|
|
||||||||
a Department of Cardiothoracic Surgery, Division of Cardiology, New York University School of Medicine, New York, New York
b Department of Medicine, Division of Cardiology, New York University School of Medicine, New York, New York
Accepted for publication May 1, 2007.
* Address correspondence to Dr Grossi, New York University Medical Center, 530 First Ave, Ste 9V, New York, NY 10016 (Email: grossi{at}cv.med.nyu.edu).
Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29–31, 2007.
| The authors disclose a financial relationship with Medtronic Inc. and Estech Corp.
|
| Abstract |
|---|
|
|
|---|
Methods: From January 1996 through March 2006, 731 patients with EuroSCOREs of 7 or higher underwent isolated AVR. In this cohort, 313 (42.8%) were septuagenarians, 322 (44.0%) were octogenarians or nonagenarians, 233 (31.9%) had had previous cardiac procedures, 237 (32.4%) had atheromatous aortas, and 127 (17.4%) had cerebrovascular disease. A minimally invasive approach was used in 469 (64.2%). Data collection was prospective. Long-term survival was computed from the Social Security Death Benefit Index.
Results: The mean EuroSCORE was 9.7 (median, 10), and the mean logistic EuroSCORE was 17.2%. Actual hospital mortality was 7.8% (57 of 731). Multivariate analysis showed ejection fraction of less than 0.30 (p = 0.002; odds ratio [OR], 3.13), chronic obstructive pulmonary disease (p = 0.019; OR, 2.14), and peripheral vascular disease (p = 0.048; OR, 2.13) were significant predictors of hospital mortality. Complication(s) occurred in 73 patients (9.9%). Freedom from all-cause death (including hospital mortality) was 72.4% at 5 years (n = 152). Age (p < 0.001), previous cardiac operations (p < 0.014; OR, 1.51), renal failure (p < 0.002; OR, 2.37), and chronic obstructive pulmonary disease (p < 0.007; OR, 1.30) were predictors of worse survival.
Conclusions: Logistic EuroSCORE greatly overpredicts mortality in these patients. Five-year survival is good, unlike suggestions from earlier EuroSCORE analyses. This raises concern about unknown long-term percutaneous prosthesis function. Clinical trials for these patients must include randomized surgical controls and have long-term end points.
It has been speculated that nearly one-third of patients with severe, symptomatic, single-valve heart disease who could benefit from valve replacement do not undergo surgical repair [1]. Most of these patients are deemed too high risk for an operation secondary to comorbidities such as advanced age and the presence of chronic obstructive pulmonary disease (COPD) or renal failure, or both [1]. With increasing life expectancy in the Western world [2], the size of this untreated cohort of high-risk patients is expected to increase.
This has fueled interest in the development of percutaneous approaches to aortic valve replacement (AVR) as an alternative treatment option for patients with perceived high operative risks. Percutaneous aortic valve replacement (PAVR) trials are currently ongoing in patients whose surgical risk is deemed excessive due to comorbidities. The European System for Cardiac Operative Risk Evaluation (EuroSCORE) is currently used as a surrogate for operative mortality in the evaluation of the relative mortality benefits of patients treated with PAVR compared with surgical repair in these trials [3–5].
The EuroSCORE model is a risk-stratification system developed from a database of adult cardiac patients from several European countries to predict in-hospital mortality after cardiac surgical procedures [6]. Since its introduction in 1999, the EuroSCORE model has been validated by databases in Europe [7, 8], North America [9], and Japan [10]. It has also been shown to be a predictor of length of intensive care unit stay [11], cost [11], mid-term outcome [12], and long-term outcome [13] of open heart surgery. In patients undergoing heart valve procedures, the EuroSCORE model has been shown to be predictive of early mortality [14], postoperative complications [15], prolonged length of stay [15], and long-term mortality [16].
The use of EuroSCORE in predicting operative mortality for high-risk patients undergoing isolated AVRs has yet to be validated. Although the logistic EuroSCORE model has been shown to be a better predictor of mortality compared with the additive EuroSCORE in high-risk populations [17, 18], several studies have found that the logistic EuroSCORE model may overestimate the mortality of such patients undergoing valve procedures [19, 20]. This study, therefore, has two aims: (1) to evaluate perioperative and long-term outcomes of isolated AVR in patients designated as "high risk" by EuroSCORE evaluation, and (2) to determine the predictive value of standard and logistic EuroSCORE in such patients.
| Patients and Methods |
|---|
|
|
|---|
Data Collection
Patient data were collected prospectively. The definitions used for preoperative risk factors and perioperative complications were those used by the New York State Cardiac Surgery Reporting System (NYS CSRS).
Hospital mortality was defined as death at any time before discharge from the hospital. Long-term survival was computed from the United States Social Security Death Benefit Index. Statistical analysis was performed using SPSS 15 software (SPSS Inc, Chicago, IL). Categoric variables were analyzed by the
2 test, and odds ratios (OR) were calculated. Predictors of time-related mortality were analyzed by Cox regression analysis, and hazard ratios (HR) for significant covariants were calculated. A value of p < 0.05 was considered to be significant. Life-table analysis was used to calculate late survival curves.
| Results |
|---|
|
|
|---|
|
|
|
|
|
|
| Comment |
|---|
|
|
|---|
However, studies looking at octogenarians undergoing valvular surgery have demonstrated that the additive and logistic EuroSCORE models overestimate the mortality in such patients [19, 20], suggesting that the EuroSCORE may not be well calibrated for high-risk patients undergoing valvular procedures. This current report tests the calibration of EuroSCORE in a cohort of high-risk patients undergoing isolated AVR. We found that the EuroSCORE is a flawed metric that greatly overestimates the hospital mortality in this patient population and that the overall outcome and long-term survival of these patients were good.
The operative mortality of 7.8% reported in this series is comparable with that in other series of high-risk patients undergoing valve procedures [20, 30–35]. The overall mortality rate for patients undergoing cardiac operations has been decreasing during the past decade, particularly for higher-risk patients [10]. This may be due to improvements in surgical technology and perioperative critical care. Our study also showed that the long-term survival for high-risk patients undergoing isolated AVR is good, with 72.4% of the patients surviving 5 years after the operation. Thus, isolated AVR can be performed with relatively low operative risk and good long-term outcome in high-risk patients.
The development of PAVRs may offer promising treatment options for patients with severe aortic stenosis who are deemed inoperable owing to comorbid conditions. Cumulative data obtained from recent trials show that current PAVR technology has a successful implantation rate of only 80% and a mortality rate of 19% in patients after successful implantation [3, 4, 36]. Most important, PAVR must therefore be compared with conventional open-heart AVR. The data presented indicate that the EuroSCORE grossly overestimates the surgical mortality of high-risk patients undergoing isolated AVR. It is therefore not an accurate standard to select for "high-risk patients" or to determine the survival benefit of PAVR compared with conventional surgery.
Limitations
Although the data were collected prospectively, this was done using the standardized methods of the NYS CSRS database. The EuroSCORE calculations and definitions vs those used by the NYS CSRS are shown in the Appendix. Definitions of most risk factors were similar between the EuroSCORE and the NYS CSRS database; however, there were some exceptions. For example, pulmonary hypertension is a variable in the EuroSCORE model, but it is not reported in the NYS CSRS database. Such differences may result in an underscoring or underestimation of the EuroSCORE as used in this article. This would only further support our finding that the EuroSCORE model grossly overestimates surgical mortality of high-risk patients undergoing isolated AVR.
|
| Discussion |
|---|
|
|
|---|
DR GROSSI: The nonsternotomy approach, in approximately 90% of the patients, was a second or third right anterior thoracotomy approach with traditional cardiopulmonary bypass, and retrograde coronary sinus cardioplegia. Ten percent of the patients had an upper ministernotomy typically teeing off into the right-hand side.
DR MICHAEL A. BORGER (Leipzig, Germany): I must mention a point about your introduction. You quoted a 20% mortality rate for percutaneous aortic valves. However, we just published this month the biggest experience to date with transapical aortic valves, with a 30-day mortality rate of only 7%. My colleague, Tommy Walter, performed five of these procedures last week, with only one patient requiring an ICU [intensive care unit] bed. The other 4 went directly to the step-down unit. So it is a very impressive procedure, once you get the kinks worked out of the system.
My question is regarding different risk scoring systems. I believe Todd Dewey from Dallas has a paper on this subject at the upcoming AATS [American Association for Thoracic Surgery]: STS [Society of Thoracic Surgeons] Risk Scoring versus EuroSCORE Risk Scoring. Several people have observed that the EuroSCORE overestimates mortality risk in high-risk patients, similar to your observations. Should we be using STS risk scoring system or some other model?
DR GROSSI: No, I think we should have randomized surgical controls with long-term end points. That is the message.
DR BORGER: But you need to somehow identify your patients prior to enrollment in such a trial, correct? So what risk-scoring system should we use for inclusion and exclusion criteria? Or do you think that we should enroll both high- and low-risk patients?
DR GROSSI: No, I would think, to be fair, you are going to want to include high-risk patients. The point is you can cull the patients to be randomized, but you have to have a randomized, controlled experiment with long-term end points to really understand what we are getting with this new technology. If three-quarters of these patients are going to be alive at 5 years, we are going to have to be very sure about what we are putting into them.
DR MICHAEL E. JESSEN (Dallas, TX): Dr Grossi, do you have any idea why the EuroSCORE seems to make the prediction too high?
DR GROSSI: We were very conservative in our approach because New York State does not include some of the risk factors. We have gone through it in detail in the manuscript.
I think it is validated on all cardiac procedures, but it is not validated in particular patient cohorts, such as very elderly with serious comorbidity.
DR OCTAVIO E. PAJARO (Jacksonville, FL): You used an expected-to-observed ratio to decide that the EuroSCORE overpredicted the mortality. Did you try going back to the original study and actually doing another statistic, comparing the probability ratios at all?
DR GROSSI: No. We just looked at the absolute prediction and what reality was. We felt that was good enough to show the model. And it has to raise questions about, as we are testing this, what our plans are going to be.
DR PAJARO: I only bring that up because your group still had an 8% mortality, which most people would still consider pretty high. So would there be some number that you would feel more comfortable that this is too high?
DR GROSSI: Youre saying 8% in patients who 50% of them are 80 or 90 years old and a third of them had previous cardiac surgery?
DR PAJARO: Exactly. Somewhere, somebody chose: Well operate. And not included in the group is the one we turned down, or the group turned down. So was that the criteria you would use then, high risk and no surgeon would operate?
DR GROSSI: No, I dont think thats a fair trial of the technology either. I am not against the technology. I am just saying for us as surgeons to understand what the technology can do for our patients, we have to do it in a way that makes sense. And the only way to make sense is to have controlled trials.
We are going to have to look at long-term outcomes. Perhaps, it is not scary for us. If we just think about the three papers we just heard, I dont think we can shortchange this patient population and say we are not going to be concerned about their long-term outcomes.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
N. Al-Attar, D. Himbert, F. Descoutures, B. Iung, R. Raffoul, D. Messika-Zeitoun, E. Brochet, F. Francis, H. Ibrahim, A. Vahanian, et al. Transcatheter Aortic Valve Implantation: Selection Strategy Is Crucial for Outcome. Ann. Thorac. Surg., June 1, 2009; 87(6): 1757 - 1763. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Zajarias and A. G. Cribier Outcomes and Safety of Percutaneous Aortic Valve Replacement J. Am. Coll. Cardiol., May 19, 2009; 53(20): 1829 - 1836. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. H. Rahimtoola The Year in Valvular Heart Disease J. Am. Coll. Cardiol., May 19, 2009; 53(20): 1894 - 1908. [Full Text] [PDF] |
||||
![]() |
S. Bleiziffer, H. Ruge, D. Mazzitelli, C. Schreiber, A. Hutter, J.-C. Laborde, R. Bauernschmitt, and R. Lange Results of percutaneous and transapical transcatheter aortic valve implantation performed by a surgical team Eur. J. Cardiothorac. Surg., April 1, 2009; 35(4): 615 - 621. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. Piazza, Y. Onuma, P. de Jaegere, and P. W. Serruys Guidelines for reporting mortality and morbidity after cardiac valve interventions--need for a reappraisal? Ann. Thorac. Surg., February 1, 2009; 87(2): 357 - 358. [Full Text] [PDF] |
||||
![]() |
N. Khaladj, M. Shrestha, S. Peterss, I. Kutschka, M. Strueber, L. Hoy, A. Haverich, and C. Hagl Isolated surgical aortic valve replacement after previous coronary artery bypass grafting with patent grafts: is this old-fashioned technique obsolete? Eur. J. Cardiothorac. Surg., February 1, 2009; 35(2): 260 - 264. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. Walther, J. Kempfert, and M. A. Borger Reply to Apostolakis et al. Eur. J. Cardiothorac. Surg., February 1, 2009; 35(2): 377 - 378. [Full Text] [PDF] |
||||
![]() |
E. E. Apostolakis, K. Akinosoglou, and D. Dougenis On-pump or off-pump transapical aortic valve implantation provides better clinical outcomes? Eur. J. Cardiothorac. Surg., February 1, 2009; 35(2): 376 - 377. [Full Text] [PDF] |
||||
![]() |
V. H. Thourani, R. Myung, P. Kilgo, K. Thompson, J. D. Puskas, O. M. Lattouf, W. A. Cooper, J. D. Vega, E. P. Chen, and R. A. Guyton Long-Term Outcomes After Isolated Aortic Valve Replacement in Octogenarians: A Modern Perspective Ann. Thorac. Surg., November 1, 2008; 86(5): 1458 - 1465. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. J. Antunes Percutaneous aortic valve implantation. The demise of classical aortic valve replacement? Eur. Heart J., June 1, 2008; 29(11): 1339 - 1341. [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 |