The Forgotten Interleaflet Triangles: A Review of the Surgical Anatomy of the Aortic Valve

Transcription

1 The Forgotten Interleaflet Triangles: A Review of the Surgical Anatomy of the Aortic Valve John P. Sutton III, MD, Siew Yen Ho, PhD, and Robert H. Anderson, MD Department of Paediatrics, National Heart and Lung Institute, London, United Kingdom Surgical descriptions of the aortic root are not always in keeping with anatomy as observed in the autopsied heart. Although all surgeons appreciate that the concept of the aortic annulus does not imply the presence of a straight ring as a hinge point, the nature of the supporting fibrous structures relative to the semilunar attachment of the leaflets has yet to be clarified. We have analyzed 50 normal aortic roots, two fetal aortic roots sectioned histologically in horizontal and sagittal planes, respectively, and two autopsied adult hearts in which prosthetic aortic valves had been inserted during life. Our results demonstrate the important interrelationships between aortic sinuses, valvar leaflets, and supporting left ventricular structures that produce the three fibrous interleaflet triangles. It is the structure and location of these triangles that is the key to the understanding of the surgical anatomy. Our results also show that the presently used definition of commissure does not reflect adequately the extent of the zones of apposition between adjacent valvar leaflets, essential for normal function when the valve is closed. (Ann Thorac Surg 1995;59:419-27) S urgeons and morphologists continue to battle with the most appropriate terminology to describe the proximal aorta and its origin from the left ventricle. Surgical texts suggest sewing prosthetics to the aortic ring or annulus [1, 2]. Morphologists, in contrast, describe the attachments of the valvar leaflets as semilunar, and stress the absence of a true "ring" of annular tissue supporting the hinge points of the leaflets in one straight circular plane [3-5]. What is the true situation, and how do we reconcile these differences? To clarify the morphology, we have reviewed the pertinent literature, assessed the gross anatomy of the aortic outflow tract, and observed both macroscopically and microscopically the consequences of placement of a prosthesis within the aortic root. Material and Methods We studied 50 hearts with normal aortic roots to review the overall arrangement of the aortic valvar leaflets and aortic sinuses, together with their mode of connection to the supporting ventricular structures. We analyzed serial sagittal and transverse sections through the aortic root of two normal human fetal hearts stained for both elastic and collagenous tissues (elastic van Gieson stain). In addition, we found two hearts in our collection in which, at a previous operation, prosthetic aortic valves had been sutured in the aortic root (a stent-mounted bioprosthesis and a ball-and-cage type mechanical prosthesis). We analyzed these hearts morphologically to identify the position of the prostheses and their relationship to the structures of the outflow tract. Then we sectioned portions of the line of attachment, colored them with elastic Accepted for publication Oct 4, Address reprint requests to Dr Anderson, Department of Paediatrics, National Heart and Lung Institute, Dovehouse SL London SW3 6LY UK. van Gieson stain, and analyzed these histologic sections. Measurements of the structures were made by opening the aortic root longitudinally through the base of the left coronary sinus, and then measuring with a thin string the circumference of the aortic root at the level of the bases of the attachments of the leaflets. The string then was compared to a millimeter rule to assess the degree of muscular versus fibrous support. Measurements of the dimensions of the fibrous interleaflet triangles found on the ventricular aspect of the leaflets were performed with needle-tipped dividers and also measured against a millimeter rule. All measurements then were compared to the overall circumference of the aortic root to yield a percentage value of the proportion occupied by the fibrous triangles. Definition of Terms The aortic root is that portion of the ventricular outflow tract that supports the leaflets of the aortic valve. It is a functioning unit with relations both to the aorta and to the left ventricle. The aortic root is made up of the sinuses of Valsalva, the valvar leaflets, and the interleaflet triangles. The sinuses of Valsalva are the expanded portions of the aortic root confined proximally by the attachments of the valvar leaflets and distally by the sinutubular ridge. The sinuses are named according to the arteries arising from within them (right, left, and noncoronary). Various names have been used for the portions of tissue separating the ventricular and aortic areas. These include valvule, cusp, scallop, and leaflet [3, 4, 6]. We prefer the term leaflet for use in description of the atrioventricular as well as the arterial valves, as the structures described subserve similar functions in both settings. The collagenous condensation at the point of attachment of each leaflet has been termed the annulus 1995 by The Society of Thoracic Surgeons /95/$ (94)00893-C

2 420 SUTTON ET AL Ann Thorac Surg SURGICAL ANATOMY OF THE AORTIC VALVE 1995;95: Fig 1. The aortic root is viewed from the arterial aspect. The zones of apposition (white broken lines) are arranged as triradiating spokes between the valvar leaflets. The peripheral attachments (*) are at the junction between the sinusal and tubular portions of the aorta (sinutubular junction) (cor. - coronary.) Table 1. Relationship of Various Portions of the Aortic Root to Surrounding Structures Portion of Aortic Root Noncoronary sinus Right coronary sinus Left coronary sinus Non-/right coronary interleaflet triangle (membranous septum) Right/left coronary interleaflet triangle Left/noncoronary interleaflet triangle (subaortic curtain) Related Structure Left and right atriums, transverse sinus Right atrium, free pericardial space Left atrium, free pericardium Right atrium, conduction system, septal leaflet of tricuspid valve, right ventricle Potential space between aorta and puhnonary trunk or infundibulum Left atrium, makes up large portion of the aortic leaflet of the mitral valve fibrosus [2, 7]. The body is the large weight-bearing surface of the leaflet. The free edge of the leaflet is constructed so that, when closed, the leaflets coapt over several millimeters. This margin of overlap defines the coapting surface of the valvar leaflet. It also has been termed the lunula. When the aortic valve is closed, each leaflet of a trifoliate valve meets its neighbor along a junction extending from the periphery to the centroid of the valvar orifice (Fig 1). There is no consensus as to the most appropriate description of this overall zone of apposition between adjacent leaflets. All of the area is important if the valve is to be a competent structure. If the word "'commissure" was used in its vernacular meaning, then the entirety of these junctions of adjacent leaflets would justifiably be called commissures. Thus, the Shorter Oxford English Dictionary on Historic Principles [8] gives four definitions for commissure, at least two of which are pertinent for the aortic valve: (1) A joining together; the place where two bodies touch or unite; a joining, juncture, seam. (2) The line of junction or angles of the two lips, eyelids, etc. Both of these definitions would fit with the description of the zones of apposition between the leaflets of the aortic valve as shown in Figure 1. Neither is consonant with the traditional and current use of cornmissures in cardiac surgery. As stated by an anonymous referee of the first draft of this manuscript, conventional wisdom holds that the commissure applies only to the peripheral attachment of the valvar leaflets to the sinutubular junction of aortic root and ascending aorta. Therefore, to circumvent these semantic problems in this work, we will simply describe the zones of apposition of the leaflets (Fig 1). These zones have upper and lower lines of coaptation. Within this concept, the aortic root can contain either one zone of apposition, as seen in the bileaflet valve, or three zones of apposition as is the norm. This view of zones of apposition is equally applicable to the atrioventricular valves. If the word commissure was used sinutubular ridge ~ J" ~'..~/ triangles X valve- "\~" ~ ~ Fig 2. A composite picture of the aortic root from the heart of a neonate. The aortic root has been opened longitudinally through the area of the left coronary sinus. (Upper panel) Part of the muscular and fibrous support of the aortic valve. The leaflets are hinged in semilunar fashion to the aortic wall. (Lower panel) The circumferential thickening of the aortic wall at the junction between the sinual and tubular portions forms the sinutubular ridge. The broken line joins the nadirs of the attachment of the leaflets and circumscribes a basal "ring," but this "r/ng" is not a continuous collagenous structure.

3 Ann Thorac Surg SUTFON ET AL ;95: SURGICAL ANATOMY OF THE AORTIC VALVE in its vernacular meaning, then the zones should logically be described as commissures. Only time will tell if this can be achieved. Beneath the apices formed by the lines of attachment of the leaflets to the aortic wall exist the fibrous components of the aortic root. Although these areas are unequivocally a part of the aortic root, they are subjected to the hemodynamics of the left ventricle and not the aorta [5]. The areas are bounded by the semilunar attachments of the valvar leaflets. These triangles of tissue have been described previously using several names (spatia intervalvularis, trigone, interleaflet triangle) [3, 4, 9]. We will call them the interleaflet triangles. The interleaflet triangles and aortic sinuses are parts of, and are related to, several important structures of the heart (Table 1). Results Normal Aortic Root The aortic root comprises the proximal portion of the aorta delineated by the sinutubular ridge superiorly and the bases of the semilunar valvar leaflets inferiofly. Its components are the sinuses of Valsalva, the intefleaflet triangles, and the valvar leaflets (Fig 2). The proportion of muscular, as opposed to fibrous, support of the aortic root provided by the left ventricular structures ranged from 25% to 85% in the hearts studied, with an average of 47%. The interleaflet triangles averaged 54% of the total circumference when measured at the level of the bases of the sinuses (Table 2). There was considerable variety in the architecture of musculature supporting the left and right sinuses. The noncoronary sinus had significant muscular support in only one instance. The transition from muscular to fibrous tissue at the membranous septum was also variable. Muscular bands could be seen crossing the interleaflet triangles in some specimens, whereas, in an equal number of specimens, the membranous septum passed under part of the right coronary sinus. The sinutubular ridge is thicker than the adjacent sinuses. When viewed from above in the autopsied heart, it is circular and delineates the beginning of the ascending aorta proper (Fig 1). Just below the sinutubular ridge, the aortic root takes on a trifoliate and scalloped appearance when seen in cross section, this portion corresponding to the sinuses of Valsalva. The sinuses expand down to the base of the aortic root where the circumference narrows slightly. The coronary arteries generally arise within the sinuses or at the ridge, although they can originate above the sinutubular ridge as a normal variation. The origins of the arteries permit use of the descriptive terms of left and right coronary sinuses, although the arterial orifices can be paired or anomalous. It is exceedingly rare, however, for a coronary artery to arise from that sinus that is not adjacent to the pulmonary trunk (the nonfacing sinus). The sinuses are confined inferiorly by the leaflets. The valvar leaflets are the portions of the aortic root that separate the hemodynamic components of aorta and ventricle, and are attached to the wall of the root in a Table 2. Analysis of Muscular Contribution of Ventricle to Aortic Root, and Percentage of Circumference of Aortic Root Occupied by lnterleaflet Triangles Muscular Percent of Total Standard Contribution Circumference Deviation Average muscular circumference Triangle A Triangle B Triangle C A ~ between right and left coronary sinuses; B between right and noncoronary sinuses; C between non- and left coronary sinuses. semilunar fashion. The apices of these attachments are at the sinutubular ridge, whereas the nadirs are at or below the anatomic ventriculoarterial junction, this junction being the straight circle where the fibroelastic aortic wall joins with the supporting structures of the left ventricle. Each leaflet contains a hinge point, a body, and a coapting surface with a thickened central nodule. The hinge point is that area where the leaflet is attached to the aortic root. There is a condensation of collagenous tissue at the hinge point that follows the semilunar contour of the valvar attachment. Previously, this has been defined as the annulus fibrosus. This annulus, which takes the form of a three-pronged coronet rather than a ring, was thickest in all the hearts studied at the nadir of the semilunar attachments. A basal ring within the ventricle can be readily constructed by joining together these nadirs, but is not a complete straight circle (Fig 3). Histology The histology of the aortic root is characterized by a gradual shift from the primarily elastic aorta to the muscular ventricle (Figs 4, 5). At the level of the sinutubular ridge, three bulges can be seen internally in the elastic lamellas (see Fig 4A). Proceeding inferiorly, the usually circular lameuar architecture is interrupted by increasing amounts of interspersed collagenous tissue (see Fig 4B-D). The collagen then coalesces to take a primarily longitudinal orientation with interdigitating elastic tissue, with the bundles of longitudinally oriented collagen being surrounded on their luminal surface by additional elastic tissue (see Fig 5). Proceeding more inferiorly, the collagen becomes more defined, and the layers of elastic tissue on the luminal surface become thin and then disappear. At this point, the collagenous structure is seen almost to "extrude" into the lumen of the aortic root. The couagenous structure takes on a bicornuate appearance concomitant with separation of the limbs of longitudinally oriented collagen bundles (see Fig 4). As the sinuses become recognizable (see Fig 5), the elastic layers diminish in number, and the collagen bundles separate into two distinct structures. Between these limbs of thickened collagen (the so-called annulus), the leaflets of the valve take shape, and can be seen to have several layers. The more densely staining aortic side

4 422 sutton ET AL Ann Thorac Surg SURGICAL ANATOMY OF THE AORTIC VALVE 1995;95: Fig 3. Diagram of the aortic root. (Inset) The coronetlike arrangement of the valvar attachments. Aortic wall within ventricle (interleaflet triangle) Sinutubular juction ~lnterleaflet triangle Ventriculo-arterial ring and juction Basal ring Ventricle within sinus blends with the so-called annulus, forming the so-called fibrous layer (fibrosa). A less densely staining spongy portion is oriented toward the ventricle (spongiosa). The spongy layer blends with the tissue between the longitudinal limbs of the annulus to form the interleaflet triangle. Each layer is covered by segments of endothelial Fig 4. Photomicrographs of sections taken transversely through a neonatal aortic valve from superior to inferior (A-D) (elastic van Geison stain). (A) At the level of the sinutubular ridge, three bulges in the aortic wall begin to demarcate the aortic sinuses. (Pulm. = pulmonary.) (B) A deeper section shows the hinge points of the three leaflets at various depths. Their insertions are to pale-staining collagenous structures. The origin of the left coronary (cor.) artery is also seen in this section. (C) The collagenous hinges (**) of the left and right leaflets are well seen in this section. The small arrow points to the interleaflet triangle. (D) The interleaflet triangles (small arrows) become larger the deeper the section is into the aortic root. The triangle between the left and right coronary leaflets is related to the tissues (curved arrow) between the aorta and the pulmonary trunk.

5 Ann Thorac Surg SUTFON ET AL ;95: SURGICAL ANATOMY OF THE AORTIC VALVE Fig 5. (A) Section inferior to that shown in Fig 4D. The aortic wall is no longer complete as the section traverses through the ventricular septum. The fibrous triangle between the left and noncoronary (cor.) leaflets (open arrow) is related to the left atrium. The arrowhead points to the interleaflet triangle between the right and noncoronary leaflets. This triangle is related to the membranous septum and through it the septal leaflet of the tricuspid valve. (B, C, D) The gradual changes in the aortic wall through an interleaflet triangle as the transverse sections are followed from superior to inferior. (B) The aortic wall is thickened at the sinutubular ridge but its major component is elastic tissue. This section is at the most superior part of the hinges of the leaflets--the periphery of the commissure. (C) The hinges of the leaflets diverge from one another at a level near the apex of the interleaflet triangle (arrowhead). (D) A section deep in the interleaflet triangle (arrowhead) shows the wide separation of the leaflets. The elastic wall is confined to the sinuses. tissue, the aortic side coated with endothelium, and the ventricular side with endocardium. Nearer the base of the sinuses, the orientation of the elastic tissue relative to the collagen is reversed (Fig 6). The elastic layers decrease in number and become oriented toward the lumen. A dense band of circularly oriented collagen fibers surrounds these tongues of elastic tissue. The area between the limbs of the annulus, corresponding to the interleaflet triangle, becomes large in relation to the area of the aortic root. This tissue of the interleaflet triangle is composed of a thinner, primarily circularly oriented, layer of fibers of collagen and light staining acellular material. A very thin layer of elastic tissue can be seen on the luminal surface that is continuous with the elastic layer in the ventricle beneath the endocardium. The fibrous portions at the base of the left Fig 6. (A) Section deep into the aortic rool The right coronary (cor.) sinus is no longer visible. The base of the left coronary leaflet (star) is seen as an amorphous mass and some elastic tissue remains in the noncoronary sinus. The interleaflet triangle (A) is enlarged (B) to show the circular orientation of the fibrous tissue. (Vent. ventricular.)

6 424 SUTTON ET AL Ann Thorac Surg SURGICAL ANATOMY OF THE AORTIC VALVE 1995;95: Fig 7. (A) The aortic root is displayed after removal of a stentmounted bioprosthetic valve. The broken lines mark the semilunar arrangement of the hinges of the leaflets. Major portions of the leaflets were removed to allow for surgical insertion of the bioprosthesis. The line of attachment crosses the interleaflet triangles. (B) The attachment of the bioprosthesis is retracted to show the aortic sinuses and its insertion to the remnants of the leaflets at their bases. coronary sinus and the noncoronary sinus can be seen to blend with the central fibrous portions of the heart. The hearts containing the prosthetic valves did not differ in their basic anatomy, although the valvar leaflets had been excised. When the prostheses were removed, the line of surgical attachment was seen to run across the interleaflet triangles sitting on the bases of the sinuses (Fig 7). This line corresponds to the anatomic ventriculoarterial junction. Histologically, large amounts of fibrous tissue could be seen surrounding the sewing ring and extending into the aortic wall. Comment Robicsek [10], in a treatise on the sinuses of the aortic root, suggests that, if a cardiac surgeon were asked to describe an aortic valve, he or she would describe three leaflets and three commissures. He goes on to suggest that a physiologist, if asked the same question could be expected to answer: three leaflets, three commissures, and three sinuses. Understanding the purpose of his suggestion, we would like to counter that the same physiologist should have answered: three leaflets, three commissures, three sinuses, and three interleaflet triangles. This latter description would then include all the hemodynamic areas of the aortic root, highlighting the fact that the root must be looked at in its entirety. Zimmerman [3] states that the heart does not lend itself to being studied in a purely anatomic sense. He goes on to recommend viewing the heart in the form of a continuum embracing structure and function. This view seems appropriate. It is the basis for our assessment that a circular, ringlike, annulus cannot, and does not, exist in the aortic root. Surgeons will, nonetheless, almost certainly continue to describe the aortic ring, therefore, it is important to establish the candidates for this descriptive title. The aortic root is a simple structure viewed from the surface. It is bordered superiorly by the sinutubular ridge. It contains the semilunar leaflets. The attachments of the leaflets define the extent of the sinuses of Valsalva and the interleaflet triangles. The base of the aortic root is defined by the nadirs of the attachments of the leaflets. The base sits on partially fibrous and partially muscular ventricular support. This description is that of a threedimensional structure (see Fig 3). Alterations in any part of the structure affect its function. There are contrary views concerning the nature of the valve. For example, Silver and Roberts [11] define the aortic valve as those thin portions of tissue that allow the passage of blood into the aorta and prevent the reflux of blood back into the ventricle. In addition, they state that the structures contacting the valve are important to its function but are not, in themselves, part of the valve. Zimmerman [3], however, has recommended the study of cardiac anatomy with regard to structure and function. We agree with this point of view and prefer to discuss the aortic root as incorporating the entire aortic valvar complex. This complex can then be analyzed at three levels: sinutubular ridge, sinus, and base. The sinutubular ridge is circular when seen from above in the autopsied heart (see Fig 1), and unequivocally constitutes a ring. This portion of the aortic root supports the peripheral attachments of the valvar leaflets. Lewis and Grant [12], using elastic stains, performed detailed histologic studies of the aortic root in the healthy heart and in the heart affected by endocarditis. Their findings are confirmed by our study. The apices of the attachment of the leaflet are surrounded by elastic aortic tissue. Gross and Kugel [7] also echoed their findings. The parabolic shape of the aortic valve has been suggested to form a suitably even support similar to that of a suspension bridge [6]. This occurs both at the attachment of the leaflet to the wall of the aortic root and along the thickened coapting portions of the free edge (Fig 8). These two areas correspond to areas of collagenous thickening as seen using light and electron microscopy [13]. The peripheral attachments to the sinutubular ridge are several millimeters above the level of coaptation of the valvar leaflet [6]. Within the analogy of the bridge, these thickenings correspond to support cables. Brewer and colleagues [14] recognized that the collagenous nature of these free edges of the valvar leaflets

7 Ann Thorac Surg SUTTON ET AL ;95: SURGICAL ANATOMY OF THE AORTIC VALVE "... Fig 8. The support of the aortic valve is likened to a suspension bridge. The collagenous thickening at the zone of coaption and the hinges of the leaflets correspond to support cables. would militate against repeated changes in their length. During systole, they showed the area of the aortic root at the level of the commissural attachments increased by about 16%. This increase in diameter corresponds to a loss of the parabolic shape of the free edge of the leaflet, and contributes in a large part to valvar opening. Again, carrying on with our analogy, as the support poles of the.. bridge tilt back, the cables are lifted, aiding in the opening of the valve. Thus, the sinutubular junction is a circular structure of primarily elastic composition, but with important collagenous supports for the valvar leaflets. A prosthetic valve placed at this point would be above the coronary arteries. This unequivocal ring is morphologically in the wrong place to be the annulus. It is unlikely that anyone would nominate this structure, despite its true circular shape, as the aortic annulus. The sinuses of Valsalva occupy the greater part of the aortic root. The importance of these sinuses was recognized by Da Vinci [10]. The sinuses are composed primarily of elastic tissue, and usually support the openings of the coronary arteries [15]. In cross section, the aortic root at this level is trilobed [3]. The sinuses are the anatomic and functional units of the aortic root. They are the pivotal point of aortic radicular expansion [16]. They accommodate the open leaflets. The formation of vortexes within them is important to valvar function [3]. But, again, the sinuses are not the level of placement of prosthetic valves. Toward the proximal end of the aortic root is the anatomic ventriculoarterial junction. This is also unequivocally of circular shape, and is below the origin of the coronary arteries. It is at this level that surgeons sew in prosthetic valves (see Fig 7). The circular ring, nonetheless, is crossed by the semilunar attachments of the valvar leaflets. These attachments dictate that portions of the aortic root are exposed to ventricular hemodynamics, whereas other portions are exposed to aortic hemodynarnics (Figs 2, 9). These discordances are much more apparent in the pulmonary valve, but are also seen in the aortic valve. The unequivocal fibrous segments supporting the nadirs of attachment of each leaflet are interrupted by the interleaflet triangles. If joined together, these nadirs constitute a basal ring, but the ring does not form a continuous collagenous circle. The confusion of the presence of an aortic annulus is suggested to have come from isolated examinations of the nadirs of the valvar attachments, with subsequent Fig 9. This aortic root from an adult is displayed after removal of the valvar leaflets to demonstrate more clearly the positions of the interleaflet triangles. The area of ftbrous continuity between the aortic and mitral valves is marked by asterisks. The right and left coronary sinuses are supported in part by muscle. Note that in this specimen the right coronary artery originates from the left coronary sinus.

8 426 SUTTON ET AL Ann Thorac Surg SURGICAL ANATOMY OF THE AORTIC VALVE 1995;95: extrapolation of these fibrous areas into an annular ring [3]. Histologic studies of the base of the aortic root have all failed to show any continuous ring at this level. Lewis and Grant [12] noted the lack of a fibrous ring, as did Gross and Kugel [7], who took pains to emphasize the lack of any ring in the form of an annulus. Nonetheless, they combined two tissue layers (fibrous and spongy) at the ventricular side of the base of the valvar leaflet and termed this the valve ring. This manipulation of the histologic findings in itself argues against the presence of a true ring. This lack of an aortic ring is well recognized and has been remarked upon by numerous modern authors [3-5, 14]. Our findings do no more than confirm this position. Yet all seem determined to describe the ring. The only way to have a fibrous circlet at the ventriculoarterial junction is to combine the nadirs of the leaflets with the interleaflet triangles, thus forming a three-peaked coronet (see Fig 3). Indeed, Zimmerman [3] attempted to reconcile this problem by calling the fibrous tissue at the base of the aortic root a coronet. He argues that such a coronet is a stable fibrous structure. This suggestion received additional support from Reid [15]. Thubrikar and co-workers ]16-18], however, have shown that the base of the valve is not a fixed structure, but that it expands as the ventricle fills and then shrinks as the ventricle contracts. This argues against the presence of a ring of fibrous tissue in the sense of Zimmerman's own structure-function continuum. Rather, the base of the aortic root can be said to behave in ventricular fashion. Thubrikar and colleagues [17] showed that the base of the aortic root contracts. They pointed out that the dog heart has 60% of the ventricular outlet composed of muscle. Anderson [5] has also emphasized the contribution of the ventricle to the aorta, and the contribution of the aorta to the ventricle. Sands and colleagues [19], in a comparison of species, found that the human aorta is supported in approximately 45% of its circumference by muscle. Our findings endorse this finding. That the base of the valve follows the ventricular pattern, therefore, is not surprising. The key to this behavior is in the interleaflet triangles. The interleaflet triangles are of different histologic structure, and subserve different physiology when compared to the bases of the sinuses. They are thinner and less collagenous than the attachments of the leaflets, the latter forming their contiguous superior and lateral borders. They are generally bordered by muscle, although this is not true for the interleaflet triangle between the left and noncoronary sinuses. This portion is part of the subaortic curtain, and its more fibrous make-up is important for ventricular function [20]. Our histologic findings concerning the triangles are in keeping with those reported by Thubrikar and colleagues [17]. In the hearts we studied, the prosthetic valves were both seen sitting at the level of the anatomic ventriculoarterial junction in the base of the aortic sinuses. The peripheral attachments of the excised leaflets at the sinutubular ridge were well above the levels of surgical attachment. The sewing ring is attached just above the nadirs of the valvar leaflets and across the interleaflet triangles. The sewing ring of the stent-rnounted valve followed the contour of the leaflet attachments and was not circular. McAlpine [4] stresses this relationship, and urges care when placing sutures in the region of the interleaflet triangles. Our review of the anatomic structure of the aortic root also emphasizes the proximity of the conduction system to the aortic root, as well as the lack of a true annular ring at its base. In summary, therefore, the aortic root is a complex hemodynamic system. Its upper portion is exposed to aortic pressures and behaves as the rest of the aorta. It expands during systole, allowing the leaflets to retract and open. The base of the valve, on the other hang is exposed to ventricular dynamics. It expands as the ventricle fills and, during the peak of systole, it contracts. This decreases the distance required for the valvar leaflets to travel to close the orifice. This action has been hypothesized as reducing the stress applied to the leaflets [17]. The aortic root is also a dynamic structure. Its overall configuration changes from a cone to a cylinder, and then to an inverted cone, as the ventricle fills and contracts. The sinus can be seen as the basic structural unit of the valve. The sinuses support the coronary arteries and the valvar leaflets. The sinuses also allow formation of vortexes, which likely aid in valvar closure. The sinus rocks out, then is dragged back toward the center, by the action of the ventricular myocardium [14, 15, 17]. Each sinus is separated from each other at its base by the interleaflet triangles. It is the interleaflet triangles that are crucial for proper valvar function. Indeed, when one or more of the interleaflet triangles are vestigial, or very small, making the overall valve a more ringlike structure, the valve becomes stenotic. This situation can be seen in congenitally deformed valves with one or two leaflets [21]. Lewis and Grant [12] also remarked on this feature, but did not stress the importance of the triangles, preferring to focus on the attachments of the leaflets. From our review, it seems that the interleaflet triangles have been relatively ignored structures in the study of function of the aortic root. These triangles allow the sinuses to act independently. The mobility of this portion of the aortic root should be of interest considering the current construction of prostheses with rigid basilar sewing rings. Future designers of valvar prosthetics, and students of the aortic root, would do well to remember the forgotten interleaflet triangles. During the course of this investigation, Dr John P. Sutton was a visiting fellow from the Medical University of South Carolina. Doctors Slew Y. Ho and Robert H. Anderson are supported by the British Heart Foundation. References 1. Sabiston DC, Spencer FC. Surgery of the chest. 5th ed. Philadelphia: Saunders, 1990: Kirklin JW, Barratt-Boyes BG. Cardiac surgery. 2nd ed. New York: Churchill Livingstone, 1993:

9 Ann Thorac Surg SUTTON ET AL ;95: SURGICAL ANATOMY OF THE AORTIC VALVE 3. Zimmerman J. The functional and surgical anatomy of the aortic valve. Isr J Med Sci 1969;5: McAlpine WA. Heart and coronary arteries: an anatomical atlas for clinical diagnosis, radiological investigation, and surgical treatment. Heidelburg: Springer, 1975: Anderson RH. Editorial note: the anatomy of arterial valvar stenosis. Int J Cardiol 1990;26: Mercer JL, Benedicty M, Bahnson HT. The geometry and construction of the aortic valve. J Thorac Cardiovasc Surg 1973;60: Gross L, Kugel MA. Topographic anatomy and histology of the valves of the heart. Am J Pathol 1931;7: Onions CT, ed. The Shorter Oxford English Dictionary on Historic Principles. 3rd ed. Oxford: Clarendon Press. 1968: Williams PL, Warwick R, eds. Gray's anatomy. 36th ed. Edinburgh: Churchill Livingstone, 1980: Robiscek F. Leonardo Da Vinci and the sinuses of valsalva. Ann Thorac Surg 1991;52: Silver MA, Roberts WC. Detailed anatomy of the normally functioning aortic valve in the hearts of normal and increased weight. Am J Cardiol 1985;55: Lewis T, Grant RT. Observations relating to subacute infective endocarditis. Heart 1923;10: Clark RE, Finke EH. Scanning and light microscopy of human aortic leaflets in stressed and relaxed states. J Thorac Cardiovasc Surg 1974;67: Brewer RJ, Deck JD, Capati B, Nolan S. The dynamic aortic root. J Thorac Cardiovasc Surg 1976;72: Reid K. The anatomy of the sinus of valsalva. Thorax 1970;25: Thubrikar M, Piepgrass WC, Shaner TW, Nolan SP. The design of the normal aortic valve. Am J Physiol 1981;241: H Thubrikar M, Nolan SP, Bosher LP, Deck JD. The cyclic changes and the structure of the base of the aortic valve. Am Heart J 1980;99: Thubrikar M. The aortic valve. Boca Raton, FL: CRC Press, Sands MP, Rittenhouse EA, Mohri H, Merendino KA. An anatomical comparison of human, pig, calf, and sheep aortic valves. Ann Thorac Surg 1969;8: Zimmerman J, Bailey CP. The surgical significance of the fibrous skeleton of the heart. J Thorac Cardiovasc Surg 1962;44: Anderson RH, Devine WA, Ho SY, Smith A, McKay R. The myth of the aortic annulus: the anatomy of the subaorfic outflow tract. Ann Thorac Surg 1991;52: INVITED COMMENTARY This article by Sutton and colleagues emphasizes the importance of interleaflet triangles in understanding the surgical anatomy of the aortic valve and reminds us that the prosthetic valves are sewn partly to this structure. They also support a view that the aortic root anatomy should be described with regard to structure and function. I would like to carry this further and state that the function should include fluid dynamics as well as the mechanical stresses in the valve. Interleaflet triangles are subjected to ventricular systolic pressures that induce complex patterns of significantly high wall stresses. In a transverse plane, the triangles do not form a circular ring; instead, they form short lines that connect a cloverleaf geometry of the sinuses. This complex geometry induces bending stresses that increase distally in the triangle. Although thin, the tissue in the interleaflet triangle, therefore, must be strong enough to sustain the imposed stresses, which enables it also to support a prosthetic valve. On the other hand, the interfeaflet triangles do not "support" the leaflets (ie, the leaflets do not pull on the triangles), whereas the stent posts of bioprosthetic valves do, which may therefore cause bending of the stent posts inward from the pressure load on the valve. The leaflet of the aortic valve forms a unique structure in conjunction with the sinus, which could be called "leaflet-sinus assembly'" [1]. This leaflet-sinus assembly behaves as an independent unit to store the diastolic pressure within. When the aorta is sepa- rated from the heart, the assembly allows the aortic valve to remain competent even if the interleaflet triangles are partly incised. This assembly gives a unique feature to the sinus geometry, in that, a sinutubular ridge forms a scalloped structure. Thus, the sinutubular junction is not circular as claimed by Sutton and associates. The concept of a suspension bridge, put forth by Sutton and associates, also should be reconsidered in the light of the previous discussion. Certainly the studies on the dynamics of the interleaflet triangle are important not only for understanding the aortic root, but also for evaluating how well it supports a prosthetic valve. Sutton and associates deserve credit for studying the interleaflet triangles. Mano J. Thubrikar, PhD Heineman Medical Research Laboratory Carolinas Medical Center Charlotte, NC Reference 1. Thubrikar MJ. The aortic valve. Boca Raton, FL: CRC Press, 1989: