Chapter 33. Bioprosthetic Aortic Valve Replacement: Stented Pericardial and Porcine Valves

This chapter provides an overview of aortic valve replacement (AVR) with stented bioprostheses. The indications for aortic valve surgery are reviewed, with an emphasis on current evidence-based guidelines, and currently available stented aortic bioprostheses are described. Clinical and physiologic outcomes of aortic valve surgery are critically evaluated to create a rational basis for prosthesis selection.

Aortic Stenosis

Natural History

Aortic stenosis (AS) may be caused by: (1) degenerative calcification; (2) congenital malformations, most commonly bicuspid aortic valve; or (3) rheumatic fever. It is uncommonly associated with systemic diseases such as Paget's disease of bone and end-stage renal disease. Degenerative calcification is common in the elderly population and moderate to severe calcific AS is the most common pathology among patients undergoing AVR. The incidence of calcific AS is expected to rise with the increase in age of population in developed countries.

Valvular degenerative calcification is characterized as a progressive reduction of orifice cross-sectional area caused by calcification of the cusps. The pathogenesis of inflammation and lipid accumulation is reminiscent of atherosclerosis.1 The normal human aortic valve area (AVA) is between 3.0 and 4.0 cm2 with minimal to no gradient. Aortic stenosis is defined as mild, moderate, and severe AS, and the corresponding AVA, mean gradients, and peak jet velocities are shown in Table 33-1. In the presence of normal cardiac output, transvalvular gradient is usually greater than 50 mm Hg when the AVA is less than 1.0 cm2.2 A rapid increase in transvalvular gradient is seen when AVA is less than 0.8 to 1.0 cm2.

Obstruction to flow created by the reduced orifice area of the aortic valve elevates intracavitary pressures, and resultant wall stress leads to compensatory concentric hypertrophy to maintain normal cardiac output.3 With progressive hypertrophy, ventricular compliance progressively decreases and end-diastolic pressure rises.4 In this situation, the contribution of atrial contraction to preload becomes more significant and loss of sinus rhythm may lead to rapid progression of symptoms.

Symptomatic Patients

As AS progresses to become hemodynamically significant, initial compensation occurs by left ventricular hypertrophy (LVH). The average AVA is 0.6 to 0.8 cm2 at the onset of symptoms.3 Gradual progression of outflow obstruction and ventricular hypertrophy lead to the cardinal symptoms of AS, which are: (1) angina, (2) syncope, and (3) dyspnea or congestive heart failure (CHF). Classic natural history studies have demonstrated that average life expectancy in patients with hemodynamically significant AS is 4 years with angina, 3 years with syncope, and 2 years with CHF.5 Symptom development in the context of AS is an absolute indication for surgical intervention.6 Excessive delay of AVR in symptomatic patients is associated with rate of sudden death of greater than 10% per year. Once a patient is symptomatic, average survival is less than 3 years typically from ventricular arrhythmia or CHF.7

Managing asymptomatic patients with hemodynamically significant AS remains controversial. The average decrease in AVA is 0.12 cm2 per year, whereas the average increase in transvalvular pressure is often 10 to 15 mm Hg per year.8 However, considerable variation does exist in the rate of disease progression and many patients do not experience any change in gradient for several years. As such, there may exist a prolonged and stochastic latent periods before symptoms emerge. During this period, concentric LVH progresses as ventricle adapts to elevated chamber pressures.

Sudden death is rather uncommon in the asymptomatic patient unless the degree of AS is severe, in which case the rate is approximately 1% per year. However, the vast majority of patients who experience sudden death will become symptomatic in the months immediately before the fatal event.9

When considering AVR, the risk of sudden death should be weighed against the contemporary STS database surgical mortality of 3.5%. Although it is difficult to predict who will need surgical intervention, asymptomatic patients with increases in peak velocity jet greater than 0.45 m/s per year are substantially more likely to need an operation.9 Overall, up to 7% of asymptomatic patients with AS experience death or aortic valve surgery 1 year after diagnosis.10 After 5 years, the incidence of death or aortic valve surgery surges to 38%. Of note, early and late outcomes were similar among patients with severe AS who underwent surgery with or without symptoms.11

Low-Gradient Severe Aortic Stenosis

The significance of AS is often unclear in patients with very poor ventricular function (ejection fraction <20%) who have severely stenotic valves but small (<30 mm Hg) transvalvular gradients. Compromised left ventricular function (LVF) in these patients may be caused by afterload mismatch created by the stenotic valve or by intrinsic cardiomyopathy, particularly in the setting of chronic ischemia from diffuse coronary disease. In these patients, measurement of transvalvular gradient and valve area at rest and with positive inotropy (eg, dobutamine infusion) may distinguish cardiomyopathy versus true valvular stenosis as the most responsible diagnosis. Most patients with poor LVF and severe AS experience a significant survival benefit from valve replacement12 (Fig. 33-1). However, a subset of patients with a preponderance of cardiomyopathy do not experience significant benefit.13 A multivariate analysis to determine factors that predict poor LVF after AVR for AS identified poor preoperative LVF as the most significant predictor, indicating suboptimal outcome in the low-gradient group of patients.14

Medical Therapy

No known medical therapy has been shown to alter the natural history of AS. Although several small nonrandomized studies have shown a reduction in disease progression with antilipid therapy,15 a recent prospective randomized clinical trial of aggressive antilipid therapy with atorvastatin did not demonstrate decreased rates of disease progression16 (Fig. 33-2).

Indications for Operation

In 1998, a joint task force of the American College of Cardiology (ACC) and the American Heart Association (AHA) developed evidence-based consensus guidelines for management of valvular heart disease.17 These were subsequently updated in 2006.18 Their recommendations for AVR in the setting of AS are summarized in Table 33-2.

Aortic valve replacement is indicated in all symptomatic patients with severe AS or those with severe asymptomatic AS who require concomitant coronary bypass, aortic surgery, or other valve replacement. At Sunnybrook Health Sciences Center, it is our practice to perform AVR on patients with moderate AS requiring concomitant cardiac surgery. We do not routinely perform AVR on patients with mild AS undergoing concomitant cardiac surgery unless the aortic valve is heavily calcified and the stenosis likely to progress rapidly. Aortic valve replacement is commonly performed in otherwise asymptomatic patients with severe AS and severe left ventricular dysfunction (LVD), exercise-induced symptoms, significant hypertrophy, or ventricular arrhythmia. Asymptomatic patients with very high transvalvular gradients (>60 mm Hg) or highly stenotic valves (valve area <0.6 cm2) are at higher risk for progression to symptoms and should have valve replacement before significant ventricular decompensation or sudden death.

Aortic Regurgitation

Acute Aortic Regurgitation

Acute aortic regurgitation (AR) is caused by: (1) acute dilatation of the aortic annulus preventing adequate cusp coaptation; or (2) disruption of the valve cusps themselves. The specific causes of AR include aortic dissection, infective endocarditis, trauma, aortic cusp prolapse secondary to VSD, aortitis (syphilitic or giant cell, and iatrogenic, such as postaortic balloon valvotomy).

The heart is rather intolerant of acute AR. Because the left ventricle is unable to compensate for the sudden increase in end-diastolic volume caused by a large regurgitant volume load, a dramatic reduction in forward stroke volume ensues.19 In the context of a hypertrophic and poorly compliant LV, hemodynamic decompensation is significantly more dramatic. Tachycardia is the initial compensatory response to acute decline in forward stroke volume. Volume overload causes the left ventricular end-diastolic pressure to acutely rise above left atrial pressure, resulting in early closure of the mitral valve.20 Although this protects the pulmonary venous circulation from excessively high end-diastolic pressures, rapid progression of pulmonary edema and cardiogenic shock are unavoidable. Death secondary to progressive cardiogenic shock and malignant ventricular arrhythmias is the common endpoint of acute AR regardless of etiology. Thus urgent surgical treatment is warranted for all causes of hemodynamically significant acute AR.

Chronic Aortic Regurgitation

Chronic AR is caused by either slow enlargement of the aortic root or dysfunction of the valve cusps. Common causes of chronic AR include congenital abnormalities (eg, bicuspid, unicuspid, quadricuspid aortic valve), calcific cusp degeneration, rheumatic fever, endocarditis, Marfan's syndrome, Ehlers-Danlos syndrome, myxomatous proliferation, osteogenesis imperfecta, and ankylosing spondylitis.

Chronic AR causes a persistent volume overload of the left ventricle. This volume burden leads to progressive chamber enlargement without increasing end-diastolic pressure during the asymptomatic phase of the disease. Chamber enlargement is accompanied by adaptive eccentric hypertrophy, associated at the cellular level with sarcomere replication and elongation of myocytes. The combination of chamber dilatation and hypertrophy leads to a massive increase in left ventricular mass. Initially, the ratio of wall thickness to chamber diameter, ejection fraction (EF), and fractional shortening are maintained.21 However, a vicious cycle of chamber enlargement, continually increasing wall stress, and maladaptive ventricular hypertrophy ensues. Development of interstitial fibrosis is an important pathogenic mechanism that limits further ventricular dilitation, elevating end-diastolic pressure and leading to left ventricular systolic dysfunction, and CHF.22 Vasodilator therapy may delay progression of ventricular dysfunction by decreasing afterload thus reducing regurgitant flow. Vasodilator therapy is currently indicated in asymptomatic patients with hypertension; asymptomatic patients with severe AR, ventricular dilatation, and preserved systolic function; and for short-term hemodynamic tailoring before operation. This therapy is not recommended in patients with severe AR and LVD, as it does not improve survival but may be used if such patients are considered inoperable.

The time course from diagnosis of AR to the development of symptoms is highly variable. Natural history studies of AR show that symptoms, LVD, or both develop in less than 6% of patients per year.23 Progression to LVD without symptoms occurs in less than 4% of patients per year, and sudden death occurs in less than 0.2% per year.24 Independent predictors of progression to symptoms, LVD, or death in asymptomatic patients include age, left ventricular end-systolic dimension, rate of change in end-systolic dimension, and resting EF LVD develops, the onset of symptoms occur at a rate exceeding 25% per year.26 Because symptoms such as angina and dyspnea develop only after significant ventricular decompensation has occurred, surgery is advocated before the symptomatic phase of the disease. Symptomatic patients experience greater than 10% mortality per year; thus surgery is absolutely indicated.27

Indications for Operation

Surgical intervention for asymptomatic patients is based on the identification of subtle but measurable changes in myocardial function before they become irreversible and negatively affect the patient's long-term prognosis. Unfortunately, such deleterious changes are often missed by current imaging modalities. Patients with more severe LVD have decreased perioperative and late survival because of irreversible ventricular remodeling, including hypertrophy and interstitial fibrosis.26,28 The decision to operate on this patient population is indeed challenging because the outcomes are poor with surgery or medical therapy.4 A summary of the revised ACC/AHA Task Force guidelines for AVR for chronic AR is presented in Table 33-3.18

Many patients requiring AVR have coexistent coronary artery disease (CAD). In North America, more than one-third of AVR procedures are accompanied with coronary bypass graft surgery. This proportion will indeed increase as the surgical population continues to age. Risk assessment for ischemic heart disease is complicated in patients with aortic valve disease because angina may be related to true ischemia from hemodynamically significant coronary lesions, or other causes such as left ventricular wall stress with subendocardial ischemia or chamber enlargement in the setting of reduced coronary flow reserve. Because traditional coronary risk stratification is unreliable in aortic valve patients, at Sunnybrook Health Sciences Center, it is our practice to routinely perform diagnostic coronary angiography on all patients greater than age 35.

Technique of Operation

Myocardial Protection and Cardiopulmonary Bypass

Isolated AVR is performed using a single, two-stage right atrial venous cannula and an arterial cannula into the ascending aorta for systemic perfusion of oxygenated blood. A retrograde cardioplegia cannula may be placed into the coronary sinus via the right atrium. A left ventricular vent cannula is placed in the right superior pulmonary vein and advanced into the left ventricle to ensure a bloodless field and prevent ventricular distention with aortic insufficiency. Once cardiopulmonary bypass (CPB) is initiated, aorta and pulmonary artery are dissected to expose the anterior aortic root to the left coronary artery. Careful dissection of the pulmonary artery from the aorta ensures that the cross-clamp will be fully occlusive on the aorta and prevents inadvertent opening of the pulmonary artery with the aortotomy incision. Care is needed to present pulmonary artery injury, as this tissue is substantially more friable than the aorta.

After the cross-clamp is applied, myocardial protection is initially delivered as a single dose of high potassium blood through the ascending aorta. This will trigger prompt diastolic arrest unless there is moderate to severe AR. Myocardial protection is maintained by continuous infusion of cold or tepid oxygenated blood cardioplegia delivered via direct cannulation of both coronary ostia after the aorta has been opened. In the case of the short left main coronary artery, antegrade cannulation may preferentially perfuse either the LAD or circumflex system. In the presence of severe coronary artery disease, antegrade cardioplegia may not perfuse myocardial segments distal to significant coronary obstruction. Furthermore, direct cannulation of left main coronary artery risks endothelial injury and potential dissection or promotion of atherosclerosis development.

An alternative method of cardioprotection in aortic valve cases is retrograde cardioplegia, in either intermittent or continuous forms, and used in isolation or in combination with antegrade cardioplegia. This is helpful in patients with significant AR or severe concomitant coronary disease. However, there are some questions regarding the quality of RV perfusion using retrograde alone. If retrograde cannula cannot be guided into the coronary sinus, conversion to bicaval cannulation will allow opening the right atrium and direct placement of the cannula into the coronary sinus. Do not place the retrograde cannula beyond the origin of the right coronary vein ostium in the coronary sinus to ensure adequate right ventricular myocardial protection.

Aortotomy, Valve Excision, and Debridement

Once the cross-clamp has been applied and cardioplegic arrest has been achieved, the aorta is opened either with a transverse or oblique aortotomy. The low transverse aortotomy is a common approach to the aortic valve when using stented bioprostheses or mechanical valves. The aortotomy is started approximately 10 to 15 mm above the origin of the right coronary artery (RCA) and extended anteriorly and posteriorly. The initial transverse incision over the RCA may also be extended obliquely in the posterior direction into the noncoronary sinus or the commissure between the left and noncoronary cusps (Fig. 33-3). The oblique incision is often used in patients with small aortic roots, in whom root enlargement procedures may be required (see the following) and may also be used to tailor a larger ascending aorta.

Morphology of the valve is then inspected (Fig. 33-4). The valve cusps are incised with scissors at the right cusp between the right coronary ostium and the commissure between the right and noncoronary cusps (Fig. 33-5). Mayo scissors or special right-angled valve scissors are usually used at this stage and the calcific deposits are removed from the aortic wall. One to two millimeters of tissue are left behind to support a sewing surface. Right cusp excision is carried first toward the left coronary cusp and then toward the noncoronary cusp and the cusp is removed as a single piece if possible. Excision is then carried toward the left and noncoronary commissure along the noncoronary cusp and then the left coronary cusp. A moistened radiopaque sponge is placed into the outflow area to catch calcific debris, which must be removed before placing the valve sutures. Thorough decalcification is then performed with a scalpel or rongeur. Debridement of all calcium deposits back to soft tissue improves seating of the prosthesis and decreases the incidence of paravalvular leak and dehiscence.

Care must be taken to prevent aortic perforation while calcific deposits are debrided from the aortic wall, particularly at the commissure between the left and noncoronary cusps, where perforation is most likely. Several anatomic relationships must be respected during valve excision (Fig. 33-6). The bundle of His (conduction system) is located below the junction of the right and noncoronary cusps in the membranous septum. Deep debridement in this area can result in permanent heart block. The anterior leaflet of the mitral valve is in direct continuity with the left aortic valve cusp. If it is damaged during decalcification, an autologous pericardial patch is used to repair the defect.

Once debridement is completed, the aortic root is copiously flushed with saline while the left ventricular vent is stopped. To prevent pushing debris into the left ventricle, saline in a bulb syringe is flushed through the left ventricular vent and out the aortic valve in an antegrade manner instead of retrograde through the valve. The irrigation solution is suctioned with the external wall suction and not into the cardiotomy suction.

Valve Implantation

After the native valve has been excised, the annulus is sized with a valve-sizer designed for the selected prosthetic device. The valve is secured to the annulus using 12 to 16 double-needled interrupted 2-0 synthetic braided pledgeted sutures that are alternating in color. The pledgets can be left on the inflow/ventricular side or the outflow/aortic side of the aortic annulus (Figs. 33-7 and 33-8). Placing the pledgets on the inside of the annulus allows supra-annular placement of the valve and generally will allow implantation of a slightly larger prosthesis. In cases in which the coronary ostia are close to the annulus, supra-annular placement may only be possible along the noncoronary cusp. Mattress sutures are first placed in the three commissures and retracted to assist visualization. Some surgeons will place the commissural suture between the right and noncoronary cusps from the outside of the aorta (ie, the pledget is left on the outside of the aorta) to prevent injury to the conduction system. Pledgeted mattress sutures are then placed in a clockwise fashion typically starting in the noncoronary cusp. Sutures may be placed into the sewing ring of the prosthetic valve with each annular suture or after all annular sutures are placed. The sutures for each of the three cusps are held separately with three hemostats and retracted while the prosthesis is slid into the annulus. Sutures are then tied down in a balanced fashion alternating among the three cusps.

Aortic Closure and De-Airing

The aorta is closed with a double row of synthetic 4-0 polypropylene sutures. The first suture line is started on the right side at the posterior end of the aortotomy, and the double-needled suture is secured slightly beyond the incision to ensure there is no leak in this region. One end of the suture is run as a horizontal mattress anteriorly to the midpoint of the aortotomy, and then the second end of the suture is run anteriorly, slightly superficial to the horizontal mattress suture, in an over-and-over manner. On the left side, a similar technique is performed, the aorta is de-aired (described in the following), and the two sutures are tied to themselves and to each other at the aortotomy midpoint.

During AVR, air is entrained into the left atrium and ventricle, and aorta. This must be removed to prevent catastrophic air embolization. Immediately before tying the suture of the aortotomy, the heart is allowed to fill, the vent in the superior pulmonary vein is stopped, the lungs are inflated, and the cross-clamp is briefly partially opened. The influx of blood should expel most air from these cavities out of the partially open aortotomy. Closure of the aortotomy is then completed and the cross-clamp is fully removed. The cardioplegia cannula in the ascending aorta and the left ventricular vent are then placed on suction to remove any residual air as the heart begins electrical activity. A small needle (21-gauge) is used to aspirate the apex of the left ventricle and the dome of the left atrium. To prevent air entrainment, the left ventricular vent must be removed while the pericardium is filled with saline irrigation. De-airing maneuvers are verified with visualization using transesophageal echocardiography to verify that all air has been removed from the left side of the heart. Vigorous shaking and careful manual compression of the heart while suctioning through the aortic vent (ie, cardioplegia tack) is helpful to remove air trapped within trabeculations. Once de-airing is complete, the aortic vent is removed. The patient is then weaned from CPB and decannulated in the standard fashion. If patients are pacemaker-dependent when weaned from CPB in the operating room, it is recommended to insert atrial pacing wires to allow for synchronous atrioventricular pacing.

Concomitant Coronary Artery Bypass Grafting

Operative technique is modified when there is concomitant CAD to optimize myocardial protection. Distal anastomoses are performed before AVR so that antegrade cardioplegia may be administered through these grafts during the operation. The left internal mammary artery should be used for revascularization of the left anterior descending artery, because this may improve long-term survival in aortic valve patients.29 This anastomosis is performed after the aortotomy is closed to ensure that the coronary circulation is not exposed to systemic circulation during cardioplegic arrest and to prevent trauma to the anastomosis during manipulation of the heart.

Concomitant Ascending Aortic Replacement

In general, the ascending aorta is replaced electively when maximal diameter exceeds 5.5 to 6.0 cm; however, with Marfan's syndrome or equivalent connective tissue disorder the cutoff is 4.5 to 5.0 cm. In the setting of concomitant AVR, aortic replacement is advised if the ascending aortic diameter is greater than 5.0 cm. Patients with bicuspid aortic valves have an underlying aortopathy that leads to significant risk of late ascending aortic complications, and these patients should have replacement of their ascending aorta if its diameter exceeds 4.5 cm at the time of AVR30 (Fig. 33-9).

Aortic Root Enlargement Procedures

Detailed descriptions of aortic root enlargement procedures are presented in a later chapter. Briefly, either an anterior or posterior annular enlargement procedure may be performed in a patient with a small aortic root to allow for implantation of a larger valve. The posterior approach is the most commonly used aortic root enlargement procedure in adults and can increase the annular diameter by 2 to 4 mm. Nicks and colleagues in 1970 described a technique of root enlargement in which the aortotomy is extended downward through the noncoronary cusp, through the aortic annulus to the anterior mitral leaftlet.31 In 1979, Manouguian and Seybold-Epting described a procedure extending the aortotomy incision in a downward direction through the commissure between the left and noncoronary cusps into the interleaflet triangle and into the anterior leaflet of the mitral valve.32 The anterior approach is generally used in the pediatric population. Described by Konno and colleagues in 1975, this technique, which is also known as aortoventriculoplasty, is used when more than 4 mm of annular enlargement is required.33 Instead of a transverse incision, a longitudinal incision is made in the anterior aorta and extended to the right coronary sinus of Valsalva and then through the anterior wall of the right ventricle to open the right ventricular outflow tract. The ventricular septum is incised, allowing significant expansion of the aortic annulus and left ventricular outflow tract.

Reoperative Aortic Valve Surgery

Repeat sternotomy after AVR may be performed for valve-related complications, progressive ascending aortic disease, or CAD. Valve-related causes include structural valve deterioration, prosthetic endocarditis, prosthesis thrombosis, or paravalvular leak. Chest re-entry is the most hazardous portion of any repeat cardiac procedure. At Sunnybrook Health Sciences Center, our routine practice is to obtain an adequate lateral chest x-ray and computed tomography scan to determine the proximity of cardiac structures to the posterior sternum. Cardiopulmonary bypass is instituted through the femoral vessels for any concerns about chest re-entry. An oscillating saw is used to open the sternum and the dissection is kept as limited as possible. Extreme caution must be employed during dissection when there are patent bypass grafts.

Once cardioplegic arrest is established, the old prosthesis is excised with sharp dissection. Care must be taken to remove all sutures and pledgets from the annulus. Annular injuries caused while excising the prosthesis are repaired with pledgeted interrupted sutures. Removal of stentless prostheses may be particularly difficult in this regard. In the setting of endocarditis, aggressive debridement of infected tissue must be performed with appropriate annular reconstruction with pericardium when root abscesses are present.34 All foreign graft material, including Dacron aortic grafts, must be excised in the presence of active endocarditis.

In the presence of a Dacron prosthesis in the ascending aorta, chest re-entry may be extremely hazardous because exsanguination will occur if the graft is inadvertently opened during dissection. To limit the systemic consequences of exsanguination at normothermia, the patient should be placed on femoro-femoral CPB and cooled to 20°C before chest reentry. If the Dacron graft is accidentally opened, local control of the bleeding is established and CPB is stopped. Under circulatory arrest, atrial venous cannulation is instituted and the graft is controlled distal to the tear. Cardiopulmonary bypass may then be restarted. In all repeat aortic procedures, rigorous myocardial protection must be applied because these procedures often have very long ischemic times. Antegrade cold blood cardioplegia is usually employed in a continuous fashion throughout the case by selective cannulation of the coronary ostia. Retrograde cardioplegia may have benefit in the setting of patent old saphenous vein grafts.35

Special consideration must be given to the underlying pathologic changes to the ventricle during the immediate postoperative period. The severely hypertrophied, noncompliant LV resulting from AS is highly dependent on sufficient preload for adequate filling. Filling pressures should be carefully titrated between 15 and 18 mm Hg with intravenous volume infusion. In such cases, subvalvular left ventricular outflow obstruction with systolic anterior wall motion of the mitral valve may occur. Intravenous beta-adrenergic blockage may relieve this obstruction by decreasing inotropy. In extreme cases, reoperation and surgical myectomy may be required.

Maintenance of sinus rhythm is also essential because up to one-third of cardiac output is derived from atrial contraction in a noncompliant ventricle. Up to 10% of patients will experience low cardiac output syndrome in the immediate postoperative period. If pacing is required postoperatively, synchronous atrioventricular pacing is beneficial in preventing low cardiac output syndrome.

Complete heart block occurs in 3 to 5% of AVR patients. This complication may be caused by suture placement or injury from debridement near the conduction system. Transient complete heart block caused by perioperative edema usually resolves in 4 to 6 days. After this time, insertion of a permanent pacemaker is recommended if there is no resolution.

Profound peripheral vasodilation, often seen in patients with aortic insufficiency, is treated with vasoconstrictors, including alpha-adrenergic agonists or vasopressin. Adequate filling of the dilated left ventricle may also require volume infusion.

Stented biologic prostheses may be constructed of porcine aortic valves or bovine pericardium. Over the past 40 years, advances in tissue fixation methodology and chemical treatments have been developed to prevent extracellular matrix and calcium deposition. All heterograft valves are preserved with glutaraldehyde, which acts by cross-linking collagen fibers to reduce tissue antigenicity. Glutaraldehyde also ameliorates in vivo enzymatic degradation and causes the loss of cell viability, thereby preventing extracellular matrix turnover.36 Glutaraldehyde fixation of porcine valves can be performed at high pressure (60 to 80 mm Hg), low pressure (0.1 to 2 mm Hg), or zero pressure (0 mm Hg). Pericardial prostheses are fixed in low- or zero-pressure conditions. Porcine prostheses fixed at zero pressure retain the collagen architecture of the relaxed aortic valve cusp.37 When comparing various bioprostheses, it is important to be aware of lack of standardization in methodologies for labeling valve sizes by the different manufacturers. In general, label sizes refer to either the internal or external diameter of the stent, not the external diameter of the sewing cuff or the maximal opening diameter of the valve leaflets. Thus, the same aortic annulus will likely fit different-sized valves from different manufacturers, depending on the convention they use and the size of their sewing cuff. Figure 33-10 compares internal and external sizes for a variety of prostheses.

First-Generation Prostheses

First-generation bioprostheses were preserved with high-pressure fixation and were placed in the annular position. They include the Medtronic Hancock Standard and Modified Orifice (Medtronic, Minneapolis, MN), and Carpentier-Edwards Standard porcine prostheses (Edwards Life Sciences, Irvine, CA).

Second-Generation Prostheses

Second-generation prostheses are treated with low- or zero-pressure fixation. Several second-generation prostheses may also be placed in the supra-annular position, which allows placement of a slightly larger prosthesis. Porcine second-generation prostheses include the Medtronic Hancock II valve (Medtronic), the Medtronic Intact porcine valve (Medtronic), and the Carpentier-Edwards Supraannular valve (SAV) (Edwards Life Sciences). Second-generation pericardial prostheses include the Carpentier-Edwards Perimount (Edwards Life Sciences), and the Pericarbon (Sorin Biomedica, Saluggia, Italy) prostheses.

Third-Generation Prostheses

Newer-generation prostheses incorporate zero- or low-pressure fixation with antimineralization processes that are designed to reduce material fatigue and calcification. Stents have become progressively thinner, have a lower profile, are more flexible, and sewing rings have become scalloped for supra-annular placement. The Medtronic Mosaic porcine valve (Medtronic) is fixed in a "physiologic" environment with equal pressure (40 mm Hg) applied to the ventricular and aortic sides of the leaflets, resulting in net zero pressure on the leaflets themselves (Figs. 33-11 and 33-12). The St. Jude Medical Epic valve (St. Jude Medical, Inc.) is a porcine valve with a very low stent post and base profile to minimize protrusion into the aortic wall and facilitate coronary clearance (Fig. 33-13). The Carpentier-Edwards Perimount Magna valve (Edwards Life Sciences) is the evolution of the Perimount pericardial valve, with a narrower sewing cuff and scalloped design for supra-annular placement (Fig. 33-14). The Mitroflow Pericardial aortic prosthesis (Carbomedics) is a pericardial valve that is unique in that the pericardium is placed around the exterior of the stent, presumably allowing for a larger opening diameter (Fig. 33-15).

Operative Mortality

Operative mortality is defined as all-cause mortality within 30 days of operation or during the same hospital admission.38 Contemporary series describe a very low operative mortality for isolated AVR between 1 and 8%, depending on the patient population, the presence of coronary disease, and the era of study.39–43 In 1999, the Society of Thoracic Surgeons' (STS) database reviewed the results of 86,580 valve procedures to find an overall mortality of 4.3% for isolated AVR, and 8.0% for AVR with CABG.44 Aortic valve replacement with ascending aortic aneurysm repair was shown to have an increased operative mortality of 9.7%. The results of this study are summarized in Table 33-4. Then in 2009, they reviewed 67,292 cases of AVR and found 3.2% mortality rate (Table 33-5).45 It is important to note that information in this database is voluntarily submitted and includes both low- and high-volume centers.

The preoperative risk factors of operative mortality for AVR derived from the STS database are presented in Table 33-6.43,46–48 Most early deaths are attributable to renal failure, emergency status, and reoperations. In Table 33-7, mortality data for AVR-related procedures from an update of the STS database are presented describing the operative risk of isolated and combined valve procedures.49

Long-Term Survival

Longitudinal analysis demonstrates no difference in survival between patients receiving mechanical and bioprosthetic valves when they are implanted in similar age cohorts over 10 years of follow-up.50 However, at 15 years' follow-up, structural valve deterioration in bioprosthetic valves leads to a survival benefit for patients with mechanical valves. In a prospective trial, 15-year mortality was 79 and 66% for bioprosthetic and mechanical valves, respectively.51 Substantially more bleeding events were found in patients with mechanical valves. It is important to note that increased bioprosthetic structural deterioration was likely influenced by the use of a first-generation prosthesis more prone to structural failure than newer devices.

In most published series, the expected survival after AVR is approximately 80 to 85% at 5 years, 65 to 75% at 10 years, and 45 to 55% at 15 years.52–54 The outcomes of AVR are highly dependent on the functional status, comorbidities, and age of each individual patient.55 The effect of age at the time of surgery on late mortality for a variety of prostheses is depicted in Fig. 33-16. The comorbidities of age, concomitant CAD, LVD, and poor functional status on middle to late survival after bioprosthetic AVR are additive (Table 33-8).56 Other risk factors for late mortality include concomitant renal disease, female gender, concomitant cardiac or vascular procedure, and atrial fibrillation.57,58 Studies of mechanical valves often show superior long-term survival but patients are significantly younger at the time of operation. No prospective series of comparable patients has shown any survival benefit comparing pericardial with porcine valves in similar eras.

Valve-Related Mortality

Long-term survival data distinguishes between valve-related mortality, non–valve-related cardiac mortality, and mortality from other causes. A revised consensus document from the STS and the American Association of Thoracic Surgeons (AATS) outlines a standardized method to report valve-related complications in prosthetic and repaired heart valves.59 This panel defined valve-related mortality as all deaths caused by structural valve deterioration, nonstructural valve dysfunction, valve thrombosis, embolism, bleeding event, operated valvular endocarditis, or death related to reoperation of an operated valve. Sudden, unexplained, unexpected deaths of patients with an operated valve are included as valve-related mortality. Deaths caused by progressive heart failure in patients with satisfactorily functioning cardiac valves are not included. In the Hammermeister series, valve-related deaths accounted for 37% of all deaths in patients with mechanical valves and 41% of all deaths in patients with bioprostheses at 15 years.51 Nonvalvular cardiac deaths accounted for 17 and 21% of deaths at 15 years in patients with mechanical and bioprostheses, respectively.38 There are no well-designed prospective randomized series comparing the long-term outcomes of specific pericardial or porcine prostheses to each other.

Nonfatal Valve Events

The joint STS/AATS panel defined specific guidelines for reporting outcomes on structural and nonstructural valve deterioration, valve thrombosis, embolic events, bleeding events, and prosthetic endocarditis.60Structural valve deterioration refers to any change in function of an operated valve resulting from an intrinsic abnormality of the valve that causes stenosis or regurgitation (eg, leaflet tears, suture line disruption). Nonstructural dysfunction is any abnormality of an operated valve resulting in stenosis or regurgitation that is caused by factors not intrinsic to the valve itself (eg, pannus overgrowth, inappropriate sizing, or paravalvular leak). Valve thrombosis is any thrombus that interferes with valve function in the absence of infection. Embolism is any embolic event that occurs after the immediate postoperative period when perioperative anesthesia has been completely reversed. Cerebral embolic events are subclassified into: transient ischemic attacks, which are fully reversible neurologic events lasting less than 24 hours; reversible ischemic neurologic deficits, which are fully reversible neurologic events lasting between 24 hours and 3 weeks; and strokes, which are permanent neurologic deficits lasting longer than 3 weeks or causing death. A bleeding event is any episode of major internal or external bleeding that causes death, hospitalization, or permanent injury, or requires transfusion, regardless of the patient's anticoagulation status. This does not include embolic stroke followed by hemorrhagic transformation and intracranial bleed. Finally, operated valvular endocarditis is any infection involving an operated valve; or any structural or nonstructural valvular dysfunction, thrombosis, or embolic event associated with operated valvular endocarditis.

Structural Valve Deterioration

Stented Bioprostheses

Several large series describe long-term follow-up of first- and second-generation stented bioprostheses. These series are not directly comparable because they took place in different patient populations or different eras. Structural valve deterioration is the most common nonfatal valve-related complication in bioprosthetic aortic valve (Table 33-9). Long-term follow-up of currently available second-generation stented bioprostheses, including the Medtronic Hancock II porcine and Carpentier-Edwards pericardial valves, show freedom from structural valve deterioration less than 90% at 12-year follow-up.52,61,62 However, beyond 15-year follow-up, freedom from SVD falls rapidly.63 No good evidence exists to suggest that comparable, well-established second-generation pericardial or porcine prostheses differ in longevity; thus decisions regarding specific prosthesis should be made based on surgeon comfort and familiarity with the prosthesis and its sizing. Although newer third-generation prostheses with advanced tissue treatments eventually may be shown to have superior longevity, data about these prostheses are currently limited to 5- to 6-year outcomes, which appear comparable with those of the second-generation prostheses.64

Importantly, younger patients are predisposed to premature bioprosthetic SVD, particularly those less than the age of 40 years (Table 33-10).65,66 Structural valve deterioration may be less common in elderly patients because of the decreased hemodynamic stress placed on the valve. The freedom from SVD may be underestimated in the literature because most series report SVD by the actuarial method instead of the actual or cumulative incidence method.67 Actuarial statistical analysis overestimates SVD in older patients because it assumes that patients who have died of other causes will continue to be at risk for SVD.

Bioprostheses have a higher rate of reoperation then mechanical valves owing to structural valve dysfunction. In large series, freedom from reoperation is greater than 95% at 5 years, greater than 90% at 10 years, but less than 70% at 15 years.52,68–73 The long-term freedom from reoperation for several commonly available valves is presented in Table 33-10.

Optimal Antithrombotic Therapy

Bioprosthetic Valves

Bioprosthetic valves thus do not require long-term anticoagulation with warfarin unless the patient is at high risk for thromboembolism or has had a thromboembolic event with his or her prosthesis. Stented bioprostheses have a linearized risk for thromboembolism between 0.5 and 1% per year. This risk appears to be lower in patients with stentless heterograft, allograft, or autograft valves.74–76 An increased hazard function exists for thromboembolism before the exposed surfaces of stented bioprostheses endothelialize.77 The current AHA guidelines recommend anticoagulation with warfarin to an INR between 2.0 and 3.0 for the first 3 months for bioprosthetic valves as a class IIB recommendation.18 This is discontinued at the end of the third month unless the patient is at high risk for thromboembolism. Low-dose aspirin is continued as monotherapy in low-risk patients. Aspirin significantly decreases the risk of thromboembolism in low-risk patients with bioprostheses versus no antiplatelet therapy.78,79 High-risk patients require lifetime aspirin and warfarin treatment, which has a survival benefit over warfarin alone.78 If a patient has identified high-risk factors for thrombosis preoperatively, a mechanical prosthesis should be implanted unless the risk factor is amenable to correction because formal anticoagulation with warfarin will still be necessary. With aspirin, bioprosthetic valves have approximately the same risk of thromboembolism as fully anticoagulated mechanical valves, with fewer bleeding complications.

Prosthesis Thrombosis

Prosthesis thrombosis is a rare but potentially devastating outcome after AVR. The incidence of prosthesis thrombosis is less than 0.2% per year and occurs more often in mechanical prostheses.80 Thrombolytic therapy may be used in some patients but it is often ineffective. Thrombolysis is recommended in patients with left-sided thrombosis who are experiencing significant heart failure (NYHA class III or IV) and are considered too high risk for surgery.81,82 Cerebral or peripheral thromboembolism occurs in 12% of patients after thrombolytic therapy.81 Surgical treatment includes replacement of the valve or open thrombectomy, and mortality from either procedure is similar at approximately 10 to 15%.81 Recurrent thrombosis after de-clotting occurs in up to 40% of patients and we recommend valve replacement in virtually all patients who are managed operatively.

Prosthetic Valve Endocarditis

Prosthetic valve endocarditis (PVE) is separated into two time frames: early (<60 days postimplantation) and late (>60 days postimplantation). Early PVE is usually a sequela of perioperative bacterial seeding of the valve, either during implantation or postoperatively from wound or intravascular catheter infections. Staphylococcus aureus, S. epidermidis, gram-negative bacteria, and fungal infections are common in this period.52,72,83–86 Although most cases of late PVE are caused by septicemia from noncardiac sources, a small proportion of late cases in the first year are attributable to less virulent organisms introduced in the perioperative period, particularly S. epidermidis. Organisms responsible for late PVE include Streptococcus and Staphylococcus species and other organisms commonly found in native valve infectious endocarditis. All unexplained fevers should be meticulously investigated for PVE with serial blood cultures and TEE and/or TTE. Transesophageal echocardiography provides more detailed anatomical information such as the presence of vegetations, abscesses, and fistulas. Transthoracic views may provide better views of the anterior portion of the valve. The annual risk of PVE in the aortic position is 0.6 to 0.9% per patient year.85–87 The 5-year freedom from PVE reported in many major series is greater than 97%.52,85 Mechanical valves may have a slightly higher early hazard for PVE than stented bioprostheses.86 However, no difference exists in risk between patients with mechanical and stented bioprosthetic prostheses after the early phase. Stentless porcine heterografts and allografts are less likely to develop PVE because they have less prosthetic material that may serve as a nidus of infection.87 These valves may be particularly helpful in valve re-replacement for PVE.

Outcome for patients with PVE is very poor. Invasive paravalvular infection occurs in up to 40% of cases of PVE.84 Early PVE is associated with 30 to 80% mortality, whereas late PVE is associated with 20 to 40% mortality.52 Surgical indications for PVE include early PVE; concomitant heart failure and valvular dysfunction; paravalvular leak or partial dehiscence; presence of a new conduction defect, abscess, aneurysm, or fistula; persistent bacteremia despite a maximum of 5 days of appropriate antibiotic therapy and no other source of infection; vegetations (>10 mm); and multiple systemic emboli. Notably, all fungal and most virulent strains of Staphylococcus aureus, Serratia marcescens, and Pseudomonas aeruginosa require operation, as these organisms are highly invasive and antibiotic therapy is generally ineffective.

Paravalvular Leak and Hemolysis

Paravalvular leak is uncommon when pledgeted sutures are routinely used outside of the setting of infective endocarditis. Technical errors may result in inappropriately large gaps between sutures, leaving a small portion of the prosthesis unattached to the annulus. If paravalvular leak is sufficient to cause significant hemolysis, surgical correction may be performed with a few interrupted pledgeted sutures. Pannus overgrowth and prosthetic structural degeneration interfering with normal valve opening and closure may cause hemolysis severe enough to warrant reoperation. Milder cases of hemolysis may be managed conservatively by dietary supplementation with iron and folic acid, and routine measurement of hemoglobin, serum haptoglobin, and lactate dehydrogenase.

Left Ventricular Mass Regression

Pressure and volume overload caused by aortic valve disease leads to increased intracavitary left ventricular pressures and compensatory LVH. In severe AS, concentric ventricular hypertrophy occurs without increasing end-diastolic dimension until late in the disease process, thus maintaining the ventricular wall thickness:cavity radius ratio. On the other hand, severe AR causes volume overload with an increase in left ventricular end-diastolic volume and eccentric hypertrophy, but may not change the ratio of ventricular wall thickness to cavity radius. Both pathologies result in an increase in LV mass (LVM), which has a strong negative prognostic effect.88,89 The overall goal of AVR is to alleviate the pressure and volume overload on the left ventricle, allowing myocardial remodeling and regression of left ventricular mass.

The clinical impact of LVM regression is not as well understood, despite its widespread acceptance as a measure of outcome after aortic valve surgery. In hypertensive patients undergoing medical treatment, reduction of LVM was associated with fewer cardiac events than those whose left ventricular mass did not change or increased.135 In patients after AVR for isolated AS, LVM generally regresses significantly over the first 18 months and returns to normal limits.90,91 Ventricular mass regression may continue for up to 5 years after valve replacement.92 However, some patients do not experience adequate ventricular mass regression. The prognostic implications of LVM regression after aortic valve surgery have not been rigorously studied but it is reasonable to suppose that lack of LVM regression is associated with poor clinical outcome. Several authors have identified a situation, referred to as patient-prosthesis mismatch (PPM), in which poor hemodynamic performance of a prosthesis results in poor regression of LVH and poor patient outcome.

Prosthesis-Patient Mismatch

Definitions

The term PPM has been applied to several different clinical situations. It has been used to describe absolute small valve size (ie, <21 mm), small valve size in a patient with a large body surface area, excessive transvalvular gradient postimplantation, increased transvalvular gradient with exercise, indexed effective orifice area, and various combinations of these variables. Rahimtoola defined PPM as a condition that occurs when the valve area of a prosthetic valve is less than the area of that patient's normal valve.93 He further described a clinical condition in which the patient experienced either no relief or worsening of symptoms owing to the obstructive nature of the prosthesis, which creates a residual stenosis resulting in an elevated transvalvular gradient. To varying degrees, all valve prostheses are inherently stenotic.94 The presence of rigid sewing rings, and in the case of stented bioprostheses, struts to hold the valve commissures, cause obstruction to outflow and will therefore cause a residual gradient despite normal prosthesis function. The EOA of commonly-available bioprosthetic valves are shown in Table 33-11. PPM is further exacerbated by annular fibrosis, annular calcification, and LVH, as seen in AS, that cause contraction of the native annulus, leading to the implantation of a smaller prosthesis. Two distinct terms are commonly used to describe the size of prosthetic valves: effective orifice area and geometric orifice area.

Effective Orifice Area

The most commonly cited definition of PPM is a low indexed effective orifice area (IEOA). The IEOA is calculated by dividing the echocardiographically determined effective orifice area (EOA) by the body surface area. Effective orifice area is calculated by a reconfiguration of the continuity equation:

EOA = (CSALVOT × TVILVOT)/TVIAO

where EOA is effective orifice area (cm2), CSALVOT is cross-sectional area of the LV outflow tract (LVOT; cm2), TVILVOT is velocity time integral of forward blood flow (cm) as derived from pulse-wave Doppler in the LVOT, and TVIAO is velocity time integral of forward blood flow (cm) as derived from software integration of transvalvular continuous wave Doppler.

The EOA and mean systolic gradient of several commonly available bioprostheses are shown in Table 33-11.

Several authors suggest that PPM occurs at an IEOA of less than 0.85 cm2/m2.95,96 Dumesnil and Pibarot have redefined PPM as a condition "when the EOA of the prosthesis is too small in relation to the patient's body size, resulting in abnormally high postoperative gradients."96 This definition is based on the assumption that transvalvular gradients begin to rise substantially at IEOAs below this value, and that these elevated gradients result in increased left ventricular work preventing regression of LVH.97

The EOA is a functional estimate of the minimal cross-sectional area of the transvalvular flow jet downstream of a valve and dependent on the: (1) geometric area of the prosthesis, (2) shape and size of the LVOT and ascending aorta, (3) blood pressure, and (4) cardiac output (Fig. 33-17). Doppler-derived EOA correlates best with catheter-derived EOA (as determined by the Gorlin formula) when the ascending aortic diameter is 4 cm, but tends to underestimate EOA in patients with smaller aortic diameters.98 The EOA cannot be known for a specific valve in a specific patient until the valve has actually been implanted. Studies examining the effect of low IEOA on clinical outcomes have typically used published tables of EOA derived from historical controls instead of actually measuring true in vivo postoperative EOA. Moreover, these tables have been derived from relatively small numbers of valves in each size for each manufacturer with wide variability between studies. EOA correlates with postoperative valve gradients, which is not surprising considering that they are mathematically related. Echocardiographic mean and peak gradients are calculated according to the Bernoulli equation:

Geometric Orifice Area

Alternatively the geometric orifice area (GOA; also known as the internal geometric area) of the valve is the maximal cross-sectional area of the valve opening that does not vary significantly between same-sized valves from the same manufacturer. It is a static measure that is known preoperatively for any given prosthesis based on manufacturer specifications or by measurement with calipers. As seen in Fig. 33-17, the GOA is always larger than the EOA for any given prosthesis.

Clinical Significance

The significance of PPM is still controversial. Several studies have found that that lower IEOA is associated with diminished short- and long-term clinical results. Pibarot and Dumesnil97 studied PPM in 1266 patients undergoing AVR and found moderate (IEOA <0.85 cm2/m2) or severe (IEOA <0.65 cm2/m2) PPM to be present in 38% of patients. Multivariate analysis showed that moderate and severe PPM are associated with a twofold and 11-fold increase of perioperative mortality, respectively. Retrospective analysis of PPM in 2154 patients undergoing AVR at two large centers showed similar overall mortality with and without PPM, but higher valve-related mortality with PPM at 10 years. A single-center analysis of 1563 mechanical and tissue aortic prostheses found that IEOA less than 0.8 cm2/m2 was associated with increased prevalence of heart failure symptoms at a mean follow-up time of 4.3 years but not with mortality.99 Ruel and colleagues showed that overall survival and LVM regression was poorer in patients with PPM and LVD than those who had LVD without PPM.100 In a series of 1400 patients, PPM affected long-term survival in patients less than 60 years old but not in older patients.101

On the other hand, several studies have also shown that PPM does not influence clinical outcomes. Medalion and colleagues studied 892 AVRs and demonstrated no differences in 15-year survival between patients with or without PPM, despite 25% of patients with an indexed internal orifice area less than two standard below predicted.102 Hanayama and colleagues reported in 1129 patients with PPM post-AVR, no differences in LVM index or survival at midterm follow-up between patients with normal and abnormal gradients over a 10-year period.103 They defined PPM as: (1) IEOA less than the 90th percentile in the study population (0.6 cm2/m2); and (2) valve gradient in the highest 90th percentile of the study population (peak gradient and mean gradient were 38 and 21 mm Hg), which would be considered severe by most groups. Fig. 33-18 shows no differences in LVM index or survival at midterm follow-up between patients with or without PPM. The average labeled size of valve in the low-IEOA and high-gradient groups were 22.4 and 23 mm, respectively. The only multivariate predictor of elevated postoperative gradient was valve size.

Blackstone and colleagues have compiled the largest and most statistically sound examination of the role of PPM in short- and long-term outcomes.104 In a multi-institutional study of over 13,000 aortic prostheses, presence of an indexed GOA in the lowest 10th percentile was associated with a 1 to 2% increase in perioperative (30-day) mortality and had no effect on medium- or long-term survival (Fig. 33-19). An important finding in this study was that virtually no stented bioprostheses were in the group of patients in the lowest 10th percentile of indexed GOA, who predominantly received mechanical prostheses. In a subsequent study from the same group, there was no relationship between indexed GOA and Duke Activity Status Index in 1108 patients undergoing AVR at a mean follow-up of 8.3 months. Predictors of postoperative functional status are presented in Table 33-12.

Small Aortic Root

Many surgeons express concern regarding postoperative outcomes in patients with small aortic roots in whom only very small (19 mm or smaller) valves can be implanted. However, several studies have shown no difference in LVM regression, NYHA functional class, CHF or survival with small diameter AVRs. 105–107 Khan and colleagues studied 19 mm to 23 mm Carpentier-Edwards pericardial valves and found that significant LVM regression occurred with each valve size, including 19 mm valves.108 Studies by DePaulis and colleagues show no difference in LVM regression between patients receiving 19 and 21 mm mechanical valves versus those with 23 or 25 mm valves.106 Kratz and associates also reported that small valve size was not predictive of CHF or late death.107

Data Synthesis

Although the literature regarding the clinical significance of PPM remains divided, proponents of the concept advocate use of mechanical prostheses. The presumption is less obstructive than stented bioprostheses, aortic root enlargement, and stentless valves when used in the subcoronary position and as full root replacement. However, whether mechanical prostheses alleviate PPM versus the current generation of tissue valves is an open question and indeed, they may even result in greater PPM.109 Aortic root enlargement procedures require significant experience operating on the aortic root and may carry excessive mortality and wide variability in surgical outcomes even among highly experienced centers.

Alternatively, a stentless prosthesis in the subcoronary position may be a lower profile alternative but requires additional technical skill and cross-clamp time. Rao and colleagues compared hemodynamic data between stented Carpentier-Edwards pericardial valves and Toronto stentless porcine valves of equivalent diameter and found no hemodynamic differences in peak or mean gradient.110

Aortic root replacement in patients without ascending aortic pathology, solely to diminish the potential long-term effects of PPM, is a potentially high-risk strategy and is discouraged.

When faced with potential PPM in the operating room, the decision to perform a more complex, higher-risk procedure must be balanced carefully with the potential benefits of implanting a larger prosthesis. Some reports have shown that transvalvular gradients in patients with lower IEOA often rise substantially with exercise.97,111 Although the majority of patients undergoing AVR are elderly and unlikely to experience functional limitations from this situation, in younger, highly active patients either root enlargement or stentless prostheses may provide better functional outcome with lower transvalvular gradients. In the rare circumstance of anticipated extreme mismatch (ie, IEOA <0.6 cm2/m2), root enlargement is an acceptable approach in the hands of an experienced surgeon. Except in these circumstances, given the paucity of long-term data to support more complex procedures and well-documented increased risk, routine AVR with modern standard prostheses is acceptable and preferable.

An ideal aortic prosthesis would be simple to implant, widely available, possess long-term durability, would have no intrinsic thrombogenicity, would not have a predilection for endocarditis, and would have no residual transvalvular pressure gradient. Such a valve does not currently exist. Currently available options include mechanical valves, stented biologic heterograft valves, stentless biologic heterograft valves, allograft valves, and pulmonary autograft valves. Among these options pulmonary autograft valves and allograft valves are the most physiologic prostheses. They are less prone to thrombosis or endocarditis and have excellent hemodynamic characteristics.112,113 The longevity of such valves is dependent on patient factors, the preparation of the valve, and the technical skill of the operating surgeon. Despite their potential benefits, these prostheses are not readily available and can be very technically demanding to implant compared with standard mechanical or stented bioprostheses. They are most beneficial in children and younger adults. Allograft valves may also improve the results of AVR in active endocarditis.114 A further discussion of these valves is presented in subsequent chapters. As their use has remained confined to a few centers that perform such operations regularly, the remainder of this discussion focuses on issues regarding selection of mechanical or bioprosthetic valves.

Mechanical versus Biologic Valves

When selecting between mechanical and biologic heart valves, the surgeon and patient together must balance the risks and benefits of each choice. Mechanical valves are much less likely to undergo structural deterioration necessitating reoperation. Mechanical valves are more thrombogenic than bioprosthetic valves and require formal anticoagulation with oral warfarin, which carries a significantly increased risk of bleeding complications. Patients with mechanical valves and adequate anticoagulation do not have a greater risk of thromboembolic events than do those with bioprosthetic valves.115 There is no difference in actuarial freedom from bacterial endocarditis between mechanical and bioprosthetic valves. Regarding overall survival, two randomized comparisons of mechanical and bioprosthetic aortic valves performed in the 1970s demonstrated equivalent survival between valve types at 12 years of follow-up.116,117 Long-term survival beyond 15 years was superior in the mechanical valve groups as the risks of reoperation and structural failures of bioprostheses became more common.118 However, the rate of structural valve deterioration was higher with first-generation prostheses than with newer-generation models.

Special Patient Groups

Patients with an absolute requirement for long-term anticoagulation such as atrial fibrillation, previous thromboembolic events, hypercoagulable state, severe LVD, another mechanical heart valve in place, or intracardiac thrombus, should receive a mechanical valve regardless of age.

Patients in whom anticoagulation with warfarin is contraindicated, such as women of child-bearing age wishing to become pregnant, patients with other bleeding disorders, or those who refuse anticoagulation should receive a bioprosthesis. Alternatively, implantation of mechanical prostheses in such women combined with anticoagulation using subcutaneous low-molecular weight heparin injections during the pregnancy is a possibility.

Patients with end-stage renal failure were previously believed to have significantly elevated risk for early bioprosthetic structural valve deterioration. However, increased anticoagulation-related complications are also more likely in this group, and the current AHA guidelines do not recommend routine use of mechanical prostheses in these patients.

Further considerations that may affect future patient decisions include testing for genetic variants of the CYP2C9 and VKORC1 genes, which influence warfarin response, use of devices to facilitate home INR monitoring, and use of novel nonwarfarin oral anticoagulants. A detailed and comprehensive discussion of these risks and benefits of prosthesis selection should occur with all patients and their families before entering the operating room.

Age Considerations

Currently available bioprostheses, such as the Medtronic Hancock II porcine and Carpentier-Edwards pericardial valve, have greater than 90% freedom from structural valve dysfunction and greater than 90% freedom from reoperation at 12-year follow-up.52,60,61 The rate of structural deterioration is lower in patients over 65 to 70 years. Hence, patients greater than 65 years at the time of surgery should receive a biologic valve. Patients less than age 60 should have a mechanical prosthesis to minimize the risk of structural failure requiring repeat AVR as an octogenarian. Patients between 60 and 65 represent the group in whom there is still considerable debate regarding prosthesis selection. Those patients who have comorbidities such as severe CAD may be less likely to outlive their prosthesis and should receive a biologic valve. Prompted by improvements in durability of bioprostheses and lower reoperation risks, a recent retrospective study of patients 50 to 70 with AVR demonstrated worse 10-year unadjusted survival (50 versus 65%) and freedom from reoperation (91 versus 98%) with tissue compared with mechanical AVR.11 On the other hand, Ruel et al.119 found no difference in overall survival for patients less than 60 years old with tissue AVR versus those with mechanical AVR over a 20-year follow-up. Although age alone usually does not preclude one from surgery, advanced age and associated comorbidities may make someone a poor candidate for a classic valve replacement. In such patients, transcatheter aortic valve replacement (TAVR), discussed later, may be an alternative to medical management or balloon aortic valvuloplasty.

Stented versus Stentless Biologic Valves

Stentless porcine valves have gained popularity in cardiac surgery owing to the pioneering work done by Tirone David at the Toronto General Hospital in 1988.120 Stentless valves lack obstructive stents and strut posts, thus have residual gradients that are similar to those of freehand allografts. However, their use involves a more complex operation as they are more difficult to implant and require a longer cross-clamp time. Cohen and colleagues found no differences in aortic root size, IEOA, or LVM regression (Fig. 33-20) or functional outcome at 1 year between patients randomized to receive Carpentier-Edwards pericardial valves and Toronto stentless porcine valves (Fig. 33-21).121 These findings challenge the notion that stentless porcine valves provide increased IEOA or hemodynamic or clinically significant benefit. In a randomized trial comparing St. Jude Toronto stentless porcine valves with Carpentier-Edwards pericardial valves, Chambers and colleagues found no difference in IEOA, LVM regression, or mortality between groups. Arenaza and colleagues also performed a multicenter randomized trial comparing the Medtronic Freestyle valve with the Medtonic Mosaic valve and found increased IEOA in the stentless group, but no differences in LVM regression or clinical outcomes at 1 year.

However, Walther and colleagues performed a small randomized trial comparing the ability of stented porcine and stentless porcine valves which cause regression of LVH.122 They showed that patients in the stentless valve group received larger valves for a given annular size and had a slightly higher degree of LVM regression. Borger and colleagues showed modestly lower mean gradients in stentless prostheses versus stented prostheses (9 versus 15 mm Hg) and LVMI (100 versus 107 g/m2) but no difference in survival.123

Hence, conflicting evidence exists regarding benefit of stentless valves in LVM regression or clinical outcomes over stented bioprostheses. There is also little evidence that incremental improvements in LVM provide additional clinical benefit. Thus, the routine use of stentless bioprostheses cannot be recommended for most patients with small aortic roots based on currently available data. At this time, stentless porcine valves are most useful in a relatively younger patient with a small aortic root who is active and likely to be limited by the elevated residual gradient a small stented bioprosthesis may create.

Percutaneous Aortic Balloon Valvotomy

As an alternative to surgical management for AS, aortic balloon valvotomy has been performed percutaneously via a femoral artery puncture to treat AS.124 Inflation of the balloon within the valve orifice can stretch the annular tissue and fracture calcified areas or open fused commissures. There is no role for valvotomy in the patient with significant AR, as this will become significantly worse after the procedure.125,126 Balloon valvotomy is rarely successful if significant calcification is present and carries a prohibitive risk of stroke from calcific emboli.127 The long-term outcomes of this procedure in adult patients are dismal, with restenosis usually occurring within 1 year.126,127 Patients with severe symptomatic AS who are too hemodynamically unstable to tolerate operation or have comorbid illnesses, such as advanced malignancy, which contraindicate operation, may benefit from palliative balloon valvotomy.128,129 More recently, techniques have developed to implant prosthetic valves using transfemoral and transapical catheter-based approaches.

Transcatheter Aortic Valve Replacement

Although surgical aortic valve replacement is the definitive treatment for aortic valve stenosis, up to one-third of patients are not candidates due to increasing age, heart failure, or other comorbidities.130 The first clinical description of human transcatheter valve implantation took place in 2002.131 Transcatheter heart valve technologies are now a clinical reality and have been shown to offer symptom relief and survival benefit to high-risk, elderly, and nonoperative patients.132 Early experience from Webb and coworkers demonstrate overall mortality of 14.3% in the initial half to 8.3% in the second half of the study, and 74% 1-year survival with both percutaneous and transapical procedures.133

Currently, the two most extensively studied valve systems include the Edwards Sapien Valve and the CoreValve, although various other valves exist in different stages of development and evaluation (Fig. 33-22). Each valve system includes: (1) the valve prosthesis, (2) the stent or frame, (3) the loading system, and (4) the delivery system. The Edwards Lifesciences SAPIEN THV valve (Edwards Lifesciences Inc., Irvine, CA) is composed of a bovine pericardial valve on a balloon-expandable stent, whereas the CoreValve (Medtronic) is a porcine pericardial valve on a self-expanding nitonol frame.

The approach is accomplished via a retrograde transarterial or transapical manner. The percutaneous retrograde femoral approach involves femoral artery access, retrograde cannulation of the aortic valve, balloon dilatation, and device delivery. Alternatively, transapical transcatheter valve delivery involves creating a small thoracotomy, direct cannulation of the left ventricular apex, and passing of a wire and stent-mounted valve under fluoroscopic and echocardiographic guidance.134 Both techniques require rapid ventricular pacing to ensure there is no cardiac output during device deployment.

At the time of writing this chapter, clinical trials are in progress in which issues regarding prosthesis durability, periprocedural stroke, coronary artery injury, and hemodynamic performance will be addressed. At present, these techniques are enabling very high-risk patients to experience some relief of their symptoms without undergoing a complex operation. Given the excellent current results of open operative AVR, the use of these procedures in lower-risk patients is not currently warranted.