‘All these arrangements are marvellous, without any question. But we have to consider that it exceeds all expectations how soon the occlusion of all these openings takes place’.
HISTORICAL CONSIDERATIONS
When Siegal 1 re-examined the original Greek text, he pointed out that Galen was familiar with many aspects of the fetal circulation, even though he did not realise that blood circulated. Galen understood that fetal blood was aerated in the placenta, and that blood was diverted away from the liver by a short vessel connecting the portal to the inferior caval vein. He also knew that blood passed through the oval foramen to bypass the right ventricle, and reach the left side of the heart directly. He realised, nonetheless, that some blood still entered the right ventricle and pulmonary trunk, from whence it was shunted into the aorta through a special fetal channel, thereby bypassing the lungs. The quotation at the head of the chapter, taken from Siegal, 1 provides a clear indication of Galen’s understanding of the dramatic readjustment of the circulation at birth.
It is Botallo, nonetheless, with whose name we usually associate the persistently patent arterial duct. In fact, Botallo described postnatal patency of the oval foramen! It was by a series of misinterpretations, and careless translations, that his name became, quite unjustifiably, attached to the arterial duct. 2 William Harvey, who was a pupil of Fabrizi d’Acquapendente in Padova for 2 years, synthesised previous detailed descriptions of fetal cardiac anatomy from his mentor in his own writings. His genius resided in proposing the concept of active circulation of the blood. He was well aware of the large size of the arterial duct prior to birth, and the fact that blood flowed through it from right to left during fetal life. 3 Harvey 3 was incorrect only in his belief that flow of blood to the lungs was completely lacking in the fetus. Subsequently, Highmore, a friend of Harvey, described closure of both the oval foramen and the arterial duct as occurring with the onset of respiration, believing the arterial duct to collapse as a consequence of blood being diverted to the lungs. 4 Since then, several ingenious theories have been put forward to explain closure of the duct, all being based on postmortem appearances, and all invoking mechanical factors. 5 It was Virchow 6 who first suggested that the closure results from contraction of its mural smooth muscle, while Gerard 7 introduced the concept of two-stage closure, in which functional constriction is followed by anatomical obliteration. Thanks to Huggett, 8 we came to understand the role of oxygen in effecting functional closure by muscular contraction.
As early as 1907, in an address to the Philadelphia Academy of Surgery, John Munro had suggested surgical ligation of the persistently patent arterial duct. 9 After an interval of 31 years, the first attempt at surgical closure was made in a 22-year-old woman with bacterial endocarditis. 10 Unfortunately, although the patient survived the surgery, she died a few days later from complications of the infection. As a result, it was Robert Gross, of Boston, who first successfully ligated a patent duct, in a 7-year-old child with intractable heart failure. 11 He thereby introduced an amazing era of progress in the surgery of congenital malformations of the heart. No historical review of the arterial duct, however brief, would be complete without recognising the importance of Gibson’s exquisite description of the murmur that typifies its persistence, 12 notwithstanding the fashion in which Gibson’s initial account has subsequently been misquoted. 13
NOMENCLATURE
From time to time, authors have argued that the terms patent and persistent are redundancies that should be avoided when describing the arterial duct. In my view, this oversimplifies the situation, and both terms retain their value. Persistence implies that the duct is present after the time of its expected closure and, therefore, distinguishes a pathological from a physiological state. The concept of patency remains useful in the perinatal period, especially in the premature infant in whom the term can be used to signify a duct that is functionally open, as opposed to one that is functionally closed, but retains the potential to reopen.
THE NORMAL FETAL CIRCULATION
About two-thirds of the fetal cardiac output originates from the right ventricle, with only 5% to 10% passing through the lungs. 14,15 As such, the majority of right ventricular output passes through the arterial duct into the descending aorta, and its presence is essential for normal fetal development, permitting right ventricular output to be diverted away from the high-resistance pulmonary circulation. Premature constriction or closure may lead to right heart failure, resulting in fetal hydrops. 16
EMBRYOLOGY AND PATHOGENESIS
During early fetal development, five arterial arches link the aortic sac with the paired dorsal aortas, although all arches are never present simultaneously ( Fig. 42-1 ).
The arches are numbered 1 through 6, albeit that there is no evidence to support the existence of the purported fifth arch, even though some congenital malformations are interpreted on the basis of its persistence. The initial symmetrical arrangement does, however, provide the basis for the pattern of the hypothetical double arch proposed by Edwards 17 to explain vascular rings and slings. This bilaterally symmetrical arrangement, with paired brachiocephalic arteries and arterial ducts ( Fig. 42-2 ), is usually transformed to the configuration seen in postnatal life by the disappearance of some arterial segments, and realignment of others ( Fig. 42-3 ). 18
The normal duct develops from the dorsal portion of the left sixth arch. From their inception, the sixth arches are associated with the developing lungs. The arteries feeding the developing lungs develop within the anterior wall of the mediastinum, and take their origin of the floor of the aortic sac, which initially feeds also the bilateral arteries of the sixth arches ( Fig. 42-4 ).
When the developing arterial segment is divided to form the intrapericardial components of the aorta and pulmonary trunk, the sixth arches, originating from the caudal part of the aortic sac, are placed in continuity with the pulmonary channel. A significant event in the appropriate connection of the pulmonary arteries with the intrapericardial pulmonary trunk is the obliteration and disappearance of the right sixth arch. On the left side, the arch persists as the arterial duct, with the pulmonary arteries left in continuity with the channel from the pulmonary trunk to the left sixth arch ( Fig. 42-5 ).
The hypothetical double arch system devised by Edwards 17 is able to explain all developmental anomalies of the aortic arch, including those associated with an abnormally situated duct, be the duct patent or represented by the arterial ligament. The various possibilities are discussed in Chapter 47 , including those in which there is persistence of both sixth arches, producing bilateral arterial ducts. Such circumstances are rare, and are always associated with intracardiac anomalies. Typically the ducts supply discontinuous pulmonary arteries in the setting of tetralogy of Fallot with pulmonary atresia, albeit that bilateral ducts can also be found in association with isolation of a subclavian artery. 19 The arterial duct does not always persist on one side. Its absence was first described as a postmortem finding in 1671, being seen in a grossly malformed infant with an extrathoracic heart and tetralogy of Fallot described by Nicolas Steno. 20 Absence of the duct is a typical finding in the syndrome of tetralogy of Fallot with so-called absent pulmonary valve and dilated pulmonary arteries. 21 It was thought that absence of the duct, and hence absence of any overflow, explained the dilated pulmonary arteries, but the pulmonary arteries can be dilated even when the duct is present and patent (see Chapter 36 ). The duct is also absent in approximately three-quarters of patients with a common arterial trunk. Absence of significant flow through the duct in the presence of a larger aortopulmonary connection permits the duct to disappear early in fetal life. In more complex varieties of common arterial trunk, however, such as those with so-called absence of one pulmonary artery, or those patients with an associated interruption or atresia of the aortic arch, patency of the duct is essential to maintain both the systemic and pulmonary circulations (see Chapter 41 ).
A number of teratogens are known to influence the development of the duct, including rubella, alcohol, amphetamines, and the anticonvulsant hydantoin, with the duct being most sensitive from 18 to 60 days of gestation. 22 Absence of the duct has been induced experimentally in chick embryos by the administration of agonists of β-adrenoceptors, leading to the suggestion that the teratogenesis is mediated by cyclic AMP. 23
ANATOMY
The Arterial Duct and Its Normal Closure
In the fetus, and in the neonate prior to its closure, the arterial duct is a short and wide vessel of variable length. It connects the pulmonary arteries to the lesser curve of the arch of the aorta, terminating at the point of transition from the isthmus to the descending aorta, distal to the origin of the left subclavian artery ( Fig. 42-6 ).
Patency is maintained by the relatively low fetal oxygen tension 24 and cyclooxygenase-mediated products of arachidonic acid metabolism, primarily prostaglandin and prostacyclin, 25 whose levels are high because of production by the placenta and decreased metabolism in the fetal lungs. Produced both locally, in ductal tissue, and circulating, these mediators cause vasodilation through interaction with prostanoid receptors.
Prior to birth, the duct is very much the direct continuation of the pulmonary trunk, with the left and right pulmonary arteries taking origin as smaller branches from the trunk. Posteriorly, the duct is related to the left main bronchus, while anteriorly it is crossed by the vagus nerve. This gives off the left recurrent laryngeal nerve, which encircles the duct before ascending behind the aortic arch into the neck ( Fig. 42-7 ). In terms of its microscopic structure, the duct is a muscular artery endowed with an intima, media and adventitia, differing markedly from the adjacent pulmonary trunk and aorta ( Fig. 42-8 ). While the media of the aorta is composed mainly of circumferentially arranged elastic fibers, the media of the duct consists largely of spirally arranged smooth muscle cells, some with circular and others with longitudinal orientation, 26 with an increased content of hyaluronic acid. The intimal layers are thicker than those of the adjoining vessels, and contain increased amounts of mucoid substance. 27 In the newborn, the tissues are rather loosely arranged, with a well-defined internal elastic lamina that may be single or focally duplicated, with small interruptions encountered regularly. No collagen is seen in the media by light microscopy, but abundant material that stains positively for acid mucopolysaccharides is observed between the muscle and elastic laminas. Electron microscopy reveals fine collagen fibrils lying between adjacent lamellas of smooth muscle cells and elastin. During the second half of gestation, the smooth muscle cells show decreasing evidence of secretory activity, and increasing maturation of their contractile elements.
It is known that vessels cannot close by isolated contraction of circularly arranged muscle, 28 so coincident shortening of the less abundant longitudinally arranged muscle fibres is critical to effective closure. The duct is innervated mostly by adrenergic fibres, supplying largely the adventitia and outer media, with cholinergic fibres being extremely sparse or totally absent. 29 Vessels are also found in its walls that may have a role in fuelling contraction at birth. 26 Some degree of hyperaemia of these vessels is common in newborn infants.
There is controversy with regard to the structure of the intimal layers during fetal life. Eccentrically placed intimal cushions, or mounds composed of smooth muscle and elastic tissue, have been described by many authors, with suggestions made that the formation of these mounds precedes normal ductal closure subsequent to birth. It is questionable, however, whether the intimal cushions are produced during normal fetal maturation. When account is taken of the high flow of blood through the fetal duct, it is difficult to conceive that prominent protrusions of intima into the lumen could exist without inducing turbulent flow and a bruit. No such bruit is heard in the undisturbed ducts of fetal lambs at term. Furthermore, in many of the studies suggesting the presence of intimal cushions, the tissues have usually been subjected to one or more perturbations, such as relatively slow fixation, mechanical stimulation, or cessation of circulation with loss of intraluminal distending pressure. On this basis, the existence of intimal cushions as prenatal structures has been challenged. In a series of experiments, 30–34 Hornblad and colleagues showed that, independent of the degree of closure, the lumen of the duct remained round, was without deformation, and showed no evidence of formation of mounds. Mural thickness increased at the time of closure, while the internal elastic lamina became corrugated, especially in the mid-portion of the vessels. A decrease in the lumen was associated with accumulation of endothelial cells within the lumen. They concluded that closure was aided by passive central displacement of endothelial and inner medial cells, but that no part of the medial layer was prepared prenatally for this process. 30–34 These findings endorsed earlier studies in the human, which suggested that the cushions appeared as a normal reparative reaction to distending forces during fetal life. 35
At birth, the vessel unequivocally constricts with the abrupt increase in oxygen tension inhibiting ductal smooth muscle voltage-dependent potassium channels, resulting in an influx of calcium and ductal constriction. 36 Levels of prostaglandin and prostacyclin fall because of metabolism in the functioning lungs, and elimination of the placenta. The intimal thickenings, or cushions, become irregular ridges protruding into the lumen, running mainly lengthwise. By their extrusion, they exert traction on the media, causing disorganisation and formation of mucoid lakes (see Fig. 42-8 ). Anatomical obliteration follows functional closure. The process begins with necrosis of the inner wall, followed by the formation of dense fibrous tissue. The lumen is progressively obliterated by a process of fibrosis, probably representing organisation of mural or occlusive thrombus. Eventually, the duct becomes converted into a fibrous strand, the arterial ligament, which may become calcified. Anatomical obliteration may take several weeks to complete. About two-thirds of ducts are normally obliterated by the age of 2 weeks, 37 and almost all by 1 year.
The Persistently Patent Arterial Duct
As stated above, in normal circumstances all ducts should be converted to an arterial ligament within the first year of life, with two-thirds closing in the initial 2 weeks. Some ducts, however, never close. These are the channels best described as showing persistent patency. Gittenberger-de Groot 38 found the internal elastic lamina to be intact in some, but not all, of persistently patent ducts studied histologically, along with a sparsity of intimal cushions ( Fig. 42-9 ). Bakker 35 had noted similar findings, describing them in terms of aortification. The duct, if persisting as a patent structure, joins the pulmonary arteries to the descending aorta in the fashion seen in the neonate ( Fig. 42-10 ). The channel itself can vary markedly in its width (compare Figs. 42-11 and 42-12 ). The duct can also vary considerably in its shape. Study of a large number of angiograms from patients with persistently patent ducts undergoing interventional closure at the Hospital for Sick Children, Toronto 39 showed that the most frequent pattern was to find a constriction at the pulmonary end of the duct. This pattern was seen in two-thirds of cases. In just under one-fifth, a constriction was found at the aortic end of the duct, and in just under one-tenth, the lumen was unrestricted. In just under one-twentieth, there was a constriction at both ends, whilst the remaining patients showed bizarre patterns not lending themselves to classification ( Fig. 42-13 ).
In the past, persistently patent arterial ducts were often the nidus for infectious endocarditis, but this complication is now extremely rare in developed countries. Ducts could also become aneurysmal and elongated ( Figs. 42-14 and 42-15 ), but this is also now a rare finding, as is ductal rupture.
EPIDEMIOLOGY
The factors responsible for persistent ductal patency beyond the first days of life are not fully understood. An increased incidence is seen in the premature neonate due to physiological factors more related to prematurity rather than an inherent abnormality of the duct. 40 In infants born at term, persistent patency occurs sporadically, but there is increasing evidence that genetic factors, or prenatal infection, play a role in many children. The estimated incidence lies between 1 in 2000 and 1 in 5000 live births. 41–43 Persistent patency of the duct accounts for about one-eighth of all congenital cardiac malformations. The most extensive study from a relatively homogeneous population was that performed by Carlgren, 43 when he charted the incidence of congenital cardiac disease to children born in the Swedish City of Gothenburg. Persistent ductal patency was the third most common lesion identified, representing about 0.04% of live births. If children with the so-called silent duct are included, found incidentally by echocardiography performed for another reason, the incidence may be as high as 1 in 500. 44 A significantly higher incidence of ductal patency is also seen in infants born with low weight, patency being found in almost half of infants weighing less than 1750 g at birth, and up to four-fifths of those weighing less than 1200 g (see also Chapter 11 ).
Genetic Factors
Unlike premature infants as noted above, in whom persistent ductal patency is more often due to developmental immaturity, in the infant born at term there is likely a structural abnormality. There is also an increased frequency in several genetic syndromes, such as those with defined chromosomal aberrations, examples being trisomy 21, 14q- and 4p- syndrome, single-gene mutations such as Carpenter’s syndrome and Holt-Oram syndrome, and X-linked mutations such as incontinentia pigmenti. Although most cases of persistent patency are sporadic, many are also believed to be due to a multi-factorial inheritance pattern, with a genetic predisposition and an environmental trigger that occurs at a vulnerable time. 45 Females predominate in a ratio of greater then 2:1. The mode of inheritance appears to be autosomal recessive with incomplete penetrance. 46 In a family having one sibling with a patent duct, there is about a 3% chance of a persistent duct in a subsequent offspring, 45 and a higher risk to the offspring given one affected parent, some 45 times greater than the that for the general population. The risk to further children in sibships where two children have been affected is probably in the order of 10% and increases with each affected child.
Infection and Environmental Factors
Maternal rubella during the first trimester of pregnancy, particularly in the first 4 weeks, is associated with a high incidence of a persistently patent arterial duct. 49–50 The histology resembles that of a very immature duct, with an extensive subendothelial elastic lamina, thought to retard anatomic sealing. 51 Additional environmental factors have been reported associated with persistence, such as in fetal valproate syndrome, 52 or after thalidomide. 53 A patent arterial duct may be more likely to be found in infants born at high altitude. 54
PATHOPHYSIOLOGY
Persistent patency results in shunting of blood from one side of the circulation to the other, the volume of flow depending on the length and internal diameter of the duct, and on the systemic and pulmonary vascular resistances. As pulmonary resistance is usually much lower than systemic, the flow is from the aorta to the pulmonary trunk. Hence, flow to the lungs is increased, and results in left atrial and left ventricular overload. If the duct is widely patent, flow depends entirely on the ratio of resistances. Right ventricular failure may occur in the presence of a large duct with pulmonary hypertension or pulmonary oedema, and an elevated left atrial pressure. In most patients, the duct is partially constricted, and the major factor limiting flow. Under these conditions, pulmonary arterial pressure is normal or only mildly elevated. Symptoms and clinical findings are largely determined by the magnitude of the shunt.
Clinical Features
Most patients are asymptomatic, the lesion recognised with detection of the characteristic murmur. Occasionally, there may be a history of prematurity or asphyxia during birth. Children with large shunts may fail to thrive, experience difficulty with feeding during infancy, and frequently suffer recurrent infections of the upper respiratory tract. Occasionally, congestive heart failure develops.
Physical Examination
There is retarded growth in about one-third of children. They are acyanotic in the absence of complicating factors. The peripheral pulses are easily palpable, with a rapid upstroke and decay. There is a widened pulse pressure, with lowering of the diastolic component. Arterial pulsation in the neck may be prominent in those with large shunts. Precordial examination reveals an active cardiac impulse, with the forceful cardiac apex displaced to the left. When the shunt is small, the only abnormal finding may be the murmur. The continuous, or machinery, murmur of the uncomplicated persistent duct is best heard in the left infraclavicular area, although it is occasionally maximal at the third left interspace. Gibson’s 12 description of the murmur is quoted by Tynan 13 : ‘It begins quite obviously after the commencement of the first sound. It is continued during the latter part of that sound and the whole of the short pause. It persists throughout the second sound and dies away gradually during the long pause. The murmur is distinctly rough and thrilling in its character. It begins, however, somewhat softly, and increases in intensity to reach its acme just about, or immediately after, the incidence of the second sound, and from that point wanes until its termination. The second sound can be heard to be loud and clanging and when carefully analysed it is the pulmonary part of that sound which is accentuated’. Turbulent flow through the duct itself causes the murmur. Additional murmurs may be present due to increased flow across the aortic valve, producing an ejection systolic sound, and across the mitral valve, giving a diastolic murmur with loud onset. The systolic component of the continuous murmur may be transmitted into the neck, may be associated with a thrill in the second left intercostal space, or may increase in intensity during inspiration due to a fall in pulmonary vascular resistance. Many patients with loud continuous murmurs also have multiple clanging sounds. These are relatively localised to the pulmonary area, and are most frequent in the second half of systole, corresponding to the period of peak flow within the duct. Neill and Mounsey 53 attributed these sounds to the turbulence caused by the head-on collision of opposed flow from the duct and the right ventricle, and named them eddy sounds. These auscultatory findings are only for an uncomplicated persistent duct in a child. It should be remembered that these features may differ in infancy, or be altered by the development of complications.
Investigations
Electrocardiogram
Patients with an isolated persistent arterial duct usually have some electrocardiographic evidence of left atrial and ventricular hypertrophy, reflecting volume overload of the left heart. Occasionally, the electrocardiogram may show combined ventricular hypertrophy, or, if the duct is small, be entirely normal. The electrical axis is usually normal, and deviation to the right, with right atrial and/or right ventricular hypertrophy, suggests the presence of additional defects or pulmonary hypertension. The electrocardiographic changes are less predictable in infants and clinically less helpful. 54 Prolongation of the PR interval, which disappears or decreases after closure, has been observed in about one-fifth of cases. 55 Atrial fibrillation may develop in adult life. 56,57 When the shunt is large enough to equalise the systemic and pulmonary arterial pressures, biventricular hypertrophy is likely to develop. With the onset of pulmonary vascular disease, the predominant findings will be those of right ventricular hypertrophy.
Chest Radiography
The chest film may be normal in patients with a small shunt. Cardiomegaly is present in those in whom flow to the lungs is close to twice systemic flow or greater. Increased pulmonary vascular markings are seen, with an obvious bulge of the pulmonary trunk at the left border of the cardiac silhouette. The aorta is also prominent. Both it and the pulmonary trunk tend to enlarge with age. Enlargement of the left atrium is usually present, and reflects increased pulmonary venous return due to the left-to-right shunt. Increased pulmonary vascularity may be more marked on the right, as is often seen with other left-to-right shunts ( Fig. 42-16 ). The duct may calcify, although this complication is more common when the vessel is closed rather than patent. The aortic end of the duct, the ductal ampulla, may be seen on the chest radiograph, and can be demonstrated angiographically during the first week of life. These findings may be modified, especially if pulmonary vascular disease develops.
Echocardiography
Persistent patency beyond the neonatal period is readily diagnosed from the characteristic clinical features. Cross sectional echocardiography will help rule out other structural cardiac malformations. The duct can be imaged throughout its length using a high left parasternal view, 58,59 allowing evaluation of ductal size, and the presence of tissue within the lumen, indicating imminent closure. In preterm infants, imaging may be difficult due to emphysematous lungs from high ventilatory pressures. A subxiphoid view can then be used. Characteristic diastolic flow in the pulmonary trunk identified by Doppler interrogation increases the confidence of diagnosing ductal shunting.
Flow through the duct can be quantified by analysis of Doppler tracings of diastolic flow in either the left pulmonary artery or the descending aorta. Colour flow Doppler techniques have been more useful in revealing ductal patency. This is, at present, the most sensitive method for detecting and semi-quantifying ductal flow 60 ( Fig. 42-17 ). Qualitatively, the presence of bidirectional, or pure right-to-left, shunting is specific for elevated pulmonary arterial pressures. 61,62 In children with high pulmonary vascular resistance, with a low-velocity Doppler signal or right-to-left flow, the duct may be very difficult to demonstrate by colour flow imaging, even if it is large. Associated findings such as septal flattening, unexplained right ventricular hypertrophy, or high-velocity pulmonary regurgitation should prompt an investigation for a patent duct. Contrast echocardiography may also be helpful in this setting, identifying microbubbles in the descending aorta in consequence of ductal right-to-left shunting, but not in the ascending aorta. Additionally, using colour flow mapping, Doppler measurements of velocity can be used to estimate pulmonary arterial pressure. 63–65 M-mode studies provide an assessment of left atrial and ventricular size, which gives some idea of the magnitude of the shunt. In children with a small duct, the chambers are usually of normal size, although mild left atrial and/or left ventricular enlargement may be seen. In children with a moderate or large duct, the left atrium and ventricle are enlarged. Echocardiography is probably most valuable in the diagnosis of ductal patency in the premature infant. It will be discussed further in that section.
Recent echocardiographic studies using colour flow Doppler have further identified the presence of small ductal communications in the absence of any typical murmur of patency, this degree of shunting giving uncharacteristic soft vibratory systolic murmur or no murmur at all. 44,66,67 These findings have a significant impact on estimates of the incidence of ductal patency and the risk of endocarditis.
Cardiac Catheterisation and Angiography
Most cardiologists would not consider catheterisation a necessary diagnostic procedure for children with typical clinical findings. If it is undertaken, it is usually possible to probe the duct from the pulmonary trunk, and to pass a catheter through the vessel and down the aorta. When the catheter apparently crosses a duct, but then turns in a headward direction, the alert investigator should consider the presence of an aortopulmonary window. The size of the shunt may be difficult to quantify by oximetry because it is difficult to obtain a truly representative sample distal to the site of shunting. Pulmonary arterial pressure is usually normal or slightly elevated. The duct can be visualised by selective aortography with injection of contrast media in the last part of the aortic arch ( Fig. 42-18 ).
Therapeutic catheterisation, however, is currently the treatment of choice for most children and adults with a patent duct. In this regard, complete haemodynamic assessment is important prior to attempting closure, particularly in the adult. In patients with an elevated pulmonary arterial pressure, assessment of pulmonary vascular resistance and its response to vasodilating agents may be helpful in determining suitability of closure. Assessment of haemodynamics during temporary test occlusion with a balloon catheter may also be a helpful manoeuvre in assessing advisability of closure in marginal cases.
Angiography defines the anatomy. Such detailed assessment is essential before attempting closure so that the proper device and size can be chosen. Important features include the minimal diameter, usually at the pulmonary arterial end; the largest diameter, usually at the aortic ampulla; the length of the duct; and its relationship to the anterior border of the tracheal shadow, the latter helping to guide positioning of the device. 39 Other imaging modalities are available to confirm the presence of ductal patency. Radionuclide scanning can be used to detect the presence of shunting, but anatomic localisation is lacking. Magnetic resonance imaging provides anatomic detail, 67 and is particularly useful in the setting of unusual ductal geometry, and in those with associated abnormalities of the aortic arch ( Fig. 42-19 ). 68,69 Examples include the patient with a ductal aneurysm presenting as a mass in the chest, the duct associated with a vascular ring, a right aortic arch, or cervical arch. With velocity encoding of cinemagnetic resonance signals, patterns of shunting can be detected. 70,71 In the adult, computed tomography can assess the degree of calcification, an important feature if surgical closure is considered. In general, however, the simpler technique of cross sectionalechocardiography with colour flow Doppler provides sufficient anatomic and haemodynamic detail to define the anatomy and its variations, and points the way to proper management.
Diagnostic Problems
Other causes of a continuous murmur may create confusion. The venous hum often causes difficulty to the inexperienced auscultator. This noise, which can be loud, is usually best heard in the supraclavicular fossa, and while audible bilaterally, it is usually louder on the right. A venous hum, when loud, may be transmitted below the clavicle, and may be misdiagnosed as being from a patent duct. This error can be avoided by exerting pressure over the root of the neck, turning the head to the side, or laying the child down. These manoeuvres readily obliterate the venous hum, while having no effect on the murmur generated by ductal flow. An aortopulmonary window rarely causes a continuous murmur similar to that of persistent ductal patency. More typically, however, the communication is large, and does not cause a continuous murmur. Even with careful aortography, this condition may be misdiagnosed. Echocardiography should now help to avoid this error. Major aortopulmonary collateral arteries, pulmonary arteriovenous fistulas, and collateral arteries associated with coarctation, all cause continuous murmurs. These seldom cause diagnostic problems because of the general clinical picture and the location of the murmur. Other causes of continuous murmurs heard in the chest include ruptured sinus of Valsalva, peripheral pulmonary arterial stenosis, common arterial trunk, coronary artery fistulas, the supracardiac form of totally anomalous pulmonary venous connection, mitral atresia, surgically created systemic-to-pulmonary arterial shunts, and rarely anomalous origin of the left coronary artery from the pulmonary trunk. Prolapse of an aortic valvar leaflet into a ventricular septal defect may also simulate persistent patency of the duct. Most, if not all, of these potential pitfalls can be avoided by careful clinical evaluation and good echocardiography.
NATURAL HISTORY
Like most congenital cardiac malformations, reliable information about the natural history of untreated patients with a persistently patent duct is non-existent. Available data stems from the short period of time which elapses between the condition being diagnosed with any frequency and to its being relieved by an operation. Campbell 72 attempted an overview of the natural history, based on his own extensive clinical experience and on the literature. Inevitably, such calculations tend to over-emphasise the number of patients who experience events, be they favourable or adverse, and underestimate the number of patients with an asymptomatic and undetected duct.
Spontaneous Closure
By definition, a persistent duct is one that remains open beyond 3 months in an infant born at full-term. Delayed closure in premature infants, or that occurring within the first 3 months, is therefore excluded from consideration in this section. Campbell 72 analysed four series of patients in which 11 examples of spontaneous closure occurred over 1842 patient-years, giving a rate of 0.6% per annum. Several of the examples, however, were based on quite tenuous clinical impressions. In none was catheterisation performed before and after the event. The figure calculated by Campbell is almost certainly an overestimate. He did not suggest that surgery should be delayed except, perhaps, in patients with small shunts and signs that the duct was already closing. Few cardiologists would now agree even with these exceptions.
Effect on Life Expectancy
By combining four series, consisting mainly of unselected schoolchildren with a persistent duct, Campbell deduced a mortality rate of 0.42% per annum during the first two decades. Thereafter, he calculated mortality rates per year as 1% to 1.5% in the third decade, 2% to 2.5% in the fourth, and 4% for each subsequent year. These calculations indicate that one-third of patients with a persistent duct die by the age of 40, in contrast to less than one-twentieth of the normal population. Many of the figures are based on data obtained in the era before antibiotics were available. As infective endocarditis is a major cause of death, the impact of antibiotics must also be taken into account. These figures agree fairly well with age at death as reported in necropsy series. Abbott, 73 for example, found the mean age at death, having excluded those who died in infancy, to be 30 years, and in another series, the mean age was 36.5 years. 74 Despite this agreement, the fact remains that calculations from autopsy series, and from clinical series, are extrapolations from rather small numbers. They undoubtedly exaggerate the adverse aspects of the natural history.
COMPLICATIONS
The important complications of persistent patency of the duct include congestive heart failure, infective endarteritis, pulmonary vascular disease, aneurysmal formation, thromboembolism, and calcification.
Congestive Heart Failure
Congestive heart failure resulting from an isolated persistent duct develops either in infancy or during adult life. Infective endarteritis may rarely precipitate heart failure during childhood. Heart failure in infancy usually has its onset before the age of 3 months. A delayed normal fall in pulmonary vascular resistance may cause the left-to-right flow to increase progressively. The clinical picture is initially that of left heart failure, with tachypnoea and pulmonary oedema. Ultimately, signs of right heart failure appear with hepatomegaly. Although initially there may be a good response to diuretics, this is seldom maintained and closure is advisable. Infants born at term do not respond to indomethacin when over 3 months of age. The occasional occurrence of sudden death in infants treated medically further encourages a policy of early intervention. Amongst adults, there used to be a group with cardiomegaly and features of left ventricular overload and strain. Such patients now are rare in countries with well-developed systems of health care, as it is unlikely their lesion would have escaped detection. Congestive heart failure may also occur as a terminal event in patients in whom severe pulmonary vascular disease complicates a persistently patent duct. If so, transcatheter closure appears to be the treatment of choice.
Infective Endarteritis
Infective endarteritis in a patient with an uncomplicated persistent duct is uncommon in childhood, and appears to be prevented by surgery or catheter-based embolisation. In the era preceding antibiotics, and interventional or surgical treatment, it was a major cause of death, accounting for almost half of all deaths in several pooled autopsy series. 73–75 Campbell 72 calculated an infection rate of between 0.45% and 1.0% per annum for patients after the first decade. The first line of treatment should be with antibiotics, following the recommendations as established by the American Heart Association, with surgery or embolisation delayed until sterilisation is completed. Occasionally, this proves impossible, in which case surgery or occlusion should be performed under continuing antibiotic therapy. 76 Vegetations are usually found at the pulmonary arterial end of the duct, and may give rise to recurrent pulmonary embolisation, with the clinical picture suggesting recurrent pneumonia. In developing nations with limited access to health care, infective endarteritis associated with the persistently patent duct continues to be a significant health issue. 77 Infection may cause some examples of ductal aneurysms, especially those occurring postoperatively. While two cases of endarteritis on a clinically silent and non-hypertensive duct have been reported, the anticipated high incidence of the silent duct has little implications for treatment. 44,78,79
Pulmonary Hypertension and the Persistent Duct
Although the pulmonary arterial pressure is usually normal, or only slightly elevated, in patients with persistently patent ducts, occasionally it is raised sufficiently to modify the physical findings. The implications of pulmonary hypertension secondary to an increased flow, as opposed to that caused by increased resistance are markedly different and the two situations should be clearly differentiated.
When the duct is widely patent, and pulmonary vascular resistance is low, systolic pulmonary arterial pressure equals systemic systolic pressure, and blood flow to the lungs is several times greater than that in the systemic circuit. The pulmonary arterial diastolic pressure may equal or be slightly lower than that in the aorta. These patients usually experience severe congestive heart failure, with failure to thrive and recurrent respiratory infections. Their electrocardiogram shows combined ventricular hypertrophy, while the chest radiograph reveals cardiomegaly with marked pulmonary plethora. The echocardiogram will reveal enlargement of the left heart chambers. Such patients respond poorly to medical therapy, and the correct management is to eliminate the shunt. Successful ligation or catheter occlusion usually restores pulmonary arterial pressure to normal. There appears to be little risk of subsequent pulmonary vascular changes in this group of patients.
Some individuals respond to pulmonary venous distension, and to the left atrial enlargement secondary to high pulmonary blood flow with reflex pulmonary vasoconstriction partially protecting themselves against the full effects of unrestricted ductal flow. If studied haemodynamically, these patients will be found to have a moderate left-to-right shunt, with pressure in the pulmonary circulation at systemic levels, with or without a high pulmonary capillary wedge pressure. Pulmonary arterial pressure usually falls with administration of oxygen, or in response to pulmonary vasodilators. Successful elimination of the shunt usually restores pulmonary arterial pressure to normal. Fixed and high pulmonary vascular resistance may result from progressive structural changes in patients who originally had large left-to-right shunts and normal pulmonary vessels. Alternatively, it may exist from birth. 80,81 Civen and Edwards 82 suggested that patients in this category represent a form of persistence of the fetal pulmonary circulation. As yet, there is poor understanding of the factors which initiate and maintain the progressive pulmonary vascular damage.
It is instructive to follow the clinical changes that accompany the rise in pulmonary vascular resistance. Initially, the pulmonary diastolic pressure approaches systemic levels, decreasing diastolic flow across the duct. As flow diminishes, the diastolic component of the continuous murmur becomes attenuated and eventually disappears. At this stage, the patient has a pansystolic murmur. With a further increase in resistance, systolic pressures begin to equalise, systolic flow diminishes, the systolic murmur shortens, and eventually it also disappears. Concurrent with these changes, the second sound becomes closely split or even single, and there is accentuation of the pulmonary component. The clinical findings become those of severe pulmonary hypertension, with marked right ventricular hypertrophy and a loud, often palpable second pulmonary sound; a pulmonary ejection click is almost always audible. The second sound is loud and difficult to split. A high-pitched early diastolic murmur, the Graham Steel murmur of pulmonary regurgitation, may be added to these sounds, as may a pansystolic murmur of tricuspid regurgitation when right heart failure supervenes. Equalisation of pressures with balanced resistances also brings reversal of the direction of flow through the duct, the magnitude of the flow increases concomitant with the rise in pulmonary vascular resistance. In some patients, it is possible to recognise differential cyanosis, the blue discolouration being confined to the lower body, and with clubbing of the toes but not the fingers. Unless there is differential cyanosis, it is not possible to recognise a duct clinically in patients with severe pulmonary hypertension and high pulmonary vascular resistance. The diagnosis will depend on cardiac catheterisation and angiography, or cross sectional echocardiography and colour flow Doppler studies.
Until heart failure develops, the chest radiograph shows at most mild cardiomegaly, with marked prominence of the pulmonary arterial segment (see Fig. 42-16 B). Right-axis deviation, right atrial hypertrophy, and right ventricular enlargement are usually evident in the electrocardiogram. In some cases, however, a picture of combined, or even left, ventricular hypertrophy may still be seen. It is impossible to calculate accurately the risk of progressive pulmonary vascular disease in patients with a large persistent duct. Surgical treatment has been available almost as long as clinical recognition. The information in terms of natural history necessary to answer the question is not available, nor would such a study now be feasible. Campbell 72 did not address this problem in his calculations, although there are several reports in the literature concerning this complication. 83–85 These are based, however, on selected groups of patients and overemphasise the frequency of the problem. Nor do they all distinguish adequately between pulmonary hypertension with high flow and true pulmonary vascular disease. The presence of pulmonary hypertension secondary to structural changes within the pulmonary vasculature increases the risk of closure, especially once there is right-to-left shunting, with a reported mortality in more than half of a small group of such patients. 83 The complication should be largely avoided by early recognition and treatment of the hypertensive duct. In patients with pulmonary vascular resistance greater then 8 Woods units per meter square, lung biopsy has been recommended to determine candidacy for closure, but unfortunately it may not be fully predictive of outcome. 86 Such patients may be haemodynamically worse after closure with the development of suprasystemic pulmonary arterial pressure, low cardiac output, and right ventricular failure. Patients have been described, nonetheless, with severe histological changes consistent with irreversible pulmonary vascular disease, which resolves completely after closure of the duct. 87 In the occasional patient who escapes early detection, cardiac catheterisation with a vasodilator challenge may be useful in determining the extent of pulmonary vascular changes, and their potential for reversal.
Aneurysm of the Duct
True aneurysm of the duct is rare. It manifests in two distinct forms. The first presents at or shortly after birth, 88,89 the so-called spontaneous aneurysm of infancy 90 ( Fig. 42-20 ). The second form presents in childhood or later life. 91,92 Recent studies suggest that the incidence may be as high as 8%. 93 The true incidence remains unknown, as the definition is not precise, and many aneurysms detected by fetal or neonatal echocardiography resolve spontaneously, without clinically apparent sequels. About one-quarter of patients will have an underlying disorder, such as trisomy 21 or 13, Smith-Lemli-Opitz syndrome, type IV Ehlers-Danlos syndrome, or Marfan’s syndrome. 90 Ductal closure usually begins at the pulmonary arterial end of the vessel, and if closure at the aortic end fails to occur, it becomes in effect an aortic diverticulum under systemic pressure. While formation of such a diverticulum is common, it is less clear why this occasionally progresses to aneurysmal formation. Structural abnormalities are possibly present in the aortic, but not the pulmonary end of the duct, such as those associated with collagen vascular disorders. Sepsis may be involved in the pathogenesis of some cases in infancy. A diverticulum arising from the pulmonary trunk is also common. Usually, the type found in infancy is asymptomatic, and may not be uncovered until autopsy for death from other causes. It presents as a tumour-like left-sided mediastinal mass. In one-fifth of cases, rupture or embolism leads to death. Dissection and infection may also occur. Regression can occur, presumably due to thrombosis and organisation, but progressive enlargement, or the onset of hoarseness because of damage to the recurrent laryngeal nerve or left bronchial obstruction, is an indication for surgical excision. In view of the frequency of life-threatening complications, prompt surgical removal is advisable. Percutaneous occlusion of the aneurysm has not been established, but a potential approach is placement of a covered stent in the aorta to exclude the aneurysm and occlude the duct 94 Aneurysm of the duct is even more uncommon in adults. 95 The duct may be patent at both ends, but is usually closed at the pulmonary arterial end. 69–98 Possible pathogenic mechanisms include arrested closure, with persistence of an aortic diverticulum, delayed spontaneous closure of the pulmonary arterial end, infective arteritis and external trauma in a patient with a persistent duct, 99 or even coil occlusion of a preexisting patent duct. 100 Aneurysm of the duct should be considered in the differential diagnosis of adult with unexplained mediastinal masses seen on chest radiography. The diagnosis can be confirmed by aortography or by computerised tomography. 101,102 Like the pattern seen in infancy, the high incidence of rupture, embolisation, and effects of pressure suggest that surgical excision is advisable. Surgical ligation may itself be followed by aneurysmal formation, 95,103 often associated with recanalisation.
