Browse the corpus

Introduction In a seminal work published nearly three-quarters of a century ago, Kramer, when considering the extant knowledge regarding the development of the outflow tracts, pointed to the need for “a redefinition of the terms used in describing the structures involved.”1 This was, in part, because of the “discrepant terms employed by previous workers.” Equally important, in his opinion, was the influence of “the rapidly changing shapes and locations in which the structures themselves are found at different stages of development.” Kramer described important new findings in his own contribution, not least the appearance of the intercalated cushions. As he showed, it was these structures that formed the primordiums of the arterial valves. It is questionable, however, whether the terms he coined to solve the problem of previous discrepancies, namely, “truncus” and “conus,” achieved the understanding he was seeking. This is because, as far as we are aware, there is no current agreement as to whether the arterial roots, formed by incorporating his newly described intercalated cushions, belong to the truncus or the conus. His terms, nonetheless, have achieved widespread acceptance, leading to the ongoing description of “conotruncal anomalies.” This practice is not without its own problems, since lesions of the arterial valves, such as the bicuspid aortic valve, the commonest congenital cardiac malformation,2 are not currently classified within the “conotruncal” category. Other lesions, in contrast, such as discordant atrioventricular connections, are included by some as representing conotruncal malformations,3 even though the primary process underlying their formation involves abnormal ventricular looping. The major reason why the suggestions of Kramer, important as they have been, have not matched his expectations is that the postnatal outflow tracts have three, rather than two, components. These are the intrapericardial arterial trunks, the arterial valvar leaflets and their supporting sinuses, collectively forming the arterial roots, and the subvalvar ventricular outflow tracts.

hey have been, have not matched his expectations is that the postnatal outflow tracts have three, rather than two, components. These are the intrapericardial arterial trunks, the arterial valvar leaflets and their supporting sinuses, collectively forming the arterial roots, and the subvalvar ventricular outflow tracts. When assessed in such tripartite fashion, development can be described in a fashion that permits direct correlations to be made with the multiple congenital cardiac lesions known to afflict the outflow tracts.4-6 In this review, we show how the tripartite approach clarifies knowledge concerning both the development and morphology of the ventricular outflow tracts. Development of the Heart Tube Our knowledge of cardiac development has changed immeasurably over the past two or three decades. In part, this has depended on the advances made by molecular biologists. These moves have been matched by an increased ability to examine the anatomy of the developing heart using techniques such as scanning electron microscopy or high-resolution episcopic microscopy.7 The latter opportunities have revolutionized our capabilities of visualizing the morphological changes. Use of these techniques confirms that the outflow tract develops in a tripartite fashion,5,6 permitting more obvious appreciation of the changes that occur early in the development of the heart tube.

pic microscopy.7 The latter opportunities have revolutionized our capabilities of visualizing the morphological changes. Use of these techniques confirms that the outflow tract develops in a tripartite fashion,5,6 permitting more obvious appreciation of the changes that occur early in the development of the heart tube. It used to be thought that all parts of the definitive heart were represented within the initial linear heart tube. Evidence had existed for some time, nonetheless, that new cells were continuously added during development to its venous and arterial poles.8 These initial investigations have now been confirmed by rigorous molecular studies.9 The source of the newly added tissues is known as the second heart field, the initial linear tube being derived from the first heart field. The addition of the new cells from the second field to the cranial pole of the initial heart tube provides the material for formation of both the right ventricle and the outflow tract. These new cells provide not only the myocardial components of the right ventricle and the outflow tract but also the nonmyocardial intrapericardial arterial trunks, along with their valves and sinuses. When first seen, the initial heart tube is straight (Figure 1A), but with the addition of the new material, it becomes S-shaped by the process of looping (Figure 1B). As the tube elongates and loops, its lumen, throughout its length, is lined with cardiac jelly. Having looped, it retains a solitary lumen. The initial steps for the production of the eventual right- and left-sided chambers involve the process that has become known as ballooning.10 At the atrial level, expansions to right and left from the common atrial component of the initial tube produce the primordiums of the atrial appendages. At the level of the ventricular loop, the process involves expansions from its outer curvature. These take place in series, with the eventual apical component of the left ventricle expanding from the inlet component of the loop and that of the right ventricle expanding from the outlet part. Subsequent to looping, the outlet component of the initial heart tube extends from the cavity of the developing right ventricle to the margins of the pericardial cavity, where its initial solitary cavity becomes continuous with the aortic sac. The outlet portion then shows a significant dog-leg bend, permitting recognition of proximal, intermediate, and distal components (Figure 2A).

ube extends from the cavity of the developing right ventricle to the margins of the pericardial cavity, where its initial solitary cavity becomes continuous with the aortic sac. The outlet portion then shows a significant dog-leg bend, permitting recognition of proximal, intermediate, and distal components (Figure 2A). Figure 1. The images, prepared using dissection and scanning electron microscopy, show the developing heart tube as seen in the mouse early (panel A) and late (panel B) during the ninth day of embryonic development. This is equivalent of around five weeks of development in man. Figure 2. The images are from an episcopic data set prepared from a human embryo at Carnegie stage 14, representing the end of the fifth week of development. The left panel (A) shows the dog-leg bend in the outflow tract, which at this early stage is supported exclusively from the developing right ventricle. The right panel (B), cut in frontal fashion at the junction of the outflow tract with the pharyngeal mesenchyme, shows the origins of the pharyngeal arch arteries from the aortic sac. The back wall of the sac, shown by the star, is the putative aortopulmonary septum. The two single-headed arrows show the extent of the pericardial cavity.

, cut in frontal fashion at the junction of the outflow tract with the pharyngeal mesenchyme, shows the origins of the pharyngeal arch arteries from the aortic sac. The back wall of the sac, shown by the star, is the putative aortopulmonary septum. The two single-headed arrows show the extent of the pericardial cavity. At this early stage, the walls of the outflow tract are exclusively myocardial, with its lumen lined circumferentially by cardiac jelly. With the further addition of nonmyocardial tissue from the second heart field, there is regression of the distal myocardial border away from the margins of the pericardial cavity, with the contained cardiac jelly also effectively being shifted from the pericardial margins. The new nonmyocardial tissues are initially seen as cranial and caudal spurs at the pericardial border. As they extend into the pericardial cavity, the spurs rotate and form rightward and leftward tongues. These components will eventually form the parietal walls of the intrapericardial arterial trunks. As the myocardial border of the tube regresses proximally relative to the pericardial margins, the cardiac jelly lining its walls becomes converted into endocardial cushions by the process of endothelial-to-mesenchymal transformation.11 The outflow cushions, further populated by cells derived from the neural crest, then extend in a spiraling fashion through the intermediate and proximal parts of the outflow tract, running within the myocardial walls of these two components (Figure 3).

ions by the process of endothelial-to-mesenchymal transformation.11 The outflow cushions, further populated by cells derived from the neural crest, then extend in a spiraling fashion through the intermediate and proximal parts of the outflow tract, running within the myocardial walls of these two components (Figure 3). Figure 3. The image shows the outflow cushions reconstructed from an episcopic data set prepared from a mouse embryo killed early during the 12th day of development. The outflow cushions, shown in green and yellow, no longer reach the margins of the pericardial cavity (white arrows with red borders). The solitary lumen of the distal outflow tract becomes continuous dorsally with the cavity of the aortic sac, from which arise the pharyngeal arch arteries, at this stage bilaterally symmetrical.

cushions, shown in green and yellow, no longer reach the margins of the pericardial cavity (white arrows with red borders). The solitary lumen of the distal outflow tract becomes continuous dorsally with the cavity of the aortic sac, from which arise the pharyngeal arch arteries, at this stage bilaterally symmetrical. Further Development of the Outflow Tract The newly formed outflow cushions, which extend through the intermediate and proximal components of the outflow tract in a spiraling fashion (Figure 3), face each other throughout their length. At the initial stage, however, they have yet to fuse. Initially, therefore, the outflow tract has a common lumen. At the margins of the pericardial cavity, the lumen is continuous with the lumens of the pharyngeal arch arteries, which arise in a bilaterally symmetrical fashion from the aortic sac. The arteries extending through the fourth pharyngeal arches, which will become the systemic vessels, arise from the cranial part of the sac, with the arteries of the sixth arches, which give rise to the developing right and left pulmonary arteries, arising caudally. As the nonmyocardial tongues of tissue extend into the distal part of the outflow tract, the right-sided tongue, which will become the parietal wall of the intrapericardial aorta, is longest in cranial-to-caudal direction. The left-sided wall, destined to be the parietal wall of the pulmonary trunk, has greater length in ventrocaudal direction. Concomitant with the growth into the heart of these parietal tongues, there has been growth of the dorsal wall of the aortic sac between the origins of the fourth and sixth arch arteries. This protrusion, which is covered by cells derived from the neural crest, but with a core derived from the second heart field,5 grows obliquely into the distal outflow tract and forms the aortopulmonary septum (Figure 4). The cushions themselves have by now fused at their distal ends, with the proximal parts remaining unfused. Fusion of the protrusion with the distal ends of the cushions then obliterates the aortopulmonary foramen, separating the distal outflow tract into intrapericardial aortic and pulmonary channels (Figure 5A). It is only subsequent to this stage that it becomes possible to recognize columns of condensed mesenchyme, which occupy the central parts of the unfused proximal cushions (Figure 5B).

terates the aortopulmonary foramen, separating the distal outflow tract into intrapericardial aortic and pulmonary channels (Figure 5A). It is only subsequent to this stage that it becomes possible to recognize columns of condensed mesenchyme, which occupy the central parts of the unfused proximal cushions (Figure 5B). Figure 4. The images are both made using episcopic data sets from mice embryos killed toward the end of the 12th day of development. The left panel (A) shows a frontal section revealing the oblique nature of the ventral protrusion, which is the aortopulmonary septum. At this stage, there is an aortopulmonary foramen (double-headed white arrow) between the proximal end of the protrusion and the fused distal end of the outflow cushions. The white arrows with dark borders show the distal margin of the myocardium, which now encloses only the intermediate and proximal parts of the outflow tract. Panel B, from a different data set, shows the aortopulmonary foramen as seen from the cavity of the aorta, having cut away the parietal aortic wall. The white arrows with dark borders in this panel show the distal extent of the pericardial cavity.

encloses only the intermediate and proximal parts of the outflow tract. Panel B, from a different data set, shows the aortopulmonary foramen as seen from the cavity of the aorta, having cut away the parietal aortic wall. The white arrows with dark borders in this panel show the distal extent of the pericardial cavity. Figure 5. Panel A is prepared using an episcopic data set from a mouse embryo killed at the end of the 12th day of development. It shows how the aortopulmonary septum, formed by the protrusion from the dorsal wall of the aortic sac, has fused with the distal ends of the outflow cushions (dashed white line). The white arrows with dark borders show the extent of the pericardial cavity. The distal outflow tract is fully divided at this stage. The fused distal cushions have also divided the intermediate part of the outflow tract. Note the appearance of the intercalated cushions (white stars with dark borders) in this middle component. The outflow cushions, however, remain unfused in the proximal part of the outflow tract. Panel B is from an episcopic data set prepared from a human embryo at Carnegie stage 16. It shows how columns of condensed mesenchyme, derived from the cells migrating from the neural crest, occupy the fused distal and the unfused proximal parts of the outflow cushions. They are not involved with separating the intrapericardial arterial trunks, which have already been separated by the aortopulmonary septum, formed from the protrusion of the dorsal wall of the aortic sac.

ls migrating from the neural crest, occupy the fused distal and the unfused proximal parts of the outflow cushions. They are not involved with separating the intrapericardial arterial trunks, which have already been separated by the aortopulmonary septum, formed from the protrusion of the dorsal wall of the aortic sac. Some investigators have considered these columns, and the cushions containing them, to represent an “aortopulmonary septal complex.”12 In reality, the distal ends of the cushions separate the intermediate outflow tract into the arterial roots, while the subsequent fusion of the proximal cushions will divide the proximal outflow tract into the subvalvar ventricular outlets. As the fusion of the cushions proceeds progressively from distal to more proximal positions along the outflow tract, the fused tissue forms an effective septum that initially separates the aortic and pulmonary channels. Fusion is, however, rapidly succeeded by physical separation of the two outflow tracts, with discrete walls forming through the fused cushion tissue. The fusion of the embryonic aortopulmonary septum with the distal ends of the fused outflow cushions then places the right-sided intrapericardial aorta in continuity extrapericardially with the fourth arch derivatives. The same process places the extrapericardial pulmonary arteries, and the left-sided sixth arch artery, now recognizable as the arterial duct, in continuity with the intrapericardial pulmonary trunk. On the basis of presumed abnormal persistence of the different parts of the initially bilateral system of extrapericardial pharyngeal arch arteries, it is possible to explain all the various forms of vascular rings.

artery, now recognizable as the arterial duct, in continuity with the intrapericardial pulmonary trunk. On the basis of presumed abnormal persistence of the different parts of the initially bilateral system of extrapericardial pharyngeal arch arteries, it is possible to explain all the various forms of vascular rings. Within the intermediate part of the outflow tract, still walled by myocardium, it is the appearance of the intercalated cushions (Figure 5A) that provides the primordiums for the formation of the arterial roots. Thus, within the intermediate part of the outflow tract, which gives rise to the arterial roots, fusion across the central portion of the major cushions leaves two distinct valvar primordiums clearly recognizable by their trifoliate pattern. In both cases, this is produced by the unfused edges from the parietal portions of the major cushions, interdigitated by an intercalated cushion (Figure 6A). The process of fusion itself is dependent on the presence of cells derived from the neural crest.13 Once fusion has taken place, however, the material derived from the neural crest becomes increasingly insignificant, eventually disappearing as the aortic and pulmonary roots separate one from the other. Separation of the roots, and also the proximal outflow tracts, occurs at right angles to the line of fusion between the central cushions. By the time the two roots have separated, the distal margins of the cushions have excavated to produce the valvar leaflets, and their semilunar hinges, these processes taking place within the supporting walls of the intermediate part of the outflow tract, which initially remain myocardial (Figures 6B and 7).

cushions. By the time the two roots have separated, the distal margins of the cushions have excavated to produce the valvar leaflets, and their semilunar hinges, these processes taking place within the supporting walls of the intermediate part of the outflow tract, which initially remain myocardial (Figures 6B and 7). Figure 6. The left panel (A) is prepared using an episcopic data set from a mouse embryo killed during the 13th day of development. The short axis of the intermediate part of the outflow tract is viewed from above, showing the developing primordiums of the aortic root. Note that, at this stage, the roots remained encased in a turret of outflow tract myocardium. The right panel (B) is from an episcopic data set prepared using a human embryo at Carnegie stage 20. The three parts of the outflow tract are shown, with the distal cushions excavating to form the leaflets of the pulmonary valve, and the surface of the fused proximal cushions muscularizing to form the subpulmonary infundibulum. At this stage, the intermediate and proximal parts of the outflow tract retain their myocardial walls. Figure 7. The image, a magnification of the section shown as Figure 6B, reveals the start of the excavation of the cushions that will eventually produce the leaflets of the arterial valves. At this stage, the intermediate component of the outflow tract remains encased within a myocardial turret.

Figure 6. The left panel (A) is prepared using an episcopic data set from a mouse embryo killed during the 13th day of development. The short axis of the intermediate part of the outflow tract is viewed from above, showing the developing primordiums of the aortic root. Note that, at this stage, the roots remained encased in a turret of outflow tract myocardium. The right panel (B) is from an episcopic data set prepared using a human embryo at Carnegie stage 20. The three parts of the outflow tract are shown, with the distal cushions excavating to form the leaflets of the pulmonary valve, and the surface of the fused proximal cushions muscularizing to form the subpulmonary infundibulum. At this stage, the intermediate and proximal parts of the outflow tract retain their myocardial walls. Figure 7. The image, a magnification of the section shown as Figure 6B, reveals the start of the excavation of the cushions that will eventually produce the leaflets of the arterial valves. At this stage, the intermediate component of the outflow tract remains encased within a myocardial turret. There is then still further migration of nonmyocardial tissues derived from the second heart field in proximal direction to produce the valvar sinuses of the arterial roots, this process in itself contributing to ongoing effective proximal regression of the distal myocardial border. Initially, therefore, the developing valvar leaflets are hinged exclusively from the myocardium. It is later in development, concomitant with the formation of the valvar sinuses, that the developing leaflets achieve their semilunar configurations, with the distal parts hinged from the newly formed arterial walls, and the levels of the most distal margins then recognizable as the developing sinotubular junctions. By the time the excavation of the leaflets is complete, the aortic root has been transferred to the left ventricle by remolding of the proximal outflow tract. This is achieved by the leftward movement of the proximal part of the outflow tract across the crest of the muscular ventricular septum, with concomitant remodeling of the embryonic interventricular communication. This communication is initially roofed by the myocardial inner heart curvature (Figure 8A). As the dorsal part of the proximal outflow tract moves toward the left ventricle, so the proximal parts of the outflow cushions, which themselves are fusing during this process, are brought into line with the crest of the muscular septum. This means that the initial embryonic interventricular communication, roofed by the inner heart curve, becomes the outflow tract for the left ventricle (Figure 8B).

o the proximal parts of the outflow cushions, which themselves are fusing during this process, are brought into line with the crest of the muscular septum. This means that the initial embryonic interventricular communication, roofed by the inner heart curve, becomes the outflow tract for the left ventricle (Figure 8B). Figure 8. The images are from episcopic data sets prepared from developing mice killed during the 13th embryonic day (panel A) and at the beginning of the 14th day of development (panel B). They show that, as the aortic root is transferred so as to take origin from the left as opposed to the right ventricle, the plane of the initial interventricular communication (dashed double-headed white arrow) becomes the left ventricular outflow tract. This plane is roofed by the inner heart curvature. The plane from the ventricular septal crest to the fused proximal outflow cushions (double-headed solid white arrow) then becomes the interventricular communication. Note that, at this stage, the distal myocardial boundary remains confluent with the edges of the excavating distal cushions.

by the inner heart curvature. The plane from the ventricular septal crest to the fused proximal outflow cushions (double-headed solid white arrow) then becomes the interventricular communication. Note that, at this stage, the distal myocardial boundary remains confluent with the edges of the excavating distal cushions. As well as fusing during this process, the proximal outflow cushions also muscularize.14 The core of the cushions, packed with cells derived from the neural crest, then attenuates. The muscularized surface will eventually form the septal component of the subpulmonary infundibulum. It is the attenuation of the core that produces the extracavitary tissue plane separating the infundibulum from the newly formed aortic root. Conversion of the core of the cushions to extracavitary fibroadipose tissue means that the proximal cushions lose their initial septal configuration. The initial embryonic interventricular communication by now having remodeled to form the subaortic outflow tract, the persisting interventricular communication is closed by the formation of the membranous septum derived from the rightward tips of the atrioventricular cushions (Figure 9A). The transfer of the aorta to the left ventricle, therefore, proceeds by alignment of the muscularizing proximal outflow cushions with the apical muscular ventricular septum. When transfer is complete, and the persisting embryonic interventricular communication has been closed by the formation of the membranous septum, the aortic valvar leaflets initially remain supported in their entirety by myocardial tissues, with the inner heart curvature continuing to interpose between the leaflets of the developing aortic and mitral valves (Figure 9B). It is later during fetal development that this muscular fold becomes transformed into the region of aortic-to-mitral valvar fibrous continuity, this being one of the features of the postnatal left ventricular outflow tract.

inuing to interpose between the leaflets of the developing aortic and mitral valves (Figure 9B). It is later during fetal development that this muscular fold becomes transformed into the region of aortic-to-mitral valvar fibrous continuity, this being one of the features of the postnatal left ventricular outflow tract. Figure 9. The images are from different episcopic data sets prepared from mice killed during the 15th day of embryonic development. They show, in the left panel (A), how the rightward margins of the atrioventricular (AV) cushions close the persisting interventricular communication. The right panel shows how, subsequent to closure of the foramen, the inner heart curvature remains interposed between the developing leaflets of the aortic and mitral valves. Note the developing sinuses of the aortic root and the three components of the left ventricular outflow tract.

ng interventricular communication. The right panel shows how, subsequent to closure of the foramen, the inner heart curvature remains interposed between the developing leaflets of the aortic and mitral valves. Note the developing sinuses of the aortic root and the three components of the left ventricular outflow tract. To summarize the processes of development, from the outset, it is possible to recognize proximal, intermediate, and distal part of the outflow component of the primary heart tube. With ongoing contributions of nonmyocardial tissues from the second heart field, the distal part of the outflow tract becomes transformed into the intrapericardial trunks. The cavities of the trunks are initially separated by an embryonic aortopulmonary septum, but this disappears as the intrapericardial aorta and pulmonary trunk develop their own discrete walls. The intermediate part of the outflow tract, subsequent to the formation within it of the intercalated cushions, becomes transformed into the arterial roots. The distal cushions, along with the intercalated cushions, excavate to form the semilunar valvar leaflets, whereas the valvar sinuses are formed from ongoing contributions of nonmyocardial tissues from the second heart field. As the roots separate one from the other, so the intermediate cushions lose their initial septal function. The proximal cushions fuse both with each other and the crest of the muscular ventricular septum. It is the muscularization of their surface that produces the dorsal component of the right ventricular infundibulum (Figure 6B). Only after the aortic component of the proximal outflow tract has been transferred to the left ventricle, can there be closure of the remodeled embryonic interventricular communication (Figure 9A). And only after this has taken place, does the left ventricular component of the inner heart curvature undergo transformation to produce the area of aortic-to-mitral fibrous continuity that forms the roof of the definitive left ventricle.

be closure of the remodeled embryonic interventricular communication (Figure 9A). And only after this has taken place, does the left ventricular component of the inner heart curvature undergo transformation to produce the area of aortic-to-mitral fibrous continuity that forms the roof of the definitive left ventricle. Morphology of the Postnatal Outflow Tracts Just as with the investigation of the developing heart, the emergence of new diagnostic techniques has improved our ability to demonstrate the nuances of living cardiac anatomy. To achieve this, we have taken advantage of the preparation of three-dimensional data sets acquired during the investigation of coronary arterial disease to perform virtual dissections of normal human hearts.15 These images not only serve to validate our earlier descriptions of the morphology of the intrapericardial outflow tracts,16-18 but at the same time they confirm the tripartite arrangement of their constituent parts. They also show well the locations of the pericardial reflections that divide the arterial trunks into their intrapericardial and extrapericardial parts (Figure 10). Further virtual dissection of the data sets then shows well how both of the ventricular outflow tracts possess three parts. These are the intrapericardial component of the arterial trunk, the arterial root, and the subvalvar ventricular outflow tract. In the morphologically right ventricle, the subvalvar component is a complete muscular sleeve, which lifts the root away from the cardiac base, whereas in the left ventricle, the posterior wall of the outflow tract is fibrous, being made up of fibrous continuity between the leaflets of the arterial and atrioventricular valves (Figure 11). In the normal heart, the intrapericardial arterial trunks spiral as they extend from the arterial roots to the margins of the pericardial cavity (Figure 12). Although the developing intrapericardial trunks had initially been separated by the aortopulmonary septum (Figure 4A), we know that the central core of the embryonic septum is attenuated by term, so that, in postnatal life, the trunks have their own discrete walls, with no septal component between them.

ty (Figure 12). Although the developing intrapericardial trunks had initially been separated by the aortopulmonary septum (Figure 4A), we know that the central core of the embryonic septum is attenuated by term, so that, in postnatal life, the trunks have their own discrete walls, with no septal component between them. Figure 10. The image is prepared from a data set obtained using multidetector computed tomography in a patient undergoing investigation for coronary arterial disease. The pericardial reflections divide the ascending aorta into intrapericardial and extrapericardial components. The pulmonary trunk branches at the margins of the pericardial cavity into the right and left pulmonary arteries. Figure 11. The virtual dissections are made from a data set obtained using multidetector computed tomography in a patient undergoing investigation for coronary arterial disease. They show how each outflow tract is formed in a tripartite fashion, with the components represented by the intrapericardial arterial trunk, the arterial root, and the subvalvar outflow tract, respectively. Note that, in the right ventricle (panel A), the subvalvar component is a completely muscular infundibular sleeve, whereas in the left ventricle (panel B), the posterior wall of the subvalvar area is formed by fibrous continuity between the leaflets of the aortic and mitral valves. Figure 12. The reconstruction, seen in short-axis projection from the cardiac apex, shows the spiraling nature of the intrapericardial arterial trunks. The aorta is shown in red, and the pulmonary trunk in blue.

Figure 11. The virtual dissections are made from a data set obtained using multidetector computed tomography in a patient undergoing investigation for coronary arterial disease. They show how each outflow tract is formed in a tripartite fashion, with the components represented by the intrapericardial arterial trunk, the arterial root, and the subvalvar outflow tract, respectively. Note that, in the right ventricle (panel A), the subvalvar component is a completely muscular infundibular sleeve, whereas in the left ventricle (panel B), the posterior wall of the subvalvar area is formed by fibrous continuity between the leaflets of the aortic and mitral valves. Figure 12. The reconstruction, seen in short-axis projection from the cardiac apex, shows the spiraling nature of the intrapericardial arterial trunks. The aorta is shown in red, and the pulmonary trunk in blue. The arterial roots occupy the middle parts of the outflow tracts and are limited distally by the sinotubular junctions, there being the areas where the semilunar hinges of the arterial valvar leaflets come together at the periphery of the valvar orifices (Figure 13). There is no discrete anatomical structure that marks the proximal boundaries of the arterial roots. Instead these are represented by virtual planes constructed by joining together the proximal attachments of the hinges of the valvar leaflets. These virtual basal planes, furthermore, are proximal to the anatomic ventriculoarterial junctions. The latter boundaries, true anatomic borders, are formed at the points where the walls of the valvar sinuses are supported by the ventricular walls. In the right ventricle, the entirety of the walls of the arterial valvar sinuses is supported by infundibular musculature.17 In the aortic root, in contrast, because the fibrous continuity found posteriorly between the leaflets of the aortic and mitral valves, only the two aortic valvar sinuses giving rise to the coronary arteries have myocardium incorporated into their bases (Figure 13B). The entrances to the arterial roots, therefore, have no discrete anatomic boundaries. These are, nonetheless, the dimensions that are usually measured by echocardiographers as representing the valvar “annulus.” The reconstructions made possible by the availability of the three-dimensional data sets show that the hinges of the leaflets of the arterial valves, unlike those of the atrioventricular valves, are arranged in the form of three-pointed coronets rather than as little rings (Figure 14).

epresenting the valvar “annulus.” The reconstructions made possible by the availability of the three-dimensional data sets show that the hinges of the leaflets of the arterial valves, unlike those of the atrioventricular valves, are arranged in the form of three-pointed coronets rather than as little rings (Figure 14). Figure 13. The virtual dissections show how the pulmonary (panel A) and aortic (panel B) roots are limited distally by the sinotubular junctions (solid double-headed arrow) but proximally by a virtual plane made by joining together the basal attachments of the semilunar leaflets (double-headed dashed arrow). Note that the virtual basal plane is proximal to the anatomic ventriculoarterial junctions (white arrows with dark borders). Figure 14. The three-dimensional data set made available by means of multidetector computed tomography from a patient undergoing investigation of coronary arterial disease has been segmented and reconstructed to show the hinges of the leaflets of the arterial and atrioventricular valves. The leaflets of the atrioventricular valves are hinged in an annular fashion, whereas the hinges of the arterial valves, when reconstructed, take the form of three-pointed coronets.

nary arterial disease has been segmented and reconstructed to show the hinges of the leaflets of the arterial and atrioventricular valves. The leaflets of the atrioventricular valves are hinged in an annular fashion, whereas the hinges of the arterial valves, when reconstructed, take the form of three-pointed coronets. These features produce the problems that continue with regard to the definition of the arterial valvar annuluses. As stated above, echocardiographers define the virtual basal plane as the annulus, even though it is not marked by any anatomic entity. In contrast, many, but not all, surgeons consider the semilunar hinge lines to represent the annulus. This is one of the disagreements contributing to the “tower of Babel” currently existing with regard to the naming of the arterial roots.19 One of us had already expressed his own opinion that the virtual plane is best considered as the annulus,20 not least since this is the measurement most frequently provided to surgeons when assessing valvar morphology. The virtual plane, furthermore, is also annular, unlike the coronet-like arrangements of the valvar hinges.

us had already expressed his own opinion that the virtual plane is best considered as the annulus,20 not least since this is the measurement most frequently provided to surgeons when assessing valvar morphology. The virtual plane, furthermore, is also annular, unlike the coronet-like arrangements of the valvar hinges. Another problem contributing to the tower of Babel is the variable use of the word “cusp” when describing the components of the root. Some use the word to describe the leaflets, whereas others use it to account for the valvar sinuses. Our own preference is to avoid its use altogether, since when used in the vernacular sense, it accounts for a point or the crossing of two curves. It is by distinguishing between leaflets and sinuses, and avoiding the use of cusp, that we are able to provide a more accurate account of the morphology of the arterial roots. Assessment in this fashion also serves to emphasize the significance of the fibrous interleaflet triangles,21 which fill the spaces on the ventricular aspect of the sinuses. It is the sinuses themselves, of course, that support the distal parts of the valvar leaflets in semilunar fashion (Figure 15). During development, as we have shown, the entirety of the developing arterial roots is enclosed with the myocardial wall of the intermediate part of the outflow tract (Figure 5A). As we have also shown, the myocardial border regresses toward the cardiac base concomitant with the formation of the nonmyocardial valvar sinuses (Figure 9B). It is because of the myocardial regression that the apexes of the interleaflet triangles, in postnatal life, are able to point to extracardiac areas. These are the tissue plane between the aortic root and the infundibulum and the transverse pericardial sinus (Figure 16). Recognition of the relationships of these triangles is achieving increasing importance with the realization that many patients with acquired disease of the aortic valve can undergo reconstructive procedures as opposed to valvar replacement.22 The triangles are poorly formed and hypoplastic in the setting of congenital bicuspid and unicuspid valves (see below).

triangles is achieving increasing importance with the realization that many patients with acquired disease of the aortic valve can undergo reconstructive procedures as opposed to valvar replacement.22 The triangles are poorly formed and hypoplastic in the setting of congenital bicuspid and unicuspid valves (see below). Figure 15. The outflow tract of the normal heart has been opened through an incision across the left coronary sinus of the aortic root. The membranous septum has been transilluminated from the right side. This shows how the triangular space between the right coronary and the nonadjacent sinuses of the root is filled by a fibrous wall. Similar triangles with fibrous walls are to be found beneath each of the valvar commissures, these being the points at which the leaflets join together at the sinotubular junction (white stars with dark borders). The relationships of these triangles are shown in Figure 16. Note how the virtual basal plane is created by joining together the basal attachments of the valvar leaflets (dotted line).

ar commissures, these being the points at which the leaflets join together at the sinotubular junction (white stars with dark borders). The relationships of these triangles are shown in Figure 16. Note how the virtual basal plane is created by joining together the basal attachments of the valvar leaflets (dotted line). Figure 16. Virtual dissection of the aortic root shows how the tips of each of the interleaflet triangles (white arrows with dark borders) point outside the cardiac cavities. Panel A shows a cut through the triangle between the right and left coronary aortic sinuses and is viewed from behind. Panel B shows a cut across the triangle between the right coronary and the nonadjacent aortic sinuses and is shown from the front. Note that the membranous septum forms the base of this triangle (see also Figure 15). Panel C, showing a cut through the triangle between the left coronary and the nonadjacent aortic sinuses, is viewed from the left side.

riangle between the right coronary and the nonadjacent aortic sinuses and is shown from the front. Note that the membranous septum forms the base of this triangle (see also Figure 15). Panel C, showing a cut through the triangle between the left coronary and the nonadjacent aortic sinuses, is viewed from the left side. The third component of the outflow tracts is the outlet parts of the ventricles. As we have emphasized, these differ in the two ventricles, since the right ventricular outlet is a completely muscular infundibular sleeve. The anterior part of the sleeve is the parietal wall of the right ventricle, with the posterior part formed by the supraventricular crest (Figure 17). In the past, we considered the entirety of the posterior part of the infundibulum to be formed by the muscular outlet septum.23 We know now that this is not the case.6 The core of the proximal outflow cushions, which initially formed an embryonic outlet septum, attenuates subsequent to muscularization of their surface to produce the posterior wall of the infundibulum (Figure 6B). At the site of the attenuating tissues, we eventually find the extracavitary fibroadipose tissue that interposes postnatally between the newly formed infundibulum and the aortic root (Figure 17B).

m, attenuates subsequent to muscularization of their surface to produce the posterior wall of the infundibulum (Figure 6B). At the site of the attenuating tissues, we eventually find the extracavitary fibroadipose tissue that interposes postnatally between the newly formed infundibulum and the aortic root (Figure 17B). Figure 17. The images show virtual dissections revealing the structure of the right ventricular infundibulum. The left panel (A) is made by cutting away the parietal wall of the right ventricle, showing how the posterior wall of the infundibulum is formed by the supraventricular crest. The larger part of this wall is no more than the inner heart curvature. The component adjacent to the aortic root, however, is the free-standing infundibular sleeve, the presence of which makes it possible to remove the root for use as an autograft in the Ross procedure. The right panel (B) shows this sleeve, revealing the tissue place that separates it from the aortic root.

art curvature. The component adjacent to the aortic root, however, is the free-standing infundibular sleeve, the presence of which makes it possible to remove the root for use as an autograft in the Ross procedure. The right panel (B) shows this sleeve, revealing the tissue place that separates it from the aortic root. Correlations With Outflow Tract Malformations We have already discussed how interpretation of inappropriate remodeling and attenuation of the initially bilaterally symmetrical arrangement of the arteries extending through the pharyngeal arches is able to provide rational explanations for the various lesions known as vascular rings and slings. These are malformations of the extrapericardial arterial trunks, including the enigmatic fifth arch artery.24 It is the remodeling of the extrapericardial arterial trunks and their branches that provides alternative explanations for several of the lesions currently interpreted on the basis of retention of the fifth arch arteries. One of the problems with the conventional interpretations is that neither the alleged fifth pharyngeal arch nor its supposed artery is to be found in developing murine embryos.25 A remnant of the artery, nonetheless, has been found in one developing human embryo,24 so the conventional interpretations can remain valid, although there are better explanations for some of the lesions interpreted in this fashion.25 All of the lesions interpreted on the basis of persistence of the fifth arch artery, nonetheless, along with lesions involving persistent patency of the arterial duct, are extrapericardial malformations.

ions can remain valid, although there are better explanations for some of the lesions interpreted in this fashion.25 All of the lesions interpreted on the basis of persistence of the fifth arch artery, nonetheless, along with lesions involving persistent patency of the arterial duct, are extrapericardial malformations. Lesions such as aortopulmonary window, or anomalous origin of the right pulmonary artery from the aorta, are intrapericardial. They are best explained on the basis of failure to close the embryonic aortopulmonary foramen, this in turn reflecting inadequate growth of the aortopulmonary septum from the dorsal wall of the aortic sac. Failure of the latter process explains well why patients with aortopulmonary window are frequently found with the right pulmonary artery arising from the aortic side of the window (Figure 18). Eccentric growth, but subsequent fusion, of the aortopulmonary septum with the distal ends of the outflow cushions provides a plausible explanation for origin of the right pulmonary artery from the ascending aorta. The separate origins of the aortic and pulmonary roots in this lesion point to the inadequacy of using the term “hemitruncus” for its description.

of the aortopulmonary septum with the distal ends of the outflow cushions provides a plausible explanation for origin of the right pulmonary artery from the ascending aorta. The separate origins of the aortic and pulmonary roots in this lesion point to the inadequacy of using the term “hemitruncus” for its description. Figure 18. The images show an aortopulmonary window (panel A), with origin of the right pulmonary artery on the aortic side of the window (panel B). This is well explained by inadequate growth of the aortopulmonary septum from the dorsal wall of the aortic sac (see Figure 4A), leaving a fold between the walls of the aorta and the pulmonary trunk at the margins of the pericardial cavity (star in panel C). Note the separate formation of the aortic and pulmonary roots, implying normal septation and separation of the intermediate and proximal components of the developing outflow tract.

gure 4A), leaving a fold between the walls of the aorta and the pulmonary trunk at the margins of the pericardial cavity (star in panel C). Note the separate formation of the aortic and pulmonary roots, implying normal septation and separation of the intermediate and proximal components of the developing outflow tract. The essence of hearts having a common arterial trunk is the commonality of the ventriculoarterial junction. This finding points to another of the problems of using truncus and conus for describing the components of the developing outflow tract. It is the ventriculoarterial junctions that are common in this setting of common arterial trunk rather than the arterial trunks. When accounting for development in tripartite fashion, there are no such problems, since we know that the abnormal mechanism is failure of fusion of the major outflow cushions (Figure 19). It had previously been suggested that the mechanism for formation of common arterial trunk, or “truncus arteriosus communis,” was failure of formation of the subpulmonary outflow tract.26 Van Mierop and colleagues, having studied a Keeshond dog exhibiting the lesion, showed that the morphogenetic culprit was failure of fusion of the major cushions within the developing outflow tract.27 Our studies using a colony of mice with knockout of the Tbx1 gene have confirmed their observations.6 The presence of both intercalated cushions in hearts destined to produce common arterial valves with four leaflets and sinuses adds further weight to the fact that the lesion reflects failure of septation of the outflow tract rather than inadequate formation of its pulmonary component. We have, nonetheless, also observed developing embryos with gross hypoplasia of one of the intercalated cushions, this feature providing the template for the formation of the trifoliate common truncal valve.6

reflects failure of septation of the outflow tract rather than inadequate formation of its pulmonary component. We have, nonetheless, also observed developing embryos with gross hypoplasia of one of the intercalated cushions, this feature providing the template for the formation of the trifoliate common truncal valve.6 Figure 19. The images are from embryonic mice in which the gene for Tbx1 has been knocked out. All these mice develop with common arterial trunks. The images are from embryos killed on the 13th day of development. The left panel (A) shows the common trunk, feeding the systemic and pulmonary arteries, with no formation of the aortopulmonary septum. It also shows failure of fusion of the outflow cushions. The right panel, from a different embryo at the same stage of development, shows a cross section through the intermediate part of the outflow tract. Both the intercalated cushions are seen, together with the distal ends of the unfused major outflow cushions. This template provides the primordiums for the formation of a common truncal valve with four leaflets. Note that the cushions are contained within the turret of myocardium that surrounds the intermediate part of the outflow tract.

d cushions are seen, together with the distal ends of the unfused major outflow cushions. This template provides the primordiums for the formation of a common truncal valve with four leaflets. Note that the cushions are contained within the turret of myocardium that surrounds the intermediate part of the outflow tract. Failure of fusion of the outflow cushions, therefore, is responsible for producing the common arterial trunk, this lesion being well accepted as a conotruncal anomaly. Excessive fusion of the ends of the major outflow cushions is then the likely causative factor in producing the aortic valve with two leaflets, the conjoined leaflet being formed from the leaflets that normally guard the aortic valvar sinuses giving rise to the coronary arteries (Figure 20A).28 Fusion between one of the major cushions and the aortic intercalated cushion then provides an explanation for the bicuspid valve with the conjoined leaflet derived from the right coronary and nonadjacent aortic valvar leaflets.29 Our reconstructions of the valvar hinges in adult patients with bicuspid aortic valves confirm the lack of formation of the interleaflet triangle, with the hinges at the site of fusion of the developing leaflets failing to rise to the sinotubular junction. Since the morphogenesis of the bicuspid arterial valves is excessive fusion of the developing outflow cushions, we see no reason why these lesions, and the others involving abnormal development of the arterial roots, should not be included in the category of conotruncal anomalies.

g to rise to the sinotubular junction. Since the morphogenesis of the bicuspid arterial valves is excessive fusion of the developing outflow cushions, we see no reason why these lesions, and the others involving abnormal development of the arterial roots, should not be included in the category of conotruncal anomalies. Figure 20. In panel A, we show the features of the bicuspid aortic valve with the conjoined leaflet derived from the leaflets that usually guard the valvar sinuses giving rise to the coronary arteries. There is formation only of the interleaflet triangles between the nonadjacent leaflet and the conjoined leaflet (white arrows with dark borders), with a raphe formed at the anticipated site of the interleaflet triangle between the two coronary aortic sinuses. Panel B shows reconstruction of the hinges of the valvar leaflets in an adult with a bicuspid aortic valve and conjunction of the leaflets guarding the coronary aortic sinuses. The anticipated point of the coronet at the site of the fused leaflets does not extend to the sinotubular junction (long double-headed arrow), in contrast to the zones of apposition with the other leaflet, which extend to the sinotubular junction (short double-headed arrows).

flets guarding the coronary aortic sinuses. The anticipated point of the coronet at the site of the fused leaflets does not extend to the sinotubular junction (long double-headed arrow), in contrast to the zones of apposition with the other leaflet, which extend to the sinotubular junction (short double-headed arrows). As well as lesions involving the distal and intermediate components of the developing outflow tracts, many congenital malformations are due to maldevelopment of the proximal outflow tract. Since, when first formed, the developing outflow tract is supported exclusively by the developing right ventricle, it is hardly surprising that double outlet right ventricle is one of these lesions. Double outlet right ventricle, however, is no more than an abnormal ventriculoarterial connection. Multiple phenotypes are to be found in this setting, most frequently categorized according to the location of the interventricular communication relative to the arterial trunks. This channel between the ventricles is most usually described as the “ventricular septal defect.” When both arterial trunks arise postnatally from the right ventricle, the channel is roofed by the inner heart curvature. In this setting, the hole is then the outflow tract from the right ventricle, rather than representing the area to be closed so as to produce potential biventricular circulations. The channel that would, most appropriately, be termed the ventricular septal defect would be roofed by the outlet septum, or its fibrous remnant. The attachments of this outlet septum, or its remnant, within the right ventricle, furthermore, determine the relationships between the interventricular communication and the arterial roots. When attached to the anterior margin of the defect, it is the aortic root that is closest to the left ventricle, whereas when attached to the posterior margin, the defect is placed in subpulmonary location. If, however, the developing outlet septum remains hypoplastic, with minimal formation of the proximal outflow cushions, but with fusion of the distal cushions, then the defect can be doubly committed. This abnormal mechanism of development was seen in our colony of mice in which we perturbed the Furin enzyme (Figure 21A). It is comparable with the variant of double outlet right ventricle with a fibrous rather than a muscular outlet septum (Figure 21B).

of the distal cushions, then the defect can be doubly committed. This abnormal mechanism of development was seen in our colony of mice in which we perturbed the Furin enzyme (Figure 21A). It is comparable with the variant of double outlet right ventricle with a fibrous rather than a muscular outlet septum (Figure 21B). In terms of morphogenesis, the lesion is more akin to common arterial trunk, but with a common ventriculoarterial junction divided into aortic and pulmonary components, as opposed to the other variants of double outlet right ventricle. The feature of the latter lesions is the separate nature of the subpulmonary and subaortic ventricular outflow tracts. The variant with the subpulmonary defect is itself related developmentally to hearts with concordant atrioventricular and discordant ventriculoarterial connections, being part of the Taussig-Bing spectrum.30 The variants with subaortic defects, in contrast, are parts of the spectrums including the Eisenmenger defect and tetralogy of Fallot.31

pulmonary defect is itself related developmentally to hearts with concordant atrioventricular and discordant ventriculoarterial connections, being part of the Taussig-Bing spectrum.30 The variants with subaortic defects, in contrast, are parts of the spectrums including the Eisenmenger defect and tetralogy of Fallot.31 Figure 21. The left panel (A) is taken from an episcopic data set prepared from a developing mouse embryo at the 15th day of development with knockout of the Furin enzyme. The aortopulmonary septum has fused with the distal cushions to produce separate intrapericardial arterial trunks. The outflow cushions themselves are also fused, but with minimal development of the proximal components, which have failed to muscularize. As a result, the arterial trunks remain supported by the right ventricle, the interventricular communication being doubly committed. This arrangement is seen postnatally in the heart with double outlet right ventricle shown in the right panel (B). It has a fibrous rather than a muscular outlet septum. The interventricular communication, opening to the right ventricle between the limbs of the septomarginal trabeculation, is doubly committed.

mmitted. This arrangement is seen postnatally in the heart with double outlet right ventricle shown in the right panel (B). It has a fibrous rather than a muscular outlet septum. The interventricular communication, opening to the right ventricle between the limbs of the septomarginal trabeculation, is doubly committed. When discussing double outlet right ventricle, it is also pertinent to consider the role of the inner heart curvature. For many years, it was considered that persistence of the inner heart curvature, in other words presence of “bilateral conuses,” was the essence of double outlet right ventricle. As we have shown, however, the inner heart curvature remains muscular even after the aorta has been transferred to the left ventricle in the setting of normal development. There is no justification, therefore, for considering persistence of the inner curvature as the phenotypic feature of double outlet right ventricle. Persistence of the inner curvature, also known as the “bulboventricular flange,” has also been correlated with the formation of the anterolateral muscle bundle of the left ventricle.32 Our findings suggest that the inner curve, prior to its conversion to fibrous tissue, forms no more than the roof of the left ventricle. It seems more likely that the anterolateral muscle bundle, when present, is derived by compaction of the trabecular layer of the ventricular walls. Like many of our suggestions, nonetheless, we accept that such a presumption is speculative.

onversion to fibrous tissue, forms no more than the roof of the left ventricle. It seems more likely that the anterolateral muscle bundle, when present, is derived by compaction of the trabecular layer of the ventricular walls. Like many of our suggestions, nonetheless, we accept that such a presumption is speculative. Conclusions Problems continue to exist when seeking logically to correlate the morphogenesis of congenitally malformed hearts with concepts of normal cardiac development and also with categorizations of the so-called conotruncal malformations. In no small part, these relate to the ongoing practice of analyzing the outflow tract in terms of the truncus and the conus. The intrapericardial outflow tracts as seen in the postnatal heart unequivocally possess three components, namely, the intrapericardial arterial trunks, the arterial roots, and the subvalvar ventricular outflow tracts. As we have shown in our review, the problems outlined above are dissipated when development of the outflow tract is similarly analyzed in a tripartite fashion. Thus, it is the distal part of the outflow tract that separates to form the intrapericardial arterial trunks. The arterial roots are formed within the intermediate part of the outflow tract, initially with myocardial walls, which become arterial concomitant with ongoing addition of nonmyocardial tissues from the second heart field. It is the proximal outflow tract that produces the subvalvar outflow tracts. This component is initially supported exclusively by the developing right ventricle. Normal development requires transfer of its posterior part to the left ventricle, thus forming the subaortic outflow tract, with muscularization of the surface of the proximal outflow cushions then producing the posterior part of the newly formed subpulmonary right ventricular infundibulum. Analysis of congenitally malformed hearts is then easily achieved on the basis of the tripartite template. Such analysis reveals the need to reconsider those lesions categorized as representing conotruncal anomalies. It is paradoxical in this regard that common arterial trunk, formed because of failure of fusion of the outflow cushions, is currently included within the conotruncal category, yet arterial valves with two leaflets, representing excessive fusion of the cushions, are currently excluded.

presenting conotruncal anomalies. It is paradoxical in this regard that common arterial trunk, formed because of failure of fusion of the outflow cushions, is currently included within the conotruncal category, yet arterial valves with two leaflets, representing excessive fusion of the cushions, are currently excluded. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.

Introduction Paradoxical hypertension after repair of coarctation of the aorta is a well-known phenomenon. It can present early postoperatively (<24 hours) or after two to four days, and this delayed response is associated with abdominal pain due to arteritis and possible bowel necrosis.1 And, even years after surgery, late hypertension can develop at long-term follow-up, which is also poorly understood but is possibly due to increased arterial rigidity.2 This late hypertension is beyond the scope of this review. Stage 1 hypertension in children and adolescents is defined as systolic blood pressure (SBP) and/or diastolic blood pressure (DBP) between the 95th percentile and 5 mm Hg above the 99th percentile.3 Stage 2 hypertension, which many postoperative coarctation patients have, is defined as SBP and/or DBP >99th percentile plus 5 mm Hg.4

tension in children and adolescents is defined as systolic blood pressure (SBP) and/or diastolic blood pressure (DBP) between the 95th percentile and 5 mm Hg above the 99th percentile.3 Stage 2 hypertension, which many postoperative coarctation patients have, is defined as SBP and/or DBP >99th percentile plus 5 mm Hg.4 The phenomenon of paradoxical hypertension is complex and has been the focus of intense study. There are several mechanisms involved in developing postoperative hypertension but its exact cause is not yet fully understood.1 First, there can be a higher baroreceptor set point due to preoperative high blood pressure as adaptation to the need for sufficient renal perfusion, which can explain the immediate postoperative hypertensive response. Second, the stretch of the baroreceptors will reduce after surgery, causing elevated sympathetic nervous activity as demonstrated by higher epinephrine/norepinephrine levels after surgery compared to operations of similar magnitude. This can be a factor in initiating the delayed response. A third mechanism is activation of the renin–angiotensin–aldosterone system (RAAS) with elevated plasma renin activity (PRA) in the first week post-coarctectomy compared to patients after other cardiovascular operations.1 A comparative study of balloon angioplasty or surgical repair of aortic coarctation also supports that paradoxical hypertension after coarctectomy is caused by sympathetic activation and RAAS activation.5 The authors found a significant increase in SBP, DBP, and heart rate with increased PRA, norepinephrine, and epinephrine levels directly postoperatively in the surgical group. Conversely, in the balloon angioplasty group, a reduction in SBP and DBP was identified, without increases in catecholamine levels.5

ctivation.5 The authors found a significant increase in SBP, DBP, and heart rate with increased PRA, norepinephrine, and epinephrine levels directly postoperatively in the surgical group. Conversely, in the balloon angioplasty group, a reduction in SBP and DBP was identified, without increases in catecholamine levels.5 Several effective treatment strategies to lower blood pressure post-coarctectomy, targeted at the three above mechanisms, have been published in the literature, but the evidence is still limited. Paradoxical hypertension is known to respond to beta-blockers, arterial smooth muscle relaxants, calcium channel blockers (CCBs), and angiotensin-converting enzyme inhibitors (ACEIs).6-19 And new strategies are under development.20 It is not known which antihypertensive strategy is the most effective, as no randomized controlled trials have been published comparing different strategies in the direct postoperative phase. The aim of this study was to describe current international practice variation surrounding pharmacological management of paradoxical hypertension following repair of coarctation of the aorta in children. This was done as part of a larger online survey regarding prevention and treatment of low cardiac output syndrome (LCOS) we performed among Pediatric Cardiac Intensive Care Society (PCICS) members (in writing). The second objective was to perform a detailed and systematic review of the literature regarding pharmacological treatment of paradoxical hypertension following coarctectomy.

ion and treatment of low cardiac output syndrome (LCOS) we performed among Pediatric Cardiac Intensive Care Society (PCICS) members (in writing). The second objective was to perform a detailed and systematic review of the literature regarding pharmacological treatment of paradoxical hypertension following coarctectomy. Methods We performed a literature search in PubMed using words “aortic,” “aorta,” “coarctation,” “hypertension,” “coarctectomy,” “treatment,” and “antihypertensive agents.” We included all pediatric reports regardless of the year of publication. In May 2014, we developed an online survey with questions regarding the prevention and treatment of LCOS (in writing]). As part of this larger questionnaire, we asked which antihypertensive medication was being used for the treatment of paradoxical hypertension directly following coarctectomy and which antihypertensive agents were being used in the chronic phase. We sent an invitation to all 197 registered PCICS members to partake in the survey. Respondents were reminded of the survey after two and four weeks.

In May 2014, we developed an online survey with questions regarding the prevention and treatment of LCOS (in writing]). As part of this larger questionnaire, we asked which antihypertensive medication was being used for the treatment of paradoxical hypertension directly following coarctectomy and which antihypertensive agents were being used in the chronic phase. We sent an invitation to all 197 registered PCICS members to partake in the survey. Respondents were reminded of the survey after two and four weeks. Results Ninety-eight (50%) people responded from 62 different medical centers across the world (center distribution: 61% United States, 15% Europe, 8% Canada, 6% Asia, 3% Australia, 3% South America, and 2% Africa; see Table 1). Fifty-six (73%) respondents were pediatric intensive care unit (PICU) consultants, 27 (35%) were pediatric cardiologists, and 23 (23%) were nurse practitioners. Several medical respondents were dually certified (mostly intensive care and cardiology). Most respondents (65%) worked in centers that perform more than 300 pediatric heart surgeries per year. Table 1. Distribution of Individual Respondents and the Different Centers.

Results Ninety-eight (50%) people responded from 62 different medical centers across the world (center distribution: 61% United States, 15% Europe, 8% Canada, 6% Asia, 3% Australia, 3% South America, and 2% Africa; see Table 1). Fifty-six (73%) respondents were pediatric intensive care unit (PICU) consultants, 27 (35%) were pediatric cardiologists, and 23 (23%) were nurse practitioners. Several medical respondents were dually certified (mostly intensive care and cardiology). Most respondents (65%) worked in centers that perform more than 300 pediatric heart surgeries per year. Table 1. Distribution of Individual Respondents and the Different Centers. Respondents, n (%) Centers, n (%) Respondents Per Center United States 65 (66) 38 (61) 1.7 (1-6) Europe (France, Germany, Italy, Netherlands, Poland, Sweden, the United Kingdom) 10 (10) 9 (15) 1.1 (1-2) Canada 8 (8) 5 (8) 1.6 (1-4) Australia 5 (5) 3 (3) 1.7 (1-2) Asia (India, Israel, United Arab Emirates) 4 (4) 4 (6) 1.0 South America (Argentina, Colombia) 5 (5) 2 (3) 2.5 (2-3) South Africa 1 (1) 1 (2) 1.0 Eighty-eight respondents (45%) answered the questions regarding blood pressure control following coarctectomy. Due to the setup of the survey, it is not possible to determine the summary statistics regarding center distribution, nationality, and occupation of this subgroup. Nitroprusside is the first drug of choice for initial blood pressure control in 66% (58/88) of respondents, esmolol in 11% (10/88), labetalol in 11% (10/88), ACEIs in 3% (3/88) of respondents (Table 2). Four (4%) respondents reported the use of other drugs, such as nicardipine, urapedil, and phentolamine. Another three (3%) respondents indicated that the choice of drug depended on left ventricle function and/or age of the patient.

l in 11% (10/88), labetalol in 11% (10/88), ACEIs in 3% (3/88) of respondents (Table 2). Four (4%) respondents reported the use of other drugs, such as nicardipine, urapedil, and phentolamine. Another three (3%) respondents indicated that the choice of drug depended on left ventricle function and/or age of the patient. Table 2. Responses to the Question, “What Drug Do You Use First to Lower Blood Pressure in Children With Hypertension Following Coarctectomy?” Response (n) Nitroprusside 58 (66%) Nitroglycerin 0 Labetalol infusion 10 (11%) ACEIs 3 (3%) Other Esmolol 10 (11%) Nicardipine 2 (2%) Urapedil 1 (1%) Phentolamine 1 (1%) Depends on LV function/age/LVH 3 (3%) Abbreviations: ACEIs, angiotensin-converting enzyme inhibitors; LV, left ventricle; LVH, left ventricular hypertrophy. For oral blood pressure control, 75% (65/87) of respondents use ACEIs, 18% (16/87) use labetalol, and 13% (11/87) use beta-blockers (propranolol, carvedilol, atenolol; Table 3). Three (3%) respondents reported the use of other pharmacological agents for oral blood pressure control, such as clonidine, nifedipine, or nitroprusside. Table 3. Responses to the Question, “What Drug Do You Use as Oral Antihypertensive Following Coarctectomy in Children?” Response (n) Labetalol 16 (18%) ACEIs 65 (75%) Beta-blocker 11 (12%) Propranolol 4 (5%) Atenolol 2 (2%) Metoprolol 1 (1%) Carvedilol 1 (1%) Not specified 3 (3%) Other Clonidin 1 (1%) Nitroprusside 1 (1%) Calcium blocker (nifedipine) 1 (1%) Depends on LV function and/or age 3 (3%) Abbreviations: ACEIs, angiotensin-converting enzyme inhibitors; LV, left ventricle.

Response (n) Labetalol 16 (18%) ACEIs 65 (75%) Beta-blocker 11 (12%) Propranolol 4 (5%) Atenolol 2 (2%) Metoprolol 1 (1%) Carvedilol 1 (1%) Not specified 3 (3%) Other Clonidin 1 (1%) Nitroprusside 1 (1%) Calcium blocker (nifedipine) 1 (1%) Depends on LV function and/or age 3 (3%) Abbreviations: ACEIs, angiotensin-converting enzyme inhibitors; LV, left ventricle. We identified 14 articles, published between 1978 and 2016, discussing a pharmacological treatment for direct postoperative hypertension following repair of coarctation of the aorta in children (Table 4). The effect of esmolol on postoperative hypertension has been described in two reports from one prospective randomized trial,16,17 one case series,18 and one prospective study. Rouine-Rapp et al performed a placebo-controlled, prospective randomized trial on the effect of enalaprilat on postoperative hypertension.14 Labetalol has been described in one very small case series.6 Clonidine therapy in postoperative hypertension is investigated in one prospective trial.23 The role of prophylactic propranolol has been described in two prospective trials.9,21 And to our knowledge, there is only one case report about the effect of captopril postoperatively.7 No studies have been published comparing different pharmacological agents. Table 4. A Summary of All Published Literature Regarding Drug Therapy for Paradoxical Hypertension in Children.

We identified 14 articles, published between 1978 and 2016, discussing a pharmacological treatment for direct postoperative hypertension following repair of coarctation of the aorta in children (Table 4). The effect of esmolol on postoperative hypertension has been described in two reports from one prospective randomized trial,16,17 one case series,18 and one prospective study. Rouine-Rapp et al performed a placebo-controlled, prospective randomized trial on the effect of enalaprilat on postoperative hypertension.14 Labetalol has been described in one very small case series.6 Clonidine therapy in postoperative hypertension is investigated in one prospective trial.23 The role of prophylactic propranolol has been described in two prospective trials.9,21 And to our knowledge, there is only one case report about the effect of captopril postoperatively.7 No studies have been published comparing different pharmacological agents. Table 4. A Summary of All Published Literature Regarding Drug Therapy for Paradoxical Hypertension in Children. Author Design Participants Treatment Indication Results Authors’ Conclusion(s) Tabbutt et al, 200816 Prospective, randomized, multicenter, double-blind, dose-ranging esmolol safety, and efficacy trial n = 118 children, <6 years, >2.5 kg, post-coarctectomy; median age: 4.8 months SBP: neonate >80 mm Hg, infant >85 mm Hg, child >95 mm Hg 59% received esmolol and SNP (as esmolol alone was not enough to normalize BP); combination of esmolol and SNP was more common in older children. Older patients were more likely to receive SNP as well as esmolol and were significantly more likely to be discharged from hospital with antihypertensive medication (ACEIs, beta-blockers). Tabbutt et al, 200817 Prospective, randomized, multicenter, double-blind, dose-ranging, randomized trial comparing three different doses of esmolol (125/250/500 mcg/kg) n = 116 children, <6 years, >2.5 kg, post-coarctectomy; mean age: 17.7 months (the same group as in reference16). SBP; neonate >80 mm Hg, infant >85 mm Hg, child >95 mm Hg. Esmolol caused a significant decrease in SBP in all dose groups, but 54% of participants met criteria for rescue medication after 5 minutes of blinded esmolol, and 34% of subjects received rescue medication. Esmolol safely and significantly decreases SBP within 5 minutes after administration postoperatively. Vincent et al, 199018 Descriptive case series. Effectivity of esmolol as an adjunct in the treatment of hypertension after COA repair n = 7 children, median age: 13 years (7-19 years), repair of COA, postoperative hypertension uncontrolled with SNP Treatment goals: MAP <95 mm Hg, SBP <140 mm Hg. Esmolol bolus + continuous infusion was added to SNP when treatment goals were not reached. Adding esmolol caused a significant decrease in HR, SBP, DBP, and MAP in 6/7 pts. SNP could be stopped after 18 hours (10-36) and esmolol after 31 hours (18-56). Esmolol is effective as an adjunct to sodium nitroprusside in the treatment of hypertension after coarctation repair. Wiest et al, 199819 Prospective efficacy trial: effect of esmolol for hypertension after cardiac operations n = 10 following COA repair.

topped after 18 hours (10-36) and esmolol after 31 hours (18-56). Esmolol is effective as an adjunct to sodium nitroprusside in the treatment of hypertension after coarctation repair. Wiest et al, 199819 Prospective efficacy trial: effect of esmolol for hypertension after cardiac operations n = 10 following COA repair. Median age 25.6 months BP > 95th percentile Esmolol significantly reduced systolic BP (mean dose 830 ±153 μg/kg/min) and HR (± 20%). Esmolol is safe and effective. Gidding et al, 19859 RCT: prophylactic propanol, dose: 1.5 mg/kg/d, 2 weeks before surgery, followed by 0.05 mg/kg/d postoperatively n = 14 children with COA n = 7 in the treatment group, mean age: 8.8 ± 1.7 years, n = 7 in the control group, mean age: 8.3 ± 1.0 years Additional postoperative antihypertensive therapy was given when postoperative ABP >160/105 mm Hg Treatment with propranolol reduced the rise in postoperative BP and PRA Treatment group: no patient needed extra antihypertensive therapy compared to 4/7 in the control group. Propranolol reduced the postoperative increase in renin activity but had no effect on increase in NE levels Prophylactic propranolol is effective in the prevention of paradoxical hypertension after repair of coarctation of the aorta and should therefore become a routine part of the operative care of patients with COA. Leenen et al, 198721 RCT: prophylactic propranolol 2.5 mg/kg/d, 2 days preoperatively, 0.025 mg/kg IV bolus at the end of surgery, and continuous infusion max 0.1 mg/kg/h n = 23 children 4-16 years, post-coarctectomy; propranolol group: n = 11, mean age: 9.4 years; control group: n = 12, mean age: 10.2 years SNP infusion would be added (0.5- 6 μg/kg/min) when SBP >160 mm Hg Propranolol group required no extra antihypertensive medication (SNP). In the placebo group, 50% (6/12) required rescue SNP infusion. In the placebo group, there was a higher postoperative increase in SBP ++, NorEpi ++, and PRA ++. Increased sympathetic activity could explain all the hemodynamic changes; prophylactic propranolol inhibited the hyperdynamic circulation after coarctectomy. Will et al, 197822 Descriptive case series: effect of SNP and oral propranolol n = 6 (5-27 years), following coarctation repair SBP >160 mm Hg or DBP >100 mm Hg In 5 of 6 participants, early control of hypertension was achieved with SNP <8 μg/kg/min. In 4 of 6 participants, long-term control with propranolol was given SNP and propranolol are effective for control of hypertension after coarctectomy.

= 6 (5-27 years), following coarctation repair SBP >160 mm Hg or DBP >100 mm Hg In 5 of 6 participants, early control of hypertension was achieved with SNP <8 μg/kg/min. In 4 of 6 participants, long-term control with propranolol was given SNP and propranolol are effective for control of hypertension after coarctectomy. Jones, 197910 Descriptive case series: intraoperative halothane and labetalol n = 9 with COA; mean age: 7.4 years All children received 1 mg/kg bolus of labetalol during coarctectomy, aiming to achieve intraoperative hypotension. MAP decreased by 30% by a combination of 1% halothane and 1 mg/kg of labetalol. The usefulness of labetalol in the postoperative period is suggested. Bojar et al, 19886 Descriptive case series: IV labetalol n = 2 adolescent males with COA Labetalol was added for postoperative hypertension unresponsive to SNP, NTG, and trimetaphan. In both participants, SNP could be weaned after 1-2 hours. Both continued on oral labetalol and ACEIs Labetalol appears effective in the control of paradoxical hypertension following coarctation repair. Casta et al, 19827 Case report describing first use of captopril 15-year-old male after COA repair Hypertension 220/100 mm Hg SNP/propranolol/ methyldopa were inadequate. However, additional captopril orally 0.1 mg/kg was successful in lowering BP. Captopril may be useful for the management of hypertension after repair of COA. Rouine-Rapp et al, 200314 Prospective, randomized, double blind. Effect of enalaprilat on postoperative hypertension after coarctectomy n = 12 pts, undergoing coarctectomy. n = 6 enalaprilat group. Age: 6.5 years; n = 6 placebo group. Age: 4.7 years Enalaprilat or placebo started intraoperatively. Rescue medication if MAP >p95 *Postoperative SBP and DBP were significantly lower in the enalaprilat group, compared to the placebo group. *Two patients in each group required rescue medication with SNP. Enalaprilat decreased BP and patients trended toward decrease in total hours of SNP, decrease in time to extubation and length of stay in the PICU.

Postoperative SBP and DBP were significantly lower in the enalaprilat group, compared to the placebo group. *Two patients in each group required rescue medication with SNP. Enalaprilat decreased BP and patients trended toward decrease in total hours of SNP, decrease in time to extubation and length of stay in the PICU. Farell et al, 19798 Descriptive case series: effect of ACEI (saralasin) n = 12 (5-17 years) with COA Paradoxical hypertension (>normal BP for age) 9 of 12 participants had a significant decrease in BP ACEI is effective in decreasing BP post-coarctectomy Mastropietro et al, 201611 Retrospective study: effect of nicardipine for postoperative hypertension n = 11 (8-94 months); 4/11: single ventricle anatomy; 7/11: COA All children received SNP and/or esmolol, and received nicardipine at day 0-6, postoperatively Significant decrease of mean SBP, DBP, and SNP dose within 6 hours after additional nicardipine. Significant decrease in arterial lactate within 24 hours Nicardipine is effective in controlling hypertension and weaning SNP and esmolol infusions. Sahu et al, 201520 Case report 2 infants following coarctectomy Dexmedetomidine was given to control blood pressure unresponsive to SNP, NTG, and metoprolol Target blood pressures were achieved after initiation of dexmedetomidine Dexmedetomidine may be a useful drug to be used as an adjunct in the management of hypertension. Schreiber et al, 198623 Descriptive case series: effect of clonidine on paradoxical hypertension 66 pediatric patients after coarctectomy Hypertension >p95; 1st 24 hours: IV clonidine infusion. Maintenance with oral clonidine. Decrease in BP in the first 2 days. In 91% of participants, therapy could be stopped in the first two postoperative weeks. In 9% of participants, paradoxical hypertension persisted and clonidine therapy had to be continued after discharge. The use of clonidine suggests a promising therapeutical approach to “paradoxical hypertension.” Abbreviations: ABP, Arterial blood pressure; ACEIs, angiotensin-converting enzyme inhibitors; BP, blood pressure; COA, coarctation of the aorta; DBP, diastolic blood pressure; E, epinephrine; HR, heart rate; IV, intravenous; MAP, mean arterial pressure; NE, norepinephrine; NTG, nitroglycerin trinitrate; PICU, pediatric intensive care unit; PO, per orum; PRA, plasma renin activity; pts, patient(s); RCT, randomized controlled trial; SBP, systolic blood pressure; SNP, sodium nitroprusside.

ure; E, epinephrine; HR, heart rate; IV, intravenous; MAP, mean arterial pressure; NE, norepinephrine; NTG, nitroglycerin trinitrate; PICU, pediatric intensive care unit; PO, per orum; PRA, plasma renin activity; pts, patient(s); RCT, randomized controlled trial; SBP, systolic blood pressure; SNP, sodium nitroprusside. Discussion Paradoxical hypertension following coarctectomy occurs in the first few days after surgery, is due to several mechanisms, and is known to respond to beta-blockers, arterial smooth muscle relaxants, and ACEIs.1 The initial phase of direct postoperative hypertension is related to sympathetic activation and usually limited to 24 hours.24 Hypertension in the initial phase can result in bleeding and/or hemodynamic compromise. The second phase occurs after two to four days and involves the renin–angiotensin system. Systolic and diastolic hypertension in the second phase can cause mesenteric arteritis, a syndrome characterized by abdominal pain, tenderness, ileus, nausea, vomiting, and leukocytosis. If untreated, this could result in peritonitis and possibly prove fatal.6 The medications used to treat paradoxical hypertension target these different mechanisms to prevent the development of hypertensive complications, such as bleeding and/or mesenteric arteritis. At the moment, it is not known whether any treatment strategy is superior as the literature is limited to pediatric hypertension in general, with no specific advice for post-coarctectomy patients.25

ferent mechanisms to prevent the development of hypertensive complications, such as bleeding and/or mesenteric arteritis. At the moment, it is not known whether any treatment strategy is superior as the literature is limited to pediatric hypertension in general, with no specific advice for post-coarctectomy patients.25 Our international survey shows that significant practice variability exists, which is also found in a very recent review of the American Pediatric Health Information System (PHIS) database.26 The authors concluded that the variability in care and the use of newer drugs might be associated with greater costs and suggested that best practice and evidence-based guidelines are warranted for this population. In the following section, we discuss the results of our survey in relation to the evidence found in our systematic literature search.

Our international survey shows that significant practice variability exists, which is also found in a very recent review of the American Pediatric Health Information System (PHIS) database.26 The authors concluded that the variability in care and the use of newer drugs might be associated with greater costs and suggested that best practice and evidence-based guidelines are warranted for this population. In the following section, we discuss the results of our survey in relation to the evidence found in our systematic literature search. Sodium Nitroprusside Our survey shows that most respondents (66%) use sodium nitroprusside (SNP), a direct vasodilator that increases cyclic guanosine monophosphate, resulting in vascular smooth muscle relaxation. Sodium nitroprusside is mainly used in the initial phase of postoperative hypertension caused by increased norepinephrine levels. Sodium nitroprusside is often used to manage hypertensive crises and hypertension after cardiac operations because of its ease of titration, short half-life, and its rapid onset and offset of action.19,27 In the United States, SNP is used in 86% of patients with hypertension following coarctectomy.26 Sodium nitroprusside reduces preload and afterload and can be beneficial in congestive heart failure induced by hypertension.27 However, SNP is an indirect stimulator of the sympathetic nervous system and may further aggravate the increased sympathetic activity frequently present in patients who have had repair of aortic coarctation.24,28 Furthermore, SNP can cause thiocyanate toxicity and may add to tachycardia, decreasing Pao 2 and oxygen saturation. Therefore, a beta-blocker is often added to attenuate the reflex tachycardia.

er aggravate the increased sympathetic activity frequently present in patients who have had repair of aortic coarctation.24,28 Furthermore, SNP can cause thiocyanate toxicity and may add to tachycardia, decreasing Pao 2 and oxygen saturation. Therefore, a beta-blocker is often added to attenuate the reflex tachycardia. The first report on the effective use of SNP following coarctectomy is a descriptive case series of six patients (5-27 years old) described by Will et al in 1978.22 In five of six patients, the authors achieved early control of hypertension with SNP <8 μg/kg/min in combination with propranolol. Unfortunately, no comment was made regarding the heart rate and/or the effect of propranolol on tachycardia, which often may be associated with SNP. In another small case series, SNP alone was ineffective in controlling postoperative hypertension, but by adding esmolol, blood pressure control could be achieved.18 In two more recent studies, SNP was used as an effective rescue medication in 34% to 59% of patients whose blood pressure could not be controlled with esmolol alone.16,17

case series, SNP alone was ineffective in controlling postoperative hypertension, but by adding esmolol, blood pressure control could be achieved.18 In two more recent studies, SNP was used as an effective rescue medication in 34% to 59% of patients whose blood pressure could not be controlled with esmolol alone.16,17 Beta-Blockers The basic mechanisms of paradoxical hypertension are sympathetic activation and renin stimulation,1 both of which can be targeted using beta-blockers.9,21 Esmolol, a selective short-acting beta-blocker, is an often-used alternative to SNP, being used in 59% of patients in the recent PHIS review.26 Esmolol is the most studied in paradoxical hypertension; it has a half-life of 10 minutes and is the first drug of choice in 11% of respondents of our survey.16-19 Vincent et al first described esmolol in 1989 in seven children with a median age of 13 years, with postoperative hypertension unresponsive to SNP. Adding esmolol significantly lowered blood pressure in six of seven patients, and SNP could be stopped after a mean of 18 hours.18 All patients were switched to oral antihypertensive therapy with propranolol and/or ACEIs. In 1998, a prospective esmolol efficacy trial showed that monotherapy with continuous esmolol, mean dose 830 ±153 mcg/kg/min, significantly reduced blood pressure in 10 children following coarctectomy as part of a larger trial of postoperative hypertension.19 Target blood pressure (<90th percentile for age) was reached after 1.7 hours (0.6-3.6). Heart rate also significantly decreased by 20%. The authors concluded that monotherapy esmolol was safe and effective in reducing blood pressure in this small group of postoperative pediatric cardiac patients (n = 10).19 In a more recent prospective, randomized, multicenter, double-blind, dose-ranging esmolol safety and efficacy trial by Tabbutt et al, esmolol was used for blood pressure control in the direct postoperative phase in 118 children under six years of age, undergoing coarctation repair via lateral thoracotomy, but 59% of patients also received nitroprusside infusion (0.5-8 mcg/kg/min).16 And, with increasing age, patients were more likely to receive SNP in addition to esmolol.

lood pressure control in the direct postoperative phase in 118 children under six years of age, undergoing coarctation repair via lateral thoracotomy, but 59% of patients also received nitroprusside infusion (0.5-8 mcg/kg/min).16 And, with increasing age, patients were more likely to receive SNP in addition to esmolol. The median age in this study was lower (4.8 months) than in the study by Vincent et al (13 years).18 The same group published the dose-ranging part of their trial separately comparing three different doses of esmolol (125/250/500 mcg/kg as bolus followed by 125/250/500 mcg/kg/min continuous infusion).17 All three doses decreased SBP quickly and significantly, without any statistical difference between groups, but again 54% met criteria for rescue medication because SBP had not decreased below the levels of treatment indication five minutes after administrating the esmolol bolus. Thirty-four percent of participants actually received rescue medication with isoflurane, SNP, fentanyl, or propofol. The authors concluded that esmolol therapy is safe and effective.17 However, in both studies by Tabutt et al, more than 50% of patients received additional therapy to esmolol, suggesting that monotherapy with esmolol is often not adequate. Esmolol is also used in other, noncoarctectomy-related, pediatric hypertensive crises, particularly when there is associated tachycardia.27 It is advocated that esmolol should not be used when a hypertensive crisis is caused by catecholamine excess, as hypertension is perpetuated by persistent alpha stimulation.19

ate. Esmolol is also used in other, noncoarctectomy-related, pediatric hypertensive crises, particularly when there is associated tachycardia.27 It is advocated that esmolol should not be used when a hypertensive crisis is caused by catecholamine excess, as hypertension is perpetuated by persistent alpha stimulation.19 In this survey, the use of beta-blockers as oral therapy in the postoperative phase was reported by 12%. Specific beta-blockers reported were propranolol, atenolol, metoprolol, or carvedilol. The only study describing the use of oral beta-blockers in the specific postoperative coarctectomy patients is the study by Will et al who describe the use of SNP in combination with oral propranolol in four of six patients with sustained hypertension following initial satisfactory control of blood pressure with SNP.22 In adults who developed late hypertension 26 (±15) years after coarctectomy, metoprolol was found to have a more antihypertensive effect than candesartan, an angiotensin II receptor blocker.29 In the United States, metoprolol is the only beta-blocker approved by the US Food and Drug Administration (FDA) for pediatric hypertension based on one trial in essential pediatric hypertension (not post-coarctectomy).25

ound to have a more antihypertensive effect than candesartan, an angiotensin II receptor blocker.29 In the United States, metoprolol is the only beta-blocker approved by the US Food and Drug Administration (FDA) for pediatric hypertension based on one trial in essential pediatric hypertension (not post-coarctectomy).25 In 1985, Gidding et al advocated that prophylactic propranolol should become a routine part of the operative care of patients with coarctation of the aorta.9 The authors had shown that two weeks of prophylactic propranolol and one week of postoperative propranolol significantly reduced the rise in postoperative blood pressure in seven children following repair of coarctation of the aorta (mean age 8.8 year ± 1.7). Leenen et al confirmed in a randomized, placebo-controlled, double-blind trial that prophylactic propranolol inhibited the hyperdynamic circulation after coarctectomy in 11 children (mean age 9.4 years).21 The authors’ strategy entailed two days of prophylactic propranolol and six days of postoperative propranolol. None of the patients received rescue therapy with SNP, compared to 6 of 12 patients in the placebo group. In our survey, no questions were asked about prophylactic strategies, as our scope was postoperative care; therefore, information on the current use of prophylactic strategies is not available. We would hypothesize, however, that the prophylactic strategy has not found its way to the current routine approach to “coarctectomy care,” possibly because of the quickly effective postoperative treatments with SNP, esmolol, and labetalol. Moreover, since patients are operated at younger ages, postoperative hypertension is possibly not as therapy resistant as it used to be in the 1970s and 1980s.

way to the current routine approach to “coarctectomy care,” possibly because of the quickly effective postoperative treatments with SNP, esmolol, and labetalol. Moreover, since patients are operated at younger ages, postoperative hypertension is possibly not as therapy resistant as it used to be in the 1970s and 1980s. Labetalol In this survey, IV labetalol, a combined alpha-1 and nonselective beta-blocker, is used by 11% of respondents for the direct treatment of paradoxical hypertension following repair of coarctation of the aorta in children, compared to 20% in the recent PHIS review.26 The IV labetalol has direct vasodilator properties, is easily titratable, and prevents reflex tachycardia. The beta to alpha blocking potency ratio is 3:1 for oral labetalol and 7:1 for IV labetalol,13 but the majority of side effects can be attributed to alpha blockade (headache, dizziness). Labetalol has a relatively long half-life of three to five hours and is often used in children with hypertensive crises but should not be used in patients with bronchospastic disease or congestive heart failure as it has a negative inotropic effect.27

e effects can be attributed to alpha blockade (headache, dizziness). Labetalol has a relatively long half-life of three to five hours and is often used in children with hypertensive crises but should not be used in patients with bronchospastic disease or congestive heart failure as it has a negative inotropic effect.27 In patients with coarctation of the aorta, labetalol was first described, in nine children, in combination with 1% halothane to control blood pressure intraoperatively.10 In 1988, Bojar et al described two adolescent males with coarctation of the aorta, with postoperative hypertension unresponsive to SNP.6 Both patients responded quickly to labetalol and SNP could be weaned off within two hours of starting labetalol. The IV labetalol could successfully be converted to oral labetalol. The authors concluded that labetalol addresses the basic mechanism of elevated sympathetic activity responsible for the postoperative hypertension and that it is an easy, effective, and safe medication for the control of paradoxical hypertension following repair of coarctation of the aorta.6 There is no literature comparing labetalol with other beta-blockers. In the recent PHIS review, esmolol is used more frequently (60%) compared to labetalol (20%).26 This is possibly due to the longer half-life of labetalol, compared to esmolol. Both agents are used by 11% of respondents of our survey.

on of the aorta.6 There is no literature comparing labetalol with other beta-blockers. In the recent PHIS review, esmolol is used more frequently (60%) compared to labetalol (20%).26 This is possibly due to the longer half-life of labetalol, compared to esmolol. Both agents are used by 11% of respondents of our survey. Oral labetalol can be used in the long-term treatment of primary and secondary hypertension (eg, renal disease, autoimmune disease, essential hypertension), with few side effects in adults and children.15 Labetalol has no pediatric FDA labeling. Oral labetalol for the control of hypertension in the second postoperative phase was reported by 18% of respondents in this survey. Angiotensin-Converting Enzyme Inhibitors Renin stimulation is mostly responsible for the second phase of paradoxical hypertension1; however, the use of ACEIs in the management of the initial phase of postoperative hypertension is reported by 3% of respondents of our survey. The survey did not ask as to which ACEIs are being used. Angiotensin-converting enzyme inhibitors block the angiotensin-converting enzyme required for the conversion of angiotensin I to angiotensin II, a potent vasoconstrictor. Angiotensin-converting enzyme inhibitors also block angiotensin II-mediated aldosterone release, thereby preventing salt and water retention.3 Blood pressure response to ACEIs is affected by ethnicity, with black children demonstrating lower antihypertensive effect compared to white children.30

potent vasoconstrictor. Angiotensin-converting enzyme inhibitors also block angiotensin II-mediated aldosterone release, thereby preventing salt and water retention.3 Blood pressure response to ACEIs is affected by ethnicity, with black children demonstrating lower antihypertensive effect compared to white children.30 The effective use of an IV infusion of saralasin, a partial agonist of angiotensin II receptors, was described once (in 1979) in 12 children with immediate hypertension following coarctation repair.7 No other reports regarding saralasin post-coarctectomy have been published. Currently, enalaprilat is the only available IV ACEI that can be effective in renin-mediated hypertension in the immediate postoperative phase with only limited data available on its use in children with hypertensive crisis.27 Rouine-Rapp et al reported 14 children between 1 year and 18 years of age, receiving enalaprilat or placebo infusion after coarctectomy.14 The enalaprilat group trended toward a decrease in the total hours of SNP and demonstrated improved control of SBP and DBP.

ilable on its use in children with hypertensive crisis.27 Rouine-Rapp et al reported 14 children between 1 year and 18 years of age, receiving enalaprilat or placebo infusion after coarctectomy.14 The enalaprilat group trended toward a decrease in the total hours of SNP and demonstrated improved control of SBP and DBP. Oral captopril has been used in the first phase of paradoxical hypertension to lower blood pressure in children unresponsive to SNP, propranolol, and methyldopa.7 But the renin–angiotensin system is attributed mostly in the second phase of postoperative hypertension, which is reflected in the fact that 75% of respondents of our survey use ACEIs as oral therapy. Captopril is one of the earliest ACEIs, with substantial evidence for its efficacy in children with renal or vascular secondary hypertension, although its usage is off-label.3,12 Enalapril and lisinopril do have pediatric FDA approval. In pediatric populations, ACEIs are the most commonly prescribed antihypertensive for both primary and secondary hypertension, especially in children with chronic kidney disease and diabetic nephropathy.3,31 Alpha-Agonist The effective use of clonidine, a centrally acting alpha-agonist, has been described in one case series of 66 pediatric patients following repair of coarctation of the aorta.23 Clonidine decreases sympathetic nervous system outflow, leading to peripheral vasodilatation and decrease in heart rate, blood pressure, and cardiac output.32

se of clonidine, a centrally acting alpha-agonist, has been described in one case series of 66 pediatric patients following repair of coarctation of the aorta.23 Clonidine decreases sympathetic nervous system outflow, leading to peripheral vasodilatation and decrease in heart rate, blood pressure, and cardiac output.32 None of the respondents of our survey reported the use of clonidine in the acute phase. One respondent (1/88 = 1%) reported the use of clonidine in the chronic treatment of paradoxical hypertension. Phentolamine, phenoxybenzamine, prazosin, and doxazocin are other alpha-blockers that are often used in catecholamine-induced hypertension, as in pheochromocytoma.27 In this survey, the use of phentolamine in the acute phase was reported by one respondent (1.1%). Dexmedetomidine, a selective alpha-2 receptor agonist with sedative and analgesic properties, is mostly used for sedation in the postoperative situation in the PICU. Common side effects are hypotension and bradycardia, suggesting a possible beneficial effect post-coarctectomy. Dexmedetomidine was shown to be effective in two infants as adjunct therapy in the treatment of paradoxical hypertension after coarctectomy.20 Dexmedetomidine can safely be used in the extubated patient and is helpful in keeping the patient out of pain, less agitated, and calm, thereby preventing the rise in blood pressure and heart rate.20

own to be effective in two infants as adjunct therapy in the treatment of paradoxical hypertension after coarctectomy.20 Dexmedetomidine can safely be used in the extubated patient and is helpful in keeping the patient out of pain, less agitated, and calm, thereby preventing the rise in blood pressure and heart rate.20 The CCBs The CCBs achieve vasodilatation by blocking calcium entry into the vascular smooth muscle cell. Amlodipine, isradipine, nicardipine, and nifedipine are common oral CCBs used in primary and secondary pediatric hypertension.3,11 The CCBs can safely be used in children with renal failure, as it is metabolized by the liver and does not have renal effects seen with ACEIs.3 Nicardipine was used in 14% of patients, according to the recent PHIS review.26 Regarding pediatric patients with coarctation of the aorta, Mastropietro et al recently demonstrated a significant decrease in mean SBP and DBP and a significant decrease in the dosage of SNP in children receiving nicardipine as a substitute for SNP or esmolol following coarctectomy.11 Nicardipine is used as the first-line therapy in hypertensive crises in children, is well tolerated, efficacious, and can be used for a longer period of time without the fear of cyanide toxicity.27 Tachycardia is a side effect of therapy. The use of nicardipine was reported by 2 (2%) of 88 respondents.

coarctectomy.11 Nicardipine is used as the first-line therapy in hypertensive crises in children, is well tolerated, efficacious, and can be used for a longer period of time without the fear of cyanide toxicity.27 Tachycardia is a side effect of therapy. The use of nicardipine was reported by 2 (2%) of 88 respondents. Other Strategies Although furosemide has no pediatric FDA approval for hypertension, diuretics play an important role in the management of hypertension in children. They reduce blood pressure by decreasing sodium and water retention, thereby reducing extracellular volume. Diuretics are usually combined with other antihypertensive medications but have not been researched specific to paradoxical hypertension.3 Hydralazine and minoxidil are FDA-approved vasodilators used in children, but none of respondents of this survey reported their use. In older literature, reserpine and trimethaphan are often mentioned, but they are not used anymore. Intramuscular reserpine blocks sympathetic activity but has a slow onset of action and could not easily be titrated. Trimethaphan, a ganglionic blocker, can control blood pressure without producing tachycardia, but tolerance to its effects usually occurred after 48 hours. None of the respondents reported the use of these two pharmacological agents.

sympathetic activity but has a slow onset of action and could not easily be titrated. Trimethaphan, a ganglionic blocker, can control blood pressure without producing tachycardia, but tolerance to its effects usually occurred after 48 hours. None of the respondents reported the use of these two pharmacological agents. Limitations This survey was designed to gain insight into the pharmacological strategies that are currently being used to treat hypertension following repair of coarctation of the aorta, as it is not known which could be the most effective strategy. It is also not known which blood pressure to target in the postoperative phase. Almost all the identified studies targeted different blood pressures (Table 3). Many centers target blood pressures below the 95th to 99th percentile for age. Unfortunately, we do not know the blood pressure targets employed as no questions were asked regarding treatment goals. Another limitation is that the survey was set up to determine provider-dependent practice variation rather than center-dependent variation and it was therefore also not possible to compare practice variation between the different continents and/or countries. However, the average number of respondents per center is approximately 1.6, which is consistent for all centers apart from the European and South African centers. Therefore, the practice variation we identified in all respondents also reflects in the different centers and probably indicates that practice variation is more center dependent than provider dependent.

r center is approximately 1.6, which is consistent for all centers apart from the European and South African centers. Therefore, the practice variation we identified in all respondents also reflects in the different centers and probably indicates that practice variation is more center dependent than provider dependent. Conclusion The results of the survey show that there is wide practice variation regarding the treatment of paradoxical hypertension following surgical repair of coarctation of the aorta. This variability in care may be associated with increased costs.26

r center is approximately 1.6, which is consistent for all centers apart from the European and South African centers. Therefore, the practice variation we identified in all respondents also reflects in the different centers and probably indicates that practice variation is more center dependent than provider dependent. Conclusion The results of the survey show that there is wide practice variation regarding the treatment of paradoxical hypertension following surgical repair of coarctation of the aorta. This variability in care may be associated with increased costs.26 The systematic review of the literature shows that there is only limited evidence regarding the treatment of paradoxical hypertension. There is no evidence that one pharmacological strategy is superior to any other and often multiple drugs (eg, esmolol and SNP) are necessary to control blood pressure in the acute postoperative phase. In the recent PHIS review, patients received a median of three antihypertensive medications.26 Sodium nitroprusside is the most widely used (66% of respondents), probably because of its ease of titration; however, tachycardia and cyanide toxicity are important concerns. There is no literature on monotherapy with SNP. Beta-blockers target the two important mechanisms, sympathetic activation and renin release, involved in paradoxical hypertension and seem a logical choice to control blood pressure.1 Esmolol is the most studied; but in up to 59% of patients, monotherapy is not adequate in controlling blood pressure.16,17 Both esmolol and labetalol are being used by 11% of respondents. Thirty years ago, the use of prophylactic beta-blockers has been shown to be effective in controlling postoperative blood pressure.9,21 As the current use of prophylactic beta-blockers was out of scope of this survey, no data were gathered on this topic. However, this would be very interesting to investigate further. Other pharmacological approaches with alpha-agonists and/or CCBs have not found widespread use but deserve attention. Especially, nicardipine shows promising results.11

tic beta-blockers was out of scope of this survey, no data were gathered on this topic. However, this would be very interesting to investigate further. Other pharmacological approaches with alpha-agonists and/or CCBs have not found widespread use but deserve attention. Especially, nicardipine shows promising results.11 In the second phase of paradoxical hypertension, most respondents use ACEIs to control sustained high blood pressure, targeting the renin–angiotensin system. In conclusion, based on the literature it is not possible to advise one pharmacological agent over another and further research is warranted. It would be very interesting to start by comparing SNP with a beta-blocker (esmolol or labetalol) in a prospective randomized controlled trial with duration of antihypertensive therapy and PICU length of stay as primary end points. Secondary end points of such a study should include the development of side effects (eg, bradycardia, cyanide toxicity, hypotension), the development of complications from hypertension (eg, postoperative bleeding, rupture of anastomosis, and the development of mesenteric arteritis), and costs of treatment. Target blood pressures should be based on age-related normograms. Authors’ Note: Both authors had full control of the design of the study, methods used, outcome parameters, analysis of data, and production of the written report Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ Note: Both authors had full control of the design of the study, methods used, outcome parameters, analysis of data, and production of the written report Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. Abbreviations and Acronyms ACEIangiotensin-converting enzyme inhibitor CCBcalcium channel blocker DBPdiastolic blood pressure FDAUS Food and Drug Administration LCOSlow cardiac output syndrome PCICSPediatric Cardiac Intensive Care Society PICUpediatric intensive care unit PHISPediatric Health Information System PRAplasma renin activity RAASrenin–angiotensin–aldosterone system SBPsystolic blood pressure SNPsodium nitroprusside

Introduction Inotropes are frequently being used in children undergoing heart surgery to prevent or treat low cardiac output syndrome (LCOS) that typically occurs 6 to 18 hours after cardiopulmonary bypass (CPB) surgery.1 There is ample evidence describing how certain inotropes affect the hemodynamics in adults and children undergoing heart surgery. But there is only limited evidence that certain inotropes actually positively influence postoperative outcome.2 A recent Cochrane review concluded there is insufficient evidence of effectiveness of prophylactic milrinone in preventing death or LCOS in children undergoing surgery for congenital heart disease, compared to placebo.3 There is no evidence in the literature in favor of one inotrope over another in improving postoperative outcomes. The aim of this study was to describe current international practice variation regarding the use of inotropes in children undergoing heart surgery. Methods Survey In May 2014, we developed an online survey for the members of the Pediatric Cardiac Intensive Care Society (PCICS), with questions regarding the prevention and treatment of LCOS in pediatric patients after cardiac surgery.

Introduction Inotropes are frequently being used in children undergoing heart surgery to prevent or treat low cardiac output syndrome (LCOS) that typically occurs 6 to 18 hours after cardiopulmonary bypass (CPB) surgery.1 There is ample evidence describing how certain inotropes affect the hemodynamics in adults and children undergoing heart surgery. But there is only limited evidence that certain inotropes actually positively influence postoperative outcome.2 A recent Cochrane review concluded there is insufficient evidence of effectiveness of prophylactic milrinone in preventing death or LCOS in children undergoing surgery for congenital heart disease, compared to placebo.3 There is no evidence in the literature in favor of one inotrope over another in improving postoperative outcomes. The aim of this study was to describe current international practice variation regarding the use of inotropes in children undergoing heart surgery. Methods Survey In May 2014, we developed an online survey for the members of the Pediatric Cardiac Intensive Care Society (PCICS), with questions regarding the prevention and treatment of LCOS in pediatric patients after cardiac surgery. The questionnaire consisted of 22 questions covering different aspects of postoperative management in children who underwent cardiac surgery. Most of the questions concerned the treatment and prevention of LCOS in children while other questions where about postoperative temperature control. The management of hypertension in patients who underwent surgery for coarctation of the aortae was part of the same survey and published separately.4 Most questions could be answered with multiple answers. Respondents were free not to answer a question if they chose not to do so, therefore not all questions were answered by all 98 respondents.

nsion in patients who underwent surgery for coarctation of the aortae was part of the same survey and published separately.4 Most questions could be answered with multiple answers. Respondents were free not to answer a question if they chose not to do so, therefore not all questions were answered by all 98 respondents. Survey Participants We sent an invitation to partake in the survey to all 197 registered PCICS members at the time. In the accompanying letter it was mentioned that all responses would be analyzed anonymously and that it was the intention to publish the results. A reminder was sent after two and four weeks if they had not yet responded. Data were collected between May and June 2014. Data Analysis The results are descriptive and were expressed as percentages. No comparative statistical tests were performed. Literature Review The authors searched PubMed for all known inotropes and cross-referenced them with key words such as “pediatric,” “child,” “neonate,” “newborn,” “cardiac,” “heart,” “congenital,” and “surgery.” The authors also scanned all references in the found articles and then read all titles and abstracts of relevant studies. Then randomized controlled trials were selected in which inotropes were compared with each other or with placebo in pediatric or congenital cardiac surgery, or animal models of congenital heart surgery.

authors also scanned all references in the found articles and then read all titles and abstracts of relevant studies. Then randomized controlled trials were selected in which inotropes were compared with each other or with placebo in pediatric or congenital cardiac surgery, or animal models of congenital heart surgery. Results Survey Participants Ninety-eight people (50%) responded from 62 different medical centers across the world (center distribution: 61% United States, 15% Europe, 8% Canada, 6% Asia, 3% Australia, 3% South America, and 2% Africa; see Table 1). Fifty-six (73%) respondents were pediatric intensive care unit (PICU) consultants, 27 (35%) pediatric cardiologists, 23 (23%) nurse practitioners, 3 (3%) cardiothoracic surgeons (3%), 5 (5%) anesthesiologists, and 4 (4%) PICU trainees. Several medical respondents were dually certified (PICU and cardiology). Most respondents (63%) worked in centers with more than 300 pediatric heart surgeries per year. Table 1. Demographics of the Respondents: Origin, Specialization, and Number of Operations Performed at Their Center per Annum.a,b,c Individual distribution N % United States 65 66 Europe 10 10 Canada 8 8 Australia 5 5 Asia 4 4 South America 5 5 South Africa 1 1 Total 98 100 Center distribution United States 38 61 Europe 9 15 Canada 5 8 Australia 3 3 Asia 4 6 South America 2 3 South Africa 1 2 Total 62 100 Center size (operations per annum) Less than 100 5 5 100-200 9 9 200-300 22 22 300-400 24 25 400-500 16 16 More than 500 22 22 Abbreviation: PICU: pediatric intensive care unit.

Individual distribution N % United States 65 66 Europe 10 10 Canada 8 8 Australia 5 5 Asia 4 4 South America 5 5 South Africa 1 1 Total 98 100 Center distribution United States 38 61 Europe 9 15 Canada 5 8 Australia 3 3 Asia 4 6 South America 2 3 South Africa 1 2 Total 62 100 Center size (operations per annum) Less than 100 5 5 100-200 9 9 200-300 22 22 300-400 24 25 400-500 16 16 More than 500 22 22 Abbreviation: PICU: pediatric intensive care unit. aEurope: France, Germany, Italy, the Netherlands, Poland, Sweden, United Kingdom. bAsia: India, Israel, United Arab Emirates. cSouth America: Argentina, Colombia. Drug Choice for the Prevention of LCOS Ninety-three (95%) of all 98 respondents answered this question. The prevention of LCOS is characterized by considerable variability (see Table 2). Eight different drugs used for the prevention of LCOS. Most respondents use milrinone (97%) routinely for the prevention of LCOS. Other frequently used drugs for the prevention of LCOS are adrenaline (45%) and dopamine (38%). Dobutamine is used by 11% and levosimendan by 5% of respondents. Table 2. Reports of Drugs Regimen for Prevention of LCOS and Timing of Administration.a Prophylactic drug regimen N % Milrinone 90/93 97 Adrenaline/epinephrine 42/93 45 Dopamine 35/93 38 Dobutamine 10/93 11 Levosimendan 5/93 5 Other 10/93 11 Not answered 5/98 5 Timing of administration Preoperatively 1/93 1 After anesthetic induction 1/93 1 When on CPB 39/93 42 While coming off CPB 59/93 63 In PICU 27/93 29 Other 5/93 5 Not answered 5/98 5 Other drugs used

Prophylactic drug regimen N % Milrinone 90/93 97 Adrenaline/epinephrine 42/93 45 Dopamine 35/93 38 Dobutamine 10/93 11 Levosimendan 5/93 5 Other 10/93 11 Not answered 5/98 5 Timing of administration Preoperatively 1/93 1 After anesthetic induction 1/93 1 When on CPB 39/93 42 While coming off CPB 59/93 63 In PICU 27/93 29 Other 5/93 5 Not answered 5/98 5 Other drugs used Noradrenaline/norepinephrine 38/69 55 Alpha blockers 7/69 10 ACE-inhibitors 10/69 15 Steroids (before CPB) 37/69 54 Steroids continued after CPB 10/69 15 Vasopressin 30/69 43 Adrenaline/epinephrine 3/69 4 Other (triiodothyronin, calcium, epinephrine) 15/69 22 Not answered 29/98 30 Abbreviations: ACE, angiotensin converting enzyme; CPB, cardiopulmonary bypass; LCOS, low cardiac output syndrome; PICU, pediatric intensive care unit. aMultiple answers possible. There is marked variability in the timing of initiating prophylactic inotropes. Respondents could give multiple answers as individual clinical indication might necessitate different timing. Forty-two percent start the inotrope at the start of CPB, 63% indicate starting while coming off CPB, and 29% started inotropes when the patient arrives in the ICU.

Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. ORCID iD: Peter P. Roeleveld, MD http://orcid.org/0000-0001-5491-9408 Abbreviations and Acronyms ACEangiotensin converting enzyme cAMPcyclic adenosine monophosphate CPBcardiopulmonary bypass DHCAdeep hypothermic circulatory arrest ECMOextracorporeal membrane oxygenation iNOinhaled nitric oxide LCOSlow cardiac output syndrome LOSlength of stay MAPmean arterial blood pressure MVmechanical ventilation NIRSnear-infrared spectrometry PAPpulmonary artery pressure PCICSPediatric Cardiac Intensive Care Society PICUpediatric intensive care unit PVRpulmonary vascular resistance SVRsystemic vascular resistance PICCOpulse index continuous cardiac output USCOMDoppler cardiac output measurement

timing of initiating prophylactic inotropes. Respondents could give multiple answers as individual clinical indication might necessitate different timing. Forty-two percent start the inotrope at the start of CPB, 63% indicate starting while coming off CPB, and 29% started inotropes when the patient arrives in the ICU. Sixty-nine respondents (70%) indicated that they routinely use other pharmacological agents besides their main inotrope. Other medications that are frequently used concomitantly are noradrenaline/norepinephrine (55%), steroids (54%), and vasopressin (43%). Thirty percent of respondents indicate that they give noradrenaline/norepinephrine routinely to maintain blood pressure while giving milrinone. Vasodilators such as angiotensin-converting-enzyme (ACE) inhibitors and alpha-agonists are routinely used by 10 (15%) of 69 and 7 (10%) of 69, respectively. Adrenaline/epinephrine is routinely used by 3 (4%) of 69 of respondents. Out of 69 who answered this question, 57 (83%) respondents specifically indicated that the use of these second-tier medications is based on clinical requirements. The strategies used vary from the use of vasopressors in vasodilated states to improve perfusion pressures, to vasodilators in order to promote afterload reduction and cardiac output. When steroids are used, they are mostly given before going on CPB (54%) and in 15% it is routinely continued after the bypass.

Out of 69 who answered this question, 57 (83%) respondents specifically indicated that the use of these second-tier medications is based on clinical requirements. The strategies used vary from the use of vasopressors in vasodilated states to improve perfusion pressures, to vasodilators in order to promote afterload reduction and cardiac output. When steroids are used, they are mostly given before going on CPB (54%) and in 15% it is routinely continued after the bypass. Monitoring Cardiac Output Ninety-one respondents indicated the use of different methods to monitor cardiac output postoperatively (see Table 3). Lactate (99%), physical examination (98%), and intermittent or continuous venous saturation (89%) are mostly used for monitoring cardiac output. Echocardiography is used by 53% of the respondents, core–peripheral temperature gap by 32%, and near-infrared spectrometry (NIRS) is used by 26%. The use of minimally invasive cardiac output monitoring (e.g. pulse index continuous cardiac output [PICCO, Pulsion, Germany] or Doppler hemodynamic cardiac output monitoring [USCOM Ltd, Australia] is limited (3%). Table 3. Monitoring of Development of LCOS. Multiple Answers Possible. NIRS: Near-infrared Spectrometry.

Monitoring Cardiac Output Ninety-one respondents indicated the use of different methods to monitor cardiac output postoperatively (see Table 3). Lactate (99%), physical examination (98%), and intermittent or continuous venous saturation (89%) are mostly used for monitoring cardiac output. Echocardiography is used by 53% of the respondents, core–peripheral temperature gap by 32%, and near-infrared spectrometry (NIRS) is used by 26%. The use of minimally invasive cardiac output monitoring (e.g. pulse index continuous cardiac output [PICCO, Pulsion, Germany] or Doppler hemodynamic cardiac output monitoring [USCOM Ltd, Australia] is limited (3%). Table 3. Monitoring of Development of LCOS. Multiple Answers Possible. NIRS: Near-infrared Spectrometry. Monitoring modality N % Lactate 90/91 99 Physical examination 89/91 98 Intermittent venous saturation 69/91 76 Echocardiography 48/91 53 Core–peripheral temperature gap 26/91 32 NIRS 24/91 26 Continuous venous saturation 12/91 13 PICCO 2/91 2 USCOM 1/91 1 Other 4/91 4 Not answered 7/98 7 Abbreviations: LCOS, low cardiac output syndrome; NIRS, near-infrared spectrometry; PICCO, pulse index continuous cardiac output; USCOM, Doppler cardiac output measurement.

–peripheral temperature gap 26/91 32 NIRS 24/91 26 Continuous venous saturation 12/91 13 PICCO 2/91 2 USCOM 1/91 1 Other 4/91 4 Not answered 7/98 7 Abbreviations: LCOS, low cardiac output syndrome; NIRS, near-infrared spectrometry; PICCO, pulse index continuous cardiac output; USCOM, Doppler cardiac output measurement. Drug Choice for Improving Cardiac Output Most respondents indicate that their choice of drugs for improving cardiac output is based on the pathophysiology of the underlying condition, the severity of the LCOS, and the medication that is already being administered. Eighty-five respondents (85/98 = 87%) answered this question and indicated increasing either milrinone (42%) or adrenaline/epinephrine (36%) when cardiac output is worsening (see Table 4). If this treatment fails, then in most of the cases adrenaline/epinephrine (40%) will be added as the second drug of choice. Under “other” 17% indicated that clinical circumstances directed their specific therapies. Extracorporeal membrane oxygenation (ECMO) was also given as an option several times but was not specified by many other respondents. The timing of initiating ECMO postcardiotomy remains challenging. It was not a part of this survey. Table 4. Reports of Drugs Regimen for Treatment of LCOS.a First Choice Second Choice

Drug Choice for Improving Cardiac Output Most respondents indicate that their choice of drugs for improving cardiac output is based on the pathophysiology of the underlying condition, the severity of the LCOS, and the medication that is already being administered. Eighty-five respondents (85/98 = 87%) answered this question and indicated increasing either milrinone (42%) or adrenaline/epinephrine (36%) when cardiac output is worsening (see Table 4). If this treatment fails, then in most of the cases adrenaline/epinephrine (40%) will be added as the second drug of choice. Under “other” 17% indicated that clinical circumstances directed their specific therapies. Extracorporeal membrane oxygenation (ECMO) was also given as an option several times but was not specified by many other respondents. The timing of initiating ECMO postcardiotomy remains challenging. It was not a part of this survey. Table 4. Reports of Drugs Regimen for Treatment of LCOS.a First Choice Second Choice N % N % Milrinone 36/85 42 13/86 15 Dobutamine 2/85 2 1/86 1 Dopamine 13/85 15 8/86 9 Adrenaline/epinephrine 31/85 36 34/86 40 Noradrenaline/norepinephrine 0/85 0 0/86 0 Levosimendan 1/85 1 3/86 3 Steroids 2/85 2 9/86 10 Vasodilators 0/85 0 4/86 5 Others (eg, ECMO) 17/98 17 14/86 16 Not answered 13/98 13 12/98 12 Abbreviations: ECMO, extracorporeal membrane oxygenation; LCOS, low cardiac output syndrome. aOne answer only.

N % N % Milrinone 36/85 42 13/86 15 Dobutamine 2/85 2 1/86 1 Dopamine 13/85 15 8/86 9 Adrenaline/epinephrine 31/85 36 34/86 40 Noradrenaline/norepinephrine 0/85 0 0/86 0 Levosimendan 1/85 1 3/86 3 Steroids 2/85 2 9/86 10 Vasodilators 0/85 0 4/86 5 Others (eg, ECMO) 17/98 17 14/86 16 Not answered 13/98 13 12/98 12 Abbreviations: ECMO, extracorporeal membrane oxygenation; LCOS, low cardiac output syndrome. aOne answer only. Managing Cardiac Output in a Mixed Circulation Eighty-nine respondents (91%) answered questions regarding specific management in mixed circulations (see Table 5). In a mixed circulation, the majority will try to decrease the systemic vascular resistance (SVR) to promote systemic circulation (79%). Forty-two percent will try to increase pulmonary vascular resistance (PVR) to promote systemic cardiac output. A combination of both is indicated by 17%, which is probably an underestimation because we did not specifically ask how many used a combination. Table 5. Strategy in Patients With Mixed Circulation.a N % Routinely decrease SVR, with 70/89 79 Milrinone 87/88 99 Nitroprusside 47/88 53 Nitroglycerin 7/88 8 Alpha-blockers 15/88 17 ACE-inhibitors 23/88 26 Others 7/88 8 Routinely increase PVR, with 37/89 42 Increasing PEEP 31/85 36 Lowering Fio 2 < 21% 38/85 45 Hypoventilation 52/85 61 Increase hemoglobin/hematocrit 39/85 46 Do not increase PVR but lower SVR 12/85 14 Abbreviations: ACE, angiotensin converting enzyme; SVR: systemic vascular resistance; PVR, pulmonary vascular resistance; PEEP, positive end expiratory pressure. aMultiple answers possible.

N % Routinely decrease SVR, with 70/89 79 Milrinone 87/88 99 Nitroprusside 47/88 53 Nitroglycerin 7/88 8 Alpha-blockers 15/88 17 ACE-inhibitors 23/88 26 Others 7/88 8 Routinely increase PVR, with 37/89 42 Increasing PEEP 31/85 36 Lowering Fio 2 < 21% 38/85 45 Hypoventilation 52/85 61 Increase hemoglobin/hematocrit 39/85 46 Do not increase PVR but lower SVR 12/85 14 Abbreviations: ACE, angiotensin converting enzyme; SVR: systemic vascular resistance; PVR, pulmonary vascular resistance; PEEP, positive end expiratory pressure. aMultiple answers possible. To achieve an increase in PVR, respondents will try to achieve hypoventilation (61%), increase hemoglobin (46%), lower inhaled oxygen concentration below 21% (45%), and/or increase positive end expiratory pressure (36%). All 85 respondents of this question indicate the use of a combination of these nonpharmacological strategies. Eighty-eight respondents answered the question regarding decreasing SVR. Almost all (87/88) aim to lower SVR by using milrinone (99%) followed by nitroprusside (53%), ACE-inhibitors (26%), alpha-blockers (17%), nitroglycerin (8%), or others like nicardipine or phenoxybenzamine (8%).

To achieve an increase in PVR, respondents will try to achieve hypoventilation (61%), increase hemoglobin (46%), lower inhaled oxygen concentration below 21% (45%), and/or increase positive end expiratory pressure (36%). All 85 respondents of this question indicate the use of a combination of these nonpharmacological strategies. Eighty-eight respondents answered the question regarding decreasing SVR. Almost all (87/88) aim to lower SVR by using milrinone (99%) followed by nitroprusside (53%), ACE-inhibitors (26%), alpha-blockers (17%), nitroglycerin (8%), or others like nicardipine or phenoxybenzamine (8%). Temperature Regulation in PICU Eighty-seven (89%) and 81 (83%) of 98 respondents answered the questions about temperature management post CPB and post deep hypothermic circulatory arrest (DHCA) respectively (see Table 6). Postoperative hyperthermia is prevented routinely in PICU in the majority of the respondents following CPB (52%) with or without DHCA (48%). Only about 8% will routinely try to achieve mild hypothermia (35°C-36°C or 95°F-97°F). Nine respondents (10%) indicate that hypothermia will be used in patients for heart rate control (eg, in case of junctional ectopic tachycardia). Table 6. Temperature Management Following Pediatric Heart Surgery. After CPB After CPB with DHCA N % N % No temperature protocol 13/87 15 19/81 23 Prevent hyperthermia >38°C 45/87 52 39/81 48 36°C-37°C 22/87 25 16/81 20 35°C-36°C 7/87 8 7/81 9 Abbreviations: CPB, cardiopulmonary bypass; DHCA, deep hypothermic circulatory arrest.

Table 6. Temperature Management Following Pediatric Heart Surgery. After CPB After CPB with DHCA N % N % No temperature protocol 13/87 15 19/81 23 Prevent hyperthermia >38°C 45/87 52 39/81 48 36°C-37°C 22/87 25 16/81 20 35°C-36°C 7/87 8 7/81 9 Abbreviations: CPB, cardiopulmonary bypass; DHCA, deep hypothermic circulatory arrest. Systematic Review of the Literature We identified 20 randomized controlled trials comparing inotropes in (animal models of) congenital pediatric heart surgery. Table 7 gives a clear overview of these studies, including design, subjects, the presented results, and the effects on clinically relevant outcomes such as mortality, duration of mechanical ventilation (MV), and length of stay (LOS). Of those 20 studies, 2 were crossover studies reporting on hemodynamic variables5,7 and 2 were animal studies.11,12 Of the remaining 16 studies in children, 6 studies reported on mortality.2,14,17,18,20,21 In none of these studies, mortality was positively (or negatively) influenced by the studied inotrope. Duration of MV or LOS (PICU and/or hospital) was reported by 10 of the 16 studies.2,13,14,16–18,20–23 Only in one report the duration of MV was positively influenced, by a combination of milrinone and inhaled nitric oxide (iNO) in children following the Fontan operation.21 In the other nine studies, no positive effect of inotropes was found on important outcome measures such as duration of MV and LOS. Table 7. All Published Randomized Controlled Trials of Inotropes in (Models of) Pediatric Cardiac Surgery.a

Systematic Review of the Literature We identified 20 randomized controlled trials comparing inotropes in (animal models of) congenital pediatric heart surgery. Table 7 gives a clear overview of these studies, including design, subjects, the presented results, and the effects on clinically relevant outcomes such as mortality, duration of mechanical ventilation (MV), and length of stay (LOS). Of those 20 studies, 2 were crossover studies reporting on hemodynamic variables5,7 and 2 were animal studies.11,12 Of the remaining 16 studies in children, 6 studies reported on mortality.2,14,17,18,20,21 In none of these studies, mortality was positively (or negatively) influenced by the studied inotrope. Duration of MV or LOS (PICU and/or hospital) was reported by 10 of the 16 studies.2,13,14,16–18,20–23 Only in one report the duration of MV was positively influenced, by a combination of milrinone and inhaled nitric oxide (iNO) in children following the Fontan operation.21 In the other nine studies, no positive effect of inotropes was found on important outcome measures such as duration of MV and LOS. Table 7. All Published Randomized Controlled Trials of Inotropes in (Models of) Pediatric Cardiac Surgery.a Author, Year of Publication Inotropes Subjects Results Effect on Mortality? Effect on Mechanical Ventilation or LOS? Booker et al 1995.5 Dopamine vs Dobutamine. Crossover study 19 children undergoing cardiac surgery No significant hemodynamic differences NA NA Wenstone, et al 19916 Dobutamine vs dopamine. 142 children undergoing CPB surgery. No difference in renal function or urine output Not available Not available Kwapisz, et al 20097 Dobutamine vs Dopexamine. Crossover design 11 children undergoing cardiac surgery Both drugs similarly increased cardiac index (measured via transpulmonary thermodilution) NA NA Jaccard et al 19848 Dobutamine vs Isoprenaline. 12 children after correction of Tetralogy of Fallot Isporenaline was more effective than Dobutamine in raising CI. Not available Not available Bailey et al 19979 Amrinone vs sodium nitroprusside (SNP). Crossover study. 10 infants after cardiac surgery Amrinone resulted in significant increase in cardiac index. (SNP did not) NA NA Innes et al 1994.10 Dobutamine vs Enoximone. 28 children following cardiac surgery Mean arterial pressure significantly higher in Dobutamine group. No difference in CI Not available Not available Stocker et al 200711 Milrinone vs levosimendan vs placebo Animal model of pediatric CPB. 16 piglets. Placebo decreased CO by 15%. Milrinone maintained CO. Levosimendan increased CO 14% Not available Not available Riordan et al 199612 Dobutamine vs Adrenaline vs Dopamine Animal model of HLHS. n=6 All three drugs increased CO. Only adrenaline significantly decreased Qp/Qs and increased systemic oxygen delivery. NA NA Hoffman et al 20032 Milrinone vs placebo (PRIMACORP) 238 children after cardiac surgery High-dose milrinone (0.75 μg/kg/min) reduced the relative risk of LCOS by 55% (P = .023). No difference in urine output, creatinine clearance, lactate, or mixed venous saturation. No difference in mortality No difference in duration of MV or LOS.

20032 Milrinone vs placebo (PRIMACORP) 238 children after cardiac surgery High-dose milrinone (0.75 μg/kg/min) reduced the relative risk of LCOS by 55% (P = .023). No difference in urine output, creatinine clearance, lactate, or mixed venous saturation. No difference in mortality No difference in duration of MV or LOS. Laitinen et al 199713 Amrinone vs dopamine & nitroglycerin 32 infants after AVSD repair Systemic and pulmonary blood flow were higher in Amrinone group with lower oxygen consumption No mortality No difference in duration of MV or LOS Laitinen et al 199914 Amrinone vs dopamine & nitroglycerin. 35 Neonates after arterial switch operation Amrinone provided better CO, lower SVR, and lower oxygen consumption Not available No difference in duration of MV, or LOS Momeni et al 201115 Levosimendan vs Milrinone. 36 children after congenital heart surgery No difference in lactate 4 hours postop. Less oxygen demand with levosimendan Not available Not available Lechner et al 201216 Levosimendan (0.1) vs Milrinone (0.5). 40 infants after corrective open heart surgery Postoperative CO similar (transesophageal Doppler). HR, systemic arterial pressure, LAP, MvsO2, lactate, fractional shortening and NIRS were all similar. Not available No difference in duration of MV or LOS Ebade et al 201317 Levosimendan vs Dobutamine. 50 children with pulmonary hypertension undergoing surgical repair of cardiac septal defects Both significantly decreased PAP and CI. But levosimendan was more potent No mortality No difference in duration of MV or ICU LOS Ricci et al 201218 Levosimendan & Milrinone & dopamine vs Milrinone & dopamine. 63 newborns after cardiac surgery LCOS in 37% of levosimendan patients and in 61% of standard regimen (P = .059). Lower lactate, heart rate, and inotrope score in levosimendan group. No improvement in mortality No improvement in duration of MV or LOS Pellicer et al 201319 Levosimendan vs Milrinone. 20 children undergoing cardiovascular surgery Infants receiving Levosimendan had higher peripheral oxygenation (NIRS) and lower inotrope scores. No differences in pro-BNP, troponin, or echo findings. Not available Not available Costello et al 201420 Nesiritide vs Milrinone vs Placebo. 106 children after Fontan surgery No difference in median days alive.

ar surgery Infants receiving Levosimendan had higher peripheral oxygenation (NIRS) and lower inotrope scores. No differences in pro-BNP, troponin, or echo findings. Not available Not available Costello et al 201420 Nesiritide vs Milrinone vs Placebo. 106 children after Fontan surgery No difference in median days alive. No difference in cardiac index, inotropes scores No difference No difference in duration of MV or LOS Cai et al 200821 Nitric oxide vs Milrinone vs nitric oxide & Milrinone 64 children with high PVR following Fontan surgery Combination of nitric oxide & Milrinone led to most significant reduction in PVR and improvement of oxygenation. All three combinations significantly improved systemic circulation. No mortality Time on MV was shortest in combined iNO and milrinone group. No difference in LOS. Bettendorf et al. 200022 Triiodothyronine vs placebo 40 children after cardiac surgery CI and systolic function improved significantly Not available No significant difference in duration of MV or LOS Chowdhury et al 200123 Triiodothyronine vs placebo 75 children after cardiac surgery No difference in outcome measures. Significantly lower inotrope score only in newborn subgroup Not available No effect on LOS or duration of MV Abbreviations: NA, not applicable (because of the design of some trials mortality could not be studied); MV, mechanical ventilation; LOS, length of stay; Vs, versus; CO, cardiac output; CI, cardiac index; iNO, inhaled nitric oxide; LCOS, low cardiac output syndrome; PVR, pulmonary vascular resistance; NIRS, near-infrared spectrometry; SVR, systemic vascular resistance; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; Qp/Qs, pulmonary to systemic blood flow ratio; CPB, cardiopulmonary bypass; HLHS, hypoplastic left heart syndrome.

w cardiac output syndrome; PVR, pulmonary vascular resistance; NIRS, near-infrared spectrometry; SVR, systemic vascular resistance; PAP, pulmonary artery pressure; PCWP, pulmonary capillary wedge pressure; Qp/Qs, pulmonary to systemic blood flow ratio; CPB, cardiopulmonary bypass; HLHS, hypoplastic left heart syndrome. aAll are randomized controlled trials unless otherwise mentioned. Discussion This is the first international survey of the use of inotropes and intensive care management after pediatric cardiac surgery. There have been one Italian and one European survey concerning pediatric LCOS.24–26 Prevention of Low Cardiac Output Syndrome The results of this survey show that, worldwide, milrinone is being used by 97% of respondents for the prevention of LCOS in children (Table 2). The other drugs used are adrenaline/epinephrine (45%), dopamine (38%), dobutamine (11%), and/or levosimendan (5%). These results include the use of multiple drugs. This diversity in pharmacological approaches reflects the current lack of sufficient guiding evidence in the literature.3

ion of LCOS in children (Table 2). The other drugs used are adrenaline/epinephrine (45%), dopamine (38%), dobutamine (11%), and/or levosimendan (5%). These results include the use of multiple drugs. This diversity in pharmacological approaches reflects the current lack of sufficient guiding evidence in the literature.3 In 2011, a European survey also reported a marked variability with milrinone as the most used inotrope (71%) in pediatric cardiac surgery patients which is significantly fewer than 97% in our survey.24 In their survey, dopamine (19%) and adrenaline (16%) were used much less frequently than in our survey (45% and 38%, respectively). A recent single-country survey reported that a combination of dopamine and milrinone was the most frequent drug regimen for prevention of LCOS after congenital heart surgery in Italy.26 This difference in the use of inotropes could reflect differences in European and American approach to the use of inotropes as most of our respondents were North American or Canadian. Due to the design of our study and the survey engine used, it was not possible to do a subanalysis and compare continents.

y in Italy.26 This difference in the use of inotropes could reflect differences in European and American approach to the use of inotropes as most of our respondents were North American or Canadian. Due to the design of our study and the survey engine used, it was not possible to do a subanalysis and compare continents. Improving Cardiac Output To improve postoperative cardiac output, there is more diversity in inotrope use, whereby milrinone, adrenaline/epinephrine, dopamine, and dobutamine are mostly used (Table 4). In the European and Italian surveys, milrinone was most often used for the treatment of LCOS with elevated SVR or elevated PVR.25,26 In the European study, dobutamine is preferred for the treatment of LCOS with low SVR, but adrenaline/epinephrine or dopamine were first and second choice in the Italian survey. In this current survey, the authors did not specify the different kinds of LCOS (high SVR, low SVR). Thus, in this and other surveys, multiple drugs are used for the prevention and treatment of LCOS in pediatric patients after cardiac surgery. But there seems to be a clear preference for milrinone despite the lack of compelling evidence for its benefit over other drug regimens.3

Improving Cardiac Output To improve postoperative cardiac output, there is more diversity in inotrope use, whereby milrinone, adrenaline/epinephrine, dopamine, and dobutamine are mostly used (Table 4). In the European and Italian surveys, milrinone was most often used for the treatment of LCOS with elevated SVR or elevated PVR.25,26 In the European study, dobutamine is preferred for the treatment of LCOS with low SVR, but adrenaline/epinephrine or dopamine were first and second choice in the Italian survey. In this current survey, the authors did not specify the different kinds of LCOS (high SVR, low SVR). Thus, in this and other surveys, multiple drugs are used for the prevention and treatment of LCOS in pediatric patients after cardiac surgery. But there seems to be a clear preference for milrinone despite the lack of compelling evidence for its benefit over other drug regimens.3 The Ideal Inotrope The ideal inotrope would consistently improve systolic and diastolic cardiac function, decrease afterload, improve cardiac output, have a favorable effect on myocardial oxygen hemodynamics, and improve survival and quality of life, with as little adverse effects and interactions as possible. Furthermore, ideally, different specific inotropes would have proven effects and benefits for specific disease states. Unfortunately, the ideal inotrope does not exist. But the (cardiac) intensivist does have a number of pharmacological options at his or her disposal with well-known hemodynamic effects.27

as possible. Furthermore, ideally, different specific inotropes would have proven effects and benefits for specific disease states. Unfortunately, the ideal inotrope does not exist. But the (cardiac) intensivist does have a number of pharmacological options at his or her disposal with well-known hemodynamic effects.27 Milrinone Multiple studies in adults, and children, have shown the positive hemodynamic effects in response to milrinone, an effect not seen in preterm infants.28–34 Milrinone is a phosphodiesterase inhibitor and improves contractility by inhibiting the breakdown of cyclic adenosine monophosphate (cAMP).29 Other effects are afterload reduction and improved diastolic function (lusotropy) with minimal increase in myocardial oxygen consumption. Together these effects lead to an increased cardiac index and decrease in the mean arterial blood pressure (MAP). Milrinone has a relatively long half-life of 2 to 4 hours.35,36 Stocker et al showed that milrinone can maintain normal cardiac output in an animal model of congenital heart surgery.11

l oxygen consumption. Together these effects lead to an increased cardiac index and decrease in the mean arterial blood pressure (MAP). Milrinone has a relatively long half-life of 2 to 4 hours.35,36 Stocker et al showed that milrinone can maintain normal cardiac output in an animal model of congenital heart surgery.11 Milrinone in the pediatric population was first described in 1995 by Chang et al in ten neonates following cardiac surgery.37 The authors, using a thermodilution pulmonary artery catheter (Baxter Edwards Critical-Care, Irvine, CA), measured an increase in cardiac index from 2.1 (± 0.5) to 3.0 (± 0.8) L/m2/min with the use of 0.5 μg/kg/min milrinone and 3 to 7 μg/kg/min dopamine, without increased myocardial oxygen consumption. This was not a controlled trial and they did not comment on clinical outcomes. Then in 1998, a pharmacokinetic study in 19 infants and children reported therapeutic plasma levels with both 0.5 and 0.75 μg/kg/min milrinone with reduced clearance compared to adults, especially in infants.38 They also did not report on clinical outcomes.

ontrolled trial and they did not comment on clinical outcomes. Then in 1998, a pharmacokinetic study in 19 infants and children reported therapeutic plasma levels with both 0.5 and 0.75 μg/kg/min milrinone with reduced clearance compared to adults, especially in infants.38 They also did not report on clinical outcomes. In the landmark Primacorp study, by Hoffman et al, a high-dose milrinone (0.75 μg/kg/min) led to a significant reduction in surrogate markers of low cardiac output by 55% in children after biventricular repair.2 However, there was no difference in lactate, duration of MV, hospital LOS, or mortality. Low-dose milrinone (0.25 μg/kg/min) showed no benefit compared to placebo. Unfortunately the dose of milrinone used was not addressed in the current survey. It might be that underdosing of milirinone can explain the lack of positive effects on clinically relevant outcomes.39

duration of MV, hospital LOS, or mortality. Low-dose milrinone (0.25 μg/kg/min) showed no benefit compared to placebo. Unfortunately the dose of milrinone used was not addressed in the current survey. It might be that underdosing of milirinone can explain the lack of positive effects on clinically relevant outcomes.39 A pharmacological study in 2013 showed that a median dose of milrinone of 0.5 μg/kg/min led to either sub- or supratherapeutic levels (<100 and >300 ng/mL, respectively) in 52% of the patients.40 The authors report that there is an association between LCOS (defined as lactate > 2 mmol/L) and supratherapeutic milrinone blood levels, there was no association between subtherapeutic blood levels and LCOS. There was a higher incidence of arrhythmias with milrinone levels >200 ng/mL, but otherwise no correlation was found between blood levels and clinical outcomes such as hypotension, tachycardia, or lactate. However, milrinone dosing was not protocolized and given at the discretion of the treating physician, which may have confounded the results. The authors report an important suboptimal use of milrinone but do not comment on clinical outcome measures such as duration of MV, hospital LOS, and/or mortality.40

, or lactate. However, milrinone dosing was not protocolized and given at the discretion of the treating physician, which may have confounded the results. The authors report an important suboptimal use of milrinone but do not comment on clinical outcome measures such as duration of MV, hospital LOS, and/or mortality.40 A pharmacokinetic drug-disease model, reported by Vogt et al, also suggests that the current dosage for prevention as well as for treatment of LCOS might not be sufficient and suggests dosing should be age stratified.39 Another pharmacological study conversely suggested the use of lower dose milrinone (0.2 μg/kg/min) because of drug accumulation post CPB surgery in children.41 The effects of these age-related doses on the cardiac output and clinical outcomes remain to be investigated. The only study in which milrinone improved an important clinical outcome was reported by Cai et al who showed that in 46 children following Fontan-type surgery, combined use of iNO and milrinone (0.5 μg/kg/min) did significantly reduce the duration of MV.21 Length of hospital stay, however, was not affected.

A pharmacokinetic drug-disease model, reported by Vogt et al, also suggests that the current dosage for prevention as well as for treatment of LCOS might not be sufficient and suggests dosing should be age stratified.39 Another pharmacological study conversely suggested the use of lower dose milrinone (0.2 μg/kg/min) because of drug accumulation post CPB surgery in children.41 The effects of these age-related doses on the cardiac output and clinical outcomes remain to be investigated. The only study in which milrinone improved an important clinical outcome was reported by Cai et al who showed that in 46 children following Fontan-type surgery, combined use of iNO and milrinone (0.5 μg/kg/min) did significantly reduce the duration of MV.21 Length of hospital stay, however, was not affected. In adults there have been concerns of chronic use of milrinone increasing mortality.42,43 However, a more recent meta-analysis of 20 randomized trials showed no increased mortality due to milrinone in adult patients undergoing cardiac surgery.44 But the higher-quality trials did suggest a trend toward increased mortality with milrinone. Although pediatric studies were included, no pediatric subanalysis was performed. Adults with chronic heart failure are very different than young children with congenital heart defects following surgery, but unlimited use of milrinone might not be without harm. Therefore, we can conclude from the literature that there is insufficient evidence that milrinone improves clinical outcomes post pediatric heart surgery, a finding confirmed by a recent Cochrane review.3

In adults there have been concerns of chronic use of milrinone increasing mortality.42,43 However, a more recent meta-analysis of 20 randomized trials showed no increased mortality due to milrinone in adult patients undergoing cardiac surgery.44 But the higher-quality trials did suggest a trend toward increased mortality with milrinone. Although pediatric studies were included, no pediatric subanalysis was performed. Adults with chronic heart failure are very different than young children with congenital heart defects following surgery, but unlimited use of milrinone might not be without harm. Therefore, we can conclude from the literature that there is insufficient evidence that milrinone improves clinical outcomes post pediatric heart surgery, a finding confirmed by a recent Cochrane review.3 However, despite this lack of evidence, the positive hemodynamic effects of milrinone do lead to a widespread use as shown in our and other surveys.24–26 One of the hemodynamic effects is vasodilation with afterload reduction leading to improved cardiac output, however, there is also a high concomitant use of vasopressors as 55% report the use of noradrenaline and 43% use vasopressin to treat low SVR and blood pressure in our survey. Therefore, we must ask ourselves how much of the beneficial effect of milrinone is being counteracted by the use of vasopressors. This may be a reason why no study has yet been able to report a positive effect on clinical outcomes.

noradrenaline and 43% use vasopressin to treat low SVR and blood pressure in our survey. Therefore, we must ask ourselves how much of the beneficial effect of milrinone is being counteracted by the use of vasopressors. This may be a reason why no study has yet been able to report a positive effect on clinical outcomes. Adrenaline/Epinephrine Epinephrine or adrenaline, is an endogenous catecholamine with dose-dependent effects. At lower doses (0.05-0.1 μg/kg/min), it stimulates the beta-adrenergic receptors leading to increased contractility, increased heart rate, and afterload reduction through peripheral vasodilation.45,46 At higher doses (>0.1 μg/kg/min), it also stimulates alpha-adrenergic receptors leading to vasoconstriction. The stimulation of the beta-receptors leads to an increased calcium influx into the myocardial cell and increased cAMP, at the cost of increased oxygen consumption. In 1979, Benzing et al reported that the cardiac index in 13 children after cardiac surgery increased significantly when epinephrine was added to sodium nitroprusside, without a significant change in systemic resistance.47 In animal models of single ventricle physiology, adrenaline increased cardiac output, systemic perfusion, and systemic oxygen delivery.12,48 Despite the fact that there are no randomized controlled trials with adrenaline in pediatric heart surgery, 45% and 40% of respondents use adrenaline for prevention and treatment of LCOS, respectively, in this survey.

hat such techniques may promote symmetrical distribution of pulmonary blood flow which could contribute to favorable growth of the branch pulmonary arteries and minimize the chance of thrombus formation. The lack of adequate length of SVC for direct anastomosis can be solved by using a graft made by rolled pericardium. Acknowledgments The authors would like to express their deepest gratitude to the Cardiovascular Center, E Hospital for supporting them in the data collection process. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. ORCID iD: Tran-Thuy Nguyen, MD http://orcid.org/0000-0001-7129-0525 Abbreviations and Acronyms IVCinferior vena cava LSVCleft superior vena cava PApulmonary artery PLSVCpersistent left superior vena cava RAright atrium RSVCright superior vena cava RVright ventricle SVCsuperior vena cava

hysiology, adrenaline increased cardiac output, systemic perfusion, and systemic oxygen delivery.12,48 Despite the fact that there are no randomized controlled trials with adrenaline in pediatric heart surgery, 45% and 40% of respondents use adrenaline for prevention and treatment of LCOS, respectively, in this survey. Dobutamine Dobutamine is a synthetic catecholamine, with a short half-life of two minutes, which asserts its effect through beta receptors, stimulating a calcium influx and increase in cAMP. It has positive inotropic and chronotropic effects with peripheral systemic vasodilatation and no pulmonary vasodilatation.8,49–52 Important side effects are arrhythmias and increased myocardial oxygen consumption. Several adult and pediatric studies have found that dobutamine led to an increase in cardiac index in patients who underwent cardiac surgery, due to an increase in heart rate without an increase in stroke volume.8,34,49 Myocardial oxygen uptake and coronary blood flow is increased with dobutamine.53 In an animal model, Ferrara has found that dobutamine at higher dosage produced an increase in cardiac output only in term animals compared to preterm animals, with a significant increase in mean systemic arterial blood pressure.54 When compared to enoximone, dobutamine also increased the MAP, but there was no difference in cardiac output in a study of 28 children following heart surgery.10 The authors did not comment on mortality, duration of MV, or LOS. There is, to our knowledge, no literature on the effect of dobutamine on clinically relevant outcomes in children after heart surgery despite its well-known hemodynamic effects. Dobutamine to prevent LCOS is used routinely by 11% in this survey.

rgery.10 The authors did not comment on mortality, duration of MV, or LOS. There is, to our knowledge, no literature on the effect of dobutamine on clinically relevant outcomes in children after heart surgery despite its well-known hemodynamic effects. Dobutamine to prevent LCOS is used routinely by 11% in this survey. Dopamine Dopamine also has dose-dependent actions but with relatively unpredictable dose response.5,55 In low doses (<5 μg/kg/min) dopamine acts on dopaminergic receptors inducing natriuresis, diuresis, and increased mesenteric flow.54,56 In medium dose (5-15 μg/kg/min) dopamine stimulates beta1-adrenergic receptors increasing cardiac output through chronotropy (mild effect), inotropy, and afterload reduction and increases glomerular filtration rate (GFR) and urine output through via dopaminergic receptors.57 At high doses (>15 μg/kg/min) peripheral resistance increased via alpha-receptors leading to an increased MAP. Side effects are increased myocardial oxygen consumption, arrhythmias, hypothyroidism, and impaired T-lymphocyte proliferation.55,58,59 In one adult study comparing dobutamine with dopamine, both drugs increased myocardial oxygen uptake but dopamine seemed to cause coronary constriction.53 In children, an increase in pulmonary artery pressure (PAP) and PVR is seen with dopamine in a dose higher than 7.5 μg/kg/min.5 Dopamine also causes an increase in cerebral, cardiac, and intestinal blood flow in term and preterm animals, which is dose dependent.54 Dopamine has not been proven useful in the prevention or alteration of the course of acute renal failure.60 When compared to dobutamine, there was no difference in renal function or urine output in 142 children after heart surgery.6 Again, to our knowledge, there is no literature on the effect of dopamine on clinically relevant outcomes in children after heart surgery. However, because of its hemodynamic effects, it is used to prevent LCOS by 38% of respondents in this survey.

in renal function or urine output in 142 children after heart surgery.6 Again, to our knowledge, there is no literature on the effect of dopamine on clinically relevant outcomes in children after heart surgery. However, because of its hemodynamic effects, it is used to prevent LCOS by 38% of respondents in this survey. Other drugs and strategies Levosimendan, a calcium-sensitizer, is a relatively new inotrope and one of the most studied.61–64 Levosimendan has a half-life of 1 hour (in adults) and is metabolized into an active metabolite, OR-1896, with a much longer half-life of 75 to 80 hours.65,66 Both levosimendan and OR-1896 improve myocardial contractility by making troponin C more sensitive to the available intracellular calcium during systole, without affecting diastolic function.67 It does not cause chronotropy and does not increase myocardial oxygen demand.68,69 By stimulating potassium channels in vascular smooth muscle cells, levosimendan causes peripheral, coronary, and pulmonary vasodilatation.70,71 Because of these effects, levosimendan could be useful in children with congenital heart disease.72 The most important side effect is hypotension due to peripheral vasodilation. Levosimendan improves outcomes in adults with acute decompensation of chronic heart failure, and in adults with poor left ventricle function undergoing coronary artery bypass surgery.63,64,73 In neonatal and infant congenital cardiac surgery patients, levosimendan was shown to be as efficacious as milrinone.15,16 However, myocardial oxygen demand was significantly lower in the levosimendan group. In a pediatric study, levosimendan lowered PAP and increased cardiac index significantly more than dobutamine.17 But when added to a standard postoperative protocol of milrinone and dopamine, levosimendan did not improve mortality, duration of MV, or hospital LOS in one study of 63 neonates.18 After heart surgery, levosimendan improved peripheral oxygenation (NIRS) and lowered inotrope scores in 20 children.19 The current level of evidence is insufficient to judge whether prophylactic levosimendan prevents LCOS and mortality in pediatric patients undergoing surgery for congenital heart disease.74 The use of levosimendan is reported by 5% of respondents of this survey. Currently, there are three studies registered at clinicaltrials.gov comparing milrinone with levosimendan in neonatal cardiac surgery (NCT00549107, NCT01576094, and NCT00695929).

ediatric patients undergoing surgery for congenital heart disease.74 The use of levosimendan is reported by 5% of respondents of this survey. Currently, there are three studies registered at clinicaltrials.gov comparing milrinone with levosimendan in neonatal cardiac surgery (NCT00549107, NCT01576094, and NCT00695929). Calciumgluconate is often used based on the knowledge that the neonatal myocardium is calcium dependent. It increases MAP and SVR, probably due to peripheral vasoconstriction but has been shown not to increase cardiac index in adults or in premature neonates.75–77 The only report of calciumgluconate increasing myocardial contractility and cardiac output was in healthy horses without changing filling pressures or MAP.78 Hypothyroidism, with prolonged ICU LOS, is well described in adults and children following CPB.23,79–81 Triiodothyronin supplementation may be a useful adjunct in the management of patients after CPB, improving cardiac index, cardiac function, and inotrope requirements, but its positive effect on clinical outcomes has not been established.22,23,82,83 New inotropes like istaroxime, nesiritide, apelin, or omecamtiv mecarbil have shown some beneficial hemodynamic effects in adults, but there is too limited evidence in children at this time.20 Costello et al compared nesiritide with milrinone and placebo in 106 children following Fontan operation and found no difference in cardiac index or survival.20 None of the respondents in this survey reported their use.

eficial hemodynamic effects in adults, but there is too limited evidence in children at this time.20 Costello et al compared nesiritide with milrinone and placebo in 106 children following Fontan operation and found no difference in cardiac index or survival.20 None of the respondents in this survey reported their use. Amrinone, a phosphodiesterase inhibitor, is not being used anymore but has been shown to increase cardiac output in children following heart surgery.9,13,14 Postoperative Temperature Control The use of hypothermia during heart surgery is well established, decreasing oxygen demand. Following congenital heart surgery, moderate hypothermia (32°C-33°C) has been used in children with refractory low cardiac output state.84 In the retrospective study, cooling led to a significant increase in MAP, pH, and urine output and a decrease in heart rate, right atrial pressure, and platelet count.

xygen demand. Following congenital heart surgery, moderate hypothermia (32°C-33°C) has been used in children with refractory low cardiac output state.84 In the retrospective study, cooling led to a significant increase in MAP, pH, and urine output and a decrease in heart rate, right atrial pressure, and platelet count. Hyperthermia is known to increase oxygen consumption by 11% for every 1°C rise in temperature above 36°C.85 Therefore, it has been advocated that normothermia should be maintained in children after cardiac surgery.55 Most respondents of this survey routinely prevent hyperthermia or try to achieve a temperature below 37°C. Only 9% report to target 35°C to 36°C degrees routinely after congenital heart surgery with DHCA. Whether targeting hypothermia will improve outcomes remains to be investigated. At the moment no clinical trials comparing postoperative temperature management are registered at clinicaltrials.gov. The use of hypothermia for heart rate control is beyond the scope of this survey and review.

ital heart surgery with DHCA. Whether targeting hypothermia will improve outcomes remains to be investigated. At the moment no clinical trials comparing postoperative temperature management are registered at clinicaltrials.gov. The use of hypothermia for heart rate control is beyond the scope of this survey and review. Future Research It would be very interesting to investigate the difference between milrinone and other inotropes on mortality, duration of MV, and LOS in future randomized trials These multicenter trials need to be large enough, which can be a challenge requiring strict study protocols as many other factors may influence cardiac output postoperatively (eg, anesthesia, surgical techniques, CPB management and duration, postoperative sedation, extubation management, fluid management, and ICU treatment goals such as temperature, pH, blood pressure, etc). Furthermore, these studies have to focus on specific congenital defects as different disease states (eg, biventricular vs univentricular heart disease, VSD vs Fallot, Norwood versus Fontan, etc) will require different inotrope approaches. But first we may have to focus on the pharmacokinetic aspect of inotropes, assessing the effect of inotropic plasma levels on cardiac output and then try to find a dose–effect relationship. This will also need to be done in different age-groups. It will be very hard to convincingly show benefit of one inotrope over another in the many specific situations intensivists are faced with.

notropes, assessing the effect of inotropic plasma levels on cardiac output and then try to find a dose–effect relationship. This will also need to be done in different age-groups. It will be very hard to convincingly show benefit of one inotrope over another in the many specific situations intensivists are faced with. Currently three studies are comparing levosimendan and milrinone, hopefully these studies will be large enough to focus on the important clinical outcomes and not only the hemodynamic effects. Limitations The survey results are limited by the reliance on the self-reports by the participants. One limitation of the survey was that it did not specifically address the dosage of the different drugs that are being been used nor did we address the different types of LCOS. Another limitation is that the survey was set up to determine provider-dependent practice variation rather than center-dependent variation and it was therefore also not possible to compare practice variation between the different continents and/or countries. However, the average number of respondents per center is approximately 1.6, which is consistent for all centers apart from the European and South African centers which both had one respondent per center. Therefore, the practice variation we identified in all respondents also reflects practice variation in the different centers and probably indicates that practice variation is more center dependent than provider dependent.

enters apart from the European and South African centers which both had one respondent per center. Therefore, the practice variation we identified in all respondents also reflects practice variation in the different centers and probably indicates that practice variation is more center dependent than provider dependent. Conclusions There are several inotropes available to the (cardiac) intensivist with, more or less, well described hemodynamic effects. The use of these inotropes seems very logical and appropriate in the prevention and treatment of low cardiac output states following pediatric heart surgery. Different inotropes can be used for different clinical situations, but the choice of inotrope is up to the treating physician as there is a lack of compelling evidence in the literature favoring one inotrope over another or even placebo. Despite this currently insufficient evidence on clinically relevant outcomes, milrinone is, by far, most often used for the treatment and prevention of low cardiac output in children worldwide. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. ORCID iD: Peter P. Roeleveld, MD http://orcid.org/0000-0001-5491-9408 Abbreviations and Acronyms ACEangiotensin converting enzyme cAMPcyclic adenosine monophosphate CPBcardiopulmonary bypass DHCAdeep hypothermic circulatory arrest ECMOextracorporeal membrane oxygenation

Introduction Supravalvar aortic stenosis (SVAS) is a rare congenital cardiac anomaly and a common feature of Williams-Beuren syndrome, but it also occurs as the result of an isolated autosomal dominant trait.1,2 A deletion of the elastin gene causes narrowing of large elastic arteries like the aorta or pulmonary arteries. The narrowing of the aorta is seen characteristically in the supravalvar aorta at the sinutubular junction, this is the so-called “discrete” type of SVAS. When the stenosis extends into the ascending aorta, aortic arch, or the origin of the arch vessels, this is called the “diffuse” type of SVAS.2,3 The incidence of discrete SVAS widely varies (14%-72%) in studies reporting their results of the surgical correction of SVAS.4–9 Several SVAS-associated lesions have been described: stenosis of the coronary arteries, bicuspid aortic valve, subvalvar aortic stenosis, and pulmonary stenosis.3 Supravalvar aortic stenosis is known to be a progressive lesion.4,5 Various variations in operative techniques have been described, which differ by the number of Valsalva sinuses that are augmented by (patch) repair. The “single-patch technique,” “two sinus augmentation with an inverted Y-patch” (both nonsymmetrical correction), and the “three-patch technique” (symmetrical correction) are the techniques implemented by the majority of surgeons for the surgical correction of SVAS (Figure 1). It is unknown whether any of these techniques leads to superior results. The aim of the present study is to review the 52 years’ experience with the surgical correction of SVAS in two of four congenital cardiothoracic surgical centers in the Netherlands

jority of surgeons for the surgical correction of SVAS (Figure 1). It is unknown whether any of these techniques leads to superior results. The aim of the present study is to review the 52 years’ experience with the surgical correction of SVAS in two of four congenital cardiothoracic surgical centers in the Netherlands Figure 1. Three surgical techniques. A, Single-patch technique. B, Two-sinus augmentation with an inverted Y-patch. C, Three-patch technique. Patients and Methods Study Design and Setting This is a retrospective observational study conducted in two centers in the Netherlands: the Center for Congenital Heart Diseases Amsterdam Leiden CAHAL that includes Leiden University Medical Center, Academic Medical Center Amsterdam, and Free University Medical Center Amsterdam ; and the University Medical Center Groningen. To collect data of all patients who underwent surgery for SVAS, the cardiology and cardiac surgery databases of the different institutions were reviewed. All patients who underwent surgery for SVAS from 1962 until 2014 in one of these centers were included. Patients were excluded when they had acquired postsurgical SVAS or when they underwent SVAS correction and aortic valve replacement in the same initial operation. This study was notified to the medical ethics committee, but because of the observational character of the study it is not part of the Wet Medisch-wetenschappelijk Onderzoek = Law Medical Research and did not need approval by the medical ethics committee.

tion and aortic valve replacement in the same initial operation. This study was notified to the medical ethics committee, but because of the observational character of the study it is not part of the Wet Medisch-wetenschappelijk Onderzoek = Law Medical Research and did not need approval by the medical ethics committee. Data Collection We collected the following demographic and preoperative data: gender, date of birth, date of initial surgical SVAS correction, syndrome determined by genetic testing, previous cardiovascular operation, form of SVAS (discrete or diffuse), and concomitant preoperative cardiovascular anomalies. Operative data included the operation technique used at the initial surgical SVAS correction (divided in single-patch technique, two sinus augmentation, and three-patch technique; sliding plasty was not performed in our centers), additional surgical procedures at initial operation, patch material used, complications during the operation, and in-hospital complications during postoperative stay. Follow-up data included restenosis on echo, reoperations in the cardiovascular area, and mortality and were collected at the last follow-up appointment with the cardiologist. Any gradient over the supravalvular area that was higher than the gradient directly postoperative was considered a restenosis. The echocardiographic data of all available echocardiographs were analyzed by one cardiologist (R.B.), Z scores were used according to Pettersen et al.6 Aortic valve disease was defined as valvular stenosis (from mild to severe), regurgitation (from mild to severe), bicuspid aortic valves, or adhesion of one or more leaflets to the supravalvular ridge. For every patient, the general practitioner was consulted in order to check the patients’ well-being. If a patient had deceased, as much as possible information were collected regarding the cause of death. End points were survival rate and freedom from reoperation.

on of one or more leaflets to the supravalvular ridge. For every patient, the general practitioner was consulted in order to check the patients’ well-being. If a patient had deceased, as much as possible information were collected regarding the cause of death. End points were survival rate and freedom from reoperation. Statistical Analysis In order to identify significant differences in baseline characteristics between the two surgical groups (symmetrical vs nonsymmetrical), an independent-sample t test was performed for parametric continuous variables. For nonparametric continuous variables, the Mann-Whitney U test was used. For categorical variables, a χ2 test was used. When expected cell count was less than 5 in >20% of the cases, the Fisher exact test was used. A P value of <.05 was considered significant. The survival and reoperation-free survival of the symmetrical and nonsymmetrical groups were described using the Kaplan-Meier curves. The binary logistic regression analysis was used to compare survival rate and freedom of reoperation rate between the 2 groups at 20 years. Due to the observational character of the study, correcting for selection bias was indicated. This was done by performing a separate logistic regression analysis to assess the propensity score for each patient. Based on selected patient characteristics that are considered to be confounding factors, the propensity score predicts the chance that a patient will belong to the symmetrical or nonsymmetrical group. We considered sex, age at first SVAS correction operation, year of operation, presence of Williams syndrome, presence of pulmonary artery stenosis, and presence of aortic valve disease as potential factors to cause selection bias (confounders). We plotted the propensity scores to check whether the overlap was sufficient. The calculated propensity scores were then added to the model to correct the odds for mortality and freedom of reoperation. We used this logistic regression analysis to compare symmetrical and nonsymmetrical operating techniques in survival and reoperation-free survival. The statistical software system IBM SPSS 22 was used for data analysis.

ty scores were then added to the model to correct the odds for mortality and freedom of reoperation. We used this logistic regression analysis to compare symmetrical and nonsymmetrical operating techniques in survival and reoperation-free survival. The statistical software system IBM SPSS 22 was used for data analysis. Results Patient and Operative Characteristics From 1962 to 2014, a total of 49 patients underwent surgical relief of SVAS. Williams syndrome was present in 24 (49%) patients and Noonan syndrome in 2 (4.1%) patients . An isolated elastin (ELN) mutation was demonstrated in three (6.1%) patients, and in two cases no genetic test was performed. In 42 (85.7%) patients, SVAS was discrete and in 7 (14.3%) patients it was diffuse (Table 1). The mean age at the time of the first operation was 8.16 (standard deviation [SD] = 9.25) years (range: 0-45 years). This wide range is the result of a change in surgical practice over time. After 1974, all patients were operated on as children. Concomitant cardiovascular anomalies are specified in Table 2. Twenty-one (42.9%) patients needed additional procedures during their first SVAS operation. Four (8.2%) needed aortic arch surgery. Others underwent enlargement of pulmonary artery stenosis (n = 5), clipping of patent ductus arteriosus (n = 3), mitral valve repair (n = 3), aortic valve repair (n = 3), valvulotomy of the aortic valve (n = 2), and valvulotomy of the pulmonary valve (n = 1). In 15 (30.6%) patients, untreated autologous pericardium was used as patch material; in 20 (40.8%) patients, glutaraldehyde-treated autologous pericardium was used; in 10 (20.4%) patients synthetic material was used; in 2 (4.1%) patients dura mater was used; in 1 (2.0%) patient xenopericardium was used; and in 1 (2.0%) patient an aortic homograft was used to create a patch. The use of dura mater took place solely in the earlier years of our experience.

logous pericardium was used; in 10 (20.4%) patients synthetic material was used; in 2 (4.1%) patients dura mater was used; in 1 (2.0%) patient xenopericardium was used; and in 1 (2.0%) patient an aortic homograft was used to create a patch. The use of dura mater took place solely in the earlier years of our experience. Table 1. Baseline Characteristics: Preoperative. Variablesa Overall, N = 49 Nonsymmetrical, n = 23 Symmetrical, n = 26 P Value Sex .445 Female 27 55.1% 14 60.9% 13 50.0% Male 22 44.9% 9 39.1% 13 50.0% Syndrome <.01 No syndrome identified 18 36.7% 15 65.2% 3 11.5% Williams-Beuren 24 49.0% 5 21.7% 19 73.1% Noonan 2 4.1% 1 4.3% 1 3.8% ELN mutation de novo 3 6.1% 0 0.0% 3 11.5% Unknown 2 4.1% 2 8.7% 0 0.0% Form of SVAS .011 Discrete 42 85.7% 23 100% 19 73.1% Diffuse 7 14.3% 0 0.0% 7 26.9% Age at operation .097 Mean 8.2 (±9.3) 8.9 (±7.1) 7.5 (±10.9) Median 6 (2.0-11) 6 (4.0-13) 4 (2.0-9.0) Year of operation <.01 Before 1978 8 16.3% 8 34.8% 0 0.0% After 1978 41 83.7% 15 65.2% 26 100% Presence of pulmonary artery stenosis .017 Yes 17 34.7% 4 17.4% 13 50.0% No 32 65.3% 19 82.6% 13 50.0% Presence of aortic valve disease .013 Yes 8 16.3% 6 26.1% 2 7.7% No 41 83.7% 17 73.9% 24 92.3% Abbreviation: ELN, elastin. aCategorical data are presented as number of patients and continuous data as mean (± standard deviation) and interquartile range (Q1-Q3). Table 2. Concomitant Cardiovascular Anomalies.

Variablesa Overall, N = 49 Nonsymmetrical, n = 23 Symmetrical, n = 26 P Value Sex .445 Female 27 55.1% 14 60.9% 13 50.0% Male 22 44.9% 9 39.1% 13 50.0% Syndrome <.01 No syndrome identified 18 36.7% 15 65.2% 3 11.5% Williams-Beuren 24 49.0% 5 21.7% 19 73.1% Noonan 2 4.1% 1 4.3% 1 3.8% ELN mutation de novo 3 6.1% 0 0.0% 3 11.5% Unknown 2 4.1% 2 8.7% 0 0.0% Form of SVAS .011 Discrete 42 85.7% 23 100% 19 73.1% Diffuse 7 14.3% 0 0.0% 7 26.9% Age at operation .097 Mean 8.2 (±9.3) 8.9 (±7.1) 7.5 (±10.9) Median 6 (2.0-11) 6 (4.0-13) 4 (2.0-9.0) Year of operation <.01 Before 1978 8 16.3% 8 34.8% 0 0.0% After 1978 41 83.7% 15 65.2% 26 100% Presence of pulmonary artery stenosis .017 Yes 17 34.7% 4 17.4% 13 50.0% No 32 65.3% 19 82.6% 13 50.0% Presence of aortic valve disease .013 Yes 8 16.3% 6 26.1% 2 7.7% No 41 83.7% 17 73.9% 24 92.3% Abbreviation: ELN, elastin. aCategorical data are presented as number of patients and continuous data as mean (± standard deviation) and interquartile range (Q1-Q3). Table 2. Concomitant Cardiovascular Anomalies. Concomitant Cardiovascular Anomalies Frequency Peripheral pulmonary artery stenosis 14a (28.6%) Localized aortic coarctation 5 (10.2%) Bicuspid aortic valve 3 (6.1%) Aortic valvular stenosis 3 (6.1%) Mitral valve insufficiency 3 (6.1%) Supravalvular pulmonary artery stenosis 3 (6.1%) Patent ductus arteriosus 3 (6.1%) Aortic valve regurgitation 2 (4.1%) Ascending aortic aneurysm 2 (4.1%) Coronary artery stenosis 2 (4.1%) Hypoplastic ascending aorta 2 (4.1%) Pulmonary valve stenosis 1 (2.0%) Left carotid artery 1 (2.0%) Right ventricle hypertrophy 1 (2.0%) Patent foramen ovale 1 (2.0%) Arteria lusoria 1 (2.0%) aNumbers refer to patients with any of the respective anomalies, and presence of more than one anomaly in one patient is possible.

poplastic ascending aorta 2 (4.1%) Pulmonary valve stenosis 1 (2.0%) Left carotid artery 1 (2.0%) Right ventricle hypertrophy 1 (2.0%) Patent foramen ovale 1 (2.0%) Arteria lusoria 1 (2.0%) aNumbers refer to patients with any of the respective anomalies, and presence of more than one anomaly in one patient is possible. We divided the patients in two groups: 23 (46.9%) patients were operated on using the nonsymmetrical techniques (22.4% using the single-patch technique and 24.5% using the two sinus augmentation) and 26 (53.1%) patients were operated using the symmetrical (three-patch) technique.

poplastic ascending aorta 2 (4.1%) Pulmonary valve stenosis 1 (2.0%) Left carotid artery 1 (2.0%) Right ventricle hypertrophy 1 (2.0%) Patent foramen ovale 1 (2.0%) Arteria lusoria 1 (2.0%) aNumbers refer to patients with any of the respective anomalies, and presence of more than one anomaly in one patient is possible. We divided the patients in two groups: 23 (46.9%) patients were operated on using the nonsymmetrical techniques (22.4% using the single-patch technique and 24.5% using the two sinus augmentation) and 26 (53.1%) patients were operated using the symmetrical (three-patch) technique. Mortality Follow-up was 100% complete. The mean follow-up time was 19.2 years (SD: 12.8, median 19). The longest follow-up time was 52 years. A total of eight (16.3%) patients died, of which two were early deaths and six late deaths (Table 3). Of the six late deaths, four were out-of-hospital cardiac arrests, the age of dead patients varied between 33 and 57 years, and only one patient was diagnosed with Williams syndrome. The actuarial survival of all patients at ten years was 94% (standard error [SE]: 0.035), at 20 years 81% (SE: 0.067), at 30 years 81% (SE: 0.067), and at 40 years 68% (SE: 0.135). In the nonsymmetrical group, actuarial survival at 10 years was 96% (SE: 0.043), at 20 years was 80% (SE: 0.091), at 30 years was 80% (SE: 0.091), and at 40 years was 67% (SE: 0.143; Figure 2 ). In the symmetrical group, actuarial survival at ten years was 92% (SE: 0.056) and at 20 years was 85% (SE: 0.085). The odds of death were not significantly higher for patients in the nonsymmetrical group with an adjusted odds ratio of 0.116 (95% CI: 0.006-2.396; P = .163).

nd at 40 years was 67% (SE: 0.143; Figure 2 ). In the symmetrical group, actuarial survival at ten years was 92% (SE: 0.056) and at 20 years was 85% (SE: 0.085). The odds of death were not significantly higher for patients in the nonsymmetrical group with an adjusted odds ratio of 0.116 (95% CI: 0.006-2.396; P = .163). Figure 2. Kaplan-Meier survival curve. The numbers of patients belonging to the nonsymmetrical group are shown in blue, the numbers of patients in the symmetrical group are shown in green. Table 3. Details of Deceased Patients.

nd at 40 years was 67% (SE: 0.143; Figure 2 ). In the symmetrical group, actuarial survival at ten years was 92% (SE: 0.056) and at 20 years was 85% (SE: 0.085). The odds of death were not significantly higher for patients in the nonsymmetrical group with an adjusted odds ratio of 0.116 (95% CI: 0.006-2.396; P = .163). Figure 2. Kaplan-Meier survival curve. The numbers of patients belonging to the nonsymmetrical group are shown in blue, the numbers of patients in the symmetrical group are shown in green. Table 3. Details of Deceased Patients. Sex Age at First Operation Age at Deatha Syndrome Form of SVAS Concomitant Anomalies Preoperative Techniqueb Reoperation Cause of Death 1. M 10 10 (in-hospital mortality) – Discrete -Aortic valve stenosis -Truncus pulmonalis dilatation 1 – Severe left ventricular hypertrophy led to subendocardial infarction 2. M 9 months 9 months (in-hospital mortality) Williams Diffuse -Supravalvular pulmonary artery stenosis -Hypoplastic aortaascenden s 3 Yes Severe left ventricular hypertrophy, arrhythmias, pulmonary dysfunction → ECMO procedure → complicated by a blood clot obstruction → cardiac arrest 3. M 19 40 – Discrete – 1 – OHCA 4. M 45 49 – Discrete -Mitral valve regurgitation 3 Yes OHCA 5. M 39 57 – Discrete – 3 Yes OHCA 6. M 15 33 Williams Discrete – 2 – OHCA 7. F 7 20 Williams Discrete -Peripheral pulmonary artery stenosis 2 – Right ventricular failure 8. F 1 39 – Discrete -Patent ductus arteriosus -Mitral valve regurgitation 1 Yes, and a second reoperation End-stage heart failure Abbreviations: F, female; M, male; OHCA, out-of-hospital cardiac arrest.

Discrete – 2 – OHCA 7. F 7 20 Williams Discrete -Peripheral pulmonary artery stenosis 2 – Right ventricular failure 8. F 1 39 – Discrete -Patent ductus arteriosus -Mitral valve regurgitation 1 Yes, and a second reoperation End-stage heart failure Abbreviations: F, female; M, male; OHCA, out-of-hospital cardiac arrest. aAge in years or months if stated. bOperating technique: 1, single-patch technique; 2, two-sinus augmentation; 3, three-patch technique. Reoperation A total of eight (16.3%) patients were reoperated, involving three early reoperations and five late reoperations (Table 4). All techniques combined, the total percentage of patients free from reoperation was 92% (SE: 0.040) after 10 years and 79% (SE: 0.067) after 20 years. In the nonsymmetrical group, freedom from reoperation after 10 years was 100% and after 20 years 88% (SE: 0.079). There were no reoperations performed on the patients at risk in the nonsymmetrical group more than 20 years after their initial operation (Figure 3). In the symmetrical group, freedom from reoperation after 10 years was 84% (SE: 0.072) and after 20 years was 71% (SE: 0 .107). The odds of freedom of reoperation were not significantly higher for patients in the nonsymmetrical group compared to the symmetrical group: adjusted odds ratio: 0.156, 95% CI: 0.004-5.773; P = .313. The Kaplan-Meier curve for freedom from reoperation (Figure 3) shows a marked lower curve for the symmetrical group, mainly due to more frequent early reoperations in the symmetrical group. Table 4. Details of Reoperated Patients.

Reoperation A total of eight (16.3%) patients were reoperated, involving three early reoperations and five late reoperations (Table 4). All techniques combined, the total percentage of patients free from reoperation was 92% (SE: 0.040) after 10 years and 79% (SE: 0.067) after 20 years. In the nonsymmetrical group, freedom from reoperation after 10 years was 100% and after 20 years 88% (SE: 0.079). There were no reoperations performed on the patients at risk in the nonsymmetrical group more than 20 years after their initial operation (Figure 3). In the symmetrical group, freedom from reoperation after 10 years was 84% (SE: 0.072) and after 20 years was 71% (SE: 0 .107). The odds of freedom of reoperation were not significantly higher for patients in the nonsymmetrical group compared to the symmetrical group: adjusted odds ratio: 0.156, 95% CI: 0.004-5.773; P = .313. The Kaplan-Meier curve for freedom from reoperation (Figure 3) shows a marked lower curve for the symmetrical group, mainly due to more frequent early reoperations in the symmetrical group. Table 4. Details of Reoperated Patients. Sex Age at First Operationa Age at Reoperationb Syndrome Form of SVAS Concomitant Anomalies Preoperative Techniqueb Reoperation Specifications 1. Mc 9 months 9 months (in-hospital reoperation) Williams Diffuse Supravalvular pulmonary artery stenosis hypoplastic aorta ascendens 3 ECMO procedure 2. Mc 45 49 – Discrete Mitral valve regurgitation 3 Mitral valve replacement 3. Fc 8 months 11 months ELN mutation de novo Discrete Peripheral pulmonary artery stenosis 3 Aortic arch repair 4. Mc 39 42 – Discrete 3 Aortic valve replacement 5. M 8 12 Noonan syndrome Discrete Aortic valvular stenosis 3 Aortic valve replacement 6. M 5 18 Noonan syndrome Diffuse Peripheral pulmonary artery stenosis 3 Replacement of ascending-, descending- and aortic arch by vascular prosthesis 7. F 23 39 – Discrete Aortic valvular stenosis 2 Aortic valve replacement 8. Fc 1 21 – Discrete Patent ductus arteriosus mitral valve regurgitation 1 First reoperation: aortic valve replacement mitral valve replacement enlargement ascending aorta second reoperation: mitral valve replacement (larger prosthesis) tricuspid valve repair Abbreviation: SVAS, supravalvar aortic stenosis; ECMO, extracorporeal membrane oxygenation.

tus arteriosus mitral valve regurgitation 1 First reoperation: aortic valve replacement mitral valve replacement enlargement ascending aorta second reoperation: mitral valve replacement (larger prosthesis) tricuspid valve repair Abbreviation: SVAS, supravalvar aortic stenosis; ECMO, extracorporeal membrane oxygenation. aAge in years or months if stated. bOperating technique: 1, single-patch technique; 2, two-sinus augmentation; 3, three-patch technique. cDeceased. Figure 3. Kaplan-Meier reoperation-free survival. The numbers of patients belonging to the nonsymmetrical group are shown in blue, the numbers of patients in the symmetrical group are shown in green. Echocardiography The echocardiographic follow-up was incomplete due to the retrospective character of the study. Some patients did not visit their cardiologist regularly and some were followed up without echocardiographic imaging. Echocardiographic images that were suitable for analysis were available in 13 patients (26.5%), 9 of whom were operated with the nonsymmetrical technique and 4 with the symmetrical technique. None of the patients showed a significant restenosis of the supravalvular area (Figure 4). Figure 4. Echocardiographic gradients. Discussion This study did not show a significant difference in mortality and reoperation rates between an asymmetrical and symmetrical correction of SVAS. Interestingly, compared to the survival in the general population, the long-term survival of SVAS patients is remarkably low.

Echocardiography The echocardiographic follow-up was incomplete due to the retrospective character of the study. Some patients did not visit their cardiologist regularly and some were followed up without echocardiographic imaging. Echocardiographic images that were suitable for analysis were available in 13 patients (26.5%), 9 of whom were operated with the nonsymmetrical technique and 4 with the symmetrical technique. None of the patients showed a significant restenosis of the supravalvular area (Figure 4). Figure 4. Echocardiographic gradients. Discussion This study did not show a significant difference in mortality and reoperation rates between an asymmetrical and symmetrical correction of SVAS. Interestingly, compared to the survival in the general population, the long-term survival of SVAS patients is remarkably low. The most common reason for reoperation identified in the present study was aortic valve replacement, which was identified in previous studies as well.7–19 Enlarging the aortic root could theoretically lead to aortic regurgitation, because the aortic cusps might not be sufficient for the “new”aortic root geometry. A nonsymmetric repair would more likely cause an unbalanced aortic valve, causing regurgitation as well. Four patients needed aortic valve replacement. Two patients were operated with the symmetrical technique, one had aortic valve regurgitation and the other patient had aortic valve stenosis. Two patients were operated on with the nonsymmetrical technique, one had aortic valve regurgitation and the other patient had aortic valve stenosis. Other studies also did not find differences between the groups, mainly due to the small reoperation rates or short follow-up time.3,4,9–19 The results of the single-center study of Metton et al actually showed that at the last follow-up the lowest incidence of moderate aortic regurgitation was found in the three-patch technique group compared to the one- and two-patch technique groups.17

due to the small reoperation rates or short follow-up time.3,4,9–19 The results of the single-center study of Metton et al actually showed that at the last follow-up the lowest incidence of moderate aortic regurgitation was found in the three-patch technique group compared to the one- and two-patch technique groups.17 Previous studies showed that the presence of diffuse SVAS is a possible predictor of early and late mortality.20 In the present study, of the patients who died (n = 8), seven had discrete SVAS and one had diffuse SVAS. Because of the small incidence of diffuse SVAS (n = 7) in this study, no conclusions can be drawn concerning the form of SVAS as a predictor of mortality. Previous studies show low early mortality rates (1%-5%) and when follow-up time increases, both mortality rates and reoperation rates rise. These studies show ten-year survival rates of 84% and 95%, and 20-year survival rates of 70% and 90%.9,10,12,16,21 Our study reports the longest follow-up. Our results show a rise of both mortality and reoperation rate after the first decade (10-year survival: 94%, 20-year survival: 81%, and 40-year survival of only 68%; 10 years of freedom from reoperation: 92% and 20 and 40 years of freedom from reoperation: 79%). The survival in 1962, 1988, and 2014 in these age groups was >99% among the general population in the Netherlands.22 Thus, compared to the survival in the general population, the survival of SVAS patients in our study is remarkable low.

reedom from reoperation: 92% and 20 and 40 years of freedom from reoperation: 79%). The survival in 1962, 1988, and 2014 in these age groups was >99% among the general population in the Netherlands.22 Thus, compared to the survival in the general population, the survival of SVAS patients in our study is remarkable low. Four of the six late deaths were out-of-hospital cardiac arrests. These four patients died at an age between 33 and 57 years. No autopsies were conducted. All were operated on for discrete SVAS and one had Williams syndrome. They all were operated on at a relatively old age (>14), therefore, we could argue that the coronary arteries were exposed to a high pressure in the prestenotic area for a relatively long period of time, which might have caused endothelial damage leading to ischemic heart disease as a cause for their sudden cardiac arrest. In two of these late deaths, echocardiography was conducted, respectively, one and two years prior to death, showing no abnormalities. This remarkably high mortality rates considering the character of the surgery, of which a high percentage out-of-hospital cardiac arrests makes us plead for a closer (preferably lifelong) follow-up of patients after surgical correction of SVAS including exercise capacity and Holter monitoring.

showing no abnormalities. This remarkably high mortality rates considering the character of the surgery, of which a high percentage out-of-hospital cardiac arrests makes us plead for a closer (preferably lifelong) follow-up of patients after surgical correction of SVAS including exercise capacity and Holter monitoring. Strength and Limitations Even though this study has the longest follow-up time (52 years) ever reported in patients operated for SVAS with a mean length of follow-up of 19.2 years (SE: 1.8), and although this study includes patients from two of four congenital heart surgery centers in the Netherlands, the study population consisted of only 49 patients. Also the majority of the patients included have a follow-up of less than equal to ten years, the number of patients declines fast after this point, and therefore it is difficult to draw conclusions after this point. A strong feature of this study is that the different operating techniques are well divided over the population: nonsymmetrical techniques (46.9%) and symmetrical technique (53.1%). Most other studies show higher percentages for one particular technique. Nevertheless, in order to compare those two groups by means of a binary logistic regression analyses, a sample size of 49 patients is too small. As Long stated, sample sizes of <100 should be avoided in these analyses and a sample size of 500 or more should be adequate for almost any situation.23 The rareness of this disease makes it very difficult to obtain a high number of patients in one country.24 As all other studies that are conducted in this field so far, all have study populations of less than 100, the question rises what these conclusions really tell us.3,4,9–19 A meta-analysis would be an option. As shown in Table 1, the two groups are significantly different on various characteristics. We corrected for these differences and potential confounding factors by means of the propensity scores. Nevertheless, this will not provide total equalization of the groups. Unmeasured confounders will still be there, for example, the form of SVAS could not be matched using propensity scoring because the number of patients was too small. Nevertheless, all seven patients with the diffuse form of SVAS were in the symmetrical group. This might be considered an unmeasured confounding factor. Also all patients operated on before 1978 belonged to the nonsymmetrical group, and this group had a higher percentage of preoperative pulmonary artery stenosis.

mall. Nevertheless, all seven patients with the diffuse form of SVAS were in the symmetrical group. This might be considered an unmeasured confounding factor. Also all patients operated on before 1978 belonged to the nonsymmetrical group, and this group had a higher percentage of preoperative pulmonary artery stenosis. This cohort is too small to state that one confounder might influence the results more than another or might influence the results at all. In our cohort, there were only three patients with SVAS and a bicuspid aortic valve, all three were operated on with a nonsymmetrical technique. Based on our present experience, we do not prefer a nonsymmetrical over a symmetrical reconstruction in patients with a bicsupid aortic valve. We would enlarge the sinuses with one or two patches depending on what looks best. We took survival and reoperation as end points. In both groups, there were eight events. The small cohort and the relatively small number of events with the end point reoperation are limitations of this study. The lack of additional information regarding those deaths is another limitation, as is the less than optimal echocardiographic follow-up percentage (74%) for this cohort. Conclusions In this patient group, there was no significant difference in survival and freedom from reoperation between the different surgical techniques for SVAS repair. Compared to the survival in the general population, the survival of SVAS patients is remarkably low. Apparently, SVAS is not a benign disease and probably patients should be followed more closely for the rest of their lives.

urvival and freedom from reoperation between the different surgical techniques for SVAS repair. Compared to the survival in the general population, the survival of SVAS patients is remarkably low. Apparently, SVAS is not a benign disease and probably patients should be followed more closely for the rest of their lives. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.

Dear Dr Jacobs, Response to the letter to the editor by Kaufman et al regarding our article “The Perspective of the Intensivist on Inotropes and Postoperative Care Following Pediatric Heart Surgery: An International and Systematic Review of the Literature” in the World Journal for Pediatric and Congenital Heart Surgery. 2018;9(1). I would like to thank Drs Kaufman and da Cruz for their interest in our article and their comments and would also like to thank the editor for providing us with an opportunity to respond. I certainly agree with our colleagues from Colorado that vasopressin can be a very useful pharmacologic agent to treat vasoplegia of different causes. However, treating vasoplegia and increasing blood pressure is not the same as increasing cardiac output, which was the focus of our review.1 The use of vasoactive medication, vasodilators, and vasoconstrictors probably merits a survey and review on its own.

l pharmacologic agent to treat vasoplegia of different causes. However, treating vasoplegia and increasing blood pressure is not the same as increasing cardiac output, which was the focus of our review.1 The use of vasoactive medication, vasodilators, and vasoconstrictors probably merits a survey and review on its own. Many centers will actually aim to reduce afterload as much as possible after cardiac surgery using vasodilators such as milrinone, low-dose dobutamine, sodium nitroprusside, α-blockers, and/or ACE inhibitors.2,3 In our survey, 79% of respondents indicated they employ a strategy of decreasing afterload in postoperative Norwood patients specifically.1 It was recently shown in an animal model with fixed loading conditions that milrinone is very poor at increasing cardiac output.4 My personal belief is that with milrinone, which owes 30% of its increase in cardiac output due to vasodilatation,5,6 we often see hypotension which then requires a vasoactive drug to increase/normalize blood pressure. In our survey, 30% of respondents indicated they added norepinephrine for this reason specifically and a total of 55% of respondents indicated the use of norepinephrine and 43% use vasopressin in their prophylactic regimen.1 Respondents did not specify why a vasoconstrictor was added so frequently, but we assume it was to increase blood pressure as vasopressin has no inotropic properties. The best approach to improve cardiac output is unknown and it still raises many questions. I wonder how much increase in cardiac output can actually be achieved with milrinone when a vasoconstrictive agent is added? Aren’t we just increasing afterload again and therefore mitigating the milrinone effect? How useful is milrinone with added vasoconstrictor for the clinically relevant outcomes of our patients? Should we even use milrinone? If we should, in which dose? And if we would not, would we need less vasoconstrictors such as vasopressin? We just don’t know and this deserves further research. However, when it comes to treatment of postoperative low cardiac output syndrome (LCOS), only 3 (3.5%) of 86 in our survey indicated the use of vasopressin as a second-tier drug for the treatment of LCOS (unpublished data). This underlines our belief that vasoconstrictors do not have a role in increasing cardiac output. They do have a role in treating (postoperative) vasoplegia, and I thank Dr Kaufman and da Cruz for pointing that out.

vey indicated the use of vasopressin as a second-tier drug for the treatment of LCOS (unpublished data). This underlines our belief that vasoconstrictors do not have a role in increasing cardiac output. They do have a role in treating (postoperative) vasoplegia, and I thank Dr Kaufman and da Cruz for pointing that out. ORCID iD: Peter Paul Roeleveld http://orcid.org/0000-0001-5491-9408

Background Systemic venous anomalies are no longer considered a major risk factor for bidirectional Glenn operation and Fontan procedure. However, the presence of persistent left superior vena cava (PLSVC) is still a challenge for safe and effective completion of superior cavopulmonary amalgamation. From the anatomical perspective, in the presence of bilateral superior caval veins, the small diameters of each vein will reduce the blood flow at the level of each anastomosis. If one superior vena cava (SVC) is considerably smaller than the other, the resultant asymmetrical distribution of pulmonary blood flow may contribute to increased risk of thrombus formation and/or the unfavorable growth of pulmonary arteries.1–3 In order to improve the SVC blood flow, many ideas have been proposed: innominate vein formation by Gore-Tex prosthetic graft,4 inferior anastomosis of two SVCs to form a venous confluence,3 or other means of converting two SVCs into a single vessel,2 including the use of an aortic homograft conduit.5 In this study, we present a technical modification for bidirectional Glenn operation in patients with bilateral superior caval veins that can possibly optimize the growth of pulmonary arteries and reduce the risk of thrombus formation. In addition, we also review the literature about modifications of bilateral cavopulmonary anastomosis.

present a technical modification for bidirectional Glenn operation in patients with bilateral superior caval veins that can possibly optimize the growth of pulmonary arteries and reduce the risk of thrombus formation. In addition, we also review the literature about modifications of bilateral cavopulmonary anastomosis. Participants and Method Two patients diagnosed with functionally univentricular heart defects with bilateral superior caval veins had the indications for bidirectional Glenn operation, treated at E Hospital. The study was reviewed and approved by the ethics committee of E Hospital. All study procedures complied with the ethical principles of biomedical research. Participants consented to take part in the study and were told that they could withdraw at any time. Participants’ information was kept secure and confidential. Results Case 1 An 11-month-old boy weighing 7.4 kg, cyanotic, with SpO2 67% on room air had undergone Blalock-Taussig shunt operation at the age of 3 months. At the time of the current assessment, complete blood count and biochemistry test results were within the normal range.

Participants and Method Two patients diagnosed with functionally univentricular heart defects with bilateral superior caval veins had the indications for bidirectional Glenn operation, treated at E Hospital. The study was reviewed and approved by the ethics committee of E Hospital. All study procedures complied with the ethical principles of biomedical research. Participants consented to take part in the study and were told that they could withdraw at any time. Participants’ information was kept secure and confidential. Results Case 1 An 11-month-old boy weighing 7.4 kg, cyanotic, with SpO2 67% on room air had undergone Blalock-Taussig shunt operation at the age of 3 months. At the time of the current assessment, complete blood count and biochemistry test results were within the normal range. Echocardiographic findings were transposition of the great arteries, aorta arising from the right ventricle (RV), atresia of the main pulmonary artery (PA), a large ventricular septal defect (functionally univentricular physiology), cor triatriatum, patent arteriosus ductus, enlarged coronary sinus due to PLSVC, and patent left Blalock-Taussig shunt. Cardiac catheterization showed a large ventricular septal defect, aorta arising anteriorly from the RV, pulmonary atresia, patent Blalock shunt, left PA: 8.3 mm, right PA: 7.8 mm, and stenosis at the level of pulmonary bifurcation.

sus ductus, enlarged coronary sinus due to PLSVC, and patent left Blalock-Taussig shunt. Cardiac catheterization showed a large ventricular septal defect, aorta arising anteriorly from the RV, pulmonary atresia, patent Blalock shunt, left PA: 8.3 mm, right PA: 7.8 mm, and stenosis at the level of pulmonary bifurcation. We decided to pursue a strategy of staged univentricular palliation. The patient underwent surgery. Median sternotomy was performed and anterior pericardium was harvested. The aorta arose from RV and was positioned anteriorly with respect to the small PA. There was a small right superior vena cava (RSVC), a large left superior vena cava (LSVC; left dominance), and right inferior vena cava (IVC). There was a small ductus arteriosus (5 mm) and a patent left Blalock-Taussig shunt. Preoperative PA pressure was 16 mm Hg.

as positioned anteriorly with respect to the small PA. There was a small right superior vena cava (RSVC), a large left superior vena cava (LSVC; left dominance), and right inferior vena cava (IVC). There was a small ductus arteriosus (5 mm) and a patent left Blalock-Taussig shunt. Preoperative PA pressure was 16 mm Hg. Cardiopulmonary bypass was established with cannulation of the aorta, each SVC, and the right atrium (RA). Under mild hypothermia, the aorta was clamped. Myocardial protection was achieved by antegrade infusion of Custodiol solution. After opening the RA, we removed the left atrial membrane after partially excising the interatrial septum. The pulmonary arterial trunk was separated from the RV and ligated at the proximal end. We then augmented the PA confluence. Each SVC was separated from the RA and dissected over as much length as possible. The length of the LSVC was not sufficient to allow it to be anastomosed directly to the RSVC. We created a tubular vascular graft using rolled pericardium on a Hegar dilator (number 12) using 6.0 Prolene suture (Figure 1). We used fresh pericardium to perform this technique. The pericardial tube graft was anastomosed in an end-to-end fashion to the LSVC, effectively lengthening it. The other end of the pericardial tube graft was anastomosed in a side-to-side fashion to the divided RSVC, over a length of approximately 2 cm, creating a Y-shaped confluence of the RSVC and LSVC and enabling anastomosis of an effectively enlarged “common SVC” to PA confluence using continuous 7.0 Prolene suture line (Figure 2).

end of the pericardial tube graft was anastomosed in a side-to-side fashion to the divided RSVC, over a length of approximately 2 cm, creating a Y-shaped confluence of the RSVC and LSVC and enabling anastomosis of an effectively enlarged “common SVC” to PA confluence using continuous 7.0 Prolene suture line (Figure 2). Figure 1. Vascular graft formation by rolled pericardium. Figure 2. A, Anastomosis between superior vena cava (SVC) confluence and pericardium. B, Diagram of the completed reconstruction. The patient had stable hemodynamics during the early postoperative period, and the satisfactory status of the cavopulmonary anastomosis was confirmed by echocardiography, showing no pressure gradient through the anastomoses and good filling of both PA branches. The arterial blood pressure was 95/50, central venous pressure 15 mm Hg, SpO2 85% at fractional inspired oxygen (Fio 2) of 0.8, heart rate 118 b/m, sinus rhythm. Unfortunately, the patient had left phrenic nerve palsy resulting in inability to wean from the ventilator and we decided to perform diaphragmatic plication. Despite this intervention, the patient died due to respiratory infections after 20 days. Case 2 A three-year-old male patient weighing 12 kg was admitted to our hospital due to cyanosis. On physical examination, the patient had cleft lip and cleft palate, cyanosis of the lips and the nail beds, and clubbed fingernails.

The patient had stable hemodynamics during the early postoperative period, and the satisfactory status of the cavopulmonary anastomosis was confirmed by echocardiography, showing no pressure gradient through the anastomoses and good filling of both PA branches. The arterial blood pressure was 95/50, central venous pressure 15 mm Hg, SpO2 85% at fractional inspired oxygen (Fio 2) of 0.8, heart rate 118 b/m, sinus rhythm. Unfortunately, the patient had left phrenic nerve palsy resulting in inability to wean from the ventilator and we decided to perform diaphragmatic plication. Despite this intervention, the patient died due to respiratory infections after 20 days. Case 2 A three-year-old male patient weighing 12 kg was admitted to our hospital due to cyanosis. On physical examination, the patient had cleft lip and cleft palate, cyanosis of the lips and the nail beds, and clubbed fingernails. Echocardiography revealed transposition of the great arteries, double outlet right ventricle (DORV), complete atrioventricular canal defect, and severe stenosis of the pulmonary valve. Cardiac catheterization confirmed transposition of the great arteries, atrioventricular canal defect, valvar and subvalvar pulmonary stenosis, and dextroposition of the heart.

of the great arteries, double outlet right ventricle (DORV), complete atrioventricular canal defect, and severe stenosis of the pulmonary valve. Cardiac catheterization confirmed transposition of the great arteries, atrioventricular canal defect, valvar and subvalvar pulmonary stenosis, and dextroposition of the heart. Observations at surgery included large aorta, aortic arch, equal-sized RSVC and LSVC, right-sided IVC, well-developed right and left branch pulmonary arteries, and a small PA trunk. Heparin was administered (1 mg/kg). The operation was performed without cardiopulmonary bypass support. The right SVC was divided just above its entrance to the RA, and it was anastomosed to the right PA in an end-to-side fashion, as in a conventional bidirectional Glenn operation. The LSVC was divided just above its entrance to the left atrium. A rolled pericardial tube graft was used to elongate the LSVC, making it possible to anastomose this lengthened LSVC to the RSVC in an end-to-side fashion, with the vascular graft of rolled pericardium situated in front of the aorta (Figure 3A and B). Figure 3. A, The rolled pericardium is used to effectively elongate the left SVC. B, Diagram of anastomoses. Ao indicates aorta; IVC, inferior vena cava; PA, pulmonary artery; RA, right atrium; RV, right ventricle; SVC, superior vena cava.

Observations at surgery included large aorta, aortic arch, equal-sized RSVC and LSVC, right-sided IVC, well-developed right and left branch pulmonary arteries, and a small PA trunk. Heparin was administered (1 mg/kg). The operation was performed without cardiopulmonary bypass support. The right SVC was divided just above its entrance to the RA, and it was anastomosed to the right PA in an end-to-side fashion, as in a conventional bidirectional Glenn operation. The LSVC was divided just above its entrance to the left atrium. A rolled pericardial tube graft was used to elongate the LSVC, making it possible to anastomose this lengthened LSVC to the RSVC in an end-to-side fashion, with the vascular graft of rolled pericardium situated in front of the aorta (Figure 3A and B). Figure 3. A, The rolled pericardium is used to effectively elongate the left SVC. B, Diagram of anastomoses. Ao indicates aorta; IVC, inferior vena cava; PA, pulmonary artery; RA, right atrium; RV, right ventricle; SVC, superior vena cava. The patient had an uneventful postoperative course. A well-functioning modified bidirectional cavopulmonary anastomosis was demonstrated in this patient by two-dimensional echocardiographic evaluation at 3, 6, and 12 months after surgery. Cardiac catheterization was performed in anticipation of Fontan completion and showed no gradient across the anastomosis. He underwent a completion Fontan procedure at 14 months after the modified bidirectional Glenn shunt.

is patient by two-dimensional echocardiographic evaluation at 3, 6, and 12 months after surgery. Cardiac catheterization was performed in anticipation of Fontan completion and showed no gradient across the anastomosis. He underwent a completion Fontan procedure at 14 months after the modified bidirectional Glenn shunt. Discussion The presence of bilateral superior caval veins is commonly seen in different forms of congenital heart diseases, but the precise pathophysiology of this systemic venous abnormality is still unknown. Anatomical studies showed that when bilateral superior caval veins are present, the different diameters of two veins may result in imbalanced distribution of blood flow after bilateral bidirectional Glenn operation and can lead to obstruction and/or thrombus formation within the SVC or at the anastomosis to the ipsilateral PA.2,3 In addition, there is a concern that in the presence of unequal flow from the left and right SVC to the pulmonary arteries, blood flow from IVC may not symmetrically distribute to the right and left PA after Fontan procedure. The asymmetrical distribution of pulmonary blood flow can result in pulmonary arteriovenous malformations, which may increase the risk of thrombus formation and may affect the growth of the pulmonary arteries.3 In 2002, Amodeo and Di Donato studied the energetics and the distribution of pulmonary blood flow after Glenn operation and Fontan procedure using models based on magnetic resonance imaging and cardiac catheterization images from 110 patients who underwent Fontan procedure.6

ffect the growth of the pulmonary arteries.3 In 2002, Amodeo and Di Donato studied the energetics and the distribution of pulmonary blood flow after Glenn operation and Fontan procedure using models based on magnetic resonance imaging and cardiac catheterization images from 110 patients who underwent Fontan procedure.6 In order to optimize blood flow from bilateral superior caval veins to the pulmonary arteries, many techniques for making anastomosis have been proposed in the literature. Vida et al in 2006 reported the case of a 14-month-old in whom a 5-mm Gore-Tex conduit was interposed between two SVCs in bidirectional Glenn operation. Despite the good reported results, there is a concern about the possibility of conduit occlusion as well as the fixed size of the prosthesis in relation to subsequent growth of the patient.4

rted the case of a 14-month-old in whom a 5-mm Gore-Tex conduit was interposed between two SVCs in bidirectional Glenn operation. Despite the good reported results, there is a concern about the possibility of conduit occlusion as well as the fixed size of the prosthesis in relation to subsequent growth of the patient.4 Many studies have shown that disorders in pulmonary blood flow from SVCs and the distribution of blood flow from the IVC are major factors in the prognosis of the results of Glenn operation and Fontan procedure. In 2007, Amodeo and Di Donato proposed a novel method that can possibly resolve the problems in patients with bilateral SVCs. The purpose of this method is to convert two superior caval veins into a unifocalized confluence in order to increase the blood flow at the bidirectional anastomosis between SVC and pulmonary arteries. A single unifocalized connection at the level of the PA confluence is considered to be more effective than two small bidirectional peripheral anastomoses with competing blood flow, a situation which may lead to asymmetrical pulmonary blood flow.3 Technically speaking, this modification of bidirectional cavopulmonary anastomosis may be affected by the aortic arch. The large aorta can lead to stretching, compression, or even blockage of the confluence of two SVCs. This is a potential disadvantage of this technique. The advantage of this method is that the location of two SVCs can be adjusted more to the right than to the left, to allow enough compensation for the anticipated location of the extracardiac conduit placed between IVC and PA at the time of a future completion Fontan procedure.3

otential disadvantage of this technique. The advantage of this method is that the location of two SVCs can be adjusted more to the right than to the left, to allow enough compensation for the anticipated location of the extracardiac conduit placed between IVC and PA at the time of a future completion Fontan procedure.3 With the efforts to resolve the problems of unequal anatomical diameters of right and left SVC leading to reduced blood flow, which may cause blood retention, thrombus formation, and unfavorable growth of two PAs, in 2014 Nakanishi proposed a new surgical technique named “unifocalization of bilateral SVCs.”2 Nakanishi proposed the hypothesis that in unifocalization of bilateral superior caval veins, the blood flow from the joined SVCs can be distributed to two lungs in a manner similar to unilateral Glenn operation. This new technique was only applied on patients with the IVC and the larger SVC on the same side. If the smaller SVC and IVC were on the same side, the conventional bidirectional Glenn operation was performed. If the RSVC and LSVC were equal in diameters, the author chose the SVC on the same side as the IVC to be the main vessel. The disadvantage of this technique is the length of SVC may still be too short to make end-to-side cavo-caval anastomosis.2

e same side, the conventional bidirectional Glenn operation was performed. If the RSVC and LSVC were equal in diameters, the author chose the SVC on the same side as the IVC to be the main vessel. The disadvantage of this technique is the length of SVC may still be too short to make end-to-side cavo-caval anastomosis.2 Our first patient had the smaller SVC on the same side as the IVC with dominant PLSVC, and pulmonary atresia diagnosed preoperatively. The goal of achieving cavopulmonary anastomosis and concomitant enlargement of the PA confluence led us to choose the method of creating the single confluence of the right and left superior caval veins at the level of the cavopulmonary anastomosis with the PA confluence. However, during dissection, we found that the LSVC was not long enough to make the confluence with the RSVC. To overcome this problem, we used the rolled pericardium technique. This method not only formed a long enough vascular graft but also provided enough material to augment the PA confluence without using artificial materials. Our second patient had equally sized LSVC and RSVC and right-sided IVC. The very large left-sided aorta covered over the LSVC. To dissect the LSVC, we had to retract the aorta to the right. We did not use cardiopulmonary bypass support, since this patient only needed cavopulmonary anastomosis. After dividing and mobilizing the LSVC, we found that it was not long enough to make the end-to-side anastomosis with right SVC. The rolled pericardium graft technique was chosen to resolve this problem.

the right. We did not use cardiopulmonary bypass support, since this patient only needed cavopulmonary anastomosis. After dividing and mobilizing the LSVC, we found that it was not long enough to make the end-to-side anastomosis with right SVC. The rolled pericardium graft technique was chosen to resolve this problem. The use of different materials with the aim to lengthen a SVC to facilitate the cavopulmonary anastomosis has been reported by many authors.4,5 Artificial conduits or rolled xenograft pericardium can be used. In our study, the application of rolled autologous pericardium to increase the length of LSVC seemed to be the most reasonable choice. It can be used to augment the PA confluence. Conclusion It is necessary to have more time to investigate the effectiveness of methods for cavopulmonary anastomosis in bidirectional Glenn operations for patients with bilateral superior caval veins before being widely applied to many patients. However, at least theoretically, the surgical solution to convert two SVC anastomoses into one venous confluence appears to have substantial advantages over other reported techniques. We hypothesize that such techniques may promote symmetrical distribution of pulmonary blood flow which could contribute to favorable growth of the branch pulmonary arteries and minimize the chance of thrombus formation. The lack of adequate length of SVC for direct anastomosis can be solved by using a graft made by rolled pericardium.

Introduction Single ventricle (SV) with unrestrictive pulmonary blood flow and (potential) systemic ventricular outflow tract obstruction remains a rare and challenging anomaly with heterogeneous underlying anatomy.1 Classical examples are double inlet left ventricle (DILV) or tricuspid atresia (TA) with associated transposition of the great arteries (TGA). In these patients, systemic blood flow needs to pass through a ventricular septal defect (VSD), sometimes called bulboventricular foramen, to a rudimentary subaortic outlet chamber. Such a VSD tends to become restrictive and has been linked to initial pulmonary artery banding (PAB) due to ventricular hypertrophy2–5 or after volume unloading surgery (Glenn or Fontan) due to altered ventricular geometry,6 particularly when associated with aortic arch (AA) obstruction.7,8 Subaortic or AA obstruction has been widely recognized as a risk factor for a good Fontan outcome. Therefore, early relief of such obstruction is mandatory. However, initial neonatal management in this patient group is not uniform, varying from an aggressive neonatal Norwood (NW) or Damus-Kaye-Stansel (DKS) approach to a more conservative initial PAB ± AA plasty with potential later relief of developed subaortic stenosis (SAS) concomitant with the Glenn or Fontan procedure. Subaortic stenosis can occur because of (1) a restrictive subaortic chamber, (2) a restrictive VSD, (3) subaortic fibrous tissue/membrane, or (4) a combination of these causes. Subaortic stenosis can be relieved directly by means of VSD and/or subaortic chamber enlargement or indirectly by means of a NW or DKS procedure, in which the actual subaortic obstruction is bypassed by connecting the pulmonary trunk with the aorta. In addition, some centers have used a palliative arterial switch procedure for this purpose.9 The direct approach has been mostly replaced by the DKS/NW procedure because of the risk of heart block, development of recurrent SAS, and ventricular dysfunction due to a ventriculotomy, thus making these patients potentially less suitable for a Fontan pathway.

a palliative arterial switch procedure for this purpose.9 The direct approach has been mostly replaced by the DKS/NW procedure because of the risk of heart block, development of recurrent SAS, and ventricular dysfunction due to a ventriculotomy, thus making these patients potentially less suitable for a Fontan pathway. However, the DKS/NW procedure can result in left pulmonary artery compression by the neoaortic root or in semilunar valve insufficiency due to altered root geometry. Moreover, these operations carry higher mortality and morbidity when performed in a neonatal period.1,10,11 There are only few reports on long-term outcomes in SV patients who underwent direct relief of SAS by enlargement of VSD and/or subaortic outflow chamber.12 In this study, we aim to revisit this partially abandoned concept by describing and analyzing our experience with the direct approach for relief of SAS in SV patients.

e are only few reports on long-term outcomes in SV patients who underwent direct relief of SAS by enlargement of VSD and/or subaortic outflow chamber.12 In this study, we aim to revisit this partially abandoned concept by describing and analyzing our experience with the direct approach for relief of SAS in SV patients. Patients and Methods We conducted a retrospective study in children with SV and unobstructed pulmonary blood flow, in whom initial or later SAS had been relieved directly via VSD and/or subaortic chamber enlargement. The study was approved by the ethics committee, and individual consent for the study was waived due to its retrospective study design. At our institution, this is the preferred approach in children who do not strictly need a NW or DKS procedure. In patients with a small aortic valve, ascending aorta, or otherwise unsuitable anatomy (eg, unbalanced atrioventricular septal defect (AVSD)), an initial DKS/NW procedure was considered the only available option. Data from medical records were analyzed for demographic, pre-, peri-, and post-operative characteristics (Table 1). Primary end points included mortality, achievement of Fontan circulation, and adequacy of SAS relief. Secondary end points were the incidence of recurrent SAS, heart block, subaortic chamber aneurysm, ventricular function, and semilunar valve function. Table 1. General Characteristics and Demographic Data.a

Patients and Methods We conducted a retrospective study in children with SV and unobstructed pulmonary blood flow, in whom initial or later SAS had been relieved directly via VSD and/or subaortic chamber enlargement. The study was approved by the ethics committee, and individual consent for the study was waived due to its retrospective study design. At our institution, this is the preferred approach in children who do not strictly need a NW or DKS procedure. In patients with a small aortic valve, ascending aorta, or otherwise unsuitable anatomy (eg, unbalanced atrioventricular septal defect (AVSD)), an initial DKS/NW procedure was considered the only available option. Data from medical records were analyzed for demographic, pre-, peri-, and post-operative characteristics (Table 1). Primary end points included mortality, achievement of Fontan circulation, and adequacy of SAS relief. Secondary end points were the incidence of recurrent SAS, heart block, subaortic chamber aneurysm, ventricular function, and semilunar valve function. Table 1. General Characteristics and Demographic Data.a Characteristic Results Age relief SAS 7.2 months (10 days to 4.7 years) Age at first operation 34 days (0 days to 11.2 months) Sex (M/F) (12/11) Weight (kg) 3.3 (2.2-4.3) Left SV morphology 20 (91) Aortic arch obstruction 12 (55) Pulmonary artery banding 23 (100) Glenn 20 (91) Age Glenn 8.3 months (3.1 months to 3.7 years) Fontan 19 (83) Age Fontan (years) 3.1 (1.4-5.3) Mortality 4 (17) Follow-up 15.6 years (34 days to 26.3 years) Abbreviations: SAS, subaortic stenosis; SV, single ventricle

SV morphology 20 (91) Aortic arch obstruction 12 (55) Pulmonary artery banding 23 (100) Glenn 20 (91) Age Glenn 8.3 months (3.1 months to 3.7 years) Fontan 19 (83) Age Fontan (years) 3.1 (1.4-5.3) Mortality 4 (17) Follow-up 15.6 years (34 days to 26.3 years) Abbreviations: SAS, subaortic stenosis; SV, single ventricle a Values are reported as median (range) or as percentage. Data are presented as medians with ranges where appropriate. The presence of SAS was established by means of preoperative echocardiography or catheterization and/or peroperative evaluation of subaortic outflow tract by the surgeon. Any measurable gradient, VSD/aortic valve ratio <1.0, or restriction at level of the subaortic chamber was considered relevant and formed an indication for relief using VSD and/or subaortic chamber enlargement.

echocardiography or catheterization and/or peroperative evaluation of subaortic outflow tract by the surgeon. Any measurable gradient, VSD/aortic valve ratio <1.0, or restriction at level of the subaortic chamber was considered relevant and formed an indication for relief using VSD and/or subaortic chamber enlargement. Surgical Technique All operations were performed via median sternotomy using cardiopulmonary bypass and cold antegrade cardioplegia. The approach of VSD enlargement could be via a right ventricle (RV) ventriculotomy, right atrium, pulmonary valve, or aorta and was based on surgeon preference. Ventricular septal defect and/or subaortic chamber enlargement was performed according to the technique firstly described by Cheung et al.13 The VSD was enlarged superiorly or apically toward the obtuse margin of the heart (Figure 1A, dashed line). When a ventricular approach was used, the vertical incision was closed with a patch (Figure 1B). A RV ventriculotomy was the preferred approach in cases in which preoperative echocardiography showed a small subaortic chamber, thereby allowing thorough inspection and muscle resection in this area.

gure 1A, dashed line). When a ventricular approach was used, the vertical incision was closed with a patch (Figure 1B). A RV ventriculotomy was the preferred approach in cases in which preoperative echocardiography showed a small subaortic chamber, thereby allowing thorough inspection and muscle resection in this area. Figure 1. A, View of a patient with double inlet left ventricle (DILV) + transposition of great arteries (TGA), where systemic blood flow needs to pass through the VSD into a rudimentary right ventricle (RV) toward the aorta. After a ventriculotomy, the ventricular septal defect (VSD) can be enlarged safely in the superior and apical directions toward the obtuse margin. The safe margin (dashed line) and suspected course of the conduction system (circles) are indicated. B, After enlarging the VSD, the subaortic chamber is enlarged with a patch.

er a ventriculotomy, the ventricular septal defect (VSD) can be enlarged safely in the superior and apical directions toward the obtuse margin. The safe margin (dashed line) and suspected course of the conduction system (circles) are indicated. B, After enlarging the VSD, the subaortic chamber is enlarged with a patch. Results In the period 1989 to 2016, 23 children (median age: 7.4 months, range: 10 days to 5.5 years) received direct relief of SAS at our institution. Patient demographics and general characteristics are shown in Table 1. Primary diagnoses are listed in Table 2. All patients underwent staged approach SV palliation with initial PAB ± AA repair (n = 11, 48%) and required primary relief of SAS at presentation or somewhere down the pathway (Figure 2, Table 3). Nine (39%) patients underwent direct relief of SAS as part of the first operation (median age: 29 days, range: 7 days to 9.5 months), nine (39%) patients at the time of Glenn procedure (median age: 7.4 months, range: 4.4 months to 3.4 years), one between Glenn and Fontan procedures (age: 3.2 years), four (17%) concomitant with Fontan procedure (median age: 3.6 years, range: 1.4-4.8 years), and in one patient 3.5 years after Fontan completion. Concomitant surgery at the time of first stage palliation was atrioventricular valve (AVV) repair in one. Surgical procedures concomitant with the Glenn procedure were Blalock-Taussig shunt (n = 4), PAB (n = 2), AVV repair (n = 2), AVV closure (n = 1), and a NW procedure (n = 1). Management of the pulmonary valve at the time of the Glenn procedure was variable and based on surgeon preference. Concomitant surgery at the time of Fontan was AVV repair (n = 3), AVV closure (n = 3), DKS (n = 1), and pulmonary artery augmentation (n = 1).

pair (n = 2), AVV closure (n = 1), and a NW procedure (n = 1). Management of the pulmonary valve at the time of the Glenn procedure was variable and based on surgeon preference. Concomitant surgery at the time of Fontan was AVV repair (n = 3), AVV closure (n = 3), DKS (n = 1), and pulmonary artery augmentation (n = 1). Table 2. Primary Diagnosis.a Diagnosis Number DILV + TGA 17 (9) TA + TGA 2 (1) ccTGA ± TA 3 (1) DORV 1 (1) Abbreviations: ccTGA, congenital corrected transposition of great arteries; DORV, double outlet right ventricle; DILV, double inlet left ventricle; TA, tricuspid atresia; TGA, transposition of great arteries a Number of patients with aortic arch obstruction are shown in parenthesis. Figure 2. Clinical pathway of single ventricle (SV) patients. Table 3. Clinical Pathway.a

Diagnosis Number DILV + TGA 17 (9) TA + TGA 2 (1) ccTGA ± TA 3 (1) DORV 1 (1) Abbreviations: ccTGA, congenital corrected transposition of great arteries; DORV, double outlet right ventricle; DILV, double inlet left ventricle; TA, tricuspid atresia; TGA, transposition of great arteries a Number of patients with aortic arch obstruction are shown in parenthesis. Figure 2. Clinical pathway of single ventricle (SV) patients. Table 3. Clinical Pathway.a No. Diagnosis Timing of SAS Relief SAS Mechanism/Gradient RE-SAS Mechanism/Gradient First Procedure SAS Relief Glenn Interstage II-III SAS Relief SAS Relief Fontan Post-Fontan SAS Relief 1 DILV + TGA PAB VSD↑ + SAC↑ - - - VSD ratio <1.0 - 2 DILV + TGA PAB + VSD↑ + SAC↑ - RE-VSD↑ - - 12 mm Hg 36 mm Hg 3 TA + TGA PAB + AAR + SAC↑ - - VSD↑ - 23 mm Hg - 4 ccTGA + TA PAB VSD↑ - - - VSD ratio 0.7 - 5 DILV + TGA PAB + MVP + AAR + VSD↑ + SAC↑ - - RE-VSD↑ - VSD ratio <1.0, 3 mm 8 mm Hg 6 DILV + TGA PAB + AAR + SAC↑ - - No Fontan - Restrictive SAC - 7 DILV + TGA PAB - - VSD↑ RE-VSD↑ 20 mm Hg 40 mm Hg 8 DILV + TGA PAB + AAR + VSD↑ - - - - VSD ratio <1.0, 3 mm - 9 DILV + TGA PAB VSD↑ - RE-VSD↑ RE-RE-VSD↑ VSD ratio <1.0 35 mm Hg, RE-RE 46 mm Hg 10 TA + TGA PAB VSD↑ - - - 30 mm Hg - 11 DILV + TGA + TA PAB No Glenn - VSD↑ - 10 mm Hg, 9 mm - 12 DILV + TGA PAB + VSD↑ + SAC↑ NW - - - VSD ratio <1.0 - 13 DILV + TGA PAB VSD↑ - - - 5 mm Hg - 14 DILV + TGA PAB + AAR + VSD↑ - - - - VSD ratio 0.7 - 15 DILV + TGA PAB + AAR + VSD↑ + SAC ↑ - - - RE-VSD↑ VSD ratio 0.6 70 mm Hg 16 ccTGA + TA PAB VSD↑ - - - VSD ratio <1.0, 6 mm - 17 DILV + TGA PAB SAC↑ - - - Restrictive SAC - 18 DILV + TGA + MA PAB + AAR - VSD↑ No Fontan - 21 mm Hg - 19 DILV + TGA PAB + AAR VSD↑ + SAC↑ - Scheduled - 25 mm Hg - 20 DILV + TGA PAB + AAR + SAC↑ No Glenn - No Fontan - Restrictive SAC - 21 ccTGA PAB + AAR - - VSD↑ Fibrous tunnel VSD ratio <1.0 49 mm Hg 22 DORV PAB + AAR VSD↑ - - - 20 mm Hg - 23 DILV + TGA PAB - - DKS VSD↑ VSD ratio <1.0 Abbreviations: AAR, aortic arch repair; ccTGA, congenital corrected transposition of the great arteries; DILV, double inlet left ventricle; DKS, Damus-Kaye-Stansel; DORV, double outlet right ventricle; MA, mitral atresia; MVP, mitral valve plasty; NW, Norwood; PAB, pulmonary artery banding; SAC↑, subaortic chamber enlargement; SAS, subaortic stenosis; TA, tricuspid atresia; TGA, transposition of the great arteries; VSD↑, ventricular septal defect enlargement; RE-VSD, redo VSD enlargement.

uble outlet right ventricle; MA, mitral atresia; MVP, mitral valve plasty; NW, Norwood; PAB, pulmonary artery banding; SAC↑, subaortic chamber enlargement; SAS, subaortic stenosis; TA, tricuspid atresia; TGA, transposition of the great arteries; VSD↑, ventricular septal defect enlargement; RE-VSD, redo VSD enlargement. a VSD ratio = VSD/aortic valve diameter ratio. Approach of VSD enlargement was via a RV ventriculotomy in 11 (42%) procedures, right atrium in 4 (15%), pulmonary artery in 1 (4%), and aorta in (38%) 10 procedures. In ten patients, a patch was used to enlarge the subaortic (RV) outlet chamber (six xenopericard, one Gore-Tex, two pulmonic homograft, and in one unknown). Two patients underwent both the direct and indirect approaches and were excluded from primary end point analysis as long-term outcome in these patients was not considered to reflect one particular approach. One patient had DILV + TGA, a severely hypoplastic AA and SAS, and underwent primary palliation with a hybrid NW procedure + VSD and subaortic chamber enlargement, thereby successfully deferring the NW procedure to stage II. The other patient received a DKS procedure concomitant with Fontan completion for relief of SAS. After 1.9 years, the DKS was closed and the VSD enlarged due to severe pulmonary regurgitation.

ith a hybrid NW procedure + VSD and subaortic chamber enlargement, thereby successfully deferring the NW procedure to stage II. The other patient received a DKS procedure concomitant with Fontan completion for relief of SAS. After 1.9 years, the DKS was closed and the VSD enlarged due to severe pulmonary regurgitation. Primary Outcome Complete follow-up was available for all patients for a median period of 15.6 years (range: 34 days to 26.3 years). Overall mortality was 17% (4/23), of which two were perioperative deaths. One patient died two days after initial PAB + AA repair + subaortic chamber enlargement due to unexpected cardiac arrest without known cause. One patient died one day after Fontan completion + VSD enlargement. This patient was reoperated for fenestration creation because of low cardiac output, which was complicated by cardiac arrest with severe neurological damage. One child with DILV + TGA, mitral atresia, and dysplastic tricuspid valve underwent two tricuspid valve repairs and eventually tricuspid valve replacement with postoperative poor ventricular function. This child died of end-stage heart failure at an age of 4.3 years. One late death occurred 16.5 years after Fontan completion after a large cerebral vascular accident. Eighteen out of 21 (excluding the two patients with both direct and indirect approaches) patients underwent Fontan completion and one (86%) patient is scheduled for completion.

tage heart failure at an age of 4.3 years. One late death occurred 16.5 years after Fontan completion after a large cerebral vascular accident. Eighteen out of 21 (excluding the two patients with both direct and indirect approaches) patients underwent Fontan completion and one (86%) patient is scheduled for completion. At latest follow-up, there was complete relief of SAS in 17 of 18 patients. In one patient, there was a stable laminar flow with a velocity of 3.0 m/s over the VSD and is monitored frequently with echocardiography. Secondary Outcome Twenty out of 23 patients received a total of 26 VSD enlargement procedures (including reobstructions), of which 3 were in the neonatal period. Two patients underwent neonatal subaortic chamber enlargement only because of adequately sized VSD. These five neonatal patients presented with adequately sized aortic valve and ascending aorta with restriction at the level of the VSD (n = 3) and/or subaortic chamber (n = 4). One (4%) patient developed iatrogenic complete heart block for which a pacemaker was implanted. Five (50%) patients developed an aneurysm at site of the ventriculotomy, which was a true aneurysm in one and false in four and was repaired concomitantly during Fontan procedure in two and in a separate operation in three patients. The patient with a true aneurysm had an aneurysm of the entire subaortic ventricle, not only of the xenopericard patch, and required plication of the subaortic RV at the time of Fontan.

in one and false in four and was repaired concomitantly during Fontan procedure in two and in a separate operation in three patients. The patient with a true aneurysm had an aneurysm of the entire subaortic ventricle, not only of the xenopericard patch, and required plication of the subaortic RV at the time of Fontan. At latest follow-up, no patients had more than trivial aortic regurgitation. Ventricular function was good in 17, mildly impaired in 1, and moderately impaired in 1 patient (NT-proBNP median: 151, range: 50-591 ng/L [n = 12], ASAT 28, range: 17-62 U/L, ALAT 27, range: 14-46 U/L [n = 15]). All surviving patients are in New York Heart Association class 1-2. In total, seven (30%) patients developed reobstruction after direct SAS relief. Six reobstructions occurred at VSD level and reobstruction was caused by a fibrous subaortic tunnel in one patient. One patient developed a second reobstruction. These reobstructions were addressed in one patient 19 days after Glenn, in three patients during Fontan procedure, and in three patients 8.9, 4.5, and 5.7 years after the Fontan procedure. The patient with a second reobstruction underwent VSD enlargement 11.0 years after Fontan completion. The risk of reobstruction was evident up until 10 years after first direct relief of SAS (Figure 3). We could not identify any risk factors for reobstruction. There was no association between age at first relief of SAS or presence of AA hypoplasia/coarctation with risk of reobstruction.

0 years after Fontan completion. The risk of reobstruction was evident up until 10 years after first direct relief of SAS (Figure 3). We could not identify any risk factors for reobstruction. There was no association between age at first relief of SAS or presence of AA hypoplasia/coarctation with risk of reobstruction. Figure 3. Freedom from recurrent subaortic stenosis (SAS) after first relief of SAS via the direct approach. The hazard of recurrent SAS is highest in the first few years, but can occur up till ten years after first relief of SAS. Comment The management of SV patients with (potential) SAS is subjected to different possible strategies. The conservative option is an initial PAB ± AA repair. Early and interim results were, however, not optimal with poor candidacy for Fontan palliation.14,15 Therefore, others use a more aggressive neonatal modified NW/DKS procedure, thereby relieving any (potential) SAS at the first operation. However, this approach is associated with substantial perioperative mortality, reaching up to 27%.15 Contemporary results of the NW procedure for non-hypoplastic left heart syndrome report considerable early mortality, with 15.4% in-hospital mortality and 14.5% interstage mortality. Although left dominant morphologies (such as DILV + TGA and TA + TGA) have superior results compared with right dominant morphologies, early mortality still approaches 13% to 20% and can be higher in low-volume centers.10,11,16

nsiderable early mortality, with 15.4% in-hospital mortality and 14.5% interstage mortality. Although left dominant morphologies (such as DILV + TGA and TA + TGA) have superior results compared with right dominant morphologies, early mortality still approaches 13% to 20% and can be higher in low-volume centers.10,11,16 When SAS occurs early, most surgeons will perform a DKS or NW procedure. Another possible option is direct relief of SAS by enlarging the VSD and/or subaortic outflow chamber, but this method has mainly been reserved for patients with pulmonary regurgitation, pulmonary stenosis, or for patients who present late after Fontan palliation in which pulmonary valves are closed and therefore are unsuitable for DKS repair.1 This is based on relative good experiences with the DKS procedure and the possible disadvantages associated with the direct approach. Because most centers abandoned this technique, there is no study describing the long-term outcome of the direct approach in a relatively contemporary patient cohort. In a large single-center series, Lan et al described the outcome in 140 patients with DILV/TA + TGA (median follow-up: 7.7 years). Overall survival was 71%, with 24% mortality in patients without relief of SAS and 32% in patients with relief of SAS (n = 95). Fifty-one patients received DKS and 44 patients underwent VSD enlargement with a similar cumulative mortality percentage of 33% versus 32%. Overall Fontan completion/suitability was 76%.12

s). Overall survival was 71%, with 24% mortality in patients without relief of SAS and 32% in patients with relief of SAS (n = 95). Fifty-one patients received DKS and 44 patients underwent VSD enlargement with a similar cumulative mortality percentage of 33% versus 32%. Overall Fontan completion/suitability was 76%.12 Recently, long-term results (median: 17 years) of a large cohort of 152 DILV + TGA and 59 TA + TGA patients from the Australia and New Zealand Fontan registry have been reported. In their experience, 5% of DILV patients had SAS at presentation and 44% developed SAS over time. Overall, 91% of DILV patients and 60% of TA patients proceeded to Fontan completion.17 In our selected cohort of mainly DILV + TGA patients with (developed) SAS, survival was 83% and the Fontan procedure was achieved or scheduled in 86% of patients. Differences with Lan et al can be explained by an earlier era of that report and the inclusion of a higher amount of TA + TGA patients, which have significantly higher mortality than DILV + TGA patients.12,17 Our results are however comparable to the results of Franken et al, where overall (with and without development of SAS) 91% of DILV + TGA patients proceeded to Fontan completion.17

report and the inclusion of a higher amount of TA + TGA patients, which have significantly higher mortality than DILV + TGA patients.12,17 Our results are however comparable to the results of Franken et al, where overall (with and without development of SAS) 91% of DILV + TGA patients proceeded to Fontan completion.17 When SAS is present at birth, many centers adopt a strategy of a neonatal NW or DKS procedure to relieve SAS and to concomitantly address AA hypoplasia when present. Subaortic stenosis at birth was associated with increased risk of death in the report by Franken et al. Association with a particular neonatal surgical strategy for relief of SAS could however not be established. In our experience, we had five patients who underwent neonatal relief of SAS by VSD ± subaortic chamber enlargement (n = 3) or by subaortic chamber enlargement only (n = 2). One of these latter two patients suffered from cardiac arrest 3 days postoperatively without known cause. The other patient underwent VSD enlargement at the time of Fontan procedure due to recurrence of SAS at VSD level. The other three patients underwent an uncomplicated Fontan procedure and are doing well without reobstruction at an age of 10.9, 7.5, and 15.6 years.

diac arrest 3 days postoperatively without known cause. The other patient underwent VSD enlargement at the time of Fontan procedure due to recurrence of SAS at VSD level. The other three patients underwent an uncomplicated Fontan procedure and are doing well without reobstruction at an age of 10.9, 7.5, and 15.6 years. In one patient with a severely hypoplastic AA and SAS, we used a hybrid NW + VSD enlargement to delay the NW procedure to the next stage with good result. In our opinion, neonatal VSD enlargement can be an acceptable option in selected patients and can thereby possibly delay DKS or NW surgery to a later stage when necessary with potential better outcome. Furthermore, we used the direct approach in one patient after DKS failure because of severe pulmonary regurgitation 1.9 years after Fontan with good outcome. The risk of complete heart block has been reported to be between 0% and 34%.12,18–22 In our experience, one (4%) patient received a pacemaker because of complete heart block out of 26 performed VSD enlargement procedures. Ventricular function was well preserved in all but one of our patients and we did not encounter cases with new aortic valve insufficiency.

n reported to be between 0% and 34%.12,18–22 In our experience, one (4%) patient received a pacemaker because of complete heart block out of 26 performed VSD enlargement procedures. Ventricular function was well preserved in all but one of our patients and we did not encounter cases with new aortic valve insufficiency. The rate of reobstruction reported in the literature varies between 11% and 44%4,18,20,21,23 and can occur many years after Fontan completion.24 This is in line with our experience, where 30% developed recurrent SAS, of which one patient had a second reobstruction. In four of these patients, reobstruction occurred 8.9, 5.7, 4.5, and 11 years after Fontan completion indicating the continuous hazard of reobstruction. Regular follow-up with echocardiography is therefore important for these patients even many years after Fontan repair.

of which one patient had a second reobstruction. In four of these patients, reobstruction occurred 8.9, 5.7, 4.5, and 11 years after Fontan completion indicating the continuous hazard of reobstruction. Regular follow-up with echocardiography is therefore important for these patients even many years after Fontan repair. In our experience, however, reoperation for reobstruction could be done without associated mortality. To address the high rate of reobstruction, one option would be to enlarge the VSD more aggressively, which in our opinion can be done without increased risk of heart block when strictly performed in the described direction.13 Another option would be to always do a DKS at the time of Glenn or Fontan procedure. As a large proportion of DILV + TGA patients develop SAS, which is reported to occur in up to 44% of patients over time,17 some centers have adopted a strategy where every patient receives a DKS procedure, independent of the occurrence of SAS, at stage II or during Fontan.17,25,26 However, although short- and midterm outcome of the pulmonary valve (neoaortic) after DKS is good,25–27 the long-term 15- to 20-year outcome is mainly unknown.28 Whether a universal DKS/NW, and therefore inherently an unnecessary DKS in some patients, is the best option for these patients is not known. To keep the option of future DKS open in these patients, we prefer to spare the pulmonary valve at the time of Glenn procedure, especially in patients with previous SAS.

known.28 Whether a universal DKS/NW, and therefore inherently an unnecessary DKS in some patients, is the best option for these patients is not known. To keep the option of future DKS open in these patients, we prefer to spare the pulmonary valve at the time of Glenn procedure, especially in patients with previous SAS. In a study by Jahangiri et al, no formation of aneurysms was reported for a series of 23 patients undergoing VSD and subaortic outflow chamber enlargement.18 Karl et al reported the formation of an aneurysm in one out of two patients in which they enlarged the VSD via ventriculotomy.8 In our experience, five out of ten patients developed an aneurysm, of which four were false aneurysms and one was a true aneurysm. Two of these patients required a separate operation to address this. We speculate that the reason for these false aneurysms, besides technically, can be the cause of the abnormal blood flow inherent to this technique, where blood has to travel from the left ventricle through the VSD in an abnormal angle. Collision of this blood flow with the patch can be the cause of the increased risk of false aneurysms and is an area for future research with 4-D flow magnetic resonance imaging.

f the abnormal blood flow inherent to this technique, where blood has to travel from the left ventricle through the VSD in an abnormal angle. Collision of this blood flow with the patch can be the cause of the increased risk of false aneurysms and is an area for future research with 4-D flow magnetic resonance imaging. Indications for the Direct Approach In our opinion, there are 3 groups of SV patients with (potential) SAS: (1) patients with small aortic valve and/or ascending aorta ± SAS, for which a NW/DKS procedure is the only available option, (2) patients with an adequately sized subaortic pathway, who undergo primary palliation with PAB ± AA repair at our center, and (3) patients with adequately sized aortic valve and ascending aorta but with restriction at the level of the VSD or subaortic chamber. For these patients, we prefer to perform PAB ± AA repair with direct relief of SAS when present or perform direct relief of SAS in a later stage when it develops. Limitations and Strengths This is a retrospective, single-center study which includes patients over a long time frame and is limited by its small sample size. Strength of this study is its relative contemporary cohort and long-term follow-up of SV patients who underwent direct relief of SAS.

Indications for the Direct Approach In our opinion, there are 3 groups of SV patients with (potential) SAS: (1) patients with small aortic valve and/or ascending aorta ± SAS, for which a NW/DKS procedure is the only available option, (2) patients with an adequately sized subaortic pathway, who undergo primary palliation with PAB ± AA repair at our center, and (3) patients with adequately sized aortic valve and ascending aorta but with restriction at the level of the VSD or subaortic chamber. For these patients, we prefer to perform PAB ± AA repair with direct relief of SAS when present or perform direct relief of SAS in a later stage when it develops. Limitations and Strengths This is a retrospective, single-center study which includes patients over a long time frame and is limited by its small sample size. Strength of this study is its relative contemporary cohort and long-term follow-up of SV patients who underwent direct relief of SAS. Conclusion In our experience, the relief of SAS via the direct approach appears to result in good long-term outcome in terms of survival and Fontan suitability. When enlarging the VSD strictly via the described approach, risk of complete heart block is low. The ventricular function remained well preserved and we did not encounter new aortic valve insufficiencies. However, risk of reobstruction is high and can occur many years after Fontan completion. Furthermore, we encountered a substantial risk of formation of an aneurysm at site of the patch and the ventriculotomy.

ow. The ventricular function remained well preserved and we did not encounter new aortic valve insufficiencies. However, risk of reobstruction is high and can occur many years after Fontan completion. Furthermore, we encountered a substantial risk of formation of an aneurysm at site of the patch and the ventriculotomy. In conclusion, direct relief of SAS appears to offer a relatively simple solution for a select group of SV patients with SAS. Although this approach is associated with a substantial risk of patch aneurysms and reobstruction, this could be treated without associated mortality and long-term results appear to be good. Optimization of our current approach by more aggressive VSD enlargement or prophylactic DKS at the time of Glenn or Fontan may reduce the burden of these complications and is subject to future study. The definitive role of VSD and/or subaortic chamber enlargement within the clinical pathway of SV patients with (potential) SAS is in our opinion therefore not limited to patients with pulmonary valve problems and can defer DKS/NW, if necessary, to an older age. Authors’ Note: The authors had full control of the design of the study, methods used, outcome parameters, analysis of data, and production of the written report. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. Abbreviations and Acronyms AAaortic arch

Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article. Abbreviations and Acronyms AAaortic arch DILVdouble inlet left ventricle DKSDamus-Kaye-Stansel DORVdouble outlet right ventricle NWNorwood PABpulmonary artery banding RVright ventricle SASsubaortic stenosis SVsingle ventricle TAtricuspid atresia TGAtransposition of the great arteries VSDventricular septal defect

Better understanding of the anatomy and optimization of surgical techniques have improved early clinical outcome of surgical correction of complete atrioventricular septal defects (CAVSDs). However, several studies report a high risk of reoperation (up to 10%).1 Technical performance score (TPS), a tool developed to determine technical adequacy of congenital cardiac repairs, has been shown to be an important predictor of both early and midterm outcomes across a wide range of congenital cardiac procedures.2 A previous study has shown that the presence of residual lesions before discharge, as measured by TPS, accurately identifies patients requiring postdischarge reinterventions (PD-RI).3 However, in this previous study, overall TPS was used to determine the association between TPS and PD-RI. The association between individual subcomponents of TPS and PD-RI was not investigated. The aim of this study is to determine which subcomponents of TPS best predict PD-RI.

g postdischarge reinterventions (PD-RI).3 However, in this previous study, overall TPS was used to determine the association between TPS and PD-RI. The association between individual subcomponents of TPS and PD-RI was not investigated. The aim of this study is to determine which subcomponents of TPS best predict PD-RI. A review of consecutive patients with balanced CAVSDs who were operated on at a tertiary care center between January 2000 and March 2016 was performed with institutional review board approval. Demographic, echocardiographic, and follow-up data were obtained. Primary complete repair of CAVSD was considered the index operation. Only patients discharged alive without heart transplantation were analyzed. Patients with associated major intracardiac anomalies or partial/transitional AVSD were excluded. Postoperative TPS was determined as previously reported3 based on predischarge echocardiographic findings and clinical status at discharge from the index operation. The TPS for CAVSD repair included the following subcomponents: size of residual atrial septal defect, size of residual ventricular septal defect (VSD), severity of right and left atrioventricular valve (AVV) stenosis and regurgitations, status of the patent ductus arteriosus, and status of the conduction system. Each subcomponent was assigned a score of class 1 (optimal, trivial, or no residua), class 2 (adequate and minor residua), or class 3 (inadequate and major residua), based on specific echocardiographic criteria.3 The final TPS was based on the subcomponent scores and was class 1 if all subcomponents received a class 1 score, class 2 if one or more of the subcomponents were class 2 but none were class 3, and class 3 if any of the subcomponents were class 3. Any unplanned surgical or catheter-based reintervention for residua in the anatomic area repaired during the CAVSD operation or the need for permanent pacemaker (PPM) placement prior to discharge from index CAVSD surgery resulted in a class 3 (inadequate) score. The outcome variable, PD-RI, was defined as surgical or catheter-based reinterventions that occurred following discharge from index CAVSD surgery on anatomic areas repaired at CAVSD operation, including placement of PPM.

PM) placement prior to discharge from index CAVSD surgery resulted in a class 3 (inadequate) score. The outcome variable, PD-RI, was defined as surgical or catheter-based reinterventions that occurred following discharge from index CAVSD surgery on anatomic areas repaired at CAVSD operation, including placement of PPM. Forward selection was used to develop a multivariable Cox regression model of baseline patient factors associated with time to PD-RI; P < .05 was required for retention in the final model. Qualitative TPS and its subcomponents were then added to this model. Hazard ratios were estimated with 95% confidence intervals. Statistical analysis was performed with SAS version 9.4 (SAS Institute Inc, Cary, North Carolina). There were 344 patients included in the analysis. There were 211 (61%) females, 67 (19%) were premature, and 305 (89%) had some form of genetic anomaly. The median age at operation was 3.2 months (interquartile range [IQR], 2.4-4.2 months). There were 34 (10%) PD-RI. Trisomy 21 and concomitant procedure were associated with PD-RI. After adjusting for these factors, among the subcomponents, left AVV stenosis and regurgitation, right AVV regurgitation, residual VSD, and abnormal conduction at discharge were significantly associated with PD-RI (Table 1). Table 1. Multivariable Model, Subcomponents of TPS, and Postdischarge Reinterventions (n = 344).a

Trisomy 21 and concomitant procedure were associated with PD-RI. After adjusting for these factors, among the subcomponents, left AVV stenosis and regurgitation, right AVV regurgitation, residual VSD, and abnormal conduction at discharge were significantly associated with PD-RI (Table 1). Table 1. Multivariable Model, Subcomponents of TPS, and Postdischarge Reinterventions (n = 344).a Number Number of RI (%) Hazard Ratio (95% Confidence Interval) P Value C Index Multivariable model, not considering TPS (n = 344) Trisomy 21 291 21 (7.2%) 0.29 (0.14-0.60) 0.001 0.679 Concomitant procedurea 29 7 (24.1%) 2.71 (1.14-6.44) 0.02 For each model below, the results shown are adjusted for trisomy 21 and concomitant procedure Final TPS 0.793 1 54 2 (3.7%) 1.00 2 217 13 (6.0%) 1.21 (0.27-5.41) .79 3 60 17 (28.3%) 5.60 (1.28-24.5) .02 ASD 0.694 No defect (<2 mm): 1 244 22 (9.0%) 1.00 Small defect (2-3 mm): 2 24 4 (16.7%) 2.91 (0.94-8.98) .06 Not reported 63 6 (9.5%) - VSD 0.701 No defect (<2 mm): 1 178 13 (7.3%) 1.00 Small defect (2-3 mm): 2 135 15 (11.1%) 1.33 (0.63-2.80) .45 RI or defect >3 mm: 3 4 2 (50.0%) 12.2 (2.51-59.6) .002 Not reported 14 2 (14.3%) - LAVV stenosis 0.746 No stenosis (mean <3 mm Hg): 1 255 16 (6.3%) 1.00 Mild stenosis (mean 3-6 mm Hg): 2 26 9 (34.6%) 3.12 (1.24-7.82) .02 RI or ≥moderate stenosis (>6 mm Hg): 3 4 1 (25.0%) 11.0 (1.37-88.0) .02 Not reported 66 6 (9.1%) - LAVV regurgitation 0.795 No/trivial regurgitation: 1 129 8 (6.2%) 1.00 Mild regurgitation: 2 164 12 (7.3%) 0.84 (0.34-2.09) .71 RI or ≥moderate regurgitation: 3 34 11 (32.4%) 3.98 (1.55-10.2) .004

Not reported 14 2 (14.3%) - LAVV stenosis 0.746 No stenosis (mean <3 mm Hg): 1 255 16 (6.3%) 1.00 Mild stenosis (mean 3-6 mm Hg): 2 26 9 (34.6%) 3.12 (1.24-7.82) .02 RI or ≥moderate stenosis (>6 mm Hg): 3 4 1 (25.0%) 11.0 (1.37-88.0) .02 Not reported 66 6 (9.1%) - LAVV regurgitation 0.795 No/trivial regurgitation: 1 129 8 (6.2%) 1.00 Mild regurgitation: 2 164 12 (7.3%) 0.84 (0.34-2.09) .71 RI or ≥moderate regurgitation: 3 34 11 (32.4%) 3.98 (1.55-10.2) .004 Not reported 4 1 (25.0%) - RAVV stenosis 0.696 No stenosis (mean <3 mm Hg): 1 153 17 (11.1%) 1.00 Mild stenosis (mean 3-6 mm Hg): 2 11 2 (18.2%) 0.47 (0.08-2.86) .42 Not reported 167 13 (7.8%) - RAVV regurgitation 0.732 No/trivial regurgitation: 1 135 7 (5.2%) 1.00 Mild regurgitation: 2 165 17 (10.3%) 1.47 (0.60-3.61) .40 RI or ≥moderate regurgitation: 3 19 5 (26.3%) 3.89 (1.23-12.3) .02 Not reported 12 3 (25.0%) - PDA Cannot estimate; no RI in category 3 No PDA: 1 24 2 (8.3%) RI or PDA open: 3 3 0 (0%) Not reported 304 30 (9.9%) Conduction 0.699 Normal conduction: 1 319 27 (8.5%) 1.00 Permanent pacemaker: 3 4 3 (75.0%) 6.68 (2.00-22.2) .002 Not reported 8 1 (12.5%) - Abbreviations: ASD, atrial septal defect; LAVV, left atrioventricular valve; PDA, patent ductus arteriosus; RAVV, right atrioventricular valve; RI, reintervention; TPS, technical performance score; VSD, ventricular septal defect. Significant values (p < 0.05) appear in bold. a Included: left ventricular outflow tract tissue resection, pulmonary artery band takedown, repair of coarctation of the aorta, other. Not included: patent ductus arteriosus ligation.

Not reported 8 1 (12.5%) - Abbreviations: ASD, atrial septal defect; LAVV, left atrioventricular valve; PDA, patent ductus arteriosus; RAVV, right atrioventricular valve; RI, reintervention; TPS, technical performance score; VSD, ventricular septal defect. Significant values (p < 0.05) appear in bold. a Included: left ventricular outflow tract tissue resection, pulmonary artery band takedown, repair of coarctation of the aorta, other. Not included: patent ductus arteriosus ligation. A potential limitation of this study is that it represents a single center’s experience with TPS using retrospective data with its inherent problems of missing and incomplete data, although only 4% of our patient population had missing qualitative echocardiographic data at discharge. We demonstrated the ability of TPS to predict PD-RI in patients who underwent CAVSD repair. Residual left and right AVV regurgitation and abnormal conduction at discharge were among the subcomponents strongly associated with PD-RI. Thus, TPS may aid clinicians in identifying children at higher risk of future reinterventions who may benefit from more frequent follow-up. Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding: The author(s) received no financial support for the research, authorship, and/or publication of this article.