Chapter 32. Aortic Valve Replacement with a Mechanical Cardiac Valve Prosthesis

In 1931, Paul Dudley White stated, "There is no treatment for aortic stenosis." Even today the medical therapy of aortic stenosis has not significantly advanced (Fig. 32-1).1 Conversely, patients may tolerate aortic insufficiency for many years, but as the ventricle starts to dilate, a progressive downhill course begins and early operation is warranted.2 Definitive therapy for aortic valve disease was unavailable until the advent of cardiopulmonary bypass. Innovative cardiovascular surgeons then began to develop cardiac valve prostheses. Over the subsequent 50 years3 the variety of prostheses that have become available for use have expanded greatly. Available aortic valve substitutes include mechanical valve prostheses, stented biologic valve prostheses, stentless biologic valve prostheses, human homograft tissue (both as isolated valve replacement and aortic root replacement), percutaneous or transapical biologic valves and a combination of a biologic valve using a pulmonary autograft, and pulmonary outflow tract replacement with heterograft prostheses (Ross procedure). This chapter focuses on the use of mechanical valve replacement in the aortic position.

In 1952, Hufnagel used an aortic valve ball and cage prosthesis heterotopically in the descending thoracic aorta to treat aortic insufficiency.4 After the advent of cardiopulmonary bypass, initial attempts at aortic valve replacement (AVR) consisted of replacement of the individual aortic cusps with Ivalon gussets sewn to the annulus. When successful, these prostheses often calcified and results were short lived. Shortly thereafter, surgical pioneers Starr, Braunwald, and Harkin began replacement of the aortic valve in the orthotopic position. First-generation aortic valve prostheses, the ball and cage, became the standard for AVR for more than a decade (Fig. 32-2). Many of these prostheses have remained durable for up to 40 years.5,6 Multiple modifications ensued, including changing the material of the ball from silastic to stellite, changes in the shape of the cage, depression of the ball occluder, the addition of cloth coating to the sewing ring and the cage, and changes in the sewing ring itself. These valves, however, required intense anticoagulation.7 Hemodynamic performance was compromised because there were three areas of potential outflow obstruction: (1) the annular size of the sewing ring (the effective orifice area of the valve); (2) the distance between the cage and the walls of the ascending aorta (particularly in the small aortic root); and (3) obstruction to outflow by the ball itself distal to the tissue annulus. Flow patterns were also abnormal (Fig. 32-3). These problems led to the development of the next generation of aortic valve prostheses—the tilting disc valve. Innovators such as Björk, Hall, Kaster, and Lillehei developed three models of tilting disc prostheses that became the second generation of commonly implanted aortic valve replacement devices between 1968 and 1980. The low-profile configuration simplified surgical implantation (Fig. 32-4). Problems with the tilting disc valve included stasis and eddy current formation at the minor flow orifice (see Fig. 32-3), and sticking or embolization of the leaflet, the latter leading to discontinuation of the Björk prosthesis in spite of otherwise good long-term results.8 The Lillehei-Kaster prosthesis has evolved into the Omniscience valve. The Medtronic Hall valve, the third tilting disc prosthesis, now discontinued (Fig. 32-5).

Kalke and Lillehei developed the first rigid bileaflet valve, but it had very limited clinical use. In 1977, the St. Jude Medical (SJM) prosthesis was developed and implanted by Nicoloff and associates (Fig. 32-6).3,9 Over the following decades, the dramatic step of a bileaflet prosthesis nearly obviated the use of all other kinds of mechanical prosthetic valves in the United States and to a large extent elsewhere. The SJM valve demonstrated low aortic gradients, minimal aortic insufficiency, and low rates of thromboembolism (TE) (Fig. 32-7).9–11 Anticoagulation continued to be necessary, but to a lesser extent than with previous design models.12 Because of the low-profile design and lesser need for orientation, surgical implant was further simplified. After the introduction of the SJM valve, several other third-generation models of bileaflet prostheses were introduced, including the Sulzer Carbomedics valve (Sorin S.p.A., Milan, Italy) (Fig. 32-8), the ATS Medical prosthesis (Medtronic, Minneapolis, MN) (Fig. 32-9), and the On-X prosthesis (On-X Life Technologies, Inc., Austen, TX) (Fig. 32-10). Since the introduction of the bileaflet valve, more than 2 million implants on a global basis have been accomplished and extensive literature has developed. Surgeons have become more confident in earlier aortic valve replacement, and guidelines for anticoagulation necessary for all mechanical valves have been developed for each generation of prosthesis at progressively decreasing target levels.12

Over the past 25 years, design and configurational changes have been made in bileaflet prostheses. The ATS Medical valve changed the "rabbit ears" pivot style of other bileaflet prostheses, incorporating a convex or open-pivot design allowing more complete washing of the moving parts of the valve and possibly a quieter valve closing.13,14 The sewing ring of the SJM valve has changed (SJM HP) to allow a larger valve size implantation for any given tissue annulus, as has ATS Medical with its AP design. The sewing ring of the Sulzer Carbomedics valve has been modified such that this valve is implanted in a supravalvular position (top hat model). The On-X valve incorporates advanced pyrolytic carbon technology using a purer, more flexible coating to allow flanging of the inflow portion of the valve housing, mimicking the normal flow pattern.

The most recent development in bileaflet valve design was the introduction of the SJM Regent valve (Fig. 32-11). This valve model not only modified the sewing ring, but also redefined the external profile in a nonintrinsic structural portion of the valve, increasing the effective flow orifice area. Thus, a larger prosthesis could be implanted for any given tissue annulus diameter. This was the first mechanical prosthesis to demonstrate left ventricular mass regression across all valve sizes.15,16 The Regent valve is seated supra-annular with only the pivot guards protruding into the aortic annulus.17

As with any medical therapy, AVR with a mechanical prosthesis is not indicated for all patients. Several prospective randomized studies have shown no difference in survival in patients having biologic or mechanical valve prostheses or among mechanical prostheses per se.18–22 However, follow-up was limited. Conversely, in other nonrandomized studies of patients followed over longer time frames, freedom from all valve-related events and from reoperation were improved in patients with mechanical valve prostheses as compared with patients with biologic prostheses.8,23

Most recently several publications have shown improved survival in patients having bileaflet mechanical valve prosthesis, most likely related to improved and longer patient follow-up.24,25 Importantly quality of life was similar to that of a biologic prosthesis; even in elderly patients.25

Although the advantages of large effective flow orifice and durability with a mechanical valve are paramount, the confounding effects resulting from the necessity for anticoagulation continue. Patients who are transient, noncompliant, or incapable of managing medications are not good candidates for long-term chronic anticoagulation, nor are those with dangerous lifestyles or hobbies.26 Patients with higher levels of education, and those from geographic areas with a sophisticated medical infrastructure and a static population have better compliance with necessary medication, anticoagulant monitoring, and fewer risk factors for TE are good candidate for AVR with a mechanical valve prothesis.27

A mechanical valve prosthesis is recommended to patients having second valve reoperations regardless of the nature of the first procedure, as re-reoperative risks are substantial.28,29 Some studies report low mortality for reoperation of patients with failed biologic valves, but failures can occur abruptly, creating more risk.30 Reoperative risk is also higher in those patients having combined procedures28,31 or after prior coronary bypass.

Many surgeons have opted for an age of greater than 70 years as the indication for bioprosthetic AVR, based on data by Akins and associates.28 In patients younger than 60 years of age, most would opt for a mechanical prosthesis based on prosthesis durability.32 In the decade between 60 and 70 years of age, other factors have to be taken into account.33,34

Implantation of mechanical valve prostheses has been previously described and is straightforward.11 Historically, high-profile aortic valve prostheses could be difficult to implant, particularly in small aortic roots. In such cases, a hockey-stick aortotomy is used to "unroll" the aorta and expose the annulus. Although the implantation of low-profile bileaflet prostheses is simpler, problems can still arise in the small aortic root. If a tilting disc prosthesis is used, orienting the major flow orifice toward the greater curve of the aorta is necessary. Because bileaflet prostheses are the most commonly used, the surgical technique for implantation of these devices is described: A midline incision and sternotomy is made and a pericardial well created. Alternatively, a right anterior thoracotomy approach with femoral cannulation has been reported. A partial sternotomy is also an alternative in thin patients, creating a sternal T at the fourth inter-space.35 These alternative techniques are particularly amenable to the implantation of low-profile aortic valve prostheses. The patient is cannulated via the aorta and a single atrial venous cannula. Most commonly, retrograde cardioplegic solution is used and a left ventricular vent is placed via the right superior pulmonary vein to maintain a dry operative field. After cross-clamping of the aorta, a transverse aortotomy is made approximately 1 cm above the takeoff of the right coronary artery, slightly above the level of the sinotubular ridge (Fig. 32-12). The incision is extended three-quarters of the way around the aorta, leaving the posterior one-quarter of the aorta intact, and allowing excellent visualization of the native aortic valve and annulus. The leaflets of the aortic valve are excised to the level of the annulus and the annulus is thoroughly debrided of any calcium. Extensive decalcification will minimize the risk of paravalvular leak, particularly in newer-generation prostheses with thinner sewing rings, and allows for better seating of the valve prosthesis. Braided 2-0 sutures with pledgets are used. Beginning at the noncoronary commissure, the annulus is encircled with interrupted mattress sutures (Fig. 32-13) extending from the aortic to the ventricular surface (everting). Alternatively, multiple single interrupted sutures may be placed. After placement, the suture bundles are divided into two equal portions and two individual sutures placed into the sewing ring at the level of the pivot guards, orienting the pivot guard toward the ostia of the left and right coronary artery (Fig. 32-14). Next, each half of the suture bundles are implanted in the sewing ring and the prosthesis seated (Figs. 32-15 and 32-16).

The pivot guard sutures are tied first followed by the sutures beginning at the left coronary cusp extending to the mid-portion of the right coronary cusp. The sutures of the noncoronary cusp are secured last, seating the valve appropriately. In a small aortic root, if a valve is not able to be seated, a paravalvular leak can be prevented if the unseated area of the valve is in the noncoronary cusp. External aortic sutures can be placed from outside the aorta to the valve sewing ring, securing the prosthesis and preventing regurgitation. Because of the low-profile nature of the leaflets, opening and closing can still occur unimpeded. Leaflet motion should always be checked, and the surgeon must be assured that the coronary arteries are not obstructed. The aortotomy is closed with a double layer of polypropylene suture consisting of an underlying mattress suture and a more superficial over-and-over suture. The patient is placed in the Trendelenburg position and the heart filled with blood and cardioplegic solution, vented, and the cross-clamp removed. After resuscitation and de-airing of the heart, the procedure is completed and the patient transferred to the intensive care unit. On the first postoperative day the chest tubes are removed if output is less than 125 mL in the previous 8 hours. After the removal of the chest tube, the patient is begun on subcutaneous heparin (5000 U every 8 hours) or low molecular weight heparin (1 mg/kg twice a day), and warfarin therapy started. Valve implantation usually can be accomplished in less than 40 minutes of aortic cross-clamping and with cardiopulmonary bypass times of approximately 1 hour, allowing limited coagulopathic and homeopathic alterations.

When coronary bypass grafting is indicated, the order of the operation changes. The diseased valve is excised, distal vein or free arterial grafts constructed, the valve is replaced, and the aortotomy closed. Proximal anastomoses are then completed, with one left untied for de-airing. The distal anastomoses of pedicled grafts (internal mammary artery, IMA) are then completed. De-airing is accomplished through the untied proximal anastomosis.

The durability and function of mechanical valve prostheses, particularly those of the modern generation, is unquestioned.23,32,36–39 It is the process of anticoagulation that is key and drives long-term success. The international normalized ratio (INR) is the standard to which anticoagulation levels should be targeted.26,40 Anticoagulation is begun slowly after removal of the chest tubes, as the danger of overshooting target INR to dangerous levels is common.41 Current data on anticoagulant regimens indicate that a one-size-fits-all recipe is inadequate to obtain excellent long-term results.12,27 Horstkotte and associates noted that complications occur during fluctuations in the INR, and less often during steady-state levels, whether high or low.42 When levels of INR increase, bleeding episodes become more common, and when levels of INR decrease, thromboembolic episodes become more common, both on the slope of the change. These events are opposite ends of the continuum of anticoagulation-related complications. The presence of a mechanical valve prosthesis is also not the only risk factor for TE.27,43 Traditional risk factors for TE listed in Table 32-1 predispose patients to thromboembolic episodes, and as such higher therapeutic INRs are warranted. Similarly, as shown in Table 32-2, nontraditional risk factors for thromboembolism will also predispose patients to embolic events.27,43 Butchart has noted that the more of these risk factors patients have, the greater the incidence of events and the greater the need for a higher target INR (Fig. 32-17).27 Thus, it is imperative in the modern era that patient risk factors be taken into account and the INR individualized for a given patient.12 Recommendations for INR target levels in our practice are shown in Table 32-3. These levels are more liberal than those offered by the American College of Cardiology/American Heart Association (ACC/AHA) and the American College of Chest Physicians (ACCP) guidelines, but more conservative than those recommended by the European self-anticoagulation trials.44–46 These later reports are especially relevant, because they demonstrate that a lower INR is consistent with a lower incidence of TE if patients are maintained in the therapeutic target range.44,47 Patients with home testing were maintained in the therapeutic range a substantially greater percentage of the time than those whose status was monitored at anticoagulation clinics.44,47 Starting self-management early after mechanical valve replacement further reduced valve-related events.41 In the United States, home testing has not become commonplace or popular. However, home testing could certainly be expected to lower the incidence of valve-related thromboembolic and bleeding events. It has recently been approved for reimbursement for weekly testing in patients with a mechanical valve prosthesis or atrial fibrillation, but only after a 3-month waiting period. Obtaining appropriate funding for access in the immediate postoperative period would be an important initiative that could acutely reduce the incidence of valve-related events.

A recent report of patients followed over 25 years noted that approximately 40% of the bleeding episodes occurred in the first year after surgery. It is thus important during this initial postoperative time frame when the patient's anticoagulant levels are more likely to fluctuate, that INR be measured more frequently.32 In the early postoperative period, INR can occasionally jump to supratherapeutic levels and result in significant bleeding events. This is an independent risk factor for mortality at 60 days.48 Furthermore, the most important independent predictor of reduced survival is anticoagulant variability.26 We and others therefore recommend in the early postoperative period that one proceed slowly to bring the INR to target levels while the patient is under the protection of subcutaneous enoxaparin (100 IU/kg twice a day) or heparin (5000 U every 8 hours) until the INR is therapeutic.12,32,49,50

The addition of aspirin to a warfarin regimen can be expected to result in a lower incidence of TE at any given therapeutic INR with a low probability for bleeding events and so is recommended.51,52

An educational program to teach patients how to manage their anticoagulation is an important part of the overall operative process. Patients should be instructed on the influence of alcohol and diet on anticoagulant levels, the need for regular dosing, and the potential impact of travel and gastrointestinal illnesses on fluctuations in anticoagulant levels. Warfarin is well known to be high-risk, as is insulin, yet both drugs can be managed with proper compliance and education so that the impact on lifestyle and quality of life is minimal.7,25,53 Patient age does not appear to be a risk factor for anticoagulation53–56 as long as there are no specific contraindications to anticoagulation. The presence of a mechanical valve is not a risk factor for long-term neurocognitive dysfunction.57 The importance of regular testing cannot be overstated.

Newer antithrombin agents may obviate several of the issues discussed in the preceding. These agents are showing promise in the treatment of atrial fibrillation, with a lower incidence of embolism and bleeding complications. The drugs are expensive, require multiple administrations per day, and may cause hepatic dysfunction, yet they do not require blood testing or physician visits to maintain the therapeutic effect.58 Their application to mechanical valve prostheses is as yet unknown.

Of importance platelet activation may be more important in the long-term therapy of an aortic prosthesis than mitral, where areas of stasis predominate.10 This is likely the reason there is no difference in the freedom from embolic events in mechanical versus biologic prosthesis over the long term.23 Thus aortic valve prostheses could in theory be managed with newer strong antiplatelet therapy. Garcia-Renaldi has tested this theory in 178 patients followed out to 7.8 years treated only with clopidogrel. Results were excellent, with few bleeding episodes and minimal thromboembolic events.59 Importantly, this group found the majority of patients having thromboembolic events when tested were resistant to the drug or had been removed from the drug by themselves or a physician. The authors stressed the importance of measurement of platelet responsiveness if only antiplatelet therapy is used.59 A prospective randomized trial is warranted to confirm those results.

Outcomes from aortic valve replacement with mechanical valve prostheses vary among reports depending on the patient population. Patients with higher-risk factors for TE and those with risk factors for anticoagulation will have a higher incidence of valve-related events, making meta-analyses less meaningful.60 Older patients are at higher risk for valve-related events, particularly thromboembolic episodes, because of the greater number of risk factors that accumulate with aging.27 The incidence of valve-related events is also determined by the intensity with which the investigators follow their patients. A higher incidence of early hemorrhagic events may also be diluted over the longer periods of follow-up.61 Compliance is vital to good long-term outcomes. Traditional and nontraditional risk factors for embolism and risk factors for anticoagulation and valve-related events must be considered.27,43 Several trials have indicated no significant differences in events among various mechanical prostheses, but follow-up was short.19,21,62 There are, however, certain standards within which one should expect a mechanical valve to perform, and within which the medical decisions for anticoagulation must be made. The majority of valve-related morbidity is related to TE and anticoagulation-related hemorrhage.26,29,39 The sections that follow deal with specific valve-related complications and acceptable current incidence.

Valve Type

Freedom from all valve-related events over the long term is shown in Fig. 32-18. In the early follow-up period, anticoagulation-related hemorrhage is the most common untoward event for mechanical valve prostheses. Thus, over the first 10 years of follow-up there is a higher incidence of valve-related events in patients with mechanical prostheses as opposed to those with biologic valves.23 However, over the subsequent period of 10 to 20 years, the incidence of biologic valve failure changes this ratio such that the valve-related complications of biologic prostheses become more common than those with mechanical valve prostheses. In a series of aortic reoperations, Potter has noted that the time to biologic valve failure was only 7.6 years and Maganti 11 years.30,63 This failure rate will increase over time.23,64 Overall, freedom from valve-related events is more strongly influenced by preexisting comorbidities than the presence of a mechanical prosthesis per se.23,27,32,36

Anticoagulant-Related Hemorrhage

Anticoagulation-related hemorrhage (ARH) is the most common valve-related event. The more intense the anticoagulation regimen, the higher the incidence of valve-related hemorrhage. Most commonly ARH occurs during fluctuations in INR values commonly related to changes in warfarin dosing or medical or drug interactions.42 The most common site for ARH is the gastrointestinal tract, and the second is the central nervous system.32 ARH also accounts for the highest incidence of patient mortality for valve-related events. Acceptable ARH rates range from 1.0 to 2.5% per patient-year in long-term reports.8,23,32,36–39 These long-term reports dilute the short-term impact, because ARH risks are higher early after valve replacement.32,39,61 With individualized and home-monitored anticoagulant regimens, both related events, TE and ARH, are diminished.44 Freedom from anticoagulation at 10 and 20 years are 75 to 80% and 65 to 70%, respectively. Importantly, one long-term study noted that nearly 40% of all ARH that occurred over a 25-year follow-up period occurred during the first year of anticoagulation (Fig. 32-19), indicating that a slow increase to therapeutic levels, coupled with close follow-up during this early period is warranted.12,32,48,49 Results of the European self-anticoagulation study indicate that a lower INR target is appropriate if home testing is initiated, because greater time is spent in the therapeutic range.44 Mortality more commonly occurs in relation to bleeding events than to thromboembolic events.12,32

Thromboembolism

Thromboembolic episodes are the second most common valve-related event and are the major reason that chronic anticoagulation is warranted. Khan and associates reported in a large series of patients that the incidence of thromboembolic events between bioprostheses and mechanical prostheses are the same (Fig. 32-20), but the mechanical valve patients are on warfarin.23 Acceptable thromboembolic rates range between 0.8 and 2.3% per patient-year.8,23,32,36–39,65 Approximately one-half of these events are neurologic, 40% are transient, and 10% peripheral.32 Freedom from thromboembolic events at 10 and 20 years is approximately 80 to 85% and 65 to 70%, respectively.

Thromboembolism is a continuous risk factor that is present throughout the life of the mechanical valve prosthesis. As patients age, risk factors for TE increase, so one must be on guard to maintain therapeutic anticoagulant levels. Changes in the target INR may be necessary as individual risks increase.

Interestingly, not all neurologic events classified historically as embolic are indeed embolic. Piper and coworkers report a study of patients prospectively admitted to a single institution with a neurologic event postmechanical valve replacement and intensively worked up. More than 75% were found to have intracerebral hemorrhage as the etiology of the event as opposed to embolic. This would indicate that target anticoagulant levels may be artificially high, and that not all neurologic events that appear to be embolic in fact are so.66

Valve Thrombosis

Valve thrombosis in the aortic position is an unusual event that occurs late after valve replacement and is most commonly a result of inadequate anticoagulation or noncompliance.67,68 In bileaflet valves, thrombus formation impinging valve function occurs at the pivot guards and in the crevices of the valve. Only one bileaflet design does not have convexities into which the leaflets fit.69 In tilting disc valves, thrombus is most common at the minor flow orifice. The incidence of thrombosis is approximately less than 0.3% per patient-year, and freedom from valve thrombosis at 20 years is greater than 97%.8,23,32,36–39

Prosthetic Valve Endocarditis

Prosthetic valve endocarditis is also a rare event in the modern era with prophylactic antibiotics. Approximately 60% of events occur early and are associated with staphylococci. The mortality for this event is high. The remainder appear late (>60 days). Prosthetic valve endocarditis is also a continuous variable, and patients must be cautioned to take prophylactic antibiotics for any invasive procedure. Freedom from endocarditis with mechanical valve prostheses is 97 to 98% at 20 to 25 years.32,39

Paravalvular Leak

Paravalvular leak is an operative complication and is most commonly related to operative technique but occasionally to endocarditis. With annular decalcification and closely placed sutures, these events can be minimized. The Silzone experience showed an increased incidence of paravalvular leak wherein the silver impregnated in the sewing ring not only impeded bacterial growth, but also healing of the annular ring, doubling the accepted rate of this complication.70 The Silzone-coated sewing ring was removed from the market. There may be an anatomical predisposition to paravalvular leak in the area of the annulus extending from the right and noncoronary commissure, one-third the distance along the right coronary cusp, and two-thirds the distance to the noncoronary cusp, owing to intrinsic weakness in this area of the annulus.71 The acceptable range of paravalvular leak is approximately less than .01% per patient-year, with early postoperative occurrence predominating.29,32

Structural Failure

Structural failure of bileaflet aortic prostheses resulting from wear has not been observed or reported in long-term studies totaling more than 50,000 patient-years of follow-up. This indicates the high structural integrity of these modern aortic devices.23,32,39 In a single study totaling 21,742 patient-years with 94% complete follow-up, there was no structural failure.49

Freedom from Reoperation

The long-term durability of modern mechanical valve prostheses is excellent, and a valve replacement rate of less than 2% over 25 years (Fig. 32-21) can be expected, and re-reoperation after AVR replacement is even more rare.29 Subvalvular pannus formation is also rare with aortic bileaflet valves.32,39 The most common reasons for prosthetic valve reimplantation are preoperative and postoperative endocarditis, paravalvular leak, and valve thrombosis.

Technical Considerations

Although the technique of implanting a mechanical valve prosthesis is very straightforward, special circumstances arise. Because of the large size of the valve housing of the St. Jude Medical Regent valve compared to the tissue annulus, entry into the aorta can sometimes be difficult. Occasionally, patients have the smallest diameter of their aortic root at the sinotubular ridge. Although the sizer will pass readily through the aortic annulus, seating the Regent valve itself into the annulus can sometimes be difficult and frustrating. It is important to gently rock the valve back and forth through the sinotubular ridge, tilting the valve circumferentially through the narrowest part. Once the valve is below the level of the sinotubular ridge, it will seat readily in the annulus if sizing has been correct. When tying the valve into the annulus, sutures in the pivot guards and the left and right coronary cusps should be completed first. The last sutures ligated are those of the noncoronary cusp for two reasons. It may seem like the Regent valve will not seat because of the large-sized valve housing; however, with gentle persistence seating can be completed as long as sizing has been correct. The Regent valve sits supra-annular and only the pivot guards lie inside the annulus (Fig. 32-22). Therefore one should leave the last sutures to be tied in the mid-part of the noncoronary cusp with the valve oriented so the leaflets are parallel to the ventricular septum.72 With proper annular decalcification and flexibility of the annulus, we have not seen a Regent valve that has not been able to be seated properly.

Similarly, when using the On-X valve one has to be sure that the flare of the valve inflow is seated properly in the left ventricular outflow tract. Gentle manipulation and patience may be necessary. Walther and colleagues note that exact sizing requires some experience.20

If an oversized bileaflet valve has been implanted purposefully in a small aortic root, the valve can be tilted and will still function without coronary obstruction as long as the highest portion of the valve is in the noncoronary cusp. Pledgeted sutures placed from outside the aorta through the sewing ring of the valve can prevent paravalvular leak, and opening and closing can still occur because of the low-profile nature of the prosthetic valve. In our review of nearly 3000 bileaflet aortic valve replacements, no annular enlarging procedures were completed.49

Patient-Prosthesis Mismatch

Patient-prosthesis mismatch (PPM) is a concept first described by Rahimtoola and popularized by Pibarot and Dumesnil.73,74 Unfortunately, views are varied on the importance of PPM.75–78 This is because virtually all contributions to the literature study mixtures of mechanical and biologic prostheses and varying types of each. In a 25-year follow-up of patients with a single model mechanical valve prosthesis, no difference was found in overall valve-related mortality for patients who had severe PPM, moderate PPM, or insignificant PPM according to the criteria described by Blais and associates.77 This similarity in long-term survival was sustained whether the effective area was measured by in vitro (internal geometric valve area) or in vivo criteria (echo-calculated valve area) as shown in Figs. 32-23 and 32-24. This study also found no difference in valve-related events, including operative mortality, long-term cumulative mortality, anticoagulant-related hemorrhage, TE, valve thrombosis, paravalvular leak, or diagnosis of congestive heart failure. Follow-up in this study was 94% complete and extended over 13,000 patient-years.79 Therefore, with bileaflet mechanical valve prostheses, PPM does not appear to be an issue. This study was limited in that it did not address patient age (ie, younger versus older), patient activity, or ventricular function. Thus, it is likely that PPM is important in patients with small biologic prostheses, because as the valve leaflets stiffen, clinical aortic stenosis becomes prominent early in the postoperative follow-up period, affecting symptoms and survival in younger and more active patients and those with depressed ventricular function. However, if one is concerned about PPM, the indexed effective orifice area can be calculated for any given prosthetic valve and a determination made whether an annular enlarging procedure or a different model prosthesis with a larger effective orifice area is warranted.80 PPM has been minimized by the new-generation Regent valve and is very rare with this prosthesis.16,49

Off Anticoagulation

Anticoagulation is recommended in all patients with mechanical valve prostheses. Limited trials have been undertaken with low-risk patients, but only after a several-month course of systemic anticoagulation. An increased incidence of valve thrombosis but with little increase in the incidence of TE has been reported in patients not taking chronic anticoagulation if antiplatelet agents are used.67,81,82 One study found no significant differences in valve-related events in patients on warfarin as compared with antiplatelet therapy alone, but follow-up was limited.83 One prospective study is currently ongoing, consisting of a randomized trial of antiplatelet therapy versus warfarin after 3 months of formal anticoagulation, but the results are not yet available.84 Certainly one can expect that highly selected patients with mechanical valve prostheses will do well off warfarin on antiplatelet agents, but this is unproved.17,85

In a prospective nonrandomized trial of clopidogrel along after bileaflet aortic valve replacement, Garcia-Rinaldi and coworkers as noted, found thromboembolic events limited to those patients who had clopidogrel discontinued or were nonresponders. Although such an approach is rational based on deductive reasoning from available data, a prospective randomized trend is required before this approach can be recommended.

When anticoagulation requires reversal electively, as for scheduled surgery, the INR is allowed to slowly drift toward normal over 5 days and the patient is admitted for intravenous heparin therapy 24 hours before the procedure. Anticoagulation is restarted after the procedure with antiplatelet therapy, and subcutaneous heparin with warfarin restarted on postoperative day one. Abrupt reversal of INR in patients who have bleeding episodes may be warranted, but carries increased risk of TE. Fresh-frozen plasma will gently reverse INR when necessary, but it is best to avoid the use of vitamin K. Frequent INR checks are warranted.

After ARH episodes, when feasible, because of the high incidence of recurrent bleeding, anticoagulant therapy is withheld up to 2 weeks or until the source of the bleeding has been identified and definitively treated, using antiplatelet agents only.82 For those patients in whom anticoagulant therapy cannot be restarted, antiplatelet therapy is warranted, but the patient should be informed of an increased incidence of thromboembolism to approximately 4% per patient-year and that of valve thrombosis to 2% per patient-year with bileaflet valves.67,81,82,86

Mechanical Valve Replacement in the Younger Patient

A major deterrent to mechanical valve replacement in the younger patient is the impact of long-term anticoagulation. Mechanical valves are, however, more ideal for younger patients because of their excellent durability characteristics. Most importantly, younger patients (i.e., patients less than age 50) are a low-risk subset for valve-related events. These individuals have very few risk factors for TE, and thus anticoagulation can be run at the lower end of the therapeutic target range, decreasing the incidence of anticoagulant-related hemorrhage without altering the incidence of TE. In fact, many infants and children have been managed with only aspirin with quite good long-term results.87 Although this is not recommended in patients older than infants, it is a feasible alternative. A recent study in patients under 50 years of age followed 254 patients for up to 20 years and found an exceedingly low rate of valve-related events (Table 32-4), an exceptional long-term overall survival of nearly 88%, and event-free survival probability of 92% at 19 years.88

The longest surviving patient in a series of bileaflet valve replacement patients is currently more than 30 years from his procedure, which was completed in his early forties and has been without complication.30 The low incidence of valve-related complications in younger patients should drive a discussion of the alternative prostheses available for such individuals, especially with new aggressive antiplatelet drugs. Operative mortality increases with each succeeding procedure; therefore, discussion of durability with patients becomes mandatory.28–30

Follow-Up

With the use of any valve prosthesis, long-term follow-up is a vital factor. Ten years is certainly not long enough to ascertain the true durability of a prosthesis. Grunkemeier reviewed several types of biologic prostheses and found that the 10-year durability was excellent, but between 12 and 18 years durability fell off and prosthetic replacement was necessary.64 These data were echoed by Khan and colleagues.23 Valve-related failure of all biologic alternatives, including stented prosthetic replacements, stentless prostheses, and homograft and autograft replacements have all shown long-term durability to be less than that of modern mechanical valve prostheses.

Even with some mechanical valve prostheses, durability after 10 years has not been adequate. In recommending mechanical valve replacement, one should be assured to have clinically available data on the prostheses that extend beyond 15 years.

In conclusion, with proper selection of patients for mechanical valve replacement, one can expect excellent long-term results, long-term survival, and a low incidence of valve-related complications. Current indications for recommending aortic valve replacement with mechanical prostheses are shown in Table 32-5.