirmed by RCTs that demonstrate that IOL at term is associated with a 67% reduction in stillbirth risk. 16

Although early delivery can be performed safely at term, this is not the case preterm. A Cochrane review16 described exactly the same reduction in the risk of perinatal death with IOL at term as was observed for stillbirth. Perinatal deaths include both stillbirths and neonatal deaths, and hence the favourable effect of IOL on stillbirth was not cancelled out by an unfavourable effect on the risk of neonatal death. However, preterm birth is one of the major determinants of neonatal death, and, therefore, if women are routinely induced preterm, reducing the risk of stillbirth will be outweighed by the increased risks of intrapartum stillbirth and neonatal death associated with prematurity. The inflection point (i.e. where the risks balance out) has previously been estimated as between 38 and 39 weeks’ gestation. 17 Hence, although 37 weeks’ gestation is, strictly, term, routinely delivering all women at 37 weeks’ gestation could increase overall perinatal mortality as a result of higher rates of intrapartum stillbirth and neonatal death. 18 It follows, therefore, that screening using a test with a high false-positive rate has the potential to cause net harm by increasing iatrogenic prematurity (or early term delivery) in false positives. 19

Evidence for screening using universal late pregnancy ultrasound

There is strong evidence to support the use of ultrasound scanning in high-risk pregnancies. A systematic review of umbilical artery Doppler has shown that this procedure reduces perinatal mortality by about 30% in high-risk pregnancies. 20 The mechanism of the effect is likely to be explained by the fact that its use is also associated with lower rates of IOL and caesarean delivery. Hence, it is likely that the use of Doppler reduces the risk of perinatal death overall by reducing unnecessary intervention. However, there is also a strong trend towards a reduced risk of stillbirth, indicating that Doppler may also be useful for targeting intervention to the highest-risk pregnancies.

The fundamental role of ultrasound scanning in the care of high-risk women led researchers to explore whether or not routinely using the same approaches might improve outcomes in low-risk women. Disappointingly, a meta-analysis of 13 RCTs comprising ≈ 35,000 women did not demonstrate any evidence that routine ultrasound scanning improved outcome. 21 It is this finding that has led to the recommendation that ultrasound should not routinely be performed in the second half of pregnancy in the UK and the USA. The cautious approach is supported by some evidence from countries where universal late pregnancy ultrasound has been introduced, despite the lack of strong evidence supporting its clinical effectiveness. A seminal study22 from France reported rates of adverse perinatal outcome in relation to women’s screening status for SGA. Each woman’s screening status was identified [screened positive for SGA or screened normal, i.e. appropriate for gestational age (AGA)] and the actual status of the infant at birth was also assessed (SGA or AGA by actual birthweight). The authors subsequently described rates of perinatal morbidity and mortality by true-positive and false-positive status. As one might have predicted, false positives had higher rates of multiple adverse outcomes than AGA infants that were true negatives, and this was explained primarily by higher rates of iatrogenic prematurity among the false positives. Interestingly, the true-positive SGA infants also had higher rates of adverse outcomes that were missed by scanning than SGA infants (false negatives). The former observation confirms that screening has the potential to result in iatrogenic harm to false positives. The latter observation questions the rationale for screening for SGA infants in late pregnancy at all.

Critical analysis of the Cochrane review16

Although it is generally accepted that a systematic review of RCTs represents the highest level of evidence, a number of features of the systematic review of RCTs of universal ultrasound21 undermine its main conclusions.

• All of the 13 studies in the meta-analysis used different definitions of ‘screen positive’. Moreover, some of the ultrasound findings were completely divergent. For example, whereas multiple studies analysed some variant of an estimation of fetal size, one large study assessed placental calcification without assessing any other features of the scan. An implicit assumption around combining these studies is that these different ultrasonic tests all had comparable effectiveness, which a subsequent systematic review of diagnostic test accuracy (DTA) studies has demonstrated is not the case. 23

• None of the studies was preceded by a high-quality assessment of the diagnostic effectiveness of the test in a low-risk population. This is problematic for a number of reasons. A key element of study design is a power calculation. It is impossible to perform a power calculation without quantitative information on the diagnostic effectiveness of a test. Moreover, the tests had generally been developed for and evaluated in high-risk populations. It is well recognised in screening that test performance differs according to the risk status of the population. One of the key outcomes of a screening test is the positive predictive value (PPV) (i.e. the proportion of women screening positive who experience the outcome). The PPV of a test is determined by the prior risk of disease multiplied by the positive likelihood ratio (LR+ = the proportional increase in the odds among screen-positive women compared with the whole population). Hence, the higher the prior risk of disease, the higher the PPV for a given positive likelihood ratio. Consequently, it is typical that a positive screening test is associated with a much lower PPV in a low-risk population. As the PPV determines the ratio of true positives to false positives, this will have a major impact on trials of screening.

• None of the 13 RCTs coupled the screening test with an intervention. In all 13 studies the result was revealed to the attending clinicians but no specific intervention was planned. It is self-evident that a screening test could have an impact on an outcome only if it is coupled with an intervention. Moreover, the tests were performed at a wide range of gestational ages. Given that the primary intervention available to the attending clinicians would have been delivery of the infant, the potential for this to result in benefit or harm would vary according to the gestational age at which the scan was performed. Hence, a positive effect of late pregnancy ultrasound and delivery could have been masked by a negative effect of preterm pregnancy ultrasound scan with higher rates of iatrogenic harm.

• Although the meta-analysis included 35,000 women, it was still underpowered for the key outcome of interest: perinatal death. The risk ratio for perinatal death from the meta-analysis was 1.01 with a 95% confidence interval (CI) of 0.67 to 1.54. Although this CI may seem quite narrow, the capacity for reducing the rate of an outcome with a screening trial is different from interventional trials in women with established disease. If we identified a screening test for perinatal death with a positive likelihood ratio of 10 and a 5% screen-positive rate, and if we applied an intervention that reduced the risk by 50%, the estimated relative risk would be 0.76, which is within the 95% CI of the systematic review. Hence, the Cochrane review16 is underpowered to detect the effect of a highly effective screening test coupled with a highly effective intervention. If we use the 5.8 per 1000 perinatal mortality rate in the control group of the Cochrane review, a power calculation indicates that a sample size of 110,000 women would be required to detect this effect with 90% power.

Parity and the risk of adverse outcome

One of the most important determinants of adverse pregnancy outcome is obstetric history (i.e. the outcome of previous pregnancies). Many conditions of pregnancy have quite high risks of recurring in subsequent pregnancies, such as pre-eclampsia,24 preterm birth,25 stillbirth26 and FGR. 27 Hence, women who have experienced complications in previous pregnancies generally receive enhanced antenatal care. Conversely, uncomplicated previous pregnancies are strongly predictive of a normal outcome in future pregnancies. Hence, women who have had a previous vaginal delivery of a normally grown liveborn infant at term following an uncomplicated pregnancy have a low absolute risk of complications in future pregnancies. 28 Obstetric history is, necessarily, not available for women who have not had previous births. Although maternal characteristics, as described above, are associated with the risk of pregnancy complications, the associations are generally rather weak and perform poorly as a screening test in isolation. 29 Moreover, first pregnancies, collectively, have higher rates of complications than second pregnancies. This increased rate of complications has identified first pregnancies as a priority area for research. Quoting a National Institutes of Health (NIH) study description of nulliparous women:

This large proportion of women lacks previous pregnancy information to guide risk assessment; as such, adverse outcomes in these first pregnancies are particularly difficult to predict and prevent.

Haas et al. 30

Summary of the rationale for the focus on nulliparous women in late pregnancy

The characteristics above provide the rationale for the focus of this review. Screening and intervention near term has less potential to cause harm than screening and intervention in the preterm period, as the primary intervention, delivery of the infant, is less likely to lead to iatrogenic injury. The need for screening is greatest in the nulliparous population because their background suggests that they are at higher risk of an adverse outcome and they lack one of the key discriminating characteristics of risk assessment: knowledge of the outcome of prior births.

The health economics of screening and intervention

A critical consideration in relation to screening and intervention using universal ultrasound is whether or not this screening is cost-effective. It is possible that, for the individual woman and infant, having a screening ultrasound scan and associated intervention leads to a better outcome but that the cost of providing the screening test and intervention results in net societal harm as it removes resources from other more cost-effective elements of the health-care system. The capacity of all health-care systems is finite; however, systems differ in their willingness to pay (WTP). These questions are addressed quantitatively in health economic analyses by calculating the sum of money required to gain one additional quality-adjusted life-year (QALY), a subject that is discussed in detail elsewhere. 31 In NHS England, interventions are considered cost-effective if the cost of each QALY is below a given threshold, and this is typically between £20,000 and £30,000.

Providing a late pregnancy ultrasound scan will clearly incur direct costs. Managing women who are assessed as high risk after screening will clearly incur further costs. However, these additional costs then have to be set against the reduction in harm (i.e. the QALYs gained by the mother or child because of being screened). Many of the individual elements required for these calculations are associated with uncertainty. Hence, these health economic analyses frequently employ a probabilistic approach, running large numbers of simulations where the different parameters for the models are sampled from the presumed plausible range of values from the literature. These methods and their interpretation are discussed in more detail in Chapter 11.

Value-of-information analysis

The health economic analyses described above relate to the economic case for implementing a given programme of screening and information. Value-of-information (VOI) analysis addresses the economic case for funding research to try to reduce the uncertainty in the evidence base. Generally speaking, a research question that will be identified as being cost-effective from this perspective will have uncertain input values (i.e. the CIs for the given parameter in the literature are wide). Moreover, questions identified as being cost-effective in a VOI analysis will often generate highly variable results in sensitivity analyses in which the input value of the parameter is varied within the range of uncertainty. This subject is again dealt with in detail in Chapter 11.

Designing a randomised controlled trial

Randomised controlled trials of screening have certain differences from RCTs of other interventions. Typically, interventions are evaluated in populations with a disease and so the individuals recruited will have high rates of complications as a result of disease. Moreover, most of the outcomes in the group are likely to be related to the disease process. By contrast, screening, by design, focuses on individuals before they manifest disease so the background rate of serious adverse outcomes is likely to be low. Moreover, adverse outcomes in the population are likely to be from diverse causes, not simply the disease being screened for. For example, a RCT studying mortality rates among people with cancer is likely to show high rates of death in the different arms of the trial and most of the deaths in both arms are likely to be related to cancer. By contrast, a RCT of screening or not screening a healthy population for the same cancer is likely to have low rates of deaths in both arms and many of the deaths would be unrelated to the experience of cancer. Both of these factors will tend to increase the sample size in the screening study as there is a low incidence of adverse outcomes and only a subset of the adverse outcomes will be preventable by the given programme of screening and intervention.

We previously reviewed the approach to screening in pregnancy32 and highlighted an alternative, namely that all women in a population be screened and that randomisation is to either revealing or masking the result. Those that have the result revealed will have an intervention as required, and those that have the result masked will receive routine care. Using this design, randomisation is performed in a group that has a higher rate of complications (by virtue of the positive screening test) and a greater proportion of the adverse events will be related to disease being screened for. This approach has the advantages that the overall number needed to screen for statistical power is substantially reduced and that the screening test can be validated in the same study design by comparing screen-negatives with screen-positives randomised to have the result masked. These issues are discussed further in Chapter 13.

Methods

Results

Meta-analysis

The literature search Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram is presented in Appendix 1, Figure 26. We identified 13 studies35,42–53 that met our inclusion criteria and these involved a total of 67,764 patients. The study characteristics are presented in Appendix 1, Table 19. Five studies35,42,48,51,52 (n = 63,436) included unselected pregnancies as part of universal screening, four studies43,46,47,53 (n = 2634) included low-risk pregnancies only and four studies44,45,49,50 (n = 1694) included mixed-risk pregnancies. Three of the studies42,51,52 that were done in the same hospitals may have had short periods of overlap. Nine studies35,43,44,46–50,53 (n = 8097) were prospective and four42,45,51,52 (n = 59,687) were retrospective. Studies varied in relation to the gestational age at scan (ranging from 28 to 41 weeks’ gestation), as well as in the indices and the cut-off points used. The majority of patients in the included studies delivered at term. The assessment of study quality is presented in Appendix 1, Figure 27. Overall, the quality was variable. The main risk of bias was that only six studies35,43,44,46,48,50 (n = 5777) blinded clinicians to the umbilical artery Doppler result. However, five of these six studies revealed other features of the scan result, such as fetal biometry. Only the POP study35 blinded participants to the results of both the uteroplacental Doppler and fetal biometry.

The summary results of the meta-analysis are presented in Table 2. The pattern of results was very similar to that in the POP study. A high-resistance pattern detected by Doppler was associated with an increased risk of delivering a SGA infant or a severely SGA infant. However, the finding was not strongly predictive, with positive LRs between 2.5 and 3.0. A high-resistance pattern of umbilical artery Doppler was not associated with an increased risk of a range of indicators of neonatal morbidity. The summary receiver operating characteristic (ROC) curves are presented in Figure 1. For some outcomes, such as 5-minute Apgar score of < 7, caesarean section for fetal distress and pre-eclampsia, the Rutter–Gatsonis model could not produce summary results despite an adequate number of studies. We also pooled DORs for all the reported outcomes (Figure 2) and illustrated the variation between studies using forest plots. Finally, we used Deeks’ funnel plot asymmetry test to assess the risk of publication bias using the outcome of neonatal unit admission for the analysis (see Appendix 1, Figure 28). The test showed no evidence of publication bias (p = 0.52).

TABLE 2 - Summary diagnostic results of meta-analysis of umbilical artery Doppler for predicting adverse pregnancy outcome

The pattern of definition varied between studies and includes one or more of the following: stillbirth, neonatal death, hypoxic–ischaemic encephalopathy, inotrope support or severe metabolic acidosis.

Outcome	Number of studies	Number of patients	Summary sensitivity (%), (95% CI)	Summary specificity (%), (95% CI)	Summary positive LR (95% CI)	Summary negative LR (95% CI)
SGA < 10th centile	8	19,203	21.7 (13.2 to 33.6)	91.8 (86.5 to 95.1)	2.65 (1.89 to 3.72)	0.85 (0.77 to 0.94)
SGA < 3rd centile	5	53,907	25.4 (14.0 to 41.5)	90.4 (78.6 to 96.1)	2.65 (1.92 to 3.66)	0.83 (0.75 to 0.91)
NICU admission	8	66,253	13.6 (6.8 to 25.3)	89.9 (83.5 to 94.0)	1.35 (0.93 to 1.97)	0.96 (0.90 to 1.03)
Neonatal acidosis	5	9629	12.0 (5.3 to 25.0)	91.1 (81.0 to 96.1)	1.34 (0.86 to 2.08)	0.97 (0.91 to 1.02)
Severe adverse pregnancy outcomea	4	58,866	9.3 (4.8 to 17.5)	88.3 (74.5 to 95.2)	0.80 (0.44 to 1.46)	1.03 (0.95 to 1.11)

FIGURE 1.

Summary ROC curves for umbilical artery Doppler for predicting (a) neonatal intensive care unit admission; (b) neonatal metabolic acidosis; (c) SGA (< 10th centile); and (d) severe SGA (< 3rd centile).

FIGURE 2.

Meta-analysis of DORs of umbilical artery Doppler at predicting (a) neonatal intensive care unit admission; (b) neonatal metabolic acidosis; (c) 5-minute Apgar score of < 7; (d) severe adverse perinatal outcome; (e) caesarean section for fetal distress; (f) pre-eclampsia; (g) SGA (< 10th centile); and (h) severe SGA (< 3rd centile).

Discussion

The main finding of this study was that the umbilical artery Doppler has moderate predictive accuracy in detecting SGA and severely SGA infants. However, it did not predict neonatal morbidity at term. The results were very similar in both the POP study and the meta-analysis that included the POP study and other published studies. The only notable difference between the analysis of the POP study and the meta-analysis including the POP study is that the association in the former was slightly stronger for severe SGA. The outcome of SGA is used as a proxy for FGR. As discussed in Chapter 1, FGR is a theoretical concept with no gold standard. SGA is used as a proxy for FGR but it is recognised that only a proportion of SGA infants are small because of FGR. As the threshold for defining SGA is lowered, the proportion of cases that are truly FGR increases. Hence, the stronger association with severe SGA is most likely explained by a true association between high-resistance patterns of umbilical artery Doppler and FGR.

The similar associations between the POP study and the meta-analysis is reassuring. Of all the studies evaluated, only the POP study blinded both the Doppler result and fetal biometry. A lack of blinding in studies could lead to bias. First, revealing the results could lead to interventions that then improve the outcome of the pregnancy. In this case, an investigation that is truly predictive for adverse outcome may not appear to be so when evaluated in a study where the result is revealed, as knowledge of the result leads to interventions that prevent the adverse outcome. However, revealing the result could also lead to a non-informative test being wrongly identified as predictive of adverse outcome. The primary intervention following a concerning ultrasound finding is to deliver the infant, which, if performed pre term or at early term, can cause iatrogenic morbidity. Hence, a non-informative test could appear to be associated with adverse neonatal outcome when evaluated in a study where the result is revealed because revealing the result leads to interventions that cause iatrogenic morbidity. Moreover, if outcomes include events that are defined on the basis of the results of the diagnostic test being evaluated, there is the risk of ascertainment bias. For example, if the presence of abnormal umbilical artery Doppler is used to define caesarean section for fetal distress, there could be an association between the two because the test was being used to classify the outcome.

The lack of association between umbilical artery Doppler and adverse neonatal outcome is likely to be explained by two reasons. First, a minority of term SGA infants have abnormal umbilical artery Doppler. This study showed that about one in five of the SGA infants born below the tenth birthweight centile and one in four of those born below the third birthweight centile had an abnormal umbilical artery Doppler. Second, only a small percentage of overall morbidity at term is associated with abnormal fetal growth. For example, previous studies of perinatal death at term have demonstrated that only one in three stillbirths at term is associated with abnormal fetal growth. 54 This association would probably be even weaker for other outcomes, such as neonatal intensive care unit (NICU) admission, which includes morbidity for various reasons not related to fetal size, such as neonatal infection. It is plausible that umbilical artery Doppler would be more strongly predictive of adverse neonatal outcome in fetuses who were actually SGA, and this has been confirmed in a previous analysis of the POP study. 8

Given that umbilical artery Doppler appears to be predictive of FGR in low-risk women, it might be regarded as surprising that the RCTs of its use as a screening test failed to demonstrate any benefit. However, a previous analysis of required sample sizes of screening and intervention to prevent stillbirth demonstrated that, even if a test had a positive LR of 5 for perinatal death, and was observed in 5% of women, and even if the test was coupled with an intervention that reduced the risk of perinatal death by 50%, a RCT of screen versus no screen would need to recruit ≈ 300,000 women to achieve 90% power (see supplementary figure 10 in Flenady et al. 55). Thus, the Cochrane meta-analysis of low-risk pregnancies is significantly underpowered to identify a reduction in perinatal death.

In conclusion, a high-resistance pattern of umbilical artery Doppler is somewhat predictive of the risk of delivering a SGA infant. The strength of prediction was similar using a blinded 36 weeks’ gestation scan in unselected nulliparous women in the POP study as it was in a systematic review of the wider literature.

Methods

Results

The literature search flow chart is presented in Appendix 2, Figure 29. We identified 16 studies42,57–71 that met the inclusion criteria, which involved a total of 121,607 patients. The study characteristics are presented in Appendix 2, Table 20. Four studies42,57,58,68 (n = 85,059) included unselected pregnancies, seven studies59,60,62,63,66,67,70 (n = 12,929) included low-risk pregnancies only and five studies61,64,65,69,71 (n = 23,619) included mixed-risk pregnancies. Nine studies (n = 87,208) were prospective and seven (n = 34,399) were retrospective. There was population overlap between the Akolekar et al. ,57 Akolekar et al. 42 and Bakalis et al. 58 studies. For the first two we reported different outcomes and for those outcomes that were the same we employed the data from the larger Akolekar et al. 42 study in the meta-analysis. In the study by Bakalis et al. ,58 ultrasound was performed at 32 weeks’ gestation, compared with the two Akolekar et al. 42,57 studies, in which ultrasound was performed at around 36 weeks’ gestation. There was also population overlap between the Khalil et al. ,62 Monaghan et al.,64 and Morales-Roselló et al. 65 studies, which reported different outcomes at the same tertiary maternity unit. Moreover, there was population overlap between the Flatley and Kumar,61 Sabdia et al. 69 and Twomey et al. 71 studies. In the study by Twomey et al.,71 ultrasound was performed at 32 weeks’ gestation, and the other two studies, in which ultrasound was performed between 35 and 38 weeks’ gestation, reported different rates of nulliparity and different gestational age at delivery (Sabdia et al. 69 included preterm deliveries), which indicates that the potential population overlap was not significant. Furthermore, there was a complete population overlap between the studies by Bligh et al.,59,60 but the two studies reported different outcomes.

The assessment of study quality was performed using the QUADAS-2 tool and is summarised in Appendix 2, Figure 30. The main risk of bias was for reference standard because of the lack of blinding in the majority of studies. Only five studies59,60,66–68 (n = 3079) blinded the clinicians to results. The second most common risk of bias was for flow and timing because of the different gestational ages at which ultrasound was performed. In the studies by Bakalis et al. ,58 Rial-Crestelo et al. 68 and Twomey et al.,71 ultrasound was performed at around 32–33 weeks’ gestation, and in Prior et al. 66,67 and Stumpfe et al. ,70 it was performed prior to IOL (interval between ultrasound and delivery of < 72 hours). Hence, the results of the above studies might not be applicable to universal screening at 36 weeks’ gestation. One study63 had unclear risk of selection bias as it did not specify whether the selection of patients was consecutive or random.

The summary results for the diagnostic accuracy of CPR at predicting adverse pregnancy outcomes are presented in Table 3. Overall, the strongest associations were with the risk of delivering a SGA or severely SGA infant and the positive LRs were in the region of 3.5–4.0, which was stronger than for umbilical artery on its own. Moreover, unlike umbilical artery Doppler in Chapter 4, a low CPR was associated with a statistically significantly increased risk of neonatal morbidity. However, the strength of prediction was weak, with positive LRs of between 1.5 and 3.0.

The summary ROC curves are presented in Figure 3. Generally, the larger studies reported lower sensitivities and higher specificities for all the outcomes. We also present the pooling of the DORs in Figure 4. These demonstrate that, for many of the outcomes, there was a very high level of heterogeneity between the studies.

FIGURE 3.

Summary ROC curves for the diagnostic performance of abnormal CPRs at predicting adverse pregnancy outcomes. (a) Neonatal unit admission; (b) 5-minute Apgar score of < 7; (c) neonatal metabolic acidosis; (d) severe adverse perinatal outcome (including stillbirth, neonatal death and hypoxic–ischaemic encephalopathy); (e) SGA (birthweight < 10th centile); (f) severe SGA (< 3rd or < 5th centile); (g) caesarean section for fetal distress; and (h) operative delivery for fetal distress (including both caesarean section and instrumental delivery).

FIGURE 4.

The diagnostic odd ratios for the diagnostic performance of abnormal CPRs at predicting adverse pregnancy outcomes. (a) neonatal unit admission; (b) 5-minute Apgar score of < 7; (c) neonatal metabolic acidosis; (d) severe adverse perinatal outcome (including stillbirth, neonatal death and hypoxic–ischaemic encephalopathy); (e) SGA (birthweight < 10th centile); (f) severe SGA (< 3rd or < 5th centile); (g) caesarean section for fetal distress; and (h) operative delivery for fetal distress (including both caesarean section and instrumental delivery).

Furthermore, we used Deeks’ funnel plot asymmetry test to assess the risk of publication bias using the outcome of neonatal unit admission for the analysis. The test showed no significant risk of publication bias (p = 0.28; see Appendix 2, Figure 31).

Discussion

The meta-analysis demonstrated that the CPR may be slightly more predictive than umbilical artery Doppler in identifying pregnancies at an increased risk of adverse outcome. In the case of SGA, the positive LRs were in the region of 3.5–4.0, compared with 2.5–3.0 for umbilical artery Doppler. Moreover, unlike umbilical artery Doppler, a low CPR was associated with an increased risk of neonatal morbidity. However, in this case the strength of prediction was weaker, with positive LRs of < 2.0. Moreover, in both analyses, there was very significant heterogeneity in relation to both birthweight-based outcomes and neonatal morbidity. Consequently, the 95% CIs for the positive LR are wide and include the point estimates observed for umbilical artery Doppler for both SGA and severe SGA. Furthermore, given that many of the studies were not blinded, it is possible that the associations with neonatal morbidity were a result of bias. However, the association between the CPR and SGA fetuses indicates that the ratio is likely to predict FGR. Overall, this analysis indicates that the CPR is indeed predictive of adverse pregnancy outcome. However, it is not clear from the present analysis whether or not the ratio performs better than simply assessing the result of umbilical artery Doppler, which is used in its calculation anyway. Of the indices assessed in these sections of the report, only MCA Doppler was not measured in the POP study; hence, unlike in the other chapters, we are unable to compare the strength of association in the POP study with the meta-analysis. Our findings contradict the previously published systematic review,72 which concluded that the CPR at term has a strong association with adverse obstetric and perinatal outcomes. We believe that this is because the systematic review by Dunn et al. 72 included studies carried out in mostly high-risk populations, did not include some large, recently published studies that offered ultrasound as part of universal screening42,57,58 and did not produce any pooled analyses.

There are other issues that should be taken into account when considering MCA Doppler as a screening test in unselected nulliparous women near term. First, the infant’s head often engages earlier in nulliparous women and it can be technically difficult to use MCA Doppler when the head is deeply engaged. Second, the safety of ultrasound has been established in RCTs but MCA Doppler was not performed in these. The main concern with ultrasound is the potential for harm caused by heating tissues. The form of ultrasound that is most strongly associated with heating is pulsed wave Doppler ultrasound. Hence, there is a theoretical safety concern about the infant’s brain being heated as a result. In high-risk pregnancies, the balance of risks and benefits probably favours gathering additional information. However, screening the entire population using this method may raise some safety concerns. Furthermore, the method also requires a certain level of training and implementation of MCA Doppler as a population-based screening method would involve some challenges in relation to implementation.

Methods

Results

The literature search flow chart is presented in Appendix 3, Figure 32. We identified 14 studies74–87 that met our inclusion criteria, which involved a total of 109,679 patients. The study characteristics are presented in Appendix 3, Table 21. Two studies77,78 (n = 30,555) included unselected pregnancies, 10 studies74–76,80–85,87 (n = 61,047) included low-risk pregnancies only and two studies79,86 (n = 18,077) included mixed-risk pregnancies. Six studies75,78,79,81,82,84 (n = 5740) were prospective, six studies74,77,80,83,85,86 (n = 97,022) were retrospective, one study76 (n = 260) was cross-sectional and one study87 (n = 6657) was carried out as part of a clinical trial.

The assessment of study quality was performed using the QUADAS-2 tool and is summarised in Appendix 3, Figure 33. The main risk of bias was for reference standard because of the lack of blinding in the majority of studies. Only two studies81,84 (n = 1892) blinded the results to clinicians, one of which blinded only the AFI result and not the other aspects of the ultrasound. The second, more common, risk of bias was for flow and timing. Two studies75,85 performed ultrasound prior to IOL or within 4 days of delivery. Two other studies77,82 did not report gestational age at either ultrasound or delivery. Hence, these results may not be applicable for universal third-trimester screening at 36 weeks’ gestation. Two studies were rated as having unclear risk of selection bias79,86 as they did not report how the patients had been selected and one study76 was rated as having high applicability concerns for patient selection as it included prolonged (> 41 weeks’ gestation) pregnancies only.

The summary results for the diagnostic accuracy of oligohydramnios at predicting adverse pregnancy outcomes are presented in Table 4. The most commonly reported outcomes were neonatal unit admission and caesarean section for fetal distress (11 and 10 studies respectively). The stronger statistically significant association was with SGA < 10th centile, with a positive LR of 2.8 (see Table 4). There were also statistically significant associations with NICU admission and caesarean section for fetal distress, with positive LRs of 1.7 and 2.2 respectively. The positive LR for neonatal death was 3.7 but, because of the small number of events, the CIs were very large and include unity. The summary ROC curves are presented in Figure 5. Generally, the larger studies reported lower sensitivities and higher specificities for all outcomes. Figure 6 illustrates forest plots of DORs. Finally, we used Deeks’ funnel plot asymmetry test to assess the risk of publication bias using the outcome of neonatal unit admission for the analysis (see Appendix 3, Figure 34). The test showed no evidence of publication bias (p = 0.54).

FIGURE 5.

Summary ROC curves for AFI < 5 cm at predicting adverse pregnancy outcome. (a) NICU admission; (b) 5-minute Apgar score of < 7; (c) neonatal metabolic acidosis; (d) caesarean section for fetal distress; (e) SGA (< 10th centile); and (f) neonatal death.

FIGURE 6.

Meta-analysis of DORs for AFI < 5 cm at predicting adverse pregnancy outcome: (a) NICU admission; (b) 5-minute Apgar score of < 7; (c) neonatal metabolic acidosis; (d) caesarean section for fetal distress; (e) SGA (< 10th centile); and (f) neonatal death.

Discussion

This meta-analysis confirms that a diagnosis of severe oligohydramnios is associated with adverse pregnancy outcome. The key finding was that severe oligohydramnios had a positive LR for SGA of between 2.5 and 3.0. The associations with admission to NICU and emergency caesarean section for fetal distress are more difficult to interpret. First, for both of these outcomes, the association was weaker than it was for SGA. Second, in both cases the association could have been a consequence of the scan rather than an outcome predicted by the scan. Only two studies, containing < 5% of the patients included in the meta-analysis, blinded the results of the scan. Revealing the results of the scan could explain both associations. In the case of NICU admission, revealing the scan result could lead to a decision to deliver the infant as a result of suspected fetal distress. If this occurs preterm or at early term gestation it could lead to NICU admission as a result of iatrogenic prematurity. In the case of caesarean delivery for fetal distress, revealing the result that there is severe oligohydramnios could be used as an indication (in whole or in part) to perform a caesarean section for suspected fetal distress. Alternatively, if a caesarean section was performed for failure to progress, it is possible that the operator may include suspected fetal distress in the indication given the scan finding.

It is, however, also possible that the negative association with adverse neonatal outcome is due to treatment paradox. Given that the diagnosis was known in > 95% of cases in the meta-analysis, the attending clinicians may well have put interventions in place that prevented adverse outcome. These could include enhanced levels of fetal monitoring, IOL, or delivery by pre-labour caesarean section. A further complexity is that the aetiology of severe oligohydramnios may differ between studies, as some excluded women with ruptured fetal membranes, whereas others did not.

In conclusion, this analysis confirms that severe oligohydramnios is associated with adverse pregnancy outcome. This can confidently be stated, as there was an association with SGA, which is much less likely to arise from biases. However, the association between oligohydramnios and neonatal morbidity is less clear. Despite the association with SGA, the positive LR was not very high, and its capacity to act as a screening test in unselected nulliparous women at 36 weeks’ gestation is limited.

Methods

Results

Meta-analysis

The literature search flow chart is presented in Appendix 4, Figure 36. We identified 11 studies88–98 (including the POP study) that met our inclusion criteria, which involved a total of 37,848 patients. The study characteristics are presented in Appendix 4, Table 23. Only the POP study88 (n = 3387) included unselected pregnancies, three studies91,97,98 (n = 1890) included low-risk pregnancies only and seven studies89,90,92–96 (n = 32,571) included mixed-risk pregnancies. Two studies97 (n = 3817) were prospective and nine studies89–96,98 (n = 34,031) were retrospective. Seven studies91,93–97 (n = 36,293) defined borderline oligohydramnios as AFI of between 5 and 8 cm and four studies89,90,92,98 (n = 1555) defined it as between 5 and 10 cm. The majority of patients in all the studies delivered at term. However, four studies89,92,95,97 reported a significantly higher rate of preterm delivery among those with borderline oligohydramnios.

The assessment of study quality was performed using the QUADAS-2 tool and is summarised in Appendix 4, Figure 37. The main risk of bias was from the lack of blinding of the ultrasound result (which we defined as high risk for reference standard), which affected all studies except the POP study. 88 We classified one study93 as being at high risk for selection bias as it used only low-risk patients for the comparison group and we classified two studies89,90 as being at unclear risk of selection bias as they did not specify whether they enrolled a consecutive or random sample of patients. Moreover, we classified five studies89,92,94,96,98 as having an unclear risk of bias for flow and timing because they did not report gestational age at ultrasound or delivery.

The summary diagnostic performance of borderline AFI at predicting adverse pregnancy outcome is presented in Table 6. The most commonly reported outcomes were SGA < 10th centile (nine studies), NICU admission (eight studies), 5-minute Apgar score of < 7 (eight studies), meconium-stained amniotic fluid (seven studies) and caesarean section for fetal distress (six studies). The meta-analysis demonstrated a statistically significant association between borderline oligohydramnios and all of the outcomes, and the strongest association was with delivery of a SGA infant (positive LR = 2.6). The summary ROC curves are presented in Figure 7. Forest plots of the DORs (Figure 8) demonstrated statistically significant heterogeneity for SGA and NICU admission. Two studies (POP and Petrozella et al. 95) reported SGA below the third centile and three studies reported perinatal death. However, we could not generate summary results for outcomes that were reported in fewer than four studies. Finally we used Deeks’ funnel plot asymmetry test to assess the risk of publication bias using the outcome of SGA < 10th centile for the analysis (see Appendix 4, Figure 38). The test showed no evidence of publication bias (p = 0.33).

TABLE 6 - Summary diagnostic performance of borderline AFI in predicting adverse pregnancy outcome

Outcome	Number of studies	Number of patients	Summary sensitivity, % (95% CI)	Summary specificity, % (95% CI)	Positive LR (95% CI)	Negative LR (95% CI)
SGA < 10th centile	9	37,132	31.6 (13.0 to 58.7)	87.9 (71.9 to 95.3)	2.60 (1.83 to 3.69)	0.78 (0.61 to 0.99)
NICU admission	8	9747	34.8 (15.9 to 60.1)	82.6 (69.1 to 91.0)	2.00 (1.41 to 2.85)	0.79 (0.61 to 1.02)
5-minute Apgar score of < 7	8	9666	34.0 (17.4 to 55.8)	82.0 (68.8 to 90.4)	1.89 (1.47 to 2.42)	0.80 (0.66 to 0.98)
Caesarean section for fetal distress	6	33,517	21.2 (7.5 to 47.2)	90.0 (74.5 to 96.5)	2.13 (1.56 to 2.90)	0.87 (0.75 to 1.02)
Meconium-stained in amniotic fluid	7	2885	42.1 (28.7 to 56.9)	74.9 (67.7 to 81.0)	1.68 (1.24 to 2.28)	0.77 (0.62 to 0.96)

FIGURE 7.

Summary ROC curves of borderline AFI at predicting (a) SGA < 10th centile; (b) NICU admission; (c) 5-minute Apgar score of < 7; and (d) caesarean section for fetal distress.

FIGURE 8.

The diagnostic odd ratios of borderline AFI at predicting: (a) SGA < 10th centile; (b) NICU admission; (c) 5-minute Apgar score of < 7; and (d) caesarean section for fetal distress. a, Alexandros A Moraitis, Ilianna Armata, Ulla Sovio, Peter Brocklehurst, Alexander EP Heazell, Jim G Thornton, Stephen C Robson, Aris Papageorghiou and Gordon CS Smith, University of Cambridge, 2021.

Discussion

The main finding of the present study is that borderline oligohydramnios is moderately predictive of SGA babies. This was observed in the meta-analysis of multiple studies of variable quality. There was also a comparable association between borderline oligohydramnios and severe SGA in the only study in which researchers were blinded to the scan results, namely the POP study.

The observation that borderline oligohydramnios was associated with severe SGA only in the POP study is of interest. One possible explanation for this is that the scan result was not revealed; hence, the finding did not lead to changes in clinical management. The success from blinding the result is evidenced by the fact that borderline oligohydramnios was not associated with increased rates of IOL in the POP study. A previous RCT of routine early term induction compared with expectant management of pregnancies in which ultrasonic fetal biometry indicated a SGA infant demonstrated that early delivery was associated with a significantly decreased risk of the infant being delivered with a birthweight < 3rd percentile. 99 A possible explanation for the POP study’s association with severe SGA and the meta–analysis association with all SGA is that a finding of borderline oligohydramnios may have led to increased rates of early delivery in studies in which the result was revealed, whereas the lack of intervention in the POP study led to growth-restricted fetuses becoming progressively smaller for gestational age as the pregnancy advanced.

The other major difference between the meta-analysis and the POP study may also relate to the lack of blinding in the other studies. Borderline oligohydramnios was associated with increased rates of neonatal morbidity in the meta-analysis but none of the outcomes of neonatal morbidity was associated with this finding in the POP study. However, the CIs were wide and one explanation could be the lower statistical power of the POP study. However, plotting the DORs demonstrates that, in relation to NICU admission, the 95% CI observed in the POP study excluded the point estimate of the meta-analysis. This result could also be explained by the absence of blinding in the other studies. If the scan result is revealed, the only disease-modifying intervention available in late pregnancy is early delivery, and this could be late preterm or early term. It is well recognised that both are associated with increased rates of neonatal morbidity and NICU admission. Hence, the association between borderline oligohydramnios and neonatal morbidity in the meta-analysis could be because the finding led to iatrogenic prematurity and the absence of the finding in the POP study could be due to the lack of this effect. Assessment of individual studies in the meta-analysis is consistent with this interpretation. Gumus et al. 92 reported higher rates of IOL in women with borderline oligohydramnios, which was associated with higher rates of preterm and early term delivery, and higher rates of NICU admission. Similarly Asgharnia et al. ,89 who offered screening after 28 weeks’ gestation, found that those with borderline AFI had a rate of preterm delivery of 40.4% (compared with 14.9% for those with normal AFI) and this is the likely explanation for the strong association between borderline oligohydramnios and NICU admission. This association was not found in studies that offered ultrasound later in pregnancy such as that by Sahin et al. 97

In conclusion, we provide strong evidence that borderline oligohydramnios is associated with an increased risk of delivering a SGA infant. However, when the finding of borderline oligohydramnios is revealed to clinicians, it may lead to increased risks of neonatal morbidity as a result of earlier delivery. Given that the prediction of SGA was not strong and that revealing the result may have led to increased risks of neonatal morbidity, the observed association with SGA does not necessarily mean that screening unselected nulliparous women near term with this method will result in better clinical outcomes.

Methods

Results

The literature search flow chart is presented in Appendix 5, Figure 39. We identified 40 studies102–141 that met our inclusion criteria, which involved a total of 66,187 patients. The study characteristics are presented in Appendix 5, Table 24. Five studies105,114,120,123,138 (n = 8088) included unselected pregnancies, nine studies110,116,118,119,122,129,131,139,140 (n = 6436) included only low-risk pregnancies and 26 studies102–104,106–109,111–113,115,117,121,124–128,130,132–137,141 (n = 51,663) included mixed-risk pregnancies.

The assessment of study quality was performed using the QUADAS-2 tool and is summarised in Appendix 5, Figure 40. The main risk of bias was for reference standard because only two studies116,138 blinded the results to the clinicians. The second most common risk of bias was for flow and timing. This is because six studies106,111,123,125,133,142 had a very short interval between ultrasound and delivery (the ultrasound was carried out either prior to IOL or < 72 hours from delivery), two studies105,114 had a long interval (the ultrasound was carried out prior to 33 weeks’ gestation) and two studies104,107 did not specify the gestational age at delivery. Finally, three studies110,134,140 included prolonged (> 41 weeks’ gestation) pregnancies only, which were classified as having ‘high applicability concerns because of patient selection’. 37

The most commonly reported outcomes were birthweight of > 4000 g (29 studies103–106,110–113,118–123,125–135,138–141) followed by birthweight > 90th centile (seven studies102,107,109,114,115,124,138), both of which we classified as LGA. We defined severe LGA as a birthweight of > 4500 g (six studies113,117,131,137,138,141) or > 95th or 97th centiles (two studies114,138). Shoulder dystocia was reported in six studies. 108,112,116,136,138,141 Finally, neonatal morbidity (any related outcomes) was reported in only two studies,112,138 and consequently we could not produce summary results for this outcome. The most commonly used formulas for EFW were those described by Hadlock et al. ,5 followed by Shepard et al. 143 The most common thresholds for suspected LGA on scan were 4000 g (21 studies103,104,106,108,110,118,119,121,125,126,128–135,139–141) and 90th centile for the gestational age (nine studies). The abdominal circumference was used in nine studies,102,105,107,109,111,112,114,115,138 with the most common threshold applied being 36 cm (five studies122,123,125,127,137).

We present the summary diagnostic performance in Table 7. An estimated EFW of > 4000 g or > 90th centile had > 50% sensitivity for predicting LGA at birth and this was similar regardless of the formula used. The positive LR ranged between 7.5 and 12 for the Hadlock formulas5,6 and was about 5 for the Shepard formula. 143 The abdominal circumference (AC) had similar performance with the EFW. Suspected LGA also had about 70% sensitivity at predicting severe LGA at birth. Finally, an EFW of > 4000 g or 90th centile had 22% sensitivity at predicting shoulder dystocia with a statistically significant positive LR of 2.1.

The summary ROC curves for LGA and shoulder dystocia are presented in Figure 9. We also present the pooling of the DORs (Figure 10). Finally, we used Deeks’ funnel plot asymmetry test to assess the risk of publication bias using the outcome of LGA for the analysis (see Appendix 5, Figure 41). The test showed potentially significant risk of publication bias (p = 0.02).

FIGURE 9.

Summary ROC curves for the diagnostic performance of EFW > 4000 g (or 90th centile) at predicting (a) LGA at birth (birthweight > 4000 g or > 90th centile); and (b) shoulder dystocia.

FIGURE 10.

The diagnostic odds ratios for the diagnostic performance of EFW > 4000 g (or > 90th centile) at predicting (a) LGA at birth (birthweight > 4000 g or > 90th centile); and (b) shoulder dystocia.

Discussion

The key findings of the present study are that suspicion of fetal macrosomia on ultrasound scan is strongly predictive of the risk of delivering a large infant, but it is only weakly, albeit statistically significantly, predictive of the risk of shoulder dystocia. In the case of delivering a LGA infant as defined by the Hadlock formula, the positive LRs were quite strong, in the region of 7–12, whereas in relation to the diagnosis of shoulder dystocia the positive LR was ≈ 2. The forest plot of DORs indicates significant heterogeneity between the studies in their ability to predict a LGA infant. The source of this heterogeneity is unclear but it could relate to differences in the quality of the performance of the diagnostic test, such as the quality of the imaging equipment, the skill and training of the sonographers, and the characteristics of the population.

In this chapter and in Chapters 4 and 7 we have focused analysis on data from the POP study, as these data are particularly applicable to the research question addressed in this report, given that late-pregnancy ultrasound was performed in a large number of nulliparous women using contemporary equipment and staff trained using the standards of NHS England. The POP study analysis of a 36 weeks’ gestation scan in the diagnosis of macrosomia had previously been published138 and this was incorporated into the meta-analysis. Interestingly, the DOR from the POP study was 17.1 (95% CI 12.0 to 24.3) and this was virtually identical to the summary estimate from all of the other studies, which was also 17.1 but with a slightly narrower 95% CI (13.3 to 22.0). These data suggest that the results from the POP study are likely to be generalisable.

A recurrent theme in all chapters has been the lack of blinding in studies of the diagnostic effectiveness of ultrasound of pregnancy screening research. Hence, generally, the POP study has been unique as a contemporary study of late pregnancy in nulliparous women. However, in this analysis there is a second comparable study: the Genesis study. This was a prospective cohort study of 2772 nulliparous pregnant women recruited across seven centres in Ireland between 2012 and 2015. Women had the ultrasound scan between ≥ 39 weeks’ gestation and < 41 weeks’ gestation (i.e. ≈ 3–4 weeks later than in the POP study). Although the scan was carried out slightly later than stated in the research question of the current report, the design makes the study particularly useful.

The analysis of fetal macrosomia from the Genesis study has been published in abstract form only. It did not report the diagnostic effectiveness of EFW as a predictor of LGA birthweight, but it did report shoulder dystocia. Interestingly, the POP study and the Genesis study were the only two large studies (i.e. comprising > 1000 women) not to demonstrate a statistically significant association between macrosomic EFW and the risk of shoulder dystocia. Overall, the meta-analysis indicated that ultrasound may be weakly predictive of shoulder dystocia. However, as with other analyses in Chapters 4–7, these findings could be explained by ascertainment bias. Specifically, if a scan is performed and the fetus is suspected to be macrosomic, the clinical staff attending the birth may be more likely to institute manoeuvres for shoulder dystocia in the event of any delay, or to document a given delay as being due to shoulder dystocia. The potential for such biases may explain why the studies with blinded ultrasound were not significantly associated with shoulder dystocia and why the meta-analysis as a whole was only weakly predictive of shoulder dystocia, whereas it was strongly predictive for macrosomia. A weak association between ultrasonic EFW and the risk of shoulder dystocia is not surprising given that the actual birthweight of the infant is not strongly predictive of shoulder dystocia and that the majority of cases of shoulder dystocia do not involve a macrosomic infant. 144

Finally, the relationship between fetal macrosomia and pregnancy outcome is an area where there is good evidence that revealing the scan result changes the experience of complications of women who are false positives. Multiple studies have demonstrated that a false-positive diagnosis of fetal macrosomia is an independent risk factor for emergency caesarean delivery. 145–147 These observations underline the potential of screening low-risk women to cause harm and that designing a study where the results are revealed to the attending physician could lead to an association that is iatrogenic (because the knowledge of the result may change clinical decision-making) rather than because of a true prediction.

Management plan for breech presentation

This search identified an existing UK-based guideline from the Royal College of Obstetricians and Gynaecologists (RCOG), Management of Breech Presentation (Green-top Guideline No. 20a). 13 In brief, women who do not have a contraindication to ECV are offered this procedure (turning of the fetus by manual manipulation without anaesthetic). Where the procedure is contraindicated, declined or unsuccessful women would then have a discussion regarding attempting vaginal breech birth. Where vaginal breech birth was contraindicated or declined, a planned caesarean section would be scheduled at 39 weeks’ gestation (in the absence of a clinical indication for earlier delivery) with the proviso that the infant would be delivered by emergency caesarean section if the woman presented in labour before the scheduled date. Women who had a successful ECV would have routine care thereafter, but with midwife checks to ensure that the infant had not reverted to breech. In practice, given that the target population is nulliparous, it would be a small minority who would opt for vaginal breech birth and no women took up this option in the POP study. 11 For the purposes of the Markov chain modelling and health economic analysis we used the effect estimates of a Cochrane review that quantified ‘the effects of planned Caesarean section for singleton breech presentation at term on measures of pregnancy outcome’. 14 Other parameters were obtained from the literature and are detailed in Chapter 11.

Management plan for diagnosis of a small for gestational age fetus

We next used the NICE evidence search engine to identify existing guidelines for the management of a SGA fetus. This search identified an existing UK-based guideline from the RCOG, The Investigation and Management of the Small-for-Gestational-Age Fetus, Investigation and Management (Green-top Guideline No. 31). 152 Much of this guideline focuses on the identification of risk factors in early pregnancy and the management of the preterm SGA fetus. The RCOG recommendations are: (1) to take into consideration and abnormal umbillical artery or MCA Doppler to time delivery, (2) to offer delivery of the SGA fetus at 37 weeks’ gestation even if the umbilical artery Doppler is normal, (3) to recommend caesarean section in the SGA fetus with umbilical artery AREDV and (4) to offer IOL and continuous fetal heart monitoring in the SGA fetus with normal umbilical artery Doppler or with abnormal umbilical artery PI but end–diastolic velocities present.

The same search also identified an NHS England care bundle that aimed to reduce rates of perinatal death, Saving Babies’ Lives Version Two: A Care Bundle for Reducing Perinatal Mortality. 153 This guideline has a section on the management of SGA fetuses at term, and the following are key recommendations: (1) in the cases of severe SGA < 3rd centile and with no other concerning features, delivery should be offered at 37⁺⁰ weeks’ gestation and no later than 37⁺⁶ weeks’ gestation. (2) Fetuses between the 3rd and 10th centile should be assessed individually and the risk assessment should include Doppler investigations, the presence of any other high-risk features, for example, recurrent reduced fetal movements. In the absence of any high-risk features IOL should be offered at 39⁺⁰ weeks.

However, the context for both the RCOG and the NHS England guidelines was the management of women who were identified through the current approach of targeting ultrasound to high-risk women. As outlined in Chapter 9, we have not found evidence that these additional ultrasound tests are diagnostically effective when used as screening tests. Hence, the management protocol for SGA infants employed in the health economic analysis is to offer IOL. For the purposes of the health economic analysis we used the effect estimates of a Cochrane review that quantified ‘the effects of immediate delivery versus expectant management of the term suspected compromised infant on neonatal, maternal and long-term outcomes’. 150 In practice, 90% of the women included in the review came from a trial of IOL for suspected FGR. 99 IOL took place in the intervention group of this trial at an average of 38 weeks’ gestation and we have incorporated this into our management protocol (see section below). This does not represent an extreme intervention as a large-scale NIH-funded RCT demonstrated no adverse effect of routine IOL at 39 weeks’ gestation in nulliparous women who did not have risk factors. 154 Other parameters were obtained from the observational literature and are detailed in Chapter 11.

Management plan following diagnosis of a large for gestational age fetus

We next used the NICE evidence search engine to identify existing guidelines for the management of a LGA fetus. The only guidelines that we identified using this search related to women with diabetes. These women are routinely scanned during pregnancy and have specific issues, and the recommendations for this group are not generalisable to the population of interest in the current report. However, the search did identify a number of systematic reviews that addressed IOL, and one of these was a Cochrane review. 151 The Cochrane review concluded that IOL for suspected fetal macrosomia results in a lower mean birthweight, fewer birth fractures and shoulder dystocia. They concluded that to prevent one fracture it would be necessary to induce labour in 60 women and that induction of labour does not appear to alter the rate of caesarean delivery or instrumental delivery. However they suggested that further trials of induction shortly before term for suspected fetal macrosomia are needed. 151

Consistent with this recommendation, the HTA programme has funded a RCT [‘Induction of labour for predicted macrosomia: the Big Baby trial’ (ISRCTN18229892)]. Given the uncertainty in the evidence base, it is not possible to develop a robust plan for management following a diagnosis of macrosomia. For the purposes of the Markov chain modelling and health economic analysis, we addressed this uncertainty by comparing multiple strategies, including expectant management, early-term IOL and planned caesarean section. The effects in relation to IOL were taken from the Cochrane review,151 as this was assessed as the highest-quality evidence available at the time of writing. About 70% of the women came from a single trial101 in which the most common week for IOL was 38 weeks’ gestation. Other parameters for the modelling and health economic analysis were obtained from the observational literature and are detailed in Chapter 11. A summary of the management plan is outlined in Figure 11.

FIGURE 11.

Summary of the management plan following the 36 weeks’ gestation scan.

Introduction

This study was commissioned to evaluate the current evidence base on the costs and clinical effectiveness of performing a routine ultrasound scan in late pregnancy in all nulliparous women combined with appropriate management plans, to identify evidence gaps, and to predict whether or not future research to fill those gaps is likely to be a cost-effective use of health-care resources. In this analysis, we use decision modelling to assess the likely outcomes from universal ultrasound screening and determine whether or not its potential benefits can be clinically and economically justified.

We present a cost–utility analysis focusing on three of the main conditions detectable by ultrasound screening that may warrant intervention: breech presentation, the fetus being SGA and the fetus being LGA. The cost-effectiveness of universal ultrasound screening for each of these conditions individually has been explored previously. 11,155 However, here we evaluate the cost-effectiveness of screening for all of these conditions at the same session. Furthermore, we use decision uncertainty to predict the expected return on further research. We have applied the simplified management plan outlined in Figure 11. In essence, women are first assessed for presentation. If the infant is in breech presentation, ECV is offered. If this is successful, the woman reverts to receiving expectant management, and, if it is unsuccessful, the baby is delivered by planned caesarean section. If the infant is in a cephalic presentation and the EFW is in the normal range, the woman receives expectant management. If the infant is either SGA or LGA, IOL is offered. However, we also compare combined assessment for presentation and fetal biometry with a scan simply for presentation. The rationale for this is that a presentation scan may be readily implemented and relatively inexpensive, and there is much less uncertainty about the usefulness of knowing the infant’s presentation than there is about the usefulness of estimating the infant’s size.

The structure of this chapter is as follows. In Methods, we first introduce the general methodology for our economic evaluation. We then summarise the clinical definitions used, as well as the competing strategies evaluated, in this study before introducing the structure of the economic simulation model underlying the analysis. Once the model structure and mechanics have been explained, we discuss how we populated the model with the best available data; complete technical details regarding how individual parameters were derived are presented in Appendix 6. Finally, we describe the base-case analyses, sensitivity analyses and VOI analysis to guide how future research in this area could be prioritised.

In Results, we present the results of the baseline economic evaluation and sensitivity analyses. The results of the VOI analysis are then presented, which include the results for the expected value of perfect information (EVPI), the expected value of partial perfect information (EVPPI) and, finally, the expected value of sample information (EVSI).

In Discussion, we summarise the key findings, explain the interpretation of our results and discuss what impact our methodological limitations may have had on the results.

Methods

To compare long-term health and cost outcomes associated with different strategies of screening in third-trimester pregnancy, we constructed an economic simulation model. We focused the model on two features for which late-pregnancy ultrasound is amenable to detect: fetal presentation and fetal size. We used a decision tree model consisting of four subtrees, one each for breech presentation, LGA, SGA and AGA. The model structure is based largely on previous economic analyses of screening for these conditions individually, and the development and key characteristics of these submodels’ models have previously been described11,155 (a brief summary is provided in Appendix 7). Chapter 10 dealt with the diagnostic effectiveness of ultrasound in this setting and outlined how a positive result on scan could influence subsequent care. This chapter focuses on how these submodels were incorporated into a joint framework, enabling a cost-effectiveness analysis of simultaneous screening for all of these conditions.

Model structure

As stated, the model structure is a decision tree. It was coded in R (The R Foundation for Statistical Computing, Vienna, Austria) version 3.4.1, using the packages BCEA, FinCal, ggplot2, gtools, readxl, tidyr and SAVI. 158,159 The code for the model is available from the corresponding author on request.

Figure 12 shows the structure of the first stages of the decision model. The [+] indicates sub-branches that have been collapsed for clarity. Nodes are named to show their relationship to one another; nodes with the same letter have identical structures to the branches of the tree beyond, whereas a different number and/or a lower-case letter indicates a different set of probabilities. The prefixes B, L and S denote nodes with probability sets specific to breech presentation or large or small for gestational age infants, respectively.

FIGURE 12.

Model overview. [+], sub-branches of model collapsed for clarity.

At commencement, the scan policy can be set to selective (i.e. status quo), a universal scan for presentation only, or a universal scan for fetal biometry and presentation. The model structure is identical in each case. The difference is in the sensitivity and specificity of the scanning policies and their cost.

A fetus will be in either breech or cephalic presentation (node A1), or be LGA, SGA or AGA (node A2). For ease of modelling, we assume that all four possibilities are mutually exclusive and structured hierarchically, beginning with presentation (breech or cephalic) and followed by size (LGA, SGA or AGA). The implications of this are considered in Discussion. The probability of breech is the prevalence of breech at the time of screening (approximately 4.6%). 11 If the scan policy is universal ultrasound (whether for fetal biometry or for presentation only), then, given the ease of interpretation of such a scan, we assume all breeches are detected (i.e. 100% sensitivity and specificity, node B_B). However, under the selective scan policy, approximately 45% of breeches will be undetected11 owing to the mother not having undergone a scan at all (for consistency with the rest of the model, we label these ‘false negatives’). Further outcomes relating to breech presentation are described in Outcomes relating to breech.

If the infant is in cephalic presentation, it may be LGA, SGA or AGA. The probabilities of each is the prevalence of the condition (node A2, by definition 10% for each). If an infant is LGA or SGA, the probability of detection is a function of the sensitivity of the scanning policy (nodes L_B and S_B; LGA: 26.55% under selective and presentation-only scan, 37.85% under universal scan for fetal size;138 SGA: 19.6% under selective and presentation-only scan, 56.53% under universal scan for fetal size8). The sensitivity and specificity of ultrasound for detecting SGA and LGA were derived from the POP study. 8,138 The rationale for using the POP study values is that this study was conducted in NHS England, it involved nulliparous women being scanned at 36 weeks’ gestation, it is the only level 1 study of the diagnostic effectiveness of ultrasound to predict SGA and LGA (i.e. where the test result was blinded) and the values of sensitivity and specificity for SGA were similar to those in a 2019 Cochrane review of DTA. 23 In addition, the DOR from the POP study for macrosomia was identical to the DOR in the meta-analysis presented in Chapter 8.

If a LGA infant is correctly diagnosed as LGA, the pregnancy is managed in accordance with the defined LGA policy of either IOL or expectant management (node ‘MGT_LGA_TP’), in either case leading to either vaginal delivery or emergency caesarean section (nodes L_C3 and L_C2a; odds ratio of emergency caesarean section compared with otherwise healthy infant, 1.79146). If a LGA infant is misdiagnosed as AGA (i.e. false-negative scan), delivery can be either vaginal or by emergency caesarean section. Further outcomes relating to LGA babies are described in Outcomes relating to large for gestational age infants.

If the infant is SGA and is correctly diagnosed as such, labour is induced, leading to either vaginal delivery or emergency caesarean section (node S_C3). False negatives may lead to vaginal delivery or emergency caesarean section (node S_C2). Further outcomes relating to SGA pregnancies are described in Outcomes relating to small for gestational age infants.

An AGA infant may be misdiagnosed as SGA or LGA (false-positive SGA and LGA, respectively), or correctly diagnosed as AGA (node B). A false-positive SGA infant will be induced unnecessarily, leading to either vaginal delivery or emergency caesarean section (node S_C4). A false-positive LGA infant will be managed in accordance with the defined LGA policy namely either IOL or expectant management (node ‘MGT_LGA_FP’). IOL and expectant management can lead to either spontaneous vaginal or emergency caesarean section delivery (nodes L_C4 and L_C1 respectively). Finally, a correctly diagnosed AGA infant (true negative) can be delivered vaginally or by emergency caesarean section (node C1).

Short- and long-term outcomes

For all parts of the model, different levels of neonatal morbidity and mortality are possible, although these outcomes are structured slightly differently between the model’s subtrees. For the breech, SGA and AGA models, delivery outcomes include no, moderate and severe neonatal morbidity, as well as perinatal death. The risks of each level of adverse outcome differ between specific branches (i.e. are affected by the true status of the infant, the mode of delivery and whether or not labour was induced early). Long-term outcomes are then modelled as a function of the level of neonatal morbidity at delivery. For the LGA model, delivery and long-term outcomes are modelled differently. This is explained in detail in Outcomes relating to large for gestational age infants.

Long-term outcomes include ‘no long-term complications’, ‘SEN’, ‘severe neurological morbidity’ (SNM) and ‘neonatal/infant mortality’. The risk of long-term complications increases with the level of neonatal morbidity (nodes E1, E2 and E3). Unlike delivery outcomes, long-term outcomes are not affected by the actual status of the infant prior to delivery, only by the level of neonatal morbidity at delivery. Importantly, this means that all screening and management options affect long-term outcomes indirectly only as a result of the impact that they have on the outcomes at delivery.

Outcomes relating to breech

Figure 13 shows the decision tree with outcomes relevant to breech expanded and the remaining branches collapsed. The prevalence of breech refers to the fetal presentation at the time of screening. We assume that sensitivity and specificity for universal ultrasound is perfect at detecting fetal presentation, whether for size or breech presentation only. The sensitivity of selective ultrasound is lower because not all women receive ultrasound screening; however, we assume that all cases of suspected breech presentation would be either confirmed or rejected by ultrasound, so false-positive diagnosis is not an option (i.e. perfect specificity).

FIGURE 13.

Outcomes associated with breech. [+], collapsed sections of the decision tree.

On diagnosis of a breech presentation, an ECV is offered (node B_ECV). If the ECV is successful (node B_ECVs) and the infant remains cephalic (node B_ECVs_rb), no further intervention will be offered (i.e. expectant management). However, the infant may spontaneously revert to breech presentation (node B_ECVs_rb). In either case, there is a probability of emergency caesarean section, which is increased if the infant has reverted to breech presentation (nodes B_C3b and B_C3a respectively). If breech presentation is not diagnosed prior to labour, delivery options include breech vaginal delivery and emergency caesarean section (node B_C2).

Following labour and delivery there is a risk of no, moderate or severe neonatal complications or perinatal death (node D1), subsequently leading to no long-term complications, SEN, SNM or perinatal mortality (node E1). Note that we assume no raised risk of neonatal morbidity associated with cephalic emergency caesarean section compared with cephalic vaginal delivery per se. We do, however, allow for a raised risk of complications with an emergency caesarean section following breech presentation compared with a vaginal breech delivery (nodes B_D2a and B_D2c). If ECV is not accepted, or fails, then elective caesarean section may be offered.

Outcomes relating to large for gestational age infants

Figure 14 shows the decision tree with outcomes relevant to LGA expanded and remaining branches collapsed. When LGA is suspected, the intervention given will be in accordance with the predetermined management strategy (IOL or expectant management) for both true-positive and false-positive LGA diagnoses. The management option will affect the likelihood of the delivery outcome, as well as the mode of delivery, which can be either vaginal or by emergency caesarean section. When LGA is not suspected, delivery can be either vaginal or by emergency caesarean section.

FIGURE 14.

Outcomes associated with LGA. [+], collapsed sections of the decision tree.

Delivery outcomes include ‘no complications’, ‘respiratory morbidity’, ‘shoulder dystocia’, ‘other acidosis’ (i.e. acidosis not caused by shoulder dystocia) and ‘perinatal death’. The risk of each adverse outcome depends on the baseline risk, as well as on the mode of delivery, and whether or not labour was induced early.

Long-term outcomes depend on the outcome at delivery. For ‘no complications’, ‘respiratory morbidity’ and ‘other acidosis’, long-term outcomes included ‘no long-term complications’, ‘SEN’, ‘SNM’ and ‘neonatal/infant mortality’. For ‘no long-term complications’ the risk was equivalent to ‘no neonatal morbidity’ (node E1), and for ‘respiratory morbidity’ and ‘other acidosis’ the risk of long-term complications was equivalent to ‘severe neonatal morbidity’ (node E3). Shoulder dystocia (node L_E1) could result in no complications, brachial plexus injury (BPI) (node L_F1) or acidosis. BPI could be either transient or permanent (node L_G), the latter carrying the same risk of long-term outcomes as no neonatal morbidity (node E1) but with a penalty in terms of quality of life. Permanent BPI, SEN and SNM were long-term events; any other morbidity was expected to be resolved within the first year of life.

Outcomes relating to small for gestational age infants

Figure 15 shows the decision tree with the outcomes relevant to SGA expanded and the remaining branches collapsed. Labour will be induced early in suspected cases of SGA, whether based on a true or a false SGA diagnosis. Deliveries can be either vaginal or by emergency caesarean section. The probability of each mode of delivery is affected by whether or not labour was induced early. However, to avoid double counting the health effects of early labour induction, the mode of delivery affects only costs and not health outcomes.

FIGURE 15.

Outcomes associated with SGA. [+], collapsed sections of the decision tree.

Delivery outcomes include no, moderate and severe neonatal morbidity, as well as perinatal death. Women with correctly diagnosed SGA pregnancies (true positives) are offered early IOL, which reduces the risk of morbidity and mortality. When SGA is unsuspected (false negatives), pregnancies are managed expectantly, with no risk reduction. Note that early labour induction may also increase the risk of morbidity if initiated needlessly (i.e. in an AGA pregnancy falsely suspected of being SGA). However, in a true SGA pregnancy, early labour induction is expected to reduce the risk of morbidity. The scenario with a false-positive diagnosis is discussed further in Outcomes relating to appropriate for gestational age infants.

Long-term outcomes include ‘no long-term outcomes’, ‘SEN’, ‘SNM’ and ‘neonatal/infant mortality’. Each outcome is possible for all levels of neonatal morbidity. However, the risk of long-term complications increases for moderate and severe neonatal morbidity (nodes E2 and E3).

Outcomes relating to appropriate for gestational age infants

Figure 16 shows the decision tree with the outcomes relevant to AGA expanded and the remaining branches collapsed. An AGA fetus may be either correctly diagnosed or incorrectly diagnosed as either SGA or LGA (node B). If correctly diagnosed, the mode of delivery can be either vaginal or emergency caesarean section (node C1), after which short- and long-term outcomes will follow as described in Short- and long-term outcomes.

FIGURE 16.

Outcomes associated with AGA. [+], collapsed sections of the decision tree.

If an AGA fetus is falsely diagnosed as SGA, early IOL is offered. Unlike in the case of a true SGA, early labour induction of AGA pregnancies increases the risk of morbidity; however, the risk of perinatal death is still reduced. 160 Short- and long-term outcomes will then follow as described in Short- and long-term outcomes. If, instead, an AGA fetus is misdiagnosed as LGA, the short- and long-term outcomes depend on the management strategy. Compared with expectant management, early IOL decreases the risk of emergency caesarean section and perinatal death but increases the risk of neonatal morbidity.

Just as for other branches of the model, long-term outcomes include ‘no long-term outcomes’, ‘SEN’, ‘SNM’ and ‘neonatal mortality’. Each outcome is possible for all levels of neonatal morbidity; however, the risk of long-term complications increases for moderate and severe neonatal morbidity (nodes E2 and E3).

Results

Cost-effectiveness results

Table 14 shows the overall costs, QALYs, net benefit and incremental net benefit for each of the six screening management strategies. Net benefit is calculated assuming a WTP of £20,000 per QALY gained. INMB is shown relative to the status quo (assumed selective ultrasound scanning and IOL for both suspected SGA and LGA). Strategies are ordered in terms of increasing cost.

TABLE 14 - Cost-effectiveness results (per woman scanned)

INB, incremental net benefit relative to current practice (selective ultrasound and IOL); NB, net benefit; P_CE, probability of being the most cost-effective strategy.

Strategy with the highest expected net benefit.

Note

Management refers to management strategy when LGA is suspected; all babies that are of suspected SGA are assumed induced.

Screening and management	Cost (£), mean (95% credibility interval)	QALYs, mean (95% credibility interval)	NB | £20,000, mean (95% credibility interval)	INB | £20,000, mean (95% credibility interval)	P_CE | £20,000 (%)
Selective ultrasound and induction	6090 (4420 to 7890)	13.640 (13.441 to 13.841)	266,719 (262,333 to 271,079)	0 (0 to 0)	0.65
Selective ultrasound and expectant	6091 (4424 to 7889)	13.639 (13.439 to 13.839)	266,682 (262,297 to 271,040)	–37.09 (–124.7 to 35.24)	0.22
Universal ultrasound for presentation and inductiona	6101 (4443 to 7887)	13.645 (13.446 to 13.846)	266,806 (262,426 to 271,154)	87.36 (4.88 to 205.68)	44.19
Universal ultrasound for presentation and expectant	6102 (4446 to 7887)	13.644 (13.444 to 13.844)	266,769 (262,389 to 271,120)	50.29 (–68.06 to 186.43)	15.63
Universal ultrasound for size and expectant	6178 (4508 to 7972)	13.646 (13.446 to 13.846)	266,734 (262,351 to 271,099)	14.47 (–133.98 to 173.31)	0.51
Universal ultrasound and induction	6180 (4498 to 7983)	13.648 (13.448 to 13.849)	266,779 (262,386 to 271,147)	60.24 (–151.43 to 281.7)	38.81

Given current evidence, and assuming a WTP of £20,000 per QALY, the strategy associated with the highest net benefit is a presentation-only scan for all women (where women with relevant indications also get a full scan). When LGA is suspected, the recommended management is IOL; on average, IOL is associated with a small improvement in QALYs compared with expectant management (SGA is assumed managed with IOL). Universal ultrasound screening for fetal size is not supported by this analysis as its added benefits do not justify its added cost. Decision uncertainty suggests that there is a 44.19% probability that this is the most cost-effective strategy (Table 14 and Figure 17).

FIGURE 17.

Cost-effectiveness acceptability curve for the chance that each strategy will be the most cost-effective as a function of WTP for an additional QALY. Mexp, expectant management; MIOL, IOL; Sbre, universal ultrasound for fetal presentation only; Ssel, selective ultrasound; Suni, universal ultrasound for fetal biometry plus presentation.

One-way and scenario analyses

Cost-effectiveness conclusions were sensitive only to the time horizon, the cost of an ultrasound scan for fetal presentation only, the background risk of stillbirth, moderate and severe perinatal complications, and the risk of SEN associated with IOL. 189

With respect to the time horizon, universal ultrasound for fetal presentation is the most cost-effective option only as long as the time horizon of the analysis is < 45 years (Figure 18). Beyond this time horizon, universal ultrasound for size and presentation becomes the most cost-effective option. With respect to the cost of a presentation scan, a presentation-only scan remains the most cost-effective option, provided that this costs no more than £90. Above this cost, status quo is the most cost-effective (Figure 19).

FIGURE 18.

One-way sensitivity analysis of model time horizon. MExp, expectant management; MIOL, IOL; Sbre, universal ultrasound for fetal presentation only; Ssel, selective ultrasound; Suni, universal ultrasound for fetal biometry plus presentation.

FIGURE 19.

One-way sensitivity analysis of the cost of a scan for fetal presentation only. MExp, expectant management; MIOL, IOL; Sbre, universal ultrasound for fetal presentation only; Ssel, selective ultrasound; Suni, universal ultrasound for fetal biometry plus presentation.

As the background risks of perinatal mortality, moderate and severe perinatal complications rise, the net benefit of a detailed universal scan rises (Figure 20). This is because the risks of complications from SGA and LGA infants are modelled relative to the baseline risks; as the baseline risk rises, the risks for SGA and LGA infants rises more than proportionately, thus the benefit from detection and intervention rises. A breech-only scan remains the most cost-effective option so long as the baseline risk of perinatal death remains < 0.28% and the risk of moderate and severe complications is < 4.8% and < 1.12%, respectively. Above these values, universal screening becomes the cost-effective option.

FIGURE 20.

One-way sensitivity analysis of baseline risk of (a) perinatal mortality; (b) severe morbidity; and (c) moderate morbidity. MExp, expectant management; MIOL, IOL; Sbre, universal ultrasound for fetal presentation only; Ssel, selective ultrasound; Suni, universal ultrasound for fetal biometry plus presentation.

Our base-case analysis assumed a linear progression through the model whereby long-term outcomes were dependent on perinatal outcomes, which were dependent on mode of delivery alone [vaginal vs. caesarean section (emergency or elective)]. However, there is evidence to suggest that IOL may increase the risk of SEN in later life. 172 We therefore explored the impact on the results via a one-way sensitivity analysis. We found that our results remained the same as long as the relative risk of SEN as a result of IOL is between approximately 0.95 and 1.3 and the estimated risk at 38 weeks’ gestation was within this range. 172 Below this risk, the most cost-effective strategy is to perform universal screening for both presentation and EFW, and to induce labour when SGA or LGA is suspected. Above this risk, then while the recommended scan remains a presentation-only scan, the most cost-effective intervention for suspected SGA or LGA is expectant management (i.e. IOL ceases to be the appropriate intervention; Figure 21). Given this, although not captured in our formal VOI analysis (because of structural assumptions), it may be worthwhile exploring the impact that inducing labour has on long-term risk of SEN in future research.

FIGURE 21.

One-way sensitivity analysis on relative risk of SEN from IOL. MExp, expectant management; MIOL, IOL; Sbre, universal ultrasound for fetal presentation only; Ssel, selective ultrasound; Suni, universal ultrasound for fetal biometry plus presentation.

Figure 18 shows the expected INMB for different strategies compared with current practice (selective ultrasound with IOL for suspected LGA) as a function of the model’s time horizon (years). Calculations are based on a WTP (i.e. valuation of one additional QALY) of £20,000.

Figure 19 shows the expected INMB for different strategies compared with current practice (selective ultrasound with IOL for suspected LGA) as a function of the cost of an ultrasound for fetal presentation only. Calculations are based on a WTP (i.e. valuation of one additional QALY) of £20,000.

Figure 20 shows the expected INMB for different strategies compared with current practice (selective ultrasound with IOL for suspected LGA) as a function of the baseline risk of perinatal mortality (see Figure 20a), severe neonatal morbidity (see Figure 20b) and moderate neonatal morbidity (see Figure 20c). Calculations are based on a WTP (i.e. valuation of one additional QALY) of £20,000.

Figure 21 shows the expected INMB for different strategies compared with current practice (selective ultrasound with IOL for suspected LGA) as a function of the relative risk of SEN if labour is induced early (compared with expectant management). Calculations are based on a WTP (i.e. valuation of one additional QALY) of £20,000.

Value-of-information analysis

Expected value of perfect information

At a WTP of £20,000 per QALY, the per-patient EVPI is £31.56. Given a beneficiary population of 1,689,663, the population EVPI to England is £53.3M. If the results of the analysis are assumed generalisable to all pregnancies in England, then the population EVPI is £172.9M. Figure 22 shows the per-patient EVPI as a function of the WTP threshold. The two local peaks indicate where the decision (i.e. which screening strategy is preferred) changes, and, thus, the impact of decision uncertainty is greatest around these thresholds.

FIGURE 22.

Per-patient EVPI as a function of the WTP for an additional QALY. EVPI is presented per person. WTP refers to monetary valuation of an additional QALY (£).

Expected value of perfect parameter information and expected value of sample information

Table 15 shows the parameters with an EVPPI exceeding £100,000 under the broader assumption that any future study will be of value to all births in England, not just low-risk singleton pregnancies. The most valuable parameter is difference in cost of delivery from IOL, accounting for 84% of the EVPI. Except for this cost, no other parameters individually account for > 1% of the total EVPI. The other parameters with the greatest contribution to EVSI are the relative risk (LGA vs. AGA) of acidosis from a vaginal delivery following IOL, the odds ratio of perinatal death (LGA vs. AGA) from an infant delivered vaginally without IOL, the relative risk (SGA vs. AGA) of emergency caesarean section following IOL and the odds ratio (SGA vs. AGA) of severe neonatal morbidity under expectant management.

TABLE 15 - The expected value of partial perfect information for individual parameters and groups of parameters

OR, odds ratio; RR, relative risk.

Assuming study results are applicable to all births in England.

Parameter	Per-patient EVPPI (£)	Standard error	Percentage of EVPI	pEVPPI (£)	pEVPPI (£)a
Cost of delivery from IOL	26.51	0.07	84	44,790,000	145,200,000
RR for acidosis in macrosomic fetuses if induced early	0.27	0.04	1	456,000	1,478,000
OR for mortality if fetus is macrosomic	0.26	0.03	1	438,900	1,423,000
Group	0.72	0.07	2	1,215,199	3,939,513
RR for emergency caesarean section among SGA fetuses following early labour induction	0.06	0.01	0	99,290	321,900
OR for severe neonatal morbidity if fetus is SGA	0.03	0.01	0	48,740	158,000
Group	0.26	0.04	1	443,104	1,436,484

These five parameters could naturally be collected from three separate studies:

1. a costing study of the difference in cost of delivery associated with IOL compared with expectant management

2. a RCT of delivery outcomes relating to LGA babies

3. a RCT of delivery outcomes relating to SGA infants.

The EVPPI of the costing study is either £44.8M or £145.2M, depending on whether the results are considered applicable to singleton nulliparous pregnancies only or to all pregnant mothers, respectively. The two RCTs have EVPPIs of up to £3.9M and £1.4M under the broader applicability criteria.

The EVSI of the costing study suggests that there is scope for the study to yield a positive return on investment. For example, a two-arm study with 1000 patients in each arm has an EVSI to England of £11.3M (or £97.2M if this information is of value to all pregnancies in England, not just to low-risk nulliparous singleton pregnancies). If such a study was to cost £1M, then it would yield a net return on investment of at least £10.3M (Figure 23).

FIGURE 23.

Population EVSI for a study on the cost of IOL.

We were not able to calculate non-zero EVSI estimates for studies on macrosomia or SGA outcomes as the per-patient EVPPI is too low.

Expected value of perfect parameter information under alternative scenarios

The EVPPI provides the value of obtaining perfect information for a parameter based on the magnitude at which perfect information would affect the decision outcome. This means that even parameters that have a great impact on overall cost and QALYs, and for which the value is highly uncertain, may have low EVPPI if perfect information would not change the decision (i.e. which screening strategy is most cost-effective). However, whether or not the exact value of a parameter affects the decision outcome is highly dependent on context. Through simulating alternative scenarios, we analysed how the EVPPI of key parameters was affected by model assumptions.

Given the uncertainty about the setting in which an ultrasound scan for fetal presentation only could be provided, there were some concerns that the cost was not correctly specified in the base-case scenario. We therefore simulated three alternative scenarios where we varied the assumptions underlying the cost calculations: (1) fetal presentation could be assessed through directly accessed diagnostic services (£52, 95% CI £24 to £91), (2) an antenatal standard routine ultrasound scan was required (£108, 95% CI £97 to £118) and (3) costs could range between those of either of these scenarios (£24–118). The results showed that EVPPI was highest where the cost was highest. In this scenario, the EVPPI was £6.07 per person. Depending on the beneficial population, the overall EVPPI was £10.3M (nulliparous women only) or £33.3M (all women). It is worth noting that the model’s assessment of the value of further studies is, in this case, at odds with cost-effectiveness. A higher cost for scanning means a lower chance that ultrasound for fetal presentation will be cost-effective, but the value of researching this parameter further increases.

The cost of IOL (specifically, the net difference in total cost between pregnancies that were induced early and that of expectant management) had the highest EVPPI in our base-case scenario, and hence the greatest expected benefit from future research. In the base-case scenario, the cost was £125 (95% CI –£1343 to £1594); more details are presented in Appendix 6. To test how sensitive the EVPPI was to the exact input values used, we simulated two alternative scenarios: (1) where the standard error of the mean was reduced by 50% and (2) where costs were instead obtained from the 35/39 trial,194 where the cost difference was –£236 (95% CI –£646 to £174);194 see Appendix 6 for details. When the standard error was reduced by 50%, the EVPPI fell by ≈ 80%. When costs were obtained from the 35/39 trial, the EVPPI was £6.3M for the beneficial population (i.e. nulliparous women).

Discussion

Conclusions

The remit of this work was to advise the National Institute for Health Research on the current body of evidence regarding the cost-effectiveness of late-pregnancy ultrasound screening and specifically whether or not there is value in commissioning further research in the area and, if so, what this research should focus on.

Our results suggest that universal ultrasound for fetal presentation only may be both clinically and economically justified, but implementation research is needed before it is adopted into routine care. Specifically, this must explore whether or not a scan can be conducted by a midwife during a routine antenatal visit. Universal ultrasound including estimation of fetal weight is of borderline cost-effectiveness and is sensitive to certain assumptions. Our formal VOI analysis suggests that future research should be focused on the net cost of IOL compared with expectant management.

Aims

The aims of this section were to:

1. assess pregnant women’s knowledge about the current antenatal care pathway for low-risk pregnancies

2. assess pregnant women’s understanding of the potential benefits and drawbacks of third-trimester screening

3. estimate pregnant women’s willingness to participate in a future randomised clinical trial, examine which trial design they would prefer to participate in, and calculate the expected recruitment rate.

Methods

To evaluate both the quantitative and the qualitative aspects of the above aims, we conducted a survey and ran focus groups. For each aim we collaborated with the National Institute for Health Research Cambridge Biomedical Research Centre Communications and patient and public involvement (PPI) department of Cambridge University Hospitals NHS Foundation Trust (CUHFT). Amanda Stranks, the head of the PPI department of CUHFT, had an active role in the writing and testing of the survey as well as the design, recruitment and running of the focus groups, as explained below.

The objective of the survey was to meet the requirements of aims 1 and 3 by involving a large and representative number of women. We planned to recruit low-risk nulliparous women after their ultrasound scan at 12 or 20 weeks’ gestation, given that the scans at these points confirm a viable pregnancy. We excluded any high-risk pregnancies with either maternal or fetal pathology. The questionnaire was approved by all of the collaborators of the study and tested by the PPI office in CUHFT to ensure that it was understood by the women. We received feedback from five anonymous individuals and modified our form accordingly. We include the final version of the questionnaire in Appendix 8. In brief, this questionnaire had three parts. The first two questions were about the woman’s knowledge of current antenatal care and her willingness to have an additional ultrasound scan in the third trimester. The second part included three questions about potential participation in a future randomised controlled trial. We discussed two possible trial designs. The first study (study A) would randomise low-risk women to have a scan at 36 weeks’ gestation or not (the latter being current standard of care). The ultrasound results would be revealed to their clinical care team and their management would be affected accordingly. In the second study (study B) all women would have an ultrasound at 36 weeks’ gestation. If there was a major problem (e.g. breech presentation or very small amount of fluid around the infant), the result would be revealed to the care team. In all other cases the result would be blinded to the women and the clinicians. Finally, we included some questions on women’s demographics, such as age, ethnicity and education, to ensure that the sample was diverse. All of the replies were anonymised.

The second part of this section was running groups in which we could discuss the qualitative aspects of all the above aims. We planned to recruit women who had recently delivered (within the last 2 years), and discuss in detail the benefits and potential risks of third-trimester screening. To advertise the focus groups, we used the mailing list of the PPI office, personal contact by midwives, and social media including Facebook (www.facebook.com; Facebook, Inc., Menlo Park, CA, USA), Twitter (www.twitter.com; Twitter, Inc., San Francisco, CA, USA) and WhatsApp (Facebook, Inc., Menlo Park, CA, USA) to address groups of mothers in the broader area of Cambridge. The focus group discussion was designed by Alexandros A Moraitis, Gordon CS Smith and Amanda Stranks.

Results

Discussion and conclusions

We were able to collect both quantitative and qualitative data about the opinions of women on third-trimester ultrasound screening. We found that there was a clear interest in having an additional ultrasound scan in the third trimester, which was also confirmed in the focus group by all but one participant. This also confirms the previously published finding by the Stillbirth Priority Setting Partnership,197 which included responses from > 300 parents and 700 professionals and concluded that the question of whether or not a third-trimester ultrasound scan can reduce the risk of stillbirth was one of the most important research priorities. We also found that the majority of women would be happy to participate in a future randomised controlled trial and we would expect a recruitment rate of at least two out of three women, which is similar to the recruitment rate of the POP study in which the ultrasound result was blinded to the women and the clinicians. In total, 66% of women who replied to our questionnaire, and all of the focus group participants, would be happy with the blinding of the ultrasound result if there was no severe problem, something that we would have to define clearly.

Reflections/clinical perspective

We managed to acquire a large number of replies (as planned) to a questionnaire that gave us an overall view of women’s opinions about and willingness to participate in a future trial. However, we found it difficult to recruit women to the focus groups. Prior to recruitment, after discussion with the collaborators and the PPI office in CUHFT, we made the decision not to include pregnant women in the focus groups as the discussion could cause them anxiety about their care. However, it was also difficult to recruit new mothers and they could not easily find the time to participate. We managed to recruit four women by arranging child care and transport (in one case). The input from those in the focus group was valuable because we had the opportunity to listen to women who were keen to have an additional scan and a woman who was sceptical about the need for those additional scans. We also gained valuable information about what to include in a future consent form and the timing of this additional form. Overall, we believe that all of the above information would affect the design and conduct of a future clinical trial.

Implications of the health economic analysis

The economic analysis demonstrated that although, on average, the most cost-effective approach was to screen all nulliparous pregnant women with a presentation-only scan, this had only a 44% probability of being true, and a scan that included fetal biometry had a ≈ 39% chance of being the most cost-effective. Moreover, if the time scale was increased, it became likely that such a scan in late pregnancy would be the most cost-effective approach. These observations indicate that implementing such a scan could be considered. However, one of the major obstacles to implementing such a policy is that there is no direct evidence from a RCT that this screening and intervention is clinically effective. The Cochrane review of universal late-pregnancy ultrasound failed to show any benefit of this to the mother or infant. 21 However, as discussed in the introduction, this review has a number of methodological issues and it is more accurate to state that it does not provide a definite answer the question of whether or not universal late-pregnancy ultrasound reduces the risk of perinatal death.

Interestingly, the VOI analysis highlighted reducing uncertainty about the costs of IOL. Given the above, this may be regarded as somewhat counterintuitive. However, the parameters used in the VOI analysis in relation to the screening performance of ultrasound and the effect of intervention were known with a degree of precision that meant that reducing their uncertainty was not the most cost-effective research question. For example, the ability of ultrasound to predict SGA, the relationship between SGA birthweight and the risk of stillbirth, and the ability of IOL to reduce the risk of stillbirth are all known quite precisely and are based on high-quality data. Consequently, even though there is no direct evidence to indicate that universal late-pregnancy ultrasound would reduce the risk of stillbirth, the model estimates quite a high chance that it is the most cost-effective approach and does not highlight reducing the uncertainty in these parameters in the VOI analysis. By contrast, previous health economic analyses of IOL have generated quite wide CIs,176,194 and hence the model has identified that reducing this uncertainty is the key question.

Case for considering a randomised controlled trial of screening and intervention

In this chapter we consider the practicalities of designing a RCT of screening and intervention using fetal biometry in nulliparous women at 36 weeks’ gestation. We have done this because, even though the parameters in the modelling were reasonably certain, these parameters were calculated from a range of different study designs (i.e. we did not perform the VOI analysis based on the uncertainty of parameters calculated from a large RCT of late pregnancy screening and intervention in nulliparous women). Rather, we performed the analysis using parameters from a range of observational studies and a range of studies of interventions in women who were deemed to be high risk for other reasons. The concern in this case is external validity. The parameters may be reasonably certain in relation to the setting where they were derived but there is an unquantifiable uncertainty in relation to how well they inform our research question. The obvious way to address this would be to perform a study in the setting of interest. Such a study could be the definitive study or it could be a pilot or a proof-of-principle study. The former might be a trial of screening compared with not screening, with perinatal death as the primary outcome. The latter might exploit alternative study designs and use of proxies. Hence, there are a number of important considerations to take into account when designing a RCT of screening and intervention using universal ultrasound, and we will consider each of these in turn.

Candidate primary outcomes

In relation to the primary outcome of a RCT, we believe that the strongest case can be made for perinatal death. First, losing an infant at term is clearly a devastating outcome for a family. In the absence of a lethal anomaly, preventing death would lead to an entire life gained which, from a health-care and health economic perspective, is a gain of unique magnitude. Second, the main intervention available is earlier delivery. There is strong evidence that IOL is effective in reducing the risk of perinatal death. Over two-thirds of perinatal deaths at term are antepartum stillbirths54 (i.e. intrauterine fetal death prior to the onset of labour). Self-evidently, antepartum stillbirth cannot occur after an infant has been delivered. 17 Delivery at or after 38–39 weeks’ gestation carries the same risk of intrapartum stillbirth and neonatal death as delivery at a later week of gestation. 17,198 These epidemiological observations underlie the 67% reduction in the risk of perinatal death associated with IOL at term. 16

Proxies

The main problem with a primary outcome of perinatal death is that the outcome is uncommon, and this will result in major issues of statistical power. Indicators of perinatal morbidity would be an alternative outcome to perinatal death. First, as the same factors might be involved in death and morbidity, the latter could be used as proxies of the former. Second, perinatal morbidity is of importance in its own right. For example, birth asphyxia is one of the major determinants of the burden of litigation in the health service as a result of devastating effects on the later health of the child, such as CP. There is evidence to support the use of a single indicator in both roles. An Apgar score of < 4 at 5 minutes was associated with a relative risk of early neonatal death of ≈ 360174 and a relative risk of CP of > 400. 173 Hence, a primary outcome based on perinatal morbidity, such as an Apgar score of < 4, could be clinically important, both as a proxy of death and as a determinant of long-term outcome. Morbidity could be a more pragmatic outcome as rates of severe morbidity are much greater than the risks of death, and hence it may be easier to design a trial with morbidity as the primary outcome.

Subgroups

A further refinement to the primary outcome is to study subgroups of the given event that were actually associated with the infant being born SGA or LGA. It is self-evident that screening for SGA or LGA will primarily have an impact on outcomes related to fetal growth disorder. Many adverse perinatal outcomes, both lethal and non-lethal, are unrelated to fetal growth abnormalities. Consequently, if a screening study of fetal biometry has a primary outcome that includes infants in the full range of birthweight, most of the primary outcomes in both arms of the trial will be unrelated to fetal growth disorder, which is not preventable by screening for fetal growth disorder and intervention. This means that the potential for screening to have an impact on the rate of death is limited and extremely large sample sizes would be required. For example, around one-third of perinatal deaths at term are related to being SGA or LGA. 54 The background rate of perinatal death at term is ≈ 2 per 1000. Even if a screening test was perfect (i.e. detected all cases of growth disorder), and even if the intervention was perfect (i.e. prevented all such deaths), a power calculation still indicates that > 100,000 women would have to be recruited to the trial. However, if the primary outcome was perinatal death of a SGA or LGA infant, the sample size would be ≈ 22,000 (note that this is used to illustrate the point that it is not a practical proposition, as the screening and intervention characteristics were assumed to be perfect). An analogy might be a trial of breast cancer screening. Screening reduces deaths related to breast cancer but does not reduce all-cause mortality. 199 This is likely to be explained by the fact that no study could be sufficiently powered to detect an effect of screening for breast cancer on all-cause mortality because most deaths are due to other causes. Consequently, one approach to addressing the problems of statistical power in trials of screening using fetal biometry would be to define primary outcomes related to fetal growth abnormalities. An insistance on evidence that shows a reduction in all-cause perinatal death would simply remove the possibility of screening and intervention being implemented, which could lead to avoidable harm that could have been prevented in a cost-effective way.

Early delivery and iatrogenic harm

Routine induction at term had less dramatic effects on the risk of neonatal morbidity, with a 12% reduction in the risk of NICU admission and a 30% reduction in the risk of a low Apgar score. Moreover, these effects may be lost or even reversed in the context of early-term IOL. Most trials in the Cochrane review of term induction were of pregnancies at 41 weeks’ gestation and beyond. 16 As post-term pregnancy is associated with an increased risk of neonatal morbidity, preventing this outcome should improve immediate neonatal outcomes as well as preventing stillbirth. In the context of IOL at < 39 weeks’ gestation, epidemiological data indicate that the intervention may actually increase neonatal morbidity. 160 The potential for earlier intervention to cause harm is increasingly recognised. The Awareness of fetal movements and care package to reduce fetal mortality (AFFIRM) study200 reported a stepped-wedge RCT of a programme to inform women about reduced fetal movements and to standardise intervention. Although it did not show a significant reduction in stillbirth, the intervention was associated with increased risks of neonatal morbidity. 200 This trial has some parallels with the current question. Despite the fact that women were selected on the basis of having a risk factor (i.e. reduced fetal movements, which is associated with stillbirth), it still failed to demonstrate a reduction in stillbirth rates, and the intervention was associated with increased rates of intervention and adverse outcomes. The result of the trial underlines two key issues: (1) the need for better predictors of adverse outcome and (2) the potential for intervention to cause harm.

Current status of screening tests

Unfortunately, the results of our systematic reviews of diagnostic effectiveness and a Cochrane DTA review23 failed to identify any ultrasonic marker that was clearly predictive of the risk of stillbirth in the context of scanning women in late pregnancy using ultrasound. Moreover, if we regard neonatal morbidity as a proxy of stillbirth, again, tests performed very poorly. Finally, actual birthweight in the < 3rd percentile was associated with a 0.9–1% risk of perinatal death at term compared with a background risk of just over 0.2%. 54 Hence, even knowing that the actual birthweight was < 3rd percentile would be associated with a positive LR of between 4 and 5. In the POP study, of 562 women whose scan indicates that their infant was SGA, only 12% of women delivered an infant with a birthweight in the < 3rd percentile; a further 23% delivered an infant ≥ 3rd and < 10th percentile but about two-thirds of the women delivered an infant ≥ 10th percentile. Hence, on the basis of the association between the EFW and the actual birthweight, and their relationship between the actual birthweight and the risk of stillbirth, it is highly unlikely that detecting a SGA infant is strongly predictive of the risk of stillbirth. Given the lack of information, we model outcomes with variable incidence and assess different screening test values to establish what characteristics would be required of a test to make a trial of screening and intervention feasible.

Possible trial designs

Broadly speaking, there are two main approaches to trial design (Figure 24). 32 First (hereinafter referred to as screen vs. no screen), women might be randomised (1) to be screened, with the offer of intervention if they screen positive, or (2) to receive routine care, which currently requires scanning only if there is a conventional clinical indication. The result of this trial design is a simple comparison between the two groups. In the event of a negative result, it is impossible to determine whether the result was because the screening test did not work or because the intervention did not mitigate the higher risks in screen-positive women. The second approach is to screen the whole population and randomise high-risk women to an intervention or to routine care (masking the result in the latter group), hereafter referred to as ‘screen all’. The advantages of the second approach are that the number of women who need to be recruited is substantially fewer and that the same trial can assess both the diagnostic effectiveness of the screening test and the clinical effectiveness of the intervention. The two approaches are illustrated in Figure 24.

FIGURE 24.

Flow charts of possible trial designs: (a) screen vs. no screen; and (b) screen all.

Acceptability of the ‘screen-all’ approach

When discussing the possibility of randomising women with a high-risk screening result, some of the co-applicants expressed concerns. Interestingly, however, when we surveyed pregnant women, they actually preferred a study design that involved all participants being scanned. In the focus group, women tended to be more concerned about being offered interventions. The observations underline the different perspectives of pregnant women and professionals. We envisaged that women who are recruited to a ‘screen all’ approach would have some information revealed irrespective of their randomisation status. For example, we do not feel that it would be practical or ethical not to reveal the presentation of the infant as cephalic or non-cephalic. Hence, this would probably be revealed in a ‘screen all’ trial design. In the POP study, although scans were blinded, breech presentation was revealed. Subsequent interviews with participants were highly positive about this element of the study where the infant was breech [Dacey 2015; www.repository.cam.ac.uk/handle/1810/280595 (accessed June 2019)]. However, a drawback of this approach is that a ‘screen all’ design, which reveals breech presentation, would not capture the health benefits of detecting breech presentation. Other features that should be considered in revealing the result are the presence of previously undiagnosed major congenital anomalies and placenta praevia. In the POP study, there was no cases of placenta praevia, but two patients had major anomalies diagnosed where revealing the result optimised care and, in one case (unilateral hydrothorax with severe mediastinal shift), is likely to have prevented intrauterine fetal demise.

Power calculations

To determine the feasibility of a RCT we performed power calculations using the two different study designs represented above. The sample size calculations are presented in Table 17. All power calculations have been performed for a p-value < 0.05 (two-sided) with 90% power to detect the effect. We selected a range of possible primary outcomes: perinatal death, severe neonatal morbidity, any neonatal morbidity and delivery of a SGA infant with complications. In relation to perinatal death, we found no adequately powered studies of the diagnostic effectiveness of ultrasound to predict this outcome and the Cochrane DTA review23 of SGA also found no data in relation to this question. Therefore, we modelled a series of possible screening performances, varying the screen-positive rate and positive LR. In relation to morbidity, we used two studies reporting data from the POP study, from The Lancet8 and The Lancet Child & Adolescent Health. 149 As described above, the POP study was one of only two studies (Perinatal Ireland Genesis study being the other) that performed blinded ultrasound scanning in late gestation in nulliparous women. Unfortunately, the Genesis study did not report the association between SGA and morbidity, and the only publication in relation to LGA is in abstract form only and addresses shoulder dystocia. The two POP study publications8,149 address the relationship between SGA, SGA combined with reduced growth velocity (which was the best-performing predictor of morbidity from a range of candidate predictors of FGR) and the Delphi consensus definition of late FGR.

In all of these calculations we assumed that the intervention would reduce the risk of the given event by 50%. Given the lack of data, a range of figures could be considered. We used this figure as we felt that it was conservative in relation to perinatal death. It could be argued, based on the discussion above, that it is optimistic in relation to neonatal morbidity. However, by concentrating the outcome of morbidity on infants that are actually SGA, it is plausible that the combined effect of making the diagnosis and intervening could substantially reduce the rate of adverse events. It should be borne in mind that in the relevant RCT, DIGITAT,99 randomisation occurred after ultrasound scanning led to suspicion of SGA. Hence, the group randomised to expectant management would still have received enhanced monitoring and high-risk care during labour as the infant was known to be SGA. By contrast, routine care in a trial of screening means that neither antenatal nor intrapartum care is tailored to the suspected SGA status of the fetus.

Implications of sample size calculations

We present the data on sample size calculations but we are not recommending a specific trial design. It is also possible that a trial may be considered where the combination of screening parameters, intervention effect and outcome are not listed in Table 17. The exact design of the trial would depend on the resources available and the research question. We do, however, discuss some of the issues that may motivate a choice.

We believe that the calculations above rule out a trial based on either perinatal death or severe neonatal morbidity as the sample size required is so great that the trial may not be feasible, but would inevitably be extremely expensive. Whether the screening test is simply for SGA or one of the FGR indicators is used will depend on the trade-off between labelling much larger numbers of women as screen positive and sample size. In all calculations, the screen-positive rate was higher for SGA, but the sample size was smaller.

Whether a ‘screen versus no screen’ or a ‘screen all’ approach is used will depend on the information required and on the screening test evaluated. A problem with the ‘screen all’ approach is that it would not capture the real world of comparing not doing something with doing it. It would also not capture the health benefits of diagnosing non-cephalic presentation at 36 weeks’ gestation. However, it would provide more information about the evidence base as it would allow the performance of the screening test and the intervention to be quantified separately. Finally, the complicated SGA outcome is delivery of a small infant where either the mother experiences pre-eclampsia or the infant experiences morbidity. This outcome has the attraction of focusing on the cases most likely to reflect true FGR and it is perhaps in this group that the intervention is most likely to yield a positive result. However, a primary outcome that includes morbidity of all infants may be preferred if the priority is to determine the overall effect of screening and intervention. It is also worth noting in the ‘complicated SGA’ outcome that the ‘screen all’ study design would actually involve performing more scans than the ‘screen versus no screen’ design if the screening test was simple SGA or the Delphi consensus definition of FGR.

Overall conclusions

• Late-pregnancy ultrasound is only weakly predictive of neonatal morbidity.

• Late-pregnancy ultrasound is strongly predictive of SGA and LGA.

• There is a strong health economic case for implementing ultrasound scan in late pregnancy to assess fetal presentation.

• There is a chance that screening for fetal size in late pregnancy may be cost-effective under the current NHS recommendations; however:

  • The balance of probabilities favours a presentation-only scan.

  • The case for including assessment of fetal size is sensitive to the assumptions of the model.

  • There is no direct evidence from a RCT or meta-analysis that screening and intervention are clinically effective.

• The main uncertainty in relation to the health economic case for universal ultrasound (including both presentation and an estimate of fetal size) is uncertainty about the net costs of IOL compared with expectant management.

• RCTs of late-pregnancy screening aimed at directly demonstrating a protective effect on the risk of perinatal death or severe morbidity are unlikely to be feasible because of the required sample size.

• RCTs of late-pregnancy screening aimed at directly demonstrating a protective effect on the risk of proxies or subgroups of outcomes could be feasible because of sample size, but would depend on the exact study design.

Consultation with the National Screening Committee

We sent the scientific summary of the project and Chapter 13 to the UK National Screening Committee (NSC) Evidence Lead, who has worked for the UK NSC for > 15 years. The UK NSC would be happy to contribute to any further HTA discussions where this is useful. Following preliminary discussion, the applicants plan to submit a proposal to the UK NSC to suggest that it recommends a screening programme for breech presentation near term. Their evidence review process is outlined on its website [www.gov.uk/government/publications/uk-nsc-evidence-review-process (accessed June 2019)].

We then discussed the case for a trial of including assessment of fetal size in the same scan. The key questions were as follows:

• If the uncertainty around the costs of IOL were reduced, how likely is it that the NSC would recommend screening for fetal size near term based on a model that lacked direct evidence from a RCT that involved screening? For example, if the currently funded HTA trial around IOL for suspected fetal macrosomia confirms improved outcomes with intervention, would the combination of the diagnostic effectiveness of ultrasound as a screening test for LGA and the clinical effectiveness of IOL as an intervention in LGA be regarded as acceptable evidence for screening? The issue of interpretation is that screened women are likely to have lower prior odds of complications than women identified as having a LGA fetus through a clinically indicated scan. Hence, extrapolation of the results of the trial may involve an assumption that is untrue.

• If direct evidence of a beneficial effect of screening from a RCT was required, would this have to come from a ‘screen versus no screen’ trial or would evidence from a ‘screen all’ trial suffice?

• What outcomes would be acceptable? Specifically –

  • Would screening be recommended on the basis of an effect on proxies?

  • Would screening be considered on the basis of an effect on a subgroup, for example, subgroups of neonatal morbidity or mortality confined to infants who were actually small or large at birth?

  • Would screening be considered on the basis of an effect on a composite outcome?

Following discussion, the overview was that the NSC does not have specific ‘hard stops’ but, as one would expect, the stronger the evidence across the 20 criteria for assessing the viability of a screening programme, the more likely it is that a programme would be recommended. For example, because the committee bases recommendations on an assessment of these criteria, it would not necessarily reject a screening programme because the main trial supporting the programme reported a composite outcome in one criterion. However, all other things being equal, a programme would be less likely to be recommended if the study was based on a composite. Hence, none of the questions above was answered by a simple yes/no. The following were key points:

• RCTs based on intervention from screen-positive women would provide much stronger support for a programme than evidence derived from RCTs of high-risk women (i.e. those not identified through screening the general population).

• Data from a ‘screen versus no screen’ study would be preferred to those from a ‘screen all’ design. However, if there were absolute methodological obstacles to ‘screen versus no screen’, one approach would be to show proof of principle with a ‘screen all’ study, consider other studies to address any shortfall arising from this design and other criteria, and then perform a stepped-wedge RCT trial when implementing the new test.

Although evidence from trials reporting proxies, subgroups and composite outcomes would be considered, a strong case for screening would involve a simple substantive outcome that reflected the totality of the effect of screening (i.e. benefit to true positives and harm to false positives).

Steering group members

David Cromwell, Gianluca Baio, Kathryn Cook, Elizabeth Duff, Neil Marlow, Tracey Mills and Dharmintra Pasupathy.

We would like to thank the members of our steering group and our co-applicants Ian White, David Fields and Charlotte Bevan for valuable advice at different stages of the project. We are also grateful to Amanda Stranks for her strategic help with PPI and engagement and Alison Dacey for analysing records on breech presentation in the POP study.

For their roles as second reviewers on one of the systematic reviews, we would also like to thank Tom Bainton (umbilical artery), Norman Shreeve (LGA), Illianna Armata (borderline AFI) and Dexter Hayes (cerebroplacental ratio).

Contributions of authors

Gordon CS Smith (https://orcid.org/0000-0003-2124-0997) (Professor, Head of Department of Obstetrics and Gynaecology) conceived the project and contributed to protocol development, management of the project, planning of the systematic reviews, conceptualisation of the economic models, clinical interpretation of findings and writing of the report.

Alexandros A Moraitis (https://orcid.org/0000-0003-4634-1129) (Research Associate, Obstetrics and Gynaecology) performed the systematic reviews of clinical effectiveness, drafted and edited the final report, designed questionnaires for determining which conditions the systematic reviews should focus on, commissioned the focus group, and contributed to the identification of data and conceptualisation of the economic models.

David Wastlund (https://orcid.org/0000-0002-5074-4740) (Research Assistant, Health Economics) conceptualised and programmed the economic models, identified and estimated data for the economic analyses, performed the cost-effectiveness and VOI analyses, performed the systematic review on AFI, and drafted and edited the final report.

Jim G Thornton (https://orcid.org/0000-0001-9764-6876) (Professor, Obstetrics and Gynaecology) provided input on which systematic reviews to undertake and the design of future research, helped design the questionnaire for identifying topics for systematic reviews, and commented on drafts of the systematic reviews, the economic analyses and the final report.

Aris Papageorghiou (https://orcid.org/0000-0001-8143-2232) (Professor, Obstetrics and Gynaecology) contributed to designing the methods for systematic reviews, provided input on the design of future research, and commented on drafts of the report.

Julia Sanders (https://orcid.org/0000-0001-5712-9989) (Professor, Clinical Nursing and Midwifery) provided input on which systematic reviews to undertake, and edited drafts of the economic analyses and final report.

Alexander EP Heazell (https://orcid.org/0000-0002-4303-7845) (Professor, Obstetrics) helped design the methods for the systematic reviews, provided input on which clinical areas the systematic reviews should target and edited the final report.

Stephen C Robson (https://orcid.org/0000-0001-7897-7987) (Professor, Fetal Medicine) reviewed chapters on systematic reviews, contributed to the design of PPI and edited the final report.

Ulla Sovio (https://orcid.org/0000-0002-0799-1105) (Senior Research Associate, Applied Medical Statistics) contributed to the statistical analysis of the systematic reviews, reviewed and commented on drafts of the systematic reviews and the economic analyses, contributed to data analysis for the economic models, and edited the final report.

Peter Brocklehurst (https://orcid.org/0000-0002-9950-6751) (Professor of Women’s Health) provided input on which systematic reviews to undertake and the design of future research, helped design the questionnaire for identifying topics for systematic reviews, and commented on drafts of the systematic reviews, the economic analyses and the final report.

Edward CF Wilson (https://orcid.org/0000-0002-8369-1577) (Senior Lecturer, Health Economics) designed and programmed the models for the economic and VOI analysis, designed methods for the quantification of data for the economic analysis, and drafted and edited the final report.

Publications

Wastlund D, Moraitis AA, Dacey A, Sovio U, Wilson EC, Smith GC. Screening for breech presentation using universal late-pregnancy ultrasonography: a prospective cohort study and cost-effectiveness analysis. PLOS Med 2019;16:e1002778.

Wastlund D, Moraitis AA, Thornton JG, Sanders J, White IR, Brocklehurst P, et al. The cost-effectiveness of universal late-pregnancy screening for macrosomia in nulliparous women: a decision-analysis. BJOG 2019;126:1243–50.

Moraitis AA, Bainton T, Sovio U, Brocklehurst P, Heazell AEP, Thornton JG, et al. The fetal umbilical artery Doppler as a tool for universal third trimester screening: a systematic review and meta-analysis of diagnostic test accuracy. Placenta 2021; in press.

Wilson ECF, Wastlund D, Moraitis AA, Smith GCS. Late pregnancy ultrasound to screen for and manage potential birth complications in nulliparous women: a cost-effectiveness and value of information analysis. Value Health 2021; in press.

Data-sharing statement

All available data are contained within the report. All requests for access to study data should be made to the corresponding author.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.

MEDLINE and EMBASE

Date range searched: inception to 19 March 2019.

Search strategy

1. exp pregnant woman/

2. exp pregnancy/

3. pregnan*.mp.

4. exp prenatal diagnosis/

5. exp fetus echography/

6. exp Doppler ultrasonography/

7. arterial doppler.mp.

8. doppler velocimetry.mp.

9. doppler ultraso*.mp.

10. umbilical arter*.mp.

11. 1 or 2 or 3

12. 4 or 5 or 6

13. 7 or 8 or 9 or 10

14. 11 and 12

15. 13 and 14.

FIGURE 25.

The POP study inclusion flow chart. USS, ultrasound scan.

TABLE 18 - Maternal characteristics and birth outcomes of POP study

DM, diabetes mellitus; IQR, interquartile range.

p-value for trend.

Characteristic	UA PI > 90th centile (N = 346)	UA PI < 90th centile (N = 3269)	p-value	Overall baseline characteristics (N = 3615)
Maternal characteristic
Age (years), median (IQR)	29.7 (26.2–32.7)	30.3 (26.8–33.3)	0.05	30.2 (26.7–33.3)
Deprivation quartile, n (%)
1 (lowest)	97 (28.0)	784 (24.0)	0.14	881 (24.4)
2	73 (21.1)	776 (23.7)		849 (23.5)
3	92 (26.6)	773 (23.7)		865 (23.9)
4 (highest)	71 (20.5)	799 (24.4)		870 (24.1)
Missing	13 (3.7)	137 (4.2)		150 (4.2)
White ethnicity, n (%)	324 (93.6)	3036 (92.9)	0.53	3360 (93.0)
Missing	6 (1.7)	56 (1.7)		62 (1.7)
Married, n (%)	229 (66.2)	2238 (68.5)	0.39	2467 (68.2)
Smoker, n (%)	24 (6.9)	152 (4.7)	0.06	176 (4.9)
Any alcohol consumption, n (%)	13 (3.8)	155 (4.7)	0.40	168 (4.7)
Missing	0 (0)	1 (0)		1 (0)
BMI (kg/m2), median (IQR)	24.3 (21.7–28.1)	24.0 (21.8–27.2)	0.44	24.0 (21.8–27.3)
One or more previous miscarriage(s), n (%)	34 (9.8)	331 (10.1)	0.86	365 (10.1)
Chronic hypertension, n (%)	25 (7.3)	161 (4.9)	0.06	186 (5.1)
Pre-eclampsia, n (%)	29 (8.4)	204 (6.2)	0.12	233 (6.5)
Missing	0 (0)	2 (0.1)		2 (0.1)
DM, n (%)
Type 1 or type 2	2 (0.6)	10 (0.3)	0.14	12 (0.3)
Gestational	20 (5.8)	124 (3.8)		144 (4.0)
Birth outcome
Birthweight (g), median (IQR)	3263 (2970–3560)	3470 (3170–3770)	< 0.001	3445 (3150–3750)
Gestational age (weeks), median (IQR)	40.4 (39.3–41.1)	40.4 (39.4–41.3)	0.74	40.4 (39.4–41.3)
< 37	3 (0.9)	34 (1.0)	0.19a	37 (1.0)
37	22 (6.4)	133 (4.1)		155 (4.3)
38	35 (10.1)	360 (11.0)		395 (10.9)
39	71 (20.5)	641 (19.6)		712 (19.7)
40	92 (26.6)	1001 (30.6)		1093 (30.2)
41	102 (29.5)	909 (27.8)		1011 (30.0)
≥ 42	21 (6.1)	191 (5.8)		212 (5.9)
IOL, n (%)	125 (36.1)	1081 (33.1)	0.25	1206 (33.4)
Mode of delivery, n (%)
Spontaneous vaginal	178 (51.5)	1662 (50.8)	0.20	1840 (50.9)
Assisted vaginal	86 (24.9)	821 (25.1)		907 (25.1)
Intrapartum caesarean	54 (15.6)	601 (18.4)		655 (18.1)
Pre-labour caesarean	27 (7.8)	176 (5.4)		203 (5.6)
Missing	1 (0.3)	9 (0.3)		10 (0.3)

FIGURE 26.

Literature search PRISMA flow diagram for the systematic review on umbilical artery Doppler.

FIGURE 27.

Risk of bias and applicability concerns using the QUADAS-2 tool for the studies included in the meta-analysis of umbilical artery Doppler.

TABLE 19 - Characteristics of studies included in the meta-analysis

IUGR, intrauterine growth restriction; PIH, pregnancy-induced hypertension; S/D ratio, systolic/diastolic ratio.

Study (first author and year of publication)	Type of study; setting	Number of fetuses and selection (all singleton, non-anomalous unless otherwise stated)	Index test	Gestational age at ultrasound	Reference standard	Gestational age at delivery	Other comments
Akolekar 201942	Prospective cohort; two NHS hospitals, UK	n = 47,211	PI > 90th centile	Between 35+ 6 and 37+ 6 weeks’ gestation	Adverse perinatal outcome (composite of stillbirth, neonatal deaths and HIE grade 2 or 3), perinatal hypoxia (cord arterial pH of < 7.0, 5-minute Apgar score of < 7, NICU admission), caesarean section for fetal compromise, SGA < 3rd centile	Median gestational age at delivery 40.0 (39.0–40.9) weeks	Nulliparous: 45.4% for those with no adverse outcome, 58.5% for those with adverse outcome
	Between March 2014 and September 2018 (potential overlap with Valino et al. studies51,52)	Universal, > 36 weeks’ gestation	Not blinded
Bolz 201343	Prospective cohort; single hospital, Germany	n = 514 Low risk, term, cephalic only	PI > 1.2 Blinded umbilical artery Doppler	Within 1 week from delivery	Neonatal acidosis (cord arterial pH of < 7.10)	Mean gestational age: 40+1 weeks	Nulliparity: not reported
				Mean gestational age 39+2 weeks			IOL: not reported
		Excluded maternal disease, SGA, RFM
Cooley 201144	Prospective cohort; single hospital, Ireland	n = 810 Mixed risk, nulliparous only. Only included Caucasians aged 18–40 years	PI > 95th centile	Around 36 weeks’ gestation (not specified)	Emergency caesarean section, PIH, pre-eclampsia, preterm delivery (< 37 weeks’ gestation), SGA < 10th centile, SGA < 3rd centile, 5-minute Apgar score of < 7, cord arterial pH of < 7.10, NICU admission, stillbirth	Not reported	Nulliparity: all
			Umbilical artery blinded but EFW not blinded				IOL: 22.4%
Filmar 201345	Retrospective cohort; single hospital, New York, NY, USA	n = 251 Mixed risk, EFW > 10th centile	S/D ratio > 90th centile (persistent), not blinded	Mean gestational age 35.3 weeks for abnormal umbilical artery group. Mean gestational age 34.4 weeks for control group	NICU admission, 5-minute Apgar score of < 7	Median gestational age: 37 weeks for the abnormal umbilical artery group, 39 weeks for the control group	Nulliparity: not reported
							IOL: not reported
Fischer 199146	Prospective cohort; single hospital, PA, USA	n = 75	S/D ratio > 3.0	Mean interval from scan to delivery: 2 days	Composite perinatal outcome: non-reassuring intrapartum fetal heart rate umbilical artery pH of < 7.15 or a venous pH of < 7.2 5-minute Apgar score of < 7 meconium-stained liquor NICU admission birthweight < 10th centile	Mean gestational age: at delivery 292.2 days	Nulliparity: 57%
		Low risk, post dates > 41 weeks’ gestation. Excluded maternal disease, suspected IUGR	S/D ratio > 2.4				IOL: not reported
			Blinded umbilical artery Doppler
Goffinet 199647	Prospective cohort; 17 hospitals, France	n = 1903 Low risk, excluded maternal disease, suspected IUGR	RI > 90th centile	Between 28 and 34 weeks’ gestation	PIH, pre-eclampsia, intervention for fetal distress, 5-minute Apgar score of < 7, NICU admission, birthweight < 3rd centile, birthweight 3–10th centile	Mean gestational age: 39.2 weeks for those with an abnormal umbilical artery, 39.4 weeks for those with a normal umbilical artery	Nulliparous: 43.0% for those with an abnormal umbilical artery, 45.3% for those with normal umbilical aretery
			Not blinded
Hanretty 198948	Prospective cohort; single hospital, Glasgow, UK	n = 395 Universal	S/D ratio > 95th centile	34–36 weeks’ gestation	PIH, SGA < 5th centile, 5-minute Apgar score of < 6, NICU admission	Mean gestational age: 38.9 weeks for those with an abnormal umbilical artery, 39.5 weeks for those with a normal umbilical artery	Nulliparity: not reported
			Blinded umbilical artery Doppler				IOL: not reported
Moraitis 202135	Prospective cohort; single hospital, Cambridge, UK	n = 3615 Universal, nulliparous only, > 36 weeks’ gestation	PI > 90th centile	Mean 36 weeks’ gestation	NICU admission, metabolic acidosis, 5-minute Apgar score < 7, composite neonatal morbidity (one or more of the above), composite severe neonatal morbidity, SGA < 10th centile, SGA < 3rd centile	40.4 (39.3–41.1) weeks’ gestation	Nulliparity: all
			Blinded				IOL: 36.1% for those with an abnormal umbilical artery Doppler, 33.1% for those with a normal umbilical artery Doppler
Schulman 198949	Prospective cohort; single hospital, NY, USA	n = 255 Mixed	S/D ratio > 3 Not blinded	Around 30 weeks’ gestation	SGA < 15th centile	Not reported	Nulliparous: not reported
							IOL: not reported
Sijmons 198950	Prospective cohort; single hospital, the Netherlands	n = 368 Mixed (randomly selected)	PI > 95th centile Blinded umbilical artery Doppler	At 28 and 34 weeks’ gestation	SGA < 10th centile, SGA < 3rd centile	Not reported	Nulliparous: not reported
							IOL: not reported
Valino 201651	Retrospective cohort; three NHS hospitals, south-east England, UK	n = 8262	PI > 95th centile	30+0–34+6 weeks’ gestation	Term pre-eclampsia, term SGA < 10th centile, stillbirth, caesarean section for fetal distress, cord arterial pH of < 7.0, 5-minute Apgar score of < 7, NICU admission	Mean 40.0 weeks’ gestation	Nulliparous: 49.2%
	May 2011–August 2014	Universal	PI > 90th centile	Mean 32.2 weeks’ gestation			IOL: 15.5%
			Not blinded
Valino 201652	Retrospective cohort; two NHS hospitals, south-east England, UK	n = 3953	PI > 95th centile	35+0–37+6 weeks’ gestation	Pre-eclampsia, SGA < 10th centile, caesarean section for fetal distress, cord arterial pH of < 7.0, 5-minute Apgar score of < 7, NICU admission	Mean 40.0 weeks’ gestation	Nulliparous: 49.7%
	February 2014–December 2014 (potential overlap with above)	Universal	Not blinded	Mean 36.1 weeks’ gestation			IOL: 19.1%
Weiner 199353	Prospective cohort; single hospital, Israel	n = 142	RI > 95th centile	After 41 weeks’ gestation	Composite adverse outcome: 5-minute Apgar score of < 7 NICU admission Caesarean section for fetal distress, SGA < 5th centile	Mean 41.8 weeks’ gestation	Nulliparous: n = 43
		Low risk, term only afert 41 weeks’ gestation	Not blinded				IOL: not reported

FIGURE 28.

Deeks’ funnel plot for publication bias for umbilical artery Doppler for the prediction of neonatal unit admission. Deeks’ funnel plot asymmetry test, p = 0.52. ESS, effective sample size.

MEDLINE and EMBASE

Date range searched: inception to 30 May 2019.

Search strategy

1. exp pregnant woman/

2. exp pregnancy/

3. pregnan*.mp.

4. exp fetus echography/

5. exp prenatal diagnosis/

6. exp Doppler ultrasonography/

7. exp fetus monitoring/

8. ultraso*.mp.

9. exp middle cerebral artery/

10. middle cerebral artery.mp.

11. uteroplacental.mp.

12. utero-placental.mp.

13. cerebroplacental.mp.

14. cerebro-placental.mp.

15. cerebroumbilical.mp.

16. cerebro-umbilical.mp.

17. fetal brain doppler.mp.

18. fetal cerebral doppler.mp.

19. 1 or 2 or 3

20. 4 or 5 or 6 or 7 or 8

21. 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18

22. 19 and 20

23. 21 and 22.

FIGURE 29.

Literature search PRISMA flow diagram for the systematic review on CPRs.

FIGURE 30.

Risk of bias and applicability concerns using the QUADAS-2 tool for the studies included in the meta-analysis of CPRs.

TABLE 20 - Characteristics of studies included in the meta-analysis of CPRs to predict adverse pregnancy outcome

HIE, hypoxic–ischaemic encephalopathy; IQR, interquartile range; MoM, multiples of median; USS, ultrasound scan.

Study (first author and year of publication)	Type of study; setting	Number of fetuses and selection (all singleton, non-anomalous unless otherwise stated)	Index test CPR = MCA PI/umbilical artery PI (unless otherwise stated)	Gestational age at ultrasound	Reference standard	Gestational age at delivery	Other comments
Akolekar 201557	Prospective cohort; two NHS hospitals (King’s College London and Medway Maritime Hospital), UK	n = 6038	CPR < 5th centile	35+0 to 37+6 weeks	Cord arterial pH of < 7.0, 5-minute Apgar score of < 7, NICU admission	Median 39.9 (IQR 39.0–40.7) weeks	Nulliparous: 49.8%
	Between February 2014 and December 2014	Universal screening	Not blinded	Median 36.1 (IQR 36.0–36.6) weeks			IOL: 20% overall
Akolekar 201942	Prospective cohort; two NHS hospitals (King’s College London and Medway Maritime Hospital), UK	n = 47,211	CPR < 10th centile	Between 35+0 and 37+6 weeks	Adverse perinatal outcome (composite of stillbirths, neonatal deaths and HIE grade 2 or 3), perinatal hypoxia (composite of cord arterial pH of < 7.0 and venous < 7.1, 5-minute Apgar score of < 7, NICU admission for > 24 hours), caesarean section for fetal compromise, SGA < 3rd centile	Median gestational age at delivery 40.0 (39.0–40.9) weeks	Nulliparous: 45.4% for those with no adverse outcome, 58.5% for those with adverse outcome
	Between March 2014 and September 2018; significant population overlap with the 2015 Akolekar et al. study57	Universal screening	Not blinded				IOL: not reported
Bakalis 201558	Prospective cohort; three NHS hospitals (King's College London, University College London, Medway Maritime Hospital), UK	n = 30,780	CPR < 5th centile	30+0 to 34+6 weeks, mean 32.3 (IQR 32.0–32.9) weeks	Stillbirth, emergency caesarean section for fetal distress, cord arterial pH of < 7.0, cord venous pH of 7.1, 5-minute Apgar score of < 7, NNU admission, NICU admission	Median 40 (IQR 39.0–40.9) weeks	Nulliparous: 50.2%
							Further analysed in SGA vs. AGA and delivery < 2 weeks from scan vs. > 2 weeks from scan
	Between May 2011 and August 2014; likely to be population overlap with Akolekar et al. studies42,57	Universal screening	Not blinded				IOL: 14.5% overall
Bligh 201859	Prospective cohort; single hospital, Brisbane, QLD, Australia (May 2014–August 2016)	n = 437	CPR < 10th centile	From 36+1 weeks’ gestation	Caesarean section for fetal distress. Composite adverse neonatal outcome (cord arterial pH of < 7.10, 5-minute Apgar score of < 7 or NICU admission)	Median 40 weeks (IQR 39.3–40.9 weeks)	Nulliparous: 87.4%
		Low risk	Blinded	Within 2 weeks of delivery			IOL: not reported
		Uncomplicated, term only
Bligh 201860	Prospective cohort; single hospital, Brisbane, QLD, Australia (May 2014–August 2016)	n = 437	CPR < 10th centile	From 36 weeks’ gestation	SGA < 10th centile	Median 40 weeks (IQR 39.3–40.9 weeks)	Nulliparous: 87.4%
		Low risk	CPR < 5th centile	Within 2 weeks of delivery	SGA < 5th centile		IOL: not reported
		Uncomplicated, term only	Blinded
Flatley 201961	Retrospective cohort; single hospital, Brisbane, QLD, Australia (2010–15) (likely to be some population overlap with Bligh et al.59,60)	n = 2425	CPR < 10th centile	Between 36 and 38 weeks’ gestation	Cord arterial pH of < 7.00, 5-minute Apgar score of ≤ 3, NICU admission, perinatal death. Composite of all of the above (SCNO) caesarean section for fetal distress. SGA < 10th centile, SGA < 5th centile	Term only, 54.5% of those with an abnormal CPR delivered < 39 weeks, 36.4% of those with a normal CPR	Nulliparous: 65.4% of those with an abnormal CPR, 48.0% of those with a normal CPR
		Mixed risk
		Excluded preterm delivery < 37 weeks’ gestation, maternal hypertension and diabetes mellitus	Not blinded				IOL: 46.4% for those with an abnormal CPR, 39.5% for those with a normal CPR
Khalil 201562	Retrospective cohort; one tertiary NHS hospital (St George’s), UK (2000–13)	n = 9772	CPR < 0.6765 MoM	Within 2 weeks of delivery	NNU admission	Median 41.1 weeks for both those admitted and those not admitted to NNU	Nulliparous: 65.2% of those admitted to NNU, 54.6% for those not admitted to NNU
		Low risk
		Term only. For the analysis of operative delivery for fetal distress, the patients who had elective caesarean section were excluded	Not blinded	Median 40.4 weeks for those admitted to NNU, 40.4 weeks for those not admitted to NNU	Operative delivery of fetal distress, (including instrumental delivery and caesarean section)		IOL: 44.1% for NNU, 39.4% for no NNU
Maged 201463	Prospective cohort; single hospital, Cairo, Egypt	n = 100	CPR < 1.05	37.8 weeks’ gestation for those with adverse outcome, 39.5 weeks’ gestation for those with normal outcome	Caesarean section for fetal distress	283.1 days for those with adverse outcome, 281.7 days for those with normal outcome	Nulliparous: not reported
		Low risk
		Included those delivered between 40 and 42 weeks’ gestation Excluded PPROM, APH, patients in labour and maternal HTN/DM	Not blinded		Composite adverse pregnancy outcome defined as one or more of caesarean section for fetal distress, 5-minute Apgar score of < 7, MAS, NICU admission		IOL: not reported
Monaghan 201864	Retrospective cohort; single NHS hospital (St George’s), UK	n = 7013	CPR < 10th centile	36.4 weeks for all live births, 37 weeks for perinatal deaths	Perinatal death	Median: 40.1 weeks’ gestation for all live births, 39 weeks’ gestation for perinatal deaths	Nulliparous: not reported
	January 2008–June 2016 (likely to be population overlap with Khalil et al.62)	Mixed risk (had ultrasound scan based on NHS indications)	CPR < 5th centile				IOL: not reported
		Only included those delivered after 36 weeks’ gestation	Not blinded
Morales-Roselló 201465	Retrospective cohort; single NHS hospital (St George’s), UK, 2002–12 (likely to be population overlap with Khalil et al.62 and Monaghan et al.64)	n = 11,576	CPR < 0.6765 MoM	Mean 40.1 ± 1.5 weeks	SGA < 10th centile	Mean 40.8 ± 1.3 weeks	Nulliparous: not reported
		Mixed risk
		Term only with ultrasound scan within 14 days of delivery	Not blinded				IOL: not reported
Prior 201366	Prospective cohort; single NHS hospital (Queen Charlotte’s and Chelsea), UK. (March 2011–March 2014)	n = 400	CPR < 10th centile	Mean 40 weeks’ gestation + 2 days (range 37+0–42+1 weeks)	Caesarean section for fetal compromise, 5-minute Apgar score of < 7, cord arterial pH of < 7.20, NNU admission	Within 72 hours from scan	Nulliparous: 65.5%
		Low risk
		Term only. Recruited before active labour. Excluded pre-eclampsia, FGR, intrauterine infection	Blinded				IOL: not reported
Prior 201567	Prospective cohort; single tertiary NHS hospital (Chelsea), UK. (likely to be population overlap with the Prior et al. study66)	n = 775	CPR < 0.6765 MoM	Median 41 weeks’ gestation (range 37–42 weeks)	Caesarean section for fetal distress, 5-minute Apgar score of < 7, cord arterial pH of < 7.20, NNU admission	Within 72 hours from scan	Nulliparous: 80.8%
		Low risk
		Term only. Recruited before active labour or IOL (for post dates or social). Excluded SGA/FGR, PIH/pre-eclampsia, PPROM	Blinded				IOL: not reported
Rial-Crestelo 201968	Prospective cohort; single hospital, Barcelona, Spain. January 2013–December 2016	n = 1030	CPR < 10th centile	Between 32+0 and 34+6 weeks, mean 33 weeks	SGA < 10th centile	Mean 40 weeks’ gestation	Nulliparous: 70% of those born SGA, 54% of those not born SGA
		Universal screening	Doppler blinded for those with EFW > 10th centile				IOL: not reported
Sabdia 201569	Retrospective cohort; single hospital, Brisbane, QLD, Australia (June 1998–November 2013)	n = 1381 Mixed risk Included cephalic with umbilical artery PI < 95th centile	CPR < 10th centile (1.20)	Between 35 and 37 weeks’ gestation	Operative delivery for fetal distress (caesarean section or instrumental), 5-minute Apgar score of < 7, NICU admission	Median gestational age 36 weeks for those with an abnormal CPR, 38 weeks for those with a normal CPR	Nulliparous: 53.9% of those with an abnormal CPR, 40.4% of those with a normal CPR
			Not blinded
							IOL: not reported
Stumpfe 201970	Retrospective cohort; single tertiary centre, Germany (January 2016–April 2017)	n = 1008	CPR < 0.6765 MoM	Term, within 72 hours of delivery	Caesarean section for fetal distress, 5-minute Apgar score of < 7, cord arterial pH of < 7.10	Term (not further specified)	Nulliparous: not specified
		Low risk
		Term only, excluded those in labour, elective caesarean section, EFW < 10th centile	Not blinded				IOL: 42.4% overall
Twomey 201671	Retrospective cohort; single hospital, Brisbane, QLD, Australia (January 2007–December 2013) (population overlap with Sabdia et al.69)	n = 1224	CPR < 1	30–34 weeks, median 32.1 weeks	Caesarean section for fetal compromise, cord arterial pH of < 7.0, 5-minute Apgar score of ≤ 3, NNU admission, SGA < 10th centile, SGA < 5th centile	Mean gestational age 32 weeks for those with a CPR < 1, 37 weeks for those with a CPR > 1	Nulliparous: 43.2%
		Mixed risk
		Excluded women who had elective caesarean section	Not blinded				IOL: not reported

FIGURE 31.

Deeks’ funnel plot for publication bias for CPRs for the prediction of neonatal unit admission. Deeks’ funnel plot asymmetry test: p = 0.28. ESS, effective sample size.

MEDLINE and EMBASE

Date range searched: 1 January 2011 to 5 June 2019.

Search strategy

1. exp Pregnant Women/

2. limit 1 to yr=“2011 -Current”

3. exp Pregnancy Trimester/

4. limit 3 to yr=“2011 -Current”

5. pregnan*.mp.

6. limit 5 to yr=“2011 -Current”

7. exp Prenatal Diagnosis/

8. limit 7 to yr=“2011 -Current”

9. exp Ultrasonography, Prenatal/

10. limit 9 to yr=“2011 -Current”

11. exp Amniotic Fluid/

12. limit 11 to yr=“2011 -Current”

13. exp Oligohydramnios/

14. limit 13 to yr=“2011 -Current”

15. oligohydramnio*.mp.

16. limit 15 to yr=“2011 -Current”

17. exp Polyhydramnios/

18. limit 17 to yr=“2011 -Current”

19. polyhydramnio*.mp.

20. limit 19 to yr=“2011 -Current”

21. amniotic fluid index.mp.

22. limit 21 to yr=“2011 -Current”

23. AFI.mp.

24. limit 23 to yr=“2011 -Current”

25. maximum pool depth.mp.

26. limit 25 to yr=“2011 -Current”

27. MPD.mp.

28. limit 27 to yr=“2011 -Current”

29. single deepest pocket.mp.

30. limit 29 to yr=“2011 -Current”

31. SDP.mp.

32. limit 31 to yr=“2011 -Current”

33. largest vertical pocket.mp.

34. limit 33 to yr=“2011 -Current”

35. LVP.mp.

36. limit 35 to yr=“2011 -Current”

37. maximum vertical pocket.mp.

38. limit 37 to yr=“2011 -Current”

39. MVP.mp.

40. limit 39 to yr=“2011 -Current”

41. amniotic fluid volume.mp.

42. limit 41 to yr=“2011 -Current”

43. anhydramnios.mp.

44. limit 43 to yr=“2011 -Current”

45. liquor volume.mp.

46. limit 45 to yr=“2011 -Current”

47. quadrants.mp.

48. limit 47 to yr=“2011 -Current”

49. biophysical profile.mp.

50. limit 49 to yr=“2011 -Current”

51. BPP.mp.

52. limit 51 to yr=“2011 -Current”

53. 2 or 4 or 6

54. 8 or 10 or 12 or 14 or 16 or 18 or 20

55. 22 or 24 or 26 or 28 or 30 or 32 or 34 or 36 or 38 or 40 or 42 or 44 or 46 or 48 or 50 or 52

56. 53 and 54 and 55

57. 8 or 10

58. 12 or 14 or 16 or 18 or 20 or 22 or 24 or 26 or 28 or 30 or 32 or 34 or 36 or 38 or 40 or 42 or 44 or 46 or 48 or 50 or 52

59. 53 and 57 and 58.

FIGURE 32.

The PRISMA flow diagram for the systematic review of severe oligohydramnios. USS, ultrasound scan.

FIGURE 33.

Risk-of-bias graph of included studies for systematic review of severe oligohydramnios using the QUADAS-2 tool.

TABLE 21 - Characteristics of studies included in the meta-analysis of severe oligohydramnios

DM, diabetes mellitus; GDM, gestational diabetes mellitus; HIE, hypoxic–ischaemic encephalopathy; IUGR, intrauterine growth restriction; IVH, intraventricular haemorrhage; MAS, meconium aspiration syndrome; PPROM, preterm premature rupture of membranes; SROM, spontaneous rupture of membranes; USS, ultrasound scan.

Study (first author and year of publication)	Type of study; setting	Number of fetuses and selection (all singleton, non-anomalous unless otherwise stated)	Index test	Gestational age at ultrasound	Reference standard	Gestational age at delivery	Other comments
Ashwal 201474	Retrospective cohort; single university hospital, Israel	n = 23,267	AFI < 5 cm	Within 1 week from delivery	Caesarean section for fetal distress, operative vaginal delivery for fetal distress, 5-minute Apgar score of < 7, umbilical artery pH of < 7.10, NICU admission, need for intubation, MAS or HIE. Also stillbirth, neonatal death, IVH, meconium amniotic fluid (not MAS)	39+8 ± 1.1 weeks for isolated oligohydramnios; 39.3 ± 1.1 weeks for normal AFI	Nulliparous: n = 442 (44.8%) for isolated oligohydramnios, n = 6848 (30.7%) for normal AFI
		Low risk
		Term only. Excluded pregnancies with hypertensive disorders, diabetes, AFI > 25 cm, and EFW < 10th centile	Not blinded				IOL: n = 273 (27.7%) for oligo hydramnios, n = 824 (3.7%) for normal
Ghosh 200275	Prospective cohort; single hospital, Sweden	n = 333	AFI < 5 cm	In early labour or before IOL	Operative delivery for fetal distress, caesarean section for fetal distress, 5-minute Apgar score of < 7, cord arterial pH of < 7.10, NICU admission	Mean gestational age 283 days for those with AFI < 5 cm, 280 days for those with AFI > 5 cm	Nulliparous: 26/49 of those with AFI < 5 cm, 134 for those with AFI > 5 cm
		Low risk	Not blinded
		Term only, in early labour or prior to IOL
Hassan 200576	Cross-sectional; single hospital, Pakistan	n = 260	AFI < 6 cm	After 41+0 weeks	Neonatal death, caesarean section, meconium-stained amniotic fluid	After 41+0 weeks	Nulliparous: 34% of those with low AFI, 19.7% of those with normal AFI
		Low risk
		Post dates (after 41+0 weeks)	Not blinded				IOL: not specified
Hsieh 199877	Retrospective cohort; single hospital, Taiwan (Province of China)	n = 27,506	AFI < 5 cm	Not specified	Stillbirth, SGA < 10th centile, 5-minute Apgar score of < 7, NICU admission, neonatal death	Not specified	Nulliparous: not specified
		Universal	Not blinded				IOL: not specified
		Excluded those with AFI > 24 cm, PPROM
Locatelli 200478	Prospective cohort; single hospital, Italy	n = 3049	AFI < 5 cm	40 weeks’ gestation	Meconium-stained amniotic fluid, caesarean section for fetal distress, SGA < 10th centile, Apgar score of < 7, cord arterial pH of < 7.0	40+0–41+6 weeks’ gestation	Nulliparous: 72% for those with low AFI, 58% for those with normal AFI
		Universal
		Routine scan at 40 weeks’ gestation
		Excluded those with PPROM and those with other indications for ultrasound scan	Not blinded				IOL: 83% for those with low AFI, 25% for those with normal AFI
Megha 201379	Prospective cohort; single centre, India	n = 200	AFI < 5 cm	34–41 weeks’ gestation	Caesarean section for fetal distress, meconium-stained fluid, 5-minute Apgar score of < 7, cord arterial pH of < 7.10. Admission to NICU for > 48 hours	Not specified. 56% of those with low AFI delivered < 37 weeks’ gestation vs. 34.3% of those with normal AFI	Nulliparous: 68% of those with low AFI, 58.9% of those with normal AFI
		Mixed
		Selection not specified	Blinded	Within 7 days of delivery			IOL: 72% of those with low AFI, 51% of those with normal AFI
Melamed 201180	Matched cohort (3 : 1); single hospital, Israel	n = 432	AFI < 5 cm	Gestational age at initial ultrasound scan: 33.9 weeks for low AFI, 33.9 weeks for normal AFI	Caesarean section for fetal distress, meconium-stained fluid, preterm delivery (< 37 weeks’ gestation), admission to NICU	37.3 ± 1.6 weeks for cases, 39.1 ± 1.8 weeks for controls	Nulliparous: 62 (57.4%) of cases, 186 (57.4%) of controls
		Low risk
		Excluded pregnancies with pre-eclampsia/DM/GDM, EFW < 10th centile, abnormal umbilical artery Doppler, and PROM	Not blinded	Gestational age at last scan not reported			IOL: 54 (50%) of cases, 31 (9.6%) of controls
Morris 200381	Prospective cohort; single hospital, Oxford, UK	n = 1584	AFI < 5 cm	At or after 40 weeks’ gestation (59% at 40 weeks)	Caesarean section for fetal distress, NICU admission, 5-minute Apgar score of < 7	At or after 40 weeks’ gestation (615 at 41 weeks’ gestation)	Nulliparous: 778 (49.1%)
		Low risk	SDP < 2 cm				IOL: 643 (40.6%)
		Term only (> 40 weeks’ gestation). Excluded non-vertex and those with clinically required ultrasound	Not blinded
Myles 200282	Prospective cohort; single hospital, FL, USA	n = 266	AFI < 5 cm	Between 37+0 and 41+6 weeks (not specified)	Caesarean section for fetal distress, NICU admission, meconium-stained amniotic fluid	Not specified	Nulliparous: not specified
		Low risk	SDP < 2.5 cm				IOL: not specified
		Term only. Excluded non-vertex, SROM, polyhydramnios, and any pregnancies with fetal or maternal complications	Not blinded
Naveiro-Fuentes 201683	Retrospective cohort; single hospital, Spain	n = 27,708	AFI < 5 cm	39 weeks’ gestation	Caesarean section for fetal distress, instrumental delivery for fetal distress, meconium-stained fluid, SGA (< 10th centile), 5-minute Apgar score of < 7, admission to NICU, umbilical artery pH of < 7.10	279 ± 7.3 days for those with oligohydramnios, 278.2 ± 7.5 days for normal	Nulliparous: 65.1% of those with low AFI
		Low risk
		Term only. Routine antenatal scan at 39 weeks’ gestation. Excluded pregnancies with maternal or fetal pathology including suspected IUGR	Not blinded				IOL: not reported
Quiñones 201284	Prospective cohort; two centres, PA, USA	n = 308	AFI < 5 cm	37–40 weeks’ gestation (mean 38.1 ± 0.9 weeks’ gestation)	Fetal vulnerability index, which is defined as one or more of the following: 5-minute Apgar score of < 3, umbilical cord pH of < 7.0, intrapartum fetal death, neonatal seizures, intubation in the absence of meconium, or NICU admission for > 24 hours	Mean gestational age 39.9 ± 0.8 weeks	Nulliparous: 50%
			AFI < 8 cm
		Low risk	AFI < 10 cm
		Between 37 and 40 weeks’ gestation, excluded pregnancies with maternal or obstetric complications (including suspected FGR)	SDP < 2cm
Rainford 200185	Retrospective cohort; single hospital, USA	n = 232	AFI < 5 cm	Within 4 days of delivery	Operative delivery for fetal distress, NICU admission, 5-minute Apgar score of < 7, meconium-stained amniotic fluid	Mean gestational age 40.1 weeks for those with oligohydramnios, 40.9 weeks for normal AFI	Nulliparous: 17% for low AFI, 20% for normal AFI
		Low risk
		Term only. Excluded those with any maternal or fetal complications	Not blinded				IOL: 98% of those with low AFI, 51% of those with normal AFI
Shanks 201186	Retrospective cohort; single centre, USA	n = 17,877	AFI < 5 cm	Mean 34.38 ± 3.04 weeks’ gestation	NICU admission	Mean 38.27 ± 2.86 weeks’ gestation	Nulliparous: n = 7069 (39.5%)
		Mixed risk	AFI < 5th centile
		Selection criteria not specified	Not blinded
Zhang 200487	Clinical trial (ultrasound scan screening vs. no screening). For this study data used by the screening group	n = 6657 in the low-risk group. All women had two research scans at 15–22 and 31–35 weeks’ gestation. Excluded multiple pregnancies and those with any maternal or fetal conditions	AFI < 5 cm	31–35 weeks’ gestation	Caesarean section for fetal distress, 5-minute Apgar score of < 7, NICU admission, perinatal mortality	Mean gestational age 39.6 weeks for those with oligohydramnios, 39.8 weeks’ gestation for those with normal AFI	Nulliparous: 53% of those with oligohydramnios, 45% of normal AFI
			Not blinded				IOL: not specified

FIGURE 34.

Deeks’ funnel plot for publication bias for severe oligohydramnios for the prediction of neonatal unit admission. Deeks’ funnel plot asymmetry test: p = 0.54. ESS, effective sample size.

MEDLINE and EMBASE

Date range searched: inception to 18 June 2019.

Search strategy

1. exp Pregnant Women/

2. exp pregnancy/

3. pregnan$.mp.

4. exp oligohydramnios/

5. oligohydramnio$.mp.

6. exp Amniotic Fluid/

7. amniotic fluid index.mp.

8. AFI.mp.

9. liquor volume.mp.

10. ow.mp.

11. borderline.mp.

12. decreased.mp.

13. perinatal.mp.

14. peripartum.mp.

15. fetal.mp.

16. 1 or 2 or 3

17. 4 or 5 or 6 or 7 or 8 or 9

18. 13 or 14 or 15

19. 16 and 17 and 18

20. 10 or 11 or 12

21. 19 and 20.

FIGURE 35.

The POP study inclusion flow chart.

TABLE 22 - Patient characteristics and birth outcomes of POP study

IQR, interquartile range.

Characteristic	Borderline AFI 5–8 cm (n = 108)	Normal AFI 8–24 cm (N = 3279)	p-value	Overall baseline characteristics (N = 3387)
Maternal characteristic
Age (years), median (IQR)	30.1 (26.7–33.2)	30.3 (26.2–33.7)	0.60	30.1 (26.7–33.2)
Deprivation quartile, n (%)
1 (lowest)	29 (26.9)	808 (24.6)	0.53	837 (24.7)
2	28 (25.9)	769 (23.5)		797 (23.5)
3	23 (21.3)	776 (23.7)		799 (23.6)
4 (highest)	25 (23.2)	783 (23.9)		808 (23.9)
Missing	3 (2.8)	143 (4.4)		146 (4.3)
White ethnicity, n (%)	96 (88.9)	3052 (93.1)	0.16	3148 (92.9)
Missing	3 (2.8)	54 (1.7)		57 (1.7)
Married, n (%)	81 (75.0)	2222 (67.8)	0.11	2303 (68.0)
Smoker	3 (2.8)	164 (5.0)	0.29	167 (4.9)
Any alcohol consumption	1 (0.9)	154 (4.7)	0.06	155 (4.6)
Missing	0 (0.0)	1(0.0)		1 (0.0)
BMI (kg/m2), median (IQR)	23.4 (21.6–26.5)	23.9 (21.8–27.1)	0.19	23.9 (21.8–27.0)
One or more previous miscarriage(s), n (%)	8 (7.4)	327 (10.0)	0.38	335 (9.9)
Chronic hypertension, n (%)	4 (3.7)	164 (5.0)	0.54
Pre-eclampsia, n (%)	9 (8.3)	201 (6.1)	0.35	210 (6.2)
Missing	0 (0)	2 (0.1)		2 (0.1)
Birth outcome
Birthweight (g), median (IQR)	3260 (3005–3520)	3460 (3150–3770)	< 0.001	3450 (3150–3760)
Gestational age (weeks), median (IQR)	40.0 (38.8–40.9)	40.4 (39.6–41.3)	< 0.001	40.4 (39.6–41.3)
IOL, n (%)	41 (38.0)	1016 (31.0)	0.12	1057 (31.2)
Mode of delivery, n (%)
Spontaneous vaginal	70 (64.8)	1685 (51.4)	0.04	1755 (51.8)
Assisted vaginal	19 (17.6)	832 (25.4)		851 (25.1)
Intrapartum caesarean	13 (12.0)	596 (18.2)		609 (18.0)
Pre-labour caesarean	6 (5.6)	157 (4.8)		163 (4.8)
Missing	0 (0.0)	9 (0.3)		9 (0.3)

FIGURE 36.

The PRISMA flow diagram for the systematic review of borderline oligohydramnios.

FIGURE 37.

Risk-of-bias and applicability concerns for included studies in systematic review of borderline oligohydramnios using the QUADAS-2 tool.

TABLE 23 - Characteristics of studies included in the meta-analysis of borderline oligohydramnios

IUGR, intrauterine growth restriction; PPROM, preterm premature rupture of membranes; SROM, spontaneous rupture of membranes; USS, ultrasound scan.

Alexandros A Moraitis, Ilianna Armata, Ulla Sovio, Peter Brocklehurst, Alexander EP Heazell, Jim G Thornton, Stephen C Robson, Aris Papageorghiou and Gordon CS Smith, University of Cambridge, 2021.

Study (first author and year of publication)	Type of study; setting	Population and selection (singletons only unless otherwise specified)	Index test	Gestational age at ultrasound	Reference standard	Gestational age at delivery (mean unless otherwise specified)	Other comments
Asgharnia 201389	Retrospective cohort; single hospital, Islamic Republic of Iran	n = 235	5 < AFI < 10 cm	> 28 weeks’ gestation (mean gestational age not reported)	RDS, 5-minute Apgar score of < 7, NICU, IUGR, SGA < 10th centile	Mean gestational age not reported	Nulliparous: BAFI 68.1%, normal AFI 58.2%
		Mixed risk
		Pregnancies > 28 weeks. Excluded PPROM, uterine anomalies, vaginal bleeding	Not blinded			Preterm: BAFI 40.4%	IOL: BAFI 22.3%, normal AFI 10.6%
						Normal AFI 14.9%
Banks, 199990	Retrospective cohort; single hospital, USA	n = 214	5 cm < AFI < 10 cm	Not reported	Intrapartum fetal distress, meconium-stained amniotic fluid, SGA < 10th centile	Not reported	Nulliparous: not reported
		Mixed risk
		Pregnancies with antepartum testing within 1 week of delivery	Not blinded				IOL: not reported
Choi 201691	Retrospective cohort; single hospital, the Republic of Korea	n = 721	5.1 ≤ AFI ≤ 8.0 cm	Within 1 week of delivery	Meconium-stained amniotic fluid, caesarean section for fetal distress, 5-minute Apgar score of < 7, NICU admission, SGA < 10th centile	BAFI: 39.2 weeks	Nulliparous: BAFI 66.1%, normal AFI 57.3%
		Low risk
		Uncomplicated, term pregnancies only				Normal AFI: 39.4 weeks	IOL: BAFI 60.7%, normal AFI 27.4%
		Excluded SROM, elective caesarean section, breech presentation, pre-eclampsia, and other maternal disease
Gumus, 200792	Retrospective cohort; single hospital, Turkey	n = 367	5 cm < AFI < 10 cm	Not reported	Intrapartum fetal distress, meconium-stained amniotic fluid, SGA < 10th centile), NICU admission, RDS	BAFI 37.7 weeks for normal AFI 38.3 weeks	IOL: BAFI 73.3%
		Mixed risk
		Excluded PROM, uterine anomalies, vaginal bleeding				Preterm: BAFI 18.9%, normal AFI 9.7%	Normal AFI 54.5%
Jamal 201693	Matched cohort (matched 1 : 1); single hospital, Islamic Republic of Iran	n = 128	5.1 ≤ AFI ≤ 8.0	37–40 weeks’ gestation	Meconium-stained amniotic fluid, 5-minute Apgar score of < 7, umbilical artery pH of < 7.0, NICU admission, SGA < 10th centile	BAFI (median): 37+5 weeks	Nulliparous: not reported
		Mixed risk
		Term only. Excluded PPROM, anomalies, maternal medical diseases, contraindications for vaginal delivery		Within 1 week of delivery		Normal AFI: 38+6 weeks	IOL: not reported
Kwon 200694	Retrospective cohort; single hospital, the Republic of Korea	n = 3740	5.1 ≤ AFI ≤ 8.0	Within 2 weeks of delivery	Perinatal death, NICU admission, caesarean section for fetal distress, 5-minute Apgar score of < 7, SGA < 10th centile	BAFI: 36.3 weeks’ gestation	Nulliparous: not reported
		Mixed risk
		Excluded fetal malformations, SROM pre-eclampsia, chromosomal anomalies, AFI > 25 cm				Normal AFI: 38.0 weeks’ gestation	IOL: not reported
The POP Studya	Prospective cohort; single centre; Cambridge, UK	n = 3387	5 cm < AFI < 8 cm	36 weeks’ gestation	NICU admission, metabolic acidosis, 5-minute Apgar score of < 7, composite morbidity (all above), composite severe morbidity		Nulliparous only
		Nulliparous only
		Universal screening	Blinded
Petrozella, 201195	Retrospective cohort; regional hospitals, USA	n = 27,601	5 cm < AFI < 8 cm	24+0 to 33+6 weeks’ gestation	Caesarean section for fetal distress, SGA < 10th centile, SGA < 3rd centile, neonatal death	BAFI 37.1 weeks’ gestation	Nulliparous: not reported
		Mixed risk
		Those who received USS between 24 and 34 weeks’ gestation		Mean gestational age 29.2 weeks		Normal AFI 39.2 weeks’ gestation	IOL: not reported
		Excluded AFI > 24 cm, SROM				Preterm: BAFI 37%, normal AFI 8%
Rutherford, 198796	Retrospective cohort; single hospital, USA	n = 286	5 cm < AFI < 8 cm	Not reported	Meconium, caesarean section for fetal distress, 5-minute Apgar score of < 7	Not reported	Nulliparous: not reported
		Mixed risk
		Those who had antepartum surveillance					IOL: not reported
		Excluded PPROM
Sahin, 201897	Prospective (matched 1 : 3); single hospital, Turkey	n = 430	5 cm < AFI ≤ 8 cm	Between 34+0 and 36+6 weeks’ gestation	5-minute Apgar score of < 7, caesarean section for fetal distress, RDS, meconium-stained amniotic fluid, meconium aspiration syndrome, NICU, neonatal death	BAFI: 37.5 weeks	Nulliparous: not reported
		Low risk		Mean 35,4 weeks’ gestation		Normal AFI: 38.6 weeks	IOL: BAFI 34.6%, normal AFI 23.8%
		Excluded maternal disease, IUGR chromosomal/fetal abnormalities, SROM, abnormal Doppler				Preterm: BAFI 15.9%, normal AFI 8.4%
Wood 201498	Retrospective cohort (matched 1 : 3); two hospitals, USA	n = 739	5 cm < AFI ≤ 10 cm	Not reported	Caesarean section for fetal distress, SGA, meconium-stained amniotic fluid, 5-minute Apgar score of < 7, NICU admission, preterm delivery	BAFI: 38.3 weeks	Nulliparous: not reported
		Low risk
		Exclusion criteria: AFI ≤ 5 cm, PPROM, pre-eclampsia				Normal AFI: 38.9 weeks	IOL: not reported

FIGURE 38.

Deeks’ funnel plot for publication bias for borderline oligohydramnios for the prediction of SGA < 10th centile. Deeks’ funnel plot asymmetry test: p = 0.33. ESS, effective sample size.

MEDLINE and EMBASE

Date range searched: inception to 22 October 2018.

Search strategy

1. exp fetus echography/

2. ultrasonography, prenatal.mp.

3. exp ultrasound/

4. ultraso*.mp.

5. sonograph*.mp.

6. exp biometry/

7. USS.mp.

8. estimated fetal weight.mp.

9. EFW.mp.

10. abdominal circumference.mp.

11. AC.mp.

12. exp macrosomia/

13. macrosomi*.mp.

14. exp fetus weight/

15. fetal weight.mp.

16. exp birth weight/

17. birthweight.mp.

18. large for gestational age.mp.

19. LGA.mp.

20. large fetus.mp.

21. exp brachial plexus injury/or brachial plexus injury.mp.

22. exp shoulder dystocia/or shoulder dystocia.mp.

23. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11

24. 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22

25. 23 and 24

26. exp pregnancy/

27. 25 and 26.

FIGURE 39.

The PRISMA flow diagram for the systematic review of macrosomia.

FIGURE 40.

Risk-of-bias applicability concerns for included studies for systematic review of macrosomia.

TABLE 24 - Characteristics of studies included in the meta-analysis of macrosomia

BPD, biparietal diameter; BW, birthweight; DM, diabetes mellitus; FL, femur length; GDM, gestational diabetes mellitus; HC, head circumference; IUFD, intrauterine fetal death; oGCT, oral glucose challenge test; SROM, spontaneous rupture of membranes; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus; USS, ultrasound scan.

Study (first author and year of publication)	Type of study; setting	Number of total fetuses (LGA fetuses), risk, and selection (all singleton, non-anomalous unless otherwise stated)	Index test (blinding)	Gestational age at ultrasound	Reference standard	Gestational age at delivery	Other comment (inclusion of T1DM, T2DM and GDM)
Aviram 2017102	Retrospective cohort; single hospital, Israel	n = 7996 (1618)	EFW (20 formulas)	Within 1 week from delivery	BW > 90th centile	Mean for LGA group: 39.4 weeks’ gestation, mean for AGA group: 38.3 weeks’ gestation	DM/GDM: included (21% for LGA, 14% for AGA)
		Risk: mixed	Hadlock (AC/FL/BPD)
		Selection: mixed risk, term only. Excluded SGA deliveries, intrapartum and SROM	Hadlock (AC/FL/HC)
			Hadlock (AC/FL/BPD/HC)
			Hadlock (AC/FL)
			Hadlock (AC/BPD)
			Shepard (AC/BPD)
			Threshold: > 90th centile
			Blinded: no
Balsyte 2009103	Retrospective cohort; single hospital, Switzerland	n = 1062 (135)	EFW	Within 1 week from delivery	BW > 4000 g	Mean 39.3 weeks’ gestation	DM/GDM: not reported
			Hadlock (AC/FL/HC)
		Risk: mixed	Threshold: > 4000 g
		Selection: term only	Blinded: no
Benecerraf 1988104	Retrospective cohort; single hospital, Boston, MA, USA	n = 1301 (324)	EFW (Birnholz)	Within 1 week from delivery	BW > 4000 g	Not specified	DM/GDM: included
		Risk: mixed	Threshold: > 4000 g, > 3800 g
		Selection: included all pregnancies apart from breech and multiples	Blinded: no
Ben-Haroush 2007105	Prospective cohort; single hospital, Israel	n = 259 (23)	EFW	Mean 32 weeks’ gestation	BW > 4000 g	Mean 39 weeks’ gestation	DM/GDM: excluded
		Risk: universal	Hadlock (AC/FL/BPD)
		Selection: routine scan. Included SGA. Excluded hypertensives and diabetics	Threshold: > 90th centile
			Blinded: no
Ben-Haroush 2008106	Retrospective cohort; single hospital, Israel	n = 1925 (140)	EFW	Interval from ultrasound scan to delivery 2.5 days	BW > 4000 g	Mean for LGA 40 weeks’ gestation, mean for normal BW 39.4 weeks’ gestation	DM/GDM: excluded
		Risk: mixed	Hadlock (AC/FL)
		Selection: term only	EFW + AFI
			Threshold: EFW > 4000 g, AFI > 95 mm (60th centile)
			Blinded: no
Benson 1991107	Retrospective cohort; Boston, MA, USA	n = 412 (32)	EFW	Within 1 week from delivery	BW > 90th centile	Not specified	DM/GDM: excluded
		Risk: mixed	Hadlock (AC/FL/BPD)
		Selection: not specified. Excluded diabetics	Threshold: > 90th centile
			Blinded: no
Burkhardt 2014108	Retrospective cohort; single hospital, Zurich, Switzerland	n = 12,794	EFW, AC	Within 1 week from delivery	Shoulder dystocia	281 days for shoulder dystocia, 278 days for no shoulder dystocia	DM/GDM: 7.5% for those with shoulder dystocia, 2.7% for those without shoulder dystocia
		Risk: mixed	Hadlock (AC/FL/BPD)
		Selection: all term, with vertex presentation with scan with 7 days	Threshold: > 4000 g, > 4500 g and > 35 cm, > 39 cm
			Blinded: no
Chauhan 2006109	Retrospective cohort; single hospital, Houston, TX, USA	n = 1954 (119)	EFW	Within 4 weeks from delivery; 64% within 7 days from delivery	BW > 90th centile	34% preterm	DM/GDM: included (13%)
		Risk: mixed	Hadlock (AC/FL/BPD)
		Selection: pregnancies undergoing fetal surveillance. Included SGA, hypertensives (22%) and SROM (5%)	Threshold: > 90th centile
			Blinded: no
Chervenak 1989110	Prospective cohort; single hospital, New Jersey, USA	n = 317 (81)	EFW	> 41 weeks’ gestation	BW > 4000 g	Mean 42 ± 0.6 weeks	DM/GDM: excluded
		Risk: low	Hadlock AC/BPD or AC/FL if BPD not available
		Selection: uncomplicated pregnancies after 41 weeks’ gestation	Threshold: > 4000 g
			Blinded: not clear
Cohen 2010111	Retrospective cohort; single hospital, Montréal, QC, Canada	n = 1099 (105)	EFW	On the same day as or next day of delivery	BW > 4000 g	Mean 275.2 days	DM/GDM: included (11.6%)
		Risk: mixed	Hadlock (AC/FL/BPD/HC)
		Selection: only included pregnancies with ultrasound scan on the same or next day as delivery	Threshold: > 90th centile
			Blinded: no
Crimmins 2018112	Retrospective cohort; single hospital, Baltimore, MD, USA	n = 945 (40)	AFG defined as EFW > 90th centile (Hadlock- AC/FL/BPD) or AC > 95th centile	> 34 weeks’ gestation	BW > 4000 g	Not specified	DM/GDM: excluded
		Risk: mixed	Polyhydramnios > 25 cm		Shoulder dystocia
		Selection: all pregnancies > 34 weeks’ gestation with normal oGCT	Threshold: as above		NICU admission
			Blinded: no
Cromi 2007113	Retrospective cohort; two hospitals, Switzerland	n = 1026 (53)	EFW, AC	Within 4 weeks of delivery	BW > 4000 g, BW > 4500 g	> 34 weeks’ gestation; mean 39.2 weeks’ gestation	DM/GDM: included (8.8%)
		Risk: mixed	Hadlock (AC/FL/BPD)	Mean 37.3 weeks’ gestation
		Selection: all singletons > 34 weeks’ gestation with ultrasound scan within 4 weeks of delivery. Excluded SROM	Threshold: > 95th centile
			Blinded: no
De Reu 2008114	Retrospective cohort; single hospital, the Netherlands	n = 3449 (285)	AC	Between 27 and 33 weeks’ gestation	BW > 90th centile, BW > 95th centile	Mean 278.7 days	DM/GDM: excluded
		Risk: universal	Threshold: > 75th/90th/95th centile
		Selection: women with no risk factors or pathology. Did not exclude SGA	Blinded: no
Freire 2010115 (article in Portuguese)	Retrospective cohort; two hospitals, Brazil	n = 114 (8)	EFW	Within 7 days of delivery	BW > 90th centile	15.6% preterm, 84.4% at term	DM/GDM: not reported
		Risk: mixed	Hadlock (AC/FL/BPD/HC)
		Selection: those with ultrasound scan within 7 days of delivery	Threshold: > 90th centile
			Blinded: no
Galvin 2017116 (GENESIS study) (abstract only)	Prospective cohort; large multicentre study, Ireland	n = 2336	EFW (not specified)	Between 39+0 and 40+6 weeks’ gestation	Shoulder dystocia	Not specified	DM/GDM: excluded
		Risk: low	Threshold: 4000 g		NICU admission
		Selection: term, uncomplicated, cephalic only	Blinded: yes
Gilby 2000117	Retrospective cohort; single hospital, FL, USA	n = 1996 (318)	AC	Within 1 week from delivery	BW > 4500 g	> 36 weeks’ gestation, mean not reported	DM/GDM: not reported
		Risk: mixed	Threshold: > 35 cm, > 38 cm
		Selection: all singleton > 36 weeks’ gestation with ultrasound scan within 1 week from delivery	Blinded: no
Hasenoehrl 2006118	Prospective cohort; single hospital, Austria	n = 200 (33)	EFW (Schild)	Mean 39.2 weeks’ gestation	BW > 4000 g	Mean interval 2.0 days	DM/GDM: not reported
		Risk: low	Threshold: > 4000 g
		Selection: included those with ultrasound scan within 1 week. Excluded only fetal anomaly	Blinded: no
Hendrix 2000119	Prospective (RCT); GA, USA	n = 367 (39)	EFW	> 37 weeks’ gestation	BW > 4000 g	Mean 39.1 weeks’ gestation	DM/GDM: not reported
		Risk: low	Hadlock AC/BPD
		Selection: term only	Threshold: > 4000 g
			Blinded: no
Henricks 2003120	Prospective cohort; SC, USA	n = 256 (21)	AC	> 37 weeks’ gestation	BW > 4000 g	Mean 39.1 weeks’ gestation	DM/GDM: not reported
		Risk: universal	Threshold: > 35 cm
		Selection: term only	Blinded: no
Humphries 2002121	Retrospective cohort; SC, USA	n = 238 (29)	EFW	Within 2 weeks of delivery	BW > 4000 g	> 37 weeks’ gestation	DM/GDM: not reported
		Risk: mixed	Combs (AC/FL/FL)
		Selection: term only, with ultrasound scan within 2 weeks	Threshold: > 4000 g
			Blinded: no
Kayem 2009122	Prospective cohort; multiple hospitals, France and Belgium	n = 1689 (124)	AC	Within 10 days of delivery	BW > 4000 g	Median 39 weeks’ gestation	DM/GDM: not reported
		Risk: low	Threshold: > 36.3 cm
		Selection: as part of a prospective cohort for breech. Term only, with ultrasound scan within 10 days of delivery	Blinded: no
Kehl 2011123	Prospective cohort; single hospital, Germany	n = 258 (30)	AC	Within 3 days of delivery	BW > 4000 g	40+5 weeks’ gestation for AC > 36 cm	DM/GDM: not reported
		Risk: universal	Threshold: > 36 cm			39+6 weeks’ gestation for AC < 36 cm
		Selection: term only with vertex presentation and ultrasound scan within 3 days of delivery	Blinded: no
Levine 1992124	Retrospective cohort; single hospital, New York, NY, USA	n = 406 (68)	EFW	5–10 days before delivery	BW > 90th centile	Mean 39.4 weeks	DM/GDM: included (22%)
		Risk: mixed	Hadlock (AC/FL/HC)
		Selection: term only. Included pregnancies with diabetes (22%) and previous caesarean section (20%)	Threshold: > 90th centile
			Blinded: no
Melamed 2011125	Retrospective cohort; single hospital, Israel	n = 4765 (431)	EFW (multiple) and AC	Within 3 days of delivery	BW > 4000 g	Mean 38.1 weeks	DM/GDM: excluded
			Hadlock (AC/FL/BPD)
			Hadlock (AC/FL/HC)
		Risk: mixed	Hadlock (AC/FL/BPD/HC)
			Hadlock (AC/FL)
			Shepard (AC/BPD)
		Selection: all deliveries with ultrasound scan within 3 days of delivery. DM/GDM and SROM excluded	Threshold: > 4000 g, > 36 cm
			Blinded: no
Miller 1986126	Retrospective cohort; single hospital, LO, USA	n = 150 (28)	EFW	Within 7 days of delivery	BW > 4000 g	Term (mean gestational age not reported)	DM/GDM: included
		Risk: mixed	Hadlock (AC/FL)
		Selection: term only, included diabetes, pre-eclampsia, prior caesarean section. Excluded SGA	Shepard (AC/BPD)
			Threshold: > 4000 g
			Blinded: no
Miller 1988127	Retrospective cohort; single hospital, LO, USA	n = 382 (58)	EFW and AC	Within 7 days of delivery	BW > 4000 g	Mean gestational age 279.1 days	DM/GDM: not reported
		Risk: mixed	Hadlock (AC/FL/BPD)
		Selection: term only, excluded SROM	Threshold: EFW > 4100 g, AC > 36.4 cm	Mean gestational age 275.8 days
			Blinded: no
Nahum 2003128	Retrospective cohort; single hospital, CA, USA	n = 74 (12)	EFW (11 formulas)	Within 3 weeks of delivery	BW > 4000 g	Term (mean gestational age not reported)	DM/GDM: included (23.0%)
			Hadlock (AC/FL/BPD)
		Risk: mixed	Hadlock (AC/FL/HC)
			Hadlock (AC/FL/BPD/HC)
			Hadlock (AC/BPD)
		Selection: only included Hispanic ethnicity, term only	Shepard (AC/BPD)
			Threshold: > 4000 g
			Blinded: no
Nahum 2007129	Retrospective cohort; single hospital, CA, USA	n = 98 (16)	EFW	Within 3 weeks of delivery	BW > 4000 g	Term (mean gestational age not reported)	DM/GDM: excluded
			Hadlock (AC/FL/BPD)
		Risk: low risk	Hadlock (AC/BPD)
			Hadlock (AC/FL)
		Selection: term only. Excluded medical complications (pre-eclampsia, DM)	Threshold: > 4000 g
			Blinded: no
Nicod 2012130 (article in French)	Retrospective cohort; single hospital, Switzerland	n = 708 (141)	EFW	Within 7 days of delivery	BW > 4000 g	Not reported	DM/GDM: not reported
			Hadlock (AC/FL/BPD/HC)
		Risk: mixed risk	Hadlock (AC/FL)
			Threshold: > 4000 g
		Selection: pregnancies with ultrasound scan within 7 days of delivery	Blinded: no
O’Reilly-Green 1997131	Retrospective cohort, single hospital; New York, NY, USA	n = 445 (107)	EFW	Within 3 weeks of delivery	BW > 4000 g, BW > 4500 g	Gestational age > 40+4 weeks’ gestation	DM/GDM: excluded
		Risk: low	Hadlock (AC/FL/BPD)
		Selection: prolonged pregnancies defined as gestational age > 40+4 weeks	Threshold: > 4000 g, > 4500 g
			Blinded: no
Pates 2007132	Retrospective cohort; single hospital, TX, USA	n = 3115 (239)	EFW and AFI	Within 7 days of delivery	BW > 4000 g	Not reported	DM/GDM: included (11%)
		Risk: mixed	Hadlock (AC/FL/BPD/HC)
		Selection: those with clinically indicated ultrasound scan within 7 days of delivery	Threshold: > 4000 g, AFI > 20 cm (95th centile)
			Blinded: no
Peregrine 2007133	Prospective cohort; single hospital, London, UK	n = 262 (48)	EFW	Exactly before IOL	BW > 4000 g	Median gestational age 41 weeks’ gestation	DM/GDM: not reported
		Risk: mixed	Hadlock (AC/FL)
		Selection: pregnancies with gestational age > 35+6 weeks undergoing IOL. Excluded those with IUD or antepartum haemorrhage	Shepard (AC/BPD)
			Threshold: > 4000 g
			Blinded: yes
Pollack 1992134	Retrospective cohort; single hospital, New York, NY, USA	n = 519 (119)	EFW	Within 7 days of delivery	BW > 4000 g	> 41 weeks’ gestation	DM/GDM: not reported
		Risk: mixed	Hadlock (AC/FL)
		Selection: postdate pregnancies > 41 weeks’ gestation	Threshold: > 4000 g, > 4500 g
			Blinded: no
Rossavik 1993135	Retrospective cohort; single hospital, OK, USA	n = 498 (36)	EFW	Within 2 weeks of delivery (if gestational age > 38 weeks) or within 1 week of delivery (if gestational age < 38 weeks)	BW > 4000 g	Not reported	DM/GDM: not reported
		Risk: mixed	Hadlock (AC/FL/HC)
		Selection: infants with ultrasound scan within 2 weeks of delivery (if gestational age > 38 weeks) or within 1 week of delivery (if gestational age < 38 weeks)	Threshold: > 4000 g
			Blinded: no
Sapir 2017136 (abstract only)	Retrospective cohort; single hospital, Israel	n = 6214	EFW, AC	Within 1 week of delivery	Shoulder dystocia	Term (not specified)	DM/GDM: excluded
		Risk: mixed	Threshold: > 4000 g, > 4500 g, AC > 39 cm
		Selection: term only; no GDM with scan within 7 days of delivery	Blinded: no
Smith 1997137	Retrospective cohort; single hospital, Glasgow, UK	n = 1213 (16)	EFW and AC	Within 7 days of delivery	BW > 4500 g	Not reported	DM/GDM: excluded
		Risk: mixed	Hadlock (AC/FL)
		Selection: non-diabetic pregnancies with ultrasound scan within 7 days of delivery	Threshold: > 4000 g, > 4500 g, AC > 36 cm, AC > 38 cm
			Blinded: no
Sovio 2018138	Prospective cohort; single hospital, Cambridge, UK	n = 3866 (177)	EFW, ACGV	Regular research scan at 36 weeks’ gestation (median 36.4 weeks’ gestation)	BW > 90th centile, BW > 97th centile	Median 40.4 weeks’ gestation	DM/GDM: included (4.3%)
		Risk: universal	Hadlock (AC/FL/BPD/HC)		BW > 4000 g, BW > 4500 g, shoulder dystocia, neonatal morbidity (composite of metabolic acidosis, 5-minute Apgar score of < 7, NICU admission), severe neonatal morbidity
		Selection: unselected number of nulliparous women who delivered after 36 weeks’ gestation	Threshold: > 90th centile (population/customised)
			Blinded: yes
Sritippayawan 2007139	Prospective cohort; single hospital, Thailand	n = 328 (3)	EFW	> 34 weeks’ gestation, mean interval 16.9 days from delivery	BW > 4000 g	Mean gestational age 39.4 weeks	DM/GDM: excluded
		Risk: low	Hadlock (AC/FL/BPD/HC)
		Selection: pregnancies > 34 weeks’ gestation. Excluded IUFD, any medical complication	Threshold: > 4000 g
			Blinded: no
Sylvestre 2000140	Retrospective cohort; single hospital, New York, NY, USA	n = 656 (147)	EFW (Hadlock or Shepard/not specified)	> 41 weeks’ gestation	BW > 4000 g	41.3 weeks’ gestation	DM/GDM: not reported
		Risk: low	Threshold: > 4000 g
		Selection: postdate pregnancies only (> 41 weeks’ gestation)	Blinded: no
Weiner 2002141	Prospective cohort; single centre, Israel	n = 315 (134)	EFW	Ultrasound scan with 3 days of delivery	BW > 4000 g	40.1 weeks’ gestation for both groups	DM/GDM: included (9.2%)
		Risk: mixed	Shepard (AC/BPD)		BW > 4500 g
		Selection: offered routine clinical screening to all women at term. Those with suspected EFW > 3700 g had ultrasound scan. Only included those with ultrasound scan with 3 days of delivery	Threshold: > 4000 g		Shoulder dystocia
			Blinded: no

FIGURE 41.

Deeks’ funnel plot for publication bias for the prediction of LGA (birthweight > 4000 g or > 90th centile). Deeks’ funnel plot asymmetry test: p = 0.02. ESS, effective sample size.

Beneficial population

An estimate of the total population is required for the VOI analyses, defined as the total population who could benefit from future research that reduces decision uncertainty. The relevant population is all singleton births to nulliparous women in England, excluding those women opting for elective caesarean section for reasons other than breech presentation.

The NHS Maternity Statistics201 states that there were 636,401 births in England in the financial year 2016–17. Of these, 91.8% were at ≥ 37 weeks’ gestation, 33.6% of which were to nulliparous women. 201 The statistics do not disaggregate by reason for elective caesarean section (specifically, whether or not because of suspected breech position). Therefore, this means that there were: