Ann Thorac Surg 2003;75:28-34
© 2003 The Society of Thoracic Surgeons
Original article: cardiovascular
Durability and outcome of aortic valve replacement with mitral valve repair versus double valve replacement
Masaki Hamamoto, MDa,
Ko Bando, MDa*,
Junjiro Kobayashi, MDa,
Toshihiko Satoh, MD, MPHb,
Yoshikado Sasako, MDa,
Kazuo Niwaya, MDa,
Osamu Tagusari, MDa,
Toshikatsu Yagihara, MDa,
Soichiro Kitamura, MDa
a Department of Cardiovascular Surgery, National Cardiovascular Center, Osaka, Japan
b Department of Public Health, Kitasato University, Kitasato, Japan
* Address reprint requests to Dr Bando, Department of Cardiovascular Surgery, National Cardiovascular Center, 5-7-1 Fujishirodai, Suita, Osaka 565-8565 Japan
e-mail: kobando{at}hsp.ncvc.go.jp
Presented at the Thirty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28–30, 2002.
 |
Abstract
|
|---|
BACKGROUND: The purpose of this study was to evaluate morbidity and mortality after double valve replacement (DVR) and aortic valve replacement with mitral valve repair (AVR + MVP).
METHODS: From 1977 to 2000, 379 patients underwent DVR (n = 299) or AVR + MVP (n = 80). Actuarial survival and freedom from reoperation were determined by the Kaplan-Meier method. Potential predictors of mortality and reoperation were entered into a Cox multiple regression model. Propensity score was introduced for the multivariable regression modeling for adjustment of a selection bias.
RESULTS: Survival 15 years after surgery was similar between the groups (DVR, 81% ± 3%; AVR + MVP, 79% ± 7%; p = 0.44). Freedom from thromboembolic event at 15 years was similar between the groups (p = 0.25). Freedom from mitral valve reoperation at 15 years was significantly better for the DVR group (54% ± 5%) as compared with the AVR + MVP group (15% ± 6%; p = 0.0006), primarily due to progression of mitral valve pathology and early structural deterioration of bioprosthetic aortic valve used for patients with AVR + MVP. After AVR + MVP, freedom from mitral reoperation at 15 years was 63% ± 16% for nonrheumatic heart diseases, and 5% ± 5% for rheumatic disease (p = 0.04).
CONCLUSIONS: Although both DVR and AVR + MVP provided excellent survival, DVR with mechanical valves should be the procedure of choice for the majority of patients because of lower incidence of valve failure and similar rate of thromboembolic complications compared with AVR + MVP. MVP should not be performed in patients with rheumatic disease because of higher incidence of late failure.
|
Introduction
|
|---|
For patients with aortic and mitral valve disease, double valve replacement (DVR) has been advocated as a standard surgical option, which has been safely performed even in elderly populations in recent years. In these patients, a majority of aortic valve disease required valve replacement because the early and late results of aortic valve repair have not been satisfactory [1,2]. In contrast, valve repair has been advocated in mitral valve disease [3, 4]. However, the durability and outcome of combined aortic valve replacement and mitral valve repair versus double valve replacement remain to be determined. The purpose of this study was to evaluate survival and late outcome after DVR and aortic valve replacement and mitral valve repair (AVR + MVP) for double valve disease.
|
Patients and methods
|
|---|
From 1977 to 2000, 379 patients underwent AVR with either mitral valve replacement (MVR) (DVR group; n =299) or MVP (n = 80). The cases of concomitant procedure with coronary artery bypass grafting or aortic surgery were excluded from the study. The preoperative clinical characteristics of each group are shown in Table 1.
The DVR group was more common in patients with atrial fibrillation (Af) (p < 0.0001). The AVR + MVP group was more common in patients with pure mitral regurgitation (MR) (p < 0.0001). The age of patients having bioprosthetic valve was 52 ± 12 years old in the AVR + MVP group and 50 ± 9 years old in the DVR group. The proportion of New York Heart Association (NYHA) functional class III and IV was similar between the two groups. In the DVR group, follow-up ranged from 6 months to 20.1 years (8.6 ± 6.1 years), with a total of 2,561 patient-years. In the AVR + MVP group, follow-up ranged from 6 months to 21.7 years (9.4 ± 7.1 years), with a total of 752 patient-years.
Surgical procedure
The majority of patients with DVR had mechanical valve prostheses in both aortic (219/299 patients, 73%) and mitral (215/299, 72%) position. In contrast, a bioprosthetic valve was more frequently used in the AVR + MVP group (49/80, 61%) (p < 0.0001). In both groups, an Ionescu-Shiley valve (58/84 [69%] in the DVR group, 28/49 [57%] in the AVR + MVP group) was primarily used as a bioprosthetic valve between 1981 and 1984 for all generations. Since then, the Carpentier Edwards valve was commonly used for the patients more than 70 years of age, and most recently, the Mosaic valve was introduced in 1999. In patients who received mechanical valves, the St. Jude Medical valve (135/215 patients, 63%) was the most frequently used in both aortic and mitral positions. The Maze procedure was more common with the DVR group as compared with the AVR + MVP group because the rate of preoperative Af rhythm was more frequently associated with the DVR group as compared with the AVR + MVP group (DVR, 228/299 [76%] vs AVR + MVP, 39/80 [49%]; p = 0.02). Several different techniques were used for MVP, including commissurotomy (36 patients, 45%), Kay procedure (33 patients, 41.3%), ring annuloplasty (10 patients, 12.5%) use of Duran or Carpentier-Edwards ring and leaflet resection (5 patients, 6.3%), or a combination of these techniques.
Data collection and follow-up
We retrospectively reviewed the data from the operation notes, anesthesia records, clinical histories, laboratory investigations, and cardiac catheterization. This retrospective study was approved by the Internal Review Board of National Cardiovascular Center. Follow-up data were collected from National Cardiovascular Center records of outpatient visits and correspondence with referring physicians. All clinical characteristics were accumulated as a computerized database and analyzed in the Appendix. The definitions of morbidity and mortality were based on the published guidelines of Society of Thoracic Surgeons and American Association for Thoracic Surgery "Guidelines for Reporting Morbidity and Mortality After Cardiac Valvular Operations" [5].
Statistical methods
Statistical analysis of the two groups were performed using the Pearson chi square test with Yates’ correction or Fisher’s exact test for categorical variables and the Mann-Whitney U test for continuous variables. Unadjusted survival curves for the two groups were generated using the Kaplan-Meier method. The log-rank test was used for the comparison of the unadjusted survival curves. Cox’s multivariate regression model with the propensity score method [6] was performed to assess the influence of surgical method on the probability of survival or reoperation. The incorporation of a propensity score balances the weight of covariates between the two groups on each patient level so that the comparisons of these two groups of patients were more significant. Hazards ratio and 95% confidence intervals (CI) were provided. All statistical analyses were performed using the software package SPSS 10.0 for Windows (SPSS Inc., Chicago, IL). Differences were considered statistically significant when p was less than 0.05.
|
Results
|
|---|
Early and late mortality
Overall hospital mortality was 4.5% (17/379), which declined to 2.9% (2/69) during the last 5 years. The major causes of hospital death were postoperative low cardiac output (LOS) (5/17, 31%) or multiple organ failure (MOF) (6/17, 31%). Three deaths (0.8%) were related to the prosthetic valve failure. There were 36 late deaths, 27 in the DVR group and 9 in the AVR + MVP group (p = 0.70). Cardiac-related late deaths were observed in 5 patients in the DVR group, whereas only 1 patient died in the AVR + MVP group. Actuarial survival at 5, 10, and 15 years was 96% ± 2%, 92% ± 4%, and 79% ± 7% for the AVR + MVP group versus 89% ± 2%, 87% ± 2%, and 81% ± 3% for the DVR group, respectively (p = 0.44) (Fig 1).
View larger version (17K):
[in this window]
[in a new window]
|
Fig 1. Long-term actuarial survival for the subgroups of patients with AVR + MVP and DVR. Solid line = AVR + MVP; dotted line = DVR. (AVR = aortic valve replacement; DVR = double valve replacement; MVP = mitral valve repair; Pts. = patients.)
|
|
Valve-related death occurred in 5 patients in the DVR group (5/299, 1.7%): valve thrombus (1), cerebral infarction (1), stuck valve (1), acute subdural hematoma (1), and prosthetic endocarditis (1). In contrast, there was only one valve-related death in the AVR + MVP group (1/80, 1.3%): anticoagulant-related cerebral hemorrhage. Actuarial freedom from valve-related death at 5, 10, and 15 years was 100%, 98% ± 2%, and 98% ± 2% in the AVR + MVP group versus 99% ± 1%, 99% ± 1%, and 97% ± 2% in the DVR group (p = 0.73).
Valve-related complications
The actuarial estimates of freedom from overall valve-related complications at 15 years were 45% ± 5% and 18% ± 7% for the DVR group and the AVR + MVP group, respectively, with significance (p = 0.02) (Table 2)
. This was probably related to the fact that 61% of patients with AVR + MVP received bioprosthetic valves, whereas only 28% of the patients with DVR had bioprosthetic valves (p < 0.0001). Subsequently, structural valvular deterioration (SVD) of the bioprosthetic valve developed in 28 of 80 patients (35%) in the AVR + MVP group compared with 50 of 299 (17%) patients in the DVR group (p = 0.0006). Actuarial freedom from SVD at 15 years in the AVR + MVP group (26% ± 8%) was significantly lower than that of the DVR group (67% ± 4%) (p = 0.002). When stratified by the type of prosthesis, DVR as well as AVR + MVP using bioprosthetic valve resulted in lower freedom from SVD at 15 years (27.7% ± 5.8% and 18.5% ± 7.7%, respectively) compared with both groups using mechanical valves (both 100%) (Fig 2).
View larger version (17K):
[in this window]
[in a new window]
|
Fig 2. Freedom from structural valve deterioration for the subgroups of patients with AVR + MVP and DVR stratified by the type of prosthesis. (AVR = aortic valve replacement; DVR = double valve replacement; MVP = mitral valve repair; Pts. = patients.)
|
|
Anticoagulant-related complications
The systemic anticoagulant therapy with warfarin sodium was initiated in all patients immediately after the extubation. The warfarin sodium was adjusted from 1.8 to 2.8 of INR for the MVR and from 1.5 to 2.5 for the AVR [7]. Three months later, warfarin sodium was discontinued in the patients having only bioprosthetic valve without atrial fibrillation. However, the patients who received a mechanical valve or remained in atrial fibrillation after valve surgery continued to take warfarin sodium. The Maze procedure has been done for 7 of 80 patients (9%) with AVR + MVP, and 60 of 299 patients (20%) with DVR since 1994. Return to sinus rhythm was obtained for 4 of 7 patients (57%) in the AVR + MVP group and 45 of 60 patients (75%) in the DVR group. Ultimately, sinus rhythm was maintained in the postoperative periods for 63% (50/80) of patients in the AVR + MVP group and 33% (98/299) of patients in the DVR group, respectively. Freedom from systemic anticoagulant therapy at 10 years was 58% for the AVR + MVP group and 23% for DVR group (p = 0.0006). All patients with atrial fibrillation after surgery continued to have warfarin sodium regardless of the type of valve received.
Postoperative thromboembolic complication was observed in 3 of 80 (3.8%) patients with the AVR + MVP group and 5 of 299 (1.7%) patients in the DVR group. All these thromboembolic events occurred in the cerebral lesion, but not in the other organs. Estimates of freedom from thromboembolism at 15 years was 95% ± 3% in the AVR + MVP group versus 97% ± 1% in the DVR group (p = 0.26).
There was only one bleeding event (1.3%) in patients with the AVR + MVP group, whereas four episodes (1.3%) occurred in the DVR group and one resulted in a fatality. Estimates of freedom from bleeding events at 15 years were 99% ± 1% in the AVR + MVP group and 98% ± 1% in the DVR group (p = 0.94).
Reoperations
Reoperation was required in 39 of 80 patients (49%) in the AVR + MVP group, and 69 of 299 patients (23%) in the DVR group (p < 0.0001). All these patients underwent mitral valve reoperations with or without second AVR. In the AVR + MVP group, the major cause of reoperation was progression of mitral stenosis or regurgitation (39/39) with the structural deterioration of bioprosthetic valves in the aortic position (32/39), or associated with the pannus formation after AVR with mechanical valves (2/39). Thus, the remaining 5 patients had only MVR without aortic valve surgery. In this group, reoperation was required for 22 of 28 patients (78.6%) with an Ionescu-Shiley valve 15 years after surgery, whereas 4 of 10 patients with the Carpentier-Edwards valve required reoperation during the same time period. In the DVR group, the SVD of bioprosthetic valve in either aortic or mitral position (54/69, 78%) was a primary cause of reoperation. Freedom from mitral reoperation in the AVR + MVP group was significantly lower as compared with that of the DVR group (p = 0.0006)(Fig 3). Although the AVR + MVP group had a higher incidence of reoperation, the mortality in reoperation was similar between the two groups: 13% (5/39) in the AVR + MVP group versus 9% (6/69) in the DVR group (p = 0.78).
View larger version (19K):
[in this window]
[in a new window]
|
Fig 3. Freedom from mitral reoperation for the subgroups of patients with AVR + MVP and DVR. (AVR = aortic valve replacement; DVR = double valve replacement; MVP = mitral valve repair; Pts. = patients.)
|
|
Reoperation in the AVR + MVP group stratified by etiology of the valve disease
In the AVR + MVP group, when freedom from survival was stratified by the etiology of the valve disease, freedom from reoperation at 5, 10, and 15 years was 85% ± 10%, 76% ± 13%, and 63% ± 16% for nonrheumatic heart disease, and 89% ± 5%, 46% ± 11%, and 5% ± 5% for rheumatic heart disease, respectively (p = 0.04) (Fig 4).
View larger version (17K):
[in this window]
[in a new window]
|
Fig 4. Freedom from mitral reoperation after AVR + MVP stratified by the rheumatic etiology. Solid line = nonrheumatic; dotted line = rheumatic. (AVR = aortic valve replacement; DVR = double valve replacement; MVP= mitral valve repair; Pts. = patients.)
|
|
Risk factor analysis for survival
By univariate analysis, significant predictors of poor survival included older age, past history of myocardial infarction, preoperative NYHA class IV, and longer cardiopulmonary bypass (CPB) time and aortic cross-clamp time (Table 3).
By multivariate analysis using propensity score, significant predictors for early and late mortality included older age, past history of myocardial infarction, NYHA class IV, early year of operation, and longer CPB time (Table 4).
Risk factor analysis for reoperation
By univariate analysis, risk for reoperation included early year of operation, mitral repair as the first operation, omission of Maze procedure for the patients with Af, and the use of bioprosthetic valve (Table 5).
By multivariate analysis using propensity score, early year of operation and the use of bioprosthetic valve were predictors for reoperation (Table 6).
|
Comment
|
|---|
For surgery in patients with double valve disease, the choice of MVP versus MVR remains controversial. Our retrospective study of more than 22 years indicated both MVP and MVR combined with AVR provided excellent long-term survival (> 85% at 10 years after surgery in both groups). This was favorably compared with other reports [8–11], probably due to younger population (mean age, 54 years) and smaller number of patients in NYHA class III/IV (46%) in our study.
Freedom from thromboembolic event was similar between the DVR and the AVR + MVP group up to 15 years after surgery. Moreover, freedom from other morbidities, including major bleeding (98% ± 1% in the DVR group vs 99% ± 1% in the AVR + MVP group at 15 years) or prosthetic valve endocarditis (93% ± 3% in the DVR group vs 97% ± 3% in the AVR + MVP group at 15 years) was also similar between the groups. These results indicated that both DVR and AVR + MVP provided excellent survival and low morbidity up to 15 years after surgery
In sharp contrast, regarding reoperation, a significantly higher incidence was observed in the AVR + MVP group as compared with the DVR group. This was probably related to the fact that two-thirds of patients with AVR + MVP received bioprosthetic valves even in younger populations; of those, 57% had the first-generation Ionescu-Shiley valve [12]. Early structural deterioration of this valve was the primary cause of reoperation for the AVR + MVP group. On the other hand, double valve surgery using mechanical valve in both groups provided better results without structural deterioration, which resulted in no reoperation (Fig 2).
In the AVR + MVP group, the patients with rheumatic heart disease had higher risk for reoperation as compared with those with nonrheumatic heart disease. Because the majority of patients with rheumatic heart disease had some component of mitral stenosis, decision to perform the MVP versus MVR is difficult [13]. In this series, open mitral commissurotomy was commonly performed for the patients with mitral stenosis; most of these patients eventually required mitral valve replacement. Thus, our results indicated that double valve replacement might be a best option for patients with rheumatic heart disease [14].
For risk factor analysis of mortality and freedom from reoperation, we have used a multivariable model with the incorporation of propensity score. This will minimize the biases in the observational study and have demonstrated that the choice of MVR or MVP does not significantly affect patient long-term survival up to 20 years (Table 5). Instead, history of myocardial infarction and preoperative NYHA class IV were the strong predictors for early and late death after surgery for double valve disease.
Moreover, multivariable risk analysis using propensity score revealed that only bioprosthesis and early year of operation were the significant predictors for reoperation. There was certainly a logistic concern that year of operation was a risk simply because cases in recent years did not have enough follow-up time. However, if we analyze the data in patients before 1995, year of operation was still the strong predictor for reoperation (data not shown). Advanced myocardial protection and improved operative technique as well as well understanding of pathophysiology of the valves may contribute to these improvements.
The major limitation of our study is that it was not randomized, and there were significant differences between the baseline characteristics of patients in the DVR group and the AVR + MVP group. The decision for MVP versus MVR reflected the surgeons’ experience. To minimize the effect of these biases, propensity score was incorporated into the multivariate analysis. Another limitation includes inability to assess precise valve function by echocardiography because late follow-up echocardiography was available in only 60% of the patients, and this was not incorporated in the current study. Instead, assessment of durability of valves was based on either survival or free from reoperation. Further prospective study is certainly warranted to elucidate the precise difference of durability of valves in the two cohorts.
In conclusion, although both DVR and AVR + MVP resulted in good survival, DVR with mechanical valves should be the procedure of choice for the majority of patients, because of higher freedom from valve failure and similar rate of thromboembolic complications as compared with AVR + MVP. MVP should not be performed in patients with rheumatic disease because of the high incidence of late failures (Appendix).
|
Footnotes
|
|---|
* All variables are dichotomous (yes/no), unless indicated to be continuous or ordinal. 
|
Appendix
|
|---|
Variables studied in multivariable analysis of risk factors for survival and reoperation*
Demography
Age (years) at operation, gender
Cardiac comorbidity
NYHA functional class, history of myocardial infarction, transient ischemic attack, chronic heart failure, rheumatic fever, preoperative atrial fibrillation
Operative procedures and time
Year of operation, surgical method of AVR with MVP, or double valve replacement, concomitant Maze procedure, cardiopulmonary bypass time, and aortic cross-clamp time.
|
Discussion
|
|---|
DR LISHAN AKLOG (Boston, MA): That was an excellent presentation. I just had a question; maybe I missed it, but I did not see where you specified what the severity of mitral regurgitation was in these patients preoperatively and, more specifically, what the functional classification was. And related to that, what was the technique of mitral repair? Did everybody just receive an annuloplasty alone or were there other concomitant techniques that were used in the mitral repair patients?
DR HAMAMOTO: Preoperatively echocardiography was performed and severity of mitral valve dysfunction was almost the same between the AVR plus MVP group and double valve replacement group. In the case of aortic valve replacement, it was mechanical valve plus mitral valve repair group; the severity of mitral valve dysfunction was a little bit milder than the aortic valve replacement with bioprosthetic mitral valve repair.
DR AKLOG: How severe was it? Did these patients have moderate MR, was it mild MR, and what was the severity, on average, even if it was the same between the two groups?
DR HAMAMOTO: In the AVR + MVP group, the severity of mitral valve dysfunction was moderate to severe. Similar results were found with the DVR group; especially in the AVR with mechanical valve plus mitral valve plasty group, they had mild to moderate regurgitation.
DR AKLOG: What was the technique of repair? Was it annuloplasty alone?
DR HAMAMOTO: The technique of mitral valve repair is commissurotomy in 40%, Kay procedure was in about 40%, and annuloplasty about 20%, and these techniques were combined, two or three techniques were combined.
DR AKLOG: Were you able to correlate the durability with the specific technique? You said only 20% got an annuloplasty ring. Was there a correlation between the technique of repair? The ones that failed, were they less likely to have had an annuloplasty ring?
DR BANDO: Dr Aklog, let me answer the question for you. I am a co-author of this paper. The patients with rheumatic heart disease and significant mitral stenosis were the majority of patients. That is why we used commissurotomy in the patients. But for the majority of the patients with pure mitral regurgitation, we do use the ring as well as the posterior leaflet repair.
DR BELHAM AKPINAR (Istanbul, Turkey): As far as I understood, you stopped giving anticoagulation after some time in both groups, more commonly in the group that you had repair of the mitral valve. Is that true, you do not give any anticoagulation after some time? You stop giving anticoagulation?
DR HAMAMOTO: In both groups just after the operation anticoagulation was started by warfarin, but after 3 months the bioprosthetic valve replacement plus mitral valve plasty group had no anticoagulation, no antiplatelet drugs.
DR JONATHAN HAMMOND (Hartford, CT): In the mitral valve repair group that was reoperated upon, maybe you said this in your presentation, but I did not catch it, which valve or valves were being redone?
DR HAMAMOTO: Aortic valve replacement with the mitral valve repair group had a higher incidence of reoperation.
DR HAMMOND: In other words, the mitral repair was durable. It was the aortic valve that had to be redone?
DR HAMAMOTO: In almost all cases the repair of the mitral valve is replacement with bioprosthetic mitral valve repair.
|
References
|
|---|
- Gillinov A.M., Blackstone E.H., Cosgrove D.M., et al. Durability of combined aortic and mitral valve repair. Ann Thorac Surg 2001;72:20-27.[Abstract/Free Full Text]
- Casselman F.P., Gillinov A.M., Akhrass R., Kasirajan V., Blackstone E.H., Cosgrove D.M. Intermediate term durability of bicuspid aortic valve repair for prolapsing leaflet. Eur J Cardiothorac Surg 1999;15:302-308.[Abstract/Free Full Text]
- Fasol R., Mahdjoobian K., Joubert-Hubner E. Mitral repair in patients with severely calcified annulus: feasibility, surgery and results. J Heart Valve Dis 2002;11:153-159.[Medline]
- Gillinov A.M., Cosgrove D.M. Mitral valve repair for degenerative disease. J Heart Valve Dis 2002;11(Suppl 1):S15-20.
- Edmunds L.H., Jr, Grunkemeier G.L., Miller D.C., Weisel R.D. Guidelines for reporting morbidity and mortality after cardiac valvular operaions. Ann Thorac Surg 1996;62:932-935.[Abstract/Free Full Text]
- Rosenbaum P.R., Robin D.B. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41-55.[Abstract/Free Full Text]
- Matsuyama K., Matsumoto M., Sugita T., et al. Anticoagulant threrapy in Japanese patients with mechanical mitral valve. Circ J 2002;66:668-670.[Medline]
- Brown P.S., Jr, Roberts C.S., Mcintosh C.L., Swain J.A., Clark R.E. Relation between choice of prostheses and late outcome in double valve replacement. Ann Thorac Surg 1993;55:631-640.[Abstract]
- Milano A., Guglielmi C., De Carlo M., et al. Valve-related complications in elderly patients with biological, and mechanical aortic valves. Ann Thorac Surg 1998;66(Suppl):S82-87.
- Armenti F., Stephenson L.W., Edmunds L.H., Jr Simultaneous implantation of St. Jude Medical aortic and mitral prostheses. J Thorac Cardiovasc Surg 1987;94:733-739.[Abstract]
- Mueller X.M., Tevaearai H.T., Stumpe F., et al. Long-term results of mitral-aortic valve operations. J Thorac Cardiovasc Surg 1998;115:298-309.
- Machida H., Ueda H., Nakano K., et al. A morphologic study of Carpentier-Edwards pericardial xenografts in the mitral position exhibiting primary tissue failure in adults in comparison with Ionescu-Shiley pericardial xenografts. J Thorac Cardiovasc Surg 2001;122:649-655.[Abstract/Free Full Text]
- Duran C.M., Gometza B., De Vol E.B. Valve repair in rheumatic mitral disease. Circulation 1991;84(Suppl 5):III125-132.
- Skudicky D., Essop M.R., Sareli P. Time-related changes in left ventricular function after double valve replacement for combined aortic and mitral regurgitation in a young rheumatic population. Circulation 1997;95:899-904.[Abstract/Free Full Text]
Top
Abstract
Introduction
