Author + information
- Received April 19, 2017
- Revision received June 27, 2017
- Accepted June 28, 2017
- Published online July 2, 2018.
- Alexander Egbe, MBBSa,
- Sorin V. Pislaru, MD, PhDa,∗ (, )
- Mahmoud A. Ali, MB, BCha,
- Arooj R. Khan, MBBSa,
- Amber N. Boler, MDa,
- Hartzell V. Schaff, MDb,
- Emmanuel Akintoye, MD, MPHc,
- Heidi M. Connolly, MDa,
- Vuyisile T. Nkomo, MDa and
- Patricia A. Pellikka, MDa
- aDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- bDepartment of Cardiovascular Surgery, Mayo Clinic, Rochester, Minnesota
- cDepartment of Internal Medicine, Wayne State University/Detroit Medical Center, Detroit, Michigan
- ↵∗Address for correspondence:
Dr. Sorin V. Pislaru, Department of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905.
Objectives The purpose of this study was to review the institutional practice of surveillance transthoracic echocardiography (TTE) for diagnosing early prosthetic valve dysfunction (PVD).
Background Bioprosthetic valve thrombosis (BPVT) is an important cause of PVD, and guidelines do not recommend routine TTE during the first 5 years after valve implantation.
Methods The authors performed a retrospective case-control study of all suspected (imaging diagnosis) or confirmed (histopathological diagnosis) cases of BPVT from January 1997 through December 2016. Patients were matched 1:2 (age, sex, prosthesis position) to patients whose prostheses were explanted because of structural failure (SF). PVD was defined as a 50% increase above baseline gradient at valve implantation and classified as early (≤5 years) or late (>5 years) after implantation.
Results There were 94 BPVT (51 suspected, 43 confirmed) and 188 SF cases; patient age 61 ± 9 years; men 61 (65%). The prosthesis positions were aortic 56%; mitral 26%; tricuspid 15%; and pulmonary 3%. Early PVD was more common in the BPVT versus SF group: 83 of 94 (88%) versus 20 of 188 (11%) (p < 0.001). Time from implantation to PVD was shorter for BPVT than SF: 26 months (interquartile range [IQR]: 12 to 43 months) versus 74 months (IQR: 48 to 102 months) (p < 0.001). At the initial PVD diagnosis, 81% of BPVT and 90% of SF patients were asymptomatic. However, BPVT patients had rapid symptomatic deterioration, requiring intervention sooner after PVD diagnosis: 6 months (IQR: 4 to 7 months) versus 51 months (IQR: 22 to 55 months) (p < 0.001).
Conclusions Most patients with PVD due to BPVT were asymptomatic at initial diagnosis, which was made based on routine surveillance TTE, often performed before 5 years. BPVT, an acute disease process, requires timely diagnosis because patient conditions rapidly deteriorate. Further studies are needed to determine whether routine surveillance TTE should be considered for patients with bioprosthetic valves to identify pre-symptomatic features of BPVT in order to provide effective, appropriate therapy.
Bioprosthetic valve thrombosis (BPVT) is increasingly being recognized as a cause of prosthetic valve dysfunction (PVD) (1–4) and accounts for 6% to 11% of cases of PVD in surgically implanted valves (1,5,6). Some studies suggest that warfarin is beneficial as a first-line therapy for suspected BPVT to avoid reoperations or reintervention in patients with BPVT (2,5–7) Because warfarin is a noninvasive and effective therapeutic option, the pre-operative or pre-intervention diagnosis of BPVT is important.
The diagnosis of BPVT is challenging because of a low level of awareness of this disease process and because of the absence of well-defined diagnostic criteria (8). A recent study from Mayo Clinic proposed a BPVT score based on the echocardiographic characteristic of BPVT (1). In a small, prospective study, the BPVT score reliably identified the patients who would respond to anticoagulation therapy (9). The current American and European valvular heart disease guidelines do not recommend routine transthoracic echocardiography (TTE) within the first 5 years after implantation of a bioprosthetic valve (10–12). However, BPVT, which is a potentially reversible cause of PVD, tends to occur within the first 5 years after valve implantation (1,5,7). Therefore, the purpose of this study was to review our institutional practice of surveillance TTE for the diagnosis of early PVD.
Patient selection and data collection
We retrospectively reviewed all cases of suspected or confirmed BPVT in adults (age ≥18 years) who had surgically implanted valves at Mayo Clinic, Rochester, Minnesota, from January 1, 1997, through December 31, 2016. Data from some of the patients included in the current study have been previously published (1,9). The Mayo Clinic Institutional Review Board approved the study and waived written informed consent for those who provided research authorization.
Suspected BPVT diagnosis was defined as a clinical diagnosis of BPVT using previously published echocardiographic criteria (50% increase in prosthesis gradient within 5 years after valve implantation, increased cusp thickness, and abnormal cusp motion) and a positive response to warfarin therapy defined as a 50% decrease in prosthesis gradient after initiation of warfarin (1,9). Confirmed BPVT diagnosis was defined as identification of fibrin-rich material on the valve cusps, either grossly or microscopically, during the pathological examination of an explanted prosthesis. Patients with the following conditions were excluded: active endocarditis; thrombophilia; mixed disease (thrombosis and degeneration), in whom the main mechanism could not be established; absence of baseline post-implantation gradient in the medical records; and incomplete follow-up, defined as no TTE for more than 24 months during follow-up.
All prostheses that were explanted for structural failure (SF) within the study period were identified from the pathology database. SF was defined as the presence of moderate or marked pannus formation—an obstructive fibrous ingrowth on either the inflow or outflow surface of the valve that was affecting the amount of cusp excursion at the time of gross examination. We selected a control group of prostheses explanted for SF by matching 1 case of BPVT to 2 cases of SF. The BPVT and SF groups were matched by age (±5 years), sex, and prosthesis position. We then compared the annual occurrence of clinically significant PVD between patients with BPVT and those with SF.
Clinical, echocardiographic, and surgical data were reviewed for all patients. The definitions of clinical and echocardiographic data and some of the patients included in this study have been previously described (1,9). We defined PVD as a 50% increase above the baseline post-implantation gradient identified during annual follow-up. PVD was classified as early if it occurred within 5 years or late if it occurred 5 years after implantation.
We studied the impact of PVD diagnosis on clinical decision making by assessing escalation of care after PVD diagnosis. Any diagnostic testing or clinical follow-up to assess prosthetic valve function that went beyond the valvular heart disease guideline recommendations for routine follow-up was considered escalation of care (10,11). We defined escalation of care as transesophageal echocardiography, cardiac computed tomography scan, hemodynamic cardiac catheterization, repeat TTE within 6 months, and scheduled outpatient cardiology visits within 6 months.
The baseline post-implantation gradient was obtained from the first outpatient TTE performed within the first 3 months after prosthesis implantation. For patients without an outpatient TTE within the first 3 months, we used the last inpatient TTE performed before hospital dismissal as the baseline echocardiogram. Using the baseline echocardiogram as time zero, all subsequent TTEs were reviewed to determine the initial occurrence of clinically significant prosthesis dysfunction (50% gradient increase from baseline post-implantation). For the analysis of annual change in gradient, only TTEs performed within 6 months of the yearly interval from valve implantation were included in the analysis. For instance, the mean prosthesis gradient based on TTE performed from 6 months to 18 months was analyzed as the annual follow-up gradient for a patient. All pertinent clinical data that could impact prosthetic valve hemodynamics, such as heart rate, heart rhythm, and high cardiac output state (anemia, hyperthyroidism, arteriovenous shunts, morbid obesity, cirrhosis, etc.) were collected and analyzed.
All statistical calculations were performed with the JMP version 11.0 software (SAS Institute, Cary, North Carolina). Categorical variables were reported as percentages, and continuous variables were reported as mean ± SD or mean (interquartile range [IQR]) for skewed data. Comparison of categorical variables was performed with the chi-square test or Fisher exact test; continuous variables were compared with a Student t test or Wilcoxon rank sum test, as appropriate. All p values were 2-sided, and p values <0.05 were considered significant.
Baseline clinical characteristics
There were 121 patients with a BPVT diagnosis identified from January 1, 1997, through December 31, 2016. Data from some of the patients included in the current study have been previously published (1,9). We excluded 25 patients for the following reasons: post-implantation gradient not in the medical records (n = 15), incomplete follow-up (n = 1), history of thrombophilia (n = 2), active endocarditis (n = 3), and diagnosis of mixed thrombosis and degeneration (n = 4). A total of 94 cases of BPVT were included in the BPVT group; of these, 51 were suspected BPVT, and 43 were confirmed BPVT. There were 33 BPVT cases in the early era (January 1, 1997, through December 31, 2006); the proportion of confirmed versus suspected BPVT was 82% versus 18%, respectively (p = 0.002). In the latter era (January 1, 2007, through December 31, 2016), there were 61 BPVT cases; the proportion of confirmed versus suspected BPVT cases was 26% versus 74%, respectively (p < 0.001). The institutional practice at Mayo Clinic is to perform surveillance TTE within the first 5 years after valve implantation, and the mean interval between TTE scans was 15 ± 3 months in this cohort.
The mean patient age at the time of valve implantation was 61 ± 9 years, and 61 (65%) patients were men. Table 1 compares the baseline characteristics of 94 patients with BPVT and 188 matched controls with SF. Paroxysmal atrial fibrillation, CHA2DS2-VASc score of at least 2, subtherapeutic International Normalized Ratio, and implantation of a porcine bioprosthesis were more common in the BPVT group. The use of post-operative antiplatelet and anticoagulation therapy was similar in both groups.
Baseline prosthetic valve hemodynamics
Table 2 compares the baseline echocardiographic data between the BPVT and SF groups; there was no significant difference between groups. The most common prosthesis position was aortic (56%), followed by mitral (26%), tricuspid (15%), and pulmonary (3%). Outpatient TTE performed within 3 months after implantation was available for 36 (38%) of the BPVT group and for 65 (35%) of the SF group, and the last TTE performed before hospital dismissal was used as the baseline TTE in 58 (62%) of the BPVT and 123 (65%) of the SF group. There was no significant difference in the mean gradients between the outpatient TTEs and the inpatient TTEs for the different prosthesis positions.
Prosthetic valve dysfunction
Early PVD (≤5 years from implantation) was more common in the BPVT group than in the SF group, 83 of 94 (88%) versus 20 of 188 (11%) (p < 0.001). The occurrence of PVD due to BPVT versus SF was 28% versus 0% (p < 0.001) in the first year; 25% versus 2% (p < 0.001) in the second year; and 16% versus 4% (p < 0.001) in the third year (Figure 1). Within the BPVT group, the yearly occurrence of significant PVD was similar in patients with suspected BPVT (n = 51) versus confirmed BPVT (n = 43): 27% versus 29% in the first year (p = 0.61); 25% versus 25% in the second year (p = 0.93); 16% versus 15% in the third year (p = 0.82); 13% versus 12% in the fourth year (p = 0.18); and 9% versus 8% in the fifth year (p = 0.75).
A separate analysis between the patients with confirmed BPVT (n = 43) versus the patients with SF (n = 188) showed a similar pattern (the groups were unmatched). Early PVD (≤5 years from implantation) was more common in the BPVT group than in the SF group, 34 of 43 (79%) versus 20 of 188 (11%) (p < 0.001). The occurrence of PVD due to confirmed BPVT versus SF was 23% versus 0% (p < 0.001) in the first year; 21% versus 2% (p < 0.001) in the second year; and 17% versus 4% (p < 0.001) in the third year.
Table 3 shows the TTE data at the time of reoperation or initiation of anticoagulation. The mean gradient and peak velocity for aortic prostheses was higher in the BPVT group, but otherwise there were no significant differences in the prosthetic valve hemodynamics between the groups. Despite similar valve hemodynamics at the time of implantation and at the time of reoperation/anticoagulation, the rate of progression of PVD was significantly different between the groups. The interval from prosthesis implantation to the diagnosis of significant PVD was shorter when the mechanism of dysfunction was due to thrombosis versus SF: 26 months (IQR: 12 to 43 months) versus 74 months (IQR: 48 to 102 months) (p < 0.001) (Figure 2). Similarly, the interval from the diagnosis of significant PVD to the time of intervention (reoperation or anticoagulation) was also shorter when the mechanism of dysfunction was due to thrombosis versus SF: 6 months (IQR: 4 to 7 months) versus 51 months (IQR: 22 to 55 months) (p < 0.001) (Figure 2).
Data about New York Heart Association (NYHA) functional class were available for 68 of 94 patients (72%) in the BPVT group. Among these 68 patients, 13 (19%) had NYHA functional class III/IV symptoms at the time of PVD diagnosis, and 61 (90%) had class III/IV at the time of reoperation/anticoagulation therapy. Data about NYHA functional class were also available for 113 of 188 patients (60%) in the SF group. Among these 113 patients, 11 (10%) had NYHA functional class III/IV symptoms at the time of PVD diagnosis, and 98 (87%) had class III/IV symptoms at the time of reoperation/anticoagulation therapy.
Escalation of care
Of the 282 patients enrolled in the study, the diagnosis of clinically significant PVD triggered 123 escalations of care in 111 patients (39%). Escalation of care included transesophageal echocardiography (n = 21), cardiac computed tomography scan (n = 2), hemodynamic cardiac catheterization (n = 1), repeat TTE within 6 months (n = 43), and outpatient visits within 6 months (n = 56). The patients with early PVD (clinically significant PVD within the first 5 years post-implantation) were more likely to have escalation of care than those with late PVD: 78 of 103 (76%) versus 33 of 179 (18%) (p < 0.001). Of the 78 patients whose care was escalated for early PVD, 49 underwent intervention within 6 months (anticoagulation therapy in 28 and reoperation in 21 patients).
Incidence of BPVT
There were 8,308 bioprosthetic valves implanted within the study period, of which 3,128 (38%) had at least 5 years of echocardiographic follow-up. On the basis of the 94 cases of BPVT identified among the 3,128 patients with follow-up, the incidence of BPVT was 3% (95% confidence interval: 2.6% to 3.5%). BPVT incidence was similar for the different prosthesis positions: aortic 53 of 1,868 (2.8%), mitral 24 of 766 (3.1%), tricuspid 14 of 387 (3.6%), and pulmonary 3 of 107 (2.8%); p = 0.86.
Moderate-to-severe valvular heart disease affects more than 10% of people older than 65 years of age, and many of these patients undergo implantation of a bioprosthetic valve (13). Recent studies suggest that BPVT is a cause of early PVD, and it usually presents with an early increase in prosthetic gradient (1–4,7). The current study analyzed the annual occurrence of PVD within the first 5 years of valve implantation and the potential role of TTE for identifying these cases.
Of 282 patients enrolled in this study, early PVD occurred in 103 (37%) patients, and BPVT was the cause of dysfunction in 81% of the cases. Although there was no difference in the baseline prosthetic valve hemodynamics between the BPVT and SF groups, the interval between valve implantation and the development of PVD was significantly shorter in the BPVT group (26 months) than in the SF group (74 months).
In a retrospective study that included 749 patients who received a porcine bioprosthesis in the aortic position, BPVT occurred in 17 patients (2%) during a mean follow-up of 3.4 years (7). The interval from valve implantation to BPVT diagnosis was 13 months. None of the patients who received pericardial aortic valve prostheses developed BPVT. In that study, the diagnosis of BPVT was based on echocardiography and confirmed by a positive response to anticoagulation in 15 patients and by histologic findings in 2 patients who had reoperations (7). In another study, Butnaru et al. (5) reported a BPVT diagnosis in 9 of 149 consecutive patients who underwent implantation of a bioprosthetic valve in the mitral position. The method of diagnosis was by transesophageal echocardiography in 7 patients and by histological findings for 2 patients. Similarly, the interval between valve implantation and BPVT diagnosis was 12 months.
BPVT is not unique to surgically implanted valves, and it can also occur after transcatheter aortic valve replacement (TAVR) (14–16). Del Trigo et al. (14) reviewed a multicenter registry of 1,521 patients who had TAVR. Prosthesis dysfunction, defined as an increase in baseline mean gradient of at least 10 mm Hg, was present in 68 patients (4.5%), and the interval from the time of implantation to the diagnosis of valve thrombosis was 20 months. Other studies that used contrast-enhanced multidetector cardiac computed tomography scan for diagnosis of valve thrombosis have reported the incidence of BPVT as high as 7% within the first 12 months after TAVR (16).
The results of our study are similar to those of the studies described above, which suggest that BPVT is an important cause of PVD, both for surgically implanted and transcatheter valves. In all of these studies, the highest incidence of BPVT occurred within the first few years after valve implantation.
Rapid symptomatic deterioration
Another important observation from the current study was the rapid symptomatic deterioration in the patients with BPVT. Although prosthetic valve hemodynamics were similar in both groups at the time of reintervention, the interval from the initial diagnosis of PVD to the time of reintervention was significantly shorter in the BPVT group (6 vs. 51 months, respectively). Among the patients in the BPVT group who had a documented NYHA functional class, 90% had NYHA functional class III/IV symptoms at the time of reintervention, even though only 19% were symptomatic at the time of initial PVD diagnosis a few months prior. A possible explanation for rapid symptomatic deterioration is that BPVT is a short-term event and does not allow adequate time for cardiovascular adaptation. By contrast, the rate of symptomatic deterioration was much slower in the patients with PVD due to SF because SF is a slow and chronic disease process.
Screening and surveillance
The current valvular heart disease guidelines recommend baseline TTE after the implantation of a bioprosthetic valve (10–12). However, routine TTE surveillance is not recommended within the first 5 years in the absence of symptoms or suspicion for clinically significant PVD (10–12). These recommendations were based on the assumption that BPVT is uncommon and that clinically significant PVD usually occurs more than 5 years after valve implantation (17,18). Recent single-center studies have reported a BPVT incidence of 6% in the mitral position and 2% in the aortic position for surgically implanted valves (5,7). For transcatheter valves, BPVT incidence is as high as 4% to 7%. However, the diagnostic criteria and modalities differ between the studies of surgical versus transcatheter valves (14–16).
The current study does not provide data about BPVT incidence. However, it shows that BPVT was responsible for 81% of all PVD within the first 5 years after valve implantation. More importantly, patients with BPVT were mostly asymptomatic at the time of initial PVD diagnosis, but they developed rapid symptomatic deterioration within a few months. If we strictly applied the current recommendations of no surveillance echocardiogram, most of these BPVT patients would not have undergone echocardiography because they were asymptomatic at the time of echocardiographic diagnosis. These diagnoses were made because of our institutional practice of performing surveillance TTE scans within the first 5 years after valve implantation. It was difficult to determine whether the clinician would have suspected substantial PVD on the basis of the physical examination and electrocardiogram because of the retrospective design of the current study. Moreover, most of the patients had their echocardiograms by the time they were examined by their cardiologists, and this may have introduced some bias.
Making recommendations on the basis of results from a single-center, retrospective study is also difficult. Nevertheless, the current study complements an emerging body of published reports about the clinical importance of BPVT, the need for heightened awareness, and the meticulous screening required for diagnosis (1–4,6,7,9,14,16). BPVT is a potentially reversible cause of PVD, and timely diagnosis and initiation of anticoagulation therapy could prevent reoperation. In the absence of validated clinical risk factors for BPVT, routine surveillance TTE may be the only practical strategy to make this diagnosis.
First, the study is prone to potential bias inherent in the single-center, retrospective study design. In addition, it was challenging to accurately determine the clinical suspicion at the time the echocardiograms were performed and even more challenging to accurately delineate the timeline for symptomatic deterioration.
BPVT was responsible for 81% of early PVD, and most of the patients in this study developed severe symptomatic obstruction over a relatively short period of time. Most of these patients were asymptomatic at the time of initial diagnosis, and the diagnosis of BPVT was made only because they had a routine surveillance echocardiogram. Recommending routine surveillance TTE within 5 years of implantation for all bioprosthetic valves will result in a large health care cost, and the data from the current study, with its limitations, are not adequate to support this recommendation. However, given the increased frequency of surgical and transcatheter bioprosthetic valve replacement in our aging population, the entity of BPVT will become increasingly important. Further studies are required to understand the best strategy to prevent, screen for, and treat this reversible disease process.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: BPVT was responsible for 81% of cases of prosthetic valve dysfunction occurring within 5 years after valve implantation. Routine echocardiography was required for diagnosis because most patients were asymptomatic. Rapid symptomatic deterioration from the time of diagnosis was common.
TRANSLATIONAL OUTLOOK: Further studies are needed to determine whether routine surveillance transthoracic echocardiography should be considered for patients with bioprosthetic valves to identify pre-symptomatic features of BPVT in order to provide effective, appropriate therapy.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bioprosthetic valve thrombosis
- interquartile range
- New York Heart Association
- prosthetic valve dysfunction
- structural failure
- transcatheter aortic valve replacement
- transthoracic echocardiography
- Received April 19, 2017.
- Revision received June 27, 2017.
- Accepted June 28, 2017.
- 2018 American College of Cardiology Foundation
- Egbe A.C.,
- Pislaru S.V.,
- Pellikka P.A.,
- et al.
- Egbe A.C.,
- Connolly H.M.,
- Pellikka P.A.,
- et al.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.,
- American College of Cardiology/American Heart Association Task Force on Practice Guidelines
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.,
- ESC Committee for Practice Guidelines (CPG); Joint Task Force on the Management of Valvular Heart Disease of the European Society of Cardiology (ESC); European Association for Cardio-Thoracic Surgery (EACTS)
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Del Trigo M.,
- Munoz-Garcia A.J.,
- Wijeysundera H.C.,
- et al.
- Hansson N.C.,
- Grove E.L.,
- Andersen H.R.,
- et al.