Early Postoperative Anticoagulation After Mechanical Valve Replacement: A Systematic Review

Transcription

1 Early Postoperative Anticoagulation After Mechanical Valve Replacement: A Systematic Review Alexander Kulik, MD, Fraser D. Rubens, MD, MS, Philip S. Wells, MD, MS, Clive Kearon, MB, PhD, Thierry G. Mesana, MD, PhD, Judith van Berkom, BA, and B-Khanh Lam, MD, MPH Division of Cardiac Surgery, University of Ottawa Heart Institute, Division of Clinical Hematology, Ottawa Hospital, Ottawa, and the Department of Medicine, McMaster University, Hamilton, Ontario, Canada The optimal approach to early postoperative anticoagulation after mechanical valve implantation remains controversial. This review article examines the pathogenesis of thrombus formation and the different strategies for early postoperative anticoagulation. The most commonly reported anticoagulation regimens had the after estimates of early postoperative thromboembolism and hemorrhage: oral anticoagulation alone (0.9%, 3.3%); oral anticoagulation with intravenous unfractionated heparin (1.1%, 7.2%); and oral anticoagulation with low molecular weight heparin (0.6%, 4.8%). Although intravenous heparin may be associated with a higher incidence of hemorrhage, a randomized trial is needed to provide the best evidence regarding early postoperative anticoagulation after mechanical valve implantation. Nearly four decades have passed since the first mechanical prosthetic valves were implanted. Frequent thromboembolic complications with the first mechanical valves led to recommendations of universal anticoagulation for these patients. Since then, several design changes and modifications have been made to improve the longevity, hemodynamics, and thrombogenicity of newer generation mechanical valves. With improved blood flow, less stasis, and less thrombogenic materials, lower rates of thromboembolism have been reported [1]. Despite these advances however, thromboembolism and anticoagulant-related bleeding continue to account for 75% of all complications after mechanical valve replacement [2]. Occurring most commonly within six months after implantation [2], these complications can adversely affect mortality and quality of life. Furthermore, the threat of their occurrence creates a psychological burden for each patient with a mechanical valve. The need for life-long anticoagulation in patients with mechanical valves is not in dispute, and the perioperative management of anticoagulation during non-cardiac surgery has been reviewed extensively [3]. However, the approach to early postoperative anticoagulation after mechanical valve implantation is still a matter of debate. The optimal intensity and timing of anticoagulation to prevent early thromboembolism after valve replacement surgery without postoperative bleeding complications is unknown. Hence, many anticoagulation protocols have been proposed, but a lack of consensus remains. The objectives of this study were (1) to reexamine the pathogenesis of thrombus formation and the need for anticoagulation; (2) to critically review the literature on early postoperative anticoagulation strategies; and (3) provide an estimate of the incidence of bleeding and thromboembolism for each approach to early postoperative anticoagulation. (Ann Thorac Surg 2006;81:770 81) 2006 by The Society of Thoracic Surgeons Material and Methods A computerized literature search for abstracts was performed in accordance with recommendations of the Cochrane collaboration using MEDLINE, EMBASE, CINAHL, BIOSIS, EICompendex, SIGLE, and the Cochrane Library from the earliest available date to June The initial keywords and MESH terms were perioperative care, perioperative, anticoagulants, heparin, warfarin, LMWH, low molecular weight heparin, aspirin, ASA, warfarin, coumadin, heart valve prosthesis, heart valve prosthesis implantation, heart valve replacement, Address correspondence to Dr Lam, University of Ottawa Heart Institute, 40 Ruskin Street, Suite H3404, Ottawa, Ontario, Canada K1Y 4W7; bklam@ottawaheart.ca. mechanical valve, mitral valve replacement, thrombosis, bleeding, hemorrhage, thromboembolism, thrombus, bleed, clot, and heart valves [surgery]. In addition, an Internet search was conducted, and sources for practice guidelines and gray literature were explored. A total of 244 citations were obtained, and relevant manuscripts were reviewed. The reference list of each pertinent article was assessed, and the Science Citation Index was examined for cited references of selected articles. Original manuscripts and review articles focusing on early postoperative anticoagulation after mechanical valve implantation were included for evaluation. No language restrictions were applied. Studies dealing with the perioperative anticoagulation of mechanical valves during noncardiac surgery were specifically excluded. The term 2006 by The Society of Thoracic Surgeons /06/$32.00 Published by Elsevier Inc doi: /j.athoracsur

2 Ann Thorac Surg REVIEW KULIK ET AL 2006;81: REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 771 early postoperative was defined as less than 30 days from valve implantation. As per the Guidelines for Reporting Morbidity and Mortality after Cardiac Valvular Operations, thromboembolism was defined as an embolic event that occurred in the absence of infection, after the immediate operative period, and a bleeding event was defined as an episode of major internal or external bleeding that caused death, permanent injury, or required transfusion [4]. Using the search strategy described above, it was evident that very few studies directly compared early postoperative anticoagulation strategies [5 9], and a meta-analysis was not possible because of the lack of data. In order to provide a qualitative comparison of anticoagulation strategies, an additional MEDLINE search was conducted to obtain data regarding the risk of thromboembolism and bleeding associated with each of the different approaches to early anticoagulation. This second search focused on randomized controlled trials, case series, and case-control studies reporting the results of the most common prosthetic mechanical valves currently available. The initial keywords and MESH terms were St. Jude Medical, Medtronic-Hall, Omnicarbon, Carbomedics, ATS valve, On-x valve, mechanical valve, anticoagulation, hemorrhage, bleeding, thromboembolism, and human. Manuscripts reporting the experience of mechanical valves in the aortic and/or mitral position in adult patients were included in this review. Series specifically focused on the outcomes of children or redo operations were excluded, and language restrictions were not applied. A total of 141 citations were obtained, and relevant manuscripts were reviewed by two authors (AK and BKL). The incidences of early postoperative bleeding and thromboembolism were extracted from each manuscript (absolute rates), as well as rates of perioperative mortality caused by hemorrhage or thromboembolism. Studies that did not clearly state the early postoperative anticoagulation strategy were excluded from further analysis. These data were summarized according to the three most common anticoagulation strategies to yield an approximate estimate of the risk of bleeding and embolism associated with each approach (Table 1). Results Mechanical Valve Thromboembolism and Anticoagulation PATHOGENESIS. The pathogenesis of prosthetic valve thromboembolism is a complex phenomenon, occurring through an interaction of a variety of prosthesis-related and patient-related factors. The pathologic events leading to thromboembolism begin immediately after surgery. Damaged perivalvular tissue and deposition of fibrinogen on the valve surface activate platelets as soon as blood starts flowing across the valve. This leads to immediate platelet adhesion and aggregation [1], and within 24 hours after surgery, platelet deposition on the Dacron sewing ring can be imaged radiographically [10]. However, these platelets are highly sensitive to the shear forces generated by the mechanical valve and are prone to continued activation and destruction. With continuing cycles of platelet aggregation, patients with mechanical valves have shortened platelet survival, a marker of thromboembolic risk [11]. Coagulation factors are also directly activated after valve implantation, leading to further clot formation as a result of the inherent thrombogenicity of the prosthetic material (suture material, Dacron sewing ring, struts, and hinge points) and sites of denuded tissue (valve excision site) [12]. Transprosthetic turbulent flow leads to regional increases in shear stress, structurally damaging the endocardium, causing a loss of local resistance to thrombosis. Recirculation areas on the outflow side of the prosthesis create flow stagnation, trapping damaged platelets and activated factors [13]. This provides an ideal milieu for thrombus formation and subsequent embolization. Thrombin can also be formed on platelet membranes after their activation, further promoting the organization and growth of platelet fibrin thrombus [14]. In sum, the particularly high thromboembolic risk that is present shortly after valve implantation involves both coagulation and platelet activation due to turbulent flow across the valve and the thrombogenicity of prosthetic material and denuded tissue. EARLY THROMBOEMBOLIC RISK AFTER MECHANICAL VALVE IMPLANTATION. The insertion of a large artificial device in contact with the bloodstream exposes the patient to a continual risk of valve thrombosis and embolism. This risk is proportional to the surface area of the foreign material, which is in contact with blood, making patients with mitral valve prostheses more prone to thromboembolic complications [11]. In general, a recent mechanical valve implantation is a strong risk factor for thromboembolic complications [15, 16], especially in the first three to six months after surgery [17, 18]. The reasons for this are threefold: first, the pathologic sequelae of the patients inherent valvular disease (atrial fibrillation, dilated left atrium, and dilated left ventricle) may predispose to areas of stasis and thrombus formation. Second, the increased thromboembolic risk early after mechanical valve implantation may reflect incomplete endothelial proliferation on the raw intracardiac surfaces, sewing ring, and suture knots in the initial postoperative period [1, 12, 18, 19]. The presence of an endothelial lining on the valve surface effectively prevents thrombus formation, but more than one year may be required after implantation for this in-growth to fully form [20, 21]. However, in the absence of endothelial organization, the development of a mature platelet fibrin coating on the valve surface may also be favorable in terms of nonthrombogenicity and explain the absence of thrombotic deposits in explanted valves in which tissue in-growth is incomplete [20, 21]. Therefore, it is believed that the lack of host endothelial cell in-growth and mature platelet fibrin coating on the valvular surface in the postoperative period may promote early thrombus formation and contribute to the elevated early thromboembolic risk.

3 Table 1. and Thromboembolic Rates With Three Common Anticoagulation Protocols Publication Oral Coumadin with Subcutaneous Heparin Ageno W, Am J Cardiol 2001 [5] Akins CW, Ann Thorac Surg [59] Anttila V, Scan Cardiovasc 2002 [60] Bernal JM, Ann Thorac [61] Surg, 1998 Dalrymple-Hay MJR, JHVD 2000 [62] de la Fuente A, JHVD 2000 [63] Demirag M, JHVD 2001 [64] Duveau D, Eur Soc Cardiol 1984 [65] Emery RW, Ann Thorac Surg, 2003 [66] Fiane AE, Ann Thorac Surg [67] Study Type Therapy Study Description Absolute TE rate TE Mortality rate TE Description Absolute Rate Mortality Rate RCT Oral coumadin RCT of 2.5 mg fixed dose coumadin versus 5 mg load (58% mechanical valve patients) 0/197 (0%) 0/197 (0%) Oral coumadin on day 391 MH 1/391 (0.3%) 1/391 (0.3%) Gl bleed one, with IV heparin or dextran on day 5 if INR subtherapeutic CC Oral coumadin on 43 MH vs 48 SJM 2/91 (2.2%) 2 CVA 5/91 (5.5%) POD2 or POD3 Oral coumadin on 1049 CM 0/1049 (0%) 15/1049 (1.4%) POD2 Oral coumadin POD0; 1350 CM 6/1350 (0.4%) 6 CVA 8/1350 (0.6%) IV heparin on POD3 if INR 2 CC CC Oral Coumadin on POD2 Oral coumadin on POD1 and ASA 150mg/d vs Oral coumadin only on POD1 3 mg/kg/d SC Heparin and oral coumadin on POD8-9 SC heparin and oral coumadin Oral coumadin on POD 1 Iguro Y, JHVD 1999 [68] Oral coumadin on POD2 Lim KHH, JTCVS, 2002 [69] Masters RG, JTCVS, 1995 [70] RCT SC heparin and oral coumadin Oral coumadin within 72 hours postop 99 CM vs 93 Monostrut 100 SJM vs 158 Biocor 0/99 (0%) vs 0/93 (0%) 1/100 (1%) vs 1/158 (0.6%) NA 1/99 (1%) vs 4/93 (4%) CVA 0/100 (0%) vs 1/158 (0.6%) Description 349 SJM 3/349 (0.9%) 1 CVA, 2 abdominal hemorrhage 1146 ATS 17/1146 (1.5%) 4/1146 (0.3%) 10/1146 (0.9%) 997 CM 1/997 (0.1%) CVA 2/997 (0.2%) major bleeding 473 CM 1/473 (0.2%) CVA 0/473 (0%) RCT of 234 CM valves versus 251 SJM valves 3/485 (0.6%) 1/485 (0.2%) CVA 1/485 (0.2%) hemopericardium 814 SJM and MH 2/814 (0.2%) CVA 8/814 (1.0%) 772 REVIEW KULIK ET AL Ann Thorac Surg REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 2006;81:770 81

4 Table 1. Continued Publication Minakata K, JHVD 2002 [71] Montalescot G, Circ 2000 [6] Nicoloff D, JTCVS, 1981 [72] Nistal JF, JTCVS, 1996 [73] Rodler SM, Ann Thorac Surg, 1997 [74] Santini F, JHVD 2002 [75] Yamak B, Thorac Cardiovasc Surgeon, [76] Study Type Therapy Study Description Absolute TE rate CC Oral coumadin on POD 1 2 Oral coumadin with SC heparin (tid) Oral coumadin and SC heparin Oral coumadin when patient started oral intake SC heparin, oral coumadin on POD4 Oral coumadin on POD1 Oral coumadin on POD1 TE Mortality rate TE Description Absolute Rate Mortality Rate 616 CM 1/616 (0.2%) CVA 1/616 (0.2%) Description 106 mechanical valve patients 1/106 (0.9%) 1 stroke 2/106 (1.9%) 2 major bleeds 232 SJM 1/232 (0.4%) CVA 1/232 (0.4%) Cardiac tamponade 504 CM 3/504 (0.6%) 1 CVA, 2 embolic MI 583 CM 0/583 (0%) 9/583 (1.5%) 14/504 (2.8%) 13 hemorrhage, 1 tamponade 942 CM 4/942 (0.4%) Hemorrhage 548 SJM 4/448 (0.7%) 4 CVA 37/548 (6.7)% Hemopericardium and cardiac 1993 tamponade TOTAL 28/3056 (0.9%) 22/8507 (0.3%) 50/1525 (3.3%) 77/9798 (0.8%) Intravenous Heparin Acar J, Circ, 1996 [7] RCT IV heparin 6 hours postoop, oral coumadin on POD2 Aoyagi S, JTCVS, 1994 [77] Baudet EM, JTCVS, 1995 [78] Camilleri LF, Cardiovasc Surg [79] Fanikos J, Am J Cardiol 2004 [8] IV heparin on POD1, oral coumadin on POD2 IV heparin 6 24 hours postop then SC heparin, oral coumadin on day 2 CC IV heparin starting 6 hours postop, then SC heparin and oral coumadin POD2 CC Oral coumadin with IV heparin RCT of INR /433 (2.7%) 3.0 versus (mechanical valves) 908 SJM 0/908 (0%) 2/908 (0.2%) 1112 SJM valve operations 134 SJM valves and 86 Sorin Bicarbon valves 34 mechanical valve patients 3/1112 (0.3%) 3/1112 (0.3%) 3 CVA 14/1112 (1.3%) 7 surgical bleeding, 3 tamponade, 3 intracranial hemorrhage, 1 Gl bleed 5/217 (2.3%) 1/217 (0.5%) 1 valve 2/217 (0.9%) surgical bleeding thrombosis, 4 CVA 2/34 (5.9%) 3/34 (8.8%) Ann Thorac Surg REVIEW KULIK ET AL 2006;81: REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 773

5 Table 1. Continued Publication Laffort P, JACC, 2000 [31] Thulin L, Arg Bras Cardiol 1987 [52] Study Type Therapy Study Description Absolute TE rate RCT IV heparin starting 6 hours postop. then SC heparin and oral coumadin on POD 2 Coumadin (intravenous and oral) with IV heparin RCT of oral coumadin with or without aspirin after SJM MVR 510 mechanical valve patients TE Mortality rate TE Description Absolute Rate Mortality Rate Description 5/229 (2.2%) 0/229 (0%) 16/229 (7.0%) 1/229 (0.4%) major hemorrhage 1/510 (0.2%) NA TOTAL 28/2535 (1.1%) 4/2466 (0.2%) 19/263 (7.2%) 19/2466 (0.8%) Low Molecular Weight Heparin Anon, Formulary 2000 [58] Fanikos J, Am J Cardiol 2004 [8] Montalescot G, Circ 2000 [6] CC CC Oral coumadin with enoxaparin Oral coumadin with enoxaparin Oral coumadin with enoxaparin 37 patients, mechanical and bioprosthetic valves 29 mechanical valve patients 102 mechanical valve patients 1/37 (2.7%) 3/37 (8.1%) 2 major bleeds, 1 minor bleed 0/29 (0%) 3/29 (10.3%) 3 pleural effusions 0/102 (0%) 2/102 (2.0%) 2 major bleeds TOTAL 1/168 (0.6%) 8/168 (4.8%) ATS ATS valve; CC case control; CM CarboMedics valve; case series; CVA cerebrovascular accident; Gl gastrointestinal; INR international normalized ratio; IV intravenous; MH Medtronic Hall valve; MI myocardial infarction; NA not available; POD postoperative day; RCT randomized controlled trial; SC subcutaneous; SJM St. Jude Medical valve; TE thromboembolic. 774 REVIEW KULIK ET AL Ann Thorac Surg REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 2006;81:770 81

6 Ann Thorac Surg REVIEW KULIK ET AL 2006;81: REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 775 Finally, as pointed out by Butchart and colleagues [17, 18], the early thromboembolic risk can also be attributed to inconsistencies of oral anticoagulation management. During the first six months, the thromboembolic risk is up to seven times greater than in the after months and years. After aortic valve replacement, the risk of thromboembolic events falls from 16 per 100 patient years in the early postoperative period to 1.4 per 100 patient years at 5 years. Similarly, after mitral valve replacement, the risk falls from 21 per 100 patient years to 2.5 per 100 patient years [17, 18]. The propensity for early formation of thrombus may be related to difficulties in achieving therapeutic anticoagulation in the initial postoperative period [18, 19]. Spot - international normalized ratio (INR) tests of anticoagulated patients report as many as 50% of INR levels outside the therapeutic range [2], and a recent study of prosthetic valve anticoagulation reported subtherapeutic and variable anticoagulation as a contributing factor in the majority of thromboembolic events [22]. Variable or inadequate anticoagulation, defined as an INR 25% or more below the therapeutic range, has been demonstrated to increase the incidence of thromboembolism 2 to 6 times, especially in the postoperative period [2]. The early thromboembolic risk associated with the placement of a mechanical prosthesis depends on the complex interactions between the recipient and valve, the intrinsic thrombogenic properties of the mechanical valve components, and the diligent management of postoperative anticoagulation. ANTICOAGULATION THERAPY. The need for lifelong oral anticoagulation therapy in patients with mechanical prosthetic valves is well-recognized. In patients not receiving long-term anticoagulation therapy, the average rate of major thromboembolism is estimated to be 4 to 8 per 100 patient-years [23, 24]. This risk is reduced to 2.2 per 100 patient-years with antiplatelet therapy, and further reduced to 1 per 100 patient-years with oral anticoagulation (warfarin). Thus, the utilization of postoperative warfarin therapy reduces the incidence of major embolism by approximately 75% and has become the standard of care for all patients with mechanical prostheses [24]. In large studies conducted on anticoagulation therapy, risk factors known to modulate the risk of thromboembolism have included the number of valves replaced, the type of valve implanted, mitral valve replacement (versus aortic valve), atrial fibrillation, the presence of left atrial enlargement or reduced cardiac output, treatment with warfarin or other oral anticoagulants, adequacy of warfarin therapy, and the addition of antiplatelet drugs such as aspirin or dipyridamole [22, 24]. The optimal intensity of oral anticoagulation, defined as the level at which the incidence of both thromboembolic and bleeding complications is lowest, is still a matter of debate. Hirsh and colleagues [25] have demonstrated that increasing the INR, in mechanical valve patients, above the range of 2.5 to 3.0 also increased considerably the rate of bleeding without a reduction in the number of thromboembolic events. The American College of Cardiology/American Heart Association and the American College of Chest Physicians recommended, in their most recent guidelines, that contemporary mechanical valves in the aortic position be anticoagulated with a target INR of 2.0 to 3.0, and mechanical valves in the mitral position be anticoagulated with a target INR of 2.5 to 3.5 [26, 27]. In view of all the clinical evidence presented, the utilization of oral anticoagulation after placement of mechanical valves will remain the standard of care. ANTIPLATELET THERAPY. In recognition of the key role of platelet activation and aggregation in thrombus formation on mechanical valve surfaces, several studies have sought to evaluate the use of antiplatelet agents for thromboembolic prophylaxis, either exclusively, or in conjunction with oral anticoagulation. Ribeiro and colleagues [28] reported on their experience with antiplatelet agents alone in patients with bileaflet aortic and mitral mechanical valves. After 22 months of follow-up, there was no thromboembolic event reported, but the valve thrombosis rate was inappropriately high at 2.1 per 100 patient years (aortic) and 10 per 100 patient years (aortic and mitral). In another report, Czer and colleagues [29] found a significant difference in the rate of valve thrombosis between patients with no anticoagulation (18.9 per 100 patient years), antiplatelet therapy exclusively (3.2 per 100 patient years) and warfarin (1.5 per 100 patient years). Similarly, a recent randomized trial comparing a combination of clopidogrel and aspirin versus warfarin in patients with mechanical aortic valves was terminated prematurely due to the increased rate of early valve thromboses in the antiplatelet arm [30]. Therefore, in the absence of anticoagulation, antiplatelet therapy seems to be insufficient to prevent the development of thrombi. Because platelet aggregation occurs despite anticoagulation, the co-administration of antiplatelet agents with anticoagulation has been advocated to reduce the frequency of thromboembolic complications. Laffort and colleagues [31] studied the addition of aspirin on postoperative day one with mechanical mitral valve patients anticoagulated with intravenous heparin, followed by oral warfarin. Patients receiving daily aspirin had a reduction in the incidence of nonobstructive periprosthetic valve thrombi from 13% to 5% nine days after surgery, and a reduced incidence of thromboembolic events in the first 5 months, from 25% to 9% (p 0.004). A number of studies have explored the long-term coadministration of antiplatelet agents to chronic oral anticoagulation after mechanical valve replacement, and a recent meta-analysis by Massel and colleagues [32] demonstrated that the addition of either aspirin or dipyridamole to warfarin reduced the risk of thromboembolic events by 59% (odds ratio [OR] 0.41, 95% confidence intervals [CI] 0.29% to 0.58%) and total mortality by 51% (OR 0.49, 95% CI 0.35% to 0.67%). Although the risk of major bleeding was increased with antiplatelet agents (OR 1.5, 95% CI 1.03% to 2.18%), this risk was lower with contemporary low dose (100 mg daily) aspirin (OR 1.28). A meta-analysis by Cappelleri and colleagues [33] inves-

7 776 REVIEW KULIK ET AL Ann Thorac Surg REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 2006;81: tigating the same issue demonstrated a 67% (OR 0.33, 95% CI 0.16% to 0.69%) reduction in embolism, but an increase in major gastrointestinal hemorrhage by approximately 250% (OR 3.47, 95% CI 1.43% to 8.40%), leading to an estimate that for every 1.6 patients who had their stroke prevented by combination therapy, there was an excess of one major gastrointestinal bleed. These meta-analyses suggest that the benefits derived from the enhanced antithrombotic potential of combined therapy outweigh the risks [32, 33]. The American College of Cardiology/American Heart Association, in their most recent guidelines, recommended that the addition of aspirin (80 to 100 mg/day) to warfarin be strongly considered for all patients with mechanical valves [27]. This is particularly important for patients who have had an embolus while on warfarin therapy, those with known vascular disease, or those known to be particularly susceptible to hypercoagulability [27]. The American College of Chest Physicians, on the other hand, recommend low-dose aspirin only for mechanical valve patients with thromboembolic risk factors, such as atrial fibrillation, or as a strategy to lower the target INR from 3.0 to 2.5 in patients with bileaflet mechanical valves in the mitral position [26]. Despite the evidence and the presence of these guidelines, a survey of North American cardiac surgeons reported that only 21% of surgeons routinely use aspirin in conjunction with oral anticoagulation after mechanical valve replacement [34]. The prothrombotic milieu encountered after mechanical valve implantation is complex and multifactorial, putting patients at risk for thromboembolic complications. These undesirable outcomes are largely prevented by long-term oral anticoagulation with warfarin and antiplatelet therapy. To date, the data in the literature have predominantly focused on the late results of oral anticoagulation strategies, with a paucity of data on the effects of early postoperative anticoagulation on bleeding and thromboembolic risk. Early Postoperative Anticoagulation EARLY ANTICOAGULATION AND THE RISK OF BLEEDING. A clear relation exists between the intensity of anticoagulation in patients with mechanical valves and the incidence of thromboembolism, on the one hand, and of bleeding, on the other [35]. Although intravenous heparin is frequently administered after mechanical valve implantation to reduce thromboembolism, heparin has the potential to induce bleeding by inhibiting blood coagulation, impairing platelet function [36], and increasing capillary permeability [35]. Pericardial effusion and cardiac tamponade are serious complications associated with the early postoperative use of heparin and warfarin, and can present days to months after surgery. While pericardial effusion may be present in up to 64% of patients after heart surgery, cardiac tamponade is a less frequent event, occurring in 0.5% to 8.5% of patients [37, 38]. Defined as fluid in the pericardial space compromising cardiac filling, cardiac tamponade is a much more serious complication than pericardial effusion, with a mortality rate up to 16% [37, 39]. Cardiac tamponade occurs more often after valve replacement (11%) than coronary artery bypass graft surgery (CABG) (2%), and is seen almost exclusively in anticoagulated patients [38, 40]. is another serious complication associated with anticoagulation after valve implantation. In the immediate postoperative period, bleeding requiring a return to the operating room is more common in valve patients than CABG patients [41]. However, such events usually occur in the early hours after cardiac surgery before the administration of heparin or warfarin and are therefore not usually attributable to anticoagulation. On the other hand, bleeding events, such as gastrointestinal and intracranial hemorrhage, usually present later in the postoperative period and are more likely to be associated with anticoagulation. Occurring in up to 3% of anticoagulated patients in the initial postoperative period, the incidence of all major bleeding events decreases to less than 3% annually in patients on stable anticoagulation [26, 42, 43]. These include any episodes of major internal or external bleeding that cause death, hospitalization, permanent injury, or necessitate transfusion [4]. Although upper gastrointestinal bleeding after cardiac surgery is a rare event (approximately 0.5%), it is associated with a high morbidity and a mortality of up to 30% [44, 45]. Important risk factors associated with gastrointestinal bleeding after cardiac surgery include valve replacement and anticoagulation [44, 46]. In their review of postoperative gastrointestinal bleeding, Heikkinen and colleagues [46] reported that 24% of bleeding events occurred in patients excessively anticoagulated. associated with the central nervous system is also a major complication of anticoagulation, representing 20% to 30% of all bleeding events [2] and the most frequent cause (56%) of bleedingrelated death [43]. Systemic bleeding complications and cardiac tamponade therefore represent serious complications associated with early postoperative anticoagulation. The ideal early postoperative anticoagulation strategy (with both high antithromboembolic efficacy and low bleeding risk) has yet to be determined in a systematic manner. Most reports on the topic are either retrospective or empirical. EARLY POSTOPERATIVE ANTICOAGULATION STRATEGIES. Recommendations for long-term anticoagulation after mechanical valve replacement have been formulated by several working groups, but there is little objective evidence supporting strategies for the management of anticoagulation in the early postoperative period. The three most common early postoperative anticoagulation strategies that have been reported in the literature include the following: (1) oral warfarin starting on postoperative day one, without the use of therapeutic heparin anticoagulation; (2) intravenous unfractionated heparin beginning in the early hours after surgery, oral warfarin starting on postoperative day one, and the continuation of heparin until a therapeutic INR has been achieved; (3) low molecular weight heparin (LMWH) beginning in the early hours after surgery, oral warfarin starting on postoperative day one, and the continuation of LMWH until a

8 Ann Thorac Surg REVIEW KULIK ET AL 2006;81: REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 777 therapeutic INR has been achieved. Unfractionated heparin represents a heterogeneous mixture of glycosaminoglycans that bind to antithrombin, catalyzing the inactivation of thrombin and other clotting factors. Unfractionated heparin has unpredictable pharmacokinetic and pharmacodynamic properties, and therefore requires frequent laboratory assessment of the activated partial thromboplastin time to properly monitor intravenous anticoagulation therapy. The LMWH is derived from unfractionated heparin by chemical or enzymatic depolymerization. Compared with unfractionated heparin, LMWH has reduced antifactor IIa activity relative to antifactor Xa activity, and lacks the nonspecific binding affinities of unfractionated heparin. As a result, LMWH preparations have more predictable pharmacokinetic and pharmacodynamic properties. These properties allow LMWH to be administered subcutaneously once daily without laboratory monitoring [47]. Although each regimen has been promoted, most of the published investigations lack data that would permit a firm conclusion about the optimal anticoagulation regimen. 1. ORAL WARFARIN ALONE. With the early postoperative risks associated with intravenous heparin, the initiation of mechanical valve anticoagulation with oral warfarin alone has been advocated as the safest regimen [5, 48]. This is based on the increased risk of cardiac tamponade, which is up to five times greater in patients treated with therapeutic intravenous heparin (1.8%) compared with those treated with low dose heparin for deep vein thrombosis (DVT) prophylaxis (0.38%) [38, 49]. The safety of this oral anticoagulation only approach was highlighted in a recent randomized trial of different loading doses of warfarin after valve replacement (58% mechanical valves, 42% bioprosthetic valves), without the use of intravenous heparin. Patients randomized to receive 5 mg loading doses achieved the therapeutic range earlier than those randomized to 2.5 mg daily (1.98 vs 2.72 days, p ). However, of those receiving 5 mg doses, 49.6% needed the dose withheld for at least one day, and the gap between the target and mean INR on day 5 was greater. There were no bleeding complications in either group, and despite the lack of intravenous heparin, there were no thromboembolic complications. Therefore, a lower loading dose of warfarin reduced the rate of excessive anticoagulation, and the absence of intravenous heparin was not associated with any thromboembolism [5]. Oral anticoagulation alone may be effective at reducing early thromboembolic and anticoagulation-related morbidity. However, oral anticoagulation requires several days of therapy before a therapeutic INR is achieved. Therefore, strategies employing oral anticoagulation alone may not be optimal in preventing the development of thrombus in patients with mechanical valves. 2. INTRAVENOUS UNFRACTIONATED HEPARIN AND ORAL WARFARIN. Because the risk of thromboembolism is greatest in the early postoperative period [18], and intracardiac thrombus may develop soon after valve implantation, many centers have supported the utilization of intravenous unfractionated heparin early after surgery (within 24 hours), with oral anticoagulation starting on the first postoperative day [1, 19, 27]. However, even with the administration of intravenous heparin three hours after surgery, thrombus formation can be observed by transesophageal echocardiography in up to 18% of patients after mechanical valve implantation [50]. In a swine model evaluating anticoagulation strategies after mechanical mitral valve replacement, 69% (9/13) of pigs died from hemopericardium within 30 days when treated with concurrent intravenous heparin ( units/kg q12 hours) and oral warfarin. In this study, heparin was discontinued when a target INR of greater than 2.0 was achieved [51]. Thulin and Olin [52] reported their clinical experience with anticoagulation after mechanical valve replacement in 510 patients. The morning after surgery, 10 mg of warfarin sodium was administered intravenously, followed by oral warfarin on ensuing days. Anticoagulation was supplemented with intravenous heparin until a therapeutic INR was achieved. A thromboembolism incidence of 0.2% within the first three months was reported, with only 1.3% of patients developing hemopericardium and cardiac tamponade. The authors concluded by stating that the quality of the anticoagulation approach was very important in the prevention of thromboembolic and bleeding complications. The most recent American College of Chest Physicians Anticoagulation Consensus Guidelines encourage the use of intravenous unfractionated heparin until a therapeutic INR is achieved for mechanical valves (grade 2C recommendation), despite the lack of strong evidence [26]. This is based on the premise that starting warfarin therapy without simultaneous therapeutic heparin may enhance thrombosis in situations of high thrombotic risk. In order to avoid hypercoagulable conditions caused by varying degrees of decay in coagulation factors, heparin is continued until the INR is stable in the therapeutic range [53]. Some centers favor a more aggressive anticoagulation regimen, with the continuation of heparin therapy for two additional days after the INR is therapeutic to prevent potential thromboembolic events [53, 54]. However, little controlled data are available to ascertain the optimal use and the adequate degree of anticoagulation with intravenous heparin [6]. This lack of information and coordination has led to the establishment of recommendations that are based primarily on empirical data [26]. 3. LOW MOLECULAR WEIGHT HEPARIN AND ORAL WARFARIN. The LMWH has many potential advantages that may be relevant for patients with mechanical heart valves. They have a better safety profile with less thrombocytopenias [55], less bleeding, a more predictable and rapidly reached anticoagulant effect [56], the possibility of selfadministration without laboratory monitoring, and shorter hospital stays and lower costs associated with outpatient administration [57]. However, unlike unfractionated heparin, LMWH is not fully reversible by protamine, and may carry an increased bleeding risk in the immediate postoperative period [47]. Nevertheless, the

9 778 REVIEW KULIK ET AL Ann Thorac Surg REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 2006;81: pharmacokinetic and biologic advantages of LMWH may be more relevant in the postoperative period of heart valve replacement because of the severe inflammation and platelet and coagulation disorders related to cardiopulmonary bypass. This may make adequate anticoagulation with unfractionated heparin more difficult [6]. Therefore, many centers have adopted the use of LMWH within 24 hours after surgery (without the use of unfractionated heparin), and oral anticoagulation starting on the first postoperative day [6, 8]. Several clinical trials have prospectively evaluated the potential use of LMWH in patients after valve replacement surgery. Bridging therapy with enoxaparin was evaluated postoperatively in a case series of 37 patients who had undergone mechanical or bioprosthetic heart valve replacements and were stable for discharge prior to achieving a therapeutic INR. Enoxaparin at 1 mg/kg every 12 hours was administered until two consecutive therapeutic INR values were reported. During the threemonth study period, bleeding related to anticoagulation was reported in three patients (2 major, 1 minor). One thromboembolic event was reported the day after discontinuation of enoxaparin, and one death due to arrhythmia (possibly related to anticoagulation) was reported two weeks after discontinuation of enoxaparin. It was determined that home LMWH therapy reduced the number of hospital days by a mean 4.6 days per patient. This translated to health system savings of $168,300, based on the average cost of $1,100/day for a surgical bed [58]. Montalescot and colleagues [6] performed a comparative case-control study of subcutaneous unfractionated heparin (500 IU/kg/day, divided over three times a day dosing) versus subcutaneous enoxaparin (100 anti-xa IU/kg every 12 hours) after heart valve replacement. Two major bleeding events occurred in each group, and one stroke occurred in the unfractionated heparin group. Another case-control study of perioperative anticoagulation after mechanical valve implantation compared 29 patients treated with enoxaparin (1 mg/kg every 12 hours) with 34 control patients treated with intravenous unfractionated heparin. There were no thromboembolic events in the LMWH group compared with two (6%) in the unfractionated heparin group (p 0.50), and there were three (10%) bleeding episodes (pleural effusions) in the LMWH group compared with three (9%) bleeding events in the unfractionated heparin group (p 1.0). The use of LMWH reduced the length of hospital stay after surgery ( to , p ) and reduced the cost per patient by an average of $6,864 [8]. Similarly, an open-label randomized trial comparing the LMWH nadroparin to intravenous unfractionated heparin after mechanical valve implantation demonstrated that patients treated with nadroparin had a shorter hospital stay, less coagulation tests, and easier dosing administration. No difference in major hemorrhagic events was reported [9]. Therefore, anticoagulation with LMWH after mechanical valve replacement appears to be a safe, feasible, and effective alternative to intravenous unfractionated heparin, and provides more predictable and rapid anticoagulation compared with unfractionated heparin. The latter is an important point when considering costcontainment issues in government funded healthcare systems. COMPARISON OF ANTICOAGULATION STRATEGIES. Few studies in the literature have directly compared the risks of bleeding and thromboembolism associated with different anticoagulation strategies after mechanical valve replacement. With limited data available, a systematic review was conducted, searching for controlled trials and case series involving the most common mechanical prosthetic valves and different approaches to early postoperative anticoagulation. Summarized in Table 1, a qualitative comparison of the three most common early postoperative anticoagulation strategies was performed with 20 studies describing the use of oral warfarin with subcutaneous heparin (DVT prophylactic dose) [5, 6, 59 76], 7 studies employing intravenous unfractionated heparin and oral warfarin [7, 8, 31, 52, 77 79], and 3 reports of the use of therapeutic LMWH with oral warfarin [6, 8, 58]. Through this analysis, an estimated incidence of early postoperative thromboembolism of 0.9%, 1.1%, and 0.6%, respectively, was determined for each of the three anticoagulation strategies. Furthermore, an estimated incidence of early postoperative hemorrhage of 3.3%, 7.2%, and 4.8%, respectively, was noted. The mortality rate secondary to thromboembolism or bleeding was similar between oral warfarin alone (0.3% and 0.8%) compared with intravenous unfractionated heparin with oral warfarin (0.2% and 0.8%). These data suggest that the use of early postoperative intravenous unfractionated heparin, compared with the use of oral warfarin alone, yields a substantially higher risk of bleeding without lowering the incidence of thromboembolism. This analysis, although simple in design, provides the most thorough up-to-date review of the literature on early postoperative anticoagulation. Nevertheless, these data cannot substitute for a well-designed randomized controlled trial that is clearly warranted at this time. Through this analysis, the strategy of subcutaneous unfractionated heparin (DVT prophylaxis) with oral coumadin after mechanical valve implantation yielded a combined incidence of early postoperative thromboembolism and bleeding of 4.2%. This compared with the 8.3% combined rate for a strategy of intravenous unfractionated heparin with oral coumadin after surgery. Setting an level of 0.05 and of 0.10, approximately 300 patients per group, or 600 patients total would be required to adequately power a trial to detect a difference in these two approaches during the first month after surgery. With most surgical centers averaging 100 to 150 mechanical valve implantations per year, a multicenter trial would be best suited to answer the question at hand. Comment The Need for a Nomenclature Although the need for anticoagulant therapy in patients with mechanical valves is not in dispute, the optimal early postoperative anticoagulation regimen after valve

10 Ann Thorac Surg REVIEW KULIK ET AL 2006;81: REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 779 implantation has been a matter of debate. The lack of consensus is largely dependent on achieving agreement on standard definitions for thromboembolic events and bleeding complications in the early postoperative period. Certainly, some physicians might argue that the presence of nonobstructive thrombi detected by transesophageal echocardiography should not necessarily be labeled as a thromboembolic event, while knowing that the risk of thrombi expansion and embolization is present. Similarly, should an episode of epistaxis be considered a minor bleeding complication? The absence of a universally accepted nomenclature pertaining to anticoagulation management may have contributed to the underreporting of anticoagulation-related morbidity in the literature. In the present setting, a comprehensive approach to data collection related to anticoagulation would improve clinical decision-making. While guidelines exist regarding the reporting of prosthesis-related complications during long-term follow-up after valve surgery [4], these recommendations do not address the early postoperative complications that may be related to anticoagulation strategies, such as the development of a nonobstructive prosthesis thrombus, pericardial effusion, or cardiac tamponade. In order to facilitate this process, we propose the following staging system for early postoperative thromboembolic and bleeding events, ranked in increasing order of clinical severity and patient impact: Thromboembolic Level (TEL) TEL 1: TEL 2: TEL 3: TEL 4: TEL 5: Presence of nonobstructive thrombus in the vicinity of the prosthetic valve as identified by clinical imaging (echocardiography); Documented evidence of peripheral limb and visceral embolization (solid organs, coronary thrombus); Documented evidence of central nervous system (CNS) embolization (transient ischemic attack or cerebrovascular accident); Prosthesis thrombosis requiring thrombolysis or emergency surgery; A fatal thromboembolism. Level (BL) BL 1: BL 2: BL 3: BL 4: BL 5: BL 6: Mild bleeding such as recurrent epistaxis or periodontal bleeding (more than two events) and gastrointestinal bleeds not requiring hospitalization or transfusions; Documented large pericardial effusion ( 1 cm thickness on echocardiography); Gastrointestinal bleeding requiring hospitalization, transfusion, endoscopic intervention or surgery; Documented cardiac tamponade requiring intervention; Documented CNS hemorrhage (cerebrovascular accident); Fatal bleeding. The Need for a Trial The initiation of an efficacious early anticoagulation protocol is important because of its potential impact on the rate of early thromboembolic complications associated with the prothrombotic state after mechanical valve implantation. The challenge in establishing a model will be to balance the risks of under-anticoagulation (thromboembolic events) against those of over-anticoagulation (bleeding complications) in an early postoperative setting when the risk of bleeding may be significant. Although a number of strategies exist, a randomized controlled trial examining the issue of early postoperative anticoagulation after mechanical valve replacement is lacking, and the paucity of knowledge in the matter is a serious impediment to adequate early care of patients with mechanical prostheses. Such a study would compare two or more early postoperative anticoagulation strategies, in addition to routine low dose antiplatelet therapy, to document the incidence of bleeding and thromboembolic complications during the first month after mechanical valve implantation. Further, it would incorporate the proposed nomenclature described above, as well as routine postoperative echocardiography to determine the incidence and clinical impact of prosthetic valve thrombi and pericardial effusions in these anticoagulated patients. Clearly, a need exists for a multicenter randomized study comparing anticoagulation regimens after mechanical valve replacement to determine the safest and most efficacious approach. References 1. Chesebro JH, Adams PC, Fuster V. Antithrombotic therapy in patients with valvular heart disease and prosthetic heart valves. J Am Coll Cardiol 1986;8(suppl B):41B 56B. 2. Edmunds LH Jr. Thrombotic and bleeding complications of prosthetic heart valves. Ann Thorac Surg 1987;44: Kearon C. Management of anticoagulation in patients who require invasive procedures. Semin Vasc Med 2003;3: Edmunds LH Jr, Clark RE, Cohn LH, Grunkemeier GL, Miller DC, Weisel RD. Guidelines for reporting morbidity and mortality after cardiac valvular operations. J Thorac Cardiovasc Surg 1996;112: Ageno W, Turpie AG, Steidl L, et al. Comparison of a daily fixed 2.5-mg warfarin dose with a 5-mg, international normalized ratio adjusted, warfarin dose initially following heart valve replacement. Am J Cardiol 2001;88: Montalescot G, Polle V, Collet JP, et al. Low molecular weight heparin after mechanical heart valve replacement. Circulation 2000;101: Acar J, Iung B, Boissel JP, et al. AREVA: multicenter randomized comparison of low-dose versus standard-dose anticoagulation in patients with mechanical prosthetic heart valves. Circulation 1996;94: Fanikos J, Tsilimingras K, Kucher N, Rosen AB, Hieblinger MD, Goldhaber SZ. Comparison of efficacy, safety, and cost of low-molecular-weight heparin with continuous-infusion unfractionated heparin for initiation of anticoagulation after mechanical prosthetic valve implantation. Am J Cardiol 2004;93: Iliuta L, Candea V, Vasilescu A, Moldovan H, Gherghiceanu DP, Macarie C. Fraxiparine versus unfractionated heparin for the perioperative anticoagulant therapy in patients undergoing mechanical prosthetic heart valve replacement. Archives de L Union Medicale Balkanique 2003;38:141 8.

11 780 REVIEW KULIK ET AL Ann Thorac Surg REVIEW OF EARLY POSTOPERATIVE ANTICOAGULATION 2006;81: Dewanjee MK, Trastek VF, Tago M, Kaye MP. Radioisotopic techniques for noninvasive detection of platelet deposition in bovine-tissue mitral-valve prostheses and in vitro quantification of visceral microembolism in dogs. Invest Radiol 1984;19: Harker LA, Slichter SJ. Studies of platelet and fibrinogen kinetics in patients with prosthetic heart valves. N Engl J Med 1970;283: Stein B, Fuster V, Halperin JL, Chesebro JH. Antithrombotic therapy in cardiac disease. An emerging approach based on pathogenesis and risk. Circulation 1989;80: Nemerson Y, Turitto VT. The effect of flow on hemostasis and thrombosis. Thromb Haemost 1991;66: Pengo V, Peruzzi P, Baca M, et al. The optimal therapeutic range for oral anticoagulant treatment as suggested by fibrinopeptide A (FpA) levels in patients with heart valve prostheses. Eur J Clin Invest 1989;19: Douglas PS, Hirshfeld JW Jr, Edie RN, Harken AH, Stephenson LW, Edmunds LH Jr Clinical comparison of St. Jude and porcine aortic valve prostheses. Circulation 1985;72(pt 2): II Murphy DA, Levine FH, Buckley MJ, et al. Mechanical valves: a comparative analysis of the Starr-Edwards and Bjork-Shiley prostheses. J Thorac Cardiovasc Surg 1983;86: Butchart EG, Lewis PA, Kulatilake EN, Breckenridge IM. Anticoagulation variability between centres: implications for comparative prosthetic valve assessment. Eur J Cardiothorac Surg 1988;2: Butchart EG, Lewis PA, Bethel JA, Breckenridge IM. Adjusting anticoagulation to prosthesis thrombogenicity and patient risk factors. Recommendations for the Medtronic Hall valve. Circulation 1991;84(suppl):III Heras M, Chesebro JH, Fuster V, et al. High risk of thromboemboli early after bioprosthetic cardiac valve replacement. J Am Coll Cardiol 1995;25: Brais M, Braunwald NS. Acceleration of tissue ingrowth on materials implanted in the heart. Ann Thorac Surg 1976;21: Bull B, Braunwald NS. Human histopathologic response to a completely fabric-covered prosthetic heart valve. Ann Surg 1971;174: Saour JN, Sieck JO, Mamo LA, Gallus AS. Trial of different intensities of anticoagulation in patients with prosthetic heart valves. N Engl J Med 1990;322: Mok CK, Boey J, Wang R, et al. Warfarin versus dipyridamole-aspirin and pentoxifylline-aspirin for the prevention of prosthetic heart valve thromboembolism: a prospective randomized clinical trial. Circulation 1985;72: Cannegieter SC, Rosendaal FR, Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994;89: Hirsh J, Dalen JE, Anderson DR, et al. Oral anticoagulants: mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 1998;114(suppl 5):445S 469S. 26. Stein PD, Alpert JS, Bussey HI, Dalen JE, Turpie AG. Antithrombotic therapy in patients with mechanical and biological prosthetic heart valves. Chest 2001;119(suppl 1): 220S 7S. 27. Bonow RO, Carabello B, de Leon AC, Jr??, et al. Guidelines for the management of patients with valvular heart disease: executive summary. A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Management of Patients with Valvular Heart Disease). Circulation 1998;98: Ribeiro PA, Al Zaibag M, Idris M, et al. Antiplatelet drugs and the incidence of thromboembolic complications of the St. Jude Medical aortic prosthesis in patients with rheumatic heart disease. J Thorac Cardiovasc Surg 1986;91: Czer LS, Matloff JM, Chaux A, De Robertis M, Stewart ME, Gray RJ. The St. Jude valve: analysis of thromboembolism, warfarin-related hemorrhage, and survival. Am Heart J 1987;114: Schlitt A, von Bardeleben RS, Ehrlich A, et al. Clopidogrel and aspirin in the prevention of thromboembolic complications after mechanical aortic valve replacement (CAPTA). Thromb Res 2003;109(2-3): Laffort P, Roudaut R, Roques X, et al. Early and long-term (one-year) effects of the association of aspirin and oral anticoagulant on thrombi and morbidity after replacement of the mitral valve with the St. Jude medical prosthesis: a clinical and transesophageal echocardiographic study. J Am Coll Cardiol 2000;35: Massel D, Little SH. Risks and benefits of adding antiplatelet therapy to warfarin among patients with prosthetic heart valves: a meta-analysis. J Am Coll Cardiol 2001;37: Cappelleri JC, Fiore LD, Brophy MT, Deykin D, Lau J. Efficacy and safety of combined anticoagulant and antiplatelet therapy versus anticoagulant monotherapy after mechanical heart-valve replacement: a metaanalysis. Am Heart J 1995;130(pt 1): Ray JG, Turpie AG. Survey of cardiac surgeons perceptions of the addition of ASA to warfarin for patients with mechanical heart valves. Can J Cardiol 1997;13: Levine MN, Raskob G, Landefeld S, Kearon C. Hemorrhagic complications of anticoagulant treatment. Chest 2001; 119(suppl 1):108S 21S. 36. Muriithi EW, Belcher PR, Day SP, Menys VC, Wheatley DJ. Heparin-induced platelet dysfunction and cardiopulmonary bypass. Ann Thorac Surg 2000;69: Malouf JF, Alam S, Gharzeddine W, Stefadouros MA. The role of anticoagulation in the development of pericardial effusion and late tamponade after cardiac surgery. Eur Heart J 1993;14: Pepi M, Muratori M, Barbier P, et al. Pericardial effusion after cardiac surgery: incidence, site, size, and haemodynamic consequences. Br Heart J 1994;72: Mangi AA, Palacios IF, Torchiana DF. Catheter pericardiocentesis for delayed tamponade after cardiac valve operation. Ann Thorac Surg 2002;73: Meurin P, Weber H, Renaud N, et al. Evolution of the postoperative pericardial effusion after day 15: the problem of the late tamponade. Chest 2004;125: Unsworth-White MJ, Herriot A, Valencia O, et al. Resternotomy for bleeding after cardiac operation: a marker for increased morbidity and mortality. Ann Thorac Surg 1995; 59: Landefeld, Goldman L. Major bleeding in outpatients treated with warfarin: incidence and prediction by factors known at the start of outpatient therapy. Am J Med 1989;87: van der Meer FJ, Rosendaal FR, Vandenbroucke JP, Briet E. complications in oral anticoagulant therapy. An analysis of risk factors. Arch Intern Med 1993;153: van der Voort PH, Zandstra DF. Pathogenesis, risk factors, and incidence of upper gastrointestinal bleeding after cardiac surgery: is specific prophylaxis in routine bypass procedures needed? J Cardiothorac Vasc Anesth 2000;14: Sakorafas GH, Tsiotos GG. Intra-abdominal complications after cardiac surgery. Eur J Surg 1999;165: Heikkinen LO, Ala-Kulju KV. Abdominal complications following cardiopulmonary bypass in open-heart surgery. Scand J Thorac Cardiovasc Surg 1987;21: Hirsh J, Raschke R. Heparin and low-molecular-weight heparin: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004;126(suppl 3):188S 203S. 48. Chesebro JH, Fuster V. Optimal antithrombotic therapy for mechanical prosthetic heart valves. Circulation 1996;94: Wilhelm M, Carrel T, Jenni R, Turina M. Incidence, clinical aspects and therapy of late pericardial tamponade following heart surgery. Swiss Surg 1996;3: Malergue MC, Temkine J, Slama M, et al. Value of early systematic postoperative transesophageal echocardiography