| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2006;30:706-713
© 2006 Elsevier Science NL
Department of Cardiac Surgery, University Hospital of Ghent, 5K12, De Pintelaan 185, 9000 Ghent, Belgium
Received 28 April 2006; received in revised form 3 July 2006; accepted 17 July 2006.
* Corresponding author. Tel.: +32 9 2403925; fax: +32 9 2403882. (Email: Thierry.bove{at}Ugent.be).
 |
Abstract |
|---|
 Top Abstract 1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
0.85 cm2/m2), and severe preoperative left ventricular (LV) hypertrophy (LVMI > 180 g/m2) affected survival adversely. But multivariate analysis determined only age, NYHA class IV and excessive LV hypertrophy as independent predictors of late mortality. Stented and stentless xenografts were equally effective in terms of transprosthetic gradients and LVMI regression. The use of a stentless valve significantly reduced the occurrence of PPM (18% vs 41%, p
< 0.01). Early LVMI regression at 1 year was optimized by the avoidance of PPM, indicated by a higher absolute (43.7 ± 28.3 g/m2 vs 58.6 ± 33.8 g/m2, p
= 0.003) and relative (25.0 ± 12.7% vs 31.4 ± 14.9%, p
= 0.004) mass regression. However, late LV remodeling was predominantly affected by systemic hypertension and severe preoperative LV hypertrophy, resulting in the incomplete LVMI resolution in 61.3% and 66.7% of these patients, respectively. Conclusion: In elderly patients, aortic valve replacement appears to be equally effective with a stentless or stented bioprosthesis. Midterm clinical outcome is mainly determined by patient-related factors such as age, advanced NYHA class, and severity of preoperative LV hypertrophy. Regarding post-AVR left ventricular remodeling, patient-prosthesis mismatch influences the early phase, whereas arterial hypertension affects the late regression more. However, the left ventricular remodeling is continuously compromised by the preoperative presence of excessive hypertrophy, despite the efficacy of the aortic valve replacement.
Key Words: Stentless prosthesis • Aortic valve replacement • Left ventricular mass regression
|
1. Introduction |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
In order to maximize the effective orifice area after aortic bioprosthesis implantation, the stentless xenograft has been favored because of a superior hemodynamical profile, suggesting an enhanced LV mass regression and better survival [3–5]. Additionally, a lot of interest has been addressed to the concept of patient-prosthesis mismatch (PPM) as a factor that could potentially jeopardize the regression of LV hypertrophy despite effective aortic valve replacement [6,7]. The clinical impact of this issue is, however, still controversial.
Otherwise, some authors focusing particularly on the long-term effect of aortic valve replacement have revealed that regression of the left ventricular mass (LVM) seemed to be significantly influenced more by patient-related factors than by prosthesis selection. Lund et al. [8] found a poorer outcome up to 10 years after AVR in patients with excessive preoperative LV hypertrophy, probably indicating irreversible myocardial damage, although the hemodynamical performance of the aortic prosthesis was excellent. Also, arterial hypertension might further compromise the efficacy of AVR on late LV mass regression [8,9].
The aim of this study was to evaluate the midterm clinical outcome of AVR in elderly patients, by comparing stented and stentless aortic valves. Likewise, the functional results were investigated by analyzing the patient- and prosthesis-related factors that might affect the midterm evolution of LV mass.
|
2. Patients and methods |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
|
The peak and mean valve gradients were calculated using the Bernoulli equation with correction for subvalvular velocities. The left ventricular mass calculation is based on the corrected ASE formula as follows:
|
|
The LV systolic function is assessed by means of the ejection fraction using the Simpson's rule.
Postoperative residual LV hypertrophy has been defined as an LVMI higher than 134 g/m2 in male patients and an LVMI of more than 110 g/m2 in female patients, according to the criteria of the Framingham study [11]. To evaluate the influence of higher degrees of preoperative LV hypertrophy on mortality, a cut-off value of preoperative LVMI has been determined on the basis of the first value of the superior quartile of the LVMI distribution in both sexes. According to this calculation, severe preoperative LV hypertrophy is considered for an LVMI > 184 g/m2 in men, and for an LVMI > 172 g/m2 in women in this study.
Early regression of LVMI after AVR is analyzed by calculation of absolute and relative LVMI regression indices. Absolute LVMI regression represents the maximal LVMI loss between the preoperative value and the lowest LVMI within the first postoperative year. Relative LVMI regression is calculated as the rate of absolute LVMI regression related to the preoperative LVMI, expressed as a percentage of LVMI loss.
The estimation of patient-prosthesis mismatch is based on the projected indexed EOA as proposed by Pibarot et al., using the in vitro EOA of both types of prostheses indexed to BSA. Moderate PPM is considered as an IEOA of 0.85 cm2/m2 or less, and severe PPM as an IEOA inferior to 0.65 cm2/m2 [12].
2.3 Patient management and follow-up
Aortic valve replacement was performed with standard techniques using moderate hypothermic cardiopulmonary bypass and cold crystalloid cardioplegia. Implantation of the stented prosthesis was done in supra-annular position with interrupted sutures. The technique of the stentless xenograft insertion has been described previously [13]. No annular enlarging procedure was added to cope with a small aortic annulus.
Postoperatively, coumadine anticoagulation was administered routinely for 3 months, and subsequently replaced by aspirin 160 mg/day, unless the patient had chronic atrial fibrillation.
Patient follow-up was completed by reviewing clinical outpatient records, including the echocardiographic results. Data concerning late death were confirmed by contact with the referring cardiologist. Follow-up was completed in 96% of the patients treated with a stentless prosthesis and in 91% of the patients with a stented prosthesis. The mean follow-up was 47 months (12–136 months) for the CE valve, yielding 987 patient-years. The mean follow-up for the Toronto valve was 56 months (12–122 months), corresponding to 1519 patient-years.
2.4 Statistical analysis
Statistical analysis was carried out with SPSS 10.1 software for Windows (SPSS Inc., Chicago, IL, USA). Continuous data are expressed as mean ± SD. Categorical data are presented as frequencies and percentages. Baseline comparisons between groups were performed with unpaired t-test for continuous data and with 
2-test or Fisher's exact test for categorical data.
Survival was analyzed using the Kaplan–Meier product-limit estimation. Investigation of preoperative and operative variables for association with mortality was performed by univariate analysis using the log-rank and Breslow indices. Multivariate analysis with the Cox's proportional regression model was done for identification of late death determinants. The interference of patient- and prosthesis-related variables with LVMI regression was examined by linear regression analysis. A p-value <0.05 was assumed as statistically significant.
|
3. Results |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
The mean size of the implanted valve during AVR was identical, i.e., 24.5 ± 2.2 mm for CE, 24.4 ± 2.3 mm for Toronto, with an analogous size distribution. However, the EOA and IEOA were significantly in favor of the stentless prosthesis 1.58 ± 0.22 cm2 and 0.89 ± 0.13 cm2/m2, respectively, for the CE valve versus 1.70 ± 0.32 cm2 and 0.98 ± 0.15 cm2/m2, respectively, for the Toronto valve (p = 0.0008 and p = 0.00001). According to the definition of PPM, moderate PPM was more prevalent in the CE group (41% vs 18%, p = 0.00005), but severe PPM was rare (2.7% for CE, 0% for Toronto, p = 0.11). The patients presenting PPM in both groups were commonly women (39.7% vs 13.2%, p = 0.00025). Their relative overweight (71.4 ± 13.3 kg in PPM vs 65.9 ± 11.8 kg in non-PPM, p = 0.003) balanced the difference in length (161.9 ± 8.1 cm in PPM vs 164.6 ± 8.2 cm in non-PPM, p = 0.018) to get a comparable BSA (1.79 ± 0.18 m2 in PPM vs 1.75 ± 0.18 m2 in non-PPM, p = 0.12). Additionally, smaller valve sizes were implanted (22.2 ± 1.6 mm vs 25.3 ± 1.8 mm, p < 0.000001) in PPM patients. However, there were no significant differences in other characteristics.
Concerning surgery, the choice of prosthesis was independent of the need for associated CABG (53.6% for CE and 46.2% for Toronto), but a stented valve was used preferably in valve reoperations (7.4% vs 1.4%, p = 0.005). The use of a stentless valve also resulted in a longer aortic cross-clamp time (57.6 ± 17.1 min for CE and 69.5 ± 14.3 min for Toronto, p = 0.001) as well as CPB time (112 ± 26 min for CE and 125 ± 24 min for Toronto, p = 0.0001). Whether PPM was introduced by the implanted prosthesis or not, this did not result in differences in operation-related variables.
3.2 Clinical follow-up
Early 30-day mortality was 7.9% after AVR with a CE valve and 4.8% after AVR with a Toronto valve (p
= 0.13). Early mortality was not affected by PPM, as operative death occurred in 7.04% of patients with PPM and 7.06% of patients without PPM (p
= 0.99). Additionally, 60% and 62% of these early deaths were cardiac non-valve-related in PPM and non-PPM patients, respectively.
Clinical improvement after successful AVR was independent of the prosthesis type and the presence of PPM (p = 0.54 and p = 0.21). The preoperative NYHA functional status improved from class III–IV in 82% to class I–II in 91% postoperatively in the CE group, and from class III–IV in 84% to class I–II in 93% in the Toronto group. In PPM patients, NYHA class evolved from status III–IV in 84% to status I–II in 89%, similarly to the NYHA class improvement from III–IV in 85% to class I–II in 95% in patients without PPM.
Within a follow-up duration of 12–122 months in the Toronto group, 42 patients (29%) died late, compared with 29 patients (29%) in the CE group for a follow-up period of 12–136 months. Late mortality was cardiac non-valve-related in 44.8% of CE patients versus 55.2% of Toronto patients (p = 0.30). None of the deaths was valve-related in the Toronto group, whereas one patient in the CE group died from a valve-related cerebral thrombo-embolic event. Actuarial survival at 5 and 8 years was 82.3% and 59.5%, respectively for the Toronto bioprosthesis, and 68.2% and 55.9%, respectively for the CE valve (Fig. 1 ). Only at 5 years, the survival was significantly in favor of a stentless valve (p < 0.001).
|
|
3.3 Echocardiographic results
Serial measurement of echocardiographic variables like LV end-diastolic and end-systolic dimensions, fractional shortening, and ejection fraction revealed no significant differences between both types of prostheses. Both peak and mean transprosthetic gradient decreased significantly after AVR (p
= 0.001), resulting in a significant early regression of LVMI (p
= 0.001)(Fig. 3a and b). Consequently, the CE valve and Toronto valve resulted in a comparable absolute (52.8 ± 21.2 g/m2 vs 54.8 ± 24.4 g/m2, p
= 0.69) and relative mass regression (28.3 ± 11.3% vs 30.7 ± 13.0%, p
= 0.49). However, beyond the third year after AVR, the LVMI tended to increase independently of the transvalvular gradients, which remained stable and suggested a normal prosthetic valve function. This phenomenon appeared in both groups, but more pronounced in the stented valve group, without reaching statistical significance (p
= 0.18).
|
|
|
|
|
4. Discussion |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
The concept of the stentless xenograft has been introduced to overcome the potential drawback of residual obstruction by the use of a stent-mounted valve. By providing a larger effective orifice area, the stentless bioprosthesis revealed a superior hemodynamical behavior, resulting in the clinical advantage in terms of functional improvement and survival [3–5]. In a case-controlled cohort study, David et al. demonstrated a superior actuarial survival of 91% for the Toronto SPV valve compared with 69% for the stented Hancock II valve at 8 years after AVR. This benefit was contributed to the hemodynamic superiority of these valves, and the subsequent favorable impact on the left ventricular mass regression [4]. In a similar study, Del Rizzo et al. [14] compared stentless and stented xenografts with particular interest for different age groups. They also found a survival gain in favor of the stentless freestyle valve at 5 years but only in the younger patients, similar to the patient population of David's study, who had a mean age of 62 years. However, in the group with patients aged more than 70 years, the prosthetic design became less important than the age itself with regard to the late survival. Survival analysis of our patient cohort revealed a 5-year survival of 82% for the stentless group and 68% for the stented group. This significant difference disappeared gradually over time, reaching a comparable rate of 59% and 56%, respectively, at 8 years. Although univariate analysis revealed the type of prosthesis as an independent predictor of late survival, this factor became insignificant by multivariate examination. But other variables like age, advanced preoperative functional and cardiac disability, and excessive preoperative LV hypertrophy seemed to be more determining. As the cause of late death was equally cardiac- and non-cardiac-related for both kind of prostheses, it is probably suggestive that age in itself acts as a stronger but ‘natural’ predictor of late survival in a specific elderly population presenting with valvular and often associated ischemic heart disease.
Recently, a lot of interest has been addressed to the interference between survival and patient-prosthesis mismatch. Most studies report on the detrimental effect of PPM on the early outcome after AVR [6,7]. Using the definition of a projected indexed effective orifice area, the Pibarot's group found that the risk of short-term mortality was increased 2.1-fold in presence of moderate PPM. Moreover, this risk increased substantially when the left ventricular function was depressed [7]. Also, in a study covering a follow-up of stented bioprostheses over 12 years, Rao et al. were able to show that PPM affected the operative mortality and the specific valve-related mortality. But regarding overall late mortality, age and NYHA class were the only multivariate predictors [6]. Conversely, some major studies could not demonstrate an adverse effect of PPM on late survival after AVR. Most of these reports revealed that patient-related factors like age, functional class, preoperative LV dysfunction, and associated coronary artery disease were influencing the late clinical result more than the valve size itself [15–17]. In our study, moderate PPM occurred in 38% of the stented bioprostheses and could significantly be reduced to 18% by the use of a stentless valve, a finding that was already revealed by Bach et al. [18].
Patient-prosthesis mismatch could not be identified as an independent, multivariate predictor of mid-term mortality. As these patients also showed the same functional recovery as the patients without mismatch, one could join the conclusion of David's editorial comment that PPM might be clinically overestimated [19], especially concerning the population of elderly and physically less active patients. According to others [6], the population presenting PPM in our series currently consisted of smaller and fatty women receiving a smaller sized prosthesis. However, patients with and without mismatch are clearly different populations by phenotype. We found an identical body surface area in both type of patients, but in PPM patients, the smaller length was largely compensated by the relative overweight. This per se should elicit the question whether the factor weight in the formula for calculation of body surface area might be probably too decisive, and subsequently might suggest a corrective intercession for each individual indexation of such specific patient-related variables.
Intermediate survival analysis of our cohort also revealed that excessive preoperative LV hypertrophy is an independent risk factor for mortality after effective AVR. More than one third of these patients died within the first postoperative year. Blais et al. [8] stated that the impact of PPM is more important during the early postoperative period, when the left ventricle is most vulnerable. However, the interaction with the degree of preoperative LV hypertrophy was not investigated in their study. Otherwise, Fuster et al. [20] evidenced a significant higher in-hospital mortality in patients with high LVMI promoted mutually by the coexistence of PPM, and the addition of a third risk factor as low LV ejection fraction even raised the operative mortality to 26.3%. In that study, the LVMI indicating severe preoperative LV hypertrophy was even higher than the cut-off point that we considered, i.e., 226 g/m2 in men and 216 g/m2 in women compared with 184 g/m2 and 172 g/m2 in our series. Although our operative mortality remained unaffected by a severely increased preoperative LVMI, the significance of this risk factor on the subsequent 1-year mortality is perhaps related to the incomplete LV mass regression as reflected by a residual LV hypertrophy in 70% of these patients at 1 year. Hence, this might underscore the hypothesis of persistent LV vulnerability beyond the early operative period, in view of probably already underlying irreversible myocardial damage despite the effective relief of the aortic stenosis by valve replacement.
As a result of the close relationship between the clinical outcome and the hemodynamics of the aortic valve prosthesis, the attention has moved to the impact of the functional performance of the prosthesis on the remodeling of the left ventricle after AVR. This time-dependent process is initiated by an early phase that is primarily achieved within the first year after the adequate relief of the aortic valve obstruction [21,22]. In this study, the functional analysis was focused on this early phenomenon by calculation of the absolute and relative LV mass regression as well as on the reflection of prosthesis- and patient-related factors on later LVM evolution.
Both stented and stentless bioprostheses in our study showed equally low transvalvular gradients and resulted in a significant and comparable early LVMI reduction. This largely corresponds with the data of two recent randomized comparisons between stentless and stented aortic valve replacements [23,24]. Although one study, comparing the Freestyle valve with the stented Mosaic prosthesis, emphasized the superior hemodynamical profile of the stentless valve by a benefit in indexed effective orifice area and peak flow velocity, the effect on the LVM regression within 1 year was identical [23]. Numerous studies, addressing the influence of PPM on the LVM regression, have come to conflicting results. Some authors found a negative correlation between PPM and/or small prostheses and a significant LV mass resolution after AVR, often secondary to persistent higher transvalvular gradients [6,7]. Others concluded that PPM is a rare phenomenon with negligible effect on the LVM regression [15]. However, the confounding variability in conclusions may be supported by the obvious discrepancies in the choice of parameter to identify PPM as well as in the use of this factor as a continuous or categorical variable.
We found that PPM and a high degree of preoperative LV hypertrophy significantly determined the early LVM resolution. The presence of even moderate PPM appeared to slow the rate of LVMI reduction but seemed not to hinder the extent of this process as a normalized absolute value of LVMI was finally attained. However, in this study, PPM was not correlated to higher transvalvular gradients, even if PPM was commonly the consequence of the implantation of smaller prostheses. Regarding the severely hypertrophic left ventricles, the early remodeling was proportional to the preoperative level, i.e., the regression was higher in patients with a high basal LVMI. The fact that normalization of the LVMI values was rarely obtained in these hypertrophic ventricles supports the theory of Lund et al. [25,8] that increasing hypertrophy involves an increased degree of irreversibility. Fuster et al. [20] observed that specifically in these excessively hypertrophic ventricles, PPM additionally impaired the early LVM regression resulting in residual hypertrophy in all cases, when both factors were present. The analysis of this combined effect on LVMI regression as well as on clinical outcome in our series was, however, not significant, probably because the small number of this specific patient subset flawed the statistical power.
Within the assumption that longevity after AVR is related to regression of LV hypertrophy, few studies have investigated the interference of determinants on the long-term changes of LV mass. At odds to the statement that hypertrophy regression is a continuing process over many years, Lund et al. found that no further change in LVMI took place after the initial major regression during the first 12 months post-AVR. Furthermore, he identified that a worse regression was more dependent on the preoperative risk profile than on the prosthetic valve function, and this result was further compromised by systemic hypertension [8]. According to his conclusion, we found that arterial hypertension predominantly influenced the late LV remodeling. Although the early regression was in favor of non-hypertensive patients, the effect of this factor became more apparent at 3 and 5 years after AVR. At both time intervals, the absolute LVMI reached even pathological values, partially sweeping the early gain of LVM reduction in spite of a perfect functioning aortic valve prosthesis. Finally, only 38.7% of these hypertensive patients developed an LVMI estimated as a normal value. A similar observation was done by Imanaka et al., namely that only 12.5% of the patients with hypertension returned to a normal range LVMI. Moreover, even the LVMI remained unchanged beyond the sixth month after AVR [9]. In their study, systemic hypertension was defined by the average systolic blood pressure value during at least three visits exceeding 130 mmHg. However, current experience has shown that blood pressure is highly labile in nature and often overshot at the time of the outpatient control because of several stress-related interferences. But we can agree with their findings that the late LVMI result is significantly influenced by two factors, i.e., the preoperative LVM index and postoperative hypertension, rather than by prosthesis-related variables.
|
5. Limitations of the study |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
Regarding the impact of prosthesis-related variables, especially PPM and valve size on functional behavior, only the hemodynamics at rest have been studied. But it is still questionable as to how much the results should have been altered by introducing exercise-related hemodynamics in this older and more sedentary population.
The LVM is commonly used to evaluate the functional efficacy of aortic valve replacement. The calculation of this variable includes the internal LV diameter, the interventricular septum, and posterior wall thickness. However, we have not differentiated the interacting variability between the posterior wall and septum thickness, especially for the severely hypertrophic ventricles. Therefore, it remains unanswered whether the more liberal addition of a septal myotomy might be beneficial to the involution of the LVMI after AVR when the detrimental effect of excessive preoperative LV hypertrophy on the outcome is taken into consideration.
Finally, the analysis of the impact of arterial hypertension on the LVMI evolution has been performed by using this factor as a categorical variable by preoperative definition. Subsequently, the development of new onset hypertension as well as the adequacy of the drug therapy has not been considered. However, we can accept that these elderly patients are probably less sensitive to the appropriate treatment because of less compliant medication intake and poorer age-related drug response.
|
6. Conclusion |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
Owing to the accelerated early left ventricular mass regression, the use of a stentless bioprosthesis appeared to be mainly indicated in patients with a high predictability of patient-prosthesis mismatch at the time of operation. This might be especially valuable in the presence of severe preoperative left ventricular hypertrophy, when the highly lethal character of this feature within the first year after AVR is recognized.
Two factors were identified as a major hindrance for further late left ventricular mass regression, resulting in a higher incidence of residual hypertrophy. Systemic hypertension has been found to inversely influence the late left ventricular remodeling. Secondly, the remodeling process appeared to be continuously compromised by the pre-existence of excessive left ventricular hypertrophy in spite of an effective aortic valve replacement, which should eventually elucidate the controversy on the surgical treatment of asymptomatic aortic stenosis associated with potentially irreversible and detrimental left ventricular hypertrophy.
|
References |
|---|
|
TopAbstract1. Introduction2. Patients and methods3. Results4. Discussion5. Limitations of the...6. ConclusionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Urso, R. Sadaba, and G. Aldamiz-Echevarria Is patient-prosthesis mismatch an independent risk factor for early and mid-term overall mortality in adult patients undergoing aortic valve replacement? Interactive CardioVascular and Thoracic Surgery, September 1, 2009; 9(3): 510 - 518. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Nozohoor, J. Nilsson, C. Luhrs, A. Roijer, and J. Sjogren Influence of Prosthesis-Patient Mismatch on Diastolic Heart Failure After Aortic Valve Replacement Ann. Thorac. Surg., April 1, 2008; 85(4): 1310 - 1317. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kallikourdis and S. Jacob Is a stentless aortic valve superior to conventional bioprosthetic valves for aortic valve replacement? Interactive CardioVascular and Thoracic Surgery, October 1, 2007; 6(5): 665 - 672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Antunes Editorial comment: Aortic stenosis in the elderly: rethinking strategies Eur. J. Cardiothorac. Surg., November 1, 2006; 30(5): 713 - 715. [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 |
