Ann Thorac Surg 1998;66:101-107
© 1998 The Society of Thoracic Surgeons
Original articles: cardiovascular
Thromboembolism and mechanical heart valves: a randomized study revisited
Carl E.E. Kuntze, MDa,
Eugene H. Blackstone, MDa,
Tjark Ebels, MD, PhDa
a Department of Cardiothoracic Surgery, University Hospital Groningen, Groningen, the Netherlands
Accepted for publication February 18, 1998.
Address reprint requests to Dr Kuntze, Thorax Centre, University Hospital Groningen, Achterberghof 18, 9752 HE Haren, the Netherlands
 |
Abstract
|
|---|
Background. This study was designed to revise and substantiate previous inferences, based on short-term follow-up, about differences in the incidence of anticoagulant-related events after heart valve replacement among patients who had been randomly assigned to receive either a Björk-Shiley, Edwards-Duromedics, or Medtronic-Hall mechanical heart valve prosthesis.
Methods. Intermediate-term follow-up to January 1995 was completed in 418 of 419 patients randomized to receive one of three types of heart valve prostheses between January 1982 and January 1987. Median follow-up was 98.5 months. Multivariable analysis in the hazard function domain was performed to identify factors that influenced the incidence of time-related thromboembolism and bleeding. These findings were compared with those made previously after a median follow-up of 37.5 months.
Results. No differences were found among the three prostheses in rates of anticoagulant-related hemorrhage. However, the incidence of thromboembolism was higher after mitral valve replacement among patients who had received the Medtronic-Hall prosthesis (linearized rate, 5.4% per patient year; 70% confidence interval, 4.0% to 7.1%), compared with Edwards-Duromedics (1.3%; 70% confidence interval, 0.4% to 3.0%) and Björk-Shiley prostheses (1.2%; 70% confidence interval, 0.6% to 2.2%).
Conclusions. At long-term follow-up, in contrast to the findings at short-term follow-up, patients with either Björk-Shiley or Edwards-Duromedics prostheses had low rates of thromboembolism, whereas higher rates occurred in patients with a Medtronic-Hall prosthesis in the mitral position.
|
Introduction
|
|---|
In 1989, we reported a surprising difference in the risk for thromboembolism among three different heart valve prostheses that had been implanted during one of the few randomized comparative studies on mechanical heart valve designs to date. Thromboembolism occurred more frequently after implantation of Medtronic-Hall and Edwards-Duromedics mechanical heart valve prostheses than after implantation of Björk-Shiley prostheses [1, 2]. However, the duration of follow-up was short (median follow-up time, 37.5 months) and the number of events few (36). Other investigators did not report such differences. Thus, we revisited the patients (up to 12 years after valve replacement) to determine whether the originally reported differences persist.
|
Patients and methods
|
|---|
Between January 1982 and January 1987, 419 patients undergoing first-time heart valve replacement were consecutively and prospectively randomized to receive either the Björk-Shiley (Shiley Inc, Irvine, CA; n = 167), the Edwards-Duromedics (Hemex Scientific Inc, Austin, TX; n = 92), or the Medtronic-Hall (Medtronic Inc, Minneapolis, MN; n = 160) mechanical valvar prostheses at the Department of Cardiothoracic Surgery, University Hospital Groningen (see Study Design). From the 434 patients reported on initially, we excluded 15 who had re-replacements with valve designs other than the ones studied.
Details on study design and randomization, the surgical technique used for valve implantation, the methods used to obtain properly risk-adjusted estimates in the face of time-varying hazards, and our comments on the strengths and limitations of the study have been detailed previously, and will be repeated here only as necessary for the sake of clarity [1, 2].
Study design
The study was divided into two eras. During the first era, from January 1, 1982, to August 11, 1984, two mechanical valve designs were implanted: the Björk-Shiley Convex-Concave (60°) valve (later replaced by the Björk-Shiley Monostrut valve), and the Medtronic-Hall valve. During the second era, from August 11, 1984, to January 1, 1987, the Edwards-Duromedics bileaflet valve was added. Therefore, Edwards-Duromedics prostheses are represented in fewer cases and have shorter follow-ups than the other two valve designs.
The effectiveness of random allocation and a critique of its adequacy were reported previously [1]. No important differences in preoperative and operative variables were found among the three patient groups (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Baseline Characteristics of Patients Undergoing a Valve Replacement by One of the Randomized Prostheses (mean ± standard deviation)
|
|
Of the 419 patients, 254 patients underwent single aortic valve replacement (72 with concomitant coronary artery bypass grafting, CABG), 111 patients underwent single mitral valve replacement (27 with concomitant CABG), and 54 underwent simultaneous aortic and mitral valve replacement (6 with concomitant CABG).
Anticoagulation management
Acenocoumarol anticoagulation in combination with a platelet-inhibitor (dipyridamole, 100 mg four times daily) began 24 hours after operation and was continued for life. Follow-up and dosage adjustments were continued by a regional anticoagulation service at least once a month, after dismissal from the hospital. Anticoagulation was monitored by the prothrombin time using the Thrombotest reagent (Nycomed Amerham Pharma, Oslo, Norway [3]). The dose of acenocoumarol was adjusted to keep the Thrombotest values between 120 and 180 seconds (international normalized ratio, 3.2 to 4.8) [4]. If the value was outside this range, the dosage was changed and more frequent checks were performed.
Definition and identification of events after valve replacement
No attempt was made to separate the events into those occurring in-hospital (or within 30 days of operation) and those occurring later, because of the artificiality imposed by such a strategy. Instead, all events were dated to the conclusion of cardiopulmonary bypass, which for convenience was referred to as the time of operation. In this report, the events of interest were all causes of thromboembolism and anticoagulant-related hemorrhage. For the definition of events we used the guidelines according to Edmunds and colleagues [5]. The investigators identified and classified these adverse time-related events after repeated and detailed review of in-hospital and follow-up documents, and after personal telephone calls with patients, their families, and their physicians. The current formal cross-sectional follow-up ended in January 1995. The median follow-up interval was 98.5 months (8.2 years). All but 1 patient was reached; therefore, a 99.8% follow-up was achieved.
Data analysis
As in the previous analysis, time-related freedom from events followed valve replacement was estimated nonparametrically by means of Kaplan and Meier’s method and parametrically by using hazard function methodology. The latter identifies and characterizes multiple phases of time-varying risk [6–9]. Because thromboembolism is a recurrent morbid event and as with each recurrence the shape of the hazard function was found to be similar, it was analysed as a so-called modulated renewal process [10, 11].
Each time-related event was analyzed by multivariable analysis in the hazard function domain. Potential risk factors (variables) were organized for entry into the analyses, as shown in Appendix 1. Preparation for the multivariable analysis included correlation analysis, univariable association of risk to each variable, and decile risk analysis of ordinal and continuous variables to identify possible scale transformations. Variables were then entered simultaneously into each of the identified hazard phases. A directed technique of stepwise entry of variables into the multivariable risk factor model was then used [12]. The variables were retained in the final analysis if their p value was less than 0.1.
General explanation of the graphic depictions
The graphs depict the two different methods used to analyze the same data set. In the first method, circles, triangles, and squares identify the occurrence of individual events, and are positioned at the time of the event along the horizontal axis and according to the life table estimates of Kaplan and Meier along the vertical axis. The vertical bars enclose their 70% confidence intervals, equivalent to one standard error. The numbers in parentheses indicate the numbers of patients still at risk at the time of these estimates. In the second method, parametric estimates of the freedom from the event is indicated by the smooth solid lines, enclosed by dashed lines indicating their 70% confidence intervals. The table within the graph presents parametric estimates at the stated intervals.
|
Results
|
|---|
Thromboembolism
Frequency of occurrence
The only thromboembolic events reported were cerebrovascular. Forty-eight patients experienced 67 thromboembolic events; 9 patients experienced two events and 5 patients experienced three events. Twenty-one events occurred among the 167 patients who had received Björk-Shiley valves, 13 among the 92 patients who had received Edwards-Duromedics valves, and 33 among the 160 patients who had received Medtronic-Hall valves.
Categorization of events
Of the 67 thromboembolic events, 30 were categorized as transient ischemic attacks (45%; 70% confidence interval, 38% to 52%), 21 as reversible ischemic neurologic deficits (31%; 70% confidence interval, 25% to 38.2%), and 16 as completed strokes with hemiparesis or aphasia or both (24%; 70% confidence interval, 18% to 30%). Four patients died within 2 weeks after a completed stroke. In 1 patient with a Medtronic-Hall valve, four events took place within 3 days; We characterized this "shower" of events as a single thromboembolic episode.
Time relatedness
The overall (non–risk-adjusted) freedom from thromboembolism among the 419 patients was 86% at 12 years (Fig 1A). The instantaneous risk (hazard function) exhibited an early, slowly declining phase that gave way to a constant hazard phase after about 4 years (Fig 1B).
View larger version (24K):
[in this window]
[in a new window]
|
Fig 1. Repeated occurrence of thromboembolism. (A) Time-related freedom from thromboembolism according to the number of previous embolic episodes. The circles, squares, and triangles depict the Kaplan-Meyer estimates of freedom from first, second, and third episode of thromboembolism, and the solid lines and their dashed confidence limits depict parametric estimates (see Material and Methods). Note that time zero for second and third episodes was the time of the preceding thromboembolism. The numbers in parentheses are the number of patients at risk at various points in time along each curve. (B) Instantaneous risk (incidence, hazard function) for the first thromboembolic event.
|
|
Factors affecting prevalence
Patient and operative variables associated with an increased prevalence of all thromboembolic events and identified by multivariable analysis are shown in Table 2. Early risk was increased 2.3 times in patients with a double valve replacement versus isolated aortic or isolated mitral valve replacement and 2.5 times in patients with a Medtronic-Hall valve implanted in the mitral position as compared to a Björk-Shiley or Edwards-Duromedics valve. The constant-phase risk was increased 3.5 times in patients with a Medtronic-Hall valve in the mitral position, five times in patients whose valve had been replaced because of native valve endocarditis, and 12 times in patients experiencing recurrent thromboembolic events.
Association with valve type
Twelve years after a mitral valve replacement, only 71% of patients with a Medtronic-Hall valve showed no thromboembolic complications, as compared to 89% of patients with a Björk-Shiley or a Edwards-Duromedics valve (Figs 2, 3). In contrast, the differences in the rate of thromboembolism among the valve types in the aortic valve replacement group were no greater than could be due to chance (Fig 4).
View larger version (28K):
[in this window]
[in a new window]
|
Fig 2. First thromboembolism after isolated mitral valve replacement. (A) Time-related freedom from thromboembolism according to the type of mechanical valve prosthesis. (B) Hazard function for thromboembolism according to the type of mechanical valve prosthesis. (BS = Björk-Shiley (circles); ED = Edwards-Duromedics (squares); MH = Medtronic-Hall (triangles).)
|
|
View larger version (29K):
[in this window]
[in a new window]
|
Fig 3. Time-related freedom from all thromboembolic events after isolated mitral valve replacement according to the type of mechanical valve prosthesis. In contrast to Figure 2, all occurrences of thromboembolism, not just the first, are depicted in the format of a modulated renewal process (see Material and Methods). (BS = Björk-Shiley (circles); ED = Edwards-Duromedics (squares); MH = Medtronic-Hall (triangles).)
|
|
View larger version (28K):
[in this window]
[in a new window]
|
Fig 4. Time-related freedom from first thromboembolism after isolated aortic valve replacement according to the type of mechanical valve prosthesis. Note the expanded vertical axis. (BS = Björk-Shiley (circles); ED = Edwards-Duromedics (squares); MH = Medtronic-Hall (triangles).)
|
|
Hemorrhage
Severe bleeding occurred in 57 patients, 11 of whom died. The overall (non–risk-adjusted) freedom from hemorrhage was 81% at 12 years. Freedom from hemorrhage was similar for all valve types (Wilcoxon’s p > 0.6).
Comparison with previous report
Thromboembolism
In addition to our earlier report of 36 thromboembolic events, we now report another 31 episodes. Seventeen patients not previously experiencing an event had suffered one or more since the last follow-up. At 5 years an estimated 89% of patients did not suffer thromboembolism as compared with 88% in our previous report. However, the shape of the hazard functions was different despite the similarity in the freedom from event curve. With additional follow-up it exhibited a more gradual decline over time to a constant hazard level similar to that projected previously. To facilitate comparison with other reports, so-called linearized rates for thromboembolism are given in Tables 3 to 5 for the current long-term follow-up study.
Hemorrhage
The pronounced cluster of severe bleeding episodes occurring between 3 and 5 years resulted in a higher estimate of the constant hazard in our earlier report. In comparison, after the current long-term follow-up, the estimate was lower because this clustering was transient.
|
Comment
|
|---|
Previously described differences in the rate of thromboembolism after heart valve replacement observed during short-term follow-up have now proven to persist after a longer period of follow-up of this study. The additional follow-up allowed us to identify more specifically a group of patients at risk for thromboembolism and the temporal relationship of the events. The patients who had received a Medtronic-Hall valve placed in the mitral position were exposed to a substantially higher risk of thromboembolism; in addition, this increased thrombogenicity was more pronounced in the first 4 to 5 years after valve replacement. In contrast, the previous finding of increased thromboembolic risk among patients with Edwards-Duromedics valves was not confirmed by further follow-up. This was the smallest group of patients and it had the shortest follow-up in the previous study. (Other events, not shown, were not found to be different for the three types of valve then or after additional follow-up.)
As demonstrated already in our 1989 report, the risk of thromboembolism was not constant over time, but showed a temporal relationship between time after implantation and a thromboembolic event. Thus, accurate assessment and comparison of thromboembolic rates for various valve designs requires prolonged follow-up. On the basis of our two follow-up studies, we conclude that all patients should be followed-up for at least 6 to 7 years to capture a stable, accurate picture of thromboembolic events. Most reports on longitudinal one valve studies and even the few prospective randomized studies provide insufficient information to evaluate accurately the risk of thromboembolism over time [13–15]. They all express the incidence of valve complications as so-called linearized rates. Although this rate is easy to calculate, it is inappropriate in the absence of a constant hazard or incidence of events across time [11]. When linearized rates were calculated (Tables 3–5), they fell within the range other investigators have found [16–22].
Thus, after valve replacement, the incidence of thromboembolism is high initially, and decreases in 4 to 5 years to a constant rate of about 1% per patient-year or less, as observed by others [14, 20]. The factors associated with thromboembolism were found to differ in these two time frames. In the first 4 to 5 years after the valve replacement (early phase), the rate of thromboembolism was more than twice as high in patients with two prosthetic mechanical valves, each valve contributing to the risk of thromboembolism [23]. Later, in the constant hazard phase, the risk of thromboembolism was higher for patients who had undergone valve replacement for native valve endocarditis. An additional phenomenon is that, in this constant phase the risk of experiencing a recurrent event became much higher than the risk of a first event, as was also found by Mitchell and colleagues [24]. The risk of thromboembolism after mitral valve replacement with a Medtronic-Hall valve is substantially higher (both in the first 5 years after replacement and in the constant hazard phase thereafter), than is the case for the other two types of valve. However, the general risk of thromboembolism in the constant phase decreased to a range also found by others researchers [23, 25, 26].
The possible explanation for our finding evidence of increased risk of thromboembolism after Medtronic-Hall valve replacement and the reason why this has not been revealed in other studies, is that we used time-related multivariable methods to analyze our data. Other reasons for our findings concerning Medtronic-Hall valves in comparison to the other types of valve studied, is protracted follow-up and meticulous use of the guidelines for reporting on morbidity and mortality after cardiac valvar operations as discussed in our previous study [1]. The inferences made in our earlier report about risk of thromboembolism after Medtronic-Hall implantation are confirmed by our present long-term follow-up study.
We are aware that factors other than valve design and location of the prosthesis affect thromboembolic complications; the most prominent of these being the intensity of oral anticoagulation therapy. As has been discussed by Cannegieter and colleagues [3], many factors may influence anticoagulant level and effectiveness and these in turn influence the incidence of thromboembolism and bleeding complications. However, these factors are bound to randomization procedure. According to Butchart and colleagues [27], our anticoagulation range (international normalized ratio 3.2 to 4.5) should be too high for the Medtronic-Hall valve. However, on the basis of our data, one could speculate that it would be wise even to increase the anticoagulation level for this valve type. Other valve-related events could conceal possible thromboembolic events. Although Butchart [28] has questioned the matter of sudden death in an editorial, we did not include sudden, unexplained deaths as possible thromboembolic events. This decision seems appropriate in light of a recent study by Rooney and associates [29] in which 70% of all sudden deaths were followed by autopsy and 90% of these autopsied sudden deaths proved not to be related to the valve. In our study sudden deaths and indeed all deaths were evenly distributed among the valve designs.
Many researchers discuss the possible contribution of atrial fibrillation to the risk of thromboembolism after valve replacement. However, we, like others, have not found evidence for this contribution possibly because all patients received life-long anticoagulation [20, 24, 30].
In summary, we have found double valve replacement, endocarditis, and mitral valve replacement with the Medtronic-Hall valve to be incremental risk factors for thromboembolism. The risk for thromboembolism after mitral valve replacement with a Medtronic-Hall valve was substantially higher in the first 4 to 5 years after operation than for either Björk-Shiley or Edwards-Duromedics prostheses. The risk then decreased to a more acceptable level, albeit still higher than for the other two types of valve.
Already discouraged by the high incidence of thromboembolism found by us 8 years ago, we abandoned the use of the Medtronic-Hall valve device at our institution in August 1989.
|
Acknowledgments
|
|---|
We thank Mr Robert N. Brown for his work with the graphics. We also thank Hans D. de Boer, MD, and Heleen E. Wiersma, MD, for their efforts in following up the patients.
|
Appendix 1. Variables entered into the multivariable analysis
|
|---|
- Demographic variables: age at operation, sex, and body surface area.
- Clinical variables: interval from acceptance to operation, remote native valve endocarditis, active native valve endocarditis, preoperative degree of coronary artery disease.
- Variables describing the symptoms related to effects of the valve disease: New York Heart Association functional class immediately before operation, preoperative symptoms of angina, dyspnea, orthopnea, syncope.
- Variables concerning preoperative rhythm disturbances: preoperative (paroxysmal) atrial fibrillation, preoperative ventricular arrhythmia
- Variables concerning the preoperative hemodynamic status: preoperative state of decompensation, need for a emergency operation.
- Variables concerning the preoperative laboratory values: preoperative hemoglobin, preoperative creatinine, preoperative lactate dehydrogenase.
- Variables concerning the catheterization and angiographic data: aortic systolic pressure, aortic diastolic pressure, left ventricular systolic pressure, left ventricular diastolic pressure, right ventricular systolic pressure, right ventricular diastolic pressure, right atrial pressure, pulmonary artery wedge pressure, venous saturation, arterial saturation, O2 consumption (L · min-1 · m-2), O2 availability (L · min-1 · m-2), cardiac index, pulmonary vascular resistance, aortic valve gradient, heart rate, aorta valve area, mitral valve area, presence of left ventricular aneurysm on angiography, presence of left ventricular hypertrophy on angiography, presence of a normal left ventricular function on angiography.
- Etiologic variables: congenital etiology, degenerative etiology, infective etiology, ischemic etiology, rheumatic etiology.
- Valve disease: pure aortic incompetence, pure aortic stenosis, combined aortic stenosis and incompetence, pure mitral incompetence, pure mitral stenosis, combined mitral stenosis and incompetence.
- Surgical variables: date of operation, surgeons, valve replacement device, size of replacement device, concomitant procedures, duration of cardiopulmonary bypass, duration of the aortic cross-clamping, Myocardial ischemic time, cardioplegia administration time, ventricular fibrillation time, circulatory arrest time.
- Variables related to the modulated renewal for thromboembolism: number of previous thromboembolic events.
|
References
|
|---|
- Kuntze C.E.E., Ebels T., Eygelaar A., Homan van der Heide J.N. Rates of thromboembolism with three different mechanical heart valve prostheses: randomized study. Lancet 1989;I:514-517.
- Kuntze C.E.E., Ebels T., Eijgelaar A., Homan van der Heide J.N. Long-term evaluation of mechanical heart valves. Lancet 1989;I:1020-1021.
- Owren P.A. Thrombotest. A new method for controlling anticoagulant therapy. Lancet 1959;2:754-758.
- Cannegieter S.C., Rosendaal F.R., Wintzen A.R., et al. Optimal oral anticoagulant therapy in patients with mechanical heart valves. N Engl J Med 1995;333:11-17.[Abstract/Free Full Text]
- Edmunds L.H., Jr, Clark R.E., Cohn L.H., Grunkemeier G.L., Miller D.C., Weisel R.D. Guidelines for reporting morbidity and mortality after cardiac valvular operations. Ann Thorac Surg 1996;62:932-935.[Abstract/Free Full Text]
- Kaplan E.L., Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-481.
- Blackstone E.H., Naftel D.C., Turner M.E. The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc 1986;81:615-624.
- Blackstone E.H. Methods and limitations of follow up assessment (clinical studies of time-related events). In: Bodnar E., Frater R.W.M., eds. Replacement cardiac valves. New York: Pergamon Press, 1991:391-434.
- Blackstone E.H., Kirklin J.W. Death and other time related events after valve replacement. Circulation 1985;72:753-767.[Abstract/Free Full Text]
- Kalbfleisch J.D., Prentice R.L. The statistical analysis of failure time data. New York: John Wiley, 1980.
- Blackstone E.H. Analysis of thrombosis, thromboembolism and bleeding after heart valve replacement as time-related outcome events. In: Butchart E.G., Bodnar E., eds. Current issues in heart valve disease, thrombosis embolism and bleeding. London: ICR Publishers, 1992.
- Baskerville J.C., Toogood J.H. Guided regression modeling for prediction and exploration of structure with many explanatory variables. Technometrics 1982;24:9-17.
- Bloomfield P., Kitchin A.H., Wheatley D.J., Walbaum P.R., Lutz W., Miller H.C. A prospective evaluation of the Björk-Shiley, Hancock, and Carpentier-Edwards heart valve prostheses. Circulation 1986;73:1213-1222.[Abstract/Free Full Text]
- Antunes M.J. Clinical performance of St. Jude and Medtronic-Hall prostheses: a randomized comparative study. Ann Thorac Surg 1990;50:743-747.[Abstract]
- Vogt S., Hoffmann J., Roth J., et al. Heart valve replacement with the Björk-Shiley and St. Jude Medical prostheses: a randomized comparison in 178 patients. Eur Heart J 1990;11:583-591.[Abstract/Free Full Text]
- Beaudet R.L., Poirier N.L., Doyle D., Nakhl G., Gauvin C. The Medtronic-Hall cardiac valve: 7
years’ experience. Ann Thorac Surg 1986;42:644-650.[Abstract]
- Akins C.W. Long-term results with the Medtronic-Hall valvular prosthesis. Ann Thorac Surg 1996;61:806-813.[Abstract/Free Full Text]
- Jamieson W.R.E., Pelletier L.C., Pomar J.L. Clinical performance of the Duromedics bileaflet pyrolite mechanical prosthesis. Can J Surg 1992;35:605-611.[Medline]
- Klepetko W., Moritz A., Khünl-Brady G., et al. Implantation of the Duromedics bileaflet cardiac valve prosthesis in 400 patients. Ann Thorac Surg 1987;44:303-309.[Abstract]
- Daenen W., Van Kerrebroeck C., Stalpaert G., et al. The Björk-Shiley Monostrut valve. J Thorac Cardiovasc Surg 1993;106:918-927.[Abstract]
- Nakano S., Kawashima Y., Matsuda H., et al. A five-year appraisal and hemodynamic evaluation of the Björk-Shiley Monostrut valve. J Thorac Cardiovasc Surg 1991;101:881-887.[Abstract]
- Alvarez L., Escudero C., Figuera D., Castillo-Olivares L. The Björk-Shiley valve prosthesis, analysis of long-term evolution. J Thorac Cardiovasc Surg 1992;104:1249-1258.[Abstract]
- Cannegieter S.C., Rosendaal F.R., Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994;86:635-641.
- Mitchell R.S., Miller D.C., Stinson E.B., et al. Significant patient-related determinants of prosthetic valve performance. J Thorac Cardiovasc Surg 1986;91:807-817.[Abstract]
- Akins C.W. Results with mechanical cardiac valvular prostheses. Ann Thorac Surg 1995;60:1836-1844.[Abstract/Free Full Text]
- Aris A., Igual A., Padro J.M., et al. The Spanish Monostrut study group: a ten-year experience with 8,599 implants. Ann Thorac Surg 1996;62:40-47.[Abstract/Free Full Text]
- Butchart E.G., Lewis P.A., Bethel J.A., Breckenridge I.M. Adjusting anticoagulation to prosthesis thrombogenicity and patient risk factors: recommendations for the Medtronic-Hall valve. Circulation 1991;84(suppl 3):61-69.
- Butchart E.G. The significance of sudden death and unwitnessed death after heart valve replacement. J Heart Valve Dis 1994;3:1-4.[Medline]
- Rooney S.J., Moreno de la Santa P., Lewis P.A., Butchart E.G. Sudden death in a large prosthetic valve series based on a single prosthesis: experience with the Medtronic-Hall valve. J Heart Valve Dis 1994;3:5-9.[Medline]
- Antunes M.J., Wessels A., Sadowski R.G., et al. Medtronic Hall valve replacement in a third-world population group. J Thorac Cardiovasc Surg 1988;95:980-993.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
H. Rasche
Haemostasis and thrombosis: an overview
Eur. Heart J. Suppl.,
December 1, 2001;
3(suppl_Q):
Q3 - Q7.
[Abstract]
[PDF]
|
|
|
|
|
|
|
E. Muller, R. Bergemann, and GELIA Study Group
Economic analysis of bleeding and thromboembolic sequelae after heart valve replacement (GELIA 7)
Eur. Heart J. Suppl.,
December 1, 2001;
3(suppl_Q):
Q65 - Q69.
[Abstract]
[PDF]
|
|
|
|
|
|
|
R. Bergemann and E. Muller
Economic outcomes after heart valve replacement surgery in Germany
Eur. Heart J. Suppl.,
December 1, 2001;
3(suppl_Q):
Q70 - Q72.
[Abstract]
[PDF]
|
|
|
Top
Abstract
Introduction
