http://pch.sagepub.com/ Congenital Heart Surgery World Journal for Pediatric and http://pch.sagepub.com/content/1/3/364 The online version of this article can be found at: DOI: 10.1177/2150135110380239 2010 1: 364 World Journal for Pediatric and Congenital Heart Surgery Richard Van Praagh Normally and Abnormally Related Great Arteries : What Have We Learned? Published by: http://www.sagepublications.com On behalf of: World Society for Pediatric and Congential Heart Surgery can be found at: World Journal for Pediatric and Congenital Heart Surgery Additional services and information for http://pch.sagepub.com/cgi/alerts Email Alerts: http://pch.sagepub.com/subscriptions Subscriptions: http://www.sagepub.com/journalsReprints.nav Reprints: http://www.sagepub.com/journalsPermissions.nav Permissions: by Jose-Antonio Quibrera on October 30, 2010 pch.sagepub.com Downloaded from
23
Embed
World Journal for Pediatric and Congenital Heart Surgery
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
http://pch.sagepub.com/
Congenital Heart SurgeryWorld Journal for Pediatric and
http://pch.sagepub.com/content/1/3/364The online version of this article can be found at:
DOI: 10.1177/2150135110380239
2010 1: 364World Journal for Pediatric and Congenital Heart SurgeryRichard Van Praagh
Normally and Abnormally Related Great Arteries : What Have We Learned?
Published by:
http://www.sagepublications.com
On behalf of:
World Society for Pediatric and Congential Heart Surgery
can be found at:World Journal for Pediatric and Congenital Heart SurgeryAdditional services and information for
Normally and Abnormally Related GreatArteries: What Have We Learned?
Richard Van Praagh, MD1
AbstractThe conus arteriosus or infundibulum was the site of the major cardiovascular evolutionary and developmental adaptation thatmade possible air-breathing and permanent land-living for vertebrates, including mammals such as ourselves. The subarterial conalfree walls perform an embryonic aortic switch procedure by 35 to 44 days of age in utero, based on growth of the left-sidedsubpulmonary conal free wall and resorption of the right-sided subaortic conal free wall, i.e., complete right-left asymmetry inthe development of the subarterial conal free walls. There is only one way of doing the developmental aortic aortic switchprocedure right (one way in situs solitus, and its mirror-image in situs inversus), and there are many ways of doing it wrong,resulting in the conotruncal anomalies. The proximal or apical part of the conus arteriosus, the septal band, was the motherof the right ventricular sinus (the lung pump). The conus transformed the single (systemic) circulation of fish into our double(systemic and pulmonary) circulations. The right ventricle (RV) is only about 36% as old as the left ventricle (LV). Mostcongenital heart disease involves anomalies of the more recently developed RV, congenital heart disease being the mostfrequent anomaly in liveborn children — almost 1 percent (0.8%).
Keywordscardiac anatomy/pathological anatomy, congenital heart disease, embryology, great vessel anomaly
Submitted May 21, 2010; Accepted July 8, 2010.Presented at the Joint Meeting of The World Society for Pediatric and Congenital Heart Surgery Honoring Dr Aldo Castaneda; July 15-17, 2010;Antigua, Guatemala.
Introduction: the Developmental AorticSwitch
The distal or subsemilunar part of the infundibulum or conus
arteriosus performs an arterial switch procedure during cardio-
genesis. Normally, the straight heart tube loops or folds to the
right. D-loop formation places the developing right ventricle
(the bulbus cordis) to the right of the developing left ventricle
(the ventricle) of the bulboventricular D-loop. D-loop forma-
tion also places the developing ascending aorta to the right of
the developing main pulmonary artery of the truncus arteriosus.
This developmental stage is reminiscent of the Taussig-Bing
malformation; that is, potentially, there is a double-outlet right
ventricle {S,D,D} with a bilateral conus, the subaortic part to
the right and the subpulmonary part to the left, closer to the
developing ventricular septum, the interventricular foramen
(the ventricular septal defect), and the developing left ventricle.
The structural problem at this critical developmental stage is
how to avoid the Taussig-Bing type of double-outlet right
ventricle. Asymmetrical conal free wall development is the
solution. On the left side, growth of the subpulmonary conal
free wall elevates the pulmonary valve superiorly and protrudes
it anteriorly, above the anterior and right-sided right ventricle,
moving the pulmonary valve and artery away from the ventri-
cular septal defect (VSD). On the right side, resorption of the
subaortic conal free wall carries the developing aortic valve
inferiorly, posteriorly, and leftward into and partly through the
interventricular foramen, resulting in aortic-mitral approxima-
tion and direct fibrous continuity. The last step in this normal
arterial switch procedure is closure of the interventricular
foramen at its rightmost end, between the anterior and septal
leaflets of the tricuspid valve and beneath the right coronary-
noncoronary commissure of the aortic valve, by atrioventricu-
lar endocardial cushion tissue that is intimately associated with
the tricuspid valve, thereby forming the membranous septum.
World Journal for Pediatric andCongenital Heart Surgery1(3) 364-385ª The Author(s) 2010Reprints and permission:sagepub.com/journalsPermissions.navDOI: 10.1177/2150135110380239http://pch.sagepub.com
364 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
Figure 1. Types of human heart, with emphasis on segmental sets (or combinations), alignments, and spatial relations. Heart diagrams are viewedfrom below, as visualized with subxiphoid 2-dimensional echocardiography. Cardiotypes depicted in broken lines had not been documented whenthe diagram was made; cardiotypes shown in solid lines have all been documented. The aortic valve is indicated by the coronary ostia; thepulmonary valve is indicated by the absence of the coronary ostia. Braces {} mean ‘‘the set of.’’ Ant ¼ anterior; Inf ¼ infundibulum; LA ¼morphologically left atrium; L ¼ left; LV ¼ morphologically left ventricle; Post ¼ posterior; R ¼ right; RA ¼ morphologically right atrium; RV ¼morphologically right ventricle. The columns (1-4) are arranged in terms of atrioventricular (AV) concordance or discordance. Column 1 has AVconcordance between the situs solitus of the viscera and atria {S,-,-} and D-loop ventricles {S,D,-}. Column 2 has AV discordance between thesolitus viscera and atria {S,-,-} and L-loop ventricles {S,L,-}. Column 3 has AV concordance between the visceroatrial situs inversus {I,-,-} and L-loopventricles {I,L,-}. Column 4 has AV discordance between situs inversus of the viscera and atria {I,-.-} and D-loop ventricles {I,D,-}. Situs ambiguus{A,-,-} of the viscera and atria in the heterotaxy syndromes with congenital asplenia and polysplenia with D-loop ventricles {A,D,-} and with L-loopventricles {A,L,-} has been omitted for reasons of accuracy: when the atrial situs is uncertain or unknown {A,-,-}, one cannot say whether D-loopventricles {A,D,-} or L-loop ventricles {A,L,-} have AV concordance or discordance. To make the diagnosis of AV concordance or discordance, onemust know both the atrial situs and the ventricular situs. The rows (1-8) are arranged according to the types of ventriculoarterial (VA) alignmentthat are present. Normally related great arteries are depicted in rows 1 to 4 inclusively. Solitus normally related great arteries {-,-,S} and inversusnormally related great arteries {-,-,I} are both shown. In row 5, some types of transposition of the great arteries (TGA) are shown but by no meansall: D-TGA, that is, TGA {-,-,D} in which the transposed aortic valve lies to the right (dextro or D) relative to the transposed pulmonary valve, andL-TGA, that is, TGA {-,-,L} in which the transposed aortic valve lies to the left (levo or L) relative to the transposed pulmonary valve, are both
Van Praagh 367
367 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
fibrous continuity (Figure 2), and the subaortic infundibular
free wall grows, elevating the developing aortic valve super-
iorly and protruding it anteriorly (Figure 3). As a result of
reversed development of the subsemilunar conal free walls, the
aortic valve remains anteriorly above the anterior and right-
sided right ventricle, while the pulmonary artery arises poster-
iorly above the left ventricle. Thus, the arterial switch operation
of the subsemilunar conus has been done incorrectly: the pul-
monary artery (instead of the aorta) has been switched into the
left ventricle, and the aorta (instead of the pulmonary artery)
remains above the right ventricle.
This reversed or ‘‘backward’’ subsemilunar conal free wall
development may be regarded as inverted development of the
subarterial conal free walls, that is, right-left reversal or mirror
imagery, without anteroposterior or superoinferior change
(Figure 3). The concept of subsemilunar conal free wall
inverted development applies accurately after the cardiac loop
stage has been reached (Figure 3); but the concept of subsemi-
lunar infundibulum free wall developmental inversion may not
apply accurately at the straight tube stage, when these 2 devel-
oping conal free walls are thought (but not definitely known) to
be anteroposterior (Figure 3) rather than right-left. Anteropos-
terior reversal is not called inversion, whereas right-left rever-
sal is.
In D-TGA, why are the great arteries relatively parallel, or
straight and uncrossed, whereas normally great arteries are
twisted about each other, or really, untwisting above each other
(Figure 2)? Compare the proximal and distal AP relationships.
In D-TGA, proximally, the aortic valve is anterior and some-
what to the right, and the pulmonary valve is posterior and
somewhat to the left (Figure 2). Distally, the aortic arch is also
anterior and superior, and the pulmonary bifurcation is also
posterior and inferior (Figure 2). Thus, in D-TGA, both proxi-
mally at the valves and distally at the aortic arch/pulmonary
bifurcation, the aorta is anterior, and the pulmonary artery is
posterior. Consequently, in D-TGA, the great arteries are rela-
tively parallel, straight, and uncrossed because the fibroelastic
great arteries have little untwisting to do as they proceed from
the valves proximally to the aortic arch and pulmonary bifurca-
tion distally. This lack of necessary untwisting is because the
proximal and distal AP relationships are similar.
If one measures the semilunar interrelationship relative to
the sagittal plane in D-TGA, the semilunar valves often display
only about 30� dextrorotation, compared with about 150� dex-
trorotation for solitus normally related great arteries at the
Figure 1 continued. depicted. A-TGA, in which the transposed aortic valve lies directly anterior (antero or A) relative to the transposedpulmonary valve, is omitted for simplicity and clarity. In row 6, anatomically corrected malposition of the great arteries (ACM) is presented. Notethat in all anatomical types of ACM, the ventricles have looped in one direction and the infundibuloarterial segment has twisted in the oppositedirection. In ACM, the subsemilunar conus is either bilateral (subaortic and subpulmonary) or subaortic only (with pulmonary-tricuspid fibrouscontinuity). In ACM, although the great arteries are very malpositioned, nonetheless, there is VA alignment concordance by definition, hence, thename ‘‘anatomically corrected malposition of the great arteries.’’ ACM may be physiologically corrected, as in ACM {S,D,L}, because all segmentshave alignment concordance; or ACM can be physiologically uncorrected, as in ACM {S,L,D}, because there is one intersegmental alignmentdiscordance at the AV level. VA concordance is not synonymous with normally related great arteries because ACM also has VA alignmentconcordance, but the great arteries are very abnormally related in space, and the conal VA connections are also very abnormal. Row 7 depictssome anatomical types of double-outlet right ventricle (DORV), but by no means all. Similarly, row 8 shows some anatomical types of double-outletleft ventricle (DOLV) but not all. In DOLV, both parts of the subsemilunar conus may, or may not, be very deficient (absorbed or involuted).11,19,27
Reproduced with permission from Foran and colleagues.24
368 World Journal for Pediatric and Congenital Heart Surgery 1(3)
368 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
semilunar valves (Figure 2).1 Thus, D-TGA is characterized by
a major failure of dextrorotation at the semilunar valves
because of reversed or inverted development of the subsemilu-
nar infundibular free walls. With bulboventricular L-loops, the
same processes occur, but in mirror image, resulting in inverted
normally related great arteries or L-TGA (Figure 3, right
panel).
To generalize, normally, what happens both with ventricular
D-loops and L-loops, the initially posterior subpulmonary
conal free wall should grow. Following loop formation, the
subpulmonary conal free wall is in the lesser curvature, both
with D-loops and with L-loops (Figure 3); this is the
subpulmonary conal free wall that normally grows and
expands. The greater curvature, both of D-loops and L-loops,
is the subaortic conal free wall that normally undergoes resorp-
tion. If this subsemilunar conal free wall development is
reversed or inverted, that is, with growth of the greater curva-
ture part and resorption of the lesser curvature part, then, TGA
results, with pulmonary-mitral fibrous continuity and a right
ventricular aorta with no aortic-AV fibrous continuity because
of the presence of an interposed subsemilunar muscular conal
free wall.
So, the first big lesson that we have learned is that the classic
truncoconal malseptation hypothesis, initially proposed by
Figure 2. The morphogenesis of normally and abnormally related great arteries. In the top row, the straight heart tube, the bulboventricularD-loop, and the bulboventricular L-loop are shown from the front (a ventral view). In the second row, the truncus arteriosus and the derivativeascending aorta (Ao) and main pulmonary artery (PA) are shown both with ventricular D-loops and L-loops, as seen from the front (a ventralview). Also depicted are the various anatomical types of conus arteriosus or infundibulum (crosshatching) that can be located beneath the PAand/or the Ao. In the third row, the aortic valve (with coronary ostia), the pulmonary valve (without coronary ostia), the subsemilunar conalmusculature (crosshatching), the mitral valve (bicuspid), and the tricuspid valve (3 leaflets) are shown as seen from below (an inferior view, ie,from caudad looking cephalad), both with ventricular D-loops and L-loops. The bottom row lists some of the relations between the greatarteries that are associated with the 4 main anatomical types of conus: subpulmonary, subaortic, bilateral (subaortic and subpulmonary), andbilaterally absent or very deficient (neither subaortic nor subpulmonary). The diagrams of the great arteries and the conus are deliberately‘‘diagrammatic’’; that is, definitive anatomy rather than developmental stages are shown, for clarity of understanding. A ¼ common (undivided)atrium; Ant ¼ anterior (ventral); AoV-TV ¼ aortic valve–to–tricuspid valve (fibrous continuity); BC ¼ bulbus cordis (from which the conusarteriosus and the right ventricular sinus develop); D-loop ¼ a bulboventricular loop that has looped or folded in a rightward (dextral or D)direction, placing the developing right ventricle to the right of the developing left ventricle; D-MGA¼ dextromalposition of the great arteries inwhich the malposed aortic valve lies dextral (or to the right) of the malposed pulmonary valve: D-MGA occurs in the double-outlet rightventricle (DORV), doublet-outlet left ventricle (DOLV), and in anatomically corrected malposition of the great arteries (ACM); D-TGA ¼dextrotransposition of the great arteries in which the transposed aortic valve lies to the right (dextro or D) relative to the transposedpulmonary valve; Inf¼ infundibulum (also known as the conus arteriosus); Lt¼ left; L-MGA¼ levomalposition of the great arteries in which themalposed aortic valve lies to the left (levo or L) relative to the malposed aortic valve: L-MGA occurs in the DORV, DOLV, and ACM (seeFigure 1); L-TGA¼ levotransposed aortic valve is to the left (levo or L) of the transposed pulmonary valve; LV¼morphologically left ventricularsinus; Post¼posterior (dorsal); PV-MV¼ pulmonary valve–to–mitral valve (fibrous continuity); RT¼ right; RV¼morphologically right ventricularsinus; Sup ¼ superior (above or cephalad); TA ¼ truncus arteriosus (from which the great arteries develop in part); V ¼ ventricle of thebulboventricular loop (from which the LV develops). Reproduced with permission from Van Praagh and colleagues.1
Van Praagh 369
369 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
Figure 3. The embryonic aortic switch procedure: how to do it right and how to do it wrong. At the straight tube stage, the future subpulmonaryconal area (stippled) is thought to be posterior (dorsal) relative to the future subaortic conal area (clear, ie, not stippled). Following D-loopformation, the future subpulmonary conal area occupies the lesser curvature of the D-loop (stippled), and the future subaortic conal area occupiesthe greater curvature of the D-loop (not stippled); the future subpulmonary conal region lies to the left, and the future subaortic conal region lies tothe right. Now, let us focus on the subsemilunar conal free walls (not the conal septum). Normally, the subpulmonary conal free wall grows andexpands, carrying the overlying developing pulmonary valve and main pulmonary artery superiorly (cephalad) and anteriorly (ventrally) on the left-hand side of the great arterial outflow tracts. At the same time, the subaortic conal free wall undergoes resorption or involution (indicated by thebroken circular line). Resorption of the subaortic infundibular free wall causes the overlying developing aortic valve and ascending aorta to sinkinferiorly (caudad), posteriorly (dorsally), and leftward. The aortic valve passes partly through the interventricular foramen and comes into fibrouscontinuity with the mitral valve via the intervalvar fibrosa. The last step in the normal morphogenesis of crossing the circulations is closure of theinterventricular foramen at its rightmost side, adjacent to the tricuspid valve. This is how the developmental arterial switch procedure is performedcorrectly in normal development. The pulmonary valve is carried up and away from the interventricular foramen, and the aortic valve passesdownward and to the left into the interventricular foramen. Formation of the membranous septum by endocardial cushion tissue of theatrioventricular canal completes the normal aortic switch procedure. D-loop formation and asymmetrical development of the subsemilunar conalfree walls result in about 150� of rotation of the semilunar valves to the right (in a counterclockwise direction), when viewed from below.Consequently, the fibroelastic ascending aorta and main pulmonary artery normally must untwist through about 150� in the opposite direction,leftward or clockwise as viewed from below, because of the fixed aortopulmonary relationship distally, where the aorta always arches anteriorly(ventrally) and superiorly (cephalad) relative to the bifurcation of the main pulmonary artery, because these are the fixed spatial relations betweenaortic arches 4 and pulmonary arches 6 in the embryonic branchial aortic arch system. So, normally related great arteries really are untwistingabout each other (not twisting about each other). D-loop formation and asymmetrical conal subsemilunar free wall development are the ‘‘engines’’that perform the aortic switch procedure, thereby normally crossing the systemic venous and the pulmonary venous circulations and achievingsolitus normally related great arteries. But look what happens when the wrong subsemilunar conal free wall grows and expands (the nonstippledsubaortic conal free wall on the greater curvature of the D-loop) and when the wrong subsemilunar conal free wall undergoes resorption (thestippled subpulmonary conal free wall in the lesser curvature of the D-loop). When subsemilunar conal free wall development is the opposite ofnormal, then, it is the aortic valve that is carried superiorly and protruded anteriorly on the right side of the great arterial outflow tracts, and it is thepulmonary valve and main pulmonary artery on the left that sink inferiorly and posteriorly, passing partly through the interventricular foramen andcoming into fibrous continuity with the mitral valve, which is made possible by the abnormal involution of the subpulmonary conal free wall (circularbroken line), resulting in D-TGA. Why are the great arteries relatively parallel, straight, or uncrossed in D-TGA? Because the aortopulmonaryrelationships both proximally at the semilunar valves and distally at the aortic arch/pulmonary artery bifurcation are very similar. Both proximallyand distally, the aorta is anterior (ventral), and the main pulmonary artery is posterior (dorsal). Consequently, in D-TGA, the great arteries haverelatively little untwisting to do as they pass from the semilunar valves proximally to the aortic arch/pulmonary bifurcation distally. An arterialswitch procedure has been performed, resulting in D-TGA, but the embryonic arterial switch procedure has been done wrong because the wrong
370 World Journal for Pediatric and Congenital Heart Surgery 1(3)
370 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
above a left-sided and left-handed right ventricle; and on the
right side, a posterior, inferior, and right-sided aortic valve in
direct fibrous continuity with a right-sided mitral valve and
right-handed left ventricle.
So, why did this patient not have the expected corrected
TGA? An important part of the answer appeared to be
because a well-developed solitus normal type of muscular
subpulmonary conus, with resorption of the subaortic conal
free wall, was present in this patient. This was an amazing
spectacle: a solitus normal type of infundibulum and great
arteries that was related as normally as possible to the under-
lying L-loop ventricles and the inverted AV valves. Suffice it
to say that the discovery of isolated ventricular inversion
{S,L,S} strengthened our hypothesis that the subsemilunar
conus is a very important determinant of whether the great
arteries are normally or abnormally related. This case seemed
to say the following: if the conus is of the solitus normal
type, then, the great arteries are solitus normally related, even
if the ventricular loop is inverted.
There were other clues also. TGA is associated almost never
with an AP septal defect (an AP ‘‘window’’). If TGA were
really caused by an anomaly of AP septation, then, definite
malformations of the AP septum, such as an AP window,
should be relatively common with TGA; in fact, an AP window
in association with TGA is so infrequent as to be reportable.
The conal maldevelopment hypothesis also could explain
the variations in semilunar valve heights, whereas the truncal
(or truncoconal) straight septum hypothesis could not. Propo-
nents of the classic straight AP septum hypothesis tried to say
that the transposed aortic valve is higher than the transposed
pulmonary valve because the conus (beneath the transposed
Figure 3 continued. subsemilunar conal free wall grew and expanded (the subaortic) and the wrong subsemilunar conal free wall underwentresorption (the subpulmonary). When L-loop formation occurs, inverted normally related great arteries also result from subpulmonary conal freewall growth and subaortic conal free wall resorption but in a mirror image when compared with what happens to produce solitus normally relatedgreat arteries with a ventricular D-loop. Similarly, typical L-TGA results from subaortic conal free wall growth and from subpulmonary conal freewall resorption (but again, in a mirror image compared with D-TGA and a ventricular D-loop). Ao¼ ascending aorta; D-loop¼ a bulboventricularheart tube that has looped or folded in a rightward (dextral or D) direction, placing the developing right ventricle (RV) to the right of the developingleft ventricle (LV), the RV being right handed and the LV being left handed; D-TGA ¼ transposition of the great arteries in which the transposedaortic valve is to the right (dextro or D) relative to the transposed pulmonary valve; L¼ left; L-loop¼ a bulboventricular heart tube that has loopedor folded in a leftward (levo or L) direction, placing the developing RV to the left of the developing LV, RV being left handed and the LV being righthanded; PA ¼ main pulmonary artery; R ¼ right. Reproduced with permission from Van Praagh.25
Van Praagh 371
371 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
Thus, both ventricles are composed of 4 developmental
components (Figure 4A and 4B). The subsemilunar or distal
part of the conus (component 4) is the ‘‘architect,’’ responsible
for crossing the circulations. But what is the proximal or apical
part (component 3) of the conus doing? What is it there for?
The lower or apical part of the conus is the ‘‘mother’’ of the
right ventricular sinus (which is, of course, the lung pump). The
septal and moderator bands never dissociate from the right ven-
tricular sinus. In other words, the right ventricle sinus always
evaginates or pouches out just beneath the septal band. By con-
trast, the conal septum and the parietal band can be ‘‘any-
where,’’ as DOLV and ACM illustrate (Figure 1).
The right ventricle is our major cardiovascular evolutionary
adaptation to air breathing and land living.16 The subsemilunar
conus normally crosses the circulations by performing the aor-
tic switch, and the right ventricle sinus is the lung pump.
Single left ventricle17,18 results from the absence of the right
ventricle sinus (Figure 4A, component 2). The associated
infundibular outlet chamber consists of components 3 and 4
in Figure 4A. Absence of the right ventricle sinus results in a
single or unpaired left ventricle, often with double-inlet left
ventricle because there is no right ventricular sinus for the tricus-
pid valve to open into. Common-inlet left ventricle also occurs,
when a common AV canal and a common AV valve coexist.
Because the anatomically right ventricle is composed partly
of the right ventricle sinus or inflow tract (Figure 4A, compo-
nents 1 and 2) and partly of the infundibulum, or conus, or out-
flow tract (Figure 4A, components 3 and 4), the composite
Figure 4. What is the subsemilunar part of the conus? Anatomically, what are we talking about? (A) This is a diagram of a normal, opened,morphologically right ventricle (RV) that shows the 4 main anatomical and developmental component of the RV. The RV inflow tract consistsof 2 components: component 1 is the atrioventricular (AV) canal contribution (the interventricular part of the AV septum, and the tricuspidvalve); and component 2 is the RV sinus, the main pumping portion of the RV. The RV outflow tract also consists of 2 components:component 3 is the proximal or apical part of the conus, which consists of the septal band and the moderator band, which is not involved in theconotruncal malformations (such as TGA, DORV, DOLV, and ACM shown in Figure 1); and component 4 is the distal or subsemilunar part ofthe conus that is involved in the above-mentioned infundibuloarterial anomalies and that consists of the conal septum, the parietal band, andthe subsemilunar conal free wall, which may be well developed, or resorbed, and which may prevent or permit semilunar-AV fibrouscontinuity, respectively. (B) This is a diagram of a normal, opened, morphologically left ventricle (LV), showing that it too consists of 4anatomical and developmental components: 1, the AV canal contribution; 2, the finely trabeculated LV sinus portion; 3, the smooth superiorleft ventricular septal surface component that is confluent with the septal band of the RV (component 3 in A); and 4, the immediately subaorticconal septum, as seen from within the LV. Again, component 4 is the subsemilunar part of the conus that we are focusing on concerninginfundibuloarterial malformations. Note the approximation of the aortic valve and the mitral valve, which is made possible by the normalresorption of the subaortic infundibular free wall. So, it is what one does not see that is most important: no subaortic conal free wallmyocardium (because it has been resorbed), making possible the normal aortic-mitral fibrous continuity. Reproduced with permission fromVan Praagh and colleagues.15
Van Praagh 373
373 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
nature of the normally right-sided ventricle often is not well
understood and hence is delineated here. Thus, normally
related and connected great arteries illustrate how the develop-
mental arterial switch should be done by the subsemilunar
infundibular free wall. Abnormally related and connected great
arteries illustrate the consequences of alterations in this devel-
opmental process.
TOF illustrates what happens when the developmental arter-
ial switch procedure is done subnormally: The pulmonary
valve remains too leftward, posterior, and inferior; the
Figure 5. The 4 main anatomical types of subsemilunar infundibulum or conus arteriosus: subpulmonary, subaortic, bilateral (subaortic andsubpulmonary), and absent or very deficient. The upper row of diagrams shows the infundibulum (crosshatched) and great arteries as seenfrom the front (frontal view). The lower row of diagrams shows the infundibulum (crosshatched), the semilunar valves—the aortic valve (AoV),indicated by the coronary arteries, and the pulmonary valve (PV), indicated by the absence of coronary arteries—and the atrioventricularvalves—the mitral valve (MV), being a 2-leaflet valve, and the tricuspid valve (TV), being a 3-leaflet valve—as seen from below (inferior view),similar to a subxiphoid 2-dimensional echocardiogram. In all diagrams, a ventricular D-loop is assumed to be present. The subpulmonary conusis normal. Resorption of the subaortic conal free wall permits aortic-mitral fibrous continuity. The presence of a subpulmonary infundibulumprevents pulmonary valve–atrioventricular valve fibrous continuity. A subpulmonary conus is associated with solitus normally related greatarteries (diagrammed here), inversus normally related great arteries (diagrammed in Figures 1-3), and in tetralogy of Fallot, both with solitusnormally related great arteries12 and with inversus normally related great arteries. A subpulmonary conus can also be associated with double-outlet right ventricle with the hypoplastic left heart syndrome (eg, with mitral atresia) and with aortic-tricuspid fibrous continuity. Thesubaortic conus is characterized by resorption of the subpulmonary conal free wall, permitting pulmonary-mitral direct fibrous continuity. Thepresence of a complete muscular subaortic conus prevents aortic-atrioventricular fibrous continuity. The subaortic conus and great arteriesshown here are associated with typical D-transposition of the great arteries, that is, TGA {S,D,D} (Figures 1-3). A subaortic conus also occurswith L-TGA, that is, TGA {S,L,L}, and with TGA {I,L,L} (Figures 1-3). A bilateral conus, being both subaortic and subpulmonary, preventssemilunar-atrioventricular fibrous continuity. A bilateral conus is associated with typical double-outlet right ventricle, both with D-loopventricles and with L-loop ventricles (Figures 1 and 2). A bilateral conus can also be associated with TGA when there is a muscularsubpulmonary outflow tract obstruction (stenosis or atresia).26 Rarely, it is possible for solitus normally great arteries to be associated with abilateral conus if the subpulmonary part of the conus is well developed and if the subaortic conal free wall is present but poorly developed, just1 or 2 mm in height between the aortic valve above and the mitral valve below; I have seen only 1 such case in my life, in a patient with theincomplete form of common AV valve canal with an ostium primum defect at the atrial level, no ventricular septal defect, and a cleft mitralvalve. So, what matters most morphogenetically is not just the anatomical type of conus that is present but rather how much the subsemilunarconal free wall is present or has been resorbed. In the rare case that I am referring to, a small amount of the subaortic conal free wall had notbeen resorbed, but not enough to disrupt the normal type of aortic valve–to–left ventricular approximation. The bilaterally absent or verydeficient conus can be associated with double-outlet left ventricle (DOLV) with aortic-mitral and pulmonary-mitral fibrous continuity, evenwith an intact ventricular septum.19 However, DOLV does not always have a bilaterally absent or very deficient conus.27 Figure 2 shows thediagram of a rare type of D-TGA with a bilaterally deficient conus, but with aortic valve–to–tricuspid valve fibrous continuity and withpulmonary valve–to–mitral valve fibrous continuity. It is also noteworthy in Figure 2 that a bilaterally absent or very deficient conus is notdiagrammed in association with L-loop ventricles. Why not? Because we have never seen this. It may occur (but I do not know that). Figure 2 isevidence based (not hypothetical, except where indicated by broken lines). AD ¼ anterior descending (coronary artery); Ant ¼ anterior(ventral); Inf ¼ inferior (caudad); Lt ¼ left; Post ¼ posterior (dorsal); Rt ¼ right; Sup ¼ superior (cephalad). Reproduced with permission fromVan Praagh.28
374 World Journal for Pediatric and Congenital Heart Surgery 1(3)
374 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
Figure 6. The 4 main hypotheses concerning the morphogenesis of transposition of the great arteries. (1) Malseptation of the great arteries,that is, straight (as opposed to spiral) development of the aortopulmonary septum, first proposed (to our knowledge) by Quain5 in 1844 andsubsequently espoused by many authors. Ao ¼ aorta; AoV ¼ aortic valve; PA ¼ pulmonary artery; PV ¼ pulmonary valve. The semilunar valvesare designated by conventional numbers for clarity because their relative positions are highly variable. There are 4 septal semilunar leaflets,adjacent to the aortopulmonary septum: aortic and pulmonary leaflets 1, and aortic and pulmonary leaflets 3. Pulmonary leaflet 2 is nonseptal,remote from the aortopulmonary septum. Aortic leaflet 4 is also nonseptal and normally is noncoronary. The nonseptal semilunar leaflets arealso known as the intercalated leaflets. Comparison of the semilunar leaflet numbers with normally related great arteries (NRGA) as opposedto those with transposition of the great arteries (TGA) indicates (in degrees) the morphogenetic movement that has occurred with NRGA andhas not occurred with TGA. Both with NRGA and with TGA, the distal aortopulmonary relations at the aortic arch and pulmonary bifurcationare the same: the Ao arch is ventral and cephalad to the PA bifurcation because this is the fixed aortic arch 4–to–pulmonary arch 6 relationshipdistally. The straight aortopulmonary septum hypothesis is considered to be wrong for several reasons: (a) In TGA, the free walls of the greatarteries are just as abnormal as is the aortopulmonary septum. This is indicated by the abnormal locations of the coronary ostia in TGA, the
Van Praagh 375
375 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from
Figure 6 continued. coronary arteries being the first branches of the aortic free wall. So, TGA is more than an anomaly ofthe aortopulmonary septum because the great arterial free walls are also very abnormally located. (b) Definite evidence of abnormality of theaortopulmonary septum, such as AP window, is very rare in TGA. (c) The straight aortopulmonary septum hypothesis cannot explain thevariations in semilunar valve heights in the so-called conotruncal malformations, such as the aortic valve sitting high above the morphologicallyleft ventricle in anatomically corrected malposition of the great arteries {S,D,L} (Figure 1, row 6, column 1).10 (2) Conal maldevelopment. In1909, Keith23 was the first to propose that TGA results from persistence and development of the subaortic part of the conus (in white) andinvolution of the subpulmonary part of the conus (in black). In 19668 and 1967,7 we independently reached the same conclusion andsubsequently have extended the conal maldevelopment hypothesis to include all of the conotruncal anomalies (Figure 1).1-4,6-13 Note thatKeith23 and Cardell32 and many others thought that the normal semilunar relationships are pulmonary valve anterior, superior, and to the rightof the aortic valve, a frequent preangiocardiographic error. (3) Atavism. In 1923, Spitzer29,30 proposed a hypothesis of evolutionary(phylogenetic) regression to explain TGA in man, back to the cardiovascular state that is normal in higher reptiles such as crocodiles andalligators. One might suppose that such a hypothesis would strike investigators as hilariously funny. Instead, it mesmerized a generation.30,33 Myreaction (with a wink) was that our patients with TGA did not have tails, their smiles were not unusually wide, and they almost never hadichthyosis or any other features that suggested reverse evolution back to the crocodilian stage. So, I decided to study Spitzer’s theory29,30 withgreat care. Lev and Vass’s30 translation into English was very helpful. Finally, I understood what was wrong with this hypothesis in terms ofpathological anatomy. The dotted line indicates where the true ventricular septum was in TGA but has disappeared, according to Spitzer29,30
(left lower panel). Spitzer’s bicuspid, apparently stenotic, transposed pulmonary valve sits above the morphologically right ventricularmyocardium, to the right of the disappeared ventricular septum in D-loop ventricles. His diagram suggests that there is a ventricular septaldefect. Spitzer29,30 also states that the apparent ventricular septum is a hugely hypertrophied crista supraventricular (supraventricular crest),not the true interventricular septum (which has disappeared). So, in terms of pathological anatomy, what is wrong with Spitzer’sinterpretation? The anteroseptal region of the left-sided ventricle from which the transposed pulmonary artery arises in typical TGA consistsonly of the morphologically left ventricular myocardium. There is no morphologically right ventricular myocardium there to the right ofSpitzer’s dotted line. Spitzer does not explain how or why the ventricular septum routinely disappears in TGA. Crocodiles and alligatorsnormally have both a left ventricular aorta and a right ventricular aorta. Spitzer hypothesizes that human TGA represents reopening of theright ventricular aorta of the higher reptiles, plus closure of the left ventricular aorta of the higher reptile and mammals. So, Spitzer’s problemwas not to explain the right ventricular aorta in human TGA: all higher reptiles have a right ventricular aorta, as Spitzer knew. However,Spitzer’s real problem was the left ventricular pulmonary artery of human TGA because there is no animal known in which the pulmonaryartery normally originates above the morphologically left ventricle. This seems to be why Spitzer29,30 contended that the ventricular septum, tothe left of the ‘‘transposed’’ pulmonary artery, has disappeared. In this way, the pulmonary artery can be diagrammed as arising above the rightventricular myocardium, which is essential in any atavistic explanation of human TGA, because there is no animal known in which thepulmonary artery normally arises above the morphologically left ventricle. This hypothesis denies the reality of human TGA and asserts thatthe double-outlet right ventricle (DORV) is really present. The fatal flaw in Spitzer’s hypothesis29,30 of evolutionary regression to explain themorphogenesis of TGA is that in human TGA, the pulmonary artery really does arise above the morphologically left ventricle, the trueventricular septum is present and has not disappeared, and TGA (ventriculoarterial alignment discordance) really is present, not DORV. Infairness to Spitzer, it must be added that at his time (1923), what we now call DORV was then regarded as a form of TGA. Previousinvestigators have attempted to assess Spitzer’s hypothesis.30,33 It has been regarded with some doubt and skepticism. We may have been thefirst to indicate that Spitzer’s hypothesis29,30 is wrong because it is not supported by the morphological anatomical data of human TGA (asabove). (4) Fibrous malattachment. In 1962, Grant31 proposed that normally, there is a fibrous tract of low growth potential that tethers thedeveloping aortic valve (A) to the developing mitral valve (M), resulting in aortic-mitral fibrous continuity; the pulmonary valve (P) is nottethered either to the mitral valve (M) or to the tricuspid valve (T), and consequently, the pulmonary valve is normally anterior to the aorticvalve, the pulmonary valve being in communication with the anterior and right-sided ventricle (right lower panel). The fibrous tract betweenthe normally related aortic valve and the mitral valve is known as the intervalvar fibrosa. Grant’s31 hypothesis is that in TGA, this fibrous tractis shifted leftward, such that the developing pulmonary valve (P) is tethered to the developing mitral valve (M), resulting in pulmonary-mitralfibrous continuity and a left ventricular pulmonary artery. The aortic valve (A) is now untethered, and consequently, the aortic valve (A) isanterior to the pulmonary valve (P), with the aorta being above the anterior and right-sided right ventricle, as in typical TGA in man. So, thequestion becomes the following: Is there anything wrong with Grant’s31 hypothesis? Why could a fibrous tract of low growth potentialbetween the mitral valve (M) and the wrong semilunar valve, the pulmonary valve (P), not be the primary morphogenetic mechanismunderlying human TGA? This fibrous malattachment hypothesis31 cannot explain TGA with a bilateral conus (subaortic and subpulmonary).26
The presence of subpulmonary conal musculature would prevent the hypothesized abnormal pulmonary-mitral fibrous continuity that occurswhen TGA has a subaortic conus only. This fibrous malattachment hypothesis31 also cannot explain those rare cases of TGA with asubpulmonary conus,26 that is, short subpulmonary muscular conus with aortic valve–to–tricuspid valve fibrous continuity, in which thetransposed aortic valve can be posterior and inferior to the transposed pulmonary valve. This fibrous malattachment hypothesis, asproposed,31 cannot explain TGA with aortic-tricuspid fibrous continuity and a short subpulmonary conus. Grant’s hypothesis31 also does noteasily explain those rare cases of TGA with aortic-tricuspid and pulmonary-mitral fibrous continuity with a bilaterally deficient or absent conusbeneath both great arteries.26 Consequently, we prefer the view that regards subarterial conal free wall development—growth orresorption—as the primary morphogenetic mechanism. We think that semilunar-atrioventricular fibrous continuity or noncontinuity issecondary to subarterial conal free wall development (resorption or growth, respectively). The latter hypothesis can explain all of theanatomical data. The infundibuloarterial (conotruncal) anomalies (Figure 1) are appropriately named in terms of their pathological anatomy.Anatomically, both the infundibulum or conus arteriosus and the great arteries are malformed. Embryologically, or developmentally, however,the infundibuloarterial (conotruncal) anomalies are importantly misnamed. They are all infundibular or conal anomalies, like tetralogy of Fallot.The great arteries per se are normally formed. The real problem is the little hollow ‘‘platforms,’’ the coni arteriosi (arterial cones) orinfundibula (funnels) on which the great arteries stand, and which connect the great arteries above to the underlying ventricles, ventricularseptum, and atrioventricular valves below. Reproduced with permission from Van Praagh.6
376 World Journal for Pediatric and Congenital Heart Surgery 1(3)
376 by Jose-Antonio Quibrera on October 30, 2010pch.sagepub.comDownloaded from