| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 1997;113:311-318
© 1997 Mosby, Inc.
SURGERY FOR ACQUIRED HEART DISEASE |
Supported by a grant from the National Heart Foundation of Australia (grant No. G90B2744).
Received for publication April 11, 1996 revisions requested June 26, 1996; revisions received August 8, 1996 accepted for publication August 12, 1996. Address for reprints: David C. McGiffin, MD, Department of Surgery, University of Alabama at Birmingham, UAB Station, Birmingham, AL 35294.
Abstract
Biologic valve re-replacement was examined in a series of 1343 patients who underwent aortic valve replacement at The Prince Charles Hospital, Brisbane, with a cryopreserved or 4° C stored allograft valve or a xenograft valve. A parametric model approach was used to simultaneously model the competing risks of death without re-replacement and re-replacement before death. One hundred eleven patients underwent a first re-replacement for a variety of reasons (69 patients with xenograft valves, 28 patients with 4° C stored allograft valves, and 14 patients with cryopreserved allograft valves). By multivariable analysis younger age at operation was associated with xenograft, 4° C stored allograft, and cryopreserved allograft valve re-replacement. However, this effect was examined in the context of longer survival of younger patients, which increases their exposure to the risk of re-replacement as compared with that in older patients whose decreased survival reduced their probability of requiring valve re-replacement. In patients older than 60 years at the time of aortic valve replacement, the probability of re-replacement (for any reason) before death was similar for xenografts and cryopreserved allograft valves but higher for 4° C stored valves. However, in patients younger than 60 years, the probability of re-replacement at any time during the remainder of the life of the patient was lower with the cryopreserved allograft valve compared with the xenograft valve and 4° C stored allografts.
Biologic valve replacement devices (xenograft and allograft valves) have an important and complementary role in conjunction with mechanical valve devices in the treatment of valvular heart disease. Freedom from anticoagulant-related hemorrhage and a low incidence of thromboembolic events remain clear advantages of biologic valves over mechanical devices.
1 Furthermore, the allograft valve has an additional advantage over both xenograft and mechanical valves for aortic valve replacement in the setting of endocarditis with a lower probability of recurrent endocarditis. 2
However, a clear disadvantage of biologic valves is their propensity for failure as a result of leaflet degeneration and mechanisms peculiar to allograft valves, including geometric distortion and changing mechanical properties of leaflets, specifically loss of radial extensibility of leaflets, which eventually results in central incompetence. 3
The usual means of presenting time-related events (such as death and re-replacement) after valve replacement is by actuarial analysis of single events. However, many patients, particularly elderly patients, die before a biologic valve requires re-replacement. In the setting of competing risks (in this context, valve re-replacement before death and death before re-replacement), interpretation of the Kaplan-Meier curve is problematic. 4 In this case, the Kaplan-Meier curve for freedom from valve re-replacement is conditional that no patient dies because the censoring process is used at the time of each death that occurs before re-replacement. Important information for the patient and the surgeon in deciding between a mechanical or biologic valve replacement device is the probability of re-replacement of a biologic valve before death, which takes into account the competing risks of death and re-replacement. Grunkemeier and colleagues 5 applied a "competing risk" analysis in a group of patients undergoing valve replacement with xenograft prostheses to determine both actuarial valve failure (conditional that no patient dies) and actual valve failure (probability of failure before death).
The purpose of this study was to analyze biologic valve re-replacement in a series of patients undergoing aortic valve replacement with cryopreserved and 4° C stored allograft valves and xenograft valves. The analysis was conducted in the "competing risk" domain to depict the unconditional probability (actual risk) of valve re-replacement before death after aortic valve replacement. The results of this analysis may provide information that could improve the precision with which a biologic valve replacement device and an individual patient are matched.
Patients and methods
The purpose of this retrospective observational study was to analyze the incidence and time relatedness of re-replacement of implanted biologic valves in the aortic position. The study group included patients who underwent aortic valve replacement (primary and subsequent valve replacements) with biologic valves, with or without concomitant procedures, at The Prince Charles Hospital between January 1, 1970, and January 31, 1990. The patient population was obtained from a previously published data set that examined survival after aortic valve replacement. 6 Patients who underwent concomitant mitral valve replacement were not included. Patients who received fascia lata and calf heterografts, which are now regarded as obsolete because of rapid failure, were not included. Patients who underwent an aortic valve replacement at another institution but subsequent aortic valve replacement at The Prince Charles Hospital were not included.
The study group comprised 1343 patients who underwent primary aortic valve replacement, of which 111 patients subsequently underwent a first re-replacement of the aortic valve prosthesis. Four of these patients underwent a second re-replacement but this second re-replacement was not considered in this analysis. At the initial aortic valve replacement, a xenograft valve was used in 950 patients, a 4° C stored allograft valve in 94 patients, and a cryopreserved allograft valve in 299 patients (Table I).
|
Operative technique.
Aortic valve replacements were done with use of the usual methods of cardiopulmonary bypass. From 1970 to 1978, myocardial protection was provided by continuous coronary perfusion and from 1978 to 1990, by cold crystalloid cardioplegia with topical cooling.
Xenograft valves were implanted with use of an interrupted suture technique. Allograft valves were inserted by a variety of methods including the subcoronary and cylindrical techniques and as an aortic root replacement. 7 Xenograft valves were implanted from 1978 to 1990, 4° C stored allografts from 1969 to 1976, and cryopreserved allografts from 1975 to 1990.
Data collection and follow-up.
A number of demographic, clinical, pathologic, and surgical variables were collected (Appendix A) by study of the patient hospital records and from collateral information provided by other hospitals and physicians. The data were collected by data managers and subsequently checked by the investigators before being entered into the database. Re-replacement of a previously implanted aortic valve and death were the end points of the study. Re-replacement of a valve device was undertaken because of leaflet failure caused by degeneration (xenografts and allografts) and changing mechanical properties of leaflets (allografts), geometric distortion (allografts), replacement valve endocarditis, miscellaneous technical reasons, and in the course of repair of an ascending aortic dissection. Re-replacement of a biologic valve during the repair of an acute or chronic ascending aortic dissection despite normal functioning of the valve occurred in five patients. One patient, who had previously undergone an aortic valve replacement with a xenograft valve, underwent cardiac transplantation 9 years later. The valve device was censored at the time of the transplantation.
Follow-up information was obtained from hospital records and by direct contact with the patient, family, cardiologist, and family physician. Follow-up was conducted through the months of May 1990 to February 1991 and the closing date for inclusion of events was February 22, 1991. Only one patient, who underwent aortic valve replacement with a xenograft prosthesis, was lost to follow-up. The number of patients followed up by time interval is outlined in Appendix Table II.
|
The analyses included estimates of the probability of re-replacement for any reason (leaflet failure, replacement valve endocarditis, and technical reasons) for xenograft valves and 4° C stored and cryopreserved allograft valves. Risk factors for re-replacement were sought by multivariable analysis using the method of maximum likelihood. A nomogram was derived for the relationship between probability of re-replacement for any reason before death and age at initial aortic valve replacement for xenografts and cryopreserved and 4° C stored allografts.
Results
A first valve re-replacement was done in 69 patients with a xenograft valve, 28 patients with a 4° C stored allograft valve, and 14 patients with a cryopreserved allograft valve. Valve re-replacement was done because of leaflet failure, replacement valve endocarditis, and a number of other miscellaneous causes (Table II).In the multivariable analysis only younger age at operation was found to be associated with increased probability of valve re-replacement.
|
|
|
|
|
When a cardiac surgeon and a patient are considering the choice of a valve device, the decision is partly influenced by an impression (in the surgeon's mind) of the likelihood that an implanted biologic valve will require re-replacement before the patient dies. This is, of course, particularly relevant in an elderly patient in whom, not infrequently, it is said that a biologic valve may "outlive" the patient. This process, which occurs informally, is taking into account in the decision the competing risks of death and re-replacement before death.
The Kaplan-Meier estimate of freedom from re-replacement (or other end point such as valve degeneration) assumes, because of the censoring process, that no patient dies and hence all patients will eventually require valve re-replacement. Consequently, the Kaplan-Meier estimate provides information regarding biologic valve durability and is a useful means to compare the durability of valve devices inasmuch as death is removed from the computation. Although this is important information, not only for valve manufacturers, but also for patients and surgeons, what is of more importance to the patient is the probability of requiring a valve re-replacement at some time during the remainder of his or her life. This information is not provided by the Kaplan-Meier estimate. Because death of the patient and re-replacement compete for the valve, both end points need to be solved for simultaneously.
The analytic tools to deal with multiple end points (competing risks) have been available for a long time. Grunkemeier and associates 5 proposed the term actuarial to describe the survivorship function (the usual Kaplan-Meier estimate) and the term actual risk to describe the competing risk depiction. Other terms that have been used to describe the actual risk include cumulative incidence function 8 and cause-specific failure probability. 9
The competing risk analysis can be depicted either as the proportion free from re-replacement before death or, alternatively, as proportion undergoing re-replacement before death. In this study, the latter depiction was chosen. These curves, particularly those that portray proportion of patients who undergo re-replacement before death, have an unusual shape because they become asymptotic (as time from operation increases, probability of re-replacement becomes constant). This is because the proportion of patients above the asymptotic part of the curve represents those who have died and hence the valve devices in those patients are no longer available for re-replacement. The competing risk plot for xenograft re-replacement (Fig. 1) has a clearly defined asymptote compared with the plots for 4° C stored allograft valve re-replacement (Fig. 2) and cryopreserved allograft valve replacement (Fig. 3). Competing risk is a process of assortment into events (in this case death and re-replacement). A clear asymptote for xenograft valves occurs at approximately 14 years, which means that after this time the assorting process has been completed and either death or re-replacement has occurred. However, there is no asymptote for 4° C stored and cryopreserved allograft valves, which indicates that the assorting process is still occurring at 15 years after operation. The estimates represented in the figures do not have confidence limits and hence carry some uncertainty.
In the multivariable analysis study, younger age at operation was found to be a risk factor for re-replacement of biologic valves. Previous studies by Magilligan and colleagues, 10 Jamieson and colleagues, 11 and Gallo, Nistal, and Artinano 12 demonstrated that younger age at operation was a risk factor for xenograft valve degeneration. Because these studies used the Kaplan-Meier estimate, censoring for death, younger age at operation reflected a biologic predisposition to tissue failure at younger age. However, in this study, by using a competing risk analysis, increased probability of re-replacement with younger age at operation reflects not only a biologic predisposition to leaflet failure, but also the fact that younger patients have a lower probability of dying than older patients and hence a higher probability of eventually requiring re-replacement. Similarly, when a competing risk analysis for re-replacement is applied to patients with allograft valves (4° C stored and cryopreserved), the effect of younger age at operation on the probability of re-replacement reflects the competing risk of death and re-replacement, as well as any biologic predisposition to leaflet failure.
The nomogram of the probability of undergoing a re-replacement sometime during the remainder of the patient's life, according to age at operation, is an important piece of information for a patient. This is depicted in Fig. 4 and provides this probability for xenograft, 4° C stored, and cryopreserved allografts. In terms of the proportion of patients who undergo re-replacement for any reason before death, cryopreserved allografts are superior to xenograft valves up until the age of approximately 60 years at the time of operation. For a patient 20 years old at the time of operation who receives a cryopreserved allograft valve, over the remainder of the patient's lifetime there is a 30% predicted probability that a re-replacement of the valve will be required. If this patient were to receive a 4° C stored allograft valve (a valve that is still widely used), the predicted probability of a re-replacement during the remainder of the patient's life would be approximately 50%. If a xenograft valve were implanted in this patient, there would be an approximately 85% probability of requiring re-replacement before death. Beyond 60 years of age at operation, there appears to be no difference in the predicted probability of re-replacement before death in patients with xenograft valves compared with patients with cryopreserved allograft valves. Beyond the age of 40 years at operation, the predicted probability of re-replacement before death for a patient who receives a 4° C stored allograft valve is higher than that for patients who receive xenograft or cryopreserved allograft valves. In this study, failure of cryopreserved allograft valves caused by mechanical and technical reasons was an important reason for re-replacement. Allograft valve failure is a complex interplay between mechanisms such as leaflet degeneration, geometric distortion, changing mechanical properties of the leaflet tissue, and aortic root dilation. However, the current preference for the cylindrical or root replacement technique of allograft valve insertion as opposed to the subcoronary technique (the method most frequently used in this series) may reduce the probability of re-replacement by decreasing geometric distortion and hence improving leaflet coaptation. The efficacy of this strategy will require testing by repeating this study in the future.
Inferences
Appendixes
Appendix A: Variables entered into the multivariate models
Demographic variables.
Age at operation; gender.
Clinical variables.
New York Heart Association functional class immediately before operation (class V refers to shock); preoperative symptoms (angina, heart failure, syncope); significant lung disease and significant renal dysfunction; significant hepatic dysfunction; left ventricular dysfunction and atrial fibrillation; complete heart block; and aortic valve physiologic features (aortic stenosis, aortic regurgitation, mixed lesion).
Significant lung disease was defined as chronic limiting symptoms caused by lung disease with objective evidence on respiratory function tests, chest radiograph, and other tests. Significant renal dysfunction was defined as level of creatinine greater than or equal to 0.2 mmol/L or level of urea greater than or equal to 15 mmol/L, or both conditions. Significant hepatic dysfunction was defined as bilirubin level greater than or equal to 35 mmol/L. Left ventricular dysfunction was defined on the basis of echocardiogram, radionuclide scan, or left ventriculogram at cardiac catheterization. Ejection fraction was categorized as follows: normal, greater than 0.54; mild, 0.37 to 0.54; moderate, 0.20 to 0.37; and severe, less than 0.20. Fractional shortening was categorized as follows: normal, greater than 30%; mild, 20% to 30%; moderate, 10% to 20%; and severe, less than 10%.
Pathologic variables.
Presence of coronary artery disease; aortic valve pathologic condition (rheumatic, calcareous aortic valve disease caused by a congenitally abnormal valve); calcareous aortic valve disease on a trileaflet valve; infective endocarditis, prosthetic valve dysfunction, aortic wall disease, or congenital valve disease (noncalcareous); and degree of left ventricular hypertrophy.
Calcareous aortic valve disease in a trileaflet valve was considered senile degenerative disease. Aortic wall disease was a pathologic condition of the aortic wall, including the aortic root, that resulted in dilation or dissection. For patients with endocarditis on a prosthetic valve or an abnormal valve (such as a rheumatic native valve) the pathologic condition was classified as endocarditis. The degree of left ventricular hypertrophy was defined as normal, mild, moderate, or severe (on the basis of the surgeon's intraoperative assessment).
Surgical variables.
Date of operation; surgeon; cardiopulmonary bypass time; crossclamp time; use of cardioplegia; valve type; valve size; associated procedures; and completeness of revascularization for patients with coronary artery disease.
Appendix B
We let S1(t;x) and f1(t;x) denote the survival function and probability density function, respectively, where t denotes the time to death after the initial valve replacement operation for a patient age x (months) at the time of this operation. That is,
where T is the random variable denoting time to death.
Concerning the other event of re-replacement, we let S2(t;x) and f2(t;x) denote the survival function and probability density function, respectively, for a patient age x at the time of initial operation, where t denotes the time to re-replacement in a hypothetical population for which there is no competing risk of death.
The survival function Si(t;x) is modeled as
Si(t;x) = exp[–exe
i(eßit–1)/ßi]
for i = 1, 2. The corresponding hazard function hi (t;x) is given by
hi(t;x) = exhoi(t)
where
hoi(t) = e
i + ßit
is the hazard function for the Gompertz distribution. This effect of the age x of the patient at the time of the initial operation is modeled additively on the log scale of the baseline hazard function, which is taken to be of Gompertz form.
The survival functions S1(t;x) and S2(t;x) were fitted separately to the data, using the method of maximum likelihood where, in the estimation of the latter, a patient who died without a re-replacement was taken to be censored with respect to this event. The values of the parameter estimates are shown in Appendix Table I.
|
It can be interpreted as the probability that a patient age x at the time of the initial operation will undergo a re-replacement by time t afterward (and before the competing risk of death occurs). Thus the quantity
2(x) = C2(
;x)
is the probability that a patient age x at the time of the initial operation will actually undergo a re-replacement before death.
Because the original data were no longer readily accessible, confidence intervals could not be presented.
Acknowledgments
We are grateful to Dr. Greg Stafford, Dr. Michael Gardner, and Dr. Peter Pohlner, The Prince Charles Hospital, for allowing us to include patients under their care. Our understanding of competing risk has been greatly assisted by productive interaction with Dr. Eugene H. Blackstone. We would like to thank Jane Owenby (University of Alabama at Birmingham) for manuscript preparation.
Footnotes
From the Departments of Cardiac Surgeryb and Cardiology,a The Prince Charles Hospital, and the Department of Mathematics,c University of Queensland, Brisbane, Queensland, Australia, and the Department of Surgery,d University of Alabama at Birmingham, Birmingham, Ala. 
References
This article has been cited by other articles:
|
|
 |
|
  J. P. Greelish, R. M. Ahmad, J. M. Balaguer, M. R. Petracek, and J. G. Byrne Reoperative Valve Surgery Card. Surg. Adult, January 1, 2008; 3(2008): 1159 - 1174. [Full Text]
|
|
|
|
 |
|
 K. R. Kanter, P. M. Kirshbom, and B. E. Kogon Redo Aortic Valve Replacement in Children Ann. Thorac. Surg., November 1, 2006; 82(5): 1594 - 1597. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.R.E. Jamieson, L.H. Burr, R.T. Miyagishima, M.T. Janusz, G.J. Fradet, H. Ling, and S.V. Lichtenstein Re-operation for bioprosthetic aortic structural failure - risk assessment Eur. J. Cardiothorac. Surg., December 1, 2003; 24(6): 873 - 878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W.R.E. Jamieson, L.H. Burr, R.T. Miyagishima, M.T. Janusz, G.J. Fradet, S.V. Lichtenstein, and H. Ling Reoperation for Bioprosthetic Mitral Structural Failure: Risk Assessment Circulation, September 9, 2003; 108(90101): II-98 - 102. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. F. Baskett, M. A. Nanton, A. E. Warren, and D. B. Ross Human leukocyte antigen-DR and ABO mismatch are associated with accelerated homograft valve failure in children: implications for therapeutic interventions J. Thorac. Cardiovasc. Surg., July 1, 2003; 126(1): 232 - 238. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Sadowski, B. Kapelak, K. Bartus, P. Podolec, P. Rudzinski, T. Myrdko, K. Wierzbicki, and A. Dziatkowiak Reoperation after fresh homograft replacement: 23 years' experience with 655 patients Eur. J. Cardiothorac. Surg., June 1, 2003; 23(6): 996 - 1001. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. Byrne, B. J. Phillips, and L. H. Cohn Reoperative Valve Surgery Card. Surg. Adult, January 1, 2003; 2(2003): 1047 - 1056. [Full Text]
|
|
|
|
|
|
R. Sodian, S. P. Hoerstrup, J. S. Sperling, S. H. Daebritz, D. P. Martin, F. J. Schoen, J. P. Vacanti, and J. E. Mayer Jr Tissue engineering of heart valves: in vitro experiences Ann. Thorac. Surg., July 1, 2000; 70(1): 140 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. S. Carr-White, S. Glennan, S. Edwards, F. D. Ferdinand, A. C. Desouza, J. R. Pepper, and M. H. Yacoub Pulmonary Autograft Versus Aortic Homograft for Rereplacement of the Aortic Valve : Results From a Subset of a Prospective Randomized Trial Circulation, November 9, 1999; 100(90002): II-103 - 106. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Lund, V. Chandrasekaran, R. Grocott-Mason, H. Elwidaa, R. Mazhar, A. Khaghani, A. Mitchell, C. Ilsley, and M. H. Yacoub PRIMARY AORTIC VALVE REPLACEMENT WITH ALLOGRAFTS OVER TWENTY-FIVE YEARS: VALVE-RELATED AND PROCEDURE-RELATED DETERMINANTS OF OUTCOME J. Thorac. Cardiovasc. Surg., January 1, 1999; 117(1): 77 - 91. [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 |
