Cardiac involvement in genetic disease. Topics covered:

• Essentials
• Introduction
• Syndromic congenital heart disease
• Connective tissue disorders
• Further reading

Essentials

Inherited diseases of the contractile machinery of the heart, which lead to familial hypertrophic and dilated cardiomyopathy, are covered in Chapters 16.7.2 and 16.7.3 and ion-channel mutations that underlie long-QT syndromes and other inherited causes of paroxysmal ventricular tachycardia are covered in Chapter 16.4.

The first part of this chapter deals with developmental syndromes that include congenital cardiac defects, with coverage restricted to a few relatively common disorders that are seen in adult patients. The second part describes the two common connective tissue disorders—Marfan’s and Ehlers–Danlos syndromes—and the recently described Loeys–Dietz syndrome that shares some pathogenetic mechanisms with Marfan’s. A number of other heritable diseases that affect the heart are listed in a table, without discussion in the text. Haemochromatosis (Chapter 12.7.1) and Friedreich’s ataxia (Chapter 24.7.5) are discussed elsewhere on this website; the others, though important to other organ systems, offer little opportunity to the cardiologist for diagnosis or management.

Syndromic congenial heart disease

Aneuploidy disorders

The two commonest chromosomal disorders in adult patients are Down’s and Turner’s syndromes, and each includes characteristic cardiac abnormalities. A third, Klinefelter’s syndrome, does not.

Some 25 to 50% of patients with Down’s syndrome have congenital heart disease. The characteristic lesion, present in about half of the affected hearts, is atrioventricular canal defect. This ranges from the relatively simple primum atrial septal defect to the complete type, in which the defect involves both the atrial and ventricular septa, between which there lies a single atrioventricular valve ring. In other patients, ventricular septal defect, tetralogy of Fallot, and persistent ductus arteriosus are seen in roughly equal numbers. Patients with Down’s syndrome undergo heart surgery most easily when they are infants, and the tendency has shifted from the nihilistic approach of past years to correcting serious cardiac malformations early in life.

Turner’s syndrome causes abnormalities of the aorta, the two principal lesions being coarctation and congenital abnormalities of the aortic valve, usually a bicuspid valve. Most patients with coarctation have a bicuspid aortic valve as well, and patients with either lesion frequently have some degree of annuloaortic ectasia. In some patients with Turner’s syndrome, the whole aorta is abnormal—either hypoplastic or weakened by the presence of cystic medial necrosis. Aortic dissection may occur, and aortic surgery, e.g. to repair coarctation, can sometimes be very difficult, owing to the fragile nature of the aortic wall. Other congenital heart abnormalities are not common in Turner’s syndrome, except for anomalies of pulmonary venous return.

Mendelian syndromes that include congenital heart disease

Noonan’s syndrome

Noonan’s syndrome (OMIM 163950) is the most common heritable syndrome that characteristically causes congenital heart disease. The syndrome shares some features with the Turner phenotype, and the two were confused between 1930 and Noonan’s studies in the 1960s, which coincided with the advent of cytogenetics. In 1963, Noonan described a small series of patients with pulmonary stenosis who shared a characteristic facial appearance. Since then, the phenotype has been well described and shown to be associated with a normal karyotype and autosomal dominant inheritance. Cardiac involvement has been recognized to include not only pulmonary stenosis, but also a wide variety of other lesions, much wider than in Turner’s syndrome. Mutations in an intracellular signalling molecule protein tyrosine phosphatase SHP-2 (the gene is called PTPN11) account for 40% of cases, but the disease is genetically heterogeneous. The pathogenetics are further complicated by both clinical and genetic overlap. Another syndrome—LEOPARD syndrome (OMIM 151100)—is also caused by PTPN11 mutations, and the Noonan phenotype is closely related to disorders caused by mutations affecting other members of the RAS-ERK intracellular signalling cascade.

Patients with Noonan’s syndrome are of short stature, with a facies that is variously described as elfin or triangular. There is ocular hypertelorism, and the palpebral fissure may slope downwards (the antimongoloid slant), which may be emphasized by ptosis or an epicanthal fold. The ears are set low and rotated forwards so that the lobes are prominent, and there is characteristic webbing of the neck—the most obvious of the features that may lead to confusion with Turner’s syndrome. Pectus deformities are common, as are other miscellaneous skeletal abnormalities, including cubitus valgus. Patients with Noonan’s syndrome are prone to develop keloid scars. Cryptorchidism is common, as is delayed sexual maturation, but not infantilism as in Turner’s syndrome. Unlike Turner’s syndrome, many patients with Noonan’s syndrome have a degree of mental retardation, but this is quite variable. Among this author’s patients with Noonan’s syndrome are a physician, an architect, a certified accountant, and a high-school mathematics teacher.

The frequency of cardiac involvement in Noonan’s syndrome is unknown because the diagnosis is so easily missed in the absence of congenital heart disease. The most characteristic lesion is pulmonary stenosis, but in contrast to the almost stereotypical cardiovascular findings in Turner’s syndrome, the range in Noonan’s syndrome is broad. In many patients, the stenotic pulmonary valve leaflets are not simply fused, as in nonsyndromic pulmonary stenosis, but may be dysplastic, thickened, and immobile—unsuitable for simple balloon or surgical valvotomy. Other congenital lesions found in Noonan’s syndrome are ventricular and atrial septal defects, tricuspid atresia, single ventricle, and abnormalities of the left ventricle, including congenital mitral stenosis, subaortic stenosis, and a combination of these two lesions. The electrocardiogram often shows a superior axis (left-axis deviation), even when there is pulmonary stenosis and right ventricular hypertrophy.

The most ominous complication of Noonan’s syndrome is cardiomyopathy, taking the form of myocardial hypertrophy complicated by progressive fibrosis. This leads, over the course of 5 to 15 years, to low cardiac output with very high ventricular diastolic pressures—the pathophysiology of restrictive cardiomyopathy. Since the valvular abnormalities are for the most part correctable, this hypertrophic restrictive cardiomyopathy is the main factor limiting life expectancy.

Familial supravalvar aortic stenosis and Williams’ syndrome

Familial supravalvar aortic stenosis is caused by loss-of-function mutation or deletion affecting the gene for elastin located on chromosome 7. Affected patients develop a tight, fleshy constriction of the aorta, or sometimes the pulmonary artery at the level of the sinotubular junction above the semilunar valve. In some patients, both great arteries are affected. Supravalvar aortic stenosis can lead to severe left ventricular outflow obstruction, with left ventricular failure or even sudden death. This is not a setting for balloon dilation or stenting, but the results of surgery are good, for either lesion or for both.

Williams’ syndrome (OMIM 194050) is one of the best-documented examples of a contiguous gene phenomenon seen in adult medicine. It is caused by macrodeletions of chromosome 7 that include the elastin gene. Hence, Williams’ syndrome includes the cardiovascular features of familial supravalvar aortic stenosis described in the previous paragraph. In addition, more far-reaching effects caused by deletion of contiguous genes accompany these vascular abnormalities. The full syndrome comprises a characteristic facial appearance, with round, blue eyes, a distinctive stellate pattern of the irises, depression of the nasal bridge, outwards tilting of the nostrils, abnormal dentition, and big lips, together with small stature, mental retardation, and a history of infantile hypercalcaemia. Mental retardation in Williams’ syndrome takes on very individual forms, the patients often being articulate and socially adept: several purported idiot savants have had Williams’syndrome. As in the purely cardiac syndrome, surgery may be required to relieve severe left (or right) ventricular outflow obstruction.

DiGeorge and velocardiofacial syndromes (chromosome 22 deletion syndrome)

DiGeorge syndrome (OMIM 188400), described in 1965, comprises abnormalities of the parathyroid glands, absence or hypoplasia of the thymus, and conotruncal abnormalities of the heart such as pulmonary atresia and severe forms of tetralogy of Fallot. A number of affected patients have learning disabilities or schizophrenia. It was recognized soon after the original description that the syndrome is generally caused by deletions in a region of chromosome 22.

Velocardiofacial syndrome (OMIM 192430), or Shprintzen’s syndrome, described in 1981, comprises similar cardiac abnormalities along with cleft palate, a characteristic facies, and learning difficulty. It has since proved to be caused by deletions in the same region of chromosome 22, now often referred to as the DiGeorge critical region (DGCR). A third syndrome, known as ‘conotruncal anomalies face’, is also linked to this site.

With a broad spectrum of phenotypic variation, and deletions that are often quite large, it was suspected for some time that these syndromes are related manifestations of a contiguous gene phenomenon, just as in Williams’ syndrome. However, it has emerged that the size of the deletion does not predict the extent of the phenotype, and that within a family the same (presumably stable) deletion can be the cause of a wide range of phenotypes. Two candidate genes lie within the DGCR—TBX1 and UFDIL; it remains to be seen whether either can be implicated as the cause of the entire group of phenotypes.

Heart–hand syndromes

The two commonly recognized heart–hand syndromes are Holt–Oram syndrome and Ellis–van Creveld syndrome.

Holt–Oram syndrome (OMIM 142900)

Holt–Oram syndrome, inherited as an autosomal dominant trait, was described in 1960. It includes a secundum atrial septal defect and skeletal abnormalities, principally affecting the upper limbs and shoulder girdle, never the legs, and usually more pronounced in the left arm. Within a family, affected individuals may have skeletal abnormalities, congenital heart disease, or both. The limb abnormalities cover a wide spectrum from just a triphalangeal thumb to phocomelia. Abnormalities of the hand and forearm always involve the radial side and thumb (in contrast to Ellis–van Creveld syndrome). The characteristic cardiac abnormality is fossa ovalis (secundum) atrial septal defect, but affected patients may have other relatively simple lesions, e.g. ventricular septal defect or pulmonary stenosis.

Holt–Oram syndrome is caused by mutation in a transcription factor, TBX5, a close homologue of a transcription factor seen as phylogenetically far away as the fruit fly, where mutations produce abnormalities of the wing.

Ellis–van Creveld syndrome (OMIM 225500)

Ellis–van Creveld syndrome is inherited as a recessive trait, hence the more complete clinical descriptions have come from studies in genetically circumscribed communities, notably the Old Order Amish of Pennsylvania where, thanks to a founder effect, the gene is common and homozygotes abound. The syndrome, described in 1940, includes dwarfism, caused mainly by shortening of the forearms and lower legs, and symmetrical polydactyly affecting the ulnar side with accessory sixth and even seventh digits attached to or beyond the little finger. Cardiac involvement is very common, present probably in three-quarters of homozygotes. The characteristic lesion is common atrium—a lesion that has the appearance, on echocardiography and to the surgeon, of a very large primum atrial septal defect. A few patients have more complete forms of atrioventricular canal defect, and—at least among the Amish—there is a high perinatal mortality rate among affected infants, suggesting the possibility of still more extensive cardiac involvement. The gene has been mapped to chromosome 4 and sequenced, but the protein’s function is unknown.

Connective tissue disorders

Marfan’s syndrome

Thanks principally to the work of McKusick and his collaborators, beginning in 1955, Marfan’s syndrome (OMIM 154700) has become the paradigm for the clinical, genetic, and molecular investigation of the heritable disorders of connective tissue. The importance of the syndrome is heightened by the fact that its recognition and treatment have had a dramatic impact on survival among those affected. Untreated, patients had a median survival into the fourth decade before death from aortic dissection and rupture. Today, affected patients have a near-normal lifespan, and there are reasons to hope that recent advances in understanding the molecular pathogenesis may yet make this genetic disease, in a sense, ‘curable’.

In 1896, Marfan described a weak, generally hypotonic child, with what he termed arachnodactyly. In the century since, it has been appreciated that the syndrome is mendelian and pleiotropic, involving several apparently unrelated organs whose common feature proves to be the importance of elastic tissue to their structural integrity. Ocular involvement, with the lens subluxed because of failure of its suspensory ligament, was recognized early in the 20th century. Cardiovascular involvement was noted incidentally in the 1940s, and studied systematically from the 1950s onwards. Skeletal involvement includes—besides long limbs and arachnodactyly—scoliosis and other abnormalities of the thoracic cage. The sternum may be pushed outwards or inwards by the abnormally long ribs, hence pectus carinatum and/or excavatum, often asymmetrical. Skin involvement is identified by light-coloured striae, which should be looked for over the deltopectoral groove and the flanks. Less common findings are dural ectasia, which can sometimes be so marked as to cause radicular symptoms, and pulmonary involvement leading to spontaneous pneumothorax or apical blebs. In severely affected children, like the one Marfan described, there may be generalized weakness and hypotonia.

The characteristic cardiovascular findings in Marfan’s syndrome are aneurysmal dilatation of the aorta, and occasionally other large arteries, and floppy mitral valve. The former was recognized in the 1920s, but not really addressed until McKusick showed that it was the principal cause of early death in the disease. Shortly afterwards, echocardiography became available to identify and follow these abnormalities, and surgical techniques were developed by Bentall and Gott to repair the aneurysms. Until then, median life expectancy for men with Marfan’s syndrome had been 45 years, for women a year or two longer.

Fibrillin-1 mutations

The syndrome (OMIM 134797) is caused by mutations of the fibrillin-1 gene (FBN1) on chromosome 15. It has recently emerged that besides a purely structural role, one that could hardly be replaced by any form of treatment, fibrillin-1 acts to modulate cell-to-cell signalling during development and, at least in a mouse model, after birth. The dominant negative hypothesis, in which the mutated fibrillin protein was believed to have its effect by interfering with polymerization of the product of the nonmutated allele, thus proves to have been an oversimplification. Rather, the pleiotropic effects of FBN1 mutations prove to be mediated through paradoxical up-regulation of the signalling pathway transforming growth factor β1 (TGFβ1), which is modulated by fibrillin-1. Such findings have led to the likelihood of pharmacological treatment for the disease. Losartan, an angiotensin-II receptor blocker, which like the other members of its class also blocks TGFβ1 signalling, has been shown dramatically to prevent aortic dilation in a mouse model, also in a small clinical cohort study.

The fibrillin molecule is large, and most of the disease-causing mutations have yet to be described, hence genetic diagnosis by screening for known mutations is often not possible and diagnosis usually depends on applying clinical criteria. There are many polymorphisms within the gene, so in some kindreds it is possible, by tracking particular haplotypes, to determine which is associated with the disease and therefore contains the pathogenetic mutation. This has allowed diagnosis of the syndrome in individual family members in whom the clinical findings were uncertain, and has been used for prenatal diagnosis. Furthermore, the technique makes it possible to infer the existence of a fibrillin-1 mutation in kindreds where the phenotype has not met clinical criteria for Marfan’s syndrome; if aortic ectasia segregates with a particular fibrillin-1 haplotype, then the chances are high that a fibrillin mutation somewhere in that copy is the pathogenetic mechanism.

Diagnostic criteria

The clinical diagnosis of Marfan’s syndrome rests on major and minor criteria. In an index case, involvement of three organ systems is required, with major criteria in two. Major criteria can be aortic aneurysm, lens subluxation, characteristic skeletal abnormalities, or dural ectasia. Minor criteria can be striae, mitral valve prolapse, joint laxity, the facies, or moderate pectus excavatum. Characteristic skeletal abnormalities can be arachnodactyly (encircling the wrist with the thumb and little finger, the ‘wrist sign’, and making a fist with a protruding thumb, the ‘thumb sign’), marked pectus deformity, increased wingspan to 5% more than the height, and scoliosis. In the relative of an index case, the positive family history becomes another major criterion.

In clinical practice, determining whether a patient satisfies these criteria may be fairly subjective and requires experience with the syndrome. Often, it is enough to know whether or not there is cardiovascular involvement, and there are numerous families with aortic aneurysms or ectasia who do not satisfy clinical criteria for Marfan’s syndrome, yet whose long-term management is identical. Indeed, in a busy cardiac surgery practice with expertise in aortic root replacement, such ‘nonsyndromic’ familial aortopathy represents a significant proportion of patients treated, and some of these families have yielded other loci as sites for the cause of their disease. On the other hand, a lanky patient who has a normal aorta needs only infrequent follow-up, even though there may be a suspicion that he has a mild case of the syndrome.

Clinical management

Patients with Marfan’s syndrome should be followed up with annual or 6-monthly echocardiograms to examine the aortic root. If there is reason to suspect that the aorta may be dilated above the echo plane, then CT scanning or MRI is required at least once to validate the echo measurement. When the maximum measurement across the aorta reaches 5 cm, specialists generally recommend surgical replacement of the aortic root, to prevent aortic dissection (see Chapter 16.14.1), which becomes a real risk once the dimension reaches 6 cm. The traditional and very successful approach is with the composite graft, whereby a mechanical aortic valve prosthesis—to which is indissolubly attached a tubular vascular prosthesis—is used to replace the entire aortic root and annulus. The coronary artery ostia are excised from the native aorta and reattached to the prosthetic root. Recently, to avoid anticoagulation in certain patients, there has been interest in a valve-sparing technique of root replacement in which a vascular prosthesis is fitted snugly over the aortic valve commissures, with the native leaflets suspended in their normal anatomical arrangement. Long-term success with this approach will depend on the degree to which the valve leaflets themselves degenerate because of the connective tissue abnormality. The Ross (pulmonary autograft) procedure is not appropriate in Marfan’s syndrome. After surgery, and especially in patients whose surgery was done as an emergency for dissection, follow-up is with periodic imaging by CT or MRI to keep the remaining aorta under surveillance. Management of mitral prolapse and regurgitation in Marfan’s syndrome is the same as in other patients. Surgery is required for severe or symptomatic regurgitation; mitral valve repair has proved surprisingly successful.

It is usual to treat patients who have aortic involvement with β-adrenergic blockers to slow the progression to aneurysm, but the benefit is probably modest. In mice with fibrillin-1 mutations in which the Marfan phenotype is well reproduced, β-adrenergic blockade had only slight effect on aortic ectasia. This was in contrast to the dramatic effect of losartan, alluded to in a previous paragraph, and many clinicians now recommend use of this drug. We generally advise against excessively demanding sports, particularly competitive basketball, but in all affected children it is important to balance the risks of aortic disease against the importance of normal psychological development. Pregnancy is not contraindicated in all women with Marfan’s syndrome, but genetic counselling should be offered, and it is advised that women not become pregnant if the aorta is enlarged to over 4 cm. Indeed, aortic dissection has been reported in a very few affected patients during pregnancy, even when they did not previously have aortic enlargement. In this autosomal dominant condition with high penetrance, the risk for the offspring of affected mothers or fathers is 50%. This can be mitigated, when the disease-causing mutation has been identified, by preimplantation genetic diagnosis or even by screening the fetus by chorionic villus sampling.

Loeys–Dietz syndrome

If the pathogenesis of Marfan’s syndrome lies with abnormal TGFβ signalling, then it should not come as a surprise that mutations in the TGFβ receptors also cause abnormalities of vascular and other tissues. Recently, this was confirmed in the description of Loeys–Dietz syndrome (OMIM 609192, 610380, 610168, 608967), a disease that shares some aspects of the Marfan’s phenotype and is associated with mutations of either of the two TGFβ receptors. Patients with Loeys–Dietz syndrome have more diffuse vascular involvement than those with Marfan’s, and may have dissection even in vessels that are only mildly dilated. In this, they resemble patients affected by the vascular form of Ehlers–Danlos syndrome, and the phenotypes may be very difficult to distinguish. Prominent nonvascular features include ocular hypertelorism with malar hypoplasia, bifid or broad uvula, cleft palate, arachnodactyly, scoliosis, and pectus excavatum, yet excessive height is uncommon.

Ehlers–Danlos syndromes

In the early part of the 20th century, Ehlers and Danlos independently described an association between hyperextensibility of the skin, atrophic scarring, and hypermobility of the large joints. In the following 75 years, numerous accounts were published of what we now recognize to be a group of related conditions, so that by 1988 a new classification of the Ehlers–Danlos syndrome included ten separate phenotypes in an unwieldy classification. For practical purposes, clinicians distinguish ‘classical’ Ehlers–Danlos, formerly types I and II, from the potentially fatal ‘vascular’ form, previously type IV.

Classical Ehlers–Danlos (OMIM 130000, 130010)

This is characterized by skin elasticity, abnormal scars, and joint hypermobility, and is inherited as a dominant trait. Skin hyperextensibility is obvious, e.g. on tugging at the side of the neck or face. Joint laxity is much more marked than in Marfan’s syndrome, and allows for tricks like placing the feet behind the head or other contortionist performances, besides permitting a remarkable span on the piano or violin. It also leads eventually to severe degenerative arthritis, often with considerable deformity of the hands. Ability to touch the nose with the tip of the tongue may also provide a clue to the diagnosis. The third aspect of the phenotype, atrophic scarring, if not immediately apparent, may be sought by inspecting the knees for the results of minor childhood injuries: there one may find characteristic wide, atrophic (‘cigarette paper’) scars still obvious from bygone years.

Cardiovascular findings in classical Ehlers–Danlos are for the most part benign. Affected patients frequently have mitral valve prolapse, as do many people with joint laxity who do not have diagnosable Ehlers–Danlos syndrome. Relatively few progress to develop severe mitral reflux or to the point of requiring surgery. Enlargement of the aortic sinuses of Valsalva may occur, but only rarely is this severe or progressive. Surgical replacement of the aortic root, as is performed in Marfan’s syndrome, is unusual in Ehlers–Danlos syndrome.

Vascular Ehlers–Danlos (OMIM 130050)

This, by contrast to classical Ehlers–Danlos, is a potentially fatal condition, with a natural history worse than Marfan’s syndrome. It is genetically and biochemically well characterized: patients have mutations in the COL3A1 gene which encodes for type III procollagen, with inheritance as a dominant trait. The collagen defect leads to excessive fragility of blood vessels, the bowel, and the uterus, and the natural history of the condition is to present with spontaneous rupture of one of these three (in the case of the uterus, during pregnancy). Because of the intrinsic weakness of the affected tissues, surgical repair is challenging and these complications frequently prove fatal. Furthermore, in patients who have once undergone vascular or bowel rupture, the likelihood of a second event is high.

The joint and skin features of the vascular phenotype are less obvious than those of the classical form. Joint hypermobility is not seen, nor the resulting arthropathy. However, the skin feels soft and thin, and is abnormally translucent such that the veins are easily seen through it as one examines the shoulders and upper chest. The face is often thin and bony and the nose pinched.

Vascular complications are hard to anticipate. Aortic ectasia and aneurysm occur only in a few patients. Moreover, arterial rupture—as common as dissection—may occur in medium-sized vessels of the brain, thorax, or abdomen just as often as the aorta. In these regards, the vascular complications of this disease are comparable to those of the Loeys–Dietz syndrome. In affected patients and their families, detailed genetic evaluation is important and should include screening of COL3A1 and biochemical analysis of type III collagen obtained from skin biopsy and cultured fibroblasts. If these prove negative, then screening the TGF-β receptor genes would be appropriate.

Other heart-related connective tissue and metabolic disorders

Osteogenesis imperfecta causes aortic and mitral regurgitation, as do several of the mucopolysaccharidoses (Table 1). It is striking, particularly in the case of osteogenesis imperfecta, how healing is almost nonexistent where there is foreign material. If the opportunity arises, even years later, to inspect the operative result in a patient who has undergone valve replacement, the sutures look as though they had only just been placed, with minimal endothelial reaction and scar-tissue formation.

Further reading