Maternal Autoimmune Diseases and Offspring Structural Congenital Heart Disease: A Nationwide Population-Based Cohort Study.

Observational OA: gold CC-BY-NC-4.0
⚙ AI-generated summary by qwen3.7-flash, 2026-11-14 ⓘ

This nationwide cohort study found that maternal autoimmune diseases, particularly systemic lupus erythematosus and Sjögren syndrome, are independent risk factors for offspring structural congenital heart disease.

One-sentence paraphrase of the abstract; not a substitute for reading it. No clinical advice. How this works

⚙ AI-generated deep summary by qwen3.7-flash, 2026-11-14 · read from full text ⓘ

This nationwide population-based cohort study analyzed over 1.2 million births in Taiwan to assess the association between maternal systemic autoimmune diseases and structural congenital heart disease in offspring. The researchers found that children born to mothers with conditions such as systemic lupus erythematosus or Sjögren syndrome had a significantly increased risk of structural cardiac malformations, even after adjusting for various maternal comorbidities and demographic factors. While hydroxychloroquine exposure showed an association with CHD, the authors noted this likely reflected underlying disease severity rather than a direct teratogenic effect of the medication. Relevance to endometriosis: listed as one of the maternal comorbidities adjusted for in the statistical models, though the paper's main focus is congenital heart disease.

Read from the paper's body, not the abstract. Not a substitute for reading the paper. No clinical advice. How this works

Abstract

BackgroundMaternal autoimmune diseases (ADs) are associated with adverse pregnancy outcomes, yet their specific impact on offspring structural congenital heart disease (CHD) remains inconclusive. This study aimed to evaluate the risk of structural CHD in children born to mothers with various ADs using a nationwide population-based database.MethodsWe conducted a nationwide retrospective cohort study using the Maternal and Child Health Database in Taiwan (2004-2019), a comprehensive data set linking the National Health Insurance Research Database with birth registries. We identified 4708 live births from mothers with ADs and matched them 1:2 with 9416 control births. Cox proportional hazards models were used to estimate adjusted hazard ratios (aHRs) for structural CHD, adjusting for maternal age, comorbidities, and pregnancy-related factors.ResultsOffspring of mothers with ADs had a significantly higher risk of CHD compared with the control group (aHR, 1.51 [95% CI, 1.24-1.85]). Specific maternal ADs significantly associated with offspring CHD included systemic lupus erythematosus (aHR, 2.01 [95% CI, 1.55-2.60]) and Sjögren syndrome (aHR, 1.47 [95% CI, 1.10-1.96]).ConclusionsMaternal ADs are independent risk factors for offspring structural CHD. These findings underscore the importance of multidisciplinary counseling and specialized fetal echocardiographic screening for pregnant women with ADs to ensure early detection and management of potential cardiac defects in their offspring.RegistrationURL: https://www.clinicaltrials.gov; Unique identifier: CS19009.
Full text 17,502 characters · extracted from pmc-nxml · 7 sections · click to expand

Methods

This retrospective cohort study used the MCHDB in Taiwan, which is a specialized longitudinal data set created by linking the National Health Insurance Research Database with birth registry data, maternal–neonatal linkage files, and cause‐of‐death registries. These data were accessed through the Health and Welfare Data Science Center. The MCHDB provides comprehensive, population‐based records, allowing precise ascertainment of maternal comorbidities and medications during pregnancy while maintaining strict patient anonymity. The study protocol was approved by the Institutional Review Board of Chung Shan Medical University Hospital (Approval No. CS19009). The requirement for informed consent was waived because the study used deidentified administrative claims data. The analytic methods used in this study will be made available from the corresponding author upon reasonable request. However, the raw data from the MCHDB are not publicly available due to legal restrictions imposed by the Health and Welfare Data Science Center in Taiwan. There were 1 459 093 birth certifications in the database from 2009 to 2016. After excluding missing data on the mother's identity, the child's nationality, foreign nationality, multiple births, and stillbirth, a final eligible cohort of 1 246 129 children was identified. The selection of specific ADs, including rheumatoid arthritis, SLE, Sjögren syndrome, ankylosing spondylitis, and psoriasis, was based on their higher prevalence among women of reproductive age and their shared systemic chronic inflammation. We focused on these common systemic autoimmune rheumatic diseases to ensure sufficient statistical power. Conditions such as multiple sclerosis and type 1 diabetes were excluded, as they are classified as organ‐specific rather than systemic disorders. Using birth year, child's sex, and maternal age as matching criteria, each of the 4708 offspring born to exposed mothers was matched in a 1:2 ratio with those born to nonexposed mothers. In this study, CHD was strictly defined as structural alterations of the heart, such as septal defects, valvular abnormalities, or conotruncal anomalies. We focused on structural malformations to investigate how maternal ADs independently affect fetal cardiogenesis. Consequently, nonstructural cardiac disorders, including congenital atrioventricular block and other arrhythmias, were excluded from our primary outcome to minimize the confounding effects of known rhythm disturbances associated with neonatal lupus (anti‐Ro/SSA and anti‐La/SSB antibodies) and to clarify the specific risk of structural CHD associated with maternal autoimmunity. A detailed list of structural CHD categories and their corresponding International Classification of Diseases , Ninth Revision , Clinical Modification ( ICD‐9‐CM ) and Tenth Revision , Clinical Modification ( ICD‐10‐CM ) codes is provided in Table  S1 . To ensure diagnostic accuracy and exclude transient neonatal conditions, cases of atrial septal defect, patent foramen ovale, and patent ductus arteriosus were included only if the diagnosis persisted beyond the age of 1 year. Both groups were followed until the first diagnosis of CHD, death, or the study end date of December 31, 2019. To address potential sources of bias, we adjusted for a comprehensive set of maternal and paternal covariates. These included demographic factors (age, urbanization, insurance unit), delivery characteristics (Apgar scores, gestational age, birth weight), and maternal comorbidities such as asthma, hypertension, diabetes, hyperlipidemia, gestational diabetes, preeclampsia/eclampsia, urinary tract infection, endometriosis, sleep disorder, and depression. Comorbidities were defined by clinical records spanning 2 years before the index date. Statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC). Cumulative incidence of CHD was estimated using Kaplan–Meier analysis, and the log‐rank test was used to determine statistical significance between groups. Independent risk, expressed as hazard ratios (HRs), was calculated using multivariable Cox proportional hazard models, with the proportionality assumption verified via Schoenfeld residuals. Regarding missing data, variables with missing values (eg, paternal age) were addressed by using a “missing” category to maintain the total sample size in the multivariable analysis without excluding significant portions of the cohort. To minimize the influence of potential confounders, 2 multivariable models were used: Model 1 adjusted for demographic and baseline factors (child's birth year, sex, maternal age, urbanization, insurance unit, and paternal age), while model 2 was a fully adjusted model accounting for all variables, including maternal comorbidities such as diabetes, hypertension, gestational diabetes, preeclampsia/eclampsia, depression, sleep disorders, asthma, and hyperlipidemia. Both groups were followed until the first diagnosis of CHD, death, or the end of the study period. To address the potential impact of competing risks, we evaluated infant death as a competing event. Infant death was rare and balanced between cohorts; therefore, death was treated as a censoring event in the multivariable Cox proportional hazards models. 16

Results

Of the initial 1 459 093 birth records from 2009 to 2016, 1 246 129 children remained after excluding those with missing maternal ID (n=67 270), foreign citizenship (n=88 837), stillbirths (n=14 639), and multiple births (n=42 218). Within this eligible cohort, 4708 offspring were born to mothers with ADs. Through a 1:2 exact matching process by birth year and sex, 9416 nonexposed offspring were identified from the remaining eligible children (Figure  1 ). Mothers with ADs had a higher prevalence of several comorbid conditions during pregnancy compared with the nonautoimmune group, including hypertension (4.61% versus 1.68%), hyperlipidemia (1.59% versus 0.52%), preeclampsia or eclampsia (6.63% versus 3.90%), sleep disorder (9.47% versus 5.11%), depression (6.52% versus 2.57%), asthma (2.06% versus 1.22%), and urinary tract infection (16.46% versus 11.68%) (all P <0.001). No significant differences were observed between the 2 groups about the child's sex, diabetes, gestational diabetes, or urbanization (Table  1 ). Baseline Characteristics ASD indicates absolute standardized difference. The incidence rate of CHD was significantly higher in the offspring of mothers with ADs compared with the control group (5.35 versus 3.39 per 10 000 person‐months). The cumulative incidence proportion of CHD was significantly elevated in the autoimmune group throughout the follow‐up period (log‐rank P <0.001; Figure  2 ). The yellow dashed line represents offspring of mothers with autoimmune diseases, and the solid dark red line represents the nonautoimmune group (log‐rank P <0.0001). Children born to mothers with ADs had a 1.57‐fold increased crude risk of CHD (HR, 1.57; 95% CI, 1.29–1.90). After adjusting for all maternal and paternal covariates in model 2, the association remained statistically significant with an adjusted hazard ratio (aHR) of 1.25 (95% CI, 1.01–1.54; Table  2 ). Incidence of Congenital Heart Disease aHR indicates adjusted hazard ratio. Incidence rate, per 10 000 person‐months. Model 1: adjusted for child's birth year, child's sex, mother's age, urbanization, insurance unit, mother's comorbidity, and father's age. Model 2: adjusted for all variables. In stratified analyses by AD subtypes, the risk of offspring CHD was particularly significant for maternal SLE (aHR, 2.01 [95% CI, 1.55–2.60]) and Sjögren syndrome (aHR, 1.47 [95% CI, 1.10–1.96]; Figure  3 ). Other significant independent predictors for CHD included low birth weight (<2500 g; aHR, 1.46 [95% CI, 1.06–2.01]), preterm birth (aHR, 1.74 [95% CI, 1.28–2.37]), and cesarean section (aHR, 1.40 [95% CI, 1.14–1.72]; Table  3 ). An association was noted between in utero hydroxychloroquine exposure and CHD, although the CI was wide (Table  4 ). Further stratification by specific CHD subtypes was not performed due to the limited number of events in certain categories. aHR indicates adjusted hazard ratio. Cox Proportional Hazard Model Analysis for Risk of Congenital Heart Disease aHR indicates adjusted hazard ratio; and HR, hazard ratio. Model 1: adjusted for child's birth year, child's sex, mother's age, urbanization, insurance unit, mother's comorbidity, and father's age. Model 2: adjusted for all variables. Incidence of Congenital Heart Disease in the Offspring of Mothers With Autoimmune Diseases on Medication aHR indicates adjusted hazard ratio; and NA, not applicable. Model 1: adjusted for child's birth year, child's sex, mother's age, urbanization, insurance unit, mother's comorbidity, and father's age. Model 2: adjusted for all variables.

Sources

None.

Discussion

In this study, infants born to pregnant women with ADs had a 1.51‐fold higher risk of developing CHD than those whose mothers did not have an AD, after adjusting for birth year, sex, parental age, urbanization, insurance unit, and maternal comorbidities. To the best of our knowledge, this is the first and largest nationwide population‐based cohort study in Taiwan to analyze the association between maternal ADs and offspring CHD. Our findings resonate with previous studies conducted in Western populations; for instance, Vinet et al reported a 2.62‐fold increased risk of CHD in offspring of mothers with SLE in a Canadian cohort. 17 Similarly, Liu et al identified maternal connective tissue disorders as significant risk factors for offspring CHD. 12 However, our study extends this evidence to an East Asian population and encompasses a broader spectrum of systemic ADs. Importantly, by strictly defining CHD as structural malformations and excluding rhythm disturbances (such as congenital heart block), we provide a more focused understanding of how maternal autoimmunity independently impacts structural cardiogenesis. Furthermore, our analysis showed an association between in utero hydroxychloroquine exposure and CHD (Table  4 ). However, due to the wide CIs, these findings remain inconclusive. It is critical to distinguish whether hydroxychloroquine use is a direct risk factor or if it acts as a proxy for maternal disease severity. Given that hydroxychloroquine is typically prescribed for more active or severe autoimmune flares, our results likely reflect the impact of higher maternal disease activity on fetal development rather than a direct teratogenic effect of the medication itself. 18 Therefore, the underlying disease activity and severity, for which hydroxychloroquine serves as a proxy, may have accounted for the observed increase in CHD risk. The mechanisms of maternal ADs that may underlie the physiopathology of CHD in newborns include cytokine imbalance and autoantibody‐mediated damage. Transplacental transfer of maternal autoantibodies, such as immunoglobulin G, might affect the fetus in both structural CHD and congenital heart block cases. It has been hypothesized that maternal ADs might play a critical role in structural CHD. 19 , 20 , 21 Antibodies, such as anti‐SSA/Ro, are associated with neonatal lupus and can lead to cardiac manifestations. Some research has demonstrated that the binding of maternal antibodies to apoptotic fetal cardiocytes triggers the release of profibrosing and proinflammatory cytokines, leading to scarring. This process may extend beyond the conduction system to the valves, endocardium, and myocardium. Supporting this, a retrospective study of autopsies from neonatal lupus cases observed histological damage, such as lymphohistiocytic infiltrate in the ventricular septum and foci of calcification in the atrial septum. In addition, 40% of deaths caused by congenital heart block showed pathological findings including calcification of the aortic, pulmonary, tricuspid, or mitral valves. 14 , 22 Distinct from the inflammatory pathway, separate mechanisms may involve the disruption of cardiac structure formation. Macrophages, a crucial component of the innate immune system, play an important role in cardiac remodeling. 23 , 24 While ventricular septal defects are thought to be associated with dysregulated cellular death within a formed ventricular septum during active remodeling, it has also been proposed that maternal autoantibodies might directly impair the physical closure of septal defects. 17 This could explain the increased risk of atrial septal defect and ventricular septal defect in offspring of pregnant women with ADs compared with the general population. Antiphospholipid antibodies (APLs), one of the types of autoantibodies found in women with ADs, can cross the placenta. A previous study found that 30% of newborns born to mothers with antiphospholipid syndrome tested positive for APLs in their cord blood. 25 APLs are associated with some valvular diseases (eg, valvular regurgitation, nodules, and verrucous endocarditis). 26 Deposits of APLs are thought to play a significant pathogenic role in the valvular disease. 26 Although other studies have shown perinatal thrombotic events in children born to mothers with APLs, the relationship with the prevalence of CHD remains unclear in these patients. 27 Therefore, it has been hypothesized that APLs may play an important role in congenital valve anomalies in fetuses due to their involvement in valvular damage and their ability to cross the placenta. TGF‐β (transforming growth factor‐β) also plays a role in endocardial cushion formation. It plays an important role in cardiac septation during cardiac embryogenesis, requiring expression of cytokines, such as TGF‐β. It has been demonstrated in animal models that both fetal and maternal TGF‐β in cardiac embryogenesis are important. 28 The mice born to mothers with TGF‐β1 nonexpression had more severe CHD compared with those born to mothers with normal TGF‐β1 expression. Investigators hypothesized that transplacental transfer of maternal TGF‐β1 occurring from mother to fetus might rescue potential heart defects in the offspring. 28 The levels of TGF‐β1 in patients with ADs are lower than in healthy controls, and these levels correlate inversely with disease activity. 29 Maternal TGF‐β–induced rescue of heart defect in defective TGF‐β fetuses might not happen in pregnant women with ADs, leading to an increased risk of CHD. Our study has several limitations that warrant consideration. First, although this retrospective cohort study used a large nationwide data set to identify significant associations, the inherent limitations of an observational design mean that causality between maternal ADs and offspring CHD could not be conclusively inferred. Second, we could not adjust for certain prenatal exposures, such as multivitamin and folic acid supplementation, as these are often obtained over the counter without a formal prescription. Additionally, we excluded stillbirths because a proportion of stillbirths result from medical terminations, for which the underlying indications or specific cardiac outcomes were not consistently recorded in the administrative database. Third, unmeasured confounders such as obesity, smoking, and alcohol use are inherent limitations of administrative data sets like the MCHDB, as they are not required for insurance reimbursement. To mitigate this, we used urbanization and insurance units as socioeconomic proxies. Interestingly, the AD group had a significantly higher proportion of government employees (8.81% versus 6.81%, P =0.0002). Given that government employment in Taiwan correlates with higher health literacy and socioeconomic stability, this group likely exhibits healthier prenatal behaviors. Consequently, any residual confounding would likely bias our findings toward the null, suggesting that our results provide a conservative estimate of the risk posed by maternal autoimmunity. Fourth, the MCHDB does not record individual serological data. The presence or specific titers of autoantibodies (eg, anti‐Ro/SSA or APLs) may be a primary determining factor for CHD risk. Therefore, the precise role of maternal cytokines and autoantibodies in the pathogenesis of structural CHD remains to be established in future prospective research. Finally, our study did not explicitly exclude maternal infections such as rubella or syphilis. Although these conditions are rare in Taiwan due to effective public health programs, their potential role as minor confounders cannot be entirely ruled out. It should be noted, however, that the high rate of pregnancy loss associated with these infections likely means that severely affected fetuses would not have met our inclusion criteria for live births, thereby minimizing their impact on our observed associations in the surviving cohort. Despite these limitations, the use of a nationwide population‐based data set strengthens the high generalizability of our findings.

Conclusions

This nationwide study identifies maternal ADs as independent risk factors for offspring structural CHD. Given the significant risk elevation associated with maternal SLE and Sjögren syndrome, clinical strategies should prioritize specialized prenatal cardiac surveillance for these high‐risk pregnancies. Further investigation into the underlying immunological pathways is warranted to refine preventive interventions.

Coi Statement

None.

Supplementary Material

Table S1 STROBE Statement

Text is read by the "Ask this paper" AI Q&A widget below. Extraction quality varies by source — PMC NXML preserves structure cleanly, OA-HTML may include some navigation residue, and OA-PDF can have broken hyphenation. The publisher copy (via DOI) is the canonical version.

My notes (saved in your browser only)

⚙ Ask this paper AI returns verbatim quotes from the full text · source: pmc-nxml ⓘ

Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works

Citation neighborhood (no data yet)

We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2026) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.

SciLite annotations

organisms 8
noordeloos 2009062 noordeloos 2009062 noordeloos 2009062 noordeloos 2009062 rodents mus sp. noordeloos 2009062 noordeloos 2009062
chemicals 7
hydroxychloroquine hydroxychloroquine hydroxychloroquine hydroxychloroquine hydroxychloroquine folic acid alcohol

Source provenance

europepmc
last seen: 2026-11-15T09:34:25.365398+00:00
scilite
last seen: 2026-11-01T10:01:09.713090+00:00
unpaywall
last seen: 2026-11-02T06:26:32.194460+00:00
License: CC-BY-NC-4.0