Effect of prenatal screening on trends in perinatal mortality associated with congenital anomalies before and after the introduction of prenatal screening: A population‐based study in the Northern Netherlands

Abstract Background Perinatal mortality in foetuses/children with congenital anomalies remains high. Prenatal diagnosis, essential for risk assessment and organisation of perinatal/postnatal care, offers parents the opportunity to consider the termination of pregnancy. In times of quick changes in prenatal screening programmes, it is relevant to evaluate the effect of prenatal screening on perinatal mortality rates. Objectives The objective of this study was to study trends in early foetal and perinatal mortality associated with congenital anomalies before/after the introduction of the Dutch prenatal screening programme. Methods This population‐based cohort study included 8535 foetuses/neonates with congenital anomalies born in the Northern Netherlands between 2001 and 2017. Total deaths were defined as sum of early foetal (before 24 weeks’ gestation) and perinatal deaths (from 24 weeks’ gestation till day 7 post‐partum). Foetal deaths were categorised into spontaneous or elective termination of pregnancy for foetal anomalies (TOPFA). Trends in total mortality as well as early foetal and perinatal mortality were studied. Joinpoint regression was used to calculate the average annual percentage chance (AAPC) and identify linear trends in mortality within subperiods. Results Total and perinatal mortality were 17% and 4%. Total mortality was higher in abnormal karyotype and central nervous system anomalies. We observed an increase in total mortality over time: 11.9% in 2001 versus 21.9% in 2017 (AAPC 2.6, 95% confidence interval [CI] 1.5, 3.7), caused by an increase in early foetal mortality from 5.5% to 19.2% (AAPC 8.7, 95% CI 4.7, 12.9) and a decrease in perinatal mortality from 6.4% to 2.7% (AAPC −5.6, 95% CI −10.0, −1.0). The increase in early foetal mortality reflects an increase in TOPFA from 3.6% to 16.9% (AAPC 8.3, 95% CI 4.2, 12.7), mostly occurring at 13–14 and 20–23 weeks’ gestation. Conclusions The introduction of the prenatal screening programme led to a decrease in perinatal mortality among foetuses and neonates with congenital anomalies and a marked increase in early foetal mortality before 24 weeks’ gestation due to higher rates of TOPFA.


| BACKG ROU N D
Congenital anomalies account for ~20% of foetal deaths in Europe and are the main contributor to perinatal death in high-income countries. 1,2 For this reason, and given their high psychological and financial burden on individuals, families, healthcare systems and society, congenital anomalies are an important focus of public health.
Congenital anomalies are mostly caused by (a combination of) demographic, genetic or environmental factors, but the aetiology remains unknown in about half of cases. 2,3 Early diagnosis during pregnancy is therefore essential for risk assessment and organisation of perinatal and postnatal care.
In the Netherlands, pregnant women are offered an early scan at about 10 weeks of gestation to establish pregnancy viability, multiplicity/chronicity and to correctly date the pregnancy. At this early stage, foetal anatomical assessment does not take place. In 2007, a nationwide prenatal screening programme was implemented in the Netherlands. 4 This programme included the combined test (CT) for trisomies 21, 13 and 18 and the 20-week anomaly scan for structural anomalies. The 20-week scan is performed following a systematic protocol for foetal anatomical assessment by qualified sonographers who meet the required quality standards. 5 In 2014, non-invasive prenatal testing (NIPT) using cell-free foetal DNA became available to women showing increased risk on the CT or with a previous child with a trisomy. From April 2017, NIPT was offered to all women as first-tier screening in the first trimester of pregnancy. However, while the 20-week anomaly scan is covered by mandatory health insurance, the woman has to pay around 170 euro for the CT and NIPT.
In the Northern Netherlands, the uptake of the 20-week scan (82%) has consistently been higher than the CT (32%) or NIPT (29%). 6 The implementation of and subsequent changes to the national prenatal screening programme have led to an increase in prenatal detection of several kinds of congenital anomalies, including neural tube defects, heart defects and urinary tract anomalies. 7-10 Earlier diagnosis offers the opportunity of timely and appropriate counselling to parents on the prognosis of their child and the various therapeutic options available. This allows parents more time to make an informed decision on the management or continuation of the pregnancy and healthcare professionals to properly organise postnatal care. While most studies have focused on prenatal detection rates of congenital anomalies, data on trends in mortality rates following prenatal screening introduction are less frequently reported in the literature.
Because our hypothesis is that these important changes in the prenatal screening programme are likely to reshape mortality rates, we set out to study trends in early foetal and perinatal mortality associated with congenital anomalies in the Northern Netherlands over the time period in which the national prenatal screening programme was implemented. Secondly, we aimed to study trends in foetal and perinatal mortality associated with the different types of foetal congenital anomalies.

| Study design and setting
This is a population-based cohort study including 8535 foetuses and neonates with congenital anomalies born in the Northern Netherlands between 2001 and 2017. Data were extracted from Eurocat Northern Netherlands (Eurocat NNL), a population-based birth defects registry covering the three Northern provinces of the a marked increase in early foetal mortality before 24 weeks' gestation due to higher rates of TOPFA.

K E Y W O R D S
congenital anomalies, foetal mortality, perinatal mortality, prenatal screening

Study question
Does the introduction of prenatal screening affect foetal and perinatal mortality rates in foetuses and neonates with congenital anomalies?

What is already known
Perinatal mortality rates in foetuses and children with congenital anomalies are high. Prenatal screening leads to earlier diagnosis of congenital anomalies during pregnancy.

What this study adds
The introduction of prenatal screening causes a clear shift from late foetal and neonatal mortality to early foetal mortality and a notable difference in the percentage of foetal and neonatal deaths. A much larger number of terminations of pregnancy for foetal anomalies are performed at lower gestational ages. Foetal mortality rates peak at 13-14 and at 21-23 weeks of gestation, in concomitance with the first-trimester screening and the second-trimester anomaly scan.
Netherlands, which together account for approximately 10% of births in the Netherlands yearly. Eurocat NNL cases are identified by active case ascertainment using multiple sources, including hospital files, prenatal diagnosis reports and post-mortem examinations.
There is no gestational age limit for registration. Terminations of pregnancy due to foetal anomalies (TOPFAs), allowed until 24 weeks of gestation in the Netherlands, and spontaneous foetal deaths are therefore included in the data. The upper age limit for inclusion is 10 years.
Cases are registered in the database after parental consent.
For pregnancies delivered/terminated from 2010 onwards, basic information was registered (anomalies type, way of delivery, time of diagnosis) when parents did not respond after a second request for consent. Information on these so-called 'non-responders' is registered to enable accurate monitoring of the occurrence of congenital anomalies. For each child/foetus, information is recorded on specific congenital anomalies, overall diagnosis, year of birth, pregnancy outcome (live birth, stillbirth, TOPFA, spontaneous foetal death), date of death (if occurred postnatally), gestational age (in weeks) at birth or at TOPFA and time of detection of congenital anomalies (prenatal, at birth, post-partum). If the anomaly was detected prenatally, information on prenatal tests and gestational age at diagnosis was recorded. All congenital anomalies in the database are coded by trained registry staff following EUROCAT

| Cohort
We included all foetuses and neonates with major congenital anomalies born between 2001 and 2017. Cases were categorised according to congenital anomalies the following subgroups: isolated structural anomalies, multiple congenital anomalies (MCA), abnormal karyotype and other syndromes. Isolated structural anomalies were further subdivided into anomalies of the following systems: central nervous, cardiac and circulatory, digestive, urogenital, musculoskeletal/other organs and orofacial clefts. MCA was defined as two or more unrelated congenital anomalies in more than one organ system. The abnormal karyotype group included trisomies 13, 18 and 21, monosomy X and triploidy detectable by QF-PCR.
Other syndromes included teratogenic and genetic syndromes and associations.
For cases registered with parental consent, we included information on maternal and pregnancy characteristics. Because maternal age ≥36 years was previously an indication for direct invasive diagnostic testing or CT free of charge, we divided maternal age into two categories: <36 years and ≥36 years. Ethnicity was defined as Western and non-Western based on the country of birth of the maternal grandparents. Educational level was classified as low (primary school, lower vocational and prevocational education), middle (secondary vocational education, general secondary education and pre-university education) and high (college or university education) based on the self-reported highest educational level achieved.
Plurality was defined as singleton and multiple births and gravidity as primigravida and multigravida.

| Outcomes
For all cases, we defined pregnancy outcomes. We used the Dutch gestational age limit for termination of pregnancy (24 weeks) in the definitions of foetal, neonatal and perinatal mortality (see Table 1).
In addition, we categorised foetal mortality according to whether death occurred by elective TOPFA or spontaneously. Induced delivery after foetal demise was categorised as spontaneous foetal death.

| Statistical analysis
We examined trends in mortality among foetuses with congenital anomalies between 2001 and 2017 for total mortality, perinatal mortality and early foetal mortality (TOPFA/spontaneous). Since the legal gestational age limit for TOPFA in the Netherlands is 24 weeks, we examined trends in mortality using 24 weeks as the cut-off point. anomaly. For maternal and pregnancy characteristics, the following data were missing: information on child's gender and maternal age at birth was missing in <1% of cases, information on pregnancy characteristics was missing in 6%-7% of cases and information on maternal ethnicity and educational level, mainly obtained from a parental questionnaire, was missing in approximately 20% of cases. Since we did not perform trend analyses in relation to these characteristics, we reported those characteristics without adjustments for missing data.

| Ethics approval
This study was performed with anonymised data. Ethical approval by the Medical Ethical Committee was waived due to the retrospective nature of the study and because parental consent was obtained for data registration.

| Mortality according to type of congenital anomaly
The highest total mortality was among cases with abnormal karyotype and isolated anomalies of the central nervous system. For the abnormal karyotype cases, mortality was more frequent before 24 weeks' gestation (54%), and the majority of these early foetal deaths were TOPFAs. Perinatal mortality was highest among cases with isolated anomalies of the central nervous system (12%) and MCA (11%). Most perinatal deaths occurred in the early neonatal period (Table 3).

| Limitations of the data
The lack of national data represents a limitation of the study.
Moreover, it is important to note that the uptake of first-trimester screening in the Northern Netherlands is below the national average (uptake CT: 32% NNL vs. 52% nationally), and this might have resulted in a higher proportion of TOPFAs being performed after 20 weeks of gestation following the second-trimester anomaly scan. 16 Additionally, early foetal mortality due to TOPFAs might be overestimated as some very early losses would not be recognised to be affected by any malformation if they just resulted in a miscarriage before first-trimester screening is performed, whereas they are counted if prenatally diagnosed and terminated. Also, even though the Dutch prenatal screening programme has considerably changed throughout the study period, the number of cases is too small to effectively study the effects of such changes on mortality rates. In addition, the process of case registration and follow-up collection takes time and depends on the moment of diagnosis of congenital anomalies. Indeed, milder anomalies presenting later in childhood might be underreported in the most recent study years. As a consequence, mortality rates might be overestimated for those years.
Thorough monitoring is therefore essential to properly evaluate data registration quality. Finally, although we presented data on maternal educational level, we did not study the effects of deprivation on participating in the prenatal screening programme and on mortality rates.

| Interpretation
The prevalence of congenital anomalies in Europe has been changing in the past decades and is still continuously evolving. While the reasons behind the rise in prevalence of some foetal anomalies still remain unclear, some factors have been recognised as contributors of higher rates of congenital anomalies. For instance, higher prevalence of maternal risk factors, such as obesity and diabetes, has been associated with higher rates of some heart defects. By allowing for earlier diagnosis during pregnancy, prenatal screening is an important contributor to the epidemiology of congenital anomalies. 17 Since TOPFAs account for a non-negligible proportion of all cases affected with congenital anomalies, substantial selection bias is likely to be present when TOPFAs are omitted in studies on risk factors or from mortality rates in evaluation studies. 18,19 The major role of prenatal screening in reshaping the distribution of foetal and neonatal deaths has been previously shown. 20 The increasing trend in early foetal mortality and decreasing trend in perinatal mortality in the Eurocat NNL data were mainly visible in the TA B L E 3 Mortality among cases with congenital anomalies according to type of anomaly (n = 8535, perinatal mortality in grey); Eurocat Northern Netherlands, 2001-2017    rates for congenital anomalies. 26,27 When discussing the contribution of TOPFAs to the redistribution of mortality rates, it is essential to address the broader purposes of prenatal screening. Prenatal screening is designed to determine whether a pregnant woman is at increased risk of carrying a foetus affected by a structural or genetic anomaly and to offer her the possibility of undergoing additional diagnostic testing to obtain a definitive diagnosis. 28 Both prenatal screening and diagnosis are on a voluntary basis and it is up to parents to decide whether they wish to opt for them. Once an anomaly is found, parents are counselled on the prognosis of their child and the therapeutic options, and are given the opportunity to make an informed decision on the management of pregnancy. The rise in total and early foetal mortality rates following the increase in early anomalies is a rapidly developing field that could further contribute to the reduction in mortality rates among children with congenital anomalies. However, only a very small proportion of structural defects, such as some cases of myelomeningocele, diaphragmatic hernia or obstructive uropathy, are eligible for in utero interventions, and the success rate of such procedures is still subject to debate. 33,34 Another suggested strategy to further reduce perinatal mortality is by optimising primary prevention of congenital anomalies, for instance through folic acid supplementation or by minimising environmental risk factors in the pre/periconceptional period. 1,35 Regardless of aetiology, early diagnosis of congenital anomalies by prenatal screening is of key importance. Not only because it allows parents more time to make a well-informed decision about termination or continuation of pregnancy but also because it offers healthcare professionals' essential information for optimising postnatal care.

| CON CLUS ION
The current study demonstrates the striking influence of prenatal screening on total mortality and the pattern of foetal and neonatal mortality rates. Earlier detection of congenital anomalies resulted in a decrease in early neonatal deaths and a marked increase in early foetal deaths, which is explained by a much higher proportion of TOPFAs. Given the often-poor prognosis of severe congenital