Developmental aspects of atrioventricular septal defects

International Journal of Cardioloa,

18 (1988) 65-18

65

Elsevier

IJC 00610

Developmental aspects of atrioventricular septal defects Arnold

C.G. Wenink and Juan-Carlos

Zevallos

Department of Anatomv and Embryology, Unwersity of Leiden, Leiden, The Netherlands: Department of Paediatric Cardiology, Universi[v of Padooti, Padooa, Ita!, (Received

Wenink

14 April 1987: revision accepted

ACG, Zevallos J-C. Developmental

fects. Int J Cardiol

23 June 1987)

aspects of atrioventricular

septal de-

1988;18:65-78.

Three human embryos with an atrioventricular septal defect were studied. Their morphology was compared with that of 67 autopsy specimens, in which particular attention was paid to the septal attachments of the bridging leaflets. The malformed embryos showed deficiency of the inlet component of the ventricular septum. They had distinct superior and inferior bridging leaflets, which were nearly completely muscular. Myocardial undermining had taken place at two independent sites but had not been able to lead to the formation of a valve of mitral morphology. Normal delamination of myocardium to form the leaflets could not continue directly below the aortic root because the rim of the inlet septum had a more apical position. From this, we conclude that the deficiency of the inlet septum is the cause of the typical morphology of the left valve in these hearts. The role of endocardial cushion tissue is probably restricted to glueing together myocardial structures, thus determining the variable septal attachment of the bridging leaflets in atrioventricular septal defect. Key words: malformation

Embryology:

Endocardial

cushion

defect;

Atrioventricular

canal

Introduction Understanding of the developmental full knowledge of normal embryology.

nature of congenital malformations requires Differences in the way in which a particular

Correspondence to: Dr. A.C.G. Wenink. Dept. Postbus 9602, 2300 RC Leiden. The Netherlands.

0167-5273/88/$03.50

of Anaromy

0 1988 Elsevier Science Publishers

and

Embryology.

B.V. (Biomedical

Division)

Wassenaarseweg

62.

66

malformation may be explained, however, depend on two different factors. Firstly, there is no general agreement on the details of cardiac development and this may cause different explanations for the same anomaly. But secondly, any explanation is greatly influenced by the way in which the anomaly is described. A change of accent in the description of the pathologic anatomy may easily cause a change in the developmental approach. These remarks are particularly pertinent in the setting of atrioventricular septal defect. Developmental considerations have led to terms such as “persistent common atrioventricular canal” [l] and “endocardial cushion defect” [2]. The counterpart of these terms may be found in the name atrioventricular septal defect [3], which is based more on the description of the mature malformation [4]. Recent work on cardiac development [5,6] has prompted more detailed ideas on the morphogenesis of atrioventricular septal defect [7,8]. Studies on normal valve development [9] led us to further speculations on this malformation [lo]. Despite the great value of studies on normal embryology, it is highly desirable to observe developmental stages of congenital malformations. The reason underscoring this study, therefore, is the fact that three human embryos having an atrioventricular septal defect recently became available to us. So as to set our observation of the embryos in the appropriate morphologic context, we have reviewed the autopsy specimens of the Leiden collection having deficiencies of atrioventricular septation. We have then correlated the findings in the malformed embryos to the morphology of the mature malformation.

Materials and Methods Sixty-seven specimens with atrioventricular septal defect were studied, 52 of them with a common atrioventricular orifice and 15 with separate right and left atrioventricular orifices. The material which forms the basis of our views on normal embryonic development comprises human embryos between 3.6 and 25 mm crown-rump length. This material has been described before [ll]. Three malformed human embryos were investigated. Their crown-rump lengths were 23, 29, and 35 mm, respectively (approximate age 7 to 8 weeks). They were sectioned serially in the transverse plane (section thickness 10 pm) and stained with haematoxylin and eosin. In one case, several graphic reconstructions were made using the method customary in our laboratory [12].

Results Autopsy Specimens The most prominent features were the deficiency anterior position of the aortic orifice. Measurements are well established [13] and were not repeated in this although examples of the anomaly with one common sented with an extremely great deficiency (Fig. 1) and

of the inlet septum and the of the inlet septal deficiency study. Significantly, however, atrioventricular orifice precases with separate right and

Fig. I. Atrioventricular septai defect with a common atrioventricular orifice. left ventricular view. There is considerable deficiency of the ventricular septum (vs). The superior bridging leaflet (sbl) is free floating. An abnormal papillary muscle (pm) is reaching to the aortic orifice (ao). alm = anterolateral muscle bundle.

Fig. 2. Atrioventricular septal defect with separate right and left valve orifices: left ventricular view. Deficiency of the ventricular septum (vs) is much less than in the case illustrated in Fig. 1. An anterolateral muscle bundle (aim) is wedged between the superior bridging leaflet (sbl) and the aortic orifice (ao).

Fig. 3. Atrioventricular septal defect with separate valve orifices, left ventricular view. The ventricular septum (vs) is very deficient. A mass of dysplastic leaflet tissue (asterisk) on the septal rim connects the superior (sbl) and inferior (ibl) bridging leaflets. alm = anterolateral muscle bundle.

left orifices had much less significant deficiency (Fig. 2) there were also specimens with two valve orifices and a greatly deficient inlet septum (Fig. 3). The criterion of having two orifices was fulfilled whenever fibrous tissue adherent to the inferior bridging leaflet (or belonging to this leaflet) was continuous with similar tissue

TABLE

1

Septal attachments

of the bridging

leaflets in 67 hearts with atrioventricular Usual atrial arrangement Two-valve (N=15) No.

orifices

septal defect. Atria1 isomerism

Common (N=35)

orifice

Common (N=17)

orifice

%

No.

%I

No.

I

100

4 15 16

11 43 46

10 3 4

59 18 23

100

15 15 5

43 43 14

14 1 2

82 6 12

Inferior bridging leaflet Free-floating Cordal attachment Membranous attachment

0 0 15

Superior bridging leaflet Free-floating . Cordal attachment Membranous attachment

0 0 15

69

derived from the superior bridging leaflet. In general, the morphology of the bridging leaflets was more variable than that of the septal structures. In all 15 cases with separate right and left valve orifices. the inferior and superior bridging leaflets showed membranous attachments to the rim of the ventricular septum. Such attachments were less frequent in the cases with a common valve orifice. In the 35 cases with usual atria1 arrangement (“situs solitus”), cordal attachments were seen in nearly half of the cases, the other half being constituted by membranous attachment of the inferior leaflet and complete detachment of the superior leaflet (“free floating”). The inferior leaflet was seldom free-floating, whereas a membranous attachment was an infrequent finding for the superior leaflet. Seventeen cases with a common valve orifice had atria1 isomerism (“situs ambiguus”). Of these, the great majority had a free floating superior bridging leaflet. More than half of them also had a free-floating inferior leaflet. In the remaining cases, membranous and cordal attachments to the ventricular septal rim were equally distributed. These findings on valve attachment are summarized in Table 1. Malformed Embryos There were no conspicuous differences between the three cases. The oldest one, however, showed slightly more differentiation, making the recognition of the boundaries of cushion tissue masses somewhat more difficult. Although the exact position of the valve orifices cannot be judged fully from serial sections, it is noted here that, in one case (29 mm), the aortic and pulmonary orifices were seen in the same section, indicating the possibility of a relatively high position of the aortic valve. In all three embryos, the rim of the inlet component of the ventricular septum (recognized because it carried the inferior endocardial cushion) was distant from the aortic orifice. Both the superior and inferior cushions could be well recognized, and they were widely spaced. The superior cushion adhered to the anterior portion of the ventricular septum (the primary septum), narrowing the subaortic outflow tract (Fig. 4). Below this attachment, the cushion became detached and extended freely from the left ventricle to the right. It did not protrude very much into the right ventricle. Instead, it was continuous here with another, more parietal, mass of cushion tissue. Towards the periphery and towards the apex, the superior cushion was continuous in both ventricles with free myocardial strands (Fig. 5). More caudally, the inferior cushion was adherent to the inlet septum. This cushion, too, was continuous with the myocardial trabeculations of both ventricular cavities. The 29 mm case was reconstructed graphically. The atria1 view of the atrioventricular junction (Fig. 6) showed the characteristic features of an atrioventricular septal defect. The ventricular septum had a free rim which was distant from the aortic valve. Superior and inferior bridging leaflets were directly recognized within the junction. Their central portions were composed of cushion tissue but their more peripheral parts were muscular. The cushion tissue was directly adherent to the ventricular septum. The two bridging leaflets were very close in the left ventricle and a parietal leaflet was not recognized between the two. A view of the right ventricle was difficult to obtain because it was considerably smaller than the left.

Fig. 4. Transverse section of one of the malformed embryos. The superior endocardial cushion (SC) is attached to the rim of the ventricular septum (vs), making the left ventricular outflow tract (ot) a narrow canal. The superior cushion is also in contact with a right lateral cushion (Ic) which hangs out above the right ventricle

(rv). ra = right atrium.

Fig. 5. Lower section of the same embryo, to show the muscular superior bridging leaflet (arrows) hanging out into the left ventricle (Iv). The inferior endocardial cushion (ic) is on top of the inlet septum (vs) or, rather, the posterior ventricular wall. This section is too caudal to show the right ventricle itself. la = left atrium;

ra = right atrium.

71

Fig. 6. Graphic reconstruction of the same malformed embryo. dorsal view after removal of the atria1 part. The superior bridging leaflet (arrow) is a large structure of which only the central portion consists of cushion tissue (SC). This adheres to the rim of the ventricular septum (vs), and it borders on the left ventricular outflow tract (ot). The inferior cushion (ic) forms the central part of the inferior bridging leaflet and adheres to the septum as well. No left parietal leaflet is present between the two bridging leaflets. p = pulmonary orifice.

The superior bridging leaflet was clearly seen in the left ventricular view of this heart (Fig. 7). Its small central portion consisted of cushion tissue and was stuck to the anterior part of the ventricular septum (the primary septum). Its larger part was completely myocardial and was fused with the free wall of the left ventricle, being very close to a similar myocardial structure which was part of the inferior bridging leaflet. After removal (in a graphical way) of the superior bridging leaflet (Fig. S), the free rim of the ventricular septum could be seen. The scooped-out appearance was striking, the inlet component of the septum being positioned much more apically than the aortic root. The inferior bridging leaflet was adherent in its central part to the inlet septum. This central part was made up of cushion tissue. Its muscular part coursed freely towards the free wall. Although it formed a separate mass from the superior leaflet, the space between the two was too small to accommodate a separate parietal leaflet.

72

Fig. 7. Graphic reconstruction. view from the left after dissection of the left parietal wall. The central portion of the superior bridging leaflet (sbl) consists of the superior endocardial cushion (SC) and borders the sub-aortic outflow tract (ao). vs = ventricular septum; sp = subpulmonary outflow tract; aa = ascending aorta; la = left atrium.

Discussion The morphology of the ventricular septum was uniform in our autopsy specimens, showing the rim of the inlet septum to be at a distance from the aortic root. This is a hallmark of atrioventricular septal defects [4]. Recently, the uniform inlet septal deficiency in cases with either separate right and left atrioventricular orifices or in those with a common orifice has been confirmed by careful measurements [ 131. In this study, no significant difference was found between the groups with separate or common orifices, a finding at variance with the measurements reported by others [16]. Recent review of the Pittsburgh data, however, has revealed that the degree of scooping is indeed greater in the hearts with a common atrioventricular orifice, although the inlet dimensions (as measured along the posterior left ventricular wall) do not differ significantly from the cases with two separate orifices [17]. The striking variation in this defect, therefore, is caused by the morphology of the leaflets of the atrioventricular valves rather than in different arrangement of the

73

Fig. 8. The same reconstruction as in Fig. 7. after dissection of the superior bridging leaflet. The free rim of the ventricular septum (vs) can be seen to be far away from the aortic orifice (ao). ic = inferior endocardial cushion.

ventricular mass. This variation, in our material, was comparable to that reported by Penkoske et al. [13]. We found a tendency towards firm attachment of both bridging leaflets to the septal rim in cases with separate valve orifices (100% in our cases), and a higher tendency for such an attachment of the inferior rather than the superior leaflet in the cases with a common orifice. The fact that valve morphology causes specific variation while septal morphology does not is in agreement with previous hypotheses on the morphogenesis of this anomaly [7,10] which were based on observations in normal embryos [5,6,9]. These observations lead to the conclusion that ventricular septation precedes formation of the atrioventricular valves. Our evidence indicates that the septum is composed of inlet and primary (or apical trabecular) septal components. A much smaller contribution is provided by the outlet septum. In the normal heart, the antero-superior part of the inlet septal component reaches to the aortic root. This is the site where, normally, the superior and inferior cushions fuse. They do not, however, contribute any material to the septum [ll]. Valve formation is then a subsequent process which involves delamination of the inner layer of myocardium of the wall of the inlet segment, including the inlet

14

septum itself. The endocardial cushions, which as indicated do not materially contribute to the septum, do not contribute materially to the valve leaflets either [9]. In the left ventricle, this delamination process starts directly below the aortic orifice. It is the appropriate fusion of the septal components -which enables the

Fig. 9. Diagrams to show the delamination process in the left ventricle to form the aortic leaflet of the mitral valve. Cushion tissue is not taken into account. A. The outflow tract towards the aorta (ao) is narrow, being completely bordered by the primary fold (pf), (the myocardial fold between the inlet and outlet segments of the ventricles). This fold, the apical trabecular septum (ts) and the inlet septum (is) meet directly below the aortic orifice (ao). m = mitral orifice. B. Myocardial delamination has started, and the subaortic outflow tract expands. C. Delamination is completed. It has continued from the apical trabecular septum (ts) to the inlet septum (is). The aortic leaflet of the normal mitral valve has two different components: an anterior part, which was initially continuous with the apical trabecular septum and a posterior part derived from the inlet septum. The two components are glued together by cushion tissue (not shown). Their fusion line points towards the aortic root.

delamination to continue from the primary to the inlet components of the septum (Fig. 9). The result is the formation of the aortic (“anterior”) leaflet of the mitral valve, which is derived from two different sources. These two components meet at the aortic root.

Fig. 10. Diagrams to show the delamination process in the case of our atrioventricular septal defect. A. Before delamination. The primary fold (pf) has no relationship with the inlet septum (is). The fusion point of inlet septum and apical trabecular septum (ts) is far away from the aortic orifice (ao). B. Delamination has started, but cannot continue from one septal component onto the other. C. Valve formation is completed. Separate anterior and posterior delaminations have led to the formation of separate superior (sbl) and inferior (ibl) bridging leaflets. Note that the gap between these leaflets does not point towards the aortic root, but towards the site of fusion between the apical trabecular (ts) and inlet (is) components

of the ventricular

septum.

76

When there is deficiency of the inlet septal component, however, the fusion point near the aortic root is missing. The inlet septal rim then has a more apical position and causes the “scoop” of the septum. In this setting, delamination of septal myocardium cannot smoothly continue from one to the other septal component. Instead, there is formation of two separate myocardial flaps. This arrangement rules out the possibility of formation of a left valve of normal “mitral” morphology (Fig. 10). The developmental approach to the morphology of atrioventricular septal defects as espoused above, therefore, accounts well for the observed myocardial nature of the valve leaflets. It does not explain their variable disposition. The way in which they are or are not attached to the septum (or to each other, thus creating so-called “ostium primum” defect with separate right and left orifices) is poorly explained by ” variations in the delamination process”. We believe that our observations on the malformed embryos throw light on the pathogenetic process. Beyond doubt, the embryos examined are examples of an atrioventricular septal defect with a common atrioventricular orifice. First of all, they confirm that the valve leaflets have a muscular origin. They show superior and inferior bridging leaflets which mainly consist of myocardium. But secondly, they show how the superior bridging leaflet is attached to the anterosuperior part of the septum (the primary septal component). This is because the superior cushion adheres to the septum, creating a relatively narrow sub-aortic outflow tract (Fig. 4). Similarly, the septal attachment of the inferior bridging leaflet is seen at the site of the inferior cushion. Therefore, we submit that the cushions, which do not substanstages, this finding tially contribute to the valves, may act as “glue ” in embryonic endorsing the hypothesis advanced by Van Mierop [14]. The degree to which they perform this adhesive function may then explain the varying morphology of the atrioventricular valves in atrioventricular septal defect. It might well be that a large degree of septal scooping (a great distance between the lower rim of the defect and the aortic root), makes it more difficult for the cushion tissue to glue the leaflets to the septal rim. Thus, cases with a common annulus would tend to have free-floating bridging leaflets [13] (see Table 1). Furthermore, since there is no significant difference between the inlet dimensions of the two groups [17], then, with increased scooping, there would be a greater tendency to have a free-floating superior than a free-floating inferior bridging leaflet [13] (see Table 1). In all cases, the bridging leaflets tend to reach each other at the lowermost rim of the defect. This is the site most distant from the aortic root. Thus, when viewed from the atrium, the gap between the two leaflets will never point at the aortic root. This gap is totally different from an isolated cleft in the aortic (“anterior”) leaflet of the normal mitral valve [18]. Such isolated clefts do point at the aortic root [lo]. Since, in the case of an isolated cleft, there is no inlet septal deficiency, it is not possible to explain its existence on the basis of abnormal septation. Instead, it is reasonable to hypothesize that the cushion tissue may have been unable to glue together the two components of the normal aortic leaflet of the mitral valve, one derived from delamination of the inlet and the other from the primary component of the septum. As we have argued previously [lo], the isolated cleft of the mitral valve would then be the only known example of a real “endocardial cushion defect”.

Finally, the reconstruction of one of the embryos revealed absence of any mural leaflet for the left valve in atrioventricular septal defect (Fig. 6). In the mature form of the anomaly, this leaflet is known to be smaller than the corresponding leaflet of the normal mitral valve [13]. It might well be that this leaflet would have developed in later stages of our embryo, had it been older, and then have remained relatively small. Valve formation takes a long time. The best example of this fact is the septal leaflet of the tricuspid valve, which is completed only in late fetal or even early postnatal stages [15].

Acknowledgements The malformed embryos were kindly made available by Dr. J.P.M. Geraerdts (Department of Human Genetics, Leiden), in a collaborative research project with the Department of Gynaecology and Obstetrics, the Center for Clinical Genetics and the Department of Anatomy, all in Leiden.

References 1 Wakai CS, Edwards JE. Developmental and pathologic considerations in persistent common atrioventricular canal. Mayo Clin Proc 1956;31:487-499. 2 Van Mierop LHS, Alley RD, Kausel HW, Stranahan A. The anatomy and embryology of endocardial cushion defects. J Thorac Cardiovasc Surg 1962;43:71-83. 3 Becker AE, Anderson RH. Atrioventricular septal defects: what’s in a name? J Thorac Cardiovasc Surg 1982;83:461-469. 4 Piccoli GP, Gerlis LM, Wilkinson JL, Lozsadi K, Macartney FJ, Anderson RH. Morphology and classification of atrioventricular defects. Br Heart J 1979;42:621-632. 5 Wenink ACG. Embryology of the ventricular septum. Separate origin of its components. Virchows Archiv (Path01 Anat) 1981;390:71-79. 6 Wenink ACG, Gittenberger-de Groot AC. Left and right ventricular trabecular patterns. Consequence of ventricular septation and valve development. Br Heart J 1982;48:462-468. 7 Wenink ACG. The ventricular septum in hearts with an atrioventricular defect. In: Wenink ACG. Oppenheimer-Dekker A, Moulaert AH, eds. The ventricular septum of the heart. Boerhaave Series Vol 21. The Hague, Boston, London: Leiden University Press, 1981;131-138. 8 Wenink ACG, Gittenberger-de Groot AC, Oppenheimer-Dekker A, et al. Septation and valve formation: similar processes dictated by segmentation. In: Nora JJ, Takao A, eds. Congenital heart disease, causes and processes. Mount Kisco, New York: Futura Publishing Comp, 1984;513-529. 9 Wenink ACG, Gittenberger-de Groot AC. Embryology of the mitral valve. Int J Cardiol 1986;11:75-84. 10 Wenink ACG, Gittenberger-de Groot AC, Brom AG. Developmental considerations of mitral valve anomalies. Int J Cardiol 1986;11:85-98. 11 Wenink ACG, Gittenberger-de Groot AC. The role of atrioventricular endocardial cushions in the septation of the heart. Int J Cardiol 1985;8:25-41. 12 Tinkelenberg J. Graphic reconstruction, microanatomy with a pencil. Audivis Media Med 1979:2:102-106. 13 Penkoske PA, Neches WH, Anderson RH, Zuberbuhler JR. Further observations on the morphology of atrioventricular septal defects. J Thorac Cardiovasc Surg 1985;90:611-622. 14 Van Mierop LHS. Embryology of the atrioventricular canal region and pathogenesis of endocardial cushion defects. In: Feldt RH, ed. Atrioventricular canal defects. Philadelphia, London, Toronto: WB Saunders Company, 1976;1-12.

78 15 Allwork SP, Anderson RH. Developmental anatomy of the membranous part of the ventricular septum in the human heart. Br Heart J 1979;41:275-280. 16 Draulans-Noe HAY, Wenink ACG, Quaegebeur JM. Ventricular septal deficiency in atrioventricular (AV) septal defect - a correlation with AV-valve morphology. Pediat Cardiol 1984;5:227-228. 17 Ebels T. Het linker ventrikeluitstroomgebied bij atrioventriculaire septum defecten, chirurgische anatomie. In: Gittenberger-de Groot AC, Ottenkamp J. Wenink ACG. eds. Morfologie en kliniek van afwijkingen van hart en grote vaten op de kinderleeftijd. Boerhaave commissie voor Postacademisch Onderwijs in de Geneeskunde, Rijksuniversiteit Leiden, 1987: pp. 97-109. 18 Anderson RH, Zuberbuhler JR. Penkoske PA. Neches WH. Of clefts. commissures and things. J Thorac

Cardiovasc

Surg 1985;90:605-610.