When defining “genetics”, it is important to make a distinction between the concepts of “genetic counseling” and “genetic testing”. We emphasize from the outset that both concepts are inextricably linked—genetic testing must always be accompanied by correct counseling—but counseling will not result in testing in all cases.

According to the World Health Organization, genetic counseling is defined as “the process through which knowledge about the genetic aspects of illnesses is shared by trained professionals with those who are at an increased risk of either having a heritable disorder or of passing it on to their unborn offspring”.

Genetic counseling in the setting of CHD was introduced more than half a century ago, whereby the most important setting was to inform parents of an affected child about their recurrence risk. Early studies already nicely demonstrated that informing parents in a dedicated counseling process had a beneficial effect.12 More recent research has confirmed the beneficial effects of individualized genetic counseling sessions to the parents of children with CHD with regard to improving knowledge about the causes of CHD and enhancing psychosocial functioning, strongly recommending their inclusion in routine clinical practice.13 With increasing numbers of adults with CHD, the indications for genetic counseling and testing have been expanded and “genetics” are listed as a requirement for adult CHD programs in the recently published American College of Cardiology/American Heart Association guidelines on adult CHD.14

Genetic counselors in CHD are graduate level trained health care professionals who have received training in both medical genetics and counseling with a particular focus on CHD. Genetic counselors will draw a 3-generation pedigree and collect all relevant clinical data from the proband and family members with special attention on miscarriages or neonatal deaths. Apart from their role in family history taking and in counseling patients and families about recurrence risk, risk for a specific syndrome and interpretation of results, genetic counselors may play an important role in triaging patients who should be referred for a complete genetic evaluation.15

Genetic elaboration of CHD requires a multidisciplinary approach in which, in addition to the (pediatric) cardiologist and genetic counselor, the clinical geneticist also plays a crucial role. Clinical geneticists are physicians who have undergone specific training in diagnostic evaluation, management, and genetic counseling. Training programs and certification are nation specific. Clinical geneticists will determine whether the heart defect is isolated or part of a syndrome, which is required to guide genetic testing and to determine the medical approach. Based on large epidemiological studies, syndromic cardiovascular malformations comprise at least 25% of all cardiovascular malformations.4,16 Research in the setting of 22q11 deletion has already shown that cardiologists are less good at assessing syndromes and that clinical genetic evaluation is therefore desirable.17 Once it has been determined whether or not a patient has a syndromic entity, the medical management of syndromic forms can also be better coordinated by a clinical geneticist in the context of a multidisciplinary team —patients with isolated CHD forms are of course best followed up by the (pediatric) cardiologist.8

Another important issue to take into account in the process of genetic counseling and testing is consent. It is essential for any genetic test that the patient (or his or her legal representative) is aware of the benefits and risks of such testing and gives written consent for the test. It is outside the scope of this article to discuss (important) aspects such as incidental findings and presymptomatic testing, but we would like to briefly address direct-to-consumer (DTC) testing.

In many countries, long gone are the days when genetic testing was confined to certified clinical genetic centers (laboratory-directed testing) for which strict rules apply for conducting clinical and molecular diagnostics. As a result of the technical progress in genetic testing on the one hand, and the increasing public demand on the other hand, significant growth has been observed in companies that offer testing known as DTC. These are tests in which samples (blood or saliva) are directly mailed to the laboratory, without preemptive counseling. DTC genetic tests may detect severe and highly penetrant monogenic disorders or genetic variants associated with increased susceptibility for common and complex diseases. There are concerns that variant interpretation from DTC testing may not always be correct. There have already been reports of cases of unnecessary treatment in healthy family members or false reassurance based on incorrect information.18 One study showed that 40% of variants in a variety of genes reported in DTC raw data were false positives.19 This false information severely impacts patients and families in the first place but also overloads genetic counseling services who are consulted to clarify and rectify the results of tests ordered elsewhere.20 It goes without saying that these issues create tension in the context of DTC genetic testing regarding the expectations and normative assessment of communication strategies.21

For these reasons, the European Society of Human Genetics has developed a policy on the advertising and provision of predictive genetic tests by such DTC companies.22 We argue against the use of DTC testing in the CHD genetic testing context.

The seminal discovery of trisomy 21 as the genetic cause of Down syndrome in 195923 introduced genetic testing in CHD. Since then, novel methodologies and technical fine tuning have been instrumental in identifying the genetic cause of (mainly syndromic forms) of CHD. Classic karyotyping with G-banding has a rather low resolution of 3-5 Mb, and is currently only performed for specific indications, such as the confirmation of (mozaic) aneuploidies (Down syndrome, Turner syndrome, mosaic trisomy 8) and familial translocations.

Fluorescent in situ hybridization (FISH) makes use of a fluorescently labeled probe targeting specific genomic regions. It is used for the targeted detection of aberrations below karyotyping resolution, such as the 22q11.2 or 7q11.2 microdeletions in velocardiofacial syndrome and Williams-Beuren syndrome, respectively. A major breakthrough came with the introduction of array comparative genome hybridization (ArrayCGH),24 also called chromosomal microarray analysis. Chromosomal microarray analysis competitively hybridizes shredded DNA of control and patient DNA, labeled with different fluorochromes on an array containing tens of thousands of molecular probes dispersed over the reference genome. Next, automatic reading of differences in color intensities detects genome-wide copy number variations (CNVs), ie, deletions or duplications, as small as 100kb This technique is also referred to as “molecular karyotyping”. Single nucleotide polymorphism (SNP) array is a similar test that uses SNPs to detect regions with a loss of heterozygosity. However, the huge amount of structural variability in the human genome has hampered straightforward interpretation of test results, especially in the years following the introduction of the test in the diagnostic setting. Initiatives such as the Database of Genomic Variants25 have been detrimental in documenting normal variation while databases such as DECIPHER26 have played a prominent role in identifying novel genomic structural defects as the basis of disease. Indeed, molecular karyotyping introduced the concept of “reversed genetics”, a strategy whereby patients with the same genetic variant are compared to identify a genotype-phenotype correlation and delineate novel clinical entities, such as Koolen-de Vries syndrome.27 The most frequently associated structural variants in CHD known to date are reported in more detail below. More recently, chromosomal microarray analysis is being replaced by methods based on low-coverage (“shallow”) genome sequencing technologies (see below).

The combined use of polymerase chain reaction and Sanger sequencing introduced the use of molecular analysis in the clinic. Nevertheless, analysis was expensive and time-consuming. Moreover, the identification of novel genes causing cardiovascular phenotypes was restricted to syndromic forms that could be investigated through linkage analysis in large families with dominant inheritance for the condition (eg, Noonan syndrome28), or in consanguineous families with recessive conditions (eg, Ellis-van Creveld29), while candidate gene approaches only sporadically identified a casual defect, often helped by the previous detection of a microdeletion encompassing the candidate gene (eg, CHARGE syndrome30). A second breakthrough came with the introduction of next-generation (or massively parallel) sequencing (NGS).31 In brief, in NGS, DNA fragments of the region(s) of interest (either a specific panel of genes, the exome, ie, the coding regions of the DNA, or the genome, ie, the whole DNA sequence) are sequenced in parallel and the obtained “reads” are aligned to the reference sequence. The coverage at a certain genomic position refers to the number of times a base at a certain genomic position is independently sequenced. For a reliable interpretation, at least a coverage of 20x is necessary. Short read sequencing methods are primarily used in clinical laboratories because of their cost-effectiveness and low per-base error rate. The application of exome sequencing may than help to either analyze an extensive number of genes known to cause CHD and even to identify novel CHD candidate genes. However, short read lengths (50–500bp) can produce misalignments and misassemblies in areas of high genome complexity, are unable to cover repeats reliably, and impair phasing of variants. Furthermore, the amplification process, which is indispensable in short read sequencing, creates an underrepresentation of bases in areas of high or low guanine-cytosine (GC) content.

Again, due to the huge variability in the human genome, variant interpretation is crucial, based on freely accessible databases and causality prediction with bioinformatic tools (see below). Similar to DECIPHER, databases such as GeneMatcher32 arose to catalyze rare disease discovery by providing a platform to connect clinicians and researchers from around the world who share an interest in the same gene.

Despite copy number analysis (with a resolution of at least 100kb) and high performant sequence analysis using short read NGS, most of the structural variation is still missed. Structural variants comprising CNVs, inversions, and translocations make up to 10% of the genome and contribute greater diversity between 2 human genomes than any other form of genetic variation and may affect expression of genes.

In addition to large chromosomal defects such as translocations, inversions and especially CNVs, smaller cryptic structural variants (ranging from 50bp to 50kb) can also cause human diseases by affecting gene function or expression; eg, structural variants >20 kbp are up to 50-fold more likely to affect the expression of a gene compared with an SNV.

Third generation genome sequencing platforms use long read sequencing (LRS, >10kb reads) and enable the elucidation of structural variations at a previously unparalleled resolution and overcome most of the shortcomings of short read sequencing.33 This will eventually lead to a single analysis to cover most genomic DNA variation within the near future.

Third generation genome sequencing can be combined with other NGS approaches, such as transcriptome sequencing (library of expressed genes in a certain cell type) to identify variation at the level of expression and splicing. Nevertheless, the main hurdle will remain the interpretation of the variants, which often requires additional validation and which will remain the main burden in the diagnostic application of these techniques (see below).

With the technical progress of genetic testing as described above, now enabling simultaneous testing of large numbers of genes, the temptation has been great to effectively set up more extensive panels for specific disorders. Commercial genetic laboratories in particular have taken this path and now offer panels for CHD that contain> 100 genes, for example.

However, some caution is advised in this trend. For multiple disorders, there is strong evidence that testing more genes inevitably leads to detection of more “variants of unknown significance” or VUSs, the interpretation of which is not easy and sometimes even risks creating unnecessary anxiety and discrimination in patients.34,35 When selecting genes for inclusion in diagnostic panels or reporting from exome/genome sequencing, the clinical validity, ie, the strength of evidence that variation in that gene predisposes to the disease, needs to be carefully considered. A framework for semiquantitative assessment of gene-disease validity has been developed for many diseases (but not yet for CHD) by the Clinical Genome Resource, or ClinGen. In this framework, genes are classified into prespecified tiers based on the clinical, genetic and experimental evidence, along with discussion and consensus of clinical domain experts. These clinically validated genes can be used to prioritize genes for research and inform which genes should be included in disease panels.36,37

The possibility of generating large amounts of genetic data through broad genetic screening has allowed rapid and more accurate genetic diagnosis. However, as already mentioned, each analysis yields multiple genetic variants (up to 50 000 variants per exome), which need appropriate interpretation. Distinguishing benign from pathogenic variants is essential for translating genetic results into clinical practice but remains challenging. Moreover, the number of VUSs is still too high. In 2015 the American College of Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) proposed a guideline to classify genetic variants for Mendelian disorders.38 Classification is based on 28 criteria to finally classify a variant as benign (class 1), likely benign (class 2), variant of unknown significance (class 3), likely pathogenic (class 4) or pathogenic (class 5), where classes 2 and 4 provide greater than 90% certainty of a variant either being benign or disease-causing. The criteria include clinical findings, data retrieved from large human exome databases such as gnomAD and ExAC and data addressing the structural effect of a variant on the DNA/protein level.

Correct and detailed clinical findings are required, not only of the proband, but often also of family members. Cooperation with—and from—family members is therefore highly important, both in a diagnostic setting and in the further communication of the results. Further segregation of identified variants (both copy number variants and single nucleotide variants) requires verification in first-degree relatives, in which, in addition to DNA studies, clinical cardiovascular examination is also required to correctly assess the status of the individual. It is important to communicate this properly to the patient from the beginning of the counseling and testing process.

If an abnormal test result is confirmed, it is also recommended to check the result further with additional family members. Caregivers are not permitted by law to contact family members; contact must go through the patient (or his or her representative) and it is also important to communicate this correctly to the patient when discussing the results.

In addition to clinical criteria in patients and family members, the molecular characteristics of the variants are also taken into account. Computational analysis of a variant with modeling of the expected effects of the gene or variant on protein structures and function can provide supporting evidence to establish pathogenicity. Despite careful interpretation and international initiatives for data curation, the clinical consequences for many variants obtained through in-depth molecular analysis remain unknown. Several tools, including transcriptome, proteomic, metabolomic, lipidomic and methylomic analysis, may help to identify the molecular consequences. However, the respective (multi)omic signatures are unknown for most genes defective in CHD. In addition, some genes may only be relevant during cardiac development and postnatal testing on other tissues might be irrelevant. Also, animal modeling for specific diseases is time-consuming, expensive, and impossible in the clinical setting. However, direct mutagenesis using the CRISPR_Cas9 techniques is becoming increasingly efficient and may eventually be helpful in the interpretation of genomic variants.

Although the ACMG/AMP guidelines definitely introduced major improvements in the interpretation of genetic variants, they often still leave room for subjective interpretation and therefore several groups have proposed more gene specific classifications.39

Since the publication of the guidelines, several tools have been developed to aid the interpretation of genetic variants (table 1). Moreover, several online repositories are also available to consult variants which have been previously classified by other laboratories (table 1). Caution, however, should be exercised when consulting these repositories since variant interpretation remains subjective and is not always performed by experienced groups. Clinvar, one of these archives, not only provides an interpretation of a specific variant, it also provides a level of review supporting the assertion of clinical significance.

Genetics is a dynamic and rapidly evolving field. For many clinical phenotypes, new genomic data are regularly discovered and test findings issued today may be outdated tomorrow. Therefore, regular and careful reconsideration of genetic counseling and testing should take place, especially in those individuals/families with a high level of suspicion but without identification of a genetic defect. In the same line, genetic variant interpretation can change over time and previously found genetic variants (especially VUSs) should be regularly reassessed in light of new published data.57

Again, and even more so than in the initial counseling process, retesting and communication of altered test results should be undertaken with care to ensure that the results are correctly interpreted by patients and their families.58

Knowledge of an underlying genetic problem can be important to diagnose and improve management of extracardiac complications in children and adults with CHD. For example, patients with 22q11 deletion may have decreased T cell immunity and therefore be at risk of severe infectious diseases; patients with Alagille syndrome can suffer from ophthalmologic and liver complications; children with Noonan syndrome have short stature and may benefit from growth hormone therapy.

Some genetic defects have been associated with an increased risk of other cardiovascular complications. A known example is the association between ASD and conduction disorders in those patients carrying a pathogenic variant in NKX2.5. These patients are more likely to develop atrioventricular block, ventricular dysfunction, and sudden cardiac death.65 Various types of CHD have been associated with genes causing familial cardiomyopathy. Some examples are the ACTC1, MYH6, and MYBPC3 genes.66–68

When discussing recurrence risk, knowledge of an underlying genetic entity is essential. Risk estimates will also vary according to the setting of siblings or offspring and, in some cases, risks differ for fathers and mothers. If there is a known genetic disorder, recurrence risk in a sibling will greatly depend on the type of inheritance and on the de novo character of the anomaly. For those genetic disorders with an autosomal recessive pattern, the recurrence risk is 25%. For those with an autosomal dominant pattern, the recurrence risk will be 50% if one of the parents is affected and up to 1% in case of a de novo variation.69 For adult patients with CHD with a known genetic condition, the recurrence risk for the offspring will be 50% if the disorder is autosomal dominant. For those patients with autosomal recessive anomalies, the recurrence risk for children is similar to that in the general population, but each child will be carrier of 1 allele with the anomaly. If no underlying genetic anomaly is found, the recurrence risk is still higher than in the general population. Considering all CHD together, the recurrence risk for siblings is estimated at 2.1% and for the offspring at 4.4%, with women in general having a higher recurrence rate than men.70Table 4 summarizes the known epidemiological recurrence risk for the most common CHD in the absence of a known underlying genetic cause. Some lesions such as heterotaxia, right ventricular outflow track obstruction and atrioventricular septal defect, present higher familial clustering. The recurrence risk is 80-fold to 25-fold higher than in the general population.4

Another important aspect in terms of counseling family members is the need for further clinical assessment. For some lesions, such as LVOT obstruction, there is known variation in the clinical spectrum ranging from an asymptomatic BAV to a severe hypoplastic left heart syndrome.75 In one study, the relative risk of having a BAV in a parent or a sibling of a patient with a LVOT obstructive lesion was 5.05 (95% confidence interval, 2.2-11.7).11 This high incidence together with the fact that many of the complications of left-sided lesions are treatable or preventable, led to the rationale that first degree family members of patients with LVOT obstruction should undergo echocardiographic assessment.76 Currently no systematic screening of family members is recommended for other forms of nonsyndromic CHD.

Prenatal diagnosis is possible whenever a genetic cause of the CHD has been identified. In this case, transmission to the next generation can be prevented through preimplantation diagnosis, a technique based on in vitro fertilization, which occurs before pregnancy and selects the unaffected embryos.

An alternative is prenatal testing, which conventionally implies chorionic villus or amniotic fluid sampling in the early stages of pregnancy. In more recent years, noninvasive prenatal testing (NIPT) has emerged as a noninvasive alternative for prenatal testing. NIPT was primarily developed to detect trisomy 21 in the fetus early in pregnancy with high specificity (> 99%) and sensitivity (> 99%) in the absence of fetal anomalies. During pregnancy, cells from the placenta (containing fetal DNA) lyse into the maternal circulation. The test is based on the relative number of reads that map to a certain chromosome in maternal plasma (cell free DNA). Hence, the test can detect other aneuploidies such as trisomy 13 and 18, Klinefelter (47,XXY), Turner (45,X0), or triple X (47,XXX) syndrome, albeit with a somewhat lower sensitivity and specificity. Technological fine tuning will eventually render NIPT suitable to detect de novo variants (both CNVs and SNVs)77,78 and targeted testing of inherited variants.79 It goes without saying that broad scale application of NIPT requires careful consideration of ethical issues. In both instances of prenatal testing, termination of pregnancy may be considered if the results of the test are abnormal test.

Some couples may choose not to undergo prenatal diagnosis, in which case, the possibility of genetic screening in the newborn should be discussed. If no underlying genetic cause of the CHD has been identified or no prenatal testing has been performed, fetal echocardiography is recommended. This should be performed in a specialized center at 18 to 20 weeks of pregnancy.80