Bioprosthetic valves are increasingly implanted, with generally consistent and durable results. Early bioprosthetic valve failure is uncommon, and most clinicians are unfamiliar with the spectrum of early structural complications involving bioprostheses. In this review, the authors organize causes of early bioprosthetic valve failure according to possible pathogenesis, demonstrate the correlation between echocardiographic and anatomic findings, and discuss potential treatments. First, they address early bioprosthetic valve stenosis secondary to thrombosis. Next, they discuss excessive pannus formation, a hitherto rarely described cause of early bioprosthetic valve failure. Finally, the authors address early structural valve deterioration mediated by calcification or primary tears. Illustrative examples with relevant echocardiographic and operative findings are provided.
Highlights
- •
Fibrin deposition on the nonflow surface of a valve can cause stenosis.
- •
Subvalvular pannus can cause stenosis, and pannus can cause regurgitation by retracting leaflets.
- •
Pannus may be difficult to appreciate on echocardiography unless extensive or combined with thrombus.
- •
Primary perforations or tears can result in eccentric regurgitant jets.
- •
Endocarditis must be excluded in cases of thrombosis or leaflet tear.
Valve replacement is frequently performed in the United States, with > 90,000 surgical procedures annually. Bioprosthetic valves are increasingly implanted for many reasons, including improved durability, lack of requisite long-term anticoagulation, and the overall aging of the population. Furthermore, patients in their fifth decade appear to do well with bioprostheses, and this practice may be more common if the promise of transcatheter valve-in-valve procedures is realized. Traditionally, the decision to select a mechanical or a bioprosthetic valve has hinged on the risks of long-term anticoagulation versus the risks of repeat valve replacement. In clinical practice, most bioprosthetic valves begin to fail in older patients after 10 to 15 years. Hence, in older patients, bioprosthetic valves are generally recommended, though patient comorbidities and preferences are emphasized.
In older adults, bioprosthetic valves rarely fail within a few years of the initial surgery. Clinicians are often aware of valve dysfunction related to endocarditis, patient-prosthesis mismatch (PPM), and technical failure. These etiologies are reviewed in detail elsewhere. Less common pathophysiologies of early bioprosthetic valve failure include valve thrombosis, excessive pannus formation, and accelerated structural valve deterioration (SVD). These failure mechanisms have not been well characterized, and echocardiographic features that may aid in diagnosis are limited. Often, by recognizing echocardiographic findings within an appropriate clinical setting, an astute clinician can make the correct diagnosis. In a recent imaging vignette, we described echocardiographic characteristics in specific pathologies. With this review, we expand on these observations, focus on the frequent corroboration of echocardiographic with in situ findings, and provide mechanistic insights for these modes of failure. Finally, we address potential treatment strategies on the basis of clinical experience, as the published evidence is limited.
Case Selection and Literature Review
On the basis of surgical experience, representative examples were selected from our institution. Early bioprosthetic valve failure was defined as symptoms leading to repeat surgery on the previously operated valve within 5 years of the initial surgery. A MEDLINE search was conducted with the following keywords and their variants: bioprosthetic valve , thrombosis , pannus , and structural valve deterioration . Results were restricted to English studies in humans between 1980 and December 31, 2014. All relevant articles and their references were reviewed.
Early Thrombosis of Bioprosthetic Valves
Possible Pathogenesis of Early Bioprosthetic Valve Thrombosis
The discussion of thrombotic events in patients with bioprosthetic valves typically focuses on systemic embolization, not thrombosis of the valve itself. In general, rates of thromboembolism with bioprosthetic valves are similar to anticoagulated patients with mechanical valves. Early after bioprosthetic valve replacement, thromboembolic complications occur more frequently, likely related to a sewing ring and suture material that are not yet fully covered with biofilm and endothelialized. Controversy regarding the duration and magnitude of this increased risk, as well as the extent to which the risk is modified with combined aspirin and warfarin treatment, has led to different conclusions regarding anticoagulation management.
In part, consensus has been limited because thromboembolic events in these patients are relatively uncommon, with an approximate risk of 1% in the first 3 months. Actual thrombosis of a bioprosthetic valve is extremely rare. Thrombosis has been described for both mitral and aortic bioprostheses as a late finding and as an early finding within the first few years of surgery. In a large series of 4,568 patients receiving bioprosthetic aortic valves, eight patients required reoperation for valve thrombosis at a median time of 1.1 years from the index surgery. All of these patients had porcine valves from one manufacturer, which led the authors to speculate that an issue related to design, namely, the presence of a rail where the leaflet attaches to the stent, could lead to stasis of blood and thrombus. However, objective evidence of this phenomenon was limited, and thrombosis of stented bovine pericardial valves has rarely been described as well.
Clinical Presentation, Echocardiographic Features, and Surgical Findings of Early Bioprosthetic Valve Thrombosis
To illustrate this rare clinical entity, we present three examples of early thrombosis of bioprosthetic aortic valves. The first patient was a 78-year-old man with a history of severe symptomatic calcific aortic stenosis (AS), obstructive coronary artery disease, and paroxysmal atrial fibrillation who underwent aortic valve replacement (AVR) with a Mosaic valve (Medtronic, Minneapolis, MN), coronary artery bypass grafting, and a maze procedure. Because he did not have recurrence of atrial fibrillation, his warfarin was stopped after 3 months, but he developed a stroke 1 year after his initial surgery. Subsequent echocardiography showed severe AS and a suggestion of echodensities on the aortic side of the valve ( Figures 1 A and 1B; Video 1 ; available at www.onlinejase.com ). On reoperation, the flow surface of the valve appeared normal ( Figure 1 C), and thrombus was present on the nonflow surface ( Figure 1 D). The thrombosed valve was removed and replaced with a 25-mm Trifecta bioprosthesis (St. Jude Medical, St. Paul, MN), and the patient was discharged on warfarin and aspirin.
The second patient was a 67-year-old man with calcific AS who received a 25-mm Biocor bioprosthesis (St. Jude Medical). He had postoperative atrial fibrillation and was discharged on warfarin as well as aspirin. Less than 1 year later, he developed marked dyspnea while riding his bicycle, and echocardiography again showed severe AS with a suggestion of echodensities on the nonflow surface of the valve ( Figure 2 A; Videos 2 and 3 ; available at www.onlinejase.com ). Reoperation showed organizing thrombus on the aortic side of the leaflets ( Figure 2 B), and the ventricular side of the valve appeared normal ( Figure 2 C). The valve was replaced with a 24-mm homograft. Even though blood cultures had been negative, culture of the valve grew Propionibacterium acnes , emphasizing the importance of valve culture and sequencing for bacterial deoxyribonucleic acid in the setting of bioprosthetic valve thrombosis.
The final patient was a 51-year-old woman with dyspnea who had severe aortic regurgitation (AR) and left ventricular outflow tract obstruction from a subvalvular membrane. The subaortic membrane was resected, and given patient preference, the aortic valve was replaced with a 21-mm Carpentier-Edwards bioprosthesis (Edwards Lifesciences, Irvine, CA). Because she had no risk factors for thrombosis, she was discharged on aspirin alone. She returned 3 years later with recurrent dyspnea on exertion, and echocardiography revealed severe AS with a large protruding mass ( Figure 3 A; Video 4 ; available at www.onlinejase.com ). Repeat surgery demonstrated a large thrombus extending from the ventricular to the aortic side of the valve ( Figures 3 B and 3C). Microbiologic assessment was unremarkable. The Carpentier-Edwards valve was removed, and a 24-mm homograft was placed.
Potential Treatment Strategies for Bioprosthetic Valve Dysfunction Related to Thrombosis
Given its infrequent occurrence, the optimal management of bioprosthetic valve thrombosis is unclear. In our patients, the decision to reoperate was based on concern regarding recurrent systemic embolization, progressive symptomatic AS despite anticoagulation, and a large mass, respectively. For patients who are stable with minimal symptoms, anticoagulation with surveillance echocardiography can be an effective treatment. Thrombolysis has also been described but may be best reserved for inoperable patients. However, early operative intervention is often considered for two additional reasons. First, the etiology of valvular obstruction is not always apparent preoperatively. Second, as illustrated, a thrombosed valve with superimposed endocarditis may be evident only when the explanted valve is thoroughly examined.
When a patient does undergo repeat surgery, the best operative procedure is also uncertain. Thrombectomy has been performed, but given the possibility of valve dysfunction as a cause of unexpected thrombosis, most surgeons elect to replace the thrombosed valve. As was done in two of our patients, in the case of aortic valve thrombosis, the threshold for root replacement may be lower as aortic homografts may have decreased thrombogenicity. However, thrombosis has been described in this setting as well. Moreover, a stentless valve or a stented valve from a different manufacturer is usually chosen, though the evidence to support one choice over another is anecdotal.
Early Thrombosis of Bioprosthetic Valves
Possible Pathogenesis of Early Bioprosthetic Valve Thrombosis
The discussion of thrombotic events in patients with bioprosthetic valves typically focuses on systemic embolization, not thrombosis of the valve itself. In general, rates of thromboembolism with bioprosthetic valves are similar to anticoagulated patients with mechanical valves. Early after bioprosthetic valve replacement, thromboembolic complications occur more frequently, likely related to a sewing ring and suture material that are not yet fully covered with biofilm and endothelialized. Controversy regarding the duration and magnitude of this increased risk, as well as the extent to which the risk is modified with combined aspirin and warfarin treatment, has led to different conclusions regarding anticoagulation management.
In part, consensus has been limited because thromboembolic events in these patients are relatively uncommon, with an approximate risk of 1% in the first 3 months. Actual thrombosis of a bioprosthetic valve is extremely rare. Thrombosis has been described for both mitral and aortic bioprostheses as a late finding and as an early finding within the first few years of surgery. In a large series of 4,568 patients receiving bioprosthetic aortic valves, eight patients required reoperation for valve thrombosis at a median time of 1.1 years from the index surgery. All of these patients had porcine valves from one manufacturer, which led the authors to speculate that an issue related to design, namely, the presence of a rail where the leaflet attaches to the stent, could lead to stasis of blood and thrombus. However, objective evidence of this phenomenon was limited, and thrombosis of stented bovine pericardial valves has rarely been described as well.
Clinical Presentation, Echocardiographic Features, and Surgical Findings of Early Bioprosthetic Valve Thrombosis
To illustrate this rare clinical entity, we present three examples of early thrombosis of bioprosthetic aortic valves. The first patient was a 78-year-old man with a history of severe symptomatic calcific aortic stenosis (AS), obstructive coronary artery disease, and paroxysmal atrial fibrillation who underwent aortic valve replacement (AVR) with a Mosaic valve (Medtronic, Minneapolis, MN), coronary artery bypass grafting, and a maze procedure. Because he did not have recurrence of atrial fibrillation, his warfarin was stopped after 3 months, but he developed a stroke 1 year after his initial surgery. Subsequent echocardiography showed severe AS and a suggestion of echodensities on the aortic side of the valve ( Figures 1 A and 1B; Video 1 ; available at www.onlinejase.com ). On reoperation, the flow surface of the valve appeared normal ( Figure 1 C), and thrombus was present on the nonflow surface ( Figure 1 D). The thrombosed valve was removed and replaced with a 25-mm Trifecta bioprosthesis (St. Jude Medical, St. Paul, MN), and the patient was discharged on warfarin and aspirin.
The second patient was a 67-year-old man with calcific AS who received a 25-mm Biocor bioprosthesis (St. Jude Medical). He had postoperative atrial fibrillation and was discharged on warfarin as well as aspirin. Less than 1 year later, he developed marked dyspnea while riding his bicycle, and echocardiography again showed severe AS with a suggestion of echodensities on the nonflow surface of the valve ( Figure 2 A; Videos 2 and 3 ; available at www.onlinejase.com ). Reoperation showed organizing thrombus on the aortic side of the leaflets ( Figure 2 B), and the ventricular side of the valve appeared normal ( Figure 2 C). The valve was replaced with a 24-mm homograft. Even though blood cultures had been negative, culture of the valve grew Propionibacterium acnes , emphasizing the importance of valve culture and sequencing for bacterial deoxyribonucleic acid in the setting of bioprosthetic valve thrombosis.
The final patient was a 51-year-old woman with dyspnea who had severe aortic regurgitation (AR) and left ventricular outflow tract obstruction from a subvalvular membrane. The subaortic membrane was resected, and given patient preference, the aortic valve was replaced with a 21-mm Carpentier-Edwards bioprosthesis (Edwards Lifesciences, Irvine, CA). Because she had no risk factors for thrombosis, she was discharged on aspirin alone. She returned 3 years later with recurrent dyspnea on exertion, and echocardiography revealed severe AS with a large protruding mass ( Figure 3 A; Video 4 ; available at www.onlinejase.com ). Repeat surgery demonstrated a large thrombus extending from the ventricular to the aortic side of the valve ( Figures 3 B and 3C). Microbiologic assessment was unremarkable. The Carpentier-Edwards valve was removed, and a 24-mm homograft was placed.
Potential Treatment Strategies for Bioprosthetic Valve Dysfunction Related to Thrombosis
Given its infrequent occurrence, the optimal management of bioprosthetic valve thrombosis is unclear. In our patients, the decision to reoperate was based on concern regarding recurrent systemic embolization, progressive symptomatic AS despite anticoagulation, and a large mass, respectively. For patients who are stable with minimal symptoms, anticoagulation with surveillance echocardiography can be an effective treatment. Thrombolysis has also been described but may be best reserved for inoperable patients. However, early operative intervention is often considered for two additional reasons. First, the etiology of valvular obstruction is not always apparent preoperatively. Second, as illustrated, a thrombosed valve with superimposed endocarditis may be evident only when the explanted valve is thoroughly examined.
When a patient does undergo repeat surgery, the best operative procedure is also uncertain. Thrombectomy has been performed, but given the possibility of valve dysfunction as a cause of unexpected thrombosis, most surgeons elect to replace the thrombosed valve. As was done in two of our patients, in the case of aortic valve thrombosis, the threshold for root replacement may be lower as aortic homografts may have decreased thrombogenicity. However, thrombosis has been described in this setting as well. Moreover, a stentless valve or a stented valve from a different manufacturer is usually chosen, though the evidence to support one choice over another is anecdotal.
Early Bioprosthetic Valve Failure Related to Pannus
Possible Pathogenesis of Early Bioprosthetic Valve Failure Related to Pannus
In response to surgical trauma or small amounts of thrombus, a host tissue response initiates pannus formation. This response is part of normal healing, but an exuberant response with increasing amounts of collagen can lead to prosthetic dysfunction. Late prosthetic dysfunction related to pannus overgrowth in mechanical valves has been well described. For bioprosthetic valves, pannus formation over stents is considered normal, and occasionally, pannus has been considered beneficial in the longevity of a bioprosthetic valve. Rarely, excessive pannus formation causes bioprosthetic valvular dysfunction. Two mechanisms include obstruction through subvalvular extension of pannus and regurgitation when pannus encroachment restricts leaflets. In an obstructed mechanical valve, early dysfunction is more likely related to thrombus, though excessive early pannus formation has been described. Here, we highlight findings of early bioprosthetic valve failure related to pannus formation and emphasize its heterogeneous presentation.
Clinical Presentation, Echocardiographic Features, and Surgical Findings with Early Bioprosthetic Valve Failure Related to Pannus
The first patient was a 27-year-old woman with a bicuspid aortic valve and severe symptomatic AR who received a 23-mm Medtronic Mosaic valve. She subsequently had a successful pregnancy but then noted shortness of breath <3 years after her initial surgery. Further evaluation revealed ST-segment depression with exercise and severe prosthetic AS without obvious echodensities to explain the valvular dysfunction ( Figures 4 A and 4B; Videos 5 and 6 ; available at www.onlinejase.com ). On reoperation, the valve leaflets appeared normal, but expansive pannus was seen, causing subvalvular obstruction ( Figures 4 C and 4D). Subvalvular pannus, even if extensive, may not be visible on echocardiography, in part because of shielding from the bioprosthesis. After removal, the patient underwent repeat valve replacement with a 21-mm On-X mechanical valve (On-X Life Technologies, Austin, TX), and an ascending aorta reduction aortoplasty was performed.
The next patient was a 57-year-old woman with severe symptomatic AR who, per patient preference, underwent AVR with a 21-mm 3F stentless valve (Medtronic). She was discharged with 3 months of aspirin and warfarin but returned 1 year later with recurrence of exertional dyspnea and chest pain. Echocardiography revealed severe prosthetic AS with leaflet thickening ( Figure 5 A; Video 7 ; available at www.onlinejase.com ), and surgery revealed a focal thrombus as well as extensive pannus ( Figures 5 B and 5C). Her bioprosthetic valve was subsequently replaced with a 21-mm CarboMedics mechanical valve (Sorin Group, Milan, Italy). In both of these patients, the bioprosthetic valve was smaller, and there may have been a component of PPM that contributed to their symptomatic stenosis and accelerated pannus formation.
The next patient was a 51-year-old woman who presented with shortness of breath and was found to have a rheumatic mitral valve with severe mitral regurgitation (MR). Per patient preference, she received a 27-mm Biocor valve and was discharged on aspirin and warfarin. She returned 4 months later with right facial numbness and left upper extremity weakness, with neuroimaging findings of several small infarcts. Repeat surgery confirmed a mass that was also seen on transesophageal echocardiography ( Figures 6 A and 6B; Videos 8 and 9 ; available at www.onlinejase.com ), and pathology demonstrated a noninfectious admixture of thrombus and pannus ( Figure 6 C). The patient underwent repeat mitral valve replacement with a 27-mm Magna bioprosthetic valve (Edwards Lifesciences). The latter two cases also highlight that pannus formation and thrombus are not mutually exclusive and may both be present. On echocardiography, the mixture of thrombus and pannus can present as abnormal echodensities, whereas predominant pannus may be echolucent.
The next patient was a 67-year-old man with severe symptomatic calcific AS who underwent AVR with a 25-mm Trifecta bioprosthesis. One year later, he developed dyspnea on exertion, and echocardiography demonstrated severe valvular AR ( Figure 7 A; Video 10 ; available at www.onlinejase.com ) with normal-appearing bioprosthetic leaflets ( Figure 7 B; Video 11 ; available at www.onlinejase.com ). During surgery, two cusps were retracted secondary to pannus ( Figure 7 C), and he underwent repeat valve replacement with a 23-mm On-X mechanical valve.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
