Prevalence of anomalies on the routine mid‐trimester ultrasound: 3172 consecutive cases by a single maternal–fetal medicine specialist

Abstract

Introduction/Purpose

The routine mid‐trimester fetal anatomy ultrasound (FAS) is offered to every pregnant woman and remains critical in the detection of structural fetal anomalies. Our study aimed to determine the prevalence of abnormalities on routine FAS performed by a single operator, who is an experienced sub‐specialist in maternal–fetal medicine.

Methods

A retrospective analysis of all routine FAS performed a tertiary private obstetric ultrasound practice in metropolitan Sydney over a 7‐year period, August 2015–July 2022. An advanced ultrasound protocol including detailed cardiac views was used in every case. Second opinion scans for suspected abnormalities were excluded. Fetal anomalies were classified into major and minor, based on the likely need for neonatal intervention.

Results

Among 14,908 obstetric ultrasound examinations, routine FAS were performed on 3172 fetuses by a single operator. More than 99% of women had screened low‐risk for fetal aneuploidy. Structural anomalies were identified in 5% (157/3172) of fetuses; the prevalence of major anomalies was 1% (30/3172). Almost 60% of total anomalies were either cardiac or renal. No differences were identified in anomaly rates for singletons compared with twins (5.0% vs. 4.2%; P = 0.75). The prevalence of placenta previa and vasa previa was 10% and 0.1%, respectively.

Discussion

The prevalence of fetal anomalies on routine FAS by a single operator using a standardised protocol was higher in our practice (5%) than in previously published studies. Although most anomalies were minor, the rate of major abnormality was 1%.

Conclusion

The routine mid‐trimester FAS remains an integral component of prenatal ultrasound screening.

Introduction

Recent years have seen significant advancements in the first trimester ultrasound assessment of fetal anatomy. ¹ , ² Despite this, the mid‐trimester fetal anatomy scan (FAS, or ‘morphology scan’) remains critical to the detection of structural fetal anomalies in pregnancy; several fetal systems, in particular the heart, spine, posterior fossa and face, often cannot be thoroughly assessed until 18–22 weeks. ³ The importance of the FAS is underscored by expert guidelines from Australia, the UK, the United States and Canada, all of which recommend that it be routinely offered to every pregnant women. ⁴ , ⁵ , ⁶ , ⁷ In Australia, minimum standards for the performance and reporting of the FAS are set out by the Australasian Society for Ultrasound in Medicine (ASUM) and International Society of Ultrasound in Obstetrics and Gynaecology (ISUOG). ⁸ , ⁹ Overseas, expert bodies in the United States, the American Institute of Ultrasound in Medicine (AIUM) and the Society for Maternal–Fetal Medicine (SMFM) have also published guidelines for a more detailed (advanced) mid‐trimester scan. ¹⁰

The sensitivity of the FAS in detecting structural fetal anomalies is determined primarily by operator experience and by the ultrasound protocol used. Because ultrasound is a real‐time, user‐dependant modality, previous studies have, unsurprisingly, demonstrated a significant association between operator experience and detection rate of abnormalities for both mid‐trimester and early fetal ultrasounds. ¹¹ , ¹² , ¹³ , ¹⁴ Most published series include operators with a wide range of clinical ultrasound experience.

All ultrasounds in our practice are performed using a standardised advanced ultrasound protocol by a single operator, who is a highly experienced sub‐specialist in maternal–fetal medicine (MFM). We undertook a retrospective analysis of all the routine FAS performed at our tertiary private obstetric ultrasound practice over a 7‐year study period, to determine the prevalence of ultrasound abnormalities in an Australian population in contemporary practice.

Materials and methods

We performed a retrospective analysis of consecutive routine mid‐trimester fetal anatomy ultrasound scans (FAS) performed at our practice over a 7 year period, 1 August 2015 to 31 July 2022. Ours is a private, tertiary‐level, obstetric ultrasound practice in metropolitan Sydney, Australia. Every ultrasound examination is performed personally by the first author, who is a certified sub‐specialist in MFM by the Royal College of Obstetricians and Gynaecologists (London, UK) and who has performed more than 20,000 referred obstetric ultrasound examinations over the past 10 years.

Ultrasound examinations in our practice are coded contemporaneously using an Australian Medicare Benefits Schedule (MBS) item number, based on the indication for referral. A computerised database of ultrasound data collected prospectively (ViewPoint 6; GE Healthcare, Chicago, Illinois, USA) was searched by MBS item number to identify scans meeting inclusion criteria. Only routine mid‐trimester FAS (MBS 55706 and 55759) were included in the present study, and both singleton and multiple pregnancies were included. Patients referred for a second opinion – after an abnormal finding on the routine FAS – and high‐risk patients referred for a fetal echocardiogram after a normal FAS (MBS 55712 and 55764) were excluded. In cases where additional demographic details were required, additional patient data were retrieved from computerised medical records (Genie version 10.1.2) on an individual basis.

All ultrasound examinations included in the present study were performed on a Medicare‐eligible Voluson™ E8 machine (GE Healthcare, Chicago, Illinois, USA). As ours is a tertiary MFM‐led ultrasound practice, our protocol is to perform an advanced mid‐trimester fetal anatomy scan in every case, even in low‐risk patients. FAS were performed consistent with the components of the detailed FAS as set out in the 2014 ASUM guidelines (subsequently updated in 2018). ⁹ All fetal long bones were examined, but only the femur length was measured routinely. An extended cardiac scan was performed in every case. ⁸ The cardiac assessment incorporated the four‐chamber view, LVOT, RVOT, three‐vessel view, three‐vessel and trachea view, aortic arch view, colour flow Doppler, fetal heart rate and rhythm, as well as venous assessment (bicaval view and pulmonary venous return).

For findings considered ‘soft markers’ for aneuploidy, in women with prior low‐risk first‐trimester aneuploidy screening, isolated intracardiac echogenic focus and choroid plexus cysts were not recorded. ¹⁵ The nasal bone and mid‐trimester nuchal fold were assessed subjectively but only formally measured in women who had previously declined aneuploidy screening or who had declined invasive testing following a high‐risk result. ¹⁵ Markers with implications for fetal growth – echogenic bowel and two‐vessel cord – and markers that might be associated with other structural anomalies – renal pelvic dilatation (AP diameter > 4 mm) ¹⁶ or ventriculomegaly (Vp ≥ 10 mm) – were always recorded.

Structural fetal abnormalities were categorised into major and minor abnormalities. We classified abnormalities as major if they had the potential to be life‐threatening and/or were likely to require a significant surgical intervention in the neonatal or childhood period. All remaining structural anomalies were classified as minor. In cases of multiple major fetal abnormalities, the most clinically meaningful abnormality was selected for categorisation purposes to avoid duplication.

The placental structure and location were always assessed. The placenta was considered low‐lying/previa if the leading placental edge was <20 mm from the internal cervical os. ¹⁷ The placental cord insertion site was not routinely assessed, consistent with ISUOG guidelines. ⁸ Maternal adnexa were not routinely examined unless the patient reported symptoms suggestive of an adnexal mass. Screening for vasa previa was performed selectively, depending on the placental site. Amniotic fluid volume was measured subjectively and, where indicated, using the deepest vertical pocket.

Cervical length screening was performed in all cases; this was typically transabdominal (TA), with transvaginal assessment reserved for cervical lengths <30 mm, poor visualisation transabdominally or a high‐risk history including prior preterm birth or cervical surgery. TA cervical length <30 mm represents the lowest 10th centile in a healthy low‐risk population. ¹⁸ Uterine fibroids were recorded at the time of the ultrasound examination but are not included in the present study due to their high prevalence. Similarly, apparent Mullerian uterine anomalies were included in the ultrasound report but are excluded from the present analysis as it was not possible to ascertain which represented clinically novel findings. Uterine synechiae, which may have implications for fetal presentation, were recorded.

Statistical analysis was performed using Statsdirect statistical package version 3.0 (Statsdirect Ltd., Birkenhead, UK). Categorical data were compared using Fisher's exact test. Two‐tailed P values were used, and the 5% level was considered significant. Ethics approval was granted by the Ramsay Health Care NSW/VIC Human Research Ethics Committee (HREC Reference Number: 2022/ETH/0127).

Results

In our ultrasound practice, 14,908 referred ultrasound examinations were performed by a single operator during the 7‐year study period, 1 August 2015–31 July 2022. Of these, 3424 were mid‐trimester fetal anomaly scans. Excluding women with a suspected fetal anomaly referred for a second opinion and women referred for a fetal echocardiogram after a normal mid‐trimester FAS, the routine FAS were performed 3052 pregnancies (3172 fetuses, Figure 1). The median gestation at the time of FAS was 19⁺³ (range 17⁺⁰ to 24⁺⁰) weeks. The median maternal age was 34 (range 19–52) years old. By plurality, 2932 routine FAS were performed in singletons and 120 in twin pregnancies (78 dichorionic, 40 monochorionic‐diamniotic, 2 monochorionic‐monoamniotic). Overall, 389 women underwent the routine FAS in two separate pregnancies and 15 women underwent the routine FAS in three separate pregnancies during the study period in our practice, for a total of 2648 individual women included in the present study.

Figure 1.

Patient flow chart during study. Abbreviations: CNS, Central Nervous System; GI, Gastrointestinal; MBS, Medicare Benefits Schedule.

In total, the routine FAS were performed on 3172 fetuses in the study period (Figure 1). Structural anomalies were identified in 5% (157/3172) of fetuses. The prevalence of major structural fetal anomalies was 1% (30/3172). No differences were identified in anomaly rates for singletons compared with twins (5.0% [147/2932] vs. 4.2% [10/240]; P = 0.75). Of 3052 pregnancies included in the present study, formal first trimester aneuploidy screening had been declined in only 0.6% (n = 19). The remaining pregnancies (99.4%, n = 3033) had all screened low risk on either combined first trimester screening or non‐invasive fetal DNA screening, which was typically followed by an early fetal structural scan at 12–14 weeks.

Structural fetal anomalies

Renal and cardiac anomalies were the most common abnormalities seen, accounting for 33% and 26% of total fetal anomalies, respectively. The overall prevalence of any renal and any cardiac issue was 1.6% and 1.3% of fetuses, respectively (Table 1). Among the fetuses classified as having a major structural defect, 27% (8/30) were cardiac, 27% (8/30) were neurological and 10% (3/30) were skeletal (Table 1). For individual anomalies, renal pelvic dilatation (1.3% of fetuses) and ventricular septal defect (0.4%) were most prevalent. Among other systems, the most frequently identified anomalies were echogenic bowel (0.4%), talipes equinovarus (0.2%), hemivertebra (0.1%), agenesis of the corpus callosum (0.1%) and congenital pulmonary airway malformation (0.1%). Rates per fetus for the most common fetal anomalies are included in Table 2 to facilitate patient counselling.

Table 1.

Individual structural fetal anomalies.

Isolated structural fetal anomaly	N
Renal anomalies	52 (16 per 1000)
Renal pelvic dilatation (>4 mm)	40
Duplex kidney	6
Unilateral renal agenesis	3
Bilateral renal agenesis	1
Bilateral cystic renal dysplasia	1
Pelvic kidney	1
Cardiac anomalies	42 (13 per 1000)
Ventricular septal defect (VSD)	12
Tricuspid regurgitation (mild–moderate)	10
Aberrant right subclavian artery (ARSA)	7
Tetralogy of fallot/double‐outlet right ventricle	2
Hypoplastic left heart syndrome (HLHS) a	2
Persistent left superior vena cava (SVC)	2
Atrioventricular septal defect (AVSD)	1
Pulmonary stenosis	1
Aortic stenosis	1
Cardiomyopathy	1
Isolated right aortic arch	1
Right‐axis deviation	1
Mitral regurgitation	1
CNS anomalies	8 (2.5 per 1000)
Dandy–Walker malformation/variant a	2
Ventriculomegaly (≥10 mm)	2
Agenesis of corpus callosum	3
Spina bifida	1
Skeletal anomalies	11 (3.5 per 1000)
Talipes equinovarus	7
Hemi‐vertebra/scoliosis	3
Skeletal dysplasia	1
Gastrointestinal anomalies	19 (6 per 1000)
Echogenic bowel	13
Echogenic stomach/liver foci	4
Abdominal cyst	1
Fetal intra‐umbilical varix	1
Other anomalies	25 (8 per 1000)
Isolated two‐vessel cord	17
Congenital pulmonary airway malformation	3
Cleft lip +/− palate	2
Congenital diaphragmatic hernia	1
Hypospadias	1
Non‐immune hydrops/cystic hygroma	1
Total	157 (50 per 1000)

Major anomalies highlighted in bold.

^a Primary anomaly in a case of multiple fetal abnormalities.

Table 2.

Patient counselling rates for the most common abnormal findings on the routine mid‐trimester anatomy ultrasound.

Structural fetal anomaly	Rate per fetus
Renal pelvic dilatation	1‐in‐80
Isolated two‐vessel cord	1‐in‐200
Echogenic bowel	1‐in‐250
Ventricular septal defect	1‐in‐250
Talipes equinovarus	1‐in‐400
Duplex kidney	1‐in‐500
Hemi‐vertebra	1‐in‐1000
Unilateral renal agenesis	1‐in‐1000
Agenesis of the corpus callosum	1‐in‐1000
Congenital pulmonary airway malformation	1‐in‐1000

Maternal anomaly	Rate per pregnancy
Low‐lying placenta at 18–20 weeks	1‐in‐10
Asymptomatic short cervix	1‐in‐250
Vasa previa	1‐in‐1000

The overall prevalence of an isolated two‐vessel cord (single umbilical artery) in our population was 0.5% (17/3172). Two additional cases of two‐vessel cord, one as part of multiple fetal anomalies (see below) and the other in association with a persistent left SVC, are not included. The prevalence of two‐vessel cord was not significantly higher in twins (1.3% [3/240]) than in singletons (0.5% [14/2932]; P = 0.13), although analysis is limited by the relatively low number of twin pregnancies in the present study.

Maternal and placental anomalies

Abnormal maternal/placental findings were found in 11% (325/3052) of pregnancies. Of the mid‐trimester FAS performed during the study period, the prevalence of placenta previa / low‐lying placenta was 10% (304/3052). The prevalence of vasa previa identified on the routine FAS was 0.1% (3/3052). Chorangiomas were identified in 0.3% of placentas (9/3052), and all were small (<4 cm). Similarly, the rate of uterine synechiae detected on the routine FAS was 0.3% (9/3052). Sonographic cervical length was assessed in all cases; the prevalence of an asymptomatic short cervix (≤25 mm) in our population was 0.4% (11/3052).

Multiple fetal anomalies

Over the 7‐year study period, in a cohort with a very high (>99%) uptake of first trimester aneuploidy screening, there were only three cases (0.1%) of multiple major fetal anomalies diagnosed at the routine FAS. There was one case of Dandy–Walker malformation, hypoplastic left heart syndrome and polydactyly in a woman following low‐risk cell‐free fetal DNA screening, who had declined an early fetal anatomy scan, where amniocentesis revealed multiple pathogenic copy number variants. In the second case, cell‐free fetal DNA screening was low risk and a 14‐week early ultrasound showed a two‐vessel cord. At 19 weeks, there was evidence of a Dandy–Walker variant and a double‐outlet right ventricle. Prenatal microarray was normal on amniocentesis, and the patient opted for termination. In the final case, there was a huge fetal cystic hygroma and hydrops in a patient who had declined all prior screening. Amniocentesis was declined, and the patient was referred to the local fetal medicine unit for ongoing care at her request.

Discussion

The mid‐trimester FAS has been an integral component of routine pregnancy care for several decades now. More recently, prenatal diagnosis of structural fetal anomalies has mostly focused on the improving detection rates at first‐trimester ultrasound, with sensitivities of 30%–60% reported prior to 14 weeks of gestation. ¹ , ² However, as the recent large meta‐analysis by Karim et al. points out, ‘there are some conditions…that are never identifiable….thus, first‐trimester anomaly screening will not replace examinations at later gestational ages completely’. ¹ In the present study, we identified a fetal structural anomaly in 5% of routine mid‐trimester FAS; in 1% of cases, this was a major anomaly. Such high prevalence rates certainly underscore the ongoing importance of a detailed mid‐trimester anatomy ultrasound for all pregnancies.

The sensitivity of the FAS in detecting fetal anomalies is determined primarily by operator experience and by the ultrasound protocol used as well as technical factors such as fetal position, maternal bladder filling, maternal size and ultrasound equipment. The probable impact of operator experience is acknowledged in many published studies but rarely elaborated upon. A 2016 Swedish study, which found that only 40% of structural anomalies were detected prior to 22 weeks of gestation, included clinicians with ultrasound experience ranging from 1 month to 17 years. The influence of operator experience was raised but not analysed. ¹⁹ The large Eurofetus study reported in 1999 made no mention of operator experience. ²⁰ In the 2017 meta‐analysis by Karim et al., the authors report that ‘an initial attempt was made to collect data with respect to the level of experience and training of sonographers in each study. However, the majority of studies reported no such data…therefore this was abandoned’. ¹ One rare study did relate sonographer experience to the detection of congenital heart disease; not surprisingly, it found a significant difference in detection by experienced operators (52%) compared to less experienced (33%). ¹² Given this previous work, a major strength of the present study is that every mid‐trimester ultrasound was performed by a single operator, with no potential for inter‐observer variability. Additionally, the operator was a sub‐specialist in fetal medicine with considerable experience in obstetric ultrasound, including fetal echocardiography. In contrast, such high detection rates may not be achievable in those obstetric ultrasound practices in which scans are performed by multiple operators with a wide range of ultrasound experience.

An approximate 2%–3% incidence of fetal abnormality on the routine FAS is often cited in clinical practice. Several large older studies substantiate this estimate, with published prevalence rates of 1.6%–3.5%. ²¹ , ²² , ²³ In the present study, prevalence rates for any structural and major structural anomalies were 5% and 1%, respectively (Table 1). The relatively high prevalence of minor structural fetal anomalies is likely due to three factors. Firstly, subtle anomalies such as aberrant right subclavian artery were never reported in older studies, but improvements in technology and expertise permit their detection in contemporary practice. Secondly, many soft markers, such as pyelectasis and echogenic bowel, were considered too trivial to include in previous studies. ²² , ²³ However, we believe that echogenic bowel warrants inclusion as the risk of stillbirth is increased almost 10‐fold, even when adjusted for CMV and aneuploidy. ²⁴ Lastly, the classification of fetal anomalies into ‘major’ and ‘minor’ is somewhat arbitrary. Sometimes, the distinction is self‐evident, while other anomalies can be more challenging to classify. Talipes, ventricular septal defects (VSD) and renal pelvic dilatation, for instance, could credibly be included in the sub‐group of major fetal anomalies.

The Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG) recommends that prenatal screening for chromosomal conditions should be discussed and offered to all women; it further recommends that acceptable first‐line screening tests include either combined first‐trimester screening or cell‐free fetal DNA screening. ²⁵ In our cohort of more than 3000 pregnancies, >99% had screened low risk for fetal aneuploidy in the first trimester. The majority of women opting for cell‐free fetal DNA screening also underwent an early structural ultrasound at 12–14 weeks. Therefore, most women presenting for the routine mid‐trimester anatomy ultrasound in our practice had undergone a first‐trimester structural scan, which may explain why the rate of major structural fetal anomalies in our low‐risk cohort – 1% – seems low in comparison to previously published studies. It is notable that no cases of anencephaly, holoprosencephaly, abdominal wall defect, limb reduction defect, body stalk anomaly or lower urinary tract obstruction were diagnosed on the routine FAS in our population throughout the study period. With advancements in first‐trimester fetal anatomy scanning, an increasing proportion of major anomalies will inevitably be diagnosed earlier, with a consequent fall in major anomalies being detected for the first time at 18–22 weeks. As such, future work should focus on those fetal anomalies, which are only amenable to detection in the second trimester.

We acknowledge certain limitations with our study. It is a retrospective analysis, albeit of data that were collected prospectively and in a contemporaneous fashion. Given that ours is a private obstetric ultrasound practice with a wide geographical referral base and affiliations to multiple tertiary maternity units, it was not practical or possible to accurately collate data on neonatal outcomes for every case included in our 7‐year study. Previous studies have shown that the specificity for anomalies detected on the mid‐trimester scan is very high. ²¹ Thus, we believe that the false‐positive rate for fetal anomalies included in our study is likely very low. However, the absence of neonatal outcome data means we cannot determine sensitivity rates in our population. Nevertheless, we believe that this study provides clinically useful information for women presenting for the routine mid‐trimester fetal anatomy scan, in a contemporary Australian population, using an advanced ultrasound protocol by a single expert operator.

Conclusion

There is a 5% prevalence of fetal anomalies on the routine mid‐trimester FAS performed by an expert operator. Although most anomalies are minor, the rate of major abnormality is 1%.

Conflict of interest

None.