- Email: [email protected]
Physician Preferences for Perioperative Anticoagulation in Patients With a Mechanical Heart Valve Who Are Undergoing Elective Noncardiac Surgery
Physician Preferences for Perioperative Anticoagulation in Patients With a Mechanical Heart Valve Who Are Undergoing Elective Noncardiac Surgery* James D. Douketis MD; Mark A. Crowther MD, MSc; Sunjay S. Cherian, BSc; and Clive B. Kearon, MB, PhD, FCCP
Study objective: To determine physicians’ anticoagulation preferences in patients with a mechanical heart valve who are undergoing elective surgery, and to determine the effect of different risks of thromboembolism (TE) and postoperative bleeding on anticoagulation preferences. Design: Mail survey of physicians who prescribe anticoagulant therapy. Methods and results: Physicians were asked to provide anticoagulation preferences in four clinical scenarios of patients with a mechanical heart valve who are undergoing elective surgery. Physicians were asked to select from three preoperative anticoagulation options (two aggressive, one less aggressive) and four postoperative anticoagulation options (two aggressive, two less aggressive). IV heparin was the most frequently selected anticoagulation option. Depending on the scenario, it was preferred by 39 to 79% of respondents for preoperative anticoagulation therapy, and by 44 to 84% of respondents for postoperative anticoagulant therapy. The risk of TE had a strong influence on anticoagulation preferences: more respondents preferred aggressive anticoagulant management in high-risk compared with low-risk TE scenarios (p < 0.001). Anticoagulation preferences were not influenced by the risk of bleeding: the proportion of respondents who preferred aggressive anticoagulant management did not differ in high-risk and low-risk bleeding scenarios (p > 0.05). Of respondents who preferred IV heparin for postoperative anticoagulation therapy, the risk of bleeding influenced the timing of heparin initiation: fewer respondents preferred early heparin initiation (within 12 h after surgery) in high-risk compared with low-risk bleeding scenarios (p < 0.01). Conclusions: (1) Preoperative and postoperative IV heparin were the most frequently selected anticoagulation options. (2) The risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management. (3) If IV heparin was selected, the risk of bleeding influenced the timing of heparin initiation. (CHEST 1999; 116:1240 –1246) Key words: anticoagulation; embolism; mechanical heart valve; surgery; thrombosis Abbreviations: LMWH 5 low-molecular-weight heparin; TE 5 thromboembolism
patients with a mechanical heart valve who are I nreceiving long-term oral anticoagulant therapy and are undergoing elective surgery, the management of perioperative anticoagulation therapy is *From the Department of Medicine (Drs. Douketis, Crowther, Mr. Cherian, and Dr. Kearon), St. Joseph’s Hospital (Drs. Douketis and Crowther), Hamilton, Ontario, Canada; and the Hamilton Health Sciences Corporation (Dr. Kearon), McMaster University, Hamilton, Ontario, Canada. Dr. Crowther is a recipient of a Research Fellowship from the Medical Research Council of Canada. Dr. Kearon is a recipient of a Research Scholarship from the Heart and Stroke Foundation of Ontario. Manuscript received January 11, 1999; revision accepted April 29, 1999. Correspondence to: James D. Douketis, MD, St. Joseph’s Hospital, Room F-513, 50 Charlton Ave East, Hamilton, ON, Canada, L8N 4A6; e-mail: [email protected] 1240
problematic. Temporary discontinuation of anticoagulants increases the risk of valve thrombosis and systemic embolism.1–3 On the other hand, continuing anticoagulants perioperatively, or stopping and restarting anticoagulants too soon after surgery, can cause life-threatening bleeding.4,5 The optimal anticoagulation strategy would minimize the risk of thromboembolism (TE), without causing excessive postoperative bleeding. To date, there is no consensus about the management of perioperative anticoagulation therapy in patients with a mechanical heart valve who are undergoing elective surgery, largely because of a lack of clinical trials investigating different anticoagulation strategies. The traditional management apClinical Investigations
proach has been to hospitalize patients 3 to 4 days before surgery to discontinue warfarin or other oral anticoagulants and to start IV heparin, which is stopped 3 h before surgery.6 –11 After surgery, IV heparin and warfarin are restarted when surgical hemostasis has been achieved, and heparin is continued until the anticoagulant effect of warfarin is within the therapeutic range. However, some authors have suggested that the risk of perioperative TE has been exaggerated,12,13 and that perioperative IV heparin should be reserved for patients at highest risk for TE, eg, previous TE.14,15 A less aggressive management approach involves temporary discontinuation of warfarin and the use of low-dose subcutaneous heparin postoperatively until patients are discharged from the hospital.14 Another approach, which is appealing for patients undergoing a surgical procedure that does not require overnight hospitalization, involves outpatient perioperative anticoagulant therapy with full-dose subcutaneous heparin or low-molecular-weight heparin (LMWH). As a first step in designing prospective studies to investigate different perioperative anticoagulation strategies, we performed a survey to determine physicians’ anticoagulation preferences in patients with a mechanical heart valve who are undergoing elective surgery and to determine the effect of different risks of TE and postoperative bleeding on anticoagulation preferences.
Table 1—Clinical Scenarios* 1. A 70-year-old woman with a mechanical mitral valve, chronic atrial fibrillation, and a previous stroke 2 years ago is to undergo elective (open) subtotal colectomy for resection of adenocarcinoma of the colon (TE risk, high; bleeding risk, high) 2. A 65-year-old man with a mechanical aortic valve is to undergo elective (open) subtotal colectomy for resection of adenocarcinoma of the colon (TE risk, low; bleeding risk, high) 3. A 75-year-old man with a mechanical mitral valve, a previous stroke 3 years ago, and chronic atrial fibrillation is to undergo elective bilateral inguinal hernia repair (TE risk, high; bleeding risk, low) 4. A 48-year-old woman with a mechanical aortic valve is to undergo elective bilateral inguinal hernia repair (TE risk, low; bleeding risk, low) *TE and bleeding risks for each scenario were not provided in the mailed survey.
requested information relating to a respondent’s medical specialty (general internal medicine, cardiology, cardiac surgery, or other) and their frequency of managing anticoagulant-related problems (frequently, infrequently, or not at all). Analysis The analysis consisted of four parts. (1) Anticoagulation preferences in different scenarios: The proportion of respondents who selected each of the anticoagulation options was determined for each scenario. (2) Effect of TE risk on preoperative and postoperative anticoagulation preferences: The proportion of respondents who preferred aggressive anticoagulant management (ie, options “a” or “b”) was compared in scenario pairs in which the TE risk was high in one scenario and low in the other
Materials and Methods Table 2—Anticoagulation Options* Survey Methods The survey was mailed to all members of the Canadian Society of Internal Medicine (n 5 420) and the Canadian Cardiovascular Society (n 5 540) in July 1997, with a repeat mailing in December 1997 to nonresponders. A priori, criteria were established to exclude returned surveys from the analysis: (1) physician does not manage anticoagulant-related problems in adults; (2) physician no longer resides at the address provided, ie, survey returned to sender; (3) survey was partially completed; and (4) physician has retired from medical practice. Survey Questions Physicians were asked to provide anticoagulation preferences in four scenarios of patients with a mechanical heart valve who are undergoing elective surgery (Table 1). The scenarios represented four different combinations of risks for TE (high or low) and bleeding (high or low). However, this information was not provided in the survey. For each scenario, physicians were asked to select from a standardized list of preoperative and postoperative anticoagulation options or, if these options were not acceptable, to provide an alternative anticoagulation strategy (Table 2). Options “a” and “b” (preoperative and postoperative) were classified as aggressive anticoagulant management, and options “c” (preoperative and postoperative) and “d” (postoperative) were classified as less aggressive anticoagulant management. We also
Preoperative anticoagulation options a: Admit to hospital 2 to 4 days preoperatively for full-dose IV heparin† b: Outpatient full-dose SC heparin or LMWH† c: Nothing else other than stopping warfarin preoperatively d: Other Postoperative anticoagulation options a: Full-dose in-hospital IV heparin until INR therapeutic† Heparin to be restarted , 6 h postoperatively 6 to 12 h postoperatively . 12 h postoperatively b: Early discharge home with full-dose SC heparin or LMWH until INR therapeutic† c: Low-dose in-hospital SC heparin or LMWH until INR therapeutic‡ d: Nothing else other than restarting warfarin postoperatively e: Other *These options are given in addition to stopping warfarin 4 to 5 days preoperatively and restarting warfarin 1 to 2 days postoperatively. INR 5 International Normalization Ratio; SC 5 subcutaneous injection. †Full-dose SC/IV heparin or SC LMWH is the dose recommended for treatment of venous TE or acute coronary syndromes. ‡Low-dose SC heparin or LMWH is the dose recommended as prophylaxis for venous TE. CHEST / 116 / 5 / NOVEMBER, 1999
1241
scenario, and the bleeding risk was the same in both scenarios. (3) Effect of bleeding risk on postoperative anticoagulation preferences: The proportion of respondents who preferred aggressive anticoagulant management (ie, options “a” or “b”) was compared in scenario pairs in which the bleeding risk was high in one scenario and low in the other scenario, and the TE risk was the same in both scenarios. (4) Effect of TE or bleeding risk on the timing of postoperative IV heparin initiation: The proportion of respondents who preferred early heparin initiation, defined as starting heparin within 12 h after surgery, was compared in scenario pairs in which the TE risk was high in one scenario and low in the other scenario, and the bleeding risk was the same in both scenarios; and in scenario pairs in which the bleeding risk was high in one scenario and low in the other scenario, and the TE risk was the same in both scenarios. McNemar’s x2 test was used to compare the proportion of respondents who preferred aggressive anticoagulant management within scenario pairs. Statistical significance was defined as p , 0.05.
Results Survey Response Rate and Characteristics of Respondents The survey was mailed to 960 physicians; 87 surveys were returned because the physician no longer resided at the address that was provided. Of 873 physicians who received the survey, 538 (62%) responded. Of 538 returned surveys, 65 were excluded for the following reasons: (1) physician does not manage anticoagulant-related problems in adults
(n 5 38), (2) physician has retired from medical practice (n 5 17), and (3) survey was partially completed (n 5 10). Thus, there were 473 surveys included in the analysis. Of the 473 respondents, 217 (46%) were cardiologists, 179 (38%) were general internists, 55 (12%) were cardiac surgeons, and 22 (4.6%) were from another specialty; 406 (86%) managed anticoagulant-related problems frequently and 67 (14%) managed them infrequently. Responses to Clinical Scenarios The anticoagulation preferences for the four scenarios are summarized in Figures 1 and 2. IV heparin was the most frequently selected anticoagulation option for preoperative and postoperative anticoagulation therapy in each of the four scenarios. Of those respondents who selected IV heparin for postoperative anticoagulation therapy, preferences for the timing of heparin initiation are summarized in Figure 3. Initiation of IV heparin 6 to 12 h after surgery was the most frequently selected timing for heparin initiation in each of the four scenarios, whereas heparin initiation , 6 h after surgery was least preferred by the respondents. The effect of TE risk on preoperative and postoperative anticoagulation preferences is summarized in Table 3. The risk of TE had a strong influence on
Figure 1. Preoperative anticoagulation management preferences. For description of different scenarios and management options, see Tables 1 and 2. 1242
Clinical Investigations
Figure 2. Postoperative anticoagulation management preferences. For description of different scenarios and management options, see Tables 1 and 2.
anticoagulation preferences. There was a significantly greater proportion of respondents who preferred aggressive preoperative and postoperative anticoagulant management in high-risk TE scenarios compared with low-risk TE scenarios. The effect of bleeding risk on postoperative anticoagulation preferences is summarized in Table 3. The risk of bleeding did not have a detectable effect on postoperative anticoagulation preferences, inasmuch as the proportion of respondents who preferred aggressive anticoagulant management was not significantly different in high-risk and low-risk bleeding scenarios. The effect of TE or bleeding risk on the timing of postoperative IV heparin initiation is summarized in Table 4. Of respondents who preferred IV heparin for postoperative management, the risk of bleeding influenced whether the timing of heparin initiation after surgery was early (ie, within 12 h) or later (ie, . 12 h). There was a significantly smaller proportion of respondents who preferred early heparin initiation in high-risk bleeding scenarios compared with lowrisk bleeding scenarios. The risk of TE did not have a detectable effect on the timing of heparin initiation, inasmuch as the proportion of respondents who preferred early heparin initiation was not significantly different in high-risk and low-risk TE scenarios. Discussion In this survey, we investigated physician preferences for perioperative anticoagulant therapy in
patients with a mechanical heart valve who are undergoing elective noncardiac surgery. There are three main findings from this survey. (1) Preoperative and postoperative IV heparin was the most frequently selected anticoagulation option. (2) The risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management. (3) If IV heparin was selected, the risk of bleeding influenced the timing of heparin initiation. Our finding that IV heparin was the most frequently selected anticoagulation option is noteworthy because the effectiveness and safety of this management approach has not been validated in clinical trials. The rationale for aggressive perioperative anticoagulation therapy with IV heparin is based on evidence from retrospective case series16,17 and one small prospective cohort study18 that did not provide reliable estimates of the risks of perioperative TE and bleeding. Furthermore, most patients in these studies had first-generation caged-ball heart valves, which are more thrombogenic than newer bileaflet tilting-disk models.5 In two methodologically rigorous studies that provided quantitative estimates of the risks of TE and bleeding for patients with a mechanical heart valve who require elective surgery, the investigators questioned the use of perioperative IV heparin, except in patients at highest risk for TE (ie, previous TE, caged-ball heart valve).14,15 One of these studies concluded that in the absence of an embolic event within the previous month, the use of IV heparin was expected to CHEST / 116 / 5 / NOVEMBER, 1999
1243
Figure 3. Postoperative IV heparin initiation preferences. For description of different scenarios, see Table 1.
markedly increase morbidity related to postoperative bleeding.14 Thus, although IV heparin was preferred by most respondents in this survey, the evidence to support this anticoagulation strategy is questionable. Table 3—Effect of TE or Bleeding Risk on Preoperative and Postoperative Anticoagulation Preferences* Anticoagulation Preferences Preoperative management Scenario 1 Scenario 2 Scenario 3 Scenario 4 Postoperative management Bleeding risk High Scenario 1 Scenario 2 Low Scenario 3 Scenario 4 p Value
TE Risk† High
Low
p Value , 0.001
434 (92) 312 (66)
, 0.001
416 (88)
Table 4 —Effect of TE or Bleeding Risk on the Timing of Postoperative IV Heparin Initiation*
288 (61)
TE Risk Bleeding Risk , 0.001
424 (90) 340 (72)
, 0.001
415 (88) 0.36
320 (68) 0.06
*Values given as No. (%). †Values represent the proportion of 473 respondents selecting aggressive anticoagulant management: IV heparin (option a); or full-dose SC heparin or LMWH (option b). For abbreviation definition, see Table 2. 1244
Another noteworthy finding is that despite convincing evidence that subcutaneous heparin or LMWH is as effective as IV heparin for the treatment of venous TE19,20 and acute coronary syndromes,21 this anticoagulation strategy, which does not require hospitalization, was selected by only 5.3 to 23% of respondents. Thus, demonstration that outpatient perioperative anticoagulation therapy with subcutaneous heparin or LMWH is effective and safe could have a major impact on clinical
High Scenario Scenario Low Scenario Scenario p Value
High
1 (n 5 399)† 2 (n 5 287)
250 (63)
3 (n 5 354) 4 (n 5 210)
268 (76)
Low
p Value 0.27
167 (58)
0.002
0.26 150 (71) 0.002
*Values given as No. (%). Numbers are given as a proportion of those respondents who selected aggressive heparin initiation (ie, within 12 h after surgery from among those who preferred IV heparin for postoperative anticoagulation therapy). †No. of respondents who selected aggressive heparin initiation. Clinical Investigations
practice. Further, this management approach is likely to be more cost-effective compared with the use of in-hospital IV heparin.15 A less aggressive anticoagulant management (ie, options “c” or “d”), which has been advocated by some authors,14 was preferred in the preoperative or postoperative setting by only 3.2 to 9.4% of respondents in high-risk TE scenarios, and by 11 to 36% of respondents in low-risk TE scenarios. It is likely that before a less aggressive anticoagulant management is adopted for wider use, the effectiveness of this approach will require validation in prospective clinical trials. Our finding that the risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management may be explained by the following considerations. First, physicians may consider the prevention of TE as the primary management objective because the clinical consequence of TE (eg, stroke), resulting from less aggressive anticoagulation therapy, is likely to be greater than that of bleeding (eg, wound hematoma), resulting from aggressive anticoagulation therapy. Second, physicians may be concerned about legal liability if a disabling stroke occurs perioperatively, whereas the occurrence of postoperative bleeding may be more defensible.22 However, there is evidence to suggest that the risk of perioperative TE has been exaggerated, and greater consideration should be given to the risk of postoperative bleeding and its clinical consequences as determinants of anticoagulation preferences. In a cohort study of 45 patients with a mechanical heart valve who underwent elective surgery and received perioperative IV heparin, 13 patients (29%) had postoperative bleeding complications, and in 1 patient (2.2%), the bleeding was fatal.23 In a meta-analysis of randomized controlled trials involving patients with a mechanical heart valve, in which one patient group did not receive anticoagulant therapy, the risk of TE in patients who were not anticoagulated was 9% per year.24 If the period in which patients are not fully anticoagulated is limited to 2 or 3 days, as in the perioperative period, the risk for TE during this period would be low (ie, 9% 3 3/365 5 0.07%). It is probable that the risk of TE would be higher in a high-risk subgroup (ie, mechanical mitral valve, atrial fibrillation). However, if most physicians are using IV heparin perioperatively, it is possible that this anticoagulation strategy is causing greater morbidity and deaths because of bleeding compared with the morbidity and deaths because of TE if IV heparin had not been used, particularly in patients at lower risk for TE (ie, mechanical aortic valve, sinus rhythm). Of those respondents who preferred IV heparin for postoperative anticoagulation therapy, the risk of bleeding influenced the timing of heparin initiation.
A significantly higher proportion of respondents preferred later heparin initiation (ie, . 12 h after surgery) in high-risk bleeding scenarios compared with low-risk bleeding scenarios. To our knowledge, there are no studies investigating when it is safe to initiate IV heparin after surgery. Presumably, this will depend on the surgical procedure, inasmuch as patients undergoing urologic, neurosurgical, or cancer-related procedures are more susceptible to postoperative bleeding.25,26 It is possible that withholding IV heparin for as little as 24 to 48 h after surgery might substantially reduce the risk of postoperative bleeding, although having a negligible effect on patients’ risk of TE. The are several points that support the validity and generalizability of our findings. First, the anticoagulation options we provided were chosen by . 98% of respondents, depending on the scenario, thereby suggesting that the responses in this survey are a valid representation of physicians’ usual clinical practice. Second, the respondents were from a broad cross-section of specialties that included general internal medicine, cardiology, and cardiac surgery. Third, because 86% of respondents managed anticoagulant-related problems frequently, our findings are generalizable to other physicians who manage anticoagulant-related problems. We acknowledge that the survey response rate of 62% might be considered suboptimal; however, this response rate is consistent with rates reported in other surveys of anticoagulant-related practices.27,28 To summarize, there is a need for large, prospective studies to investigate the efficacy and safety of the anticoagulation preferences identified in this survey, and to provide reliable estimates of the risks of perioperative TE and bleeding associated with different anticoagulation strategies. ACKNOWLEDGMENT: We thank Mrs. Teresa Tomlinson for her invaluable administrative assistance.
References 1 Stein PD, Alpert JS, Dalen JE, et al. Antithrombotic therapy in patients with mechanical and biological prosthetic heart valves. Chest 1998; 114(suppl 5):602S– 610S 2 Silber H, Kahn SS, Matloff JM, et al. The St. Jude valve: thrombolysis as the first level of therapy for cardiac valve thrombosis. Circulation 1993; 87:30 –37 3 Edmunds LH Jr. Thromboembolic complications of current cardiac valve prostheses. Ann Thorac Surg 1982; 34:96 –106 4 Levine M, Raskob GE, Landefeld S, et al. Hemorrhagic complications of anticoagulant treatment. Chest 1998; 114(suppl 5):511S–523S 5 Vongpatanasin W, Hillis D, Lange RA. Prosthetic heart valves. N Engl J Med 1996; 335:407– 416 6 Bryan AJ, Butchart EG. Prosthetic heart valves and anticoagulant management during non-cardiac surgery. Br J Surg 1995; 82:577–578 CHEST / 116 / 5 / NOVEMBER, 1999
1245
7 Travis S, Wray R, Harrison K. Perioperative anticoagulation control. Br J Surg 1989; 76:1107–1108 8 Madura JA, Rookstool M, Wease G. The management of patients on chronic Coumadin therapy undergoing subsequent surgical procedures. Am Surgeon 1994; 60:542–547 9 Busuttil WJ, Fabri BM. The management of anticoagulation in patients with prosthetic heart valves undergoing noncardiac operations. Postgrad Med J 1995; 71:390 –392 10 Tiede DJ, Nishimura RA, Gastineau DA, et al. Modern management of prosthetic valve anticoagulation. Mayo Clin Proc 1998; 73:665– 680 11 Blaustein AS. Preoperative and perioperative management of cardiac patients undergoing noncardiac surgery. Cardiol Clin 1995; 13:149 –160 12 Kuwada SK, Balm R, Gostout CJ. The risk of withdrawing chronic anticoagulation because of acute GI bleeding. Am J Gastroenterol 1996; 91:1116 –1119 13 Tinker JH, Tarhan S. Discontinuing anticoagulant therapy in surgical patients with cardiac valve prostheses. JAMA 1978; 239:738 –739 14 Kearon C, Hirsh J. Management of anticoagulation before and after elective surgery. N Engl J Med 1997; 336:1506 – 1511 15 Eckman MH, Beshansky JR, Durand-Zaleski I, et al. Anticoagulation for noncardiac procedures in patients with prosthetic heart valves. JAMA 1990; 263:1513–1521 16 Katholi RE, Nolan SP, McGuire LB. Living with prosthetic heart valves: subsequent noncardiac operations and the risk of thromboembolism of hemorrhage. Am Heart J 1976; 92:162– 167 17 Shean FC, Austen WG, Buckley MJ, et al. Survival after Starr-Edwards aortic valve replacement. Circulation 1971; 41:1– 8 18 Katholi RE, Nolan SP, McGuire LB. The management of anticoagulation during noncardiac operations in patients with prosthetic heart valves. Am Heart J 1978; 96:163–165 19 Low-molecular-weight heparin is an effective and safe treat-
1246
20
21
22
23
24 25 26
27
28
ment for deep vein thrombosis and pulmonary embolism: The Columbus Investigators. N Engl J Med 1997; 337:657– 662 Hommes DW, Bura A, Mazzolai L, et al. Subcutaneous heparin compared with continuous intravenous heparin administration in the initial treatment of deep vein thrombosis. Ann Intern Med 1992; 116:279 –284 Cohen M, Demers C, Gurfinkel EP, et al. A comparison of low-molecular-weight heparin with unfractionated heparin for unstable coronary artery disease: Efficacy and Safety of Subcutaneous Enoxaparin in Non-Q-Wave Coronary Events Study Group. N Engl J Med 1997; 337:447– 452 McIntyre K. Medicolegal implications of consensus statements with special attention to the Fifth Antithrombotic Therapy Consensus Conference. Chest 1998; 114(suppl 5): 742S–747S Darmon D, Enriquez-Sarano M, Acar J. Cardiac complications after subsequent non-cardiac operations in patients with non-biological prosthetic heart valves [abstract]. Eur J Cardiol 1983; 4(suppl 1):30 Cannegieter SC, Rosendaal FR, Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994; 89:635– 641 Bigg SW, Catalona WJ. Prophylactic mini-dose heparin in patients undergoing radical retropubic prostatectomy: a prospective trial. Urology 1992; 39:309 –313 Turpie AGG, Hirsh J, Gent M, et al. Prevention of deep vein thrombosis in potential neurosurgical patients: a randomized trial comparing graduated compression stockings alone or graduated compression stockings plus intermittent pneumatic compression. Arch Intern Med 1989; 149:679 – 681 Kutner M, Nixon G, Silverstone F. Physicians’ attitudes toward oral anticoagulants and antiplatelet agents for stroke prevention in elderly patients with atrial fibrillation. Arch Intern Med 1991; 151:1950 –1953 Chang HJ, Bell JR, Deroo DB, et al. Physician variation in anticoagulating patients with atrial fibrillation. Arch Intern Med 1990; 150:81– 84
Clinical Investigations
Study objective: To determine physicians’ anticoagulation preferences in patients with a mechanical heart valve who are undergoing elective surgery, and to determine the effect of different risks of thromboembolism (TE) and postoperative bleeding on anticoagulation preferences. Design: Mail survey of physicians who prescribe anticoagulant therapy. Methods and results: Physicians were asked to provide anticoagulation preferences in four clinical scenarios of patients with a mechanical heart valve who are undergoing elective surgery. Physicians were asked to select from three preoperative anticoagulation options (two aggressive, one less aggressive) and four postoperative anticoagulation options (two aggressive, two less aggressive). IV heparin was the most frequently selected anticoagulation option. Depending on the scenario, it was preferred by 39 to 79% of respondents for preoperative anticoagulation therapy, and by 44 to 84% of respondents for postoperative anticoagulant therapy. The risk of TE had a strong influence on anticoagulation preferences: more respondents preferred aggressive anticoagulant management in high-risk compared with low-risk TE scenarios (p < 0.001). Anticoagulation preferences were not influenced by the risk of bleeding: the proportion of respondents who preferred aggressive anticoagulant management did not differ in high-risk and low-risk bleeding scenarios (p > 0.05). Of respondents who preferred IV heparin for postoperative anticoagulation therapy, the risk of bleeding influenced the timing of heparin initiation: fewer respondents preferred early heparin initiation (within 12 h after surgery) in high-risk compared with low-risk bleeding scenarios (p < 0.01). Conclusions: (1) Preoperative and postoperative IV heparin were the most frequently selected anticoagulation options. (2) The risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management. (3) If IV heparin was selected, the risk of bleeding influenced the timing of heparin initiation. (CHEST 1999; 116:1240 –1246) Key words: anticoagulation; embolism; mechanical heart valve; surgery; thrombosis Abbreviations: LMWH 5 low-molecular-weight heparin; TE 5 thromboembolism
patients with a mechanical heart valve who are I nreceiving long-term oral anticoagulant therapy and are undergoing elective surgery, the management of perioperative anticoagulation therapy is *From the Department of Medicine (Drs. Douketis, Crowther, Mr. Cherian, and Dr. Kearon), St. Joseph’s Hospital (Drs. Douketis and Crowther), Hamilton, Ontario, Canada; and the Hamilton Health Sciences Corporation (Dr. Kearon), McMaster University, Hamilton, Ontario, Canada. Dr. Crowther is a recipient of a Research Fellowship from the Medical Research Council of Canada. Dr. Kearon is a recipient of a Research Scholarship from the Heart and Stroke Foundation of Ontario. Manuscript received January 11, 1999; revision accepted April 29, 1999. Correspondence to: James D. Douketis, MD, St. Joseph’s Hospital, Room F-513, 50 Charlton Ave East, Hamilton, ON, Canada, L8N 4A6; e-mail: [email protected] 1240
problematic. Temporary discontinuation of anticoagulants increases the risk of valve thrombosis and systemic embolism.1–3 On the other hand, continuing anticoagulants perioperatively, or stopping and restarting anticoagulants too soon after surgery, can cause life-threatening bleeding.4,5 The optimal anticoagulation strategy would minimize the risk of thromboembolism (TE), without causing excessive postoperative bleeding. To date, there is no consensus about the management of perioperative anticoagulation therapy in patients with a mechanical heart valve who are undergoing elective surgery, largely because of a lack of clinical trials investigating different anticoagulation strategies. The traditional management apClinical Investigations
proach has been to hospitalize patients 3 to 4 days before surgery to discontinue warfarin or other oral anticoagulants and to start IV heparin, which is stopped 3 h before surgery.6 –11 After surgery, IV heparin and warfarin are restarted when surgical hemostasis has been achieved, and heparin is continued until the anticoagulant effect of warfarin is within the therapeutic range. However, some authors have suggested that the risk of perioperative TE has been exaggerated,12,13 and that perioperative IV heparin should be reserved for patients at highest risk for TE, eg, previous TE.14,15 A less aggressive management approach involves temporary discontinuation of warfarin and the use of low-dose subcutaneous heparin postoperatively until patients are discharged from the hospital.14 Another approach, which is appealing for patients undergoing a surgical procedure that does not require overnight hospitalization, involves outpatient perioperative anticoagulant therapy with full-dose subcutaneous heparin or low-molecular-weight heparin (LMWH). As a first step in designing prospective studies to investigate different perioperative anticoagulation strategies, we performed a survey to determine physicians’ anticoagulation preferences in patients with a mechanical heart valve who are undergoing elective surgery and to determine the effect of different risks of TE and postoperative bleeding on anticoagulation preferences.
Table 1—Clinical Scenarios* 1. A 70-year-old woman with a mechanical mitral valve, chronic atrial fibrillation, and a previous stroke 2 years ago is to undergo elective (open) subtotal colectomy for resection of adenocarcinoma of the colon (TE risk, high; bleeding risk, high) 2. A 65-year-old man with a mechanical aortic valve is to undergo elective (open) subtotal colectomy for resection of adenocarcinoma of the colon (TE risk, low; bleeding risk, high) 3. A 75-year-old man with a mechanical mitral valve, a previous stroke 3 years ago, and chronic atrial fibrillation is to undergo elective bilateral inguinal hernia repair (TE risk, high; bleeding risk, low) 4. A 48-year-old woman with a mechanical aortic valve is to undergo elective bilateral inguinal hernia repair (TE risk, low; bleeding risk, low) *TE and bleeding risks for each scenario were not provided in the mailed survey.
requested information relating to a respondent’s medical specialty (general internal medicine, cardiology, cardiac surgery, or other) and their frequency of managing anticoagulant-related problems (frequently, infrequently, or not at all). Analysis The analysis consisted of four parts. (1) Anticoagulation preferences in different scenarios: The proportion of respondents who selected each of the anticoagulation options was determined for each scenario. (2) Effect of TE risk on preoperative and postoperative anticoagulation preferences: The proportion of respondents who preferred aggressive anticoagulant management (ie, options “a” or “b”) was compared in scenario pairs in which the TE risk was high in one scenario and low in the other
Materials and Methods Table 2—Anticoagulation Options* Survey Methods The survey was mailed to all members of the Canadian Society of Internal Medicine (n 5 420) and the Canadian Cardiovascular Society (n 5 540) in July 1997, with a repeat mailing in December 1997 to nonresponders. A priori, criteria were established to exclude returned surveys from the analysis: (1) physician does not manage anticoagulant-related problems in adults; (2) physician no longer resides at the address provided, ie, survey returned to sender; (3) survey was partially completed; and (4) physician has retired from medical practice. Survey Questions Physicians were asked to provide anticoagulation preferences in four scenarios of patients with a mechanical heart valve who are undergoing elective surgery (Table 1). The scenarios represented four different combinations of risks for TE (high or low) and bleeding (high or low). However, this information was not provided in the survey. For each scenario, physicians were asked to select from a standardized list of preoperative and postoperative anticoagulation options or, if these options were not acceptable, to provide an alternative anticoagulation strategy (Table 2). Options “a” and “b” (preoperative and postoperative) were classified as aggressive anticoagulant management, and options “c” (preoperative and postoperative) and “d” (postoperative) were classified as less aggressive anticoagulant management. We also
Preoperative anticoagulation options a: Admit to hospital 2 to 4 days preoperatively for full-dose IV heparin† b: Outpatient full-dose SC heparin or LMWH† c: Nothing else other than stopping warfarin preoperatively d: Other Postoperative anticoagulation options a: Full-dose in-hospital IV heparin until INR therapeutic† Heparin to be restarted , 6 h postoperatively 6 to 12 h postoperatively . 12 h postoperatively b: Early discharge home with full-dose SC heparin or LMWH until INR therapeutic† c: Low-dose in-hospital SC heparin or LMWH until INR therapeutic‡ d: Nothing else other than restarting warfarin postoperatively e: Other *These options are given in addition to stopping warfarin 4 to 5 days preoperatively and restarting warfarin 1 to 2 days postoperatively. INR 5 International Normalization Ratio; SC 5 subcutaneous injection. †Full-dose SC/IV heparin or SC LMWH is the dose recommended for treatment of venous TE or acute coronary syndromes. ‡Low-dose SC heparin or LMWH is the dose recommended as prophylaxis for venous TE. CHEST / 116 / 5 / NOVEMBER, 1999
1241
scenario, and the bleeding risk was the same in both scenarios. (3) Effect of bleeding risk on postoperative anticoagulation preferences: The proportion of respondents who preferred aggressive anticoagulant management (ie, options “a” or “b”) was compared in scenario pairs in which the bleeding risk was high in one scenario and low in the other scenario, and the TE risk was the same in both scenarios. (4) Effect of TE or bleeding risk on the timing of postoperative IV heparin initiation: The proportion of respondents who preferred early heparin initiation, defined as starting heparin within 12 h after surgery, was compared in scenario pairs in which the TE risk was high in one scenario and low in the other scenario, and the bleeding risk was the same in both scenarios; and in scenario pairs in which the bleeding risk was high in one scenario and low in the other scenario, and the TE risk was the same in both scenarios. McNemar’s x2 test was used to compare the proportion of respondents who preferred aggressive anticoagulant management within scenario pairs. Statistical significance was defined as p , 0.05.
Results Survey Response Rate and Characteristics of Respondents The survey was mailed to 960 physicians; 87 surveys were returned because the physician no longer resided at the address that was provided. Of 873 physicians who received the survey, 538 (62%) responded. Of 538 returned surveys, 65 were excluded for the following reasons: (1) physician does not manage anticoagulant-related problems in adults
(n 5 38), (2) physician has retired from medical practice (n 5 17), and (3) survey was partially completed (n 5 10). Thus, there were 473 surveys included in the analysis. Of the 473 respondents, 217 (46%) were cardiologists, 179 (38%) were general internists, 55 (12%) were cardiac surgeons, and 22 (4.6%) were from another specialty; 406 (86%) managed anticoagulant-related problems frequently and 67 (14%) managed them infrequently. Responses to Clinical Scenarios The anticoagulation preferences for the four scenarios are summarized in Figures 1 and 2. IV heparin was the most frequently selected anticoagulation option for preoperative and postoperative anticoagulation therapy in each of the four scenarios. Of those respondents who selected IV heparin for postoperative anticoagulation therapy, preferences for the timing of heparin initiation are summarized in Figure 3. Initiation of IV heparin 6 to 12 h after surgery was the most frequently selected timing for heparin initiation in each of the four scenarios, whereas heparin initiation , 6 h after surgery was least preferred by the respondents. The effect of TE risk on preoperative and postoperative anticoagulation preferences is summarized in Table 3. The risk of TE had a strong influence on
Figure 1. Preoperative anticoagulation management preferences. For description of different scenarios and management options, see Tables 1 and 2. 1242
Clinical Investigations
Figure 2. Postoperative anticoagulation management preferences. For description of different scenarios and management options, see Tables 1 and 2.
anticoagulation preferences. There was a significantly greater proportion of respondents who preferred aggressive preoperative and postoperative anticoagulant management in high-risk TE scenarios compared with low-risk TE scenarios. The effect of bleeding risk on postoperative anticoagulation preferences is summarized in Table 3. The risk of bleeding did not have a detectable effect on postoperative anticoagulation preferences, inasmuch as the proportion of respondents who preferred aggressive anticoagulant management was not significantly different in high-risk and low-risk bleeding scenarios. The effect of TE or bleeding risk on the timing of postoperative IV heparin initiation is summarized in Table 4. Of respondents who preferred IV heparin for postoperative management, the risk of bleeding influenced whether the timing of heparin initiation after surgery was early (ie, within 12 h) or later (ie, . 12 h). There was a significantly smaller proportion of respondents who preferred early heparin initiation in high-risk bleeding scenarios compared with lowrisk bleeding scenarios. The risk of TE did not have a detectable effect on the timing of heparin initiation, inasmuch as the proportion of respondents who preferred early heparin initiation was not significantly different in high-risk and low-risk TE scenarios. Discussion In this survey, we investigated physician preferences for perioperative anticoagulant therapy in
patients with a mechanical heart valve who are undergoing elective noncardiac surgery. There are three main findings from this survey. (1) Preoperative and postoperative IV heparin was the most frequently selected anticoagulation option. (2) The risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management. (3) If IV heparin was selected, the risk of bleeding influenced the timing of heparin initiation. Our finding that IV heparin was the most frequently selected anticoagulation option is noteworthy because the effectiveness and safety of this management approach has not been validated in clinical trials. The rationale for aggressive perioperative anticoagulation therapy with IV heparin is based on evidence from retrospective case series16,17 and one small prospective cohort study18 that did not provide reliable estimates of the risks of perioperative TE and bleeding. Furthermore, most patients in these studies had first-generation caged-ball heart valves, which are more thrombogenic than newer bileaflet tilting-disk models.5 In two methodologically rigorous studies that provided quantitative estimates of the risks of TE and bleeding for patients with a mechanical heart valve who require elective surgery, the investigators questioned the use of perioperative IV heparin, except in patients at highest risk for TE (ie, previous TE, caged-ball heart valve).14,15 One of these studies concluded that in the absence of an embolic event within the previous month, the use of IV heparin was expected to CHEST / 116 / 5 / NOVEMBER, 1999
1243
Figure 3. Postoperative IV heparin initiation preferences. For description of different scenarios, see Table 1.
markedly increase morbidity related to postoperative bleeding.14 Thus, although IV heparin was preferred by most respondents in this survey, the evidence to support this anticoagulation strategy is questionable. Table 3—Effect of TE or Bleeding Risk on Preoperative and Postoperative Anticoagulation Preferences* Anticoagulation Preferences Preoperative management Scenario 1 Scenario 2 Scenario 3 Scenario 4 Postoperative management Bleeding risk High Scenario 1 Scenario 2 Low Scenario 3 Scenario 4 p Value
TE Risk† High
Low
p Value , 0.001
434 (92) 312 (66)
, 0.001
416 (88)
Table 4 —Effect of TE or Bleeding Risk on the Timing of Postoperative IV Heparin Initiation*
288 (61)
TE Risk Bleeding Risk , 0.001
424 (90) 340 (72)
, 0.001
415 (88) 0.36
320 (68) 0.06
*Values given as No. (%). †Values represent the proportion of 473 respondents selecting aggressive anticoagulant management: IV heparin (option a); or full-dose SC heparin or LMWH (option b). For abbreviation definition, see Table 2. 1244
Another noteworthy finding is that despite convincing evidence that subcutaneous heparin or LMWH is as effective as IV heparin for the treatment of venous TE19,20 and acute coronary syndromes,21 this anticoagulation strategy, which does not require hospitalization, was selected by only 5.3 to 23% of respondents. Thus, demonstration that outpatient perioperative anticoagulation therapy with subcutaneous heparin or LMWH is effective and safe could have a major impact on clinical
High Scenario Scenario Low Scenario Scenario p Value
High
1 (n 5 399)† 2 (n 5 287)
250 (63)
3 (n 5 354) 4 (n 5 210)
268 (76)
Low
p Value 0.27
167 (58)
0.002
0.26 150 (71) 0.002
*Values given as No. (%). Numbers are given as a proportion of those respondents who selected aggressive heparin initiation (ie, within 12 h after surgery from among those who preferred IV heparin for postoperative anticoagulation therapy). †No. of respondents who selected aggressive heparin initiation. Clinical Investigations
practice. Further, this management approach is likely to be more cost-effective compared with the use of in-hospital IV heparin.15 A less aggressive anticoagulant management (ie, options “c” or “d”), which has been advocated by some authors,14 was preferred in the preoperative or postoperative setting by only 3.2 to 9.4% of respondents in high-risk TE scenarios, and by 11 to 36% of respondents in low-risk TE scenarios. It is likely that before a less aggressive anticoagulant management is adopted for wider use, the effectiveness of this approach will require validation in prospective clinical trials. Our finding that the risk of TE, but not the risk of bleeding, influenced the aggressiveness of anticoagulant management may be explained by the following considerations. First, physicians may consider the prevention of TE as the primary management objective because the clinical consequence of TE (eg, stroke), resulting from less aggressive anticoagulation therapy, is likely to be greater than that of bleeding (eg, wound hematoma), resulting from aggressive anticoagulation therapy. Second, physicians may be concerned about legal liability if a disabling stroke occurs perioperatively, whereas the occurrence of postoperative bleeding may be more defensible.22 However, there is evidence to suggest that the risk of perioperative TE has been exaggerated, and greater consideration should be given to the risk of postoperative bleeding and its clinical consequences as determinants of anticoagulation preferences. In a cohort study of 45 patients with a mechanical heart valve who underwent elective surgery and received perioperative IV heparin, 13 patients (29%) had postoperative bleeding complications, and in 1 patient (2.2%), the bleeding was fatal.23 In a meta-analysis of randomized controlled trials involving patients with a mechanical heart valve, in which one patient group did not receive anticoagulant therapy, the risk of TE in patients who were not anticoagulated was 9% per year.24 If the period in which patients are not fully anticoagulated is limited to 2 or 3 days, as in the perioperative period, the risk for TE during this period would be low (ie, 9% 3 3/365 5 0.07%). It is probable that the risk of TE would be higher in a high-risk subgroup (ie, mechanical mitral valve, atrial fibrillation). However, if most physicians are using IV heparin perioperatively, it is possible that this anticoagulation strategy is causing greater morbidity and deaths because of bleeding compared with the morbidity and deaths because of TE if IV heparin had not been used, particularly in patients at lower risk for TE (ie, mechanical aortic valve, sinus rhythm). Of those respondents who preferred IV heparin for postoperative anticoagulation therapy, the risk of bleeding influenced the timing of heparin initiation.
A significantly higher proportion of respondents preferred later heparin initiation (ie, . 12 h after surgery) in high-risk bleeding scenarios compared with low-risk bleeding scenarios. To our knowledge, there are no studies investigating when it is safe to initiate IV heparin after surgery. Presumably, this will depend on the surgical procedure, inasmuch as patients undergoing urologic, neurosurgical, or cancer-related procedures are more susceptible to postoperative bleeding.25,26 It is possible that withholding IV heparin for as little as 24 to 48 h after surgery might substantially reduce the risk of postoperative bleeding, although having a negligible effect on patients’ risk of TE. The are several points that support the validity and generalizability of our findings. First, the anticoagulation options we provided were chosen by . 98% of respondents, depending on the scenario, thereby suggesting that the responses in this survey are a valid representation of physicians’ usual clinical practice. Second, the respondents were from a broad cross-section of specialties that included general internal medicine, cardiology, and cardiac surgery. Third, because 86% of respondents managed anticoagulant-related problems frequently, our findings are generalizable to other physicians who manage anticoagulant-related problems. We acknowledge that the survey response rate of 62% might be considered suboptimal; however, this response rate is consistent with rates reported in other surveys of anticoagulant-related practices.27,28 To summarize, there is a need for large, prospective studies to investigate the efficacy and safety of the anticoagulation preferences identified in this survey, and to provide reliable estimates of the risks of perioperative TE and bleeding associated with different anticoagulation strategies. ACKNOWLEDGMENT: We thank Mrs. Teresa Tomlinson for her invaluable administrative assistance.
References 1 Stein PD, Alpert JS, Dalen JE, et al. Antithrombotic therapy in patients with mechanical and biological prosthetic heart valves. Chest 1998; 114(suppl 5):602S– 610S 2 Silber H, Kahn SS, Matloff JM, et al. The St. Jude valve: thrombolysis as the first level of therapy for cardiac valve thrombosis. Circulation 1993; 87:30 –37 3 Edmunds LH Jr. Thromboembolic complications of current cardiac valve prostheses. Ann Thorac Surg 1982; 34:96 –106 4 Levine M, Raskob GE, Landefeld S, et al. Hemorrhagic complications of anticoagulant treatment. Chest 1998; 114(suppl 5):511S–523S 5 Vongpatanasin W, Hillis D, Lange RA. Prosthetic heart valves. N Engl J Med 1996; 335:407– 416 6 Bryan AJ, Butchart EG. Prosthetic heart valves and anticoagulant management during non-cardiac surgery. Br J Surg 1995; 82:577–578 CHEST / 116 / 5 / NOVEMBER, 1999
1245
7 Travis S, Wray R, Harrison K. Perioperative anticoagulation control. Br J Surg 1989; 76:1107–1108 8 Madura JA, Rookstool M, Wease G. The management of patients on chronic Coumadin therapy undergoing subsequent surgical procedures. Am Surgeon 1994; 60:542–547 9 Busuttil WJ, Fabri BM. The management of anticoagulation in patients with prosthetic heart valves undergoing noncardiac operations. Postgrad Med J 1995; 71:390 –392 10 Tiede DJ, Nishimura RA, Gastineau DA, et al. Modern management of prosthetic valve anticoagulation. Mayo Clin Proc 1998; 73:665– 680 11 Blaustein AS. Preoperative and perioperative management of cardiac patients undergoing noncardiac surgery. Cardiol Clin 1995; 13:149 –160 12 Kuwada SK, Balm R, Gostout CJ. The risk of withdrawing chronic anticoagulation because of acute GI bleeding. Am J Gastroenterol 1996; 91:1116 –1119 13 Tinker JH, Tarhan S. Discontinuing anticoagulant therapy in surgical patients with cardiac valve prostheses. JAMA 1978; 239:738 –739 14 Kearon C, Hirsh J. Management of anticoagulation before and after elective surgery. N Engl J Med 1997; 336:1506 – 1511 15 Eckman MH, Beshansky JR, Durand-Zaleski I, et al. Anticoagulation for noncardiac procedures in patients with prosthetic heart valves. JAMA 1990; 263:1513–1521 16 Katholi RE, Nolan SP, McGuire LB. Living with prosthetic heart valves: subsequent noncardiac operations and the risk of thromboembolism of hemorrhage. Am Heart J 1976; 92:162– 167 17 Shean FC, Austen WG, Buckley MJ, et al. Survival after Starr-Edwards aortic valve replacement. Circulation 1971; 41:1– 8 18 Katholi RE, Nolan SP, McGuire LB. The management of anticoagulation during noncardiac operations in patients with prosthetic heart valves. Am Heart J 1978; 96:163–165 19 Low-molecular-weight heparin is an effective and safe treat-
1246
20
21
22
23
24 25 26
27
28
ment for deep vein thrombosis and pulmonary embolism: The Columbus Investigators. N Engl J Med 1997; 337:657– 662 Hommes DW, Bura A, Mazzolai L, et al. Subcutaneous heparin compared with continuous intravenous heparin administration in the initial treatment of deep vein thrombosis. Ann Intern Med 1992; 116:279 –284 Cohen M, Demers C, Gurfinkel EP, et al. A comparison of low-molecular-weight heparin with unfractionated heparin for unstable coronary artery disease: Efficacy and Safety of Subcutaneous Enoxaparin in Non-Q-Wave Coronary Events Study Group. N Engl J Med 1997; 337:447– 452 McIntyre K. Medicolegal implications of consensus statements with special attention to the Fifth Antithrombotic Therapy Consensus Conference. Chest 1998; 114(suppl 5): 742S–747S Darmon D, Enriquez-Sarano M, Acar J. Cardiac complications after subsequent non-cardiac operations in patients with non-biological prosthetic heart valves [abstract]. Eur J Cardiol 1983; 4(suppl 1):30 Cannegieter SC, Rosendaal FR, Briet E. Thromboembolic and bleeding complications in patients with mechanical heart valve prostheses. Circulation 1994; 89:635– 641 Bigg SW, Catalona WJ. Prophylactic mini-dose heparin in patients undergoing radical retropubic prostatectomy: a prospective trial. Urology 1992; 39:309 –313 Turpie AGG, Hirsh J, Gent M, et al. Prevention of deep vein thrombosis in potential neurosurgical patients: a randomized trial comparing graduated compression stockings alone or graduated compression stockings plus intermittent pneumatic compression. Arch Intern Med 1989; 149:679 – 681 Kutner M, Nixon G, Silverstone F. Physicians’ attitudes toward oral anticoagulants and antiplatelet agents for stroke prevention in elderly patients with atrial fibrillation. Arch Intern Med 1991; 151:1950 –1953 Chang HJ, Bell JR, Deroo DB, et al. Physician variation in anticoagulating patients with atrial fibrillation. Arch Intern Med 1990; 150:81– 84
Clinical Investigations