Introduction

Atrial septal defect has been considered a relatively benign lesion, but recent studies have shown significantly higher long-term mortality in patients with an atrial septal defect diagnosis than in the background population. ^{Reference Nyboe, Karunanithi, Nielsen-Kudsk and Hjortdal1} Patients with atrial septal defects also have increased risks of arrhythmia, stroke, pneumonia, and impaired exercise capacity, as well as a higher incidence of cognitive issues and psychiatric diagnoses. ^{Reference Nyboe, Olsen, Nielsen-Kudsk and Hjortdal2–Reference Udholm, Nyboe and Karunanithi5} In the fetal circulation, there is a physiological right-to-left shunt between the atria through the foramen ovale, which normally closes at birth or shortly after. Why the shunt fails to close in some infants, resulting in an atrial septal defect, is not fully understood.

Maternal pre-eclampsia during pregnancy is associated with CHD in the offspring. Cardiac development in the fetus occurs within the first 8 weeks of pregnancy, whereas pre-eclampsia is defined as new-onset hypertension and other characteristic signs of organ dysfunction after gestational week 20. Placental abnormalities, genetics, maternal cardiovascular dysfunction, and other maternal risk factors are involved in the aetiology. Pathophysiological mechanisms involved in both pre-eclampsia and CHD include oxidative stress, inflammation, and angiogenic and metabolic biological processes. ^{Reference Sliwa and Mebazaa6–Reference Sitras, Fenton and Acharya8}

Several studies have demonstrated the association between pre-eclampsia and a broad spectrum of offspring CHDs. ^{Reference Boyd, Basit and Behrens9–Reference Liu, Zhao, Xie, Wu, Li and Yao12} The associations found in these large registry-based studies are very solid. The validity of especially major CHDs in medical registries is known to be excellent. However, for minor CHDs, there is some uncertainty, due to the natural history and clinical presentation of these defects. Minor CHDs, like atrial septal defects, can be underrepresented in medical registries, as they are often asymptomatic or not diagnosed until later in life. The atrial septal defects that are diagnosed in the registries represent the worst end of the spectrum as those patients who have symptoms, complications, or comorbidity will most likely be overrepresented. It has been subject to debate, to what extent the estimates of association between pre-eclampsia and CHD are also valid for minor CHDs. ^{Reference Brodwall, Greve and øyen13}

Given these considerations, we sought to investigate whether the well-established association between pre-eclampsia and CHDs also extends to atrial septal defects, which may be underreported or misclassified in registry-based studies due to diagnostic challenges and milder early clinical presentation.

We investigated the association between pre-eclampsia and the risk of atrial septal defects in the offspring in a large population-based cohort of neonates systematically examined using echocardiography within 1 month of birth.

Materials and methods

Results

Of the 25,590 neonates examined with echocardiography in the Copenhagen Baby Heart Study, ^{Reference Vøgg, Basit and Raja15} assessment for atrial septal defect according to the new diagnostic algorithm ^{Reference Dannesbo, Blixenkrone-Moeller and Pihl16} was feasible in 12,388 neonates. We excluded neonates with missing data on hypertensive disorders of pregnancy (n = 34).

Of the 12,354 neonates included in this study, 462 (3.7%) were exposed to maternal pre-eclampsia. Maternal and newborn characteristics of the study population are shown in Table 1. The proportion of multiple pregnancies was higher in pregnancies with pre-eclampsia (n = 42, 9.1%) than in pregnancies without pre-eclampsia (n = 299, 2.5%). Neonates exposed to maternal pre-eclampsia were more likely than unexposed neonates to be born to mothers who were overweight or obese. Neonates exposed to maternal pre-eclampsia were more often born preterm than unexposed neonates, and accordingly, a higher proportion of neonates exposed to maternal pre-eclampsia had lower birthweights and shorter birth lengths.

Table 1.

Maternal and newborn characteristics of the study cohort, by maternal pre-eclampsia status and presence or absence of atrial septum defects, Copenhagen Baby Heart Study, 2017–2018

ASD = atrial septal defect; BMI = body mass index; SD = standard deviation.

Overall, 5.9% (n = 732) of the study cohort had an atrial septal defect; 7.4% (34/462) of neonates exposed to maternal pre-eclampsia had an atrial septal defect, while an atrial septal defect was present in 5.9% (698/11,892) unexposed neonates.

Neonates exposed to maternal pre-eclampsia had a 19% increased risk of atrial septal defect, compared with unexposed neonates (Table 2), although the confidence interval was wide, reflecting the relatively small number of neonates with atrial septal defect whose mothers had had pre-eclampsia (adjusted risk ratio 1.19, 95% confidence interval 0.83, 1.64). Additional adjustment for maternal body mass index did not materially change the estimates (adjusted risk ratio 1.17, 95% confidence interval 0.81, 1.59).

Table 2.

Risk ratio (RR) of ASD in the neonate by maternal pre-eclampsia. Unadjusted and adjusted analyses, and sensitivity analyses

^a RR adjusted for infant sex, gestational age at birth (continuous, days), infant birthweight (continuous, g), infant birth length (continuous, cm), infant age at echocardiographic examination (continuous, days), and maternal age (groups).

^b Concomitant other CHD in this cohort included neonates with coarctation of the aorta (n = 1), VSD (n = 408), pulmonary stenosis (n = 5), and cardiac tumour (n = 1).

ASD = atrial septal defect; PFO = patent foramen ovale.

Severe pre-eclampsia resulted in a risk ratio for atrial septal defect of 1.35 (95% confidence interval 0.81, 2.24), while mild to moderate pre-eclampsia resulted in a risk ratio for atrial septal defect of 1.20 (95% confidence interval 0.78, 1.84).

When we treated neonates with patent foramen ovale as a separate outcome category (rather than treating them as neonates without atrial septal defects), we found very little evidence of an association between maternal pre-eclampsia and offspring patent foramen ovale (adjusted risk ratio 0.91, 95% confidence interval 0.73, 1.14), whereas pre-eclampsia remained associated with offspring atrial septal defects (adjusted risk ratio 1.19, 95% confidence interval 0.95, 1.49). When we considered exposure to gestational hypertension separately (rather than treating these neonates as unexposed), we observed the following risk ratios: gestational hypertension, adjusted risk ratio 1.10, 95% confidence interval 0.68, 1.77; pre-eclampsia, adjusted risk ratio 1.26, 95% confidence interval 0.90, 1.75.

Absolute prevalences for the above-mentioned sub-analyses are shown in Supplementary Tables 1–3.

Various sensitivity analyses did not materially change our results, except for excluding neonates exposed to maternal gestational diabetes, which slightly attenuated the risk ratio estimate of atrial septal defect for neonates exposed to pre-eclampsia.

Discussion

In this large, population-based cohort study, we investigated exposure to maternal hypertensive disorders of pregnancy, principally pre-eclampsia, and the risk of offspring atrial septal defects in a cohort of 12,354 neonates systematically examined with echocardiography. We found evidence of a 19% increase in the risk of atrial septal defect in neonates exposed to maternal pre-eclampsia, compared to unexposed neonates. We found no association between exposure to maternal pre-eclampsia and patent foramen ovale in the offspring.

Several large registry-based studies have shown strong associations between maternal pre-eclampsia and CHD in the offspring. ^{Reference Boyd, Basit and Behrens9–Reference Brodwall, Leirgul and Greve11} The risk of specifically atrial septal defect after exposure to maternal pre-eclampsia was assessed in two of those studies. A Canadian population-based study ^{Reference Auger, Fraser, Healy-Profitós and Arbour10} found a prevalence ratio for atrial septal defect in infants exposed to pre-eclampsia of 1.91 (95% confidence interval 1.73, 2.10) in analyses adjusted for maternal age, parity, comorbidity, multiple birth, socio-economic deprivation, and period. Data were derived from hospital discharge abstracts, and only defects diagnosed before discharge from the hospital after birth were included. A Danish study, ^{Reference Boyd, Basit and Behrens9} likewise based on hospital discharge registers, found strong associations with atrial septal defect for both early preterm pre-eclampsia (odds ratio 12.0; 95% confidence interval 8.96, 16.1) and late preterm pre-eclampsia (odds ratio 3.82; 95% confidence interval 2.53, 5.77), but not for term pre-eclampsia (odds ratio 0.92; 95% confidence interval 0.68, 1.23).

Whether the strong associations previously observed between pre-eclampsia and non-critical heart defects, including atrial septal defects, are the product of detection bias has been the topic of debate. ^{Reference Brodwall, Greve and øyen13} Atrial setal defects are among the most common minor CHDs and are often asymptomatic until late childhood or even adulthood. If infants born to mothers with pre-eclampsia were more often examined with echocardiography – increasing the rates of atrial septal defect detection – associations between maternal pre-eclampsia and offspring atrial septal defects might have been inflated. In our population-based study cohort, neonates were recruited and examined with echocardiography within 30 days of birth regardless of whether or not they had been exposed to maternal pre-eclampsia (by operators who did not know their exposure status), and all echocardiograms were systematically assessed for atrial septal defects, again in a blinded fashion, minimising the risk of detection bias. Contrary to previous studies, our cohort included infants with a range of defect sizes, including small and asymptomatic defects that might never have been diagnosed otherwise and would probably not have been found in registries. We find evidence of an association between maternal pre-eclampsia and atrial septal defect in the neonate even in a population with a fair representation of small defects, and we consider this important since morbidity and mortality are seen in patients with atrial septal defects even if the defect is considered small and haemodynamically insignificant and even after spontaneous, surgical, or percutaneous closure of the defect. ^{Reference Nyboe, Karunanithi, Nielsen-Kudsk and Hjortdal1,Reference Udholm, Nyboe and Karunanithi5,17,18}

A strength of our study is good data validity for both the exposure (pre-eclampsia) and the outcome (atrial septal defects). The diagnoses in the obstetrical database have been validated against medical records, showing excellent agreement.¹⁹ Ascertainment of the diagnosis of atrial septal defect has been performed based on a standardised echocardiographic protocol and using an algorithm that has been validated in the Copenhagen Baby Heart Study, showing substantial interobserver agreement. ^{Reference Dannesbo, Blixenkrone-Moeller and Pihl16}

It has not yet been fully defined what links hypertensive disorders of pregnancy and CHD, but multiple mechanisms are probably involved. Imbalances in angiogenic factors like vascular endothelial growth factor, placental growth factor, and transforming growth factor beta have been shown in both pre-eclampsia and CHD. ^{Reference Sliwa and Mebazaa6,Reference Llurba, Sánchez and Ferrer7,20,21} The placenta plays a crucial role in both pre-eclampsia and the development of the fetal heart and might also be part of the possible shared pathophysiological pathway.^22–24 Also, common traits in gene expression profiles for pre-eclampsia and CHD have been found. ^{Reference Sitras, Fenton and Acharya8} The genes found to be common for pre-eclampsia and CHD are involved in pathophysiological processes regulating oxidative stress, inflammation, and angiogenic and metabolic biological processes.

In the aforementioned Danish registry-based study, ^{Reference Boyd, Basit and Behrens9} the association between pre-eclampsia and atrial septal defects was most profound for early preterm pre-eclampsia, defined as pre-eclampsia with delivery at <34 weeks gestation. Preterm neonates were underrepresented in the Copenhagen Baby Heart Study (and there were only 60 infants delivered <34 weeks in our study population), ^{Reference Vøgg, Basit and Raja15} and as early-onset pre-eclampsia often results in preterm birth, neonates born to mothers with early-onset pre-eclampsia are likely to be underrepresented in our cohort. Early-onset and late-onset pre-eclampsia are often considered to be two different diseases with distinct pathophysiology.²⁵ In our cohort, information on the date for the diagnosis of pre-eclampsia was not available, and thus we could not differentiate between early-onset and late-onset pre-eclampsia. However, because of the underrepresentation of preterm neonates in the Copenhagen Baby Heart Study, our cohort is probably skewed towards late preterm and term pre-eclampsia, possibly weakening our estimates. Yet, we still find evidence of an association in our cohort.

The Copenhagen Baby Heart Study is a population-based study; nevertheless, our results might have limited generalisability to other ethnicities or populations with different compositions in terms of socio-economic status.

Our study cohort is to date the largest cohort of neonates to be systematically assessed for atrial septal defects based on echocardiographic data. However, we have wide confidence intervals, reflecting the relatively small number of neonates with atrial septal defect who had been exposed to pre-eclampsia. Even though we had limited power, our estimates were robust to multiple exclusions and additional adjustments, suggesting that there is evidence of an association between maternal pre-eclampsia (later-onset) and the risk of atrial septal defect in the offspring.

Conclusion

In a large population-based cohort of asymptomatic neonates systematically examined by echocardiography, we found evidence of an increased risk of atrial septal defect in neonates born to mothers with pre-eclampsia.