Author + information
- Received August 29, 2016
- Revision received November 7, 2016
- Accepted November 17, 2016
- Published online February 20, 2017.
- Alexander C. Egbe, MD, MPHa,∗ (, )
- Heidi M. Connolly, MDa,
- Patricia A. Pellikka, MDa,
- Hartzell V. Schaff, MDb,
- Richard Hanna, MDa,
- Joseph J. Maleszewski, MDa,c,
- Vuyisile T. Nkomo, MD, MPHa and
- Sorin V. Pislaru, MD, PhDa
- aDepartment of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
- bDepartment of Cardiovascular Surgery, Mayo Clinic, Rochester, Minnesota
- cDepartment of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota
- ↵∗Address for correspondence:
Dr. Alexander C. Egbe, Mayo Clinic, Department of Cardiovascular Diseases, 200 First Street SW, Rochester, Minnesota 55905.
Objectives The aim of this study was to assess the efficacy of warfarin in the treatment of bioprosthetic valve thrombosis (BPVT) of surgically implanted valves.
Background There are limited data about treatment outcomes for BPVT.
Methods This was a prospective study of patients with suspected BPVT of surgically implanted valves who received warfarin therapy at the Mayo Clinic from January 2013 to January 2016. BPVT score was calculated using previously described echocardiographic criteria. One point was assigned for each criterion: a 50% increase in prosthetic gradient within 5 years of implantation, increased cusp thickness, and abnormal cusp motion.
Results Fifty-two patients were enrolled in the study (mean age 61 ± 18 years, 34 men [66%]). The mean follow-up duration from presumed BPVT was 86 ± 24 weeks. Prosthesis positions were aortic in 31 patients (60%), mitral in 17 (32%), pulmonary in 2 (4%), and tricuspid in 2 (4%). Positive responses (defined as a 50% reduction in prosthesis gradient) occurred in 43 patients (83%) within 11 weeks (interquartile range [IQR]: 6 to 22 weeks) of anticoagulation with warfarin. Nine patients (17%) did not respond to warfarin, and these patients underwent surgical valve replacement (n = 5), transcatheter valve replacement (n = 1), and intervention (n = 3). BPVT scores were calculated for 48 patients (92%) with good-quality echocardiographic images; 9 had BPVT scores of 2, and 39 had BPVTs score of 3. A BPVT score of 3 predicted a positive response to anticoagulation therapy with sensitivity of 88% and specificity of 93%.
Conclusions A trial of anticoagulation was effective in 83% of patients with suspected BPVT, and most patients responded within 3 months. BPVT score was predictive of response to therapy and should be considered during patient selection.
The predominant mechanism of bioprosthetic valve dysfunction is structural failure (1,2), but bioprosthetic valve thrombosis (BPVT) is increasingly being recognized as an important and potentially reversible cause (3,4). BPVT is present in 11% of bioprosthetic valves explanted because of prosthesis dysfunction (4). The treatment strategies for suspected BPVT are reoperation, thrombolytic therapy, and anticoagulation. Data from several small case series, including 1 from our group, suggest that warfarin may be beneficial as a first-line therapy for suspected BPVT (3,5,6).
A recent study from the Mayo Clinic described the echocardiographic characteristics of BPVT, which are a 50% increase in prosthesis gradient within 5 years after implantation, increased cusp thickness, and presence of abnormal cusp motion (4). A BPVT risk score calculated using these 3 criteria was shown to increase the diagnostic sensitivity and specificity of BPVT.
There are no prospective studies evaluating the role of anticoagulation therapy in the treatment of BPVT. The purpose of this study was to describe the outcomes of warfarin therapy for the management of BPVT of surgically implanted valves on the basis of a prospective case registry.
Patient selection and study design
After our initial report (3), a strategy of warfarin anticoagulation for suspected BPVT was recommended to the cardiologists and cardiac surgeons at the Mayo Clinic. We maintained prospectively a registry of patients who had tentative diagnoses of BPVT of surgically implanted valves between January 2013 and January 2016. Clinical, echocardiographic, and surgical data were reviewed, and pathology data for explanted prostheses were examined. Risk factors for thromboembolic complications were reviewed, and the CHA2DS2-VASc score was calculated for each patient.
The criteria for inclusion in this study were suspected diagnosis of BPVT, trial of anticoagulation therapy with warfarin, and follow-up transthoracic or transesophageal echocardiography performed at least 4 weeks after the initiation of anticoagulation therapy. A suspected BPVT diagnosis was on the basis of the impression of the attending cardiologist using echocardiographic features such as cusp thickness, abnormal cusp mobility, increased transvalvular gradient, and presence of intracardiac thrombus. The decision to initiate anticoagulation was left at the discretion of the primary cardiologist and the patient. The Mayo Clinic Institutional Review Board approved the study protocol.
The primary endpoint was a positive response to anticoagulation therapy, defined as a 50% reduction in prosthesis gradient after the initiation of warfarin. This is the same endpoint used in our prior retrospective series of outcome of anticoagulation for suspected BPVT (3). The secondary endpoint was to assess the performance of previously described echocardiographic criteria for BPVT diagnosis (4) in predicting response to anticoagulation therapy.
All patients received warfarin with or without heparin bridging at the time of suspected BPVT diagnosis. Warfarin dose was adjusted to maintain an international normalized ratio (INR) of 2.0 to 3.0, and patients were monitored for bleeding complications. All patients were followed until the endpoint of response to anticoagulation therapy, valve replacement, or the end of study period.
Major bleeding events were defined as intracranial bleeding, pericardial or pleural hematoma requiring drainage, or any bleeding requiring transfusion. Minor bleeding events were defined as cutaneous bleeding, epistaxis, gastrointestinal bleeding, or any bleeding event that did not meet the criteria for major bleeding events. These definitions are derived from previous studies (7,8).
Digitally stored images of transthoracic echocardiograms (n = 52) and transesophageal echocardiograms (n = 38) were systematically analyzed by one author (A.C.E.) for the presence of echocardiographic characteristics of BPVT (4): a 50% increase in prosthesis gradient within 5 years after implantation, increased cusp thickness (defined as thickness >2 mm or significantly increased thickness compared with the initial post-operative echocardiogram), and presence of abnormal cusp motion. BPVT risk score was calculated on the basis of the presence of these 3 criteria, and 1 point was assigned for each criterion. Thus, a prosthesis with all 3 echocardiographic criteria had a BPVT score of 3.
Two echocardiographers (P.A.P. and H.M.C.) with experience in valvular disease and BPVT reviewed randomly selected studies in one-half of the cohort. All reviewers were blinded to the interpretations of others and the outcomes of anticoagulation therapy.
Surgically resected specimens were reviewed by a cardiovascular pathologist (J.J.M.), and the gross and histological features were documented. Thrombus formation was documented if fibrin-rich material was identified either grossly or microscopically present on the valve cusps. Pannus formation was noted when obstructive fibrous ingrowth was present on either the inflow or outflow surface of the valve and was noted as mild if it did not affect the movement of the bioprosthetic cusps at the time of gross examination. Moderate or marked pannus formation was documented if the fibrous ingrowth affected the amount of cusp excursion at the time of gross examination.
All statistical analysis was performed using JMP version 10.0 (SAS Institute, Cary, North Carolina). Categorical variables are expressed as number (percentage) and continuous variables as mean ± SD or median (interquartile range [IQR]) for skewed data. Categorical variables were compared using the chi-square test or Fisher exact test; continuous variables were compared using a 2-sided unpaired Student t test or Wilcoxon rank sum test, as appropriate. A receiver-operating characteristic curve was used to assess the performance of BPVT score for predicting response to anticoagulation therapy. All p values were 2-sided, and p values <0.05 were considered to indicate statistical significance.
Patient and prosthesis characteristics
There were 55 patients who received a trial of anticoagulation for suspected BPVT within the study; 3 patients were excluded because they did not undergo follow-up echocardiography after the initiation of anticoagulation therapy. The final cohort comprised 52 patients (mean age 62 ± 18 years, 34 [66%] men) (Table 1).
The most common prosthesis position was aortic, in 31 patients (60%), followed by mitral in 17 (32%), pulmonary in 2 (4%), and tricuspid in 2 (4%). Porcine valves accounted for 56% of all prostheses, and the most common mechanism of prosthesis dysfunction was stenosis, in 46 patients (89%) (Table 1). The median time from implantation to prosthesis dysfunction was 34 months (IQR: 13 to 80 months), and 38 of 52 episodes of BPVT (73%) occurred more than 1 year after valve implantation (Figure 1).
Almost all prostheses (51 of 52 [98%]) were implanted at the Mayo Clinic, and 42 of 51 of these patients (83%) received at least 3 months of anticoagulation with warfarin immediately post-operatively. All 9 patients not treated with early post-operative anticoagulation received aortic valve prostheses and were treated with only aspirin.
Risk factors for thromboembolism
Among the 52 patients in the study, 18 (35%) had CHA2DS2-VASc scores ≥2. Eight patients (15%) had histories of paroxysmal atrial fibrillation, 3 of whom were receiving warfarin at the time of BPVT diagnosis. Of the 5 patients with paroxysmal atrial fibrillation who were not receiving anticoagulation, 4 had CHA2DS2-VASc scores of 1, and 1 had a CHA2DS2-VASc score of 2. Three patients (6%) had histories of thrombotic or embolic events (deep venous thrombosis, n = 1; pulmonary embolism, n = 1; and ischemic stroke, n = 1).
At the time of BPVT diagnosis, 3 of 52 patients (6%) were on warfarin; INR assay at the time of diagnosis showed INRs <2.0 in 2 patients and an INR of 2.0 to 3.0 in 1 patient. Forty-nine of 52 patients (94%) were receiving antiplatelet therapy at the time of BPVT diagnosis (aspirin, n = 47; clopidogrel, n = 16; aspirin plus clopidogrel, n = 14). The reason for not being on antiplatelet therapy was a history of gastrointestinal bleeding in 1 patient and noncompliance in 2 patients.
Warfarin was initiated at the time of diagnosis, and warfarin dose was adjusted to maintain an INR of 2.0 to 3.0. One patient with mitral BPVT received unfractionated heparin for 4 days pending therapeutic INR. All other patients had their warfarin doses titrated until the target INR was achieved. Once the target INR was achieved, each patient had at least 1 INR assay per month; the median number of INR assays was 9 (IQR: 6 to 14) per patient. An INR <2.0 was documented at least once in 14 patients (27%), but the proportion of patients with subtherapeutic INRs was similar in responders versus nonresponders.
Positive responses to anticoagulation therapy were achieved in 43 of 52 patients (83%), and the median duration of trial of anticoagulation therapy was 11 weeks (IQR: 6 to 22 weeks). For the entire cohort, the mean aortic prosthesis gradient decreased from 46 ± 5 to 21 ± 5 mm Hg, the mean mitral prosthesis gradient decreased from 14 ± 4 to 5 ± 2 mm Hg, the mean pulmonary prosthesis gradient decreased from 39 to 26 mm Hg, and the mean tricuspid prosthesis gradient decreased from 13 to 8 mm Hg. Figure 2 shows transesophageal echocardiograms (2-dimensional, color Doppler, and continuous-wave Doppler images) before and after warfarin therapy for mitral BPVT.
One patient with a pulmonary prosthesis initially showed reduction of prosthetic gradient from 42 to 21 mm Hg in response to anticoagulation, but his gradient increased again to 39 mm Hg after anticoagulation was discontinued. This patient was restarted on warfarin, and his gradient decreased to 23 mm Hg, and his anticoagulation was continued indefinitely. All other responders remained on anticoagulation until the end of the study period.
Two patients had bleeding complications (gastrointestinal bleeding, n = 1; and epistaxis, n = 1), and warfarin was discontinued in both patients. There were no thromboembolic complications.
Nine patients did not respond to anticoagulation therapy (aortic, n = 5; mitral, n = 3; and tricuspid, n = 1). The time interval from prosthesis implantation to BPVT diagnosis was significantly longer in the nonresponders compared with the responders (median 96 months vs. 31 months, p = 0.01). There was no difference in the baseline characteristics and mechanism of prosthetic dysfunction between responders and nonresponders.
Among the 9 nonresponders, 5 patients underwent surgical valve replacement (aortic, n = 3; mitral, n = 1; and tricuspid, n = 1), and 1 patient underwent transcatheter mitral valve replacement. The other 3 nonresponders did not undergo valve replacement by the end of the study period. One of the 3 patients remained on warfarin until the end of the study period, and the other 2 patients had their anticoagulation discontinued because of bleeding (n = 1) and after a third echocardiographic examination showed minimal change in prosthesis gradient (n = 1). Examination of the 5 explanted bioprostheses showed pannus formation in 4 and calcification of the cusps in the remaining device.
Predictor of response to anticoagulation therapy
The diagnosis of BPVT was suspected on transthoracic echocardiography in all 52 patients; 38 patients (73%) also underwent transesophageal echocardiography.
Forty-eight patients (92%) had echocardiographic images of sufficient quality to assess for echocardiographic characteristics of BPVT, 9 of whom had BPVT scores of 2, and 39 patients had BPVT scores of 3 (Table 2). Two echocardiographers (P.A.P. and H.M.C.) reviewed randomly selected studies in one-half of the cohort, and the BPVT scores calculated by both echocardiographers were concordant.
A BPVT score of 3 reliably predicted a positive response to anticoagulation therapy with sensitivity of 88%, specificity of 93%, positive predictive value of 91%, and negative predictive value of 85%. The patients with BPVT scores of 3 had greater reductions in prosthesis gradient after anticoagulation compared with those with BPVT scores of 2 (p < 0.01) (Figure 1).
To the best of our knowledge, this is the first systematic prospective study of BPVT of surgically implanted valves. Our main findings are as follows: 1) a previously proposed echocardiographic diagnostic score predicted response to warfarin anticoagulation in BPVT with excellent accuracy; 2) warfarin significantly reduced prosthetic gradients in 83% of the patients; and 3) this diagnostic and therapeutic strategy was associated with minimal side effects and no mortality.
We based this study on the results of our retrospective case series. We were able to show that improvement in prosthetic gradients and New York Heart Association functional class occurs in the majority of patients (87%) treated with warfarin for BPVT (3). Our findings were similar to those of 2 other retrospective case series that reported thrombus resolution in 6 of 9 patients (67%) and 7 of 10 patients (70%) treated with warfarin from suspected BPVT (5,6).
After our initial observation, we decided to prospectively maintain a registry of patients with suspected BPVT of surgically implanted valves at our institution, which forms the basis of the present report. The diagnosis of BPVT was considered whenever a significant increase in gradient (>50% over baseline) was observed. The results of the present study are concordant with our retrospective series and confirm the efficacy of anticoagulation therapy in patients with suspected BPVT on the basis of prospective data. The important clinical implication is that we were able to avoid early reoperation in the majority of patients, at minimal risk (2 minor bleeding events); there were no strokes or deaths.
The occurrence of BPVT is not restricted to surgically implanted bioprosthetic valves but has also been observed after transcatheter aortic valve replacement (TAVR). A recent study reported reduced leaflet motion in 22 of 55 patients receiving the Portico transcatheter valve (40%) and in 17 of 132 (13%) in 2 TAVR registries (9). Restoration of leaflet motion was noted in all patients receiving warfarin, suggesting that the bioprosthetic dysfunction was due to BPVT. In addition, the study suggested that abnormal cusp motion was not associated with an increase in prosthetic gradients, raising the question of subclinical BPVT in TAVR patients. Del Trigo et al. (10) reported that deterioration in hemodynamic function of transcatheter valves (>10 mm Hg increase in mean gradient) was observed in 4.5% of patients during a mean follow-up of 20 months. In a multivariate analysis, failure to anticoagulate patients at hospital discharge was associated with a 3-fold increase in the risk for valve dysfunction, lending additional support to the importance of anticoagulation for BPVT. Although the present study was on the basis of a cohort of patients with surgically implanted valves, our result is concordant with emerging data of BPVT in the TAVR population (9–11). This study provides some preliminary data supporting the safety and efficacy of warfarin in the treatment suspected BPVT.
Duration of therapy
The median time from the initiation of anticoagulation therapy to subsequent echocardiography was 11 weeks in the present study. The duration of anticoagulation therapy reported in previous studies ranged from 7 to 10 weeks (3,5,6). The time interval between the initiation of warfarin and subsequent echocardiography in this study was longer compared with the initial case series from our group, 11 weeks versus 7 weeks, respectively (3). This may be because the patients in the present study were less symptomatic, with only 12% having New York Heart Association functional class III or IV symptoms compared with 38% in the initial series. One possible explanation for the less symptomatic patients in the present study may be a greater level of awareness and early recognition of BPVT because of emerging data about this diagnosis. We must emphasize that the response to warfarin may occur earlier than reported. Indeed, the timing of follow-up was left at the discretion of the primary physician; an earlier response cannot be ascertained from our registry data.
The optimal trial duration of anticoagulation therapy for BPVT is unknown, but the present study suggests that response to anticoagulation therapy usually occurs within the first 3 months. Interestingly, none of the 3 nonresponders who continued warfarin showed significant reduction in gradient after additional 4 to 6 weeks of anticoagulation. This further supports our speculation that gradients decrease in responders within the first few weeks of anticoagulation, whereas nonresponders will not have a gradient reduction regardless of the duration of anticoagulation.
The next unresolved issue is how long anticoagulation should be continued after an episode of BPVT. All the responders in this cohort remained on anticoagulation until the end of the study period. There is no consensus on the duration of therapy for these patients, but we favor anticoagulation for the life span of the prosthesis in patients without an excess risk for bleeding.
Our initial study reported paroxysmal atrial fibrillation and inadequate anticoagulation as risk factors for BPVT (4). However, only 8 patients (15%) in the present study had atrial fibrillation, suggesting the existence of other unidentified risk factors for BPVT. It will be difficult to make a case for discontinuation of anticoagulation if we are unable to identify and correct reversible risk factors for BPVT. In contrast, lifelong anticoagulation is undeniably associated with increased risk for bleeding (8,12). An important factor influencing the decision to implant a bioprosthetic valve is the desire to avoid anticoagulation; hence recommending lifelong anticoagulation for a bioprosthetic valve defeats this purpose. Shared decision making between the patient and the cardiologist is crucial in planning duration of therapy because of the paucity of data.
We have previously proposed a set of echocardiographic criteria for the diagnosis of BPVT (4). The present study was a prospective validation of this set of diagnostic criteria, albeit in a small cohort. In the absence of pathology data to confirm BPVT diagnosis, the best surrogate for confirmation of BPVT diagnosis is thrombus resolution (improved leaflet mobility, normalization of leaflet thickness and of prosthetic gradients) with anticoagulant therapy. This study shows that the presence of all 3 criteria (BPVT score of 3) predicted response to anticoagulation (surrogate pathology data) with 88% sensitivity and 93% specificity.
BPVT is often unrecognized because of the absence of validated diagnostic criteria and a lack of awareness among medical providers (3–5). In our previous reports, we identified a total of 65 BPVT cases spanning a 16-year interval. After sharing our observations with the cardiologists and cardiac surgeons at our institution, the apparent incidence of BPVT increased significantly, with 43 confirmed cases identified within a 3-year period, as reported in this study. We believe our findings reflect the improved recognition of this disease.
Even when BPVT is recognized, there is a general reluctance to initiate a trial of anticoagulation, especially in symptomatic patients, because of the paucity of data supporting its efficacy for treatment of BPVT. The present study adds to the growing body of research highlighting the efficacy and safety of anticoagulation therapy in patients with suspected BPVT (3,5,6,9,13,14). We went a step further to provide a tool for predicting response to therapy.
An important observation from this study was that the nonresponders had a longer interval from implantation to prosthetic dysfunction compared with the responders (96 months vs. 31 months). Pathology data showed the presence of pannus in 4 of 5 explants and calcification in 1 explant. These observations emphasize the inability of imaging tools to provide a precise tissue diagnosis. There are recent studies showing that cardiac computed tomography is very sensitive for detection of BPVT in the TAVR population (9,11,15). Whether a systematic strategy of additional imaging with transesophageal echocardiography or cardiac computed tomography would have avoided warfarin anticoagulation in these patients with suspected BPVT of surgically implanted valves remains unknown and should be explored in future studies. The present study suggests that incorporating BPVT score in the assessment of these patients will improve patient selection of anticoagulation therapy and hopefully avoid anticoagulation in patients with low BPVT scores who most likely may not have BPVT. However, none of these patients experienced complications from the trial of anticoagulation. It is important to maintain close clinical monitoring during trials of anticoagulation, especially for those with severe prosthetic valve stenosis.
The cellular biology underlying the pathophysiology of BPVT is not well understood, but subclinical prosthesis thrombus may be a precursor of pannus formation and accelerated structural failure. Further studies are required in this area.
There are several important unresolved issues in the prevention and management of BPVT, and these include the following: 1) there is a need for a better clinical risk stratification tool to identify patients at risk for BPVT; 2) the optimal duration of therapy for the patients who showed positive response to trial of anticoagulation remains unknown; 3) the role of novel anticoagulant agents in the prevention and treatment of BPVT is uncertain.
Prevention of BPVT
The American College of Cardiology and American Heart Association guidelines recommend warfarin for the first 3 months after implantation of a bioprosthetic valve, followed by low-dose aspirin for the life span of the prosthesis (Class IIa) (16). The European Society of Cardiology guidelines, in contrast, recommend warfarin for the first 3 months after the implantation of a bioprosthetic valve in the mitral or tricuspid position and low-dose aspirin for bioprosthetic valves in the aortic positions for the first 3 months (Class IIa) (17). The patients in this cohort received appropriate guideline-directed prophylactic therapy.
Current guidelines recommend routine imaging after bioprosthetic valve implantation after 5 years (European Society of Cardiology) or 10 years (American College of Cardiology/American Heart Association). Patients in the present study were minimally symptomatic, and identification of BPVT would not have been possible without our current strategy of annual imaging after bioprosthetic valve replacement. In a previous report we showed that 11.6% of reoperations for bioprosthetic valve dysfunction are due to BPVT, with the median time to BPVT of 2 years (4). We propose that current guidelines be revised to include yearly echocardiographic evaluation of bioprosthetic valve for the first 3 years post-implantation.
The limitation of this study is that it was a single-center study with a relatively small sample size and short-term follow-up. However, this is the first prospective study on a strategy to diagnose and treat BPVT of surgically implanted valves. Because this study was on the basis of a cohort of patients with surgically implanted valves, the performance of the proposed BPVT score in the transcatheter valve population is unknown.
We did not have a standardized schedule for follow-up echocardiography after the initiation of anticoagulation, making it difficult to determine the optimal duration for a trial of anticoagulation. Also, the screening for thromboembolic risk factors such as paroxysmal atrial fibrillation was not standardized in this study, and fewer than one-half of the patients had Holter monitoring from the time of implantation to the time of BPVT diagnosis. Finally, the role of the novel oral anticoagulant agents was not addressed in this study.
This prospective study shows that 83% of patients with suspected BPVT diagnosis responded to anticoagulation and, as a result, avoided early reoperation. A response to a trial of anticoagulation usually occurs within 3 months, and a positive response can be predicted with reasonable sensitivity and specificity using our proposed BPVT score.
WHAT IS KNOWN? BPVT occurs in surgical and transcatheter valves, and the diagnosis can be made on the basis of a set of echocardiographic criteria (BPVT score).
WHAT IS NEW? Warfarin is safe and effective in the treatment of BPVT, and the response to therapy can be predicted using the BPVT score.
WHAT IS NEXT? Further studies are required to validate the performance of this tool in the transcatheter valve population and the role of novel oral anticoagulant agents in the treatment of BPVT.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bioprosthetic valve thrombosis
- international normalized ratio
- interquartile range
- transcatheter aortic valve replacement
- Received August 29, 2016.
- Revision received November 7, 2016.
- Accepted November 17, 2016.
- American College of Cardiology Foundation
- Valfre C.,
- Ius P.,
- Minniti G.,
- et al.
- Pislaru S.V.,
- Hussain I.,
- Pellikka P.A.,
- et al.
- Egbe A.C.,
- Pislaru S.V.,
- Pellikka P.A.,
- et al.
- Schulman S.,
- Kearon C.,
- for the Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis
- Del Trigo M.,
- Munoz-Garcia A.J.,
- Wijeysundera H.C.,
- et al.
- Hansson N.C.,
- Grove E.L.,
- Andersen H.R.,
- et al.
- Fitzmaurice D.A.,
- Blann A.D.,
- Lip G.Y.
- Latib A.,
- Naganuma T.,
- Abdel-Wahab M.,
- et al.
- De Marchena E.,
- Mesa J.,
- Pomenti S.,
- et al.
- Pache G.,
- Schoechlin S.,
- Blanke P.,
- et al.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.