Genetic detection of congenital heart disease

2.1 Overview

The etiology of CHD is multifactorial, and both genetic and environmental factors are contributors.^9,19,28 Genetic variation contributes to an estimated 40% of CHD cases, environmental factors contribute to an estimated 5%, and the remaining 55% of CHD cases are of unknown etiology (Fig. 1).

2.2 Genetic causes in CHD

2.3 Environmental contributors to CHD

Environmental causes are implicated in an estimated 5% (range from 2% to 10%) of CHD cases (Fig. 1).⁹ CHD has been associated with environmental risk factors including maternal/paternal environmental exposures to air pollution or toxic chemicals, parental smoking, maternal history of infectious diseases during pregnancy, pre-gestational and gestational diabetes mellitus, maternal obesity, maternal drug use, and pregnancy through artificial reproductive technologies.^9,91–94 The association of maternal alcohol or coffee consumption with CHD was not significant, however, maternal folic acid supplementation seems to have a preventive effect on CHD.⁹⁵ Although the underlying mechanisms by which environmental factors disrupt molecular pathways during cardiac development to cause CHD remain unknown, CHD has been shown to be caused by gene-environment interactions resulting in increased incidence of CHD in animal models.⁹⁶ Table 1 highlights some of the most common environmental/teratogenic explanations and their associated CHDs. For pregnancies with known risk factors such as those listed in the table, fetal echocardiography may be indicated to screen for CHDs.

Patients should also be counseled on the availability and potential utility of carrier screening for the parents of a fetus with a CHD. Carrier screening uses various methodologies such as targeted genotyping and/or gene sequencing to attempt to identify individuals who are at increased risk of having a child with an inherited autosomal recessive or X-linked disease. This screening can be performed before or during a pregnancy to identify at-risk couples, who then have the option to pursue alternative reproductive options to reduce the risk of having an affected child, such as prenatal diagnosis, pre-implantation genetic testing of embryos, use of a sperm/oocyte donor, or adoption. Historically, ethnicity-based carrier screening was performed for a handful of the most common disorders in a given ethnic group. Advances in genetic testing technology and decreased cost now allow for high throughput testing of several hundred genes simultaneously with a rapid turnaround time, and many carrier screening laboratories can now test for several hundred diseases via a single blood and/or saliva sample. The ACMG now recommends an ethnicity-agnostic, tiered approach to carrier screening.¹²³ Tier 1 screening is defined as screening for cystic fibrosis, spinal muscular atrophy, and risk-based screening as determined by personal and family history, as well as other risk factors. Tier 2 screening is defined as screening for conditions with a carrier frequency of 1 in 100 or greater. Tier 3 screening is defined as screening for conditions with a carrier frequency of 1 in 200 or greater. Tier 4 screening is defined as screening for conditions with a carrier frequency of less than 1 in 200. For the general population, ACMG recommends offering tier 3 screening. Tier 4 screening should be considered in cases of parental consanguinity, or when a family or personal medical history warrants it.

While the yield of carrier screening for the purposes of identifying a diagnosis in a fetus with a prenatally detected CHD is often low, there are numerous autosomal recessive disorders which have been associated with CHD and, in the setting of non-specific fetal anomalies, large carrier screening panels inclusive of several hundreds of genes can provide a potentially cost-effective means of “casting a wide net” for fetal disease. Disorders that may be included on carrier screening panels that have also been implicated in CHD include Smith-Lemli-Opitz syndrome (OMIM 270400), Bardet-Biedl syndrome (OMIM PS209900), Barth syndrome (OMIM 302060), Ellis-van Creveld (OMIM 225500), Fanconi anemia (OMIM PS227650), some heterotaxy conditions such as ZIC3 (OMIM 300265), and more. Identification of a couple at-risk for a disease on a carrier screening panel before/during early-stage of pregnancy, which later fits the fetal phenotype, can help direct targeted single gene testing on a prenatal specimen obtained via chorionic villus sampling (CVS) or amniocentesis.

2.4 Complex inheritance, incomplete penetrance, and genetic modifiers of CHD

Over half of CHD cases have unknown etiology and some known CHD genes display an inheritance pattern characterized by incomplete penetrance,⁸ which highlights the genetic complexity of CHD. Incomplete penetrance of CHD genes (e.g., NOTCH1, ROBO4, and SMAD6) happens in some families when certain individuals express CHD phenotype while others do not, even though they carry a pathogenic (disease-causing) variant in a CHD gene. There can also be a sex bias in penetrance.^97–99 Genetic or environmental modifiers could contribute to clinical variabilities (severities) in CHD presentation. Besides classic recessive or dominant transmitted models, more complex genetic models (e.g., epistatic interactions of multiple variants/genes¹⁰⁰) could play a role in genetic heterogeneity of CHD.¹⁰¹ Clinical CHD phenotype could be influenced by susceptibility loci as well as genetic modifiers. Common genetic risk variants in multiple loci have been reported in CHD.^102,103 In a family with childhood–onset cardiomyopathy, three missense variants in MKL2, MYH7, and NKX2-5 were required to cause disease.¹⁰⁴ Interaction between DNAH6 and other primary ciliary dyskinesia genes was demonstrated to lead to CHD along with primary ciliary dyskinesia in a study by Li et al.¹⁰⁵ Furthermore, the severity of CHD phenotype may be associated with mosaicism, including the proportion of abnormal cell numbers in the disease-relevant tissues, allele frequencies of pathogenic variants, and types of variants.¹⁰⁶ The allele frequency of a pathogenic variant found in either blood or saliva specimen may not reflect the proportion of abnormal cell numbers/percentages in the patient's cardiac tissue. Various variants (such as missense, nonsense, frame-shift, splicing-site, indel, insertional variants) may have different impacts on the severity of CHD phenotype.

Identification of a fetus with left ventricular outflow tract obstruction (LVOTO) malformations should prompt cardiac evaluation for other family members, especially parents and siblings even when genetic testing is not pursued or the specific genetic cause in the family is not identified. This is because bicuspid aortic valve (BAV)/aneurysm syndrome may present with variations of left-sided defects, including: BAV with aneurysm, BAV without aneurysm, aortic aneurysm and normal aortic valve, as well as more severe defects including coarctation of the aorta or hypoplastic left heart in family members. BAV/aneurysm syndrome is inherited in an autosomal dominant pattern and can show decreased penetrance and therefore may appear to “skip generations” in affected families. Thus, all first-degree relatives in families with BAV/aneurysm syndrome should obtain echocardiograms to check for valve abnormalities as well as aortic root and ascending aortic aneurysms.¹⁰⁷ Deleterious variants in NOTCH1 are associated with aortic valve disease 1 (OMIM #109730) which can present with BAV, aortic calcification/stenosis, mitral valve atresia, ventricular septal defects, double-outlet right ventricle, and hypoplastic left heart.

3.1 Karyotype and FISH for CHD

Conventional chromosome analysis/karyotype recognized chromosomal aneuploidies a few decades ago.¹³ Karyotype visualizes the chromosomes during metaphase for diagnosing numerical chromosome abnormalities (aneuploidies, gross deletions and duplications) and structural chromosome abnormalities (balanced and unbalanced translocations, inversions, insertions, ring chromosomes, etc.) with a resolution of 5–10 ​Mb. It can detect balanced structural chromosome abnormalities, which are not detectable by CMA. Furthermore, it is important to perform parental karyotypes of children with chromosomal abnormalities (such as trisomy 13 or trisomy 21 caused by Robertsonian translocations, deletions, duplications, or balanced or unbalanced reciprocal translocations, etc.) for potential chromosomal abnormalities to assess risk of recurrence in future pregnancies. Inherited translocations have been reported to increase risk of aneuploidy recurrence and spontaneous abortion.¹⁰⁸

FISH targets a specific chromosomal region using fluorescent probes, and can detect aneuploidies, deletions, and duplications with a resolution of 100 kilobase pairs. FISH can be performed on cells during any stage of the cell cycle without cell culture and is therefore a useful means of rapid diagnosis. Aneuploidy screening using FISH probes for chromosomes 13, 18, 21, X, and Y is used prenatally and postnatally to determine various trisomies or sex chromosome abnormalities in fetuses or newborn babies.²⁵ FISH is also used to detect pathogenic CNVs that result in well-known deletion/duplication syndromes such as chromosome 1p36 deletion syndrome, 7q11.23 deletion syndrome (Williams-Beuren syndrome) and 22q11.2 deletion syndrome (DiGeorge syndrome).^14,17 FISH is useful to reveal mosaicism when cells with chromosomal aneuploidies or CNVs are present along with normal cells. False negative results may occur when atypical or smaller deletions/duplications are present either proximal or distal to the FISH probes.¹⁰⁹ Therefore, a normal FISH result cannot definitively rule out the aforementioned deletions, and CMA remains the gold standard for diagnosis of CNVs.

3.2 Chromosomal microarray for CHD

Microarray-based comparative genomic hybridization (CGH) and single nucleotide polymorphism (SNP) microarrays can identify sub-microscopic CNVs, which are too small to be seen by standard cytogenetic analysis via karyotype. Both array CGH and SNP microarrays can detect genome-wide CNVs; however, SNP microarray can also detect triploidy, loss of heterozygosity, uniparental disomy, and mosaicism.¹¹⁰ The widespread use of CMA as a first-line test for the diagnosis of congenital anomalies, including cardiac defects, has led to the discovery of many CNVs associated with CHD. CMA is routinely used in both the prenatal and postnatal settings for patients with both syndromic and non-syndromic CHDs, and the American College of Obstetricians and Gynecologists (ACOG) states that CMA may be used for prenatal diagnosis in the setting of fetal anomalies, as well as for any person undergoing prenatal diagnosis, in light of its increased diagnostic yield as compared to standard karyotype.¹¹¹ Unlike aneuploidies, the incidence of CNVs is not dependent upon maternal age. In the general population (i.e. fetuses without sonographic anomalies or other risk factors), the incidence of CNVs is estimated to be approximately 0.4%.¹¹¹ In the setting of a fetal CHD, the yield of CMA is estimated to be increased to approximately 3%–25%, but varies greatly depending upon the type of CHD and the presence of other anomalies, soft markers, and/or growth restriction. Many recurrent CNVs are the result of flanking repeat sequences, which predispose to non-allelic homologous recombination and recurrent de novo deletions or duplications encompassing the same genomic interval. CNV-mediated CHDs are associated with a poorer prognosis, as compared to cases of CHD without CNVs. Specifically, the presence of pathogenic CNVs has been associated with significantly decreased transplant-free survival after surgery, worse linear growth, and worse neurocognitive outcomes.²⁸

3.3 Targeted next-generation sequencing (pathogenic variants) for CHD

NGS is a massively parallel sequencing technology used to identify small genetic variation down to the nucleotide level. NGS panels are designed to include multiple genes of interest for a specific disorder (e.g., Noonan syndrome/RASopathy gene panels) or a group of related disorders (e.g., heterotaxy/primary ciliary dyskinesia panels) or may even be a broad CHD panel including several genes associated with non-syndromic and syndromic CHD. Some broad panels may additionally contain one or more heterotaxy/primary ciliary dyskinesia genes (Table 2), and even contain genes associated with other cardiac disorders such as cardiomyopathies, arrhythmias, etc. Test offerings can vary greatly among laboratories, not just in which condition(s) the panel targets, but also in the genes surveyed for the condition(s). Massively parallel sequencing involves simultaneous sequencing of all genes on the panel, precluding the need for consecutive testing of genes of interest. Sequencing is targeted to the coding regions and exon-intron boundaries (≤10 bp of the flanking intronic sequence on either side of each exon) in each gene, as these regions harbor the majority of monogenic disease-causing variants.⁸⁹ NGS panel testing can detect single nucleotide variants and small deletions or insertions (INDELs) in the targeted genes, which are not detectable using karyotyping and CMA. Additional technologies that enable detection of exon-level deletions and duplications are also typically incorporated.

NGS panel testing for CHD is sometimes utilized in the postnatal setting, more so for syndromic CHD, but in general, it is not the standard of care for sporadic non-syndromic CHD as the majority of these cases are expected to have multifactorial etiology with low yield on current genetic testing technologies.¹¹² Studies to determine and compare the utility and detection rate of NGS panel testing in specific CHD cohorts, i.e., non-syndromic vs. syndromic, adult vs. pediatric, and/or familial vs. sporadic have not yet been performed.

Two postnatal studies have utilized NGS panels to test families with seemingly non-syndromic CHD.^61,62 The NGS panel in both studies included 57 CHD-associated genes (the gene set included in each panel was not identical) and identified likely disease-causing variants in 5/16 families (31%) and 6/13 families (46%) respectively. TBX5, TFAB2B, ELN, NOTCH1, and MYH6 were implicated in these two studies, with disease-causing variants in NOTCH1 and TBX5 being common to both studies. The yield from these two studies is likely inflated as they only investigated multiplex families, and they did not utilize the more stringent criteria of American College of Medical Genetics & Genomics (ACMG) for variant interpretation. Nevertheless, the two studies seem to demonstrate the utility of NGS panels for familial CHD.

Currently, NGS panel testing is only sometimes employed in the testing of fetuses with CHD.¹¹³ While a family history of CHD is a consistent risk factor in the identification of fetal CHD, etiology is still more likely to be multifactorial rather than monogenic.¹¹⁴ Prenatal studies to assess diagnostic yield of NGS panel testing in fetuses with familial CHD have not yet been performed. To-date, a single prenatal study assessing the utility of NGS panel testing in the diagnosis of fetuses with sporadic CHD has been conducted.⁷⁰ The study employed a panel of 77 CHD-associated genes to test 44 fetuses with either non-syndromic or syndromic CHD, after they had undergone karyotyping and CMA with negative results. Seven fetuses (15.9%) had a positive result, all of them attributable to de novo variants, while 79.5% had variants of uncertain significance (VUSs). Positive results in this cohort were attributable to pathogenic or likely pathogenic variants in the CHD7 gene for CHARGE syndrome, CITED2 (associated with atrial septal defect and ventral septal defect), MYH6 (associated with atrial septal defect, dilated cardiomyopathy and hypertrophic cardiomyopathy), JAG1 (associated with tetralogy of Fallot and Alagille syndrome), and in two fetuses, KMT2D -associated Kabuki syndrome. The authors noted that the detection rate in the prenatal cohort was lower compared to the two postnatal studies, possibly attributable to the inclusion of sporadic rather than familial CHD cases, differences in type of cardiac lesions tested for, differences in accuracy of diagnosis owing to limitations with in utero phenotyping, and the application of ACMG criteria for variant interpretation in the prenatal study.^61,62

In 2018, the American Heart Association recognized the utility of gene panel testing in cases of suspected monogenic disease with a small differential diagnosis.²⁸ Following this, the ACMG in 2020 published a statement recommending single-gene testing or a phenotype-based gene panel test as the initial or first-line test to be performed when anomalies in the fetus strongly suggest a specific monogenic disorder.¹¹⁵ When a phenotype is genetically heterogeneous (e.g., Noonan syndrome), panel testing has increased detection rate over single gene testing, is less expensive than an approach of reflexing to the next gene of interest after a negative single gene test, and is also less expensive than WES or WGS. Panel tests typically have a short turnaround time of 2–4 weeks, with the added advantage of identifying far fewer VUSs compared to WES/WGS, making result interpretation that much more manageable. Notably, the VUS rate increases incrementally with the inclusion of additional genes on a panel. It would be important for the ordering provider to be aware that the associated risk for CHD for each gene on a panel may vary, and that including low-risk genes or candidate genes will likely result in a greater number of VUSs being identified.

3.4 Whole exome sequencing and whole genome sequencing for CHD

Similar to NGS panels, WES and WGS both utilize massively parallel NGS technology, but in comparison with panels, interrogation by WES is more large-scale while WGS is the most comprehensive (Fig. 2). WES limits examination to the exons (protein-coding regions) and exon-intron boundaries of the approximately 20,000 genes that together comprise 1–2% of the nuclear genome yet harbor 85% of variants contributing to monogenic traits (Fig. 2).¹¹⁶ WES can detect single nucleotide variants (SNVs), small INDELs, and CNVs involving one or more exons (Fig. 2).¹¹⁷ WGS involves testing of most of the 3 billion base pairs in the nuclear genome, and in addition to all exons, also covers non-coding regions (deep intronic regions, untranslated regions or intergenic regions) that harbor an estimated 15% of variants contributing to monogenic disease (Fig. 2).²² Another strength of WGS is the ability to identify copy number variants and gross chromosomal abnormalities or SVs not identifiable by other sequencing methods including panel tests and WES.^22,117,118

WES has been offered as a clinical test since 2011. Some laboratories limit analysis to the coding sequence of known disease-associated genes alone (called a clinical exome), but others may include analysis of variants identified in candidate genes as well.¹¹⁵ WGS is also available as a clinical test, but is not frequently utilized at this time mainly owing to increased cost and longer turnaround time.¹¹⁵ However, based on the ability of WES and WGS to facilitate novel gene discovery, and the extra benefit of identifying variant types by WGS that are not identifiable by other sequencing tests, WES and WGS are commonly used in research.

WES has been utilized in the testing of fetuses with CHD, typically after karyotype and/or CMA have yielded negative results and has demonstrated incremental yield in non-syndromic CHD as well as syndromic CHD with extra-cardiac anomalies. The CODE study (COngenital heart disease and the Diagnostic yield with Exome sequencing) involved antenatal trio WES of a prospective CHD cohort of 197 fetuses, as well as systematic review of 18 published studies in which WES was performed on a total of 636 prenatally diagnosed CHD cases.⁵⁷ In the prospective cohort the overall detection rate was 12.7%, 11.5% in the isolated CHD prospective cohort, and 14.7% in the prospective cohort with CHD and extra-cardiac anomalies; the corresponding pooled yields from systematic literature review were 21%, 11% and 37% respectively. When sub-analysis of studies with at least 20 cases was performed, the incremental yields were similar except for CHD associated with extra-cardiac anomalies, which had higher yield at 49%. Higher WES detection rate for CHD with extra-cardiac anomalies over isolated CHD was demonstrated. The yield was highest for cardiac shunt lesions (41%), followed by right-sided lesions (26%), complex lesions (23%), and left-sided obstructive lesions (18%). Kabuki syndrome and CHARGE syndrome were the most frequently identified disorders in this study. The majority (approximately 70%) of the pathogenic variants identified occurred de novo, and in genes associated with autosomal dominant disease.

A recent study utilized trio WGS in 111 fetuses with structural or growth anomalies including cardiac defects, and was able to detect every pathogenic variant that was identified in the same cohort by concurrently performed CMA plus WES (22/111 cases; 19.8%).¹¹⁷ Additionally, WGS was also able to detect a balanced translocation in a parent which resulted in CNVs in twin fetuses, another case with a dual diagnosis owing to presence of a CNV along with a heterozygous SNV, and a third case of intrauterine cytomegalovirus infection, which are genetic and non-genetic etiologies not detectable by CMA, WES or NGS panel testing. A subsequent prenatal WGS study by Wang et al. also resulted in a detection rate of 19%, identifying not only sequencing variants but also CNVs detected by CMA.¹¹⁹ These results suggest a role for WGS as a first-tier test replacing multiple consecutive genetic tests such as microarray, gene panels, and WES, to provide the most comprehensive analysis in a timely manner essential in the prenatal setting.

Recognizing that diagnostic rate from WES and WGS are comparable to that of karyotyping and CMA, and even higher for certain indications, the International Society for Prenatal Diagnosis (ISPD), the Society for Maternal Fetal Medicine (SMFM), and the Perinatal Quality Foundation (PQF) published a joint position statement acknowledging the use of panel testing, WES and WGS for fetal diagnosis in certain situations. The position statement was recently updated and considers the presence of a single major fetal anomaly or involvement of multiple organ systems suspicious for genetic etiology as indications for prenatal sequencing.⁷¹ They recommend a trio WES approach, and that it be performed for the above prenatal indications when CMA results are non-informative, or that CMA be run in parallel with WES as the latter cannot detect CNVs. The authors of the position statement consider WGS as still applicable in the research setting only for many reasons including difficulty in interpretation of non-coding variants and SVs. The position statement also recognizes a role for the application of sequencing methods in the testing of parents who present for preconception counseling with history of recurrent anomalies in more than one pregnancy, and a sample from the affected proband or fetus cannot be obtained. The ACMG in 2020 has also published a statement that trio WES be considered in the testing of fetuses with ultrasound anomalies for which karyotype and CMA do not provide informative results.¹¹⁵ Further, the ACMG and ISPD publications also provide guidance regarding points to consider in the reporting of WES and WGS findings that may go beyond testing of known CHD-associated genes, to include candidate genes, secondary findings, and incidental findings.^71,115

3.5 Secondary genomic findings for CHD

The ACMG has published recommendations for reporting incidental findings in clinical WES or WGS since 2013.¹²⁰ The most recent version is the ACMG SF v3.0 list published in 2021.¹²¹ It includes 73 genes based on the medical actionability of the associated condition and maximizing the potential to reduce morbidity and mortality. Thirty-three genes related to cardiovascular phenotypes are in the list, which include aortopathy genes (FBN1, TGFBR1, TGFBR2, SMAD3, ACTA2, MYH11), arrhythmogenic cardiomyopathy genes (PKP2, DSP, DSC2, TMEM43, DSG2), catecholaminergic polymorphic ventricular tachycardia genes (RYR2, CASQ2, TRDN), dilated cardiomyopathy genes (TNNT2, LMNA, FLNC,TTN), vascular type of Ehlers-Danlos syndrome (COL3A1), familial hypercholesterolemia genes (LDLR, APOB, PCSK9), hypertrophic cardiomyopathy genes (MYH7, MYBPC3, TNNI3, TPM1, MYL3, ACTC1, PRKAG2, MYL2), and genes for long QT syndrome types 1 to 3 (KCNQ1, KCNH2, SCN5A). It is recommended that pathogenic or likely pathogenic variants in these CHD-related genes be reported in clinical exome and genome sequencing, unrelated to the indication for testing, and with the patient's consent.

3.6 Noninvasive prenatal testing for CHD

Since its launch in 2011, circulating cell-free DNA screening (cfDNA) has revolutionized non-invasive screening for aneuploidy. cfDNA methodology analyzes extracellular DNA fragments in maternal circulation (typically 150–200 base pairs in length), released by the trophoblast cells of the placenta, to screen for chromosome abnormalities using various methodologies, such as massively parallel shotgun sequencing (MPSS) or SNP-based methods. cfDNA most commonly screens for the common viable aneuploidies: trisomy 21, trisomy 18, trisomy 13, and sometimes sex chromosomes aneuploidies. However, the scope of this screening continues to grow, and some laboratories now offer expanded panels which may include other autosomal aneuploidies (such as trisomy 16, trisomy 22, or even “all-chromosome” cfDNA), select microdeletion syndromes, single gene disorders, and fetal Rh status.

cfDNA screening is clinically available as early as ∼9 weeks of gestation and is known to have high sensitivity and specificity for the common aneuploidies. Given this, cfDNA has rapidly become a first-line aneuploidy screening tool for both high- and average-risk pregnancies alike. High-risk indications for cfDNA screening include advanced maternal age, positive serum screening results, or abnormal fetal ultrasound findings such as soft markers or congenital anomalies. While prenatal diagnosis via chorionic villus sampling (CVS), amniocentesis, or fetal blood sampling remains the gold standard for the diagnosis of genetic disease in pregnancy, many patients may opt for cfDNA screening over diagnostic testing, given the potential risk for miscarriage associated with a diagnostic procedure. As such, cfDNA screening is an available option in pregnancies diagnosed with CHD when the pregnant person declines diagnostic testing and remains an excellent screen for some of the most common causes for CHD, including Down syndrome, monosomy X, trisomy 18, and trisomy 13. A study by Salzer-Sheelo et al., in 2021 found that, surprisingly, 44% (n ​= ​24/55) of cases with a prenatal CHD would be detectable by cfDNA screening inclusive of 5 chromosomes: 21, 18, 13, X, and Y.¹²² An additional 15% of cases attributable to 22q11.2 deletion syndrome would (theoretically) also be detectable by cfDNA screening. Of note, the performance of cfDNA screening for other conditions such as microdeletions and single-gene disorders is less well-established but is known to be less sensitive and specific. The positive predictive value is often significantly lower for sub-chromosomal copy number variants, given the lower incidence of these conditions. Screening for select, frequently de novo, autosomal dominant disorders associated with advanced paternal age or abnormal ultrasound findings is also a newly available option; though performance data of this type of screen is even sparser. Therefore, patients opting for cfDNA screening over diagnostic testing should be thoroughly counseled about the limitations of this screening, in that a normal cfDNA result cannot rule out all possible genetic etiologies associated with CHD. Table 3 highlights select disorders that are known to cause CHDs and may be detectable by cfDNA screening.