| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2002;123:21-32
© 2002 The American Association for Thoracic Surgery
Surgery for Acquired Cardiovascular Disease |
Sponsor: Antonio Maria Calafiore, MD
From the Bristol Heart Institute,a University of Bristol, Bristol, United Kingdom, and Welsh Heart Research Institute,b University of Wales College of Medicine, Cardiff, United Kingdom.
St Jude Medical and Sulzer-CarboMedics provided financial support for a clerical assistant for the duration of this study.
Received for publication June 14, 2001. Accepted for publication July 16, 2001. Address for reprints: Alan J. Bryan, MD, FRCS, Bristol Heart Institute, Bristol Royal Infirmary, Bristol BS2 8HW, United Kingdom (E-mail: alan.bryan{at}ubht.swest.nhs.uk).
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
1-5 In a collective review, Akins 6 concluded that rates for valve-related complications were essentially highly comparable between CM and SJM valves, although the data for the CM valve from a smaller study population 7,8 did show a slightly higher rate of thromboembolic complications in the mitral position, despite adequate anticoagulation. This finding has been supported by others. 9-11 In contrast, Jamieson and colleagues 12,13 found no significant differences in thromboembolic tendency between the CM and SJM prostheses in isolated mitral valve replacement (MVR) or double valve replacement (DVR). Long-term randomized prospective studies to evaluate the experience with these prostheses have been lacking to date. Although there is general agreement that randomized prospective trials are not essential in the evaluation of prostheses, the conflicting and occasionally worrisome outcomes reported in other observational studies do make randomized prospective studies in a particular patient population, with the prevailing level of anticoagulation, an important contribution to the assessment of the clinical performance of particular valve prostheses. This report describes the midterm results of a prospective randomized study comparing the clinical outcomes of patients who received either CM or SJM mechanical heart valves implanted at a single institution.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
Surgery
All operations were performed through a median sternotomy with cardiopulmonary bypass and mild systemic hypothermia (28°C-32°C). Myocardial protection consisted of intermittent antegrade cold (6°C) St Thomas crystalloid or blood cardioplegia. The prostheses used in both the CM and SJM groups were of the standard design. Interrupted or continuous suturing technique was used at the discretion of the operating surgeon. All patients received postoperative subcutaneous heparin until the international normalized ratio (INR) was above 2 with warfarin administration. On discharge, anticoagulation was managed in the community by general medical practitioners and at anticoagulation clinics at local hospitals according to the British Society for Haematology guidelines, 14 as published in the British National Formulary, which essentially recommend a target INR range of 3 to 4.5.
Follow-up
Annual follow-up was performed primarily by patient questionnaire on the anniversary of the operation and by direct contact as necessary for the clarification of details. The family practitioners or hospital cardiologists were contacted, and the hospital health records were used as appropriate. The death registry of the General Register Office for UK National Statistics was used to provide details of deaths that were otherwise unobtainable. When adverse events were reported, these were categorized by an independent clinician in association with the study coordinator (J.W.), both of whom were blinded to the type of valve implanted. Anticoagulation data, drug dosages, and INR values were obtained through an anticoagulation history booklet, which is carried by the patient and filled in by the responsible physician during each monitoring episode. For the purpose of this interim report, data are presented for follow-up until the end of September 1999.
Statistical analysis and data reporting
The "Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Operations" were strictly followed for definitions of valve-related complications, analysis, and presentation of data and results. 15 The gathered information was recorded onto a standard form and stored in a computerized database (Microsoft Access for Windows; Microsoft Corporation, Redmond, Wash). Subsequent data analysis was performed in consultation with a qualified statistician. Continuous data are represented as mean ± SD. Intergroup comparisons were performed with the unpaired t test or nonparametric Mann-Whitney test as appropriate. Categorical data were analyzed by crosstabulation with the 
2 test or the Fisher exact test. The actuarial method was used to estimate event-free survival with 95% confidence limits, and Mantel-Haenszel log-rank test was used to compare event-free survival. Early deaths within 30 days were included in reported 5-year survivals. Calculated linearized rates (percentage per patient-year) during the entire follow-up period include early and late events. Statistical analyses were performed with the aid of StatView version 5 for Windows (SAS Institute, Inc, Cary, NC).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
|
|
|
|
|
|
|
|
Bleeding events
A total of 35 bleeding events were reported for the CM group during the study period, compared with 43 such events in the SJM group. Bleeding-associated deaths occurred in 3 patients from the SJM group and 1 from the CM group. The 5-year event-free rates for bleeding were 87.3% ± 2.5% (3.6% per patient-year) and 82.6% ± 2.8% (4.3% per patient-year) for the CM and SJM groups, respectively (log-rank P = .25,Figure 4).
|
andTable 5). The DVR group had a higher rate of hemorrhagic events, although not statistically significantly different, with a 5-year event-free rate of 68.6% ± 10.4% (CM 76.7% ± 13.3%, SJM 60.6% ± 15.7%, log-rank P = .58) compared with 86.9% ± 2.2% for AVR (CM 89.4% ± 2.9%, SJM 84.6% ± 3.3%, log-rank P = .24) and compared with 84.1% ± 3.4% for MVR (CM 85.4% ± 4.7%, SJM 83.1% ± 4.9%, log-rank P = .84). Patients received anticoagulation with an INR between 3 and 4.5 in 49% of total hemorrhagic events. The INR value was greater than 4.5 in 30% of cases and less than 3 in 5%. There was no difference in anticoagulation pattern during hemorrhagic events either according to valve position or between CM and SJM groups. All patients received warfarin during follow-up, and 6.2% of patients (CM n = 2, SJM n = 2) were also receiving antiplatelet therapy during hemorrhagic events. This compares with 7.7% patients overall who were receiving both therapies(Figure 5).
|
.
Nonstructural dysfunction and other valve-related complications
Echocardiographic evidence of paraprosthetic leak was recorded in 12 patients (CM n = 7, SJM n = 5, P = .57), with no significant difference observed after stratification according to valve position or operating surgeon. In 3 cases (CM n = 2, SJM n = 1), reoperation was indicated. One patient of the SJM group had an aorta–right ventricular fistula develop during AVR, and this had to be repaired in a second operation. There were no cases of structural valve deterioration detected in this series during this follow-up period. The 5 year event-free rates for nonstructural valve dysfunction were 96.1% ± 1.4% (0.7% per patient-year) and 96.0% ± 1.3% (0.6% per patient-year) for the CM and SJM groups, respectively (log-rank P = .57).
Nyha functional class
Symptomatic improvement after valve replacement can be expressed by the changes observed in NYHA class during follow-up throughout the 5 years, as shown inFigure 6 and Table 6. Cross-table analyses did not show any significant differences between the CM and SJM groups at any time during follow-up. The proportion of patients in functional classes III and IV decreased from 56% before the operation to 11% 1 year afterward (P < .001), but this had increased to 19% by year 3 (year 1 vs year 3, P < .001). Thereafter from year 3 to year 5 there was no significant change in the proportion of patients in NYHA functional classes III and IV.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
6 Many studies have investigated the standard-design CM and SJM bileaflet mechanical prosthetic valves independently and have shown them to perform well, with excellent durability and acceptable rates of valve-related complications. Recent comparative nonrandomized studies have shown no significant real differences between the prostheses described. 12,13 However, some have reported that the CM valve performed less well than the SJM prosthesis, specifically in relation to thrombosis and embolism in mitral implantations. 9,10 It was recognized, however, that it might be difficult to draw firm conclusions from comparisons made in uncontrolled nonrandomized studies. To our knowledge this is the first report of a randomized study to compare the clinical outcome with the use of the CM and SJM standard design bileaflet valve prostheses.
In this study the simple random assignment method used achieved an excellent degree of uniformity in the distribution of major preoperative and perioperative variables. However, it does have limitations. There was a baseline imbalance, with more patients in the CM group with infective endocarditis and congenital valve pathologic conditions(Table 3). In the SJM group overall (all valve positions), more patients underwent associated tricuspid valve repair. In hindsight, we agree that although perfectly balanced randomization is difficult to achieve in reality, such baseline imbalances could have been avoided through the use of a more sophisticated method of randomization that included strategic stratification, block randomization, or minimization. 16,17
In addressing the issue of variation in reported rates of complications, this study recorded data according to the definitions agreed on in the internationally published guidelines. All event rates quoted include early (<30 days) postoperative complications, with operation designated as time zero. The actual rates of valve-related events in this study were comparable in most respects with those reported by other centers.
With respect to isolated AVR, the linearized rates (percentage per patient-year) of thromboembolism in our series at midterm were 1.5% for the CM valve and 1.3% for the SJM valve, with published ranges of 0.9% to 1.9% 6,18 for the CM valve and 0.6% to 2.4% 19,20 for the SJM. For bleeding events in the AVR group, our rates were 3.2% for CM and 4.1% for SJM, with other reported rates ranging from 0.7% to 2.3% 6,18 for CM and 0.1% to 3.5% 21,22 for SJM. Similarly, our rate for thromboembolism among patients who underwent MVR was 3.2% per patient-year (CM 3.3%, SJM 3.1%), comparable with other reports, which have ranged from 0.4% to 5.0%. 6,12 The incidence of hemorrhagic events in our MVR group was 3.9% per patient-year (CM 3.7%, SJM 4.0%); previous reports have ranged from 1.5% to 6.4%. 6,12
It is generally believed that thromboembolic and hemorrhagic events are the same entity at opposite ends of the spectrum in which anticoagulation control is a major determinant of outcome. Anticoagulation management is usually community based in the United Kingdom, and most follow the British National Formulary (BNF) published guidelines as recommended by the British Society for Haematology. The British National Formulary specifies a general target INR range of 3 to 4.5 for mechanical heart valve prostheses, which may be higher than necessary and may predispose these patients toward more bleeding events, at least according to some reports. 23,24 Despite a higher target, the actual recorded mean INR during follow-up was 3.0, as shown inFigure 5. Of the group who bled, however, 30% had an INR in excess of 4.5 during the acute event, as compared with 5% overall during follow-up (673 of 13,612 readings) in the study population.
Despite the lack of major differences in bleeding or thromboembolic indices between CM and SJM valves and the fact that no mitral valve thrombosis was detected up to the time of this report, we recognize the need to apply some caution in interpretation of such data and to remember the interim nature of this report. With further follow-up across the projected 10-year period, and as the patient-years accumulate, the levels of power and statistical confidence will increase to provide a more accurate comparison between the two valve types, 25 particularly for such uncommon outcome events as valve thrombosis. Thus the total number of 197 prosthetic mitral valves implanted in this study (n = 160 MVR and n = 37 DVR), with the associated levels of linearized hazard rates and the presented duration of follow-up, may not yet provide statistical confidence with a high enough level of power to exclude with certainty any differences in thrombogenicity, or other measured end points, that might be found to exist over the long term between CM and SJM MVRs.
In summary, this randomized study showed no major differences in clinical outcomes between patients who received standard CarboMedics and St Jude Medical bileaflet mechanical valve prosthesis after 5 years' follow-up. Choices with respect to these two valves can therefore be based on other considerations.
|
Appendix: Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
Prospective randomized trials are the criterion standard for allowing one to seek answers to clinical questions univariably, in this case the clinical performance of two mechanical valves. However, prospective randomized trials are only of value if they contain sufficient statistical power; that is enough patients and long enough follow-up. In this planned 10-year trial there are only 234 total CM recipients and 251 SJM recipients. Will these numbers yield sufficient statistical power? Let's begin with the mitral valves.
In my 1995 review of the literature on mechanical valves, the composite thromboembolism and bleeding rates for mitral valves were 4.16% for SJM valves and 4.99% for CM valves. A simple power calculation performed with just these numbers documents that to distinguish a significant difference between the two mitral valves would require more than 10,000 patient-years of follow-up in each group; that is, 1000 patients in each group followed up for 10 years without deaths or, if one accounts for mortality, 1400 patients in each group. Even according to the more disparate clinical results from Rosengart's 1998 smaller comparative study of the two valves, one would require a complete 10-year follow-up of 700 patients in each group, not just 75 CM valves and 85 SJM valves in the mitral position. Predictable comparison of the results between aortic SJM and CM valves, which have more similar composite thromboembolism and bleeding rates, would require more aortic valve replacements than are performed in the entire United Kingdom in any decade. Unfortunately, I am left to conclude that the results of the study are predictable merely by the virtue by the study design.
Several other features of the study warrant comment. Mean patient age was 60 years, and the 5-year actuarial survival of 81.1% did not differ significantly between the two groups. However, the predicted survival for the general population of the same age at 5 years is 92.1%. Thus patients receiving either of these state-of-the-art bileaflet prostheses had a mortality rate at 5 years that was almost 2.5 times that for the general population.
In addition, although the INR for these patients was quite high, at 3 to 4.5, linearized thromboembolic rates were 3.1% for SJM MVR patients and 3.3% for CM MVR patients. When combined with the linearized bleeding rates of 4% for SJM and 3.7% for CM, the composite thromboembolism bleeding rates for both valves are 7% per patient-year.
In addition to discussing the statistical power of the study, I would like to ask Mr Bryan to comment on the somewhat disappointing mortality and morbidity rates in this study with these two state-of-the-art prostheses.
Finally, AVR recipients differ from MVR recipients, so why include DVR patients in the trial? When a late complication occurs in a patient who has undergone DVR, to which prosthesis do you assign the blame?
Mr Bryan. I thank Dr Akins for his comments and also for some of the literature that he has provided through the years with respect to the evaluation of the mechanical heart valve prosthesis, which we have used as a standard reference in the conduct of this study.
With respect to the power of the study, the original statistical advice was based on a projected total population of 500 patients followed up for 10 years, with a projection of more than 2000 patient-years of follow-up. It was anticipated that with respect to composite end points, such as overall event-free survival at 10 years, a clinically meaningful difference between these two valves could be demonstrated with a high level of confidence. I do agree with Dr Akins that in studies like this there are difficulties in demonstrating statistically significant differences with respect to end points that occur infrequently, but that was not the original stated aim of this study. I think that a more realistic end point is event-free survival at 10 years.
Second, Dr Akins raised a question with regard to the mortality. I am not sure whether he is referring to the early mortality or the late valve-related mortality. With respect to the early mortality, I think that the institutional mortality, considering the significant proportion of patients undergoing reoperative surgery and the proportion of patients undergoing associated coronary bypass, certainly conforms to the time-related mortalities produced by both the Society of Thoracic Surgeons and the UK cardiac surgical register, and indeed all of the mortalities, which were not different between surgeons, were better than predicted when adjusted for standard Parsonnet.
With respect to the late mortality after mechanical prosthetic valve replacement, I think we all must acknowledge that mechanical valve replacement is not necessarily today viewed as the criterion standard. We, like everybody else, prefer to repair mitral valves rather than replace them.
There is no doubt that if you follow up patients prospectively every year laboriously in detail, you will collect a much higher rate of complications than if you send these same patients a letter at 5 or 10 years. That was drawn to my attention by Dr Akins' own review in 1996 in the Annals of Thoracic Surgery. I suggest that here we have at least a partial reflection of that kind of finding.
Dr Verdi J. DiSesa (Chicago, Ill). Your data inevitably led you to the conclusion that there was no difference in patient outcome with either valve and that the choice between them should therefore be based on other considerations. Can you enlighten us as to what some of those considerations might be?
Mr Bryan. They might relate to individual surgeon preference; for example, they might relate to differences in implantation techniques that surgeons might perceive. When we are all talking in the coffee room after operations, we may express certain preferences when we are implanting prosthetic valves. These preferences may not necessarily, unfortunately, be scientifically based. But I suggest that with mechanical bileaflet prostheses we can be secure in the knowledge that, at least as far as our interim data are concerned, whichever of these two prostheses we choose, there does not seem to be any actual difference in clinical outcome.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix: DiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Ascione, C.A. Rogers, C. Rajakaruna, and G.D. Angelini Inadequate Blood Glucose Control Is Associated With In-Hospital Mortality and Morbidity in Diabetic and Nondiabetic Patients Undergoing Cardiac Surgery Circulation, July 8, 2008; 118(2): 113 - 123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Gudbjartsson, T. Absi, and S. Aranki Mitral Valve Replacement Card. Surg. Adult, January 1, 2008; 3(2008): 1031 - 1068. [Full Text]
|
|
|
|
 |
|
 A. J. Bryan, C. A. Rogers, K. Bayliss, J. Wild, and G. D. Angelini Prospective randomized comparison of CarboMedics and St. Jude Medical bileaflet mechanical heart valve prostheses: Ten-year follow-up J. Thorac. Cardiovasc. Surg., March 1, 2007; 133(3): 614 - 622. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. M. Palatianos, A. M. Laczkovics, P. Simon, J. L. Pomar, D. E. Birnbaum, H. H. Greve, and A. Haverich Multicentered European Study on Safety and Effectiveness of the On-X Prosthetic Heart Valve: Intermediate Follow-Up Ann. Thorac. Surg., January 1, 2007; 83(1): 40 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kulik, F. D. Rubens, P. S. Wells, C. Kearon, T. G. Mesana, J. van Berkom, and B.-K. Lam Early Postoperative Anticoagulation After Mechanical Valve Replacement: A Systematic Review Ann. Thorac. Surg., February 1, 2006; 81(2): 770 - 781. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Bernal, J. Lorca, D. Prieto-Salceda, I. Pulitani, A. Ponton, I. Garcia, and J. M. Revuelta Performance at 10 years of the CarboMedics 'Top-Hat' valve. Postclamping time is a predictor of mortality Eur. J. Cardiothorac. Surg., February 1, 2006; 29(2): 144 - 149. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rajakaruna, C.A. Rogers, G.D. Angelini, and R. Ascione Risk factors for and economic implications of prolonged ventilation after cardiac surgery J. Thorac. Cardiovasc. Surg., November 1, 2005; 130(5): 1270 - 1277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. H. Kang, H. Ahn, K. H. Kim, and K.-B. Kim Long-Term Result of 1144 CarboMedics Mechanical Valve Implantations Ann. Thorac. Surg., June 1, 2005; 79(6): 1939 - 1944. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Dagenais, P. Cartier, P. Voisine, D. Desaulniers, J. Perron, R. Baillot, G. Raymond, J. Metras, D. Doyle, and P. Mathieu Which biologic valve should we select for the 45- to 65-year-old age group requiring aortic valve replacement? J. Thorac. Cardiovasc. Surg., May 1, 2005; 129(5): 1041 - 1049. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Shuhaiber, M. Massad, and A. Geha Study design in valve surgery and outcome J. Thorac. Cardiovasc. Surg., November 1, 2003; 126(5): 1660 - 1661. [Full Text] [PDF]
|
|
|
|
|
|
N. C. de Oliveira, T. E. David, J. Ivanov, S. Armstrong, M. J. Eriksson, H. Rakowski, and G. Webb Results of surgery for aortic root aneurysm in patients with Marfan syndrome J. Thorac. Cardiovasc. Surg., April 1, 2003; 125(4): 789 - 796. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. L. Grunkemeier and Y. Wu "Our complication rates are lower than theirs": Statistical critique of heart valve comparisons J. Thorac. Cardiovasc. Surg., February 1, 2003; 125(2): 290 - 300. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
