|
|
||||||||
Eur J Cardiothorac Surg 2004;25:364-370
© 2004 Elsevier Science NL
a Division of Cardiac Surgery, University of Ottawa Heart Institute, 40 Ruskin Street, Suite 3403, Ottawa, Ont., Canada K1Y 4W7
b Department of Epidemiology and Community Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada K1H 8M5
Received 19 October 2003; received in revised form 7 December 2003; accepted 15 December 2003.
* Corresponding author. Address: Division of Cardiac Surgery, University of Ottawa Heart Institute, 40 Ruskin Street, Suite 3403, Ottawa, Ontario, Canada K1Y 4W7. Tel.: +1-613-761-4893; fax: +1-613-761-4713
e-mail: mruel{at}ottawaheart.ca
| Abstract |
|---|
|
|
|---|
Key Words: Valves Follow-up studies Prosthesis Surgery Complications
| 1. Introduction |
|---|
|
|
|---|
| 2. Methods |
|---|
|
|
|---|
18 years of age) who underwent AVR or MVR between 1970 and 2002 at our institution were followed annually in a dedicated valve clinic. Patients had a history focused on the determination of functional status and the occurrence of valve-related complications, physical examination, ECG and chest radiograph, complete blood count, serum chemistries, and international normalized ratio determinations when applicable. The study cohort included 3233 consecutive adult patients who underwent a total of 3633 operations for replacement of the aortic, mitral, or both valves. Of these 3233 patients, no information is available on 291 (9.0%) who did not attend the follow-up valve clinic at any point in time, and only partial information could be obtained on an additional 350 patients (10.8%) who were later lost to follow-up at a mean of 4.7±4.3 years after valve replacement. Total follow-up was 21179 patient-years (mean 6.6±5.0 years; maximum 32.4 years).
The preoperative characteristics of the cohort by site of implant are presented in Table 1. Two thousand three hundred forty-eight operations consisted of AVR, 1062 of MVR, and 223 of concomitant AVR and MVR. Table 2 shows the types of prostheses implanted in the study population.
|
|
2.3. Statistical analyses
Data were imported and analyzed in Intercooled Stata 8.0 (Stata, College Station, TX). For actuarial analyses, patients were censored at the time of their last follow-up visit or death if they had not yet been reoperated. Censoring was assumed to be independent of predictors and outcomes.
Actual freedom from all-cause reoperation or reoperation for SVD was determined for each site of implantation and each class (i.e. mechanical versus bioprosthetic) of valve prostheses [6,10]. Crude actuarial reoperation rates are not reported. To increase the study's generalizability with respect to current cardiac surgical practice, separate analyses were also performed that only included patients who were implanted with currently available prostheses.
Potential univariate predictors were individually tested for equality of actuarial freedom from reoperation with a Log-Rank test. These predictors were also tested for 15-year actual freedom from reoperation with logistic regression. Univariate effect estimates and P-values were used to guide multivariate model selection and are not reported.
2.3.1. Reoperation due to structural valve deterioration hazard (actuarial)
No a priori distributional assumption was made with respect to the freedom from reoperation over time, and the proportional hazard assumption was tested for each covariate with generalized CoxSnell residuals and -ln[-ln(survival)] probabilities.
Cox proportional hazards models were developed by incorporating variables that had a P-value of 0.10 or less on univariate analysis; in order to account for positive or negative confounding, no automated model selection procedure was used and all reported variables, unless collinear with a Spearman's rank correlation coefficient
0.30 and a P-value <0.005, were used simultaneously. When collinear covariates were identified, the one most readily interpretable was selected for entry in the final models. Non-significant variables used in the models are presented in the footnotes of Tables 3 and 4. Each model was evaluated with a score test and rejected if P
0.05.
|
|
| 3. Results |
|---|
|
|
|---|
3.1.2. Reoperation for structural valve deterioration
Fig. 1
displays the actual freedom from reoperation for SVD in patients with aortic (A) and mitral (B) bioprostheses, and separately depicts discontinued and current bioprosthesis models. In both aortic and mitral positions, current bioprostheses had significantly better durability than discontinued bioprostheses (odds-ratio of 15-year reoperation for SVD 0.11±0.04; 95% CI 0.06, 0.23; P<0.001 for current versus discontinued aortic bioprostheses, and 0.42±0.14; 95% CI 0.22, 0.80; P=0.009 for current versus discontinued mitral bioprostheses, respectively). Current bioprostheses were significantly more durable in the aortic position than in the mitral position (14.3±6.8% more freedom from 15-year reoperation for SVD; 95% CI 1.1, 27.7%; P=0.018, current aortic versus mitral bioprostheses). For example, in 91 patients at risk 10 years after AVR with the Medtronic Hancock II aortic prosthesis, 95.8% were free (94.8±3.2% by actuarial methods) from reoperation due to SVD. There was no significant difference in the durability of discontinued bioprosthesis between aortic and mitral implant positions.
|
3.2.2. Age
Older age was protective against reoperation for SVD after bioprosthetic AVR and MVR, with a hazard ratio of 0.97 and 0.98 per year increase in age, respectively.
3.2.3. Body surface area
A larger body surface area (BSA) was associated with a lower freedom from reoperation for SVD after AVR (hazard ratio 1.84 per m2 increase), but not after MVR.
3.2.4. Coronary artery disease
Concomitant coronary artery disease was protective against reoperation in patients undergoing bioprosthetic MVR. The hazard ratio was 0.35.
3.2.5. Smoking
Smoking was an independent risk factor for reoperation due to SVD after AVR as well as after MVR. The hazard ratios were 2.58 and 1.78, respectively.
3.2.6. Left ventricular hypertrophy
Persistent left ventricular hypertrophy at follow-up was associated with a significantly higher incidence of subsequent reoperation for SVD in patients with aortic bioprostheses (hazard ratio 2.38).
3.2.7. Prosthesis size
In patients with aortic bioprostheses, larger prosthesis size significantly predicted an increased freedom from reoperation due to SVD. Both a linear effect with respect to manufacturer prosthesis size (hazard ratio 0.82 per increase of one valve size, independent of BSA), and a dichotomous effect of the ratio of prosthesis size over BSA at or below the 10th percentile value of the cohort were observed (hazard ratio 1.79 for patients with prosthesis size/BSA ratio smaller or equal to the 10th percentile value of the cohort). These relationships were not observed in patients with mitral bioprostheses.
| 4. Discussion |
|---|
|
|
|---|
The study also identified a relationship between persistent left ventricular hypertrophy determined by objective ECG criteria at follow-up after AVR, and a higher incidence of reoperation for aortic SVD. Although the mechanistic basis of this relationship is unclear, it could relate to the coprevalence of left ventricular hypertrophy in patients with smaller aortic prostheses, in itself a risk factor for reoperation, or to the coprevalence of left ventricular hypertrophy with hypertension, which may cause chronically increased diastolic closure stress on the bioprosthesis.
BSA also correlated with the risk of reoperation for aortic SVD. This is in contrast with a previous study that showed an inverse relationship between body mass index and native aortic valve stenosis (although valve disease-related weight loss could have confounded this relationship) [12], and could potentially be explained by higher hemodynamic stresses on the prosthesis in patients with an elevated BSA or by their lower tolerance to the effects of stenosis or regurgitation as the bioprosthesis progressively fails.
Coexistent coronary artery disease at the time of bioprosthetic MVR resulted in a significantly increased freedom from reoperation in this study. This has been identified by other authors previously [10], and likely relates to competing risks of death in patients with concomitant mitral and coronary disease rather than to a true protective biological effect of coronary disease on the development of SVD.
4.1. Previous related work
Structural deterioration of a bioprosthesis is the leading cause and most frequent indication for reoperation in patients with tissue valves [1316]. Bioprosthetic valves are known to undergo a time-dependent process of structural deterioration secondary to stress-related tears, perforations or dystrophic calcifications [17]. Independent risk factors previously found to be associated with bioprosthetic SVD include younger age, mitral valve position [1823], renal insufficiency [21], and poor ventricular function [3], although no relationship between left ventricular function and freedom from reoperation was demonstrated in the present study. Larger valve size had previously been suspected to be a possible predictor of SVD [6,24], but other studies have found the opposite to be true, with small valve size being associated with increased calcification [7] and stenotic failure [8].
4.2. Limitations
The findings of this cohort study may not necessarily be generalizable to all patients with bioprosthetic valves, because the study represents a single institution's experience and may have been affected by regional referral and patient management patterns. In addition, unsuspected demographic and selection factors unique to the cohort may have resulted in overfitted statistical associations. Finally, generalizability to current bioprostheses may be in part limited by the fact that most reoperations for SVD occurred in older generation bioprostheses, despite this factor having been accounted for in the analyses.
No prospective data was available from the present cohort on serum cholesterol level and statin treatment status. Evidence suggests that hypercholesterolemia plays a role in the progression of native aortic stenosis [11,25], and recent findings unpublished as of this writing have identified a potential benefit for statins in preventing reoperation in patients with bioprosthetic valves. The impact of serum creatinine and renal failure could also not be evaluated, as too few dialysis-dependent patients had, until recently, been implanted with bioprosthetic valves at our institution.
Because this study was observational it is possible, despite the use of a hard end-point like reoperation, that clinical factors or selection bias may have influenced results, such as in those patients for whom reoperation might have been delayed because of patient preferences or perceived surgical risk. In addition, although the effect of multiple potential confounders was forced into the statistical models, it remains possible that unknown confounders may ultimately have impacted on some findings.
4.3. Conclusions
Despite the aforementioned limitations, this study identified several risk factors for reoperation due to structural valve dysfunction in patients with aortic and mitral bioprostheses. A major, potentially modifiable factor shown to be of detrimental impact in both aortic and mitral implant positions was smoking. In addition, the study revealed that larger aortic bioprosthesis size was associated with an increased freedom from late reoperation for SVD, therefore constituting an additional impetus to implant as large a valve as possible in all patients undergoing bioprosthetic AVR. Other risk factors for reoperation due to SVD included younger age, male gender, BSA, and persistent left ventricular hypertrophy for patients with aortic bioprostheses, and younger age and coronary artery disease for patients with mitral bioprostheses. The study also formally determined that current models of bioprostheses have markedly better durability than older models that are now discontinued, suggesting that a revision process on the recommended minimum age guidelines for the implantation of tissue versus mechanical prostheses may be indicated. Overall, the results of this study may provide useful information for clinicians and patients in determining what prosthesis class may be best suited for a given patient, how the prosthesis may fare with respect to reoperation for SVD, and how some risk factors for reoperation due to SVD may potentially be avoided.
| Acknowledgments |
|---|
| Footnotes |
|---|
| Appendix A. Conference discussion |
|---|
|
|
|---|
Dr Ruel: No, we did not. The relationship we observed with aortic bioprostheses was not present with mitral bioprostheses; there was not even a trend.
Dr S. Bolling (Ann Arbor, Michigan): Were you able to analyze how many of these patients were on statins? There is sort of a trend in some of the literature now to analyze reoperative rates, valve degeneration rates on the basis of statins, and even statin dose in some of these patients. If this is a modern series we should be able to analyze which patients were on statins and does that have an influence on valve degeneration.
Dr Ruel: You are absolutely right that both native and prosthetic valve literatures now suggest that hypercholesterolemia may have a detrimental effect on valve deterioration. Unfortunately we do not have this data available.
Dr K. Wrobel (Krakow, Poland): Did you analyze the rate of urgency reoperation compared between the groups and did you try to identify any risk factors for urgent reoperation in both groups?
Dr Ruel: No, we did not look at that. As one would expect, the vast majority of those reoperations were electively performed for progressive structural valve deterioration. There were a few urgent reoperations, especially with discontinued models like the Ionescu-Shiley whose failure could happen very suddenly, but in the recent part of the series there were quite few reoperations which were performed urgently.
Dr E. Al-Mashat (Baghdad, Iraq): Have you tried to correlate the degeneration with the possibility of the etiologic cause for the pathology? It has been our observation in Iraq that no matter how good the bioprosthesis is, it gets degenerated quicker, perhaps our main pathology being rheumatic heart disease and age of our patients is much younger than yours. What is your comment on this?
Dr Ruel: It is an interesting hypothesis that perhaps some preformed antibodies as the result of rheumatic fever may play a role in bioprosthetic valve degeneration, but we have not been in a position to look at this in our series.
| References |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |