| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2005;129:1024-1031
© 2005 The American Association for Thoracic Surgery
Surgery for Acquired Cardiovascular Disease |
Division of Cardiovascular Surgery, Mayo Clinic, Rochester, Minn
Read at the Thirtieth Annual Meeting of The Western Thoracic Surgical Association, Maui, Hawaii, June 23–26, 2004.
Received for publication June 21, 2004; revisions received November 18, 2004; accepted for publication November 23, 2004. * Address for reprints: Thoralf M. Sundt III, MD, Division of Cardiovascular Surgery, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (E-mail: sundt.thoralf{at}mayo.edu).
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionDiscussionReferences |
|---|
METHODS: Between 1993 and 2000, 1151 patients underwent bioprosthetic aortic valve replacement with (641) or without (510) associated coronary artery bypass. By surgeon preference, 624 had early postoperative anticoagulation (AC+) and 527 did not (AC–). In the AC– group, 410 patients (78%) received antiplatelet therapy. Groups were similar with respect to gender (female, 36% AC+ vs 40% AC–, P = .21), hypertension (64% AC+ vs 61%, P = .27), and prior stroke (7.6% AC+ vs 8.5% AC–, P = .54). The AC+ group was slightly younger than the AC– group (median, 76 years vs 78 years, P = .006).
RESULTS: Operative mortality was 4.1% with 43 (3.7%) cerebrovascular events within 90 days. Excluding 18 deficits apparent upon emergence from anesthesia, we found that postoperative cerebrovascular accident occurred in 2.4% of AC+ and 1.9% AC– patients. By multivariable analysis, the only predictor of operative mortality was hypertension (P < .0001). Postoperative cerebrovascular accident was unrelated to warfarin use (P = .32). The incidence of mediastinal bleeding requiring reexploration was similar (5.0% vs 7.4%), as were other bleeding complications in the first 90 days (1.1% vs 0.8%). No variables were predictive of bleeding by multivariate analysis.
CONCLUSIONS: Although these data do not address the role of antiplatelet agents, early anticoagulation with warfarin after bioprosthetic aortic valve replacement did not appear to protect against neurologic events.
Early anticoagulant therapy is intended to provide protection from thromboembolic events while the cloth sewing ring is endothelialized. The recommended duration of warfarin therapy may be viewed as brief; unfortunately, the risks of anticoagulant-related hemorrhage are highest during the first 3 months of therapy.6 The recommended strategy, therefore, subjects those patients we are most reluctant to anticoagulate to a significant risk of complications.
Despite the published guidelines, opinion and practice with regard to anticoagulation in the first 90 days after bioprosthetic valve replacement remains divided. Indeed, the aforementioned guidelines themselves are somewhat inconsistent, with the AHA/ACC panel judging early anticoagulation a class I indication, whereas the authors of the ACCP guidelines grade the evidence for this to be only 2C. Furthermore, the British Society of Hematologists recommends only antiplatelet therapy after biological AVR.6
The practice in our own institution reflects this diversity of opinion. After publication of a Mayo Clinic study advocating intravenous heparin starting the day after surgery followed by warfarin therapy for 3 months,7 some surgeons in our group adopted this approach, whereas others did not. This diversity in practice pattern, therefore, provides 2 groups of patients for retrospective analysis of the impact of warfarin therapy on cerebral ischemic events during the first 90 days after bioprosthetic AVR.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
Follow-up was obtained between September 6, 2002, and December 1, 2003. Hospital charts of all patients were reviewed and abstracted for in-hospital events and any subsequent events noted in the hospital record. In-hospital follow-up was 100% complete. Ninety-day follow-up was not routinely collected prospectively. Accordingly, postal questionnaires were mailed to patients alive at the time of follow-up and telephone calls made to those not responding to the postal questionnaire. Information regarding anticoagulant use and thromboembolic or other neurologic events out to 90 days was obtained through one of these means in 87% of patients.
All suspected cerebral ischemic events were reviewed and adjudicated by one of the authors (T.M.S.) blinded to the treatment group. Postoperative delirium was not scored as stroke. Patients in whom a deficit developed after they had awakened intact from anesthesia were scored as having a postoperative event, while those awakening with a deficit or those never regaining consciousness postoperatively were scored as having intraoperative events. Because of the realities of clinical practice, a strict definition of thromboembolic stroke was frequently not possible from the hospital record. In some cases, the consulting neurologist speculated with regard to etiology, distinguishing embolic from hypertensive or hypotensive infarction, but this was not uniform. Accordingly, in the interest of an inclusive data set, either a clinical diagnosis of stroke or evidence of cerebral infarction by computed tomographic or magnetic resonance imaging was accepted as evidence of a cerebral ischemic event. There was only 1 reversible cerebral ischemic event identified within the 90-day interval. No peripheral embolic events were identified.
Members of the Division of Cardiovascular Surgery, Mayo Clinic Rochester, performed all operations. Mild hypothermic or normothermic cardiopulmonary bypass was routine. Cardioplegic arrest was achieved by blood or crystalloid cardioplegia, per the individual surgeon’s preference. The prosthesis type inserted was also at the discretion of the operating surgeon.
Postoperative management was at the discretion of the individual surgeon. Some members of the division routinely anticoagulated patients, beginning intravenous heparin within 24 hours of surgery and converting to warfarin before discharge from the hospital. Others were treated with aspirin only or no anticoagulant at all. Of patients not receiving warfarin, 78% received antiplatelet therapy. Patients experiencing atrial fibrillation were routinely administered anticoagulants before discharge. Patients were allocated to AC+ (early postoperative anticoagulation with heparin and warfarin) or AC– (no anticoagulation) groups on the basis of medications at the time of discharge from the hospital or at the time of in-hospital stroke, if such occurred.
Statistical analysis
Categorical factors were compared between groups by using Fisher exact tests. For groups with severity classifications, such as aortic valve regurgitation, the P value is based on an absent versus present response. Logistic regression models were used to construct a multivariate model to predict operative mortality. A stepwise selection technique was used to identify factors for the final multivariate model. Potential risk factors for long-term endpoints such as stroke outside the intraoperative period were evaluated by the Cox proportional hazards models. Age was adjusted for 10-year intervals.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
|
|
|
|
Univariate analysis of risk factors for operative mortality yielded only hypertension as a predictor (Table 5). No risk factors for postoperative stroke within 90 days, including history of atrial fibrillation, atrial fibrillation at he time of surgery, ejection fraction, or age, were apparent (Table 6) This may be due in part to the small number of events recorded. Postoperative warfarin did not appear protective. Similarly, there was no apparent benefit to aspirin, although the majority of patients were on this medication. As shown in Table 7, no risk factors achieved significance for postoperative bleeding.
|
|
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
Our findings are consonant with those of other investigators. In a retrospective review of patients after bioprosthetic AVR, Moinuddeen and associates9 observed no difference in incidence of stroke risk between 109 patients receiving heparin followed by 3 months of warfarin, compared with 76 patients receiving no anticoagulation. In their study, as in ours, the bleeding risk was no higher in the warfarin group. In a review of patients receiving no anticoagulation after tissue AVR, Babin-Ebell and colleagues10 observed a stroke rate of only 1.75% in the first 6 months among 54 patients. No control group was provided; however, this rate compares favorably with other series in which early anticoagulation was employed.1 These results are consistent with Turpie’s11 observations that a less-intensive anticoagulant regimen was equally effective and, in fact, safer for patients with tissue valves. Most recently, Gherli and associates12 have conducted a prospective randomized study of anticoagulation after bioprosthetic AVR. In their study, there was no difference in cerebral ischemic events between a group of 141 patients receiving warfarin for 3 months and 108 receiving only aspirin. Again, there was also no difference in major bleeding events between the groups.
The results of our current study are at variance with AHA/ACC and ACCP recommendations, as well as with a previous report from our own institution.7 This study, published by Heras and colleagues,7 provides much of the support for the guidelines as written. In that study, the authors identified a significant risk of neurologic events during the first 3 months after bioprosthetic valve replacement. The overall risk was higher after mitral than AVR and was particularly high in the first 10 days postoperatively. In this interval, there were 5 strokes among 424 patients undergoing bioprosthetic AVR, all of which occurred among patients not receiving anticoagulation. The risk fell precipitously thereafter to a linearized rate of 3.6% per year between 10 and 90 days, and 1.9% per year thereafter. The authors concluded that early anticoagulation to an international normalized ratio of 3.0 to 4.5 may be advisable. They did not, however, distinguish between prostheses in the mitral and aortic positions. Curiously, the protective effect of anticoagulation was only slight after mitral valve replacement and was not seen at all among double valve replacement patients. Furthermore, while approximately one third of the patients were given anticoagulants postoperatively, in two thirds of those cases the prothrombin time was less than 1.3 seconds during the high-risk first 10 days when an apparent protective effect was observed.
The explanation for the differences in outcome between our current study and the previous report from our own institution is a matter of speculation. Variations in follow-up are always a consideration, although information for the earliest postoperative interval—when the greatest risk of stroke was identified—should be comparable. Differences in the criteria for stroke could play a role as well, although we attempted to use as inclusive a definition as possible. The most tempting explanation is that the liberal use of platelet inhibitors in current clinical practice has reduced these early events. We support this approach and routinely administer aspirin after bioprosthetic AVR.
Our current study suffers from significant weaknesses to be sure. The retrospective design makes "intention to treat" analysis impossible, complicating the interpretation of data for the AC+ and AC– groups. The subtle effects of selection bias may also impact any nonrandomized study. Practice patterns during the study interval were relatively consistent within any individual surgeon’s practice, however, and referral for AVR within our institution is relatively evenly spread among surgeons. While referral is not random, it is certainly not based on anticoagulation philosophy. Surgeons with larger and smaller total case volumes were also represented in both groups. Follow-up was incomplete as well. Unfortunately, because of current government regulations related to privacy, it is unlikely that more complete information can be obtained in a retrospective study such as this.
As in other retrospective analyses, imprecise definition of the etiology of neurologic events is a difficulty in this study. Review of the hospital records showed that neurologic consultation was not uniform. The choice of imaging studies varied, and the assessment of embolic versus nonembolic etiology was not routinely documented. Subtle events may have been missed. The attribution of some events to the intraoperative period carries the potential for error as well. In contrast to the observations made in the setting of coronary bypass surgery by Hogue and colleagues,13,14 a great many of the cerebrovascular accidents observed in the current study appeared to be intraoperative. In their analysis, 65% of perioperative strokes were delayed. This difference is likely due to the open versus closed nature of the procedures, the potential for embolization of debris from the valve annulus, and the lower incidence of generalized atherosclerosis among patients undergoing AVR.
Finally, despite being the largest study in the literature, our sample size remains far too small to definitively answer the question posed. The data supporting early anticoagulation are remarkably sparse as well, however. The authors of the AHA/ACC guidelines reference only the Heras7 paper in support of early anticoagulation. In their consensus report for the ACCP, Stein and colleagues4 cited the event rate of 1.8% without anticoagulation reported by Babin-Ebel10 (54 patients), compared with Turpie’s11 report of no events with heparin and warfarin (109 patients). Although the event rates differ, the exceedingly small numbers of patients in both series make conclusions tenuous at best.
The results presented here may be fairly considered a test of the hypothesis generated by the earlier work from our institution. On the basis of that work, some surgeons in the group were stimulated to pursue an aggressive approach to anticoagulation. The results obtained do not support the notion that early anticoagulation after bioprosthetic AVR reduced stroke risk. Conversely, there was no increased risk of complications. We believe that one can reasonably adopt a posture of equipoise on the matter. The results of this study further suggest that efforts to reduce the risk of cerebral events after AVR might be best directed toward the intraoperative period. Recently introduced devices intended to capture15 or deflect16,17 debris liberated from the aorta or valve intraoperatively may find particular application in this regard.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
I applaud the study group for directing our attention to the guidelines and questioning its validity. I must admit that our cardiac unit does not believe all patients receiving bioprosthetic aortic valves should be anticoagulated. Unfortunately, given the presented findings of an event incidence of 2% to 7% with a 1.2% difference in stroke and 2% difference in bleeding, this study like all those previously published is insufficiently powered to answer the question is early anticoagulation necessary after bioprosthetic AVR. A clinical trial of more than 40,000 patients would be required to answer this question. I have the following questions.
Is your nonanticoagulated group sicker than the anticoagulated group due to the fact that the anticoagulated group had a lower pump run and a higher operative mortality?
My second question refers to the percentage of anticoagulated patients receiving heparin prior to oral anticoagulation.
Finally, did your target INR vary depending on the valve brand used?
Dr Sundt. Thanks very much for those questions.
The first one, are the 2 groups really different, is obviously a difficult one and that is the Achilles heel in any retrospective study. You do the best you can looking back in the charts to see if they are comparable or not. As I said in the presentation, I think that the reason for the difference in the operative mortality is to some degree the patients that were particularly ill and died early on never made it to being anticoagulated. That is, however, only conjecture. I think the difference in the pump run is frankly a difference in which surgeons practice which approach. One of the surgeons who is particularly expeditious in his procedures is one of the ones who does not anticoagulate patients routinely. Ultimately I agree it is an almost unanswerable question, are the 2 groups different? Despite this, I was still struck by how similar the distribution of events was in the 2 groups.
I respect the criticism that the study cannot be adequately powered to reverse the current guidelines. I would argue, however, that it is at least as strong as the studies that support the guidelines. We are really left once again with the logical but unproven argument that we should administer anticoagulants until the sewing ring is endothelialized. I think that I would agree that this paper cannot answer the question any more rigorously than the other papers that claim one should anticoagulate.
As for the percentage of patients receiving heparin, I do not know that. We could go back and look at the charts for that. I can tell you, however, that this is the routine for the surgeons who do anticoagulate. This was quite striking to me. I had not seen that done before. I think the majority of those patients were treated with heparin early on and we could go back and look at that.
Did the target INR vary by prosthetic device? No, it didn’t. The target INR was similar regardless of the manufacturer of the bioprosthetic valve.
Dr Chen. Were the valves studied pretty consistent across your entire group, that is, was it one particular valve used or multiple valves used?
Dr Sundt. Multiple valves were used over the course of that time period, so I suppose that is another potential confounding variable if there is a difference in thromboembolic risk among the valves. Perhaps there are differences in how much Dacron is on the prostheses, so I suppose that is a possibility. I would not have expected that, but I suppose that’s a possibility.
Dr Vaughn Starnes (Los Angeles, Calif). It is a very timely topic, and I am glad you brought it to this meeting. We are finding our valve patients are very much older than they were 5 or 10 years ago. I think a lot of us have adopted this policy just because of the aged population we are operating on. Do you have any recommendations in terms: Is there age criteria that you are more prone to do this protocol versus a younger age patient, for example?
Dr Sundt. I actually do not anticoagulate early, and I think that this supports that practice. I think as the population ages it is more supportive of the notion of not anticoagulating people. That’s really what my practice has been.
Dr Royce Calhoun (Sacramento, Calif). One thing to me that was just as impressive, as you pointed out, that the complication rate (ie, bleeding) is no higher, which I think is encouraging, so I guess one question for you and anyone that may know this, given that is sort of the official guideline that are often cited and where I trained we didn’t use routine warfarin as well.
Dr Sundt. You did not.
Dr Calhoun. Did not, no. What is this situation if we do that and do not use warfarin and then somebody has a stroke and is litigious? Is that something that we have to worry about defending against? Do you know that?
Dr Sundt. Presumably, and that was part of the driver for doing this. It would be difficult to defend, I suppose, and these guidelines have enormous impact, enormous medicolegal implications. That is a factor as The Society of Thoracic Surgeons comes out with their practice guidelines as well. I think that the guidelines can really drive your practice and it is important to have good data for those.
Dr Calhoun. What is your protocol for mitral valves?
Dr Sundt. For mitral valves, I do exactly what the guidelines suggest actually, which is I anticoagulate them with warfarin. I have not used early heparin. In fact, at Mayo there is a tendency to use early heparin with mechanical or bioprosthetic valves, and I was accustomed to waiting until we were 5 days out, and, if we were not therapeutic, then starting heparin, but a more aggressive approach is used by some. I was surprised to see no more bleeding episodes than this.
Dr Chen. I’d like to address the issue of the guidelines. Because of this paper, I did sit down and read the entire 35 pages of this thing, and I must say that this is worth reading. The guidelines do specifically state that these are guidelines. These are not rules that are meant to be followed. They do allow room for individual patient treatment, and so I would refer to that, to say that these are the recommendations, but it is not an absolute rule.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Dunning, M. Versteegh, A. Fabbri, A. Pavie, P. Kolh, U. Lockowandt, S. A.M. Nashef, and on behalf of the EACTS Audit and Guidelines Commit Guideline on antiplatelet and anticoagulation management in cardiac surgery. Eur. J. Cardiothorac. Surg., July 1, 2008; 34(1): 73 - 92. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. N. Salem, P. T. O'Gara, C. Madias, and S. G. Pauker Valvular and Structural Heart Disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition) Chest, June 1, 2008; 133(6_suppl): 593S - 629S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Brueck, W. Kramer, P. Vogt, N. Steinert, P. Roth, G. Gorlach, M. Schonburg, and M. C. Heidt Antiplatelet therapy early after bioprosthetic aortic valve replacement is unnecessary in patients without thromboembolic risk factors Eur. J. Cardiothorac. Surg., July 1, 2007; 32(1): 108 - 112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. di Marco, S. Grendene, G. Feltrin, D. Meneghetti, and G. Gerosa Antiplatelet therapy in patients receiving aortic bioprostheses: A report of clinical and instrumental safety J. Thorac. Cardiovasc. Surg., June 1, 2007; 133(6): 1597 - 1603. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 W.R. E. Jamieson, S. D. Moffatt-Bruce, P. Skarsgard, M. A. Hadi, J. Ye, G. J. Fradet, J. G. Abel, M. T. Janusz, A. Cheung, and E. Germann Early Antithrombotic Therapy for Aortic Valve Bioprostheses: Is There an Indication for Routine Use? Ann. Thorac. Surg., February 1, 2007; 83(2): 549 - 557. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. El-Husseiny, K. Salhiyyah, S. G. Raja, and J. Dunning Should warfarin be routinely prescribed for the first three months after a bioprosthetic valve replacement? Interactive CardioVascular and Thoracic Surgery, October 1, 2006; 5(5): 616 - 623. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. di Marco, M. Giordan, and G. Gerosa Early antithrombotic therapy after aortic valve replacement with tissue valves: When the practice diverges from the guidelines J. Thorac. Cardiovasc. Surg., June 1, 2006; 131(6): 1223 - 1223. [Full Text] [PDF]
|
|
|
|
 |
|
 S. H. Rahimtoola The Year in Valvular Heart Disease J. Am. Coll. Cardiol., January 17, 2006; 47(2): 427 - 439. [Full Text] [PDF]
|
|
|
|
|
|
F. di Marco, G. Meneghetti, and G. Gerosa Early anticoagulation after aortic valve replacement with bioprostheses: Time to abandon it? J. Thorac. Cardiovasc. Surg., November 1, 2005; 130(5): 1482 - 1483. [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 |
