| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 1995;59:462-468
© 1995 The Society of Thoracic Surgeons
Sections of Cardiovascular Surgery and Biostatistics, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
Accepted for publication October 14, 1994.
 |
Abstract |
|---|
 TopFootnotes Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
0.003), older age (73 years; p 0.02), and preoperative atrial fibrillation or paced rhythm (p 0.01). This pattern was similar for both transient ischemic events and strokes and rapidly decreased over the first few months of the first year and the first few years of the 12-year follow-up. These patients were not routinely anticoagulated. Although, in general, patients receiving a bioprosthesis in the aortic position do not require anticoagulants, a subset of patients have been identified who should receive short-term anticoagulation in an attempt to reduce the high early incidence of neurologic events. |
Introduction |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Freedom from events are reported to vary both by valve position and model [1]. A limitation of these reports is their duration of follow-up, in that overall prosthetic freedom from events approach equity for both mechanical prostheses and bioprostheses when viewed over a sufficient period of time. Despite this apparent interchangeable quality, there are situations more favorable for one prosthesis over another, especially in consideration of late events.
Bioprosthetics are valuable alternatives for patients unable to comply with or tolerate the regimentation necessary for long-term systemic anticoagulation [6]. Original thinking with bioprostheses virtually eliminated the need of any systemic anticoagulation, relying on the ``non''-thrombogenic nature of the bioprosthesis. Some recent reports found lower survival rates in patients after aortic valve replacement with a bioprosthesis when the patients were receiving warfarin anticoagulants [4, 16]. The current review identifies a window of time whereby patients are at an accelerated risk for neurologic events in the early postoperative period such that a period of anticoagulation may reduce these thromboembolic events.
|
Material and Methods |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Survival and survivorship free of late events was calculated using the Kaplan-Meier method [19]. Survival and survivorship free of late events was originally calculated both including and excluding 30-day mortalities, but because the relationships of survival to risk factors were similar, all survivals presented in this paper include patients who died within 30 postoperative days. For survivorship free of late events, patients without a late event at the time of a reoperation unrelated to their aortic valve were censored at the time of reoperation. Univariate analysis of risk factors were tested for significance using log-rank tests [20]. The Cox proportional hazards model [21] was used to evaluate whether continuous variables such as age or ejection fraction were significantly related to survival or survivorship free of late events. Continuous distributions were divided at their median to graphically demonstrate the effect of continuous variables on survival.
The estimated exponential yearly death rates and event rates (number of patients with events/person-years of exposure) also are presented for various subgroups of patients.
Values of p less than 0.05 were considered to be statistically significant.
|
Results |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
The influence of specific valve pathology, aortic stenosis, regurgitation, or mixed disease did not affect late survival or events. The influence of previous aortic valve operation was examined, and there was no effect on late survival following a reoperation for AVR.
Actuarial survival including 30-day mortality at 5 and 10 years was 74.8 ± 2.4 and 52.9 ± 4.9 years for all AVR patients, 81.8 ± 2.8 and 63.0 ± 5.6 years for AVR alone, and 63.0 ± 4.4 and 47.4 ± 6.8 years for AVR + CABG, respectively (Fig 1
). The linearized risk of death was 6.3%/patient-year and includes 30-day mortality. Perioperative patient risk factors that adversely affected late survival (any death or cardiac death) were advanced (IV) New York Heart Association (NYHA) class (Fig 2), age of 73 years or older (Fig 3), the addition of CABG or other associated procedures (or both) (Fig 4), and a compromised ejection fraction (0.54; p 0.001) (Fig 5). The numeric age and ejection fraction above were based on the median value derived from the overall group of patients.
|
|
|
|
|
0.005). When both CABG and an associated procedure (14.2%/patient year) were included, the risk of late death was significantly increased over AVR alone, AVR with CABG, or AVR with associated procedures (p 0.002) (see Fig 4
).
Sixty-seven patients (11.9%) sustained a cardiac death (3.4%/patient year). The 5- and 10-year freedom from cardiac death was 85.8% ± 1.9% and 76.6% ± 3.6%, respectively. The risk of late cardiac death was increased in patients 73 years of age or older (p 0.001), with an ejection fraction of 0.54 or less (p 0.001), and with the operative addition of associated procedures (p 0.001). The effect on late cardiac death by advanced (IV) NYHA class did not reach statistical significance (p 0.09).
Endocarditis developed in 13 patients (0.77%/patient year). Three patients needed reoperation for control of their infective process. At 5 and 10 years, the freedom from endocarditis was 95.5% ± 1.4% and 95.5% ± 1.4%, respectively.
Twenty-nine patients came to reoperation for all causes. The freedom from any subsequent cardiac reoperation for all patients was 91.9% ± 1.7% and 85.4% ± 3.6% at 5 and 10 years, respectively (1.4%/patient year). Fourteen patients (48%) required AVR at the time of reoperation secondary to the following: a periprosthetic leak (6), valve failure (4), endocarditis (3), and explantation due to stenosis caused by thrombus in the valve sinuses and extending from the sewing ring into the left main coronary orifice (1). The incidence of reoperation did not change when separate cohorts (AVR ± CABG) were considered, although any reoperation involving repair or replacement of the aortic prosthesis was more likely in the younger ( 73 years) patients (p 0.023). The 30-day mortality rate for patients coming to reoperation after implantation of the Carpentier-Edwards valve was 15%. It is of interest that during the follow-up period, 6 additional patients (mean age, 70.6 years) required an abdominal aortic operation for aneurysm.
Five patients came to reoperation for primary tissue failure (0.23%/patient year). Three were in their sixties at the time of valve implantation, and the valve pathology in 2 was leaflet tear or perforation without calcification. The third was a patient who presented within 3 weeks of operation with recurrent symptoms of aortic stenosis and at reoperation had organizing thrombus on the sewing ring obstructing the orifice of the left main coronary orifice. There was also organizing thrombus in the sinuses of the prosthesis. The final 2 patients were 16 and 18 years of age at implantation, and one valve failed with calcification and stenosis and the second with a tear and mild calcification. No patient died at the time of reoperation. The mean time from implantation to primary tissue failure in the 4 patients was 73 months (range, 55–88 months). The 5- and 10-year freedom from primary tissue valve failure in all patients was 98.9% ± 0.8% and 93.6% ± 3.1%, respectively. None of the valve failures occurred in the AVR + CABG group, and no patient older than 70 years had valve failure. Since the original review was completed, a second patient was encountered who had subacute thrombosis of their aortic porcine prosthesis. This latter patient also came to reoperation and was successfully managed with re-replacement of the aortic valve. Neither of the patients with the thrombosed valve had received warfarin anticoagulation at the time of hospital dismissal from their first operation.
Episodes that may cause mild lightheadedness or dizziness (TIA) but do not leave or lead to definite neurologic deficits were very difficult to interpret. Our review was inclusive, and any complaint that was not otherwise convincingly explained was considered to be valve-related. Many patients related prior historic events from memory, and precise clinical features of these events were left to speculation; thus, we believed that inclusion of these would construct a worst case senario. The hazard curves are parallel, but the TIA volume exceeded in total number those of strokes and reinforced our inclusion of these ``questionable'' events. The 5- and 10-year freedom from TIAs was 70.1% ± 2.8% and 48.2% ± 6.3%, respectively. A pattern of incidence of TIAs was noted whereby a hazard function of events depicted a high early occurrence with a definite and progressive reduction in the incidence of TIAs. This pattern of 12 months of the first year was reproduced during the 12 years of follow-up (Fig 6).
|
demonstrates the early increased incidence of major neurologic events. This graph depicts the higher rate of stroke in the first few months and during the first few years after operation for AVR and shows a similar pattern to that of patients having a TIA. The majority of events occurred early and diminished steadily thereafter, with only isolated events (7 major events between years 5 and 12). More importantly, the incidence of stroke was statistically increased in patients over the age of 73 years (p 0.02), in patients with a depressed ejection fraction (0.54; p 0.003), and in patients coming to operation in atrial fibrillation or a paced rhythm (p 0.01). The influence of preoperative atrial fibrillation on stroke occurrence was most important in patients having both AVR and CABG (p 0.001), but it was not statistically important in those patients having isolated AVR (p 0.45). Long-term anticoagulation in patients with an aortic bioprosthesis was not the routine therapeutic policy at our institution. Of all patients having a major event during the first year, only 12% were dismissed from hospital on a regimen of warfarin. Of the 27 patients who sustained a major postoperative neurologic event, 66% had preoperative atrial fibrillation or a paced rhythm or had postoperative low cardiac output.
|
0.11). This is in contrast to the presence of preoperative atrial fibrillation and paced rhythms mentioned above. The freedom from all major events (death from all causes, endocarditis, any reoperation, and major neurologic complications) at 5 and 10 years was 67.9% ± 2.6% and 42.4% ± 5.1%, respectively. The linear rate of these events was 8.8%/patient year.
|
Comment |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
The mortality figures are comparable with other reports which quote a 30-day mortality of 2.1% [17] to 19% [22]. A noticeable contrast is the older mean age of our patient population, which is a reflection of current practice and a bias toward implanting bioprostheses in older-age patients. It has been shown that the bioprosthesis is preferred in patients 70 years of age or older and in those patients unable to tolerate warfarin anticoagulation [6, 17]. It has been reported [9, 11, 12, 15] and confirmed in our study that late survival falls when CABG is added to the operative procedure (47% versus 63% at 10 years) and is reflective of the severity of the patient's heart disease and the impact of additional procedures at the time of operation. Our figures are quite similar to those reported by Cohn and colleagues [17] and Revuelta and Duran [15], which demonstrated that the addition of CABG or other associated procedures to the operation significantly reduces late survival from 71% to 58% at 9 years. In that report, the mean age of their patients was 60.6 years. There are reports that address the safety of aortic valve replacement on patients over the age of 70 and 80 years [16]. These reports acknowledge the increased risk but also point out the improved long-term survival postoperatively in patients with severe aortic valve disease. Our risk factors affecting late survival (advanced NYHA class, advanced age, reduced ejection fraction, and the addition of associated procedures) are strikingly similar to other reports in the literature that also identify these as a negative influence. The low operative mortality (4.8%) attests to improved current operative techniques and postoperative management despite the older age of the patients. It is also apparent that the number or extent of associated procedures has a cumulative determination on late survival (see Fig 4). It may be necessary to recognize the occasional need to prioritize the components of an operation, and if possible, accept a mild residual aspect of the pathology without correcting it completely.
We did not identify any difference in survival or subsequent events based on gender. Borkon and co-workers [14] identified male gender as an independent risk factor for hospital mortality in a small comparison of mechanical (n = 68) and bioprosthetic (n = 73) valves in 141 patients greater than 70 years of age. Lytle and associates [16], on the other hand, found female patients to be at increased risk for in-hospital mortality in their sample of 500 patients undergoing combined AVR and CABG. The discrepancy may be due to differences in sample size or to the added factor of different prostheses within the same series. In our study, which included both patients older than 70 years and patients with combined procedures, there was no apparent gender difference in the operative or late results for AVR ± CABG.
The risk of endocarditis demonstrated in our review (0.77%/patient year) is near the lower limits of the range reported in the literature (0.5% – 1.6%/patient year). The majority did not come to operation for treatment of their disease.
Our incidence of reoperation is comparable with other studies. The freedom from any cardiac reoperation (CABG) at 10 years was 85.4% ± 3.5%, whereas the freedom from subsequent reoperation on the aortic prosthesis for any reason (perivalvular leak, etc) was 90.9% ± 3.2% and freedom from primary tissue valve failure at 10 years was 93.6% ± 3.1%. The mortality rate for reoperation in our patients was 15% and is comparable with reports in the literature with reoperative mortality rates of 11% [4] to 16% [13].
Our experience with the Carpentier-Edwards bioprosthesis in relation to primary tissue failure has been excellent. All 5 of the patients requiring valve replacement were less than 70 years of age, and no patient more than 70 years of age came to reoperation for tissue valve failure. This experience parallels that of others with their results of low primary tissue failure with the Carpentier-Edwards porcine valve [1, 3, 23], especially in the patients over the age of 70 years [6, 24]. The reliable identification and reporting of neurologic events can be difficult and confusing. It may be impossible to accurately identify an episode as a TIA rather than as hypoglycemic- or medication-induced. These episodes are better, but not infallibly distinguished, while the patient is still in the hospital following operation, but it becomes nearly impossible to verify them weeks or months after the event by examination, much less by letter or phone in the late follow-up period. All follow-up was coordinated and recorded by a registered nurse and reviewed by one of us (T.A.O.). Although the legitimate criticism can be made that the episodes were not neurologic in origin, we believe strongly that the event-free rate is accurate. The thorough screening of these events identifies the lowest event-free rate, and we acknowledge that it is possible that the event-free rate may in fact be better, but we are confident it is not worse. There is little in the literature to compare events less than major events or strokes. The pattern is similar, if not identical, to that of major events or stroke (see Figs 6 and 7), albeit on a different scale. It is not possible to say that the TIAs are related to or are a precursor to a major neurologic event because only 7.6% of patients with any neurologic event sustained both a TIA and a stroke, and they occurred in varying order. Their similarity in rate and distribution signify a truly vulnerable time in the early postoperative period.
Added to the early incidence of TIAs is the parallel hazard of major neurologic events or strokes. The incidence of major events or stroke found in our review (0.77%) is within the range of postoperative stroke rate quoted in the literature (0.5 to 3.9%) [6, 14, 17, 24]. The majority of our patients were not receiving warfarin anticoagulation at the time of dismissal from hospital or at the time of follow-up. We were able to identify a pattern of events whereby the incidence of events is greatest in the first few months and in the first few years after operation (see Fig 7). This pattern has been described by Bernal and colleagues [5] in a smaller number of patients [16]. Our patients were not routinely anticoagulated; anticoagulation varied by surgeon from warfarin anticoagulation to no anticoagulation or platelet inhibitors in the postoperative period, although patients having AVR + CABG were dismissed on antiplatelet therapy. The combination of the early incidence of neurologic events and the isolated but potentially catastrophic occurrence of thrombotic obstruction would lend support to an interim period of anticoagulation. This may be tailored to specific patients in that our review demonstrated that patients with preoperative low ejection fractions, atrial arrhythmias or paced rhythms, or advanced heart class or low cardiac output states are in greater jeopardy of an early postoperative neurologic event and would benefit from anticoagulation.
The impact of atrial fibrillation on the incidence of thromboemboli has been alluded to in the literature [14, 22, 23, 24]. The reports by Borkon and colleagues [14, 22] identified preoperative atrial fibrillation as an independent risk factor for valve-related mortality, but no correlation was found for morbidity. Louagie and coworkers [24] found an increased rate of thromboemboli (3.8%/patient year) in older patients having valve replacement and attributed some of the risk to advanced age, advanced atherosclerotic disease, and the presence of atrial fibrillation in the postoperative period. This correlation is weak due to the fact that there were patients receiving AVR and mitral valve replacement in this report, and it is not clear if any of the AVR patients had the arrhythmia and only those patients with preoperative atrial fibrillation received warfarin. Our current report examined the influence of atrial fibrillation and the presence of a permanent pacemaker as a predictive marker for postoperative stroke. It appears that the preoperative presence of these rhythm disturbances bears more concern than does the postoperative occurrence. The explanation may be based on the duration of these arrhythmias, as those in the postoperative period may only have been transient, lasting for a few hours or days before conversion. The frequency of atrial fibrillation in the postoperative period of 20% to 30% in this patient group is without much comparison in the literature but does not seem inordinate for this population. The lack of influence of this postoperative arrhythmia alleviates some of the concern for patients with transient atrial fibrillation such that the need for warfarin anticoagulation for transient postoperative atrial fibrillation is reduced, especially in this elderly patient population.
The report by Bernal and associates [5] and our current data vary somewhat from a previous report of Lytle and colleagues [16]. In that report, a strong correlation was observed between patients receiving warfarin anticoagulation after AVR and a reduced long-term survival. It was not possible to identify precisely the causes of death in their report, and hemorrhage and emboli could not specifically be distinguished. There was no information regarding the timing of the neurologic events that occurred for purposes of constructing a hazard function of events. It is important that the patients receiving bioprostheses were all anticoagulated for 6 weeks after the operation; however, the adequacy of anticoagulation was not documented.
It does not appear that permanent or long-term warfarin anticoagulation is necessary for patients with bioprostheses in the aortic position. With the high rate of early events in our patient group and others [5], we would recommend early postoperative warfarin anticoagulation for a specific period of time (2 to 3 months). The older age population may not tolerate this due to their associated frailties; however, our data identify a high-risk group that should benefit and should have aggressive in-hospital and early postoperative anticoagulation with warfarin and platelet inhibition. The predictive increased risks of advancing age, reduced ejection fraction, NYHA class IV, and preoperative atrial fibrillation or a paced rhythm in those patients in our series who sustained a major neurologic event help focus the indications for anticoagulant therapy in a more logical algorithm. The absence of these potential markers would support antiplatelet therapy alone or no anticoagulant in a candidate with a hemorrhagic tendency who has recently undergone an aortic valve replacement with a bioprosthesis.
|
Acknowledgments |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
We wish to acknowledge the excellent effort of Allen Kunselman for data programming and Darcey Reinartz for manuscript organization and preparation, without whose help the project could not have been completed.
|
Footnotes |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopFootnotesAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Colli, J.-P. Verhoye, A. Leguerrier, and T. Gherli Anticoagulation or antiplatelet therapy of bioprosthetic heart valves recipients: an unresolved issue Eur. J. Cardiothorac. Surg., April 1, 2007; 31(4): 573 - 577. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Nowell, E. Wilton, H. Markus, and M. Jahangiri Antithrombotic therapy following bioprosthetic aortic valve replacement Eur. J. Cardiothorac. Surg., April 1, 2007; 31(4): 578 - 585. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Lund and M. Bland Risk-corrected impact of mechanical versus bioprosthetic valves on long-term mortality after aortic valve replacement J. Thorac. Cardiovasc. Surg., July 1, 2006; 132(1): 20 - 26. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J P A Puvimanasinghe, J J M Takkenberg, M B Edwards, M J C Eijkemans, E W Steyerberg, L A van Herwerden, K M Taylor, G L Grunkemeier, J D F Habbema, and A J J C Bogers Comparison of outcomes after aortic valve replacement with a mechanical valve or a bioprosthesis using microsimulation Heart, October 1, 2004; 90(10): 1172 - 1178. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. N. Salem, P. D. Stein, A. Al-Ahmad, H. I. Bussey, D. Horstkotte, N. Miller, and S. G. Pauker Antithrombotic Therapy in Valvular Heart Disease--Native and Prosthetic: The Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy Chest, September 1, 2004; 126(3_suppl): 457S - 482S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Gherli, A. Colli, C. Fragnito, F. Nicolini, B. Borrello, S. Saccani, R. D'Amico, and C. Beghi Comparing Warfarin With Aspirin After Biological Aortic Valve Replacement: A Prospective Study Circulation, August 3, 2004; 110(5): 496 - 500. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Ruel, R. G. Masters, F. D. Rubens, P. J. Bedard, A. L. Pipe, W. G. Goldstein, P. J. Hendry, and T. G. Mesana Late incidence and determinants of stroke after aortic and mitral valve replacement Ann. Thorac. Surg., July 1, 2004; 78(1): 77 - 83. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. A. Orszulak Invited commentary Ann. Thorac. Surg., October 1, 2002; 74(4): 1113 - 1114. [Full Text] [PDF]
|
|
|
|
|
|
T. A. Orszulak, H. V. Schaff, F. J. Puga, G. K. Danielson, C. J. Mullany, B. J. Anderson, and D. M. Ilstrup Event Status of the Starr-Edwards Aortic Valve to 20 Years: A Benchmark for Comparison Ann. Thorac. Surg., March 1, 1997; 63(3): 620 - 626. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
