Impact of National Screening Programs on Down syndrome prevalence and outcomes

Authors: Cecilie Bryn Nordklev, Denise Reis Costa, Roy Miodini Nilsen, Nils‐Halvdan Morken, Siri Vangen, Vasilis Sitras

Abstract

Introduction

Over the past five decades, antenatal screening programs have evolved significantly, primarily through advancements in ultrasound technology and the shift from invasive to non‐invasive prenatal testing methods. This study aimed to evaluate the impact of these developments in prenatal screening for Down syndrome (DS) on the prevalence of live and stillbirths, as well as terminations of pregnancy (TOP) with DS.

Material and Methods

This population‐based registry study included all pregnancies in Norway from 1967 to 2021, with data on TOPs available from 1999. Information on DS pregnancies was obtained from the Medical Birth Registry of Norway. Logistic regression models were applied to evaluate time trends and maternal characteristics associated with TOP.

Results

Among 3 231 159 pregnancies, 4764 (0.147%) were affected by DS. The prevalence of DS pregnancies increased from 0.165% in 1999 to 0.251% in 2021. During this period, the proportion of TOP rose from 20% to 55%, while prenatal detection rates improved from 18% to 70%. Despite increased detection, the proportion of TOP following prenatal diagnosis remained stable at approximately 80%. The prevalence of live births with DS remained stable at approximately 0.106%, while the prevalence of stillbirths was around 0.008%.

Maternal age, parity, country of birth, and region of residence were associated with TOP decisions. Mean gestational age at termination decreased from 17.5 to 15.7 weeks.

Conclusions

Over recent decades, prenatal detection and termination of DS pregnancies have increased. However, the proportion of women continuing pregnancies after a prenatal diagnosis and the prevalence of live‐ and stillbirths with DS have remained stable.

Key messageOver the past two decades, pregnancies with Down syndrome, prenatal diagnosis, and subsequent terminations have increased. Yet, the proportion continuing pregnancy after diagnosis and the prevalence of live‐ and stillbirths with Down syndrome have remained stable throughout the same period.

INTRODUCTION

Down syndrome (DS) is the most common chromosomal disorder found in live born infants with a prevalence rate of 0.16%. ^1^ It is linked to cognitive impairment along with various physical health issues. ^2^ Although the number of pregnancy terminations (TOP) following a DS diagnosis has risen, the rate of live‐born babies with DS in Europe has remained stable. ^3^ In contrast, nations such as France, Spain, and Denmark have experienced a marked decline in the prevalence of live‐born infants with DS, with reported rates decreasing to approximately 0.05%. This reduction is attributed to extensive national antenatal screening programs. ^1^ , ^4^

Since 1986, the Norwegian antenatal screening program has included a single second‐trimester ultrasound examination, adhering to international standards. ^5^ , ^6^ In 2004, updates of the Norwegian Biotechnology Act ^7^ introduced expanded prenatal screening options for women aged 38 years and older, including combined first‐trimester screening (cFTS) for trisomies. ^8^ In 2018, non‐invasive prenatal testing (NIPT) using cell‐free fetal DNA obtained from maternal blood was implemented as a contingency test to cFTS for this age group. ^9^ These policy changes significantly increased prenatal detection rates and led to higher termination rates of pregnancies with DS. ^10^

According to the Norwegian Abortion Act, TOP is permitted at the woman's request up to the 12th week of gestation. ^11^ Such procedures must be carried out exclusively in public hospitals. Given Norway's universal healthcare system, all pregnancy terminations are provided free of charge. Between the 13th and 22nd weeks of pregnancy, termination requests require approval by an abortion commission, typically composed of two physicians. Furthermore, terminations performed after 12 weeks, as well as all stillbirths and live births, are systematically recorded in the Medical Birth Registry of Norway (MBRN).

We aimed to study the prevalence and time trends in live‐ and stillbirths and TOP with DS in the period 1967–2021, reflecting the impact of evolving national prenatal screening programs. A secondary objective was to examine maternal characteristics associated with the decision to terminate a pregnancy following a prenatal diagnosis of DS.

MATERIAL AND METHODS

Study population and design

This study was based on data from the national population‐based MBRN that includes data on all pregnancies from 1967 through 2021. Reporting to the MBRN has been mandatory for all births in Norway since 1967. This includes live births and stillbirths from the 16th week of gestation. From 1999 onwards, it also includes terminated pregnancies, and since 2002, all miscarriages from the 12th week of gestation have been recorded. ^12^ Information is obtained from antenatal health cards and hospital medical records, including information on the mother's health before and during pregnancy, the course of delivery, and the health of the child upon discharge from the hospital. All antenatal care in Norway is standardized and free of charge for residents. ^13^ All pregnancies with known DS were included.

Variables

Year was categorized according to the periods of prenatal screening programs, 1967–1985 (no screening program), 1986–2003 (second trimester screening program), 2004–2017 (cFTS for high‐risk pregnancies) and 2018–2021 (NIPT as contingency to cFTS). Women's age was categorized as ≤29, 30–37 or ≥38 years old. The number of previous childbirths after pregnancy week 22 was coded 0, 1 or ≥2. Marital status was categorized as married/cohabiting or unmarried/single/divorced/widow. Region of residence was defined according to Norway's health South‐Eastern, Western, Central and Northern. Maternal country of birth was classified in three Norway, high‐income countries or other countries, according to Global Burden of Disease (GBD) super‐regions. ^14^ For comparative purposes, live births and stillbirths were aggregated into one outcome category and contrasted with pregnancies resulting in TOP.

Statistical analyses

Descriptive statistics regarding categorical variables were presented as actual counts and percentages, and continuous variables as means and standard deviations (SD).

Time trend analyses in DS prevalence were explored using simple binary logistic regression models, including year of childbirth as a continuous model term. Changes in DS prevalence by year of childbirth were quantified as prevalence OR with 95% CI. Given that the registration of TOP started in 1999, our trend analyses for TOP and overall DS prevalence were restricted to the period from 1999 to 2021. In contrast, trend analyses for live‐ and stillbirths used data available across the entire period from 1967 to 2021.

For live‐ and stillbirth analyses, we used segmented logistic regression techniques due to the observation of multiple inflection points. We hypothesized that these inflection points reflected the impacts of the major screening programs introduced in 1986 and 2004, and we used these years as starting values in the segmented regression analyses. In our analyses of TOP and total DS, we did not observe any inflection points; therefore, we used standard logistic regression for these analyses. In all trend analyses, year of childbirth was included as a continuous model term and changes in DS prevalence by year of childbirth were reported as prevalence OR with 95% CI. Visualizations of prevalence trend lines with 95% CI for all DS outcomes were obtained from the fitted regression models.

In the sub‐period from 1999 to 2021, we further investigated potential factors influencing TOP prevalence among pregnancies with DS. This analysis was conducted using logistic regression models that included relevant risk factors as either continuous or categorical variables, as appropriate. The effects of these risk factors on TOP prevalence were reported as odds ratios (ORs) with 95% confidence intervals (CIs), derived from both crude and adjusted logistic regression models. Covariate adjustments for each risk factor were evaluated using directed acyclic graphs (DAGs), and adjustments were made only if the covariate did not lie within the causal pathway. These data were analyzed using Stata SE 18 (StataCorp. 2023. Stata Statistical Software: Release 18. College Station, TX: StataCorp LLC).

We also conducted descriptive analyses to quantify sample characteristics during the period 1999–2021. For each year of childbirth, we further described the DS distribution of TOP vs. live‐ and stillbirth, the prevalence of prenatal diagnostics for DS, and the prevalence of TOP following DS prenatal diagnosis. Additionally, we calculated the mean gestational age at TOP by year of childbirth. Time‐trend analyses and visualizations were done using R version 4.3.2 for Windows. Segmented regression analyses were performed using the R package “segmented”. ^15^

RESULTS

A total of 3 231 159 pregnancies were registered in Norway from 1967 to 2021. Among these, 4764 (0.147%) had DS. Maternal characteristics of the pregnancies with DS are presented in Table 1. Maternal age above 35 years has increased the last 40 years, as shown in Figure 1. The prevalence of live birth with DS was relatively stable until 1993 (~0.095%), followed by an increase from 1994 to 2003 (0.135%), and then a decline from 2004 to 2021 (0.107%). The prevalence of stillbirth with DS has increased from 1967 (0.005%) to 1995 (0.017%), followed by a decline from 1996 to 2008 (0.004%), and increasing from 2009 to 2021 (0.008%). From 1999, when TOP was first recorded in the MBRN, through 2021, a total of 2791 pregnancies with DS were identified among 1 363 049 pregnancies, corresponding to a total prevalence of 0.205%. Over this period, the total prevalence of DS increased from 0.165% in 1999 to 0.251% in 2021. In parallel, the prevalence of TOP following a prenatal diagnosis of DS among all pregnancies rose from 0.037% to 0.143% (Figure 2). Details of ORs and corresponding 95% CI for each outcome are provided in Table S1.

Among all DS pregnancies between 1999 and 2021, the proportion resulting in termination of pregnancy (TOP) increased from 20% to 55% (Figure 3A). During the same period, the prenatal detection rate of DS rose steadily from 18% to nearly 70% by the end of the study. Despite this increase in detection, the proportion of TOPs following a prenatal diagnosis remained stable throughout the study period, with an average of approximately 80% (Figure 3B).

When comparing pregnancies resulting in live or stillbirths with TOP, we found that women older than 30 years were more likely to terminate the pregnancy. In contrast, multiparous women, women residing in the North health region, and women born outside Norway or other high‐income countries were less likely to terminate a pregnancy with DS (Table 2).

Notably, gestational age at termination for pregnancies with DS was reduced from 17.5 (SD: 1.80) to 15.7 (SD: 2.56) gestational weeks during the study period (Figure 3C).

DISCUSSION

Over the past two decades, the prevalence of pregnancies with DS, prenatal diagnosis of DS and subsequent TOP have increased. However, the proportion of women who continue their pregnancy following a prenatal diagnosis, as well as the prevalence of live‐ and stillbirths with DS, have remained stable over time.

The total prevalence of DS in Norway during the period 1967–2021 was 0.147% of all pregnancies, with an increasing trend to 0.251% at the end of the study period. This is consistent with other European countries, reporting a prevalence of 0.22%, and is related to increasing maternal age the last decades (Figure 1). ^16^ , ^17^ Interestingly, we found a significant increase in the prevalence of DS from 1994 (Figure 2 and Table S1), which coincides with the seminal publication of Nicolaides et al. on ultrasonographically detectable markers of fetal chromosomal abnormalities. ^18^ We hypothesize that increased global awareness among ultrasonographers from the mid‐1990s, including those in Norway, has contributed to improved prenatal detection rates and has significantly influenced the transition from invasive to non‐invasive prenatal screening strategies. However, national screening policies and regulations have varied greatly across Europe, leading to increased rates of prenatal diagnosis, which has also influenced the live‐birth prevalence of DS over time. ^4^ , ^19^ In fact, until 2004 only older women and those with previous pregnancies with trisomies or fetal malformations were offered invasive testing in Norway. Therefore, the low detection rate might be explained by screening only high‐risk pregnancies and women declining invasive testing due to risk of miscarriage. ^20^ , ^21^ In fact, we found an increase in prenatal detection from 2004 when cFTS was offered in Norway (Figure 3B). However, there has been a stable number of live births with DS (Figure 2), with a slight decrease after 2004. These trends are comparable to those observed in the Netherlands, ^17^ which has a similarly structured healthcare system and a national prenatal screening program akin to Norway's. Comparable patterns have also been noted in other Northern and Western European countries, though variations in screening policies may account for differences in prenatal diagnosis and birth prevalence of DS. ^4^ In particular, in Denmark, prenatal screening is considered a routine part of health care for both expectant mothers and stakeholders and is implemented using an opt‐out system. By contrast, in the Netherlands, counselors emphasize that prenatal screening should be a personal, well‐informed choice; many women abstain from screening because they do not consider a diagnosis of DS to be grounds for terminating a pregnancy. ^22^ , ^23^ Interestingly, in Japan, only about 20% of pregnancies undergo prenatal genetic diagnosis, and among cases detected prenatally the live‐birth rate of DS is as low as 2.6%. ^24^ , ^25^ These cultural differences and variations in screening practices contribute to international disparities in the birth prevalence of DS.

Importantly, abortion laws, along with societal values and support provided to families raising children with DS, vary significantly between countries. ^26^ We found a significant increase in TOP due to DS in the last 20 years in Norway (Figure 2), reflecting the implementation of a national prenatal screening program for trisomies from 2004 onwards. On the extreme scale, some European countries, such as Denmark, Island, Portugal, Italy and some eastern European countries reached almost a 100% termination rate for pregnancies with DS. ^27^ On the contrary, we found that the proportion of women opting to continue pregnancy after a prenatal diagnosis of DS in Norway remained unchanged, around 20% in the past years. This is also reported in several studies where the termination rates following the detection of DS by NIPT remained unchanged or decreased compared to historical termination rates. ^28^ This might indicate that many women opt for prenatal diagnosis to be prepared for having a child with specific needs, not necessarily to terminate pregnancy. The Norwegian health care system is founded on core principles including women's autonomy, informed decision making, social equity, and universal access to healthcare services across the country. Nevertheless, our findings indicate that maternal characteristics such as age, parity, ethnic background, and region of residence influenced maternal decisions on TOP after prenatal diagnosis of DS. In fact, studies from countries with comparable health care systems reported lower uptake of prenatal screening among women residing in remote areas or belonging to lower socioeconomic groups. ^29^ , ^30^ , ^31^ , ^32^

Obstetric ultrasound was invented in the 1980s as a supplement to standard obstetric care to improve maternal and neonatal outcomes. Prenatal diagnostics developed rapidly over the last 40 years affecting not only perinatal outcomes, but also maternal health during pregnancy. ^33^ Accordingly, we found that the detection of DS at the first trimester resulted concurrently in earlier terminations during the study period, from 17 to 15 weeks gestation. Earlier diagnosis offers women safer options for TOP and reduced psychological distress that typically develops as pregnancy advances. ^34^ , ^35^ , ^36^ In accordance with our findings, a German study conducted in 2017 showed that the gestational age at termination remained unchanged around week 15 over a 10‐year period. ^37^ It can be assumed that earlier gestational age at the time of prenatal diagnosis might affect parental decisions for TOP. Indeed, a study by Britt et al. found that gestational age less than 16 weeks increases the chances of parents deciding to terminate the pregnancy. ^38^ However, in our study population, the rate of pregnancy termination remained stable, despite diagnosis occurring at an earlier gestational age.

A key strength of this study is its utilization of a large, population‐based dataset derived from a comprehensive national registry within a public healthcare system. The dataset encompasses robust and reliable data collected over a 50‐year period, capturing temporal variations in prenatal screening policies and practices. ^39^ In fact, the Norwegian health care system offers prenatal screening free of charge to all pregnant women, regardless of their socioeconomic background and region of residency. Importantly, prenatal screening is regulated by law and solely offered in fetal medicine centres at the country's five University hospitals. Moreover, deliveries and TOPs are only allowed in public hospitals and reporting to the MBRN is mandatory and regulated by law.

There are several limitations in our study. Firstly, data on miscarriage rates due to DS are probably of lesser quality and underreporting is likely. Secondly, our dataset does not contain cytogenetic analyses. Therefore, some babies with DS might not be reported to the MBRN due to late genetic testing. Thirdly, the binary prenatal testing variable (yes/no) may be subject to underreporting, particularly following birth or miscarriage, which could lead to incomplete case ascertainment. Moreover, we do not have precise data on the uptake of the various screening programs. However, according to two recent Norwegian audits, 76% of women believe that NIPT should be offered to all pregnant women as part of early prenatal screening within the public health service. ^40^ , ^41^ Lastly, access to comprehensive phenotypic information for DS, including structural malformations and markers such as increased nuchal translucency, would likely enhance estimates of prenatal detection rates. A limitation of the present database is the absence of such detailed phenotypic data.

CONCLUSION

Revisions to Norway's national prenatal screening program have influenced both the prevalence and outcomes of pregnancies involving DS. In 2020, the Biotechnology Act was amended to offer first‐trimester ultrasound to all pregnant women (without maternal biomarkers) and to provide NIPT for chromosomal aneuploidies as a first‐tier test to women aged 35 and above. These changes are likely to further impact the prenatal detection rate and prevalence of DS. Consequently, it is essential to conduct updated studies evaluating the impact of the revised legislation and screening program as it becomes integrated into clinical practice nationwide. This is particularly important given that this specific screening model is not implemented in any other country.

AUTHOR CONTRIBUTIONS

Vasilis Sitras conceived and designed the study and wrote the paper. Cecilie Bryn Nordklev conceived and designed the study, performed the statistical analyses, and wrote the paper. Denise Reis Costa and Roy Miodini Nilsen performed the statistical analyses in R. Nils‐Halvdan Morken and Siri Vangen conceived and designed the study. Critical revision of the intellectual content and final approval of the all authors.

FUNDING INFORMATION

Oslo University Hospital.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

ETHICS STATEMENT

The study had no patient involvement, as it was based on data from the mandatory national registry. The Regional Committee for Medical and Health Research Ethics South‐East Norway approved the study (REK nr. 2017/682) on February 14, 2024. Informed consent was not required. Results are reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations. ^42^

Supporting information