Bioprosthetic Aortic Valve Replacement: Stented Pericardial and Porcine Valves

	INTRODUCTION

Top Next

	NATURAL HISTORY AND INDICATIONS FOR OPERATION

Top Previous Next

 
Aortic Stenosis

Natural history

Aortic stenosis (AS) may be caused by degenerative calcification, congenital malformations, or rheumatic fever. It may also be found in association with systemic diseases such as Paget’s disease of bone and end-stage renal disease. Congenital malformations include unicommisural, and more commonly, bicuspid valves. Detailed descriptions of these pathologies are presented elsewhere in this volume. Degenerative calcification is common in the elderly population and although limited calcific deposits often do not have any clinical significance, severe degenerative calcification is the most common pathology among patients undergoing AVR. Regardless of the initial pathology, there is a progressive reduction of orifice cross-sectional area caused by calcification of the cusps that resembles atherosclerotic processes that include inflammation and lipid accumulation.^1,2 Commissural fusion is typically seen in rheumatic AS but not in degenerative calcification of tricuspid or bicuspid valves. The normal human aortic valve has an area between 3.0 and 4.0 cm². Mild, moderate, and severe AS are defined as aortic valve areas (AVAs) greater than 1.5 cm², 1.0 to 1.5 cm², and less than 1.0 cm², respectively.³ This corresponds to a mean gradient of 25 mm Hg and peak jet velocity of less than 3.0 m/s in mild cases, a mean gradient of 25 to 40 mm Hg and peak jet velocity of 3.0 to 4.0 m/s in moderate cases, or a mean gradient of greater than 40 mm Hg and peak jet velocity of greater than 4.0 m/s in severe cases. In the presence of normal cardiac output, transvalvular gradient is usually greater than 50 mm Hg when the AVA is less than 1.0 cm².⁴ There is a rapid increase in transvalvular gradient when the AVA is less than 0.8 to 1.0 cm².

Exposure to elevated intracavitary pressures causes increased wall stress leading to parallel replication of sarcomeres and concentric hypertrophy.^5,6 These mechanisms compensate for the obstruction to flow created by the reduced orifice area of the aortic valve in order to maintain normal cardiac output. With progressive hypertrophy, the compliance of the ventricle decreases and end-diastolic pressure rises.^7,8 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.⁶

Symptomatic patients

Hemodynamically significant AS is initially counteracted by left ventricular hypertrophy (LVH). Progression of outflow obstruction and ventricular hypertrophy lead to the cardinal symptoms of AS: angina, syncope, and congestive heart failure (CHF). The average AVA is 0.6 to 0.8 cm² at the onset of symptoms.⁶ Classic natural history studies have shown that the average life expectancy in patients with hemodynamically significant AS is 4 years if anginal symptoms are present, 3 years if they have experienced syncope, and 2 years with the onset of CHF.⁹ Symptomatic patients should therefore undergo AVR in a timely fashion.¹⁰ Excessive waiting periods for AVR in symptomatic patients are associated with increased mortality and the rate of sudden death is >10% per year in these symptomatic patients. Once a patient is symptomatic, average survival is less than 3 years.^11–14 The typical modes of death in untreated severe AS are sudden death from ventricular arrhythmia or CHF.

Asymptomatic patients

Managing asymptomatic patients with hemodynamically significant AS can be a challenging problem, as there is often a prolonged latent period before symptoms emerge. During the latent period, there is a progression of concentric LVH as the ventricle adapts to elevated chamber pressures. Studies by Otto and colleagues have shown that up to 7% of asymptomatic patients experience death or aortic valve surgery 1 year after diagnosis.¹⁵ After 5 years, the incidence of death or aortic valve surgery increases to 38%. The average decrease in AVA is 0.12 cm² per year, while the average increase in transvalvular pressure is often 10 to 15 mm Hg per year.¹⁶ Sudden death is quite uncommon in the asymptomatic patient and occurs at a rate of 0.4% per year. The vast majority of patients who experience sudden death will become symptomatic in the months immediately prior to the fatal event.^17,18

There is considerable variation in the rate of disease progression and many patients do not experience any change in gradient for several years. To provide improved guidance on which asymptomatic patients need closer follow-up or early operation, a study by Rosenhek and colleagues identified that asymptomatic patients with an increase in peak velocity jet greater than 0.45 m/s per year on serial echocardiography were substantially more likely to need operation than patients with lesser changes in jet velocity.¹⁹

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. The 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 (i.e., dobutamine infusion) may distinguish whether cardiomyopathy or true valvular stenosis is the most responsible lesion. Although some patients with a preponderance of cardiomyopathy do not experience significant benefit from valve replacement,²⁰ a recent study by Pereira and colleagues suggests that in a balanced, propensity matched comparison, patients with poor LVF and severe AS experience a significant survival benefit from valve replacement²¹ (Fig. 34-1). Hwang and colleagues, using a multivariate analysis to determine factors that predict poor LVF after AVR for AS, identified that poor preoperative LVF was the most significant predictor, indicating suboptimal outcome in the low-gradient group of patients.²²

View larger version (14K): [in this window] [in a new window]   Figure 34-1 Impact of aortic valve replacement on survival in patients with low-gradient aortic stenosis. Survival by Kaplan-Meier analysis among all propensity-matched patients in the aortic valve replacement (AVR) and control (No AVR) groups (p <0.0001). The number of patients at risk during follow-up is shown on the x axis. (Reproduced with permission from Pereira et al.21)

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,^23–25 a recent prospective randomized clinical trial of aggressive antilipid therapy with atorvastatin did not demonstrate decreased rates of disease progression as determined by changes in aortic jet velocity or valve calcification score²⁶ (Fig. 34-2).

View larger version (16K): [in this window] [in a new window]   Figure 34-2 Progression of aortic valve jet velocity and calcification in patients treated with intensive atorvastatin therapy or matched placebo. (Reproduced with permission from Cowell et al.26)

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.²⁷ These were updated in 2006.²⁸ Their recommendations for AVR in the setting of AS are summarized in Table 34-1. A class I recommendation indicates there is good evidence and general agreement that the treatment is beneficial, useful, and effective. Class IIA recommendation indicates there may be disagreement but the weight of evidence supports the usefulness/efficacy of the treatment in that setting, Class IIB recommendation indicates that the usefulness/efficacy of the treatment is less well established, and class III recommendation indicates that the treatment is either not useful/effective or potentially harmful.

Aortic Regurgitation

Acute aortic regurgitation

Acute aortic regurgitation (AR) may be caused by acute dilatation of the aortic annulus preventing adequate cusp coaptation or by disruption of the valve cusps themselves. This typically occurs in the setting of acute aortic dissection, infective endocarditis, trauma, active connective tissue disease, aortic cusp prolapse associated with ventricular septal defects (VSDs), aortitis (syphilitic or giant cell), Marfan syndrome, Ehlers-Danlos syndrome, or iatrogenically after aortic balloon valvotomy.⁶ The heart cannot readily tolerate acute AR, as the normal left ventricle is unable to compensate for the sudden increase in end-diastolic volume caused by the large regurgitant volume load, thereby overwhelming the Frank-Starling mechanism.²⁹ A dramatic reduction in forward stroke volume then occurs. If there is poor left ventricular compliance from hypertrophy prior to the onset of acute AR, hemodynamic decompensation is significantly more dramatic.

To compensate for the acute decline in forward stroke volume, tachycardia ensues. Volume overload causes the left ventricular diastolic pressure to acutely rise above left atrial pressure resulting in early closure of the mitral valve.³⁰ Although early mitral valve closure protects the pulmonary venous circulation from high end-diastolic pressures, rapid progression of pulmonary edema and cardiogenic shock are often unavoidable.

Death is the common endpoint of all etiologies of acute AR. Progressive cardiogenic shock and malignant ventricular arrhythmias are common causes of death. 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 etiologies include congenital abnormalities, calcific cusp degeneration, rheumatic fever, endocarditis, degenerative aortic dilatation as seen in the elderly, Marfan syndrome, Ehlers-Danlos syndrome, myxomatous proliferation, osteogenesis imperfecta, ankylosing spondylitis, Behçet syndrome, Reiter syndrome, psoriatic arthritis, severe systemic hypertension, and idiopathic aortic root dilatation.⁶ The anorectic drugs fenfluramine and dexfenfluramine have been implicated in left- and right-sided valvular disease, including AR.^31,32 Bicuspid aortic valve is the most common congenital abnormality but unicommissural, quadricuspid, and fenestrated valves may also occur.³³

Chronic AR causes a chronic volume overload of the left ventricle. The volume load leads to progressive chamber enlargement without increasing end-diastolic pressure during the asymptomatic phase of the disease.³⁴ Progressive chamber enlargement is accompanied by eccentric hypertrophy, 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 all maintained.³⁵ However, this degenerates into a repetitive cycle of enlarging chamber radius with continually increasing wall stress. This wall stress is compensated by ventricular hypertrophy. Interstitial fibrosis limits the ability of the ventricle to further dilate and this cycle becomes overwhelmed, leading to elevated end-diastolic pressure, left ventricular systolic dysfunction, and CHF.³⁶ Vasodilator therapy may delay progression of ventricular dysfunction by decreasing afterload and decreasing 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 prior to operation.³⁷ This therapy is not recommended in patients with severe AR and LVD, as it does not improve survival but may be used in these patients if they are not considered operable.³⁸

Symptomatic chronic aortic regurgitation

The time course from diagnosis of AR to the development of symptoms is highly variable. Since symptoms, such as angina and dyspnea, develop only after significant ventricular decompensation has occurred, surgery is advocated prior to the symptomatic phase of the disease. Symptomatic patients experience >10% mortality per year without surgical management.^38,39

Asymptomatic chronic aortic regurgitation

Natural history studies of asymptomatic AR show that symptoms, LVD, or both develop in <6% of patients per year.⁴⁰ Progression to LVD without symptoms occurs in <4% of patients per year. Sudden death occurs in <0.2% per year.⁴¹ Age, left ventricular end-systolic dimension, rate of change in end-systolic dimension, and rest EF are all independent predictors of progression to symptoms, LVD, or death in asymptomatic patients.⁴² Asymptomatic patients with left ventricular systolic dysfunction experience onset of symptoms at a rate exceeding 25% per year.⁴³

Indications for operation

In cases of isolated AR, timing of surgery on the asymptomatic patient is predicated on the identification of subtle changes in myocardial function before they become irreversible and negatively affect the patient’s long-term prognosis. Unfortunately, such changes are often recognized by current imaging modalities only after irreversible damage and fibrosis has occurred. Patients with more severe LVD have decreased perioperative and late survival due to irreversible changes to the ventricle including hypertrophy and interstitial fibrosis.^43–45 The decision to operate on such patients is dependent on individual variables since the outcomes are poor with surgery or medical therapy. Advances in imaging such as cardiac magnetic resonance imaging (MRI) may allow identification of candidates for AVR prior to the onset of LVD.⁴⁶

A summary of the revised ACC/AHA Task Force guidelines for AVR for chronic AR is presented in Table 34-2.²⁸ Aortic valve replacement is indicated in all symptomatic patients with severe AR, regardless of LVF, and in all asymptomatic patients with severe AR with left ventricular ejection fraction (LVEF) <50%. Aortic valve replacement is also recommended in asymptomatic patients with normal LVF but significant left ventricular dilatation (end-diastolic diameter >75 mm or end-systolic diameter >55 mm). Aortic valve replacement is recommended in patients with moderate to severe AR who require concomitant coronary artery bypass surgery (CABS) or ascending aortic surgery.

	CORONARY ANGIOGRAPHY AND AORTIC VALVE REPLACEMENT

Top Previous Next

 
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 may increase as the surgical population continues to age. Risk assessment for ischemic heart disease is complicated in patients with aortic valve disease since 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. Since traditional coronary risk stratification is unreliable in aortic valve patients, it is our practice to routinely perform diagnostic coronary angiography on all patients over the age of 35. ACC/AHA Task Force guidelines for preoperative angiography are presented in Table 34-3.²⁸

Myocardial protection and cardiopulmonary bypass

Isolated AVR is performed using a single two-stage venous cannula inserted into the right atrium for venous return and a standard 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 through a purse-string suture in the right superior pulmonary vein and advanced into the left ventricle to ensure a bloodless field and to prevent ventricular distention if there is aortic insufficiency. Once cardiopulmonary bypass (CPB) is initiated, the 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. Pulmonary artery injuries may be difficult to repair, 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.^47–49 This will achieve 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.⁵⁰ Many surgeons do not routinely use retrograde cardioplegia for aortic valve cases, but this strategy is helpful in patients with significant AR or severe concomitant coronary disease.⁵¹ In cases in whom the retrograde cannula cannot be placed into the coronary sinus, conversion to bicaval cannulation and opening the right atrium to directly place the cannula into the coronary sinus is possible. If retrograde perfusion is employed, this is also used in a continuous manner. Right ventricular myocardial protection may be inadequate with only retrograde cardioplegia and can lead to significant right ventricular dysfunction after CPB is discontinued. This may be avoided by ensuring that the retrograde cannula is not placed beyond the origin of the right coronary vein ostium in the coronary sinus.^52–55 The patient’s systemic body temperature is allowed to drift downward, but active cooling for noncirculatory arrest cases is unnecessary.

Aortotomy, valve excision and débridement

After 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 millimeters 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. 34-3). The oblique incision is often used in patients with small aortic roots, in whom root enlargement procedures may be required (see below) and may also be used to tailor a larger ascending aorta.

View larger version (30K): [in this window] [in a new window]   Figure 34-3 Exposure and aortotomy incision. A two-stage venous cannula is in place in the right atrial appendage. The aortotomy (dashed line) may be made in the transverse or oblique direction.

View larger version (30K): [in this window] [in a new window]   Figure 34-4 The exposed aortic valve.

View larger version (28K): [in this window] [in a new window]   Figure 34-5 The aortic valve after leaflet excision.

View larger version (27K): [in this window] [in a new window]   Figure 34-6 Anatomic relationships of the aortic valve.

Valve implantation

After the native valve has been excised, the annulus is sized with a valve-sizer designed exactly 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. 34-7 and 34-8). 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 whom 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 (i.e., 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.

View larger version (30K): [in this window] [in a new window]   Figure 34-7 Placement of sutures with pledgets below the annulus.

View larger version (37K): [in this window] [in a new window]   Figure 34-8 Placement of sutures with pledgets above the annulus.

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 below), 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 prior to 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 subsequent 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 transesophageal echocardiographic visualization 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 (i.e., 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.

Concomitant coronary artery bypass grafting

Operative technique is modified when there is concomitant CAD to optimize myocardial protection. Distal anastomoses are performed prior to AVR so that antegrade cardioplegia may be administered through these grafts during the operation. The left internal thoracic artery should be used for revascularization of the left anterior descending artery, as this may improve long-term survival in aortic valve patients.⁵⁶ 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 in patients without Marfan syndrome, and 4.5 to 5.0 cm in patients with Marfan syndrome. In the setting of concomitant aortic valve replacement, aortic replacement is advised if the ascending aortic diameter is >5.0 cm. Recent data suggest that 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 AVR ⁵⁷ (Fig. 34-9).

View larger version (24K): [in this window] [in a new window]   Figure 34-9 Freedom from ascending aortic complications for patients with a bicuspid aortic valve with an ascending aortic diameter of <4 cm, 4.0 to 4.5 cm, and 4.5 to 4.9 cm at the time of aortic valve replacement. (Reproduced with permission from Borger MA, et al: Should the ascending aorta be replaced more frequently in patients with bicuspid aortic valve disease? J Thorac Cardiovasc Surg 2004; 128:677.)

Detailed descriptions of aortic root enlargement procedures are presented in a later chapter. 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.⁵⁸ 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.^59,60 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.⁶¹ 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 valverelated 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. It is our routine practice to obtain an adequate lateral chest x-ray and computed tomography (CT) scan to determine the proximity of cardiac structures to the posterior sternum. Cardiopulmonary bypass is instituted through the femoral vessels when there is concern 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. When only the aortic valve needs reoperation, advanced port-access technologies such as percutaneous coronary sinus cannulation and pulmonary artery venting can allow the operation to proceed with only a limited upper sternotomy.

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 débridement of infected tissue must be performed with appropriate annular reconstruction with pericardium when root abscesses are present.^62,63 All foreign graft material, including Dacron aortic grafts, must be excised in the presence of active endocarditis.⁶⁴

In the presence of a Dacron prosthesis in the ascending aorta, chest re-entry may be extremely hazardous since exsanguination will occur if the graft is accidentally 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 prior to chest reentry.⁶⁵ 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 since 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.⁶⁶

Aortic balloon valvotomy

Aortic balloon valvotomy may be performed percutaneously via a femoral artery puncture in the interventional angiography suite to treat AS.⁶⁷ 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.^68–70 Balloon valvotomy is rarely successful if significant calcification is present and carries a prohibitive risk of stroke from calcific emboli.^71,70 The long-term outcomes of this procedure in adult patients are dismal, with restenosis usually occurring within 1 year.^70,72,73 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.^74–77 More recently, techniques have developed in implant prosthetic valves using transcatheter approaches. These procedures are discussed later in this chapter.

	POSTOPERATIVE MANAGEMENT

Top Previous Next

 
Special consideration must be given to the underlying pathologic changes to the ventricle during the immediate postoperative period. The severely hypertrophied, noncompliant left ventricle found in 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. Maintenance of sinus rhythm is also essential since 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.⁷⁸ If pacing is required postoperatively, synchronous atrioventricular pacing is beneficial in preventing low cardiac output syndrome. 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.

In cases of severe hypertrophy, subvalvular left ventricular outflow obstruction with systolic anterior wall motion of the mitral valve may occur. Intravenous beta-adrenergic blockade may relieve this obstruction by decreasing inotropy. In extreme cases reoperation and surgical myectomy may be required.⁷⁹

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.

Complete heart block occurs in 3 to 5% of AVR patients. This complication may be due to suture placement or injury from débridement 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. Echocardiographic evaluation of valve function should be performed in the operating room with trans-esophageal echocardiography (TEE) and then at discharge, at 3 months, and yearly with transthoracic echocardiography (TTE).

	STENTED BIOPROSTHETIC AORTIC VALVE REPLACEMENT DEVICES

Top Previous Next

 
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 to prevent calcification have yielded improvements in the longevity of bioprostheses. All heterograft valves are preserved with glutaraldehyde, which cross-links collagen fibers and reduces the antigenicity of the tissue. Glutaraldehyde also reduces the rate of in vivo enzymatic degradation and causes the loss of cell viability, thereby preventing normal turnover and remodeling of extracellular matrix tissues.^80,81 Calcification occurs when nonviable glutaraldehyde-fixed cells cannot maintain low intracellular calcium.⁸¹ Calcium phosphate crystals form at the phospholipid-rich membranes and their remnants, and the collagen matrix also calcifies.⁸²

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.⁸³ Higher fixation pressures cause tissue flattening and compression with loss of transverse cuspal ridges and collagen crimp, which may lead to earlier calcification.^84,85

Multiple chemical treatments have been proposed to decrease the calcification process that invariably leads to material failure and valvular dysfunction. These include sodium dodecyl sulfate, polysorbate-80, Triton X-100, N-lauryl sarcosine, amino-oleic acid, aminopropane-hydroxydiphosphonate, toluidine blue, controlled-release diphosphonates, ferric chloride, aluminum chloride, and phosphocitrate.^86–100

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 34-10 compares internal and external sizes for a variety of prostheses.

View larger version (48K): [in this window] [in a new window]   Figure 34-10 Comparison of the internal and external diameter of prosthetic aortic valves to the manufacturers’ labeled size of each valve. No two manufacturers’ valves have the same internal diameter for a given labeled size. (CE = Carpentier-Edwards Pericardial valve; CM = Carbomedics Standard valve; CS = Carbomedics Supra-annular valve; HM = Medtronic Hall valve; HT = Hancock II Bioprosthesis; MO = Hancock Modified Orifice Bioprosthesis; SJ = St. Jude Standard valve; SJ-HP = St. Jude Hemodynamic Plus valve; SPV= stentless porcine valve.) (Reproduced with permission from Christakis GT, et al: Inaccurate and misleading valve sizing: a proposed standard for valve size nomenclature. Ann Thorac Surg 1998; 66:1198.)

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, Minneapolis, MN), the Medtronic Intact porcine valve (Medtronic, Minneapolis, MN), and the Carpentier-Edwards Supraannular valve (SAV) (Edwards Life Sciences, Irvine, CA). Second-generation pericardial prostheses include the Carpentier-Edwards Perimount (Edwards Life Sciences, Irvine, CA), and the Pericarbon (Sorin Biomedica, Saluggia, Italy) prostheses.

Third-Generation Prostheses

Newer-generation prostheses incorporate zero- or low-pressure fixation with anti-mineralization processes that are designed to reduce material fatigue and calcification. Stents have become progressively thinner, have a lower profile, and are more flexible, and sewing rings have become scalloped for supra-annular placement. The Medtronic Mosaic porcine valve (Medtronic, Minneapolis, MN) 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. 34-11 and 34-12). By pressurizing the root without placing pressure on the leaflets, this process stabilizes the root to maintain its anatomic shape. Alpha-amino oleic acid, which permanently covalently binds to free aldehyde groups, is used as the anticalcification treatment.

View larger version (70K): [in this window] [in a new window]   Figure 34-11 "Physiologic" fixation process. Simultaneous application of pressure to the inflow and outflow portions of a porcine bioprosthesis place zero net pressure on leaflets within a pressurized root. (Figure courtesy of Medtronic Inc., Minneapolis, MN.)

View larger version (122K): [in this window] [in a new window]   Figure 34-12 The Medtronic Mosaic porcine aortic prosthesis. (Figure courtesy Medtronic Inc., Minneapolis, MN.)

View larger version (108K): [in this window] [in a new window]   Figure 34-13 The St. Jude Epic porcine aortic prosthesis. (Figure courtesy of St. Jude Medical Inc., Minneapolis, MN.)

View larger version (83K): [in this window] [in a new window]   Figure 34-14 The Carpentier-Edwards Magna pericardial prosthesis. (Figure courtesy of Edwards Life Sciences Inc., Irvine, CA.)

View larger version (109K): [in this window] [in a new window]   Figure 34-15 The Mitroflow Pericardial aortic prosthesis (Figure courtesy of Carbomedics Inc., Austin, TX.)

Transcatheter valve replacement technologies are evolving and have recently become a clinical reality, applied to high-risk patients who are not operative candidates. Valves comprised of equine pericardium have been mounted onto stents which can be delivered by three different techniques (Fig. 34-16). The antegrade approach involves femoral venous access, transseptal puncture, dilation of the atrial septum, guiding a flotation balloon through the mitral valve, antegrade cannulation of the aortic valve, snaring of a guidewire from opposite the femoral artery with exteriorization of an arteriovenous wire loop, balloon dilatation, and device delivery.¹⁰¹ Initial results of this technique were associated with many mechanical and arrhythmic complications and a high periprocedural mortality, leading to virtual abandonment of this technique.^102–105 The retrograde femoral approach involves percutaneous femoral artery access, retrograde cannulation of the aortic valve, balloon dilatation, and device delivery. This technique has been applied with acceptable early results in elderly and high-risk patients, and represents an emerging option for nonoperative candidates.¹⁰⁶

View larger version (81K): [in this window] [in a new window]   Figure 34-16 The Cribier-Edwards aortic percutaneous heart valve (upper panel) and fluoroscopic view of this prosthesis deployed in the aortic root (lower panel). (Figure courtesy of Edwards Life Sciences, Irvine, CA.)

The ultimate role of these catheter-based techniques of AVR has yet to be determined. Questions regarding prosthesis durability, periprocedural stroke, coronary artery injury, and hemodynamic performance remain unanswered, and will continue to evolve with successive iterations of the technology. 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.

	OUTCOMES OF AORTIC VALVE REPLACEMENT

Top Previous Next

 
Operative Mortality

Operative mortality is defined as all-cause mortality within 30 days of operation or during the same hospital admission.¹⁰⁸ Contemporary series describe a very low operative mortality for isolated AVR. The mortality from AVR varies between 1 and 8%, depending on the patient population, the presence of coronary disease, and the era of study.^109–120 A publication from the Society of Thoracic Surgeons’ (STS) database reviewing the results of 86,580 valve procedures found an overall mortality of 4.3% for isolated AVR, and 8.0% for AVR with CABS.¹²¹ Aortic valve replacement with ascending aortic aneurysm repair had an operative mortality of 9.7%.¹²¹ The results of this study are summarized in Table 34-4. It is important to note that information in this database is voluntarily submitted and includes both low-volume and high-volume centers.

Longitudinal analysis shows that there is 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.¹³⁹ 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 by Hammermeister and colleagues, 11-year mortality was 62 and 57% for bioprosthetic and mechanical valves, respectively.¹⁴⁰ At 15 years, mortality in the bioprosthetic group rose to 79% while mortality in the mechanical valve group rose to 66%.¹⁴⁰ There were substantially more bleeding events in patients with mechanical valves. It is important to note that the effect of bioprosthetic structural deterioration on mortality in this series was influenced by the actual prosthesis used in the study (Medtronic Hancock porcine). This is a first-generation prosthesis and is more prone to structural failure than newer devices.^141,142

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.^143–146 The outcomes of AVR are highly dependent on the functional status, comorbidities, and age of each individual patient.^146,147 The effect of age at the time of surgery on late mortality for a variety of prostheses is depicted in Fig. 34-17. Cohen and colleagues studied the impact of age, concomitant CAD, LVD, and poor functional status on middle to late survival after bioprosthetic AVR. Their results, presented in Table 34-7, show an additive risk for each of these comorbid factors.¹⁴⁸ Other studies have shown that concomitant renal disease, female gender, concomitant cardiac or vascular procedure, and atrial fibrillation are also risk factors for late mortality.^149–154 Studies of mechanical valves often show superior long-term survival, as these patients are significantly younger at the time of operation. No prospective series of comparable patients has shown any survival benefit comparing pericardial to porcine valves in similar eras.

View larger version (18K): [in this window] [in a new window]   Figure 34-17 Projected long-term survival after aortic valve replacement stratified by age at time of surgery. For each age group the age-, race-, and ethnicity-matched population life table curve is shown as a dot-dashed line. Note that younger patients show more marked departure from normal life expectancy. (Reproduced with permission from Blackstone et al.34)

Long-term survival data distinguish 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) was published in 1996 outlining a standardized method to report valve-related complications in prosthetic and repaired heart valves.¹⁵⁵ 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.¹⁵⁶ Nonvalvular cardiac deaths accounted for 17 and 21% of deaths at 15 years in patients with mechanical and bioprostheses, respectively.¹⁵⁶ 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.¹⁵⁷ These definitions are summarized below.

1. Structural valve deterioration: Any change in function of an operated valve resulting from an intrinsic abnormality of the valve that causes stenosis or regurgitation, such as wear, leaflet tears, or suture line disruption of components.
   
2. Nonstructural dysfunction: Any abnormality of an operated valve resulting in stenosis or regurgitation that is caused by factors not intrinsic to the valve itself, such as pannus overgrowth, inappropriate sizing, or paravalvular leak.
   
3. Valve thrombosis: Any thrombus attached near an operated valve that interferes with valve function in the absence of infection.
   
4. Embolism: Any embolic event that occurs after the immediate postoperative period when perioperative anesthesia has been completely reversed. Emboli may be peripheral (noncerebral) or cerebral. Myocardial infarction is excluded unless the event occurs after the perioperative period and coronary artery embolus is unequivocally documented.
   Cerebral embolic events are subclassified into:
   1. Transient ischemic attacks: Fully reversible neurologic events lasting less than 24 hours.
      
   2. Reversible ischemic neurologic deficit: Fully reversible neurologic events lasting more than 24 hours and less than 3 weeks.
      
   3. Stroke: Permanent neurologic deficit lasting longer than 3 weeks or causing death.
      
   
   
5. Bleeding event: 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.
   
6. Operated valvular endocarditis: Any infection involving an operated valve; any structural or non-structural valvular dysfunction, thrombosis, or embolic event associated with operated valvular endocarditis is included only in this category.
   

Time-related analysis of nonfatal complications is often expressed by the Kaplan-Meier actuarial method, or by the recently described cumulative incidence method proposed by Grunkemeier and Wu.¹⁵⁸ Cumulative incidence (or actuarial) reporting provides additional information, as it removes the impact of mortality on nonfatal outcomes. This is most relevant in higher-risk groups such as elderly patients, a group in which many patients will die from other causes prior to the occurrence of nonfatal events. Kaplan-Meier actuarial methods tend to overestimate nonfatal events by continuing to assume that patients who have died are still at risk for these nonfatal events.

Structural Valve Deterioration

Mechanical prostheses

Currently available mechanical prostheses are extremely resistant to material fatigue or structural valve deterioration (SVD). Very long-term follow-up of the Starr-Edwards caged-ball prosthesis, the Medtronic Hall tilting disc prosthesis, and the St. Jude bileaflet mechanical prosthesis show that these valves are exceedingly resistant to structural failure. Some discontinued mechanical prostheses, such as the Bjork-Shiley convexo-concave valve, had high rates of structural failure due to fracture of the outlet strut.¹⁵⁹

Stented bioprostheses

There are several large series describing long-term follow-up of first- and second-generation stented bioprostheses. These series are not comparable to each other since they took place in different patient populations or in different eras. Structural valve deterioration is the most common nonfatal valve-related complication in bioprosthetic aortic valves. Table 34-8 summarizes the long-term SVD outcomes of commonly used first- and second-generation stented bioprostheses. Long-term follow-up of currently available second-generation stented bioprostheses, including the Medtronic Hancock II porcine and Carpentier-Edwards pericardial valves, shows that these prostheses have a freedom from structural valve deterioration <90% at 12-year follow-up.^160–162 However, beyond 15-year follow-up, freedom from SVD falls rapidly.¹⁶³ There is no good evidence to suggest that comparable, well-established second-generation pericardial or porcine prostheses differ in longevity, and decisions to choose a specific prosthesis should be made based on surgeon comfort and familiarity with the prosthesis and its sizing and holding mechanisms. While newer third-generation prostheses with advanced tissue treatments may eventually be shown to have superior longevity, data about these prostheses are currently limited to 5- to 6-year outcomes, which appear comparable to those of the second-generation prostheses.¹⁶⁴

Longevity of stentless bioprostheses appears to be significantly dependent on the native valve pathology and implantation technique. Retrospective studies have shown that such valves are prone to early structural deterioration when used in the subcoronary position in patients with AR or bicuspid pathology as the primary indication for operation. In these patients, progressive root dilatation due to aortic pathology often leads to poor leaflet coaptation, leaflet tears, and recurrent AR.^169–171 Such modes of failure have not been observed when stentless valves are used as root replacements, which appear to have similar mid-term durability to the stented prostheses.^172,173

Freedom from Reoperation

Freedom from reoperation for currently available mechanical valves is >95% at 10 years and >90% at 15 years.^174–180 Bioprostheses have a significantly higher rate of reoperation due to structural valve dysfunction. In large series, freedom from reoperation is >95% at 5 years, >90% at 10 years, but <70% at 15 years.^181–200 The long-term freedom from reoperation for several commonly available valves is presented in Table 34-9.

Optimal Antithrombotic Therapy

Mechanical valves

All mechanical valves require formal anticoagulation with warfarin for the lifetime of the patient since these valves are inherently thrombogenic. The overall linearized risk of thromboembolism in patients on warfarin therapy is 1 to 2% per year in most published series.^201–219 A meta-analysis of over 13,000 patients showed that the incidence of major embolism in mechanical valves without antithrombotic therapy was 4 per 100 patient-years.²²⁰ With antiplatelet therapy this risk was 2.2 per 100 patient-years, and with warfarin therapy it was reduced to 1 per 100 patient-years. The incidence of thromboembolism was slightly higher in caged-ball prostheses than other mechanical prostheses.²²¹

The embolic risk is highest in the first few months, before the exposed cloth sewing ring and valve components have fully endothelialized.²²² Antithrombotic therapy is initiated on the second postoperative day with oral warfarin. In the presence of complicating factors such as gastrointestinal or mediastinal bleeding or perioperative neurologic injury, anticoagulation may be held initially. If the target level of anticoagulation is not achieved by the fourth postoperative day, intravenous heparin is instituted to achieve a partial thromboplastin time between 60 and 90 seconds. If the patient is at high risk for thromboembolism, heparin is started concurrently with warfarin on the second postoperative day. High-risk patients include those with atrial fibrillation, intracardiac thrombus, left atrial enlargement, severe LVD, and history of systemic emboli or hypercoagulable state.²²³

The target level of anticoagulation for each individual patient is dependent on their thrombotic risk profile and the type of valve employed. It is our practice to establish an International Normalized Ratio (INR) of 3.0 (acceptable range 2.5 to 3.5) for high-risk patients with mechanical valves and additionally institute low-dose aspirin therapy (80 to 100 mg once daily). For lower-risk patients, the target INR is 2.5 (acceptable range 2.0 to 3.0) and low-dose aspirin is started on an individual basis.

Several articles suggest that caged-ball valves in the aortic position require a higher level of anticoagulation than tilting monoleaflet or bileaflet prostheses.^224,225 These valves should be anticoagulated to an INR of 3.5 to 4.5. Although older tilting monoleaftlet prostheses were associated with an increased rate of thromboembolism, currently available prostheses such as the Medtronic Hall valve do not appear to have an increased thromboembolism risk versus bileaflet prostheses and should have the same target INR.^226–228

Aspirin, an antiplatelet agent, is routinely used at a low dose to minimize the risk of thromboembolic events. Randomized trials have shown that low-dose aspirin significantly reduces fatal cardiovascular and embolic events for all patients with mechanical valves, particularly in patients with concomitant coronary or vascular disease.^229–232 There are increased bleeding events when higher doses of aspirin are used in patients formally anticoagulated with warfarin.^233,234 The linearized risk of a bleeding event is 0.1 to 3.5% per year, depending on how these events are defined, the range of anticoagulation used, and how often coagulation parameters are measured.^235–238

The On-X valve (Medical Carbon Research Institute, Austin, TX) is purported to have lower thrombogenicity due to use of a purer form of carbon and enhanced contours over the pivot guides. This device is currently being investigated in a Food and Drug Administration–approved clinical trial for reduced anticoagulation with only clopidogrel and aspirin versus standard warfarin.

Bioprosthetic valves

Bioprosthetic valves are less thrombogenic than mechanical prostheses and do not require long-term anticoagulation with warfarin unless the patient is at high risk for thromboembolism or has had a thromboembolic event with their prosthesis.²⁵² Stented bioprostheses have a linearized risk for thromboembolism between 0.5 and 1% per year.^240–246 This risk appears to be lower in patients with stentless heterograft, allograft, or autograft valves.^247–252 Anticoagulant management of bioprostheses during the first 3 months after implantation remains variable between institutions. There is an increased hazard function for thromboembolism before the exposed surfaces of stented bioprostheses endothelialize.²⁵³ The current ACC/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.²⁸ 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, and in conjunction with warfarin in high-risk patients for the lifetime of the patient. Combined aspirin and warfarin have a survival benefit over warfarin alone in highrisk patients with bioprosthetic heart valves.²³⁶ Aspirin significantly decreases the risk of thromboembolism in low-risk patients with bioprostheses versus no antiplatelet therapy.^254–257 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 since 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.²³⁶

Prosthesis Thrombosis

Prosthesis thrombosis is a rare but potentially devastating outcome after AVR. The incidence of prosthesis thrombosis is <0.2% per year and it occurs more often in mechanical prostheses.^258–260 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 higher) and are considered too high risk for surgery.^261,262 Cerebral or peripheral thromboembolism occurs in 12% of patients after thrombolytic therapy.²⁶¹ Surgical treatment includes replacement of the valve or open thrombectomy, and mortality from either procedure is similar at approximately 10 to 15%.²⁶¹ 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.²⁶³ Staphylococcus aureus, S. epidermidis, gram-negative bacteria, and fungal infections are common in this period.^264–285 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.²⁶⁹ 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 TE and TTE. Transesophageal echocardiography provides more detailed anatomic information such as the presence of vegetations, abscesses, and fistulas, but often does not provide adequate views of the anterior portion of the valve.²⁷⁰ Transthoracic views may be helpful in these cases. Mechanical valves are particularly difficult to visualize by echocardiography due to shadowing created by valve components.

The annual risk of PVE in the aortic position is 0.6 to 0.9% per patient year.^{215,242,271–283} The 5-year freedom from PVE reported in many major series is >97%.^284–287 Mechanical valves may have a slightly higher early hazard for PVE than stented bioprostheses.²⁸⁷ However, there is no difference 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 since they have less prosthetic material that may serve as a nidus of infection.^288–295 These valves may be particularly helpful in valve rereplacement for PVE.

Outcome for patients with PVE is very poor. Invasive paravalvular infection occurs in up to 40% of cases of PVE.²⁹⁶ Early PVE is associated with 30 to 80% mortality, while late PVE is associated with 20 to 40% mortality.^297,298

Surgery is indicated for PVE in the following circumstances:

1. All cases of early (<60 days) of PVE.
   
2. Concomitant heart failure and valvular dysfunction.
   
3. Paravalvular leak or partial dehiscence, even in a stable patient, requires operative management, particularly if more than 40% of the valve’s annular circumference is involved.
   
4. The presence of a new conduction defect, abscess, aneurysm, or fistula mandates operative management. All fungal, and most virulent strains of Staphylococcus aureus, Serratia marcescens, and Pseudomonas aeruginosa also require operation, as these organisms are highly invasive and antibiotic therapy is generally ineffective.
   
5. Any case of persistent bacteremia despite a maximum of 5 days of appropriate antibiotic therapy and no other source of infection.
   
6. Vegetations >10 mm are not penetrated well by antibiotics and usually need operative management.
   
7. Multiple systemic emboli.
   

Paravalvular Leak and Hemolysis

Paravalvular leak is uncommon outside of the setting of infective endocarditis when pledgeted sutures are routinely used. 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. Hemolysis is uncommon with the currently available mechanical or stented bioprosthetic valves if they are functioning well. 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.

	HEMODYNAMIC PERFORMANCE AND VENTRICULAR REMODELING

Top Previous Next

 
Left Ventricular Mass Regression

Pressure and volume overloading 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. 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 left ventricular mass (LVM). Studies from the hypertension literature indicate that increased LVM has a strong negative prognostic effect. Several reports, including the Framingham Heart Study, indicate that increased LVM was a predictor of all cardiac events, including sudden cardiac death.^299–301 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. Smaller studies have shown that in hypertensive patients undergoing medical treatment, patients with a reduction of left ventricular mass had fewer cardiac events than those whose left ventricular mass did not change or increased.³⁰² The prognostic implications of LVM regression after aortic valve surgery have not been rigorously studied, but logic would suggest that lack of LVM regression is associated with poor clinical outcome. There are no studies showing that there is incremental clinical benefit with greater degrees of ventricular mass regression.

Generally, LVM regresses significantly over the first 18 months and returns to within normal limits in many patients after AVR for isolated AS.^303–308 Ventricular mass regression may continue for up to 5 years after valve replacement.³⁰⁹ However, some patients do not experience adequate ventricular mass regression and may experience poorer prognoses. Several authors have identified a situation, referred to as patient-prosthesis mismatch, in which the poor hemodynamic performance of a prosthesis results in poor regression of LVH and poor patient outcome.

Prosthesis-Patient Mismatch

Definitions

The term prosthesis-patient mismatch has been applied to several different clinical situations. It has been used to describe absolute small valve size (i.e., <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 prosthesis-patient mismatch as a condition that occurs when the valve area of a prosthetic valve is less than the area of that patient’s normal valve.³¹⁰ He described a clinical condition in which the patient experienced either no relief or worsening of symptoms due to the obstructive nature of the prosthesis, which creates a residual stenosis resulting in an elevated transvalvular gradient. To varying degrees, all mechanical and stented tissue prostheses and stentless tissue prostheses used in the subcoronary or inclusion root position are inherently stenotic.³¹¹ 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. This was more true of first-generation prostheses used during the 1970s, when this condition was first described. The problem is 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 prosthesis-patient mismatch 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:

where EOA is the effective orifice area in square centimeters, CSA_LVOT is the cross-sectional area of the left ventricular outflow tract(LVOT) in square centimeters as determined by two-dimensional measurement of the LVOT diameter, TVI_LVOT is the velocity time integral of forward blood flow in centimeters as derived from pulse-wave Doppler in the LVOT, and TVI_AO is the velocity time integral of forward blood flow in centimeters as derived from software integration of transvalvular continuous wave Doppler.³¹² The EOA and mean systolic gradient of several commonly available bioprostheses are shown in Table 34-10.

EOA is an in vivo, functional estimate of the minimal cross-sectional area of the transvalvular flow jet downstream of a valve (Fig. 34-18). The EOA is dependent on several factors including the geometric area of the prosthesis, the shape and size of the LVOT and ascending aorta, blood pressure, and cardiac output. Mechanistic studies have shown that Doppler-derived EOA correlates best with catheter-derived EOA(as determined by the Gorlin formula) when the ascending aortic diameter is 4 cm, and tends to underestimate EOA in patients with smaller aortic diameters.³¹⁶ 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 has been shown to correlate with postoperative valve gradients.³¹⁷ Gradients and EOA are, in fact, mathematically related. Echocardiographic mean and peak gradients are calculated according to the Bernoulli equation:

View larger version (17K): [in this window] [in a new window]   Figure 34-18 Diagrammatic representation of effective orifice area and geometric orifice area in relation to the left ventricular outflow tract and aortic root.

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. 34-18, the GOA is always larger than the EOA for any given prosthesis.

Clinical significance

The significance of prosthesis-patient mismatch is controversial since there is conflicting evidence that lower IEOA causes diminished short or long-term clinical results.

Pibarot and Dumesnil studied 1266 patients undergoing AVR at a single institution. They defined moderate prosthesis-patient mismatch as IEOA <0.85 cm²/m² and severe prosthesis-patient mismatch as IEOA <0.65 cm²/m². They found that moderate or severe prosthesis-patient mismatch was present in 38% of patients. Multivariate analysis showed that moderate prosthesis-patient mismatch was associated with a doubling of perioperative mortality, and severe prosthesis-patient mismatch was associated with an 11-fold increase of perioperative mortality. Of note, the poor outcomes in the severe mismatch group were based on only 27 patients with 7 perioperative mortalities. These patients also had substantially longer intraoperative cardiopulmonary bypass times and concurrent coronary surgery. Rao and colleagues retrospectively studied prosthesis-patient mismatch in patients undergoing AVR at two large centers.³¹⁸ A total of 2154 patients were reviewed including 227 patients with prosthesis-patient mismatch, and 1927 patients without prosthesis-patient mismatch. Overall mortality was similar in both groups, but valve-related mortality was higher in the prosthesis-patient mismatch group at 10 years. Many valve-related mortalities included mechanisms of death that were unrelated to prosthesis-patient mismatch (embolic stroke, valve failure, endocarditis, bleeding, and reoperation) and supportive echocardiographic data were not provided. Effective orifice area was obtained from in vitro data supplied by the manufacturers according to the valve size implanted.

Ruel and colleagues performed a single-center analysis of 1563 mechanical and tissue aortic prostheses and found that IEOA <0.8 cm²/m² was associated with increased prevalence of heart failure symptoms at a mean follow-up time of 4.3 years.³¹⁴ This relationship was not seen when prosthesis-patient mismatch was defined as IEOA <0.85 cm²/m². Prosthesis-patient mismatch was not associated with early or late mortality in this series. Ruel and colleagues in a subsequent analysis showed that the effect of prosthesis-patient mismatch was confined to patients with preoperative LVD.³²⁰ Using multivariate methods, they determined that overall survival and LVM regression was poorer in patients with prosthesis-patient mismatch and LVD than those who had LVD without prosthesis-patient mismatch. A confounding factor in this study was that patients with prosthesis-patient mismatch and LVD were 12 years older and had more significant comorbidities including concomitant coronary disease, which are independently associated with poorer survival and LVM regression.

Mohty-Echahidi and colleagues examined prosthesis-patient mismatch in patients receiving 19 mm and 21 mm mechanical valves and found that severe mismatch, defined as IEOA <0.60 cm²/m², was associated with increased late mortality (hazard ratio 2.18) and increased CHF symptoms (hazard ratio 3.1) versus patients without mismatch. In a series of 1400 patients, Moon and colleagues reported that prosthesis-patient mismatch affected long-term survival in patients less than 60 years old but not in older patients.³²¹

Several studies have also shown that prosthesis-patient mismatch does not influence clinical outcomes. Medalion and colleagues studied 892 AVRs and demonstrated that although 25% of patients received valves with an indexed internal orifice area less than two standard deviations below predicted normal aortic valve size, there were no differences in 15-year survival between patients with or without patient-prosthesis mismatch.³²²

Hanayama and colleagues recently reported a single institutional experience with patient-prosthesis mismatch in 1129 patients who were followed over a 10-year period.³²³ They defined patient-prosthesis mismatch using two definitions: (1) IEOA less than the 90th percentile in the study population; and (2) valve gradient in the highest 90th percentile of the study population. The cut-off value for IEOA to define patient-prosthesis mismatch was 0.6 cm²/m², which would be considered very severe by most groups. The cut-off values for peak gradient and mean gradient were 38 mm Hg and 21 mm Hg. In this study, the average labeled size of valve in the low-IEOA and high-gradient groups were 22.4 mm and 23 mm, respectively. There were no differences in left ventricular mass index (LVMI) or survival at midterm follow-up between patients with normal and abnormal gradients. The only multivariate predictor of elevated postoperative gradient was valve size. Thus, although valve size is a predictor of abnormal postoperative gradient, no clinical significance could be correlated with this finding. Figure 34-19 shows that there were no differences in LVMI or survival at midterm follow-up between patients with or without patient-prosthesis mismatch defined as IEOA <0.6 cm²/m².

View larger version (26K): [in this window] [in a new window]   Figure 34-19 Actuarial survival and freedom from NYHA class III or IV in patients with and without prosthesis-patient mismatch. There were no significant differences in the two groups. (Reproduced with permission from Hanayama et al.324)

View larger version (19K): [in this window] [in a new window]   Figure 34-20 (A) Effect of indexed orifice area on non–risk-adjusted survival. (B) Time-related survival stratified by indexed orifice area. These are Kaplan-Meier estimates of 1-, 5-, and 10-year survivals in finely grouped strata of indexed orifice area. (Reproduced with permission from Blackstone et al.344)

Many surgeons have expressed concern about postoperative outcomes in patients with small aortic roots in whom only very small (19 mm or smaller) valves can be implanted. Adams and colleagues reported significantly elevated perioperative, but not late, mortality among men, but not women, receiving 19 mm prostheses.³²⁵ Conversely, Sawant and colleagues demonstrated that in patients with small aortic roots, body surface area and valve size were not determinants of long-term survival.³²⁶ Studies by DePaulis and colleagues have shown that there was no difference in LVM regression between patients receiving 19 mm and 21 mm mechanical valves versus those with 23 mm or 25 mm valves.³²⁷ Foster and colleagues showed that in patients with 17 mm and 19 mm prostheses who had resting transvalvular gradients >30 mm Hg, 93% were in NYHA class I at late follow-up.³²⁸ Kratz and associates also reported that small valve size was not predictive of CHF or late death.³²⁹ 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.³³⁰

Data synthesis

While the literature regarding the clinical significance of prosthesis-patient mismatch remains divided, proponents of the concept advocate use of mechanical prostheses, which are presumably less obstructive than stented bioprostheses, aortic root enlargement, and stentless valves used in the subcoronary position and as full root replacement to achieve improved hemodynamics. Mechanical prostheses are not likely to alleviate prosthesis-patient mismatch versus the current generation of tissue valves. Mechanical valves may even be more often associated with prosthesis-patient mismatch.³³¹ Aortic root enlargement procedures require significant experience operating on the aortic root and may carry excessive mortality. Sommers and David reported a doubling of perioperative mortality among patients receiving annular enlargement procedures with AVR.³³² However, Castro and colleagues reported no increase in perioperative mortality among patients receiving annular enlargement procedures, implying that there is wide variability in surgical outcomes even among highly experienced centers.³³³

Use of a stentless prosthesis in the subcoronary position 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.³³⁴ Additionally, concerns about durability discussed previously for patients with bicuspid pathology make subcoronary stentless bioprostheses a less viable option.

Aortic root replacement in patients without ascending aortic pathology, solely to diminish the potential long-term effects of prosthesis-patient mismatch is a potentially high-risk strategy. Recent data from the Society of Thoracic Surgeons’ database on over 200,000 patients suggests that mortality for aortic root replacement for this indication was 9.5 versus 5.7% for isolated AVR.^138,335 This report strongly discouraged the use of aortic root replacement for this indication.

When faced with potential prosthesis-patient mismatch 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.^336,337 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 (i.e., 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.

	PROSTHESIS SELECTION

Top Previous Next

 
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.^338–346 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 to 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.³⁴⁷ 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 will focus 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 must balance the risks and benefits of each choice. Mechanical valves are much less likely to undergo structural deterioration than bioprosthetic valves, and reoperation for structural valve deterioration is more common in patients with bioprosthetic valves. Mechanical valves are more thrombogenic than bioprosthetic valves and require formal anticoagulation with oral warfarin. Anticoagulated patients have a significantly increased risk of bleeding complications. Patients with mechanical valves and adequate anticoagulation do not have significantly greater risk of thromboembolic events than do those with bioprosthetic valves.³⁴⁸ There is no difference in actuarial freedom from bacterial endocarditis between mechanical and bioprosthetic valves. In recent years a significant trend toward increasing use of bioprostheses has occurred in North America.

As stated earlier, the two randomized comparisons of mechanical and bioprosthetic aortic valves performed in the 1970s showed equivalent survival between valve types at 12 years of follow-up.^349,350 Long-term survival beyond 15 years was superior in the mechanical valve groups due to the risks of reoperation and structural failures of bioprostheses became more common.³⁵¹

In these studies, the bleeding risk was significantly higher than the current standard in the anticoagulated group due to monitoring of the prothrombin time instead of the INR. Also, the rate of structural valve deterioration was higher than currently expected, since first-generation prostheses were used in these studies. It would not be practical or ethical to conduct a similar randomized study today with current prostheses. Analyses based on mathematical modeling of historic data currently suggest that at approximately 60 years of age, patients derive improved life expectancy and event-free life expectancy regardless of the need for concomitant coronary surgery.^352–354

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. There is growing interest in using mechanical prostheses in women of child-bearing age and providing anticoagulation with subcutaneous low-molecular weight heparin injections.

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 ACC/AHA guidelines do not recommend routine use of mechanical prostheses in these patients.

Age considerations

Currently available bioprostheses, such as the Medtronic Hancock II porcine and Carpentier-Edwards pericardial valve, have >90% freedom from structural valve dysfunction and >90% freedom from reoperation at 12-year follow-up.^355–357 The rate of structural deterioration is lower in patients over 65 to 70 years.

Hence, patients over 65 years at the time of surgery should receive a biologic valve. Patients under the age of 60 should have a mechanical prosthesis to minimize the risk of structural failure requiring repeat AVR in 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. A detailed discussion of these risks and benefits of prosthesis selection should occur with all patients and their families prior to entering the operating room.

Stented versus Stentless Biologic Valves

Stentless porcine valves have gained popularity in cardiac surgery due to pioneering work by Dr. Tirone David at the Toronto General Hospital in 1988.³⁵⁸ Since they lack obstructive stents and strut posts, stentless valves provide residual gradients that are similar to those of freehand allografts. Stentless valves, however, are more difficult to implant and require a longer cross-clamp time such that the risks of a more complex operation must be matched to a specific benefit the patient may receive with a stentless valve. Cohen and colleagues randomized patients to receive Carpentier-Edwards pericardial valves and Toronto stentless porcine valves and compared clinical outcomes.³⁵⁹ There were no differences in the measured size of the aortic root between the two groups. Postoperative echocardiography showed that there was no difference in indexed EOA or LVM regression between groups (Fig. 34-21). They also found no difference in functional outcome between valves at 1-year follow-up (Fig. 34-22). These findings challenge the notion that stentless porcine valves provide increased IEOA or hemodynamic or clinically significant benefit. In a randomized trial comparing Sr. Jude Toronto stentless porcine valves to Carpentier-Edwards pericardial valves, Chambers and colleagues found no difference in hemodynamic function including gradients or IEOA, LVM regression, or mortality between groups. Arenaza and colleagues also performed a multicenter randomized trial comparing the Medtronic Freestyle valve to the Medtonic Mosaic valve and found increased IEOA in the stentless group, but no differences in LVM regression or clinical outcomes at 1-year.

View larger version (18K): [in this window] [in a new window]   Figure 34-21 Indexed ventricular mass regression in stentless and stented valve patients over time. There were no significant differences in the two groups. CE = Carpentier-Edwards stented valve; LVMI = left ventricular mass index; SPV = Toronto stentless porcine valve. (Reproduced with permission from Cohen et al.359)

View larger version (21K): [in this window] [in a new window]   Figure 34-22 Change in Duke Activity Status Index (D.A.S.I.) scores in stentless and stented valve patients over time. There were no significant differences in the two groups. CE = Carpentier-Edwards stented valve; SPV = Toronto stentless porcine valve; Preop = preoperative. (Reproduced with permission from Cohen et al.359)

Hence, there is conflicting evidence that the use of stentless valves results in improved LVM regression or clinical outcomes over stented bioprostheses. Several studies have shown adequate LVM regression in patients receiving even small 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. There are reports of decreased thromboembolic events in stentless valves.³⁶²