Chapter 36. Aortic Valve Repair and Aortic Valve-Sparing Operations

The aortic valve is a complex structure that is best described as a functional and anatomic unit, the aortic root. The aortic root has four components: aorto-ventricular junction or aortic annulus, aortic cusps, aortic sinuses, and sinotubular junction. In addition, the triangles beneath the commissures of the aortic valve, although part of the left ventricular outflow tract, are also important for valve function.

The aortic annulus unites the aortic cusps and aortic sinuses to the left ventricle. It is attached to ventricular myocardium (interventricular septum) in approximately 45% of its circumference and to fibrous structures (anterior leaflet of the mitral valve and membranous septum) in the remaining 55% (Fig. 36-1). The aortic annulus has a scalloped shape. Histologic examination of the aortic annulus reveals that it is a fibrous structure with strands attaching itself to the muscular interventricular septum and has a fibrous continuity with the mitral valve and membranous septum. The fibrous structure that separates the aortic annulus from the anterior leaflet of the mitral valve is the intervalvular fibrous body. An important structure immediately below the membranous septum is the bundle of His. The atrioventricular node lies in the floor of the right atrium between the annulus of the septal leaflet of the tricuspid valve and the coronary sinus. This node gives origin to the bundle of His, which travels through the right fibrous trigone along the posterior edge of the membranous septum to the muscular interventricular septum. At this point, the bundle of His divides into left and right bundle branches that extend subendocardially along both sides of the muscular interventricular septum.

The aortic cusps are attached to the aortic annulus in a scalloped fashion (see Fig. 36-1). The aortic cusps have a semilunar shape whereby the length of the base is approximately 1.5 times longer than the length of the free margin, as illustrated in Fig. 36-2. There are three cusps and three aortic sinuses: left, right, and noncoronary. The aortic sinuses are also referred to as sinuses of Valsalva. The left coronary artery arises from the left aortic sinus and the right coronary artery arises from the right aortic sinus. The left coronary artery orifice is closer to the aortic annulus than is the right coronary artery orifice. The highest point where two cusps meet is called the commissure, and it is located immediately below the sinotubular junction. The scalloped shape of the aortic annulus creates three triangular spaces underneath the commissures. The two triangles beneath the commissures of the noncoronary cusp are fibrous structures, whereas the triangular space beneath the commissure between the left and right cusps is mostly muscular. These three triangles are seen in Fig. 36-1. The aortic annulus evolves along three horizontal planes within a cylindrical structure. Thus, the annulus of each cusp inserts itself in the aortic root along a single horizontal plane. The sinotubular junction represents the end of the aortic root. It is an important component of the aortic root because the commissures of the aortic valve are immediately below it and changes in the diameter of the sinotubular junction affect the function of the aortic cusps.

The geometry of the aortic root and its anatomic components varies among individuals, but the geometry of these components is somewhat interrelated. For instance, the larger the aortic cusps, the larger are the diameters of the aortic annulus and sinotubular junction. The aortic cusps are semilunar (crescent shape) and their bases are attached to the annulus; the free margins extend from commissure to commissure, and the cusps coapt centrally during diastole. The size of the aortic cusps varies among individuals and within the same person, but as a rule the noncoronary cusp is slightly larger than the right and left. The left is usually the smallest of the three. Because of the crescent shape of the aortic cusps and the fact that their free margins extend from commissure to commissure, the diameter of the aortic valve orifice must be smaller than the length of the free margins. Indeed, anatomic studies of fresh human aortic roots demonstrated that the average length of the free margins of the aortic cusps was one-third longer than the diameter of the aortic orifice. The diameter of the aortic annulus is 15 to 20% larger than the diameter of the sinotubular junction in children, but this relationship changes with age and the diameter of the aortic annulus is often smaller than the diameter of aortic annulus in older persons (see Fig. 36-2).

The aortic annulus, the aortic cusps, and the sinotubular junction play an important role in maintaining valve competence. On the other hand, the aortic sinuses play no role in valve competence, but they are believed to be important in minimizing mechanical stress on the aortic cusps during the cardiac cycle by creating eddies currents between the cusps and the aortic sinuses.

All components of the aortic root are very elastic and compliant in children but compliance decreases with aging as elastic fibers are replaced by fibrous tissue. Expansion and contraction of the aortic annulus during the cardiac cycle are heterogeneous probably because of its attachments to contractile myocardium as well as to fibrous structures such as the membranous septum and intervalvular fibrous body. On the other hand, the expansion and contraction of the sinotubular junction are more uniform. The aortic root also displays some degree of torsion during isovolumic contraction and ejection of the left ventricle. The movements of the aortic annulus, cusps, sinuses, and sinotubular junction also change with aging as elastic fibers are replaced by fibrous tissue.

Anatomically normal aortic cusps may become calcified late in life and cause aortic stenosis. This type of lesion is called dystrophic calcification, senile calcification, or degenerative calcification. The range of histopathologic lesions includes calcification, chondroid and osseous metaplasia, neorevascularization, inflammation, and lipid deposition.

Bicuspid aortic valve is believed to occur in 1 to 2% of the population. Movahed and colleagues1 recently reviewed the echocardiograms of 24,265 patients who had the study for various clinical reasons and 1742 teenage athletes in Southern California, and found a prevalence of bicuspid aortic valve of 0.6% in the large cohort and 0.5% in the smaller one. Males are affected more than females at a ratio of 4:1. There is a relatively high incidence of familial clustering, which suggests an autosomal dominant inheritance with reduced penetrance.2 Extensive research in the genetics of bicuspid aortic valve is being presently conducted and this disorder is likely heritable. Most patients with bicuspid aortic valve have three aortic sinuses and two cusps of different sizes. The larger cusp often contains a raphe instead of a commissure. The raphe extends from the mid portion of the cusp to the aortic annulus, and its insertion in the aortic root is at a lower level than the other two commissures. Bicuspid aortic valves with two aortic sinuses and no raphe are least common and called "type 0"; the most common is with one raphe and is called "type 1"; and finally with two raphes is "type 2."3 Types 1 and 2 can be subclassified according to the fused cusps: L-R is the most common form (a raphe in between the left and right cusps). Most patients with bicuspid aortic valves have a dominant circumflex artery and a small right coronary artery. Bicuspid aortic valve may function satisfactorily until late in life when it may become calcified and stenotic.4 It may also become incompetent, particularly in younger patients and is often associated with dilated aortic annulus and cusp prolapsed.

Other congenital anomalies of the aortic valve are the unicusp and quadricuspid valves. Subaortic membranous ventricular septal defect can cause aortic insufficiency because of distortion of the aortic annulus and prolapse of the right cusp.

Numerous connective tissue disorders (ankylosing spondylitis, osteogenesis imperfecta, rheumatoid arthritis, Reiter's syndrome, lupus, etc.) can cause aortic insufficiency. The anorexigenic drugs phentermine and fenfluramine can also cause aortic insufficiency. Rheumatic aortic valve disease is still prevalent in developing countries and can cause aortic stenosis and/or insufficiency by causing fusion, fibrosis, and contraction of the cusps.

Degenerative diseases of the media with aneurysm formation are the most common disorders of the aortic root and ascending aorta. A broad spectrum of pathologic and clinical entities is grouped under degenerative disorders, and it ranges from severe degeneration of the media, which can become clinically important early in life in cases such as Loyes-Dietz syndrome, to cases of the not so important mild dilation of the ascending aorta in elderly patients. Bicuspid and unicusp aortic valve disease often display premature degeneration of the media with dilation of the aorta. Atherosclerosis, infectious and noninfectious aortitis are other pathologic entities.

Aneurysms of the ascending aorta are often caused by cystic medial degeneration (cystic medial necrosis). Histologically, necrosis and disappearance of muscle cells in the elastic lamina, and cystic spaces filled with mucoid material are often observed. Although these changes occur more often in the ascending aorta, they may affect any portion or the entire aorta. These changes weaken the arterial wall, which dilates and forms a fusiform aneurysm. The aortic root may be involved in this pathologic process, and in patients with Marfan syndrome, the aneurysm usually begins in the aortic sinuses. A large proportion of patients with aortic root aneurysms do not fulfill the criteria of diagnosis of Marfan syndrome, but the gross appearance of the aneurysm and the histology of the arterial wall may be indistinguishable from that of Marfan syndrome. These cases are referred to as forma frusta of Marfan syndrome.

Patients with aortic root aneurysms are usually in their second or third decade of life when the diagnosis is made. These patients develop aortic insufficiency because of dilation of the sinotubular junction and/or aortic annulus (Fig. 36-3). Annuloaortic ectasia is a term used to describe dilation of the aortic annulus.

Other patients have relatively normal aortic roots but develop ascending aortic aneurysms. These patients are usually in their fifth or sixth decade of life. Finally, certain patients have extensive degenerative disease of the entire aorta and develop the so-called mega-aorta syndrome with dilation of the entire thoracic and abdominal aorta. Ascending aortic aneurysm may cause dilation of the sinotubular junction with consequent aortic insufficiency (Fig. 36-3).

Marfan Syndrome

Marfan syndrome is an autosomal dominant variably penetrant inherited disorder of the connective tissue in which cardiovascular, skeletal, ocular, and other abnormalities may be present to a variable degree. The prevalence is estimated to be around 1 in 5000 individuals. It is caused by mutations in the gene that encodes fibrillin-1 (FBN1) on chromosome 15. This is a large gene (approximately 10,000 nucleotides in the mRNA), and identification of the mutation is a complex task. More than 1000 mutations in FBN1 have been identified. The phenotype presents a highly variable degree because of varying genotype expression.

The clinical features of Marfan syndrome were thought to be a result of weaker connective tissues caused by defects in fibrillin-1, a glycoprotein, and principal component of the extracellular matrix microfibril. This concept was inadequate to explain the overgrowth of long bones, osteopenia, reduced muscular mass, and adiposity and craniofacial abnormalities often seen in Marfan syndrome.5 Dietz and colleagues5,6 showed in an experimental mouse with Marfan syndrome that many findings are the result of abnormal levels of activation of tranforming growth factor beta (TGF-β), a potent stimulator of inflammation, fibrosis, and activation of certain matrix metalloproteinases, especially matrix metalloproteinases 2 and 9. Excess TGF-β activation in tissues correlates with failure of lung septation, development of a myxomatous mitral valve, and aortic root dilation in mice. This combination of structural microfibril matrix abnormality, dysregulation of matrix homestasis mediated by excess TGF-β, and abnormal cell-matrix interactions is responsible for the phenotype features of the Marfan syndrome. Ongoing destruction of the elastic and collagen lamelae and medial degeneration result in progressive dilation of the aortic root, as well as a predisposition to aortic dissection from the loss of appropriate medial layer support. Loss of elasticity in the media causes increased aortic stiffness and decreased distensibility.

The diagnosis of Marfan syndrome is made on clinical grounds, and it is not always simple because of the variability in clinical expression. A multidisciplinary approach is needed to diagnose and manage patients afflicted with this syndrome. Table 36-1 shows the criteria for the diagnosis of Marfan syndrome. The presence of major criteria in two separate systems and involvement of a third (minor or major) are needed to establish the diagnosis.7

The most common cardiovascular features are aortic root aneurysm and mitral valve prolapse. These anatomical abnormalities may cause aortic rupture, aortic dissection, aortic insufficiency, and mitral insufficiency.

Mutations in the genes encoding TGF-ß receptors 1 and 2 have been found in association with a continuum of clinical features. On the mild end, the mutation have been found in association with presentation similar to that of the Marfan syndrome or with familial thoracic aneurysm and dissection, and on the severe end, they are associated with a complex phenotype in which aortic dissection or rupture commonly occurs in childhood.8 This complex phenotype is characterized by the triad of hypertelorism, bifid uvula or cleft palate, and generalized arterial tortuosity with widespread vascular aneurysm and dissection. This phenotype has been classified as Loeys-Dietz syndrome. Affected patients have a high risk of aortic dissection or rupture at an early age and at relatively small diameters. CT angiograms should be obtained from head to pelvis.

Ehlers–Danlos Syndrome

Vascular Ehlers–Danlos syndrome is a rare autosomal dominant inherited disorder of the connective tissue resulting from mutation of the COL3A1 gene encoding type III collagen. Spontaneous rupture without dissection of large and medium-caliber arteries such as the abdominal aorta and its branches, the branches of the aortic arch, and the large arteries of the limbs accounts for most deaths. Aortic root dilation was present in 28% in a series of 71 patients with Ehlers–Danlos syndrome.9 Aortic dissection is uncommon. Diagnosis is confirmed either by biochemical assays showing qualitative or quantitative abnormalities in type III collagen secretion or by molecular biology studies demonstrating mutation of the COL3A1 gene. Varied molecular mechanisms have been observed with different mutations in each family. No correlation has been established between genotype and phenotype. Diagnosis should be suspected in any young person presenting with arterial or visceral rupture or colonic perforation.

Other Pathologies

Atherosclerotic aneurysms of the ascending aorta are uncommon. They are more common in the abdominal aorta and to a lesser degree in the descending thoracic aorta. Atherosclerosis often causes irregular and saccular aneurysms of the ascending aorta rather than a more fusiform shape as those caused by degenerative disease of the media.

Infectious aneurysms of the ascending aorta are rare. Syphilis was a common cause of aneurysm of the ascending aorta but it is seldom seen. The spirochetal infection destroys the muscular and elastic fibers of the media, which are replaced by fibrous and other inflammatory tissues. The ascending aorta is the most common site of involvement and the aneurysm is usually saccular. The wall of the ascending aorta is frequently calcified. Syphilitic aortitis also causes coronary ostial stenosis and aortic valve insufficiency. Other bacteria can also cause aneurysm of the ascending aorta.

Various types of aortitis may involve the ascending aorta. Giant cell arteritis is among the more common and it involves medium-sized arteries, but the aorta and its branches are involved in approximately 15% of the cases. The etiology is unknown. The characteristic lesion is a granulomatous inflammation of the media of large and medium-caliber arteries such as the temporal artery. Occasionally the inflammatory process weakens the aorta leading to aneurysm formation, aortoannular ectasia, and aortic insufficiency.

Ankylosinsg spondylitis, Reiter's syndrome, psoriatic arthritis, and polyarteritis nodosa can cause aortic insufficiency because of annuloaortic ectasia. Behçet's disease can cause aneurysm of the ascending aorta.

Aortic Stenosis

Asymptomatic patients with aortic stenosis have a good prognosis.10 Sudden death in asymptomatic patients is uncommon. However, when symptoms develop, the prognosis becomes poor and the average survival is 2 to 3 years for patients with symptoms of angina and syncope, and 1 to 2 years for those with congestive heart failure.11

Aortic Insufficiency

The prognosis of symptomatic patients with aortic insufficiency is poor with death occurring within 4 years after development of angina and within 2 years after the onset of congestive heart failure.12

Bicuspid Aortic Valve Disease

A prospective follow-up on 642 adult patients (mean age 35±9 years) with bicuspid aortic valve during a mean 9±5 years in our institution revealed a late survival similar to that of the general population in spite the fact that 161 patients had an adverse event (cardiac death, aortic valve/ascending aorta surgery, dissection or congestive heart failure).13 Age greater than 30 years, and moderate or severe aortic stenosis or insufficiency were independent predictors of adverse event.13 In a study from the Mayo Clinic on 212 patients (mean age 32±20 years) with normal or minimally dysfunctional bicuspid aortic valve who lived in Olmsted County and followed for a mean of 15±6 years revealed a 20-year survival similar to that of the general population but the incidence of surgery on the aortic valve and/or ascending aorta was 27±4% and the total adverse cardiovascular events was 42±5%.4

Aortic Root and Ascending Aortic Aneurysms

Aortic root/ascending aortic aneurysm can cause aortic insufficiency, aortic dissection or rupture. The transverse diameter of the aneurysm is the most important predictor of rupture or dissection. In a study by Coady and associates14 on 370 patients with thoracic aneurysms (201 ascending aortic aneurysms), during a mean follow-up of 29.4 months the incidence of acute dissection or rupture was 8.8% for aneurysms less than 4 cm, 9.5% for aneurysms of 4 to 4.9 cm, 17.8% for 5 to 5.9 cm, and 27.9% for those greater than 6 cm. The median size of the ascending aortic aneurysm at the time of rupture or dissection was 5.9 cm. The growth rates of thoracic aneurysms are exponential.14 In Coady's study, the growth rate ranged from 0.08 cm/year for small (<4 cm) aneurysms to 0.16 cm/year for large (8 cm) aneurysms.14 The growth rates for chronic dissecting aneurysms were much higher than for chronic nondissecting aneurysms.

The growth rates for aortic root aneurysms may be higher than in ascending aortic aneurysms, particularly in patients with Marfan syndrome. Aortic dissection is rare in aortic root aneurysm of less than 50 mm, unless the patients have family history of aortic dissection. Without surgery most patients with Marfan syndrome die in the third decade of their lives from complications of aortic root aneurysm such as rupture, aortic dissection, or aortic insufficiency.15 Pregnancy in women with Marfan syndrome has two potential problems: the risk of having a child who will inherit the disorder and the risk of acute aortic dissection during the third trimester, parturition, or the first month postpartum. The offspring has a 50% risk of inheriting the syndrome.

Patients affected with Loeys-Dietz syndrome have a high risk of aortic dissection or rupture at an early age and at relatively small aortic root diameters. Surgery is recommended in adults when the aortic root exceeds 4 cm.

Patients with bicuspid aortic valve had larger aortas than patients with tricuspid aortic valve and the rate of dilation of the ascending aorta was higher (0.19 cm versus 0.13 cm per year respectively) in a study by Davies et al.16 Among patients with bicuspid aortic valve, those with aortic stensois had a higher risk of rupture, dissection, or death.

Patients with aortic stenosis remain asymptomatic for many years. Symptoms usually appear late in the course of the disease. The symptoms are: angina pectoris, syncope, congestive heart failure, and syncope. The clinical presentation of patients with aortic insufficiency is dependent on the rapidity with which aortic insufficiency develops. Patients with chronic aortic insufficiency remain asymptomatic for many years while the heart slowly enlarges. Palpitations and head pounding may occur during exertion. Angina pectoris may occur, but it is not as common as with aortic stenosis. Syncope is rare. Symptoms of congestive heart failure are usually an indication of left ventricular dysfunction. Acute aortic insufficiency is frequently associated with cardiovascular collapse, with extreme fatigue, dyspnea, and hypotension resulting from reduced stroke volume and elevated left atrial pressure. Echocardiography confirms the diagnosis of aortic valve dysfunction and provides information regarding its mechanism. Radionuclide imaging is useful in assessing the left ventricular function at rest and during exercise, valuable information in asymptomatic patients.

Most patients with aortic root aneurysm are asymptomatic and have no physical signs if the aortic valve is competent. Some patients may complain of vague chest pain. Severe chest pain is suggestive of rapid expansion or intimal tear with dissection. Echocardiography establishes the diagnosis and provides information regarding the aortic cusps. CT scan and MRI of the chest are also diagnostic and it is useful to provide information regarding the thoracic aorta.

Surgeons must be familiar with the guidelines for heart valve surgery established by a joint committee of the ACC and AHA.17

Most candidates for aortic valve repair have aortic insufficiency or normally functioning aortic valve with aortic root aneurysm. Transesophageal echocardiography is the best diagnostic tool to study the aortic valve and the mechanism of aortic insufficiency. Each component of the aortic root must be carefully interrogated, and in particular the aortic cusps. The number of cusps, their thickness, the appearance of their free margins, and the excursion of each cusp during the cardiac cycle must be examined in multiple views. The lines of coaptation of the aortic cusps should be interrogated by color Doppler imaging. The direction and size of the regurgitant jets recorded in many views. Information regarding the morphologic features of the aortic annulus, aortic sinuses, sinotubular junction, and ascending aorta should be obtained.

Obviously, the aortic cusps are the most important determinant of aortic valve repair. If the cusps are thin, mobile, and smooth free margins, the feasibility of aortic valve repair is very high, including bicuspid aortic valve. Calcified or scarred and fibrotic aortic cusps preclude aortic valve repair unless autologous or xenograft glutaraldehyde fixed pericardium is used for cusp augmentation.

Patients with aortic root aneurysm often have normal or minimally stretched aortic cusps and reconstruction of the aortic root with preservation of the native aortic cusps is feasible. Grossly dilated aortic annulus and/or sinotubular junction are likely to have overstretched, thinned out aortic cusps with stress fenestrations along the commissural areas and may not be suitable for repair.

Cusp Perforation

Occasionally a cups perforation is the sole reason for aortic insufficiency. The perforation may be iatrogenic, a sequelae of healed endocarditis, or the result of resection of a papillary fibroelastoma. A simple patch of fresh or glutaraldehyde fixed autologous pericardium is adequate to correct the problem. Fresh autologous pericardium can also be used to repair small holes (<5 mm) but the patch should be larger than the defect because it retracts during healing. We use a continuous 7-0 polypropylene to suture the patch around the defect on the aortic side of the cusp.

Cusp Extension

Cusp augmentation has been used to repair incompetent aortic valves due rheumatic and congenital disease. Glutaraldehyde-fixed bovine or autologous pericardium has been used for this purpose.

Cusp Prolapse

Cusp prolapse is caused by elongation of the free margin. This is corrected by plication along the nodule of Arantius as illustrated in Fig. 36-4. The degree of shortening is determined by examining the other cusps and their level of coaptation.

Cusp with Stress Fenestration

Dilation of the sinotubular junction increases the mechanical stress along the free margin of the cusp near the commissures and may cause a stress fenestration with thinning and sometimes even detachment from the commissure. This type of lesion has been successfully corrected by weaving a double layer of 6-0 expanded polytetrafluoroethylene suture along the free margin of the cusp as illustrated in Fig. 36-5.

Bicuspid Aortic Valve

The most commonly performed aortic valve repair in adults is for bicuspid aortic valve with prolapse of one cusp. Although the anatomic arrangement of the bicuspid aortic valve varies, most patients have an anterior cusp attached to the interventricular septum and a posterior cusp attached to the fibrous components of the left ventricular outflow tract. The anterior cusp often contains a raphe at approximately where the commissure between the right and left cusps would be (Type 1, L-R). This cusp is usually the one that is elongated and prolapsed. As long as the posterior cusp is normal, repair is feasible and relatively simple. The raphe is excised and the free margin of the anterior cusp is shortened with plicating sutures as illustrated in Fig. 36-6. The lengths of the free margins of both cusps should be similar and should coapt at the same level. Suspending the arterial walls immediately above the commissures and observing the level of coaptation of each cusp gives an estimate of the level of coaptation of the cusps.

Because most patients with incompetent bicuspid aortic valves have dilated aortic annulus, aortic valve sparing operation may be more appropriate procedure than simple cusp repair. However, if the aortic sinuses are not aneurismal and the annulus is only mildly dilated, a reduction annuloplasty to increase the coaptation area of the cusps can be done. This is accomplished by plicating the subcommissural triangles using horizontal mattress sutures of 4-0 polypropylene with Teflon felt pledgets on the outside of the aortic root (Fig. 36-7). The suture is first passed from the outside to the inside of the aorta through the aortic annulus of both cusps 2 mm below their commissure. The same suture is passed again through the annulus and subcommissural triangle 4 or 5 mm below the first one and the ends are tied together over Teflon felt pledgets on the outside of the aorta.

Aortic valve sparing operations include various procedures used to preserve the aortic cusps in patients with aortic root aneurysm or ascending aortic aneurysm with aortic insufficiency.18,19

Ascending Aortic Aneurysm with Aortic Insufficiency

Dilation of the sinotubular junction displaces the commissures of the aortic valve outward and prevents the cusps from coapting during diastole (see Fig. 36-3). These patients are often on the sixth, seventh, or eighth decade of their lives and have ascending aortic aneurysm. The dilation of the sinotubular junction is often asymmetrical, and the commissures of noncoronary aortic cusp are more affected than the other two. If the other components of the aortic root are normal, simple adjustment of the sinotubular junction restores valve competence. This is accomplished by transecting the ascending aorta 5 mm above the sinotubular junction and pulling the three commissures upward and close to each other until the cusps coapt. The three commissures form an imaginary triangle. The diameter of a circle that contains this imaginary triangle is the diameter of the graft that should be used to reconstruct the sinotubular junction. Because the aortic cusps and sinuses have different sizes, this triangle is not always equilateral and the commissures must be spaced according to the length of the free margin of each cusp. The diameter of the graft and the space between commissures are facilitated by sizing the diameter of the circle that contains all three commissures with a transparent valve sizer, such as the one used to size the aortic annulus for a stentless porcine aortic valve. Those valve sizers have three equidistant marks, and one can determine the space between the commissures by comparing to the distance between marks. The tubular Dacron graft is sutured right at the level of the sinotubular junction with a continuous 4-0 polypropylene suture (see Fig. 36-7). If after adjusting the sinotubular junction, the cusps do not coapt at the same level, one or more cusps may be elongated and the free margin has to be shortened as illustrated in Fig. 36-4. Aortic valve competence can be tested at this time by injecting cardioplegia into the graft under pressure and observing the left ventricle for distension.

If the noncoronary aortic sinus is dilated or altered by aortic dissection, a neosinus can be created by tailoring the graft with tongue of tissue that is sutured directly to the aortic annulus as illustrated in Fig. 36-8. The height of the neosinus of Dacron should be 3 or 4 mm more than the diameter of the graft, and the width should be 3 or 4 mm more than the estimated intercommissural distance to allow the graft to bulge and create a neoaortic sinus.

Small diameter grafts (<24 mm) should be avoided in adult patients because they may increase left ventricular afterload, particularly if long segments of aorta are replaced such as with concomitant transverse arch replacement using the elephant trunk technique. If the estimated diameter of the sinotubular junction is less than 24 mm, a larger graft should be used and the end that is anastomosed to the aortic root to recreate the sinotubular junction should be reduced by plicating the graft in the area of the anastomosis.

Aortic Root Aneurysm

A large proportion of patients with aortic root aneurysm have normal or minimally stretched aortic cusps and an aortic valve sparing operation is feasible. There are basically two types of aortic valve-sparing operations for patients with aortic root aneurysm: remodeling of the aortic root and reimplantation of the aortic valve.18,19

Remodeling of the Aortic Root

The ascending aorta is transected and the aortic root is dissected circumferentially down to the level of the aortic annulus. All three aortic sinuses are excised, leaving approximately 4 to 6 mm of arterial wall attached to the aortic annulus and around the coronary artery orifices as illustrated in Fig. 36-9. The three commissures are gently pulled vertically and approximated until the cusps coapt. The three commissures form a triangle and the diameter of the circle that contains that triangle is the diameter of the graft to be used for remodeling. In our experience, most grafts are 24, 26, or 28 mm in diameter. Here again the stentless valve sizers are very useful to determine the diameter of the graft and also the distance between commissures because they may not be equidistant. The spaces in between the commissures are marked in one of the ends of the graft, and the graft is tailored to create three neo-aortic sinuses (Fig. 36-10). The heights of these neosinuses should be approximately equal to the diameter of the graft. The three commissures are suspended in the graft (Fig. 36-11), which is then sutured to the remnants of the aortic wall as close as possible to aortic annulus with continuous 4-0 polypropylene sutures. The coronary arteries are reimplanted into their respective neosinuses. The aortic cusps are inspected to make sure that all three coapt at the same level and well above the nadir of the aortic annulus. If one or more cusp is prolapsing, the free margin is shortened as described previously. If one or two cusps have stress fenestrations, the free margin should be reinforced with a fine expanded polytetrafluoroethylene suture. Aortic valve competence can be assessed by injecting cardioplegia under pressure into the reconstructed aortic root and observing the left ventricle for distension or by turning the echocardiography machine on. The graft is then anastomosed to the distal ascending aorta or transverse aortic arch graft depending on the extent of the aneurysm (Fig. 36-11).

Remodeling of the aortic root may be inappropriate for patients with Marfan syndrome or annuloaortic ectasia because the annulus may continue to dilate and cause aortic insufficiency. An aortic annuloplasty along the fibrous component of the left ventricular outflow tract18 did not prevent late dilation of the aortic annulus in patients with Marfan syndrome in our experience. Thus, reimplantation of the aortic valve may be a better operative procedure for patients with annuloaortic ectasia (Fig. 36-12) because this operative procedure corrects and prevents annular dilation whereas remodeling of the aortic root may be more suitable for patients with normal aortic annulus.

Reimplantation of the Aortic Valve

This procedure can be performed in all patients with aortic root aneurysm, but it is particularly valuable in patients with annuloaortic ectasia and in those with acute type A aortic dissection. It is technically more demanding than remodeling of the aortic root because it requires greater knowledge of the functional anatomy of the aortic root. This is because the aortic annulus, the aortic sinuses, the sinotubular junction, and even the aortic cusps are reconstructed. In the original description of this procedure, the aortic valve was reimplanted into a tubular Dacron graft and no neoaortic sinuses were created. Some investigators have suggested that the presence of the aortic sinuses is important for normal cusp motion, and potentially, cusp durability. Several modifications to the reimplantation procedure were introduced to create neoaortic sinuses. There is now a commercially available graft with sinuses of Valsalva (Vascutek Ltd. Renfrewshire, Scotland); however, this graft distorts the aortic annulus because its sinuses are spherical whereas the normal aortic annulus develops along a crescent shape on a single horizontal plane. A cylindrical graft with neoaortic aortic sinuses that permits reimplantation of the aortic annulus along single horizontal planes has also been developed in Germany.20

Following is a description of this procedure as we have been performing it during the past decade, with excellent functional results. The three aortic sinuses are excised as described for the remodeling procedure (see Fig. 36-9). Multiple horizontal mattress sutures of 2-0 or 3-0 polyester are passed from the inside to the outside of the left ventricular outflow tract, immediately below the nadir of the aortic annulus, through a single horizontal plane along the fibrous portion of the outflow tract and along its scalloped shape in the interventricular septum as illustrated in Fig. 36-13. If the fibrous portion is thin, sutures with Teflon felt pledgets should be used. A tubular Dacron graft of diameter equal to double the average height of the cusps is selected and three equidistant marks placed in one of its ends. A small triangular segment is cut off along the mark that corresponds to the subcommissural triangle of the left and right cusps. If the diameter of the aortic annulus is smaller than the diameter of the graft by 10 mm or more, the diameter of the graft is reduced by plicating it in the areas corresponding to the nadir of the aortic annulus. The sutures previously placed in the left ventricular outflow tract are now passed through the graft. The sutures should be spaced symmetrically if the aortic annulus is not dilated. If there is obvious dilation of the aortic annulus, the sutures should be spaced symmetrically along the muscular interventricular septum and around the nadirs of the aortic annulus but closer together beneath the subcommissural triangles of the noncoronary cusp because that is where dilation occurs in patients with connective disorders. The sutures are then tied on the outside of the graft. Care must be exercised not to purse-string this suture line. The graft is then cut in a length of approximately 5 cm and pulled gently, and the three commissures are also pulled vertically and temporarily secured to the graft with transfixing 4-0 polypropylene sutures buttressed on small Teflon felt pledgets, but these sutures are not tied. Once all three commissures are suspended inside the graft, the commissures and the cusps are inspected to make sure they are all correctly aligned. The subcommissural triangles should also be inspected to make sure they are as narrow as the diameter of the graft allows, i.e., the triangles should have a narrower base than before surgery. Next, the sutures are tied on the outside of the graft and used to secure the aortic annulus into the graft. This is accomplished by passing the suture sequentially from the inside to the outside right at the level of the annulus and from the outside to the inside at the level of the remnants of the arterial wall. We start at the level of the commissure and stop at the nadir of the aortic annulus where the sutures are tied together. The coronary arteries are reimplanted into their respective sinuses (Fig. 36-14). The coaptation of the aortic cusps is inspected and prolapse is corrected if necessary. It is important that the coaptation level is well above the aortic annulus. Neoaortic sinuses are created by plicating the graft at the level of the commissure as illustrated in Fig. 36-15. Valve competence can be assessed by occluding the distal graft and injecting cardioplegia under pressure. If the ventricle does not distend, no more than trace aortic insufficiency is present. The aortic valve can also be examined by echocardiography during injection of cardioplegia. The distal anastomosis is performed either to the distal ascending aorta or transverse aortic arch depending on the pathology. The graft sizes in our patients ranged from 26 to 34 mm, mean of 31 mm.

One of the earliest series of aortic valve repair for aortic insufficiency caused by prolapse of bicuspid aortic valve came from the Cleveland Clinic.21 In a series of 94 patients with a mean age of 38 years, the freedom from reoperation was 84% at 7 years.21 The only factor predictive of reoperation was residual aortic insufficiency at the time of repair.21

The appropriateness of aortic valve repair in patients with incompetent bicuspid aortic valve remains unclear. Competent bicuspid aortic valves appear to be durable because a large proportion of patients who require aortic valve replacement for aortic stenosis in their fifth, sixth, or seventh decades of life are found to have bicuspid aortic valve. Thus, aortic valve repair for incompetent bicuspid aortic valves is a reasonable surgical approach in young adults but the best type of repair remains to be determined. Incompetent bicuspid aortic valves are often associated with dilated aortic annulus and subcommissural plication may be inadequate to prevent future dilation and recurrent aortic insufficiency. A conservative aortic root procedure may be more appropriate for some of these patients.

Ascending Aortic Aneurysm with Aortic Insufficiency

We reported our experience with aortic valve repair in patients with ascending aortic aneurysm, normal or minimally dilated aortic sinuses, and moderate or severe aortic insufficiency.22 There were 103 patients whose mean age was 65±12 years and 53% were men. The aneurysm extended into the transverse aortic arch in 60% of the patients and 20% had mega-aorta syndrome. The aortic valve repair consisted in adjusting the diameter of the sinotubular junction in all patients. In addition, repair of cusp prolapse was needed in 36 patients, and replacement of the noncoronary aortic sinus in 8. Associated procedures were: replacement of the transverse aortic arch in 62 patients, coronary artery bypass in 28, and mitral valve repair or replacement in 7. The follow-up was complete at 5.8±2.3 years. There were 2 operative and 30 late deaths. The survival at 10 years was 54±7%. Independent predictors of late death were transverse arch aneurysm, the use of elephant trunk technique to replace the arch and mega-aorta syndrome. Only 2 patients required aortic valve replacement: one for endocarditis and one for severe aortic insufficiency. The freedom from aortic valve replacement at 10 years was 98%. Only one patient developed severe and six developed moderate aortic insufficiency during the follow-up. The freedom from severe or moderate aortic insufficiency at 10 years was 80±7%. These findings suggest that aortic valve repair in these patients is an excellent alternative to valve replacement and the repair remains stable in most patients during the follow-up. Late survival was suboptimal because of the extensiveness of the vascular disease.

Aortic Root Aneurysm

We reported our current experience with aortic valve sparing operations for aortic root aneurysm in 220 patients.23 Their mean age was 46±15 years, 78% were men and 40% had the Marfan syndrome. In addition, 17% had type A aortic dissection, 7% had bicuspid aortic valve, and 22% had transverse arch aneurysm. Previous replacement of the ascending aorta had been done in 10 patients and the Ross procedure in 2. Sixteen patients had severe mitral insufficiency. Approximately one-half of the patients had moderate or severe aortic insufficiency before surgery. The technique of remodeling of the aortic root was used in 53 patients and the reimplantation of the aortic valve in 167. The follow-up was complete at 5.2±3.7 years. All patients had echocardiographic studies during the follow-up. There were 3 operative and 13 late deaths. Patients' survival at 10 years was 88±3% and was similar to that of the general population of Ontario. Age greater than 65 years, advanced functional class, and ejection fraction less than 40% were independent predictors of death. Seven patients developed moderate and 5 developed severe aortic insufficiency. Overall freedom from moderate or severe aortic insufficiency at 10 years was 85±5%, but it was 94±4% after reimplantation of the aortic valve and 75 ±10% remodeling of the aortic root (p = 0.04). Five patients required aortic valve replacement; the freedom at 10 years was 95±3%. One patient developed endocarditis 11 years postoperatively, and 8 suffered thromboembolic events. At the latest follow-up 88% of the patients were in functional class I, and 10% were in class II. These findings suggested that reimplantation of the aortic valve provides more stable aortic valve function than remodeling of the aortic root.

Yacoub et al.24 who used exclusively the remodeling of the aortic root to treat 158 patients with aortic root and ascending aortic aneurysms, reported a freedom from aortic valve replacement of 89% at 10 years, and moderate aortic insufficiency in one-third of the patients.

Aicher et al.25 who also used the remodeling of the aortic root in 274 patients (mean age 59±15 years, 193 tricuspid and 81 bicuspid valves) reported a freedom from aortic insufficiency grade II or greater of 91% for bicuspid and 87% for tricuspid at 10 years, and a freedom from aortic valve replacement of 98% at 5 and at 10 years.

We recently reported the long-term results with aortic valve sparing operations in patients with Marfan syndrome.26 A cohort of 103 consecutive patients with Marfan syndrome with a mean age of 37±12 years underwent either remodeling of the aortic root (26 patients) or reimplantation of the aortic valve (77 patients) and were prospectively followed for a mean of 7.3±4.2 years with annual echocardiography. The survival at 15 years was 87.2%, somewhat lower than the 95.6% for the general population largely because of complications of aortic dissection. The freedom from aortic insufficiency greater than mild was 79.2%. Only 3 patients required aortic valve replacement, two for aortic insufficiency and one for endocarditis.

There are several reports on early clinical and hemodynamic outcomes on these two types of aortic valve sparing procedures.27-30 Most comparative studies suggest that the reimplantation of the aortic valve provides a more stable aortic valve function than the remodeling of the aortic root, particularly in patients with dilated aortic annulus, Marfan syndrome, and acute type A aortic dissection. Hemodynamic studies suggest that cusp motion and flow patterns across the reconstructed aortic root are more physiologic after remodeling of the aortic root than reimplantation of the aortic valve.29 In patients who had the reimplantation procedure, flow patterns and cusp motion are better with neoaortic sinuses than without.30 However, in our series of reimplantations of the aortic valve, there was no difference in aortic valve function after 10 years in patients with a straight tube or with neoaortic sinuses. Aortic sinuses seem to decrease the mechanical stress on the aortic cusps but it is not clinically apparent during the first decade of follow-up as far as durability of the procedure.

Aortic valve sparing operations may be inappropriate for young children because of future mismatch between the size of the graft and somatic growth.

Another important question regarding aortic valve sparing operations is whether they are better than the Bentall procedure with mechanical valves. There has been no randomized clinical trial comparing these two procedures for the treatment of aortic root aneurysms or ascending aortic aneurysm but retrospective studies in patients with Marfan syndrome suggest that the outcomes may be similar.

We believe that aortic valve sparing operations offer an ideal method for treating patients with aortic root aneurysm and normal or minimally diseased aortic cusps. When correctly performed, they provide excellent results and are associated with very low rates of valve-related complications. However, because they are technically demanding operations, only surgeons with extensive experience in aortic surgery should perform them. The surgeon must have a sound knowledge of anatomy and pathology of the aortic valve and be able to apply the concepts of functional anatomy to create an anatomically and functionally satisfactory new aortic root.