The number of patients undergoing reoperation for valvular heart disease is increasing and will continue to increase as the general population ages.¹ These reoperations most commonly involve structural deterioration of a bioprosthesis or progression of native-valve disease after nonvalve cardiac surgery. Structural failure of a biologic valve should be considered a part of the natural evolution of tissue valves and should be fully appreciated by both the surgeon and the patient prior to implantation.² Reoperations are technically more difficult than primary operations because of adhesions around the heart with an associated risk of reentry, the presence of more advanced cardiac pathology, and the existence of more frequent comorbidities such as pulmonary hypertension. Perhaps most importantly, reoperative valve replacement operations often are performed in functionally compromised patients who tolerate complications poorly or who have little reserve.³ As a consequence of these and other factors, reoperative valve surgery historically has been associated with a considerably higher operative mortality than primary valve surgery, particularly in patients who have had multiple prior replacements.⁴ In the modern era, however, with the use of alternative surgical approaches and advanced perioperative care, there has been significant improvement in outcomes.^5-9

Reductions in operative risk and postoperative morbidity after reoperative valve surgery have been made in the past few years through advances in myocardial protection, as well as alternative perfusion strategies such as the proper use of deep hypothermic cardiac arrest.¹⁰ In addition, use of peripheral cannulation techniques to institute cardiopulmonary bypass (CPB) has become a relatively standard practice in reoperative cases.^11-13 Early institution of CPB prior to reentry prevents injury to the right ventricle (RV) or patent coronary artery bypass grafts during reoperative sternotomy.

Successful replacement of degenerated cardiac valves usually results in symptomatic and hemodynamic improvement. In this regard, improvements in valve design have mitigated but not eliminated primary bioprosthetic failure.^14-16 As such, the risk of re-replacement for bioprosthetic failure remains a significant factor to be considered in the selection of valve type for implantation.¹⁷

MECHANICAL VERSUS BIOLOGIC VALVES

The most appropriate valve substitute for an individual patient remains a source of much controversy. This choice should be adapted to each individual patient depending on age, life expectancy, life style, valve size, and cardiac as well as noncardiac comorbidities.¹⁸ Some studies comparing the long-term outcomes between biologic and mechanical aortic valve prostheses have yielded similar results with regard to overall valve-related complications.^19-22 However, most recent large studies have documented that anticoagulant-related bleeding with mechanical valves must be balanced against life expectancy and the risk of biologic valve re-replacement.^23-25 Bioprosthetic valves are known to undergo a time-dependent process of structural deterioration that results in freedom of reoperation of 80% at 15 years.²⁰ Consequently, structural degeneration of a bioprosthesis is the most frequent indication for reoperation in patients with tissue valves.^19,26

Despite this, recently improved durability of tissue valves, as well as the availability of stentless valves and homografts, has led to surgeons placing bioprostheses in progressively younger age groups.^18,27-30 Contributing to this trend, many patients do not want to accept the risk of anticoagulant-related hemorrhage associated with mechanical valves: major events 0.5% per patient-year and minor events 2 to 4% per patient-year.³¹ In addition, new concepts of valve-in-valve (VinV) interventions are suggesting to patients that there may be re-replacement of bioprosthetic valves in the future.

Mechanical prostheses usually are selected for younger recipients because of their proven durability over time. However, the risk of anticoagulant-related bleeding, as well as thromboembolic events (TEs), in these valves is not trivial and depends on valve design, structural materials, and host-related interactions.³¹ In a 12-year comparison of Bjork-Shiley versus porcine valves, Bloomfield et al documented severe bleeding complications in 18.6 versus 7.1%, respectively.³² Moreover, although endocarditis, dehiscence, perivalvular leak, and pannus formation are associated with both biologic and mechanical valves, acute prosthetic thrombosis is exclusively a complication of mechanical valves.^33,34 In considering mechanical valve durability, these associated risks cannot be ignored and must be weighed against the anticipated rate of tissue-valve failure and the need for reoperation.

RISK FACTORS IN REOPERATIVE VALVE SURGERY

In evaluating patients for reoperative valve surgery, certain factors are associated with added risk. For example, Husebye et al, in a review of their 20-year experience¹⁷ with reoperative valve surgery, found specific issues to carry higher risk (see Table 46-1). Overall operative mortality was 7% for the second and 14% for the third reoperation. Operative mortality for the first reoperation (n = 530 patients) was 5.9% in the aortic position and 19.6% in the mitral position. In the aortic position, operative mortality was 2.4% for New York Heart Association (NYHA) class I, 1.6% for NYHA class II, 6.3% for NYHA class III, and 20.8% for NYHA class IV emphasizing the significance of early referral. Regarding the urgency of surgery, the mortality for elective mitral valve reoperations was 1.4%; for urgent procedures, 8%; and for emergency procedures, 37.5%. Based on these findings, the authors have recommended that referral for reoperation be made when valve dysfunction is first noted, that is, before a significant decrement in myocardial function.¹⁷ Similarly, Jones et al reviewed their experience with first heart valve reoperations in 671 patients between 1969 and 1998.⁶ Their overall operative mortality for first-time heart valve reoperation was 8.6%, similar to the results published by Lytle et al³⁵ (10.9%), Cohn et al⁴ (10.1%), Akins et al³⁶ (7.3%), Pansini et al² (9.6%), and Tyers et al³⁷ (11.0%). In the Jones and colleagues series, mortality increased from 3.0% for reoperation on a new valve site to 10.6% for prosthetic valve dysfunction or periprosthetic leak; mortality was highest (29.4%) for associated endocarditis or valve thrombosis. Concomitant coronary artery bypass grafting (CABG) carried a higher associated mortality (15.4%) than when it was not required (8.2%). Among the 336 patients requiring re-replacement of prosthetic valves, mortality was 26.1% for re-replacement of a mechanical valve compared with 8.6% for re-replacement of a tissue valve. The authors concluded through multivariable analysis that significant predictors of mortality were year of reoperation, patient age, indication, concomitant CABG, and the replacement of a mechanical valve rather than a tissue valve.⁶

PREOPERATIVE IMAGING IN REOPERATIVE VALVE SURGERY

Reoperative cardiac surgery is associated with higher morbidity and mortality than the first operation^38-40 partially related to the risk of surgical reentry. The main technical concern during surgical reentry is the risk of lethal injury to vital structures such as the aorta, RV, and patent coronary artery bypass grafts. The key to mitigating the risk of reentry is via careful planning of the surgical approach. This starts with knowing the exact proximity of the RV, aorta, and prior grafts to the posterior sternal table. The standard preoperative coronary angiography and chest x-ray is usually uninformative about the precise anatomic relationship of these vital structures with the sternum.

Retrospective electrocardiographic (ECG)-gated multidetector computed tomographic (MDCT) scanning has arisen as the modality of choice to assess not only the heart’s location in relation to the sternum, but also graft’s location and patency.^41-43 Use of ECG-gated MDCT meticulously evaluates the structures of interest to the surgeon, including sternum, mediastinal structures, bypass grafts, and their relationship to each other.^44-46 At our institution, all reoperative cardiac surgeries are preceded by a preoperative cardiac computed tomography (CT) scan assessment. For patients with prior valvular (non-CABG) surgeries in which assessment of bypass graft is not necessary, noncontrast CT scan is adequate. In 2010, the Appropriate Use Criteria of the American College of Cardiology rated this application of cardiac CT as an appropriate use of this technology.^45,47

Preoperative CT scan will lead to a modification of surgical approach in 20% of patients undergoing redo cardiac surgery.^41,44 For example, if CT shows a distended RV adherent to the posterior sternum, it might be prudent to initiate CPB prior to attempting sternal reentry. In high-risk redo patients, MDCT is associated with a higher adoption of preventive maneuvers aimed at mitigating the risk of reentry.⁴⁸ There is evidence that preoperative MDCT reduces the risk associated with redo cardiac surgery⁴⁹ as well as shorter perfusion and cross-clamp time, shorter intensive care unit stays, and less frequent perioperative myocardial infarction (MI).⁵⁰

REOPERATIVE AORTIC VALVE SURGERY

Historically, aortic valve surgery typically involved the placement of a mechanical valve. In the past, there were only a few generally accepted indications to use a bioprosthesis for primary, isolated aortic valve replacement (AVR): (1) the presence of well-established contraindications to anticoagulation, (2) the inability to monitor prothrombin levels adequately, and (3) patients whose survival was limited and more dependent on nonvalve-related issues.^18,26 In recent years, however, the use of biologic valves in the aortic position has become more common.^24,51

As mentioned, reoperations are technically demanding, and many patients present in a poor functional state that further increases their mortality, in some series up to 19%.^32,52,53 Generally, optimal planning for reoperation prior to deterioration to NYHA class III to IV levels and before unfavorable comorbid conditions have arisen is imperative to ensure good outcomes.⁹ Following these guidelines in the modern era, elective re-replacement of malfunctioning aortic bioprostheses can be performed with results similar to those of the primary operation.^18,23,54 The Mayo Clinic, for example, recently reviewed its experience with 162 reoperative AVRs. Early mortality for reoperative AVR was not statistically different from that for primary AVR.⁵⁵ In light of recent lower operative mortality in reoperative valve surgery, a more conservative approach toward issues such as “prophylactic” AVR in patients with asymptomatic mild-to-moderate aortic stenosis (AS) at the time of CABG also may be more appropriate.⁵⁶

In evaluating the reoperative patient, the presence of concomitant coronary artery disease and pulmonary hypertension has been shown consistently to be independent risk factors.¹⁸ Patients with these risk factors therefore need careful surveillance once the probability of bioprosthetic dysfunction begins increasing (ie, 6 to 10 years after implantation).¹⁶ Regarding valve surveillance and timing of reoperation, the following variables are relevant to the clinical management of patients with an aortic bioprosthesis: a history of endocarditis before the first operation, perioperative infectious complications, coronary artery disease acquired after the first operation, an increase in pulmonary artery pressure, and a decrease in left ventricular function during the interoperative interval.¹⁸ Proper timing of the reoperation therefore is paramount because duration of clinical signs with a dysfunctional aortic bioprosthesis may be misleading. This is further supported by the fact that the need for emergency reoperation is the most ominous risk factor and consistently yields a high early mortality rate of 25 to 44%.⁵⁷

Reoperative Aortic Valve Surgery after Prior Coronary Revascularization

Reoperative valve surgery in patients with a prior patent left internal thoracic artery (LITA) to left anterior descending (LAD) graft is challenging for surgeons because of specific considerations for myocardial protection and prevention of patent graft injury. Unlike mitral valve surgery or CABG, in which cross-clamping of the aorta may be optional, aortic valve surgery mandates aortic cross-clamp unless hypothermic circulatory arrest is used. As such reoperative AVR in patients with a patent LITA graft presents a unique challenge of myocardial protection.

The incidence of LITA injury in the literature varies from 5 to 9%^58-61 and is associated with a 40% risk of perioperative MI⁵⁹ and 50% mortality rate.^62-64 This high morbidity and mortality rate warrants a preoperative decision concerning the most appropriate and safe surgical approach.

This is an area of recent controversy. The most traditional approach involves resternotomy, dissection of the LITA graft, and occlusion of the patent graft with a small bulldog and then aortic cross-clamping. This strategy has the advantage of short CPB time, maximizing uniform myocardial protection without cardioplegia washout but with an obvious risk of LITA injury during dissection. Efforts to reduce the risk of LITA injury have led to the “no-dissection” surgical technique in which the graft is left unclamped with use of deep hypothermic cardioplegic arrest.^58,65-67 In this approach, the LITA graft is not dissected and myocardial protection is achieved by using moderate-to-deep hypothermic cardioplegic arrest. This has the advantage of minimizing (but not eliminating) graft injury and its main disadvantage is that an unclamped LITA graft can lead to poor/uneven myocardial protection because of cardioplegia washout from the LITA territory. Another disadvantage is poor visualization due to continuous flow through the LITA. Other less utilized and studied approaches involve supraclavicular or endovascular control of the patent LITA graft.^68,69

It is worth mentioning that another strategy of handling patent LITA grafts involves deep hypothermic circulatory arrest without controlling the graft or aortic cross-clamping. This technique is occasionally indicated in patients with “porcelain” aortas.⁷⁰

All these concerns reiterate the importance of meticulous preoperative assessment and planning, to determine the optimal surgical approach for each individual patient. The “no-dissection” technique with hypothermic cardioplegia is a preferred approach for redo AVR in patients with patent LITA graft. Several studies have documented the safety of this method without any increased risk of perioperative mortality.^58,65-67

Approaches and Techniques

CONVENTIONAL RESTERNOTOMY

The evolution of cardiac surgery through the last few decades has led to the popularization of various surgical approaches. Thoracotomy, for example, once was used extensively to gain access to mediastinal structures. Then median sternotomy became the standard approach. In reoperative cases, however, repeating the sternotomy carries definite surgical risks. Before proceeding with a resternotomy, the relationship between certain anterior mediastinal structures (eg, the RV and the aorta) and the posterior aspect of the sternum must be assessed carefully.⁷¹ As described above, a CT scan with or without contrast is now the modality of choice.^41,43,62

Exposure of the femoral vessels and preparation for emergency femoral-femoral CPB should be performed before resternotomy. In cases of heightened concern for RV-graft injury or in cases in which a left internal mammary artery (LIMA) graft is patent, the surgeon should consider the use of CPB before chest reentry. Sternal wires from the previous operation should be undone carefully but left in place as a posterior safeguard during initial sternal division. An oscillating (not reciprocating) bone saw can be used to divide the anterior sternal table. An Army-Navy retractor, placed inferiorly in-line with the sternotomy can be used to stent open the wound during opening of the posterior table. Most authors recommend dividing the posterior table using a combination of scissors and anterolateral rake retraction.^62,71,72 Following this, bilateral pleural spaces should be entered inferiorly, followed by careful dissection of other mediastinal structures. The pericardial dissection plane can be developed by starting at the cardiophrenic angle and advancing slowly cephalad and laterally on the surface of the right side of the heart. Cephalad dissection should start with freeing the innominate vein before spreading the retractor to avoid its injury. Further dissection then is carried down to the superior vena cava (SVC), being careful to note the location of the right phrenic nerve. An area of consistently dense adhesions is the right atrial appendage, and caution should be used here. In addition, great care should be taken to avoid “deadventializing” the aorta. The area where the aorta apposes the pulmonary artery is another site of potential injury.

Repairing small ventricular or atrial lacerations should not be attempted before releasing the tension of the surrounding adhesions. Repair of great vessel injuries or severe RV injuries is best done on CPB.⁷¹ Severe active hemorrhage during a second sternotomy usually is caused by adherence of the heart or great vessels to the posterior sternum. Prevention of this complication by interposition of pericardium or other mediastinal tissue at the time of the first operation has been suggested but has debatable relevance.⁶² The incidence of resternotomy hemorrhage is between 2 and 6% per patient reoperation.^73-75 In a report of 552 patients who had undergone reoperative prosthetic valve surgery, 23 patients (4%) had complications related directly to sternal opening.¹⁷ Of these, five patients had injury to the right atrium, seven patients had lacerated RVs, nine patients had injuries to the aorta, and two patients had a previously placed coronary artery graft divided. Nineteen of the 23 complications occurred during a first reoperation. Overall, there were two operative deaths related to resternotomy. The first death involved division of a previously placed coronary artery graft during reentry. The second death was caused by laceration of the aorta with subsequent exsanguination.¹⁷ Of note, prior use of a right internal mammary artery (RIMA) graft can be particularly challenging because it frequently crosses the midline, and extreme caution must be used in first dissecting out this vessel.

Macmanus et al reviewed their experience with 100 patients undergoing repeat median sternotomy.⁷⁴ Eighty-one patients had one repeat sternotomy, whereas the others had undergone multiple sternotomies. All had had a previous valve procedure and were reoperated on for progressive rheumatic valvular disease or for complications related to the prosthesis. Complications included hemorrhage during reentry in eight patients, postoperative hemorrhage in two, seroma in four, and dehiscence, wound infection, and hematoma in one patient each. There was one operative death directly related to resternotomy hemorrhage.⁷⁴ When major hemorrhage does occur on sternal reentry, attempts at resternotomy should be abandoned, and the chest should be reapproximated by pushing toward the midline. The patient should be heparinized immediately while obtaining femoral arterial and venous cannulation. Blood loss from the resternotomy should be aspirated with cardiotomy suction and returned to the pump. Once bypass has been established, core cooling should be commenced with anticipation of the need for circulatory arrest. Once cool, flow rates can be reduced briefly for less than a minute, and the remaining sternal division can be completed, followed by direct repair of the underlying injury.⁶² Anticipating the possibility of this scenario, we frequently expose peripheral cannulation sites prior to beginning a resternotomy. In cases of heightened concern for RV or graft injury, or in patients with a patent LIMA to LAD artery graft, CPB and cardiac decompression may be initiated before sternal reentry. After safe sternal entry, the patient may be weaned from bypass for further dissection of adhesions to avoid prolonged pump times.

MINIMALLY INVASIVE REOPERATIVE AORTIC VALVE REPLACEMENT

Reoperative procedures are challenging owing to diffuse mediastinal and pericardial adhesions. A large incision that increases the operative exposure also has been associated with a higher risk of injury to cardiac structures and coronary artery bypass grafts and results in greater bleeding with its associated transfusion requirements.^76-79 A smaller incision with a limited sternotomy, on the other hand, reduces the area of pericardiolysis, thus mitigating these effects. The intact lower sternum that remains also preserves the integrity of the caudal chest wall, thereby enhancing sternal stability and promoting earlier extubation.^80,81 Minimally invasive valve procedures gradually have become more accepted as new technologies and instrumentation have been developed.⁸⁰ Reoperative procedures in which there is risk for graft injury are an area where minimally invasive strategies may be of direct benefit.^82,83 Our surgical approach in reoperative AVR is shown in Fig. 46-1.⁸⁰ In our series of patients, peripheral cannulation sites were exposed or cannulated before beginning the partial upper resternotomy. An external defibrillator is placed on the patient before draping for anticipated defibrillation as necessary. Transesophageal echocardiography (TEE) was used in every patient. A partial upper resternotomy was carried out to the third or fourth intercostal space depending on the estimated position of the aortic valve as documented by CT scan/TEE and then was “T’d to the right.”⁸⁴ The oscillating saw was used to divide the anterior sternal table, whereas the straight Mayo scissors, under direct visualization, was used to divide the posterior sternal table. In the setting of a patent LIMA-LAD graft or other anterior coronary artery bypass grafts, patients were placed on CPB before partial resternotomy. Mediastinal dissection was limited to only the ascending aorta as was necessary for clamping and aortotomy. The right atrium was dissected only if it was cannulated. Although intrathoracic cannulation was preferred, we frequently used peripheral cannulation to avoid clutter in the chest. Retrograde cardioplegia, if necessary, was delivered via a transjugular coronary sinus catheter or with right atrial placement under TEE guidance. Vacuum assistance of venous drainage was used in the majority of patients. Once on CPB, all patients were systemically cooled to 20 to 25°C. Patients with patent LIMA-LAD grafts were cooled routinely to 20°C for additional myocardial protection and in so doing avoided the need and potential hazard of dissecting out the LIMA for clamping in an attempt to avoid cardioplegia washout. If flow from the patent LIMA-LAD graft led to significant blood flow out of the coronary ostium and obscured the operative field, pump flows were turned down temporarily to allow visualization. Venting was accomplished by placing a pediatric vent through the aortic annulus. The aortic valve surgery then was performed based on patient indications. While closing the aortotomy, intracardiac air was removed by insufflating the lungs and decreasing flows on CPB. Carbon dioxide was used and flooded the operative field. Patients also were tilted from side to side to help with de-airing, and the ascending aortic vent was left open until separation from CPB. Temporary epicardial pacing wires were placed on the anterior surface of the RV while the heart was decompressed and before the aortic cross-clamp was removed. Two 32 French right-angled submammary chest tubes then were placed through the right pleural space, one angled medially into the mediastinum and one angled posterior into the pleural space. Decannulation and closure then were performed in the standard manner.

With our increasing experience in minimally invasive reoperative AVR, we have refined our technique as an alternative to conventional full resternotomy.⁸⁰ Technical details of the partial upper resternotomy approach are presented in Table 46-2. By following these guidelines, we have yet to convert any patient to a full resternotomy. CT scan and/or TEE are helpful in locating the level of the aortic valve and determining the proximity of the aorta to the posterior aspect of the sternum.⁸⁴ Also, extension of the sternal incision laterally on both sides through the intercostal spaces helps to later reapproximate the sternum. We have tried to limit mediastinal and pericardial dissection primarily to the aorta, believing that this is the principal reason for decreased bleeding and transfusion requirements postoperatively.^80,82,85,86 The RV, which often is attached to the sternum, does not need to be dissected. Also, injuries to patent but atherosclerotic vein grafts can be reduced with this “no touch” technique.⁸⁷

Arterial and venous cannulation sites can vary considerably, reflecting the individual choice of the operating surgeon and the sufficiency of intrathoracic space. Possible cannulation sites, other than standard ones, include the axillary artery, innominate vein, and percutaneous femoral vein.^13,88 Innominate vein or percutaneous femoral vein cannulation, as well as the use of TEE to place the retrograde cardioplegia catheter, has been extremely helpful in minimizing dissection of the right atrium. At present, we consider this approach to be useful for isolated, elective reoperative aortic valve surgery.⁸⁰

REOPERATIVE AORTIC VALVE REPLACEMENT AFTER HOMOGRAFT/ROOT/ALLOGRAFT

AVR with homografts and autografts was performed increasingly because of excellent freedom from thromboembolism, resistance to infection, and reasonable hemodynamic performance.²⁷ Although improved durability of current tissue valves has slowed this trend, autografts and, to a lesser degree, homografts remain popular in younger patients owing to durability and, in the case of autografts, the potential for growth.^30,89 Consequently, many patients will require aortic valve re-replacement for structural degeneration of their homograft or autograft valve.⁹⁰ It is expected that about one-third of patients younger than 40 years of age will require aortic valve re-replacement within 12 years of homograft placement. This is owing primarily to calcification and structural valve degeneration. As such, the issue of homograft or autograft durability is particularly pertinent in this subgroup of younger patients who are expected to live beyond 15 years from the time of operation.⁸⁹

The incidence of patients with homografts or autografts in need of a second valve operation is expected to increase owing to the aforementioned recent popularity and availability of these conduits. Also, there is varied opinion as to the optimal surgical method of primary homograft AVR, with increased rates of aortic insufficiency (AI) in patients with the subcoronary implantation technique. Importantly, the selected technique of primary homograft operation may have relevance at reoperation because calcification or aneurysmal dilatation of the homograft may pose surgical challenges at reoperation. Despite these challenges, Sundt and others^29,91,92 have documented the feasibility of aortic valve re-replacement after full-root replacement with a homograft. In our own series of 18 patients, full-root, mini-root, and subcoronary techniques all were amenable to valve re-replacement.²⁷

How to best approach the reoperative root scenario and which valve to reimplant, however, have been debated. At one extreme, Hasnat et al documented the results of 144 patients who underwent a second aortic homograft replacement with a hospital mortality rate of only 3.5%.⁹⁰ Although Kumar et al, in a multivariate analysis of reoperative aortic valve surgery, did not show that a previous homograft added significant risk,⁹³ the technical aspects of reoperative AVR in this patient population consistently have been found to be challenging owing to the heavy calcific degeneration that invariably occurs. With this in mind, and owing to the typical absence of the need for a second root operation, we and others⁹⁴ believe that a more simplified approach to reoperative aortic valve surgery in patients with previously placed homografts may be optimal. Our approach has been to perform aortic valve re-replacement using a mechanical valve or, less commonly, a stented xenograft while reserving a second homograft and root operation for specific indications such as endocarditis, associated root pathology, or a very young patient with contraindications to a mechanical valve.

An open valve-in-homogragft approach involves resection of the degenerated or infected aortic homograft leaflets and a new valve is seated within the aortic homograft valve annulus without a need for root reconstruction.⁹⁵ The procedure is done via a median sternotomy, which is preceded by peripheral cannulation for CPB via femoral vein and artery or right axillary artery. The heart is dissected out, CPB is initiated, the patient is fibrillated and cross-clamped; antegrade (and occasional retrograde) cardioplegia is initiated.⁹⁵ The aorta is opened between the homograft root and the ascending aorta. If there is heavy calcification, a vertical or S-shaped aortotomy is made. The homograft leaflets are resected and the calcified annulus debrided with careful endarterectomy of the proximal ascending aorta. The endarterectomy is carried out starting at the annulus and coming up to allow stitches on in the softer aorta.⁹⁵ The valve prosthesis (bioprosthetic or mechanical) is then sized and implanted using interrupted noneverting 2.0 Ethibond (Ethicon, Somerville, NJ) pledgeted sutures. The valve is then seated and the sutures tied and cut. The aorta is then closed with a running 3-0 Prolene (Ethicon) suture. The cross-clamp is removed and the heart defibrillated into normal sinus rhythm.⁹⁵

Homograft re-replacement nonetheless is performed but it is much less common, and hospital mortality varies widely across many centers, ranging between 2.5 and 50%.^29,91,92 David and colleagues, for example, recently reviewed their experience with root operations in 165 patients who previously had undergone cardiac surgery. Of these, 28 had a previous root operation. Overall, 12 operative (7%) and 20 late deaths (12%) occurred.⁹⁶ Variations in sample size, valve selection, surgical techniques, and patient factors, as well as the experience of the surgeons, may account for these wide differences.

AORTIC VALVE BYPASS SURGERY

Aortic valve bypass (AVB) surgery, also known as apical aortic conduit surgery, is an alternative for high-risk and “inoperable” reoperative patients with AS. It has been used in patients with reduced LV function,⁹⁷ porcelain aorta,⁹⁸ severe patient prosthetic mismatch,⁹⁹ excessive comorbidities, and vulnerable functional grafts¹⁰⁰ leading to prohibitive risk for conventional reoperative AVR. AVB surgery works by shunting blood from the apex of the left ventricle to the descending aorta through a surgically placed valve conduit (Fig. 46-2). This approach avoids cross-clamping, cardioplegic arrest, potentially CPB,¹⁰¹ debridement of the native valve, and injury to patent grafts because the procedure is performed through the left chest. Patient-prosthesis mismatch is also unlikely as the indexed effective orifice area (EOAi) is the sum of the valves in both the native position and conduit.