|Home | About | Journals | Submit | Contact Us | Français|
In recent years, there have been many advances in the treatment of cardiac disease in children with Marfan's syndrome. Early diagnosis, meticulous echocardiographic follow‐up and multidisciplinary assessment are essential. Medical treatment with β‐blockers is probably helpful in most children with aortic root dilatation. Research on TGFβ signalling and the potential treatment role of TGFβ antagonists may lead to exciting new treatments, but the results of clinical trials are awaited. In managing the cardiovascular complications of Marfan's syndrome, the paediatrician has to walk a difficult path. On the one hand, restrictive lifestyle advice and drugs may need to be prescribed, often in the context of a family history of major surgery or even sudden death. On the other hand, it is essential to encourage the often asymptomatic child to develop and mature as normally as possible.
Marfan's syndrome was first described in 1896 by the French paediatrician, Professor Antoine Marfan.1 He described a 5‐year‐old girl, Gabrielle, who had the typical phenotype we now associate with this condition. In 1912, Salle described mitral valve abnormalities and heart dilatation in an infant with heart failure, but it was not until 1943 that the typical cardiac abnormalities (aortic dilatation and dissection) were linked to the Marfan phenotype.2 Cardiovascular disease accounts for >90% of premature deaths in patients with Marfan's syndrome.3 Over the past few years, there have been important advances in the understanding of the development of Marfan's syndrome, and this has led to the investigation of new therapeutic targets. The incidence, pathophysiology, diagnosis and treatment of the cardiovascular abnormalities that occur in Marfan's syndrome are reviewed in this paper. In particular, the diagnosis and treatment of cardiovascular complications in children with Marfan's syndrome is emphasised.
Marfan's syndrome is an autosomal dominant disorder of connective tissue, which has both high penetrance and variable severity. The incidence of Marfan's syndrome is around 2–3 per 10000 individuals.4 In 25% of individuals there is no family history, which suggests that the condition has presented de novo.
Marfan's syndrome is caused by an abnormality of fibrillin, a 350‐kD glycoprotein, which is the main structural component of microfibrils. Microfibrils provide a supporting scaffold for the deposition of elastin throughout the body. Fibrillin is present in many other tissues including lung, dura mater, skin, tendons, ciliary zonules of the lens, myocardium, heart valves and periosteum. Abnormalities in these fibrillin‐containing tissues are found in most patients with Marfan's syndrome.
In 1991, mutations in the fibrillin‐1 gene (15q21.1) were found to cause Marfan's syndrome.5 For many years, this was thought to be the only cause of the Marfan phenotype. In 2005, however, it was reported that mutations in transforming growth factor β (TGFβ) receptors 1 and 2 on chromosome 3 caused a vascular phenotype similar to that seen in Marfan's syndrome.6 TGFβ cytokines have a major role in tissue development and cellular regulation.7 A full description of the advances in understanding of genotype–phenotype correlations is beyond the scope of this paper, but excellent up‐to‐date reviews are available.8,9,10 In summary, it seems that there is a regulatory relationship between extracellular microfibrils and TGFβ signalling. An abnormality in either can lead to the development of the Marfan phenotype.
Multiple organ systems are affected, including the skeleton, eyes, heart, lungs and blood vessels. There are several excellent descriptions of the typical features in both adults and children.11,12,13 Marfan's syndrome is diagnosed using the Ghent nosology (table 11),), which combines clinical and genetic factors.14
Marfan's syndrome may be suspected on antenatal ultrasound,15 but the diagnosis is often not made until late childhood or in adulthood. In the young child, it can be difficult to make a definitive diagnosis. Children often have an evolving phenotype and may need to be followed up for several years before the diagnosis can be confirmed or refuted.16 All possible cases should be regularly assessed by echocardiography, optometry and skeletal survey as the child grows. A complete family history and assessment of other family members also gives clues to the diagnosis. The American Academy of Paediatrics has produced detailed recommendations for the follow‐up of children with Marfan's syndrome, which takes this difficulty into account.17
“Neonatal” Marfan's syndrome is a severe form of Marfan's syndrome often associated with a deletion in the exon 24–32 region of the fibrillin 1 gene (fig 11).). This rare condition differs from the more usual infantile Marfan's syndrome in the severity of the cardiac and pulmonary manifestations.18 Infants with the “neonatal” form often have severe mitral and tricuspid regurgitation in addition to aortic root dilatation. Similarly, the usual arachnodactyly and tall stature may be accompanied by ectopia lentis, very loose skin “as if two sizes too big”, emphysema and joint contractures. The cardiovascular features often require surgical intervention in infancy and this may be complicated by scoliosis and pulmonary hypertension. The long‐term prognosis is very poor, usually because of progressive valve dysfunction or lung abnormalities.18,19
Recently, an important Marfan‐like syndrome has been described: the Loeys Dietz syndrome.20 This is associated with aggressive aortic vascular disease and can be distinguished by the presence of hypertelorism, low‐set ears and a bifid uvula or cleft palate. This condition is associated with abnormalities of TGFBR1 and TGFBR2. In comparison with Marfan's syndrome, there is a much higher risk of dissection at a young age, at smaller vessel dimensions and in non‐aortic vessels. Genotyping can be used to guide treatment, and most patients tolerate cardiac surgery well.21 Other Marfan‐like syndromes do exist,22 and this emphasises the importance of regular follow‐up and assessment (using the Ghent criteria) of all possible cases of Marfan's syndrome.
Virtually all adults with Marfan's syndrome have an abnormal cardiovascular system. In early childhood, however, the features may be mild and easily missed. The most common cardiovascular abnormalities are dilatation of the aorta and mitral regurgitation (table 22).). Most children with Marfan's syndrome have aortic root dilatation. The reported frequency of other valve abnormalities depends to some extent on the rigour of the method of assessment. Moreover, some abnormalities (eg, mitral regurgitation and prolapse) can be intermittent and vary from mild to severe at different times in the same patient. Children with valvular complications are at increased risk of infective endocarditis. Recommendations for antibiotic prophylaxis are changing, but good dental hygiene and early treatment of skin sepsis remain vital. A dilated pulmonary artery is a minor criterion used in diagnosis, but this is of relatively little importance in the paediatric Marfan's population.
Cardiac arrhythmias are an under‐recognised cause of morbidity in both children and adults. A link between Marfan's syndrome and Wolff–Parkinson–White syndrome has been postulated, and atrial fibrillation has been reported in children and adults.39,40 Minor ECG abnormalities may be present in up to 50% of children with Marfan's syndrome.23 In addition, ventricular arrhythmias may occur and can lead to sudden death.24,25,41 This is not surprising given the extensive fibrillin network that extends throughout the myocardium.26 For the same reason, paradoxical septal motion is common. There is also an important subgroup that has marked left ventricular dysfunction unrelated to valve regurgitation.27,42
Echocardiography is the mainstay of assessment of children with Marfan's syndrome. Table 33 shows a protocol for cardiovascular assessment. Detailed echocardiographic assessment should include a full study of left ventricular function, aortic root dimensions and intracardiac valves. Structural lesions should be excluded—in particular, atrial septal defect. Each unit should have a standardised protocol for measurement of the aortic root to allow reproducible sequential measurements, which can be plotted against the body surface area (fig 22).43 Normal values are available for aortic root dimensions in children and adults.43 These nomograms have been criticised as they do not reflect the normal aortic root dimensions in tall, slim children in whom Marfan's syndrome has been excluded. Rozendaal et al44 suggested that an adjusted nomogram derived from tall children without Marfan's syndrome be used to take this into account. The same group devised a discrimination score that showed that the rate of aortic root growth in children and adolescents with Marfan's syndrome differs from the normal population with a sensitivity and specificity of 84% and 73%, respectively.45 Perhaps the most important factor is the need for each echocardiography unit to develop a standardised measurement technique that allows reproducible measurements to be recorded sequentially in comparison to somatic growth. This allows discrimination between normal aortic growth and progressive dilatation and also enables implementation of the appropriate treatment (fig 33).
The pattern of root dilatation should also be noted as diffuse dilatation, with loss of the sinotubular junction is associated with an increased risk of dissection.46 In some children it is not possible to fully assess the aorta owing to a poor acoustic window. This may be exacerbated by marked scoliosis. In such a situation, MRI scanning should be used. This has the benefit of allowing an assessment of the lumbar dura. Dural ectasia is present in 40% of children and in >90% of adults with Marfan's syndrome.47,48
The frequency of cardiovascular assessment will depend on the age of the child, the underlying cardiovascular abnormalities and treatment. In general, most children should be assessed every 6–12 months12,17. This may need to be more frequent when starting treatment or if there is a rapid growth phase.
Most authorities advise children and adolescents with Marfan's syndrome to avoid isometric exercise and competitive or contact sports,4,49 because of the small risk of aortic dissection on exercise.4,50,51 Unfortunately, this advice can occasionally lead to complete avoidance of recreational exercise. Regular exercise has many psychosocial and general health benefits.52 Moreover, although studies on Marfan's syndrome have not been performed, regular exercise is known to attenuate poor vascular compliance in conditions such as diabetes and hypertension.53,54 Consequently, children with Marfan's syndrome should be encouraged to remain active, and a specific aerobic exercise prescription may be beneficial. Similarly, adherence to a healthy “Mediterranean diet” and avoidance of obesity and cigarette smoking should be recommended, as this may prevent exacerbation of the increased vascular stiffness that occurs in the Marfan aorta.55,56,57
Owing to its autosomal dominant inheritance, relatives are also at risk from Marfan's syndrome and should be offered medical assessment. Genetic counselling for would‐be parents explaining the 50% risk to their child and the potential complications during pregnancy, especially increasing aortic root dilatation, should also be discussed.
The diagnosis of Marfan's syndrome itself, with its increased mortality and morbidity also raises psychosocial issues, and the early involvement of clinical psychologists and support groups such as The Marfan Association UK can help in many cases.
Risk stratification in children is difficult. In adults, excessive aortic root dilatation (>1.7 mm/year), increased aortic stiffness, aortic root diameter >55 mm58,59 and dilatation at the aortic sinotubular junction46 are important risk factors for dissection. A family history of aortic dissection is one of the most important risk factors. The absence of lens dislocation has been reported as a risk factor for aortic dissection, although this may simply reflect delay in diagnosis and treatment.58
There are no large, randomised controlled trials of medical treatment in Marfan's syndrome. Despite this, four categories of drugs have been used to delay or prevent aortic root dilatation: β‐blockers, calcium antagonists, angiotensin converting enzyme inhibitors and, most recently, angiotensin‐receptor blockers.
Many small studies have concluded that β‐blockade is successful in slowing aortic root growth and improving survival in some children and adults with Marfan's syndrome.60,61,62,63 Multiple protective mechanisms have been proposed, including a reduction in inotropy, chronotropy and change in aortic pressure over time (dP/dT).4 Other studies have suggested that β‐blockade may lead to a paradoxical increase in vascular stiffness in some patients.64 In the only randomised trial of β‐blockade, 32 adolescents and young adults were treated with high‐dose propranolol for 10 years.60 In comparison with the control group, fewer patients (5 v 9) reached the clinical end points of aortic regurgitation or dissection, cardiac surgery or death. In addition, there was a slower rate of aortic dilatation in the control group. This study suggested that propranolol was ineffective in patients with increased body weight or aortic dilatation >40 mm, implying that early onset of treatment with an adequate dose is probably important. Interestingly, there were two sudden deaths (without dissection) in the control group, suggesting that propranolol may have an additional protective effect against lethal arrhythmias.
In patients who cannot tolerate β‐blockers, calcium antagonists have been suggested as a substitute as they are negative inotropes and also have a vasodilator effect. Rossi‐Foulkes et al61 published a small series of 44 children on medical treatment (β‐blockers or calcium antagonists) and found that the medicated group had a slower aortic growth rate. However, only six children were taking calcium antagonists. No other studies have been published.
Angiotensin‐converting enzyme inhibitors cause systemic vasodilatation and reduce aortic stiffness.65 They can also prevent aortic dissection associated with cystic medial necrosis in an animal model.66 In a recent study, Yetman et al62 compared the effect of enalapril and β‐blockade in 58 patients with Marfan's syndrome. Over a 3‐year period, enalapril was associated with a smaller increase in aortic root diameter and fewer clinical end points than β‐blockade, although the β‐blocker dosage may have been suboptimal.
The most exciting advance in the study of Marfan's syndrome in recent years has been the appreciation that enhanced signalling of TGFβ is an integral link in the pathogenesis of the disorder. Losartan, an angiotensin 11 type 1 receptor blocker used in the treatment of hypertension and heart failure, is a TGFβ antagonist. In a seminal study by Habashi et al,67 the administration of losartan to a mouse model of Marfan's syndrome prevented the development of aortic aneurysm and partially reversed the other somatic features. A randomised trial coordinated by the US National Heart, Lung and Blood Institute is now in progress to establish whether this will be effective in humans.
Few children with Marfan's syndrome require cardiac surgery before reaching the teenage years. In the neonatal form, surgery may be necessary to repair or replace the mitral or tricuspid valves and to replace the aortic root.68 Outside of infancy, tricuspid valve surgery is rarely necessary, and mitral valve repair or replacement is uncommon in childhood. Tsang et al68 reported only seven children referred over a 7‐year period to a large cardiac surgical centre. Of these, three had the infantile form of the syndrome, two required mitral valve replacements (aged 2 and 7 years, one of whom also underwent tricuspid valve repair) and one aortic root replacement (aged 2 years). The four older children all required aortic root surgery (aged 15, 17 and 18 years).
The traditional form of elective aortic root replacement is the composite (Bentall) graft.69 This involves the resection of the aneurysmal portion of the ascending aorta and replacement with a prosthetic valve incorporated in a Dacron tube. More recent alternatives include using a valve‐sparing procedure or novel exostent technique. The valve‐sparing procedure involves resecting the ascending aorta and replacing it with a sculpted Dacron tube, which sits above the native aortic valve.70,71 This has the major advantage of avoiding the need for anticoagulation. Although the aortic valve leaflets are abnormal in patients with Marfan's syndrome, the data on adults suggest that survival is similar to the composite graft, and valve‐related complications are lower.72 Preliminary results in children suggest that the valve‐sparing technique has excellent short‐term results, but this is dependent on the precise valve‐sparing method used.73 The exostent is a new concept, which involves creating a three‐dimensional model of the dilated aorta and producing a computer‐designed stent, which is placed on the outside of the dilated root.74 This avoids the need for bypass surgery, but prevents any form of growth, thus cannot be used in a young child. Long‐term data are awaited. The final option is the use of a human donor aorta (homograft). This has the advantage of avoiding anticoagulation but is complicated by the shortage of homografts and the poor longevity of the graft in young patients.75,76
Perhaps the most important decision is the timing of aortic root surgery. The ideal time to replace the root is “one or two months before it dissects”.77 Although aortic dissection is rare in childhood, the success of elective replacement (>95% survival) is much lower than if emergency surgery is needed. In adults, it is usually recommended that the root be replaced when the sinus of Valsalva measures 5 cm, although this figure is reduced if there are additional risk factors such as a strong family history of dissection.4,78 In older children, most authors recommend root replacement at 5 cm, when enlargement is >1 cm/year or if there is progressive aortic regurgitation.28
TGFβ - transforming growth factor β
Competing interests: AGS is a Medical Advisor to the Marfan Association UK.