Perioperative anticoagulant management. SC indicates subcutaneously; BID, twice daily; and LMWH, low-molecular-weight heparin.
Douketis JD, Johnson JA, Turpie AG. Low-Molecular-Weight Heparin as Bridging Anticoagulation During Interruption of WarfarinAssessment of a Standardized Periprocedural Anticoagulation Regimen. Arch Intern Med. 2004;164(12):1319-1326. doi:10.1001/archinte.164.12.1319
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
The treatment of patients at increased risk for arterial thromboembolism who require temporary interruption of warfarin sodium therapy is a common clinical problem. We investigated the efficacy and safety of a standardized periprocedural anticoagulation regimen with low-molecular-weight heparin.
We studied 650 consecutive patients with a mechanical heart valve, chronic atrial fibrillation, or embolic stroke who required interruption of warfarin therapy because of an invasive procedure. Warfarin was stopped 5 or 6 days before the procedure, and patients received subcutaneous dalteparin sodium, 100 IU/kg twice daily, starting 3 days before the procedure. The risk of postprocedural bleeding determined postprocedural anticoagulant management. In patients undergoing a non–high-bleeding-risk procedure who had adequate postprocedural hemostasis, warfarin was resumed on the evening of the procedure, and dalteparin sodium, 100 IU/kg twice daily, was resumed on the next day and continued until the international normalized ratio was 2.0 or more. If postprocedural hemostasis was not secured, the resumption of dalteparin was delayed. In patients undergoing a high-bleeding-risk procedure, warfarin was resumed on the evening of the procedure, but dalteparin was not given after the procedure.
Patients were followed up during the preprocedural and postprocedural period for a mean of 13.8 days (range, 10-18 days). In 542 patients who underwent a non–high-bleeding-risk procedure, there were 2 thromboembolic events (0.4%), 4 major bleeding episodes (0.7%), and 32 episodes of increased wound-related blood loss that precluded postprocedural dalteparin administration (5.9%). In 108 patients who underwent a high-bleeding-risk procedure, there were 2 deaths (1.8%) possibly due to thromboembolism and 2 major bleeding episodes (1.8%).
In patients at increased risk for arterial thromboembolism who require temporary interruption of warfarin therapy, a standardized periprocedural anticoagulant regimen with low-molecular-weight heparin is associated with a low risk of thromboembolic and major bleeding complications.
The treatment of patients with a mechanical heart valve or chronic atrial fibrillation who require temporary interruption of warfarin sodium therapy because of surgery or another invasive procedure is a frequently encountered but underinvestigated clinical problem.1- 4 A major gap in knowledge is a lack of reliable estimates as to the incidence of thromboembolic events associated with warfarin therapy interruption.1 It is well established, however, that such events can have devastating clinical consequences: thrombosis of a mechanical heart valve is fatal in 15% of patients,5,6 and embolic stroke results in a major neurologic deficit or death in 70% of patients.7 Consequently, despite disagreement on the optimal periprocedural anticoagulation strategy during interruption of warfarin therapy,8- 12 several authorities1- 4 and consensus groups13,14 advocate, for most patients, some form of bridging therapy with a short-acting anticoagulant. The rationale for bridging anticoagulation is to minimize the time before and after a procedure that patients are not receiving therapeutic anticoagulation and, thereby, minimize the risk of thromboembolism. The conventional periprocedural anticoagulation approach is to hospitalize patients 4 to 5 days before surgery to stop warfarin and administer intravenous unfractionated heparin while the anticoagulant effect of warfarin recedes.15,16 Intravenous heparin is stopped 3 to 4 hours before the procedure to avoid a residual anticoagulant effect at the time of the procedure. After the procedure, warfarin and intravenous heparin are resumed, the latter administered for 4 to 5 days until therapeutic anticoagulation with warfarin is reestablished. This approach is difficult to implement because of the current constraints on hospital bed availability and the increasing number of surgical and other invasive procedures that are being performed without hospitalization.
An alternative periprocedural management strategy is the use of low-molecular-weight heparin (LMWH) for bridging anticoagulation. This approach is appealing because LMWH can be administered subcutaneously, in a fixed weight-based dose, and without the need for laboratory monitoring,17 thereby obviating the need for hospitalization to administer anticoagulants. Furthermore, this strategy has the potential to substantially reduce health care costs.18 However, there is little evidence to support the efficacy and safety of LMWH as bridging anticoagulation. Previous studies of periprocedural anticoagulation are limited because no standardized anticoagulant regimen was investigated,19- 22 anticoagulants other than LMWH were investigated,19- 23 or studies of LMWH as bridging therapy involved fewer than 30 patients.24- 26
We report on a prospective registry of 650 patients at risk for arterial thromboembolism who required temporary interruption of warfarin therapy and received bridging anticoagulation with LMWH. This is the first large-scale study, to our knowledge, assessing the efficacy and safety of a standardized periprocedural anticoagulation regimen with LMWH.
Starting in November 1997, a structured clinical management protocol for patients who required temporary interruption of warfarin therapy was initiated at the Hamilton Health Sciences, General Hospital, a tertiary care teaching hospital and regional center for cardiovascular and neurologic diseases in Hamilton, Ontario. Patients received a standardized regimen of periprocedural bridging anticoagulation with LMWH that was administered, whenever feasible, out-of-hospital by the patient or another caregiver. This management protocol was developed in response to an increasing number of elective surgical and other invasive procedures that were being undertaken without hospitalization in patients receiving warfarin, and because of limited hospital bed availability that precluded periprocedural intravenous heparin administration. To assess the efficacy and safety of this novel periprocedural anticoagulation strategy and to allow internal auditing of patient quality of care, we established a prospective registry of patients who received bridging anticoagulation with LMWH. Because this periprocedural anticoagulation regimen became the standard of care at our institution, informed consent was not required for inclusion into this patient registry.
Consecutive adult patients, 18 years or older, at risk for arterial thromboembolism who were assessed in the hospital Anticoagulation Clinic between November 1, 1997, and June 30, 2002, for temporary interruption of warfarin therapy were eligible for this registry. Patients satisfied the following criteria for inclusion: (1) receiving warfarin therapy, with a target international normalized ratio (INR) of 2.0 to 3.5; (2) mechanical heart valve, chronic atrial fibrillation, or a previous stroke or transient ischemic attack with a presumed embolic source; and (3) undergoing an elective surgical or other invasive procedure that requires normalization of the INR. Patients were excluded from the registry if one or more of the following characteristics were present: (1) renal insufficiency (serum creatinine level >2.0 mg/dL [>178 mmol/L]) that precluded therapeutic-dose LMWH administration; (2) previous heparin-induced thrombocytopenia; (3) pregnancy; (4) treatment with an anticoagulant other than the prespecified LMWH regimen; or (5) undergoing a minor dental procedure, such as teeth cleaning or single-tooth extraction, in which reversal of anticoagulation is not required.27 Patients who were to receive spinal anesthesia were eligible for this registry but were excluded if they had an indwelling epidural catheter after the procedure for analgesia, which precluded coadministered LMWH.28 Patients who required rapid reversal of anticoagulation because of an urgent procedure were not included in this registry.
Patients received a standardized anticoagulation regimen, as outlined in Table 1, that involved therapeutic-dose LMWH as bridging therapy because less intense regimens, with low-dose LMWH, had not been previously assessed for efficacy. Patients were instructed in subcutaneous self-injection of LMWH by the Anticoagulation Clinic nurse practitioner, and syringes were prefilled to minimize dose errors. Patients were provided with information about potential bleeding and thromboembolic complications, and were asked to report such events if they occurred or were suspected. A 24-hour telephone contact number was provided if patients had difficulty with the LMWH injections or other health-related concerns. If a patient or caregiver was unable to administer LMWH, it was undertaken by a visiting nurse.
For preprocedural anticoagulation, in general, warfarin therapy was interrupted on the fifth day before the procedure in patients receiving warfarin with a target INR of 2.0 to 3.0, and on the sixth day before the procedure in patients receiving warfarin with a target INR of 2.5 to 3.5. The INR testing was done on the third or fourth day before the procedure. If the INR was less than 2.5, treatment was initiated with dalteparin sodium, 100 IU/kg twice daily by subcutaneous injection, with the first dose supervised by the Anticoagulation Clinic nurse. If the INR was 2.5 or more, thereby precluding the initiation of dalteparin therapy, an injection of isotonic sodium chloride solution was administered subcutaneously to observe the patient's self-injection technique, and the first dose of dalteparin was given on the next day. To minimize the likelihood of a residual anticoagulant effect at the time of the procedure, the last preprocedure dalteparin dose was administered not less than 12 hours before the start of the procedure, and if the INR was 3.0 or more on the third or fourth day before the procedure, patients received phytonadione, 1 mg orally. If patients were receiving antiplatelet therapy, usually aspirin, this was interrupted 7 days before the procedure.
The postprocedural anticoagulation regimen was determined on the basis of 2 considerations: (1) the bleeding risk associated with the procedure and (2) the adequacy of postprocedural hemostasis. Patients were classified as being at high risk or nonhigh risk of postprocedural bleeding. This classification of bleeding risk was based on a subjective assessment of the anticipated procedure-related bleeding, which was determined by input from the anticoagulation physician and nurse-practitioner and the attending surgeon or interventionist. The adequacy of postprocedural hemostasis was determined by the same caregivers and was based on a subjective assessment of wound-related blood loss.
In patients undergoing a high-bleeding-risk procedure, warfarin therapy was resumed on the evening after the procedure, but patients did not receive dalteparin at any time after the procedure. In patients undergoing a non–high-bleeding-risk procedure who had adequate postprocedural hemostasis, warfarin was resumed on the evening of procedure, and dalteparin sodium, 100 IU/kg twice daily, was resumed on the day after the procedure, with the first dose administered approximately 24 hours after the procedure. Dalteparin was continued until the INR was 2.0 or more. In patients undergoing a non–high-bleeding-risk procedure who had inadequate postprocedural hemostasis, the resumption of warfarin was delayed until the first postprocedural day, and dalteparin was delayed until the second or third postprocedural day when hemostasis was secured. In all patients, the initial postprocedural dose of warfarin corresponded to the patient's usual dose for that day of the week, and antiplatelet therapy was resumed on the same day as dalteparin.
Three primary clinical outcomes were assessed in all patients: (1) thromboembolism, (2) major bleeding, and (3) death. Thromboembolism was defined as a stroke or transient ischemic attack, an acute coronary syndrome, or systemic embolism involving a limb or viscera that was ascertained on the basis of clinical findings and objective diagnostic testing. Major (or clinically important) bleeding was defined as bleeding that was clinically overt, with new symptoms (eg, hematemesis) or signs (eg, wound hematoma), and was associated with one or more of the following characteristics: hemoglobin level decrease of more than 20 g/L over 24 hours; transfusion of 2 U or more of packed red blood cells; or bleeding at a critical site (retroperitoneal, intracranial, or body cavity).29 Patients who had a decrease in hemoglobin level or received a blood transfusion because of expected procedure-related blood loss, but with no clinically overt bleeding, were not considered to have major bleeding. Death was documented on the basis of information from patients' medical records or death certificate, and if it occurred suddenly, it was classified as possibly due to thromboembolism.
A secondary clinical outcome was increased wound-related blood loss, which was assessed only in patients undergoing non–high-bleeding-risk procedure in whom the presence of increased blood loss would delay or suspend the postprocedural resumption of dalteparin therapy. This subjective outcome was characterized by greater-than-expected oozing from a wound site or blood accumulation in a wound drain.
Patients underwent follow-up from the time warfarin therapy was interrupted before the procedure until therapeutic anticoagulation with warfarin, defined by an INR 2.0 or more, was reestablished after the procedure. Subsequent warfarin dosing was undertaken by the Anticoagulation Clinic or the patient's primary care physician.
Statistical analyses were done with SAS 8.2 software (SAS Institute Inc, Cary, NC). Baseline patient characteristics and periprocedural anticoagulation regimens were expressed as a mean with a corresponding standard deviation. Clinical outcome rates were based on the number of patients with a thromboembolic or bleeding event divided by the number of patients at risk and were expressed as a proportion with a corresponding 95% confidence interval.
Of 798 patients at risk for arterial thromboembolism who were assessed for temporary interruption of warfarin because of an elective surgical or other invasive procedure, 104 were excluded from the registry because the patient was undergoing a minor dental procedure (n = 89), had renal insufficiency (n = 14), or had previous heparin-induced thrombocytopenia (n = 1). Of 694 registry patients who were scheduled to receive the prespecified anticoagulation regimen, 35 were excluded because they received an LMWH other than dalteparin before or after the procedure, and 9 were excluded because they received intravenous heparin after the procedure at the discretion of the treating physician (Figure 1). Thus, there were 650 patients scheduled to receive the prespecified anticoagulation regimen, of whom 542 (83%) were undergoing a non–high-bleeding-risk procedure, and 108 (17%) were undergoing a high-bleeding-risk procedure. The baseline patient characteristics are described in Table 2. The type of procedure patients underwent and the associated bleeding risk category are described in Table 3. All patients received out-of-hospital dalteparin injections before the procedure, which was administered by the patient (self-injection), a family member, or a visiting nurse in 88%, 7%, and 5% of cases, respectively. After the procedure, dalteparin was administered by the patient, a family member, or a visiting nurse in 85%, 7%, and 8% of cases, respectively.
The periprocedural anticoagulation dosing regimen patients received is outlined in Table 4 according to patients' indications for warfarin therapy and their corresponding target INR.
For preprocedural anticoagulation, warfarin therapy was interrupted a mean of 5.7 days before the procedure, and patients received a mean of 5.4 doses of dalteparin.
For postprocedural anticoagulation, in 108 patients who underwent a high-bleeding-risk procedure and were not scheduled to receive only postprocedural dalteparin, warfarin was resumed on the evening of or the day after the procedure. In the 542 patients who underwent a non–high-bleeding-risk procedure, there was excessive postprocedural bleeding in 32 patients that precluded postprocedural dalteparin administration, and these patients received only warfarin, starting the day after the procedure. In 510 patients who underwent a non–high-bleeding-risk procedure and received postprocedural dalteparin, it was resumed on the first postprocedure day in 454 patients (89%), at least 24 hours after the procedure, on the second postprocedural day in 46 patients (9%), on the third postprocedural day in 5 patients (1%), and on the fourth postprocedural day in 5 patients (1%). This delay in resumption of postprocedural dalteparin in 56 patients was necessitated by delayed hemostasis (23 patients) and concerns about bleeding after implantation of a permanent pacemaker or internal cardiac defibrillator (30 patients) or bowel polypectomy (3 patients).
Patients underwent clinical follow-up during the preprocedural and postprocedural period for a mean of 13.8 days (range, 10-18 days). The rates of adverse clinical outcomes are outlined in Table 5. In 108 patients undergoing a high-bleeding-risk procedure who received only preprocedural dalteparin, there were 2 nonfatal major bleeding episodes (1.8%) (upper gastrointestinal tract bleeding and wound hematoma), neither of which was fatal; no confirmed thromboembolic events; and 2 deaths (1.8%) (cardiac arrests) possibly due to thromboembolism. In 542 patients undergoing a non–high-bleeding-risk procedure who were scheduled to received preprocedural and postprocedural dalteparin, there were 4 major bleeding episodes (0.7%) (3 wound hematomas and a rectus abdominus sheath hematoma), none of which were fatal; 2 thromboembolic events (0.4%) (systemic embolism and transient ischemic attack); and 32 episodes (5.9%) of excessive wound-related bleeding that precluded postprocedural administration of dalteparin. The adverse clinical outcomes, which all occurred during the postprocedural period, are described in Table 6.
In this registry of 650 patients at increased risk for arterial thromboembolism who required temporary interruption of warfarin therapy for an elective procedure, a standardized bridging anticoagulation regimen with LMWH was associated with a low risk of thromboembolic (<1%) and major bleeding (1%-2%) complications in the immediate periprocedural period, and was feasible for out-of-hospital administration. This anticoagulation strategy involved 2 main components: (1) preprocedural and postprocedural dalteparin sodium therapy, 100 IU/kg twice daily, to minimize the duration without therapeutic anticoagulation; and (2) avoidance of postprocedural dalteparin in patients at high risk of bleeding or with inadequate postprocedural hemostasis, and delay in dalteparin resumption in patients with delayed hemostasis. The second component was of paramount importance because the development of postprocedural major bleeding would delay the resumption of anticoagulation and, thereby, increase the time that patients would be exposed to the risk of thromboembolism.
This patient registry has several features that support the validity of the results. First, unlike most registries in which patient treatment is left to the discretion of the treating physician,30,31 patients in this registry received a standardized periprocedural anticoagulation regimen. This would reduce the potential for variability in the aggressiveness of periprocedural anticoagulation8,11,12 or cointerventions such as periprocedural aspirin therapy that could influence the risk of thromboembolic and bleeding events.32,33 Second, patients underwent a standardized duration of clinical follow-up, thereby limiting variability in patient surveillance that would influence the likelihood of capturing clinical outcomes. Third, unlike most registries that allow broad patient eligibility criteria, this registry assessed a prespecified patient group at increased risk for arterial thromboembolism, in whom most physicians would consider some form of periprocedural bridging anticoagulation.1- 4,13,14
Clinically important (or major) bleeding complications were infrequent, occurring in 1% to 2% of patients. The low incidence of major bleeding is likely attributable to 3 factors. First, 52% of patients underwent nonsurgical invasive procedures, such as cardiac catheterization or gastrointestinal endoscopy, which are associated with a low risk of bleeding complications.34,35 Second, measures were taken to minimize the risk of bleeding. Postprocedural dalteparin was avoided in patients undergoing a high-bleeding-risk procedure such as open heart surgery,36 abdominal vascular surgery,37 neurosurgery,38 major cancer surgery,39,40 or procedures involving the prostate or kidney.41- 43 In patients undergoing a non–high-bleeding-risk procedure, the start of dalteparin was delayed for 24 hours until the day after the procedure, when there was adequate wound hemostasis. This was done because in patients who receive LMWH after a surgical procedure, more than 90% of major bleeding episodes occur at the surgical wound site,44- 46 and such bleeding is more likely to occur with early postprocedural initiation of anticoagulant therapy, 4 to 12 hours after surgery, than delayed initiation of anticoagulation, more than 12 hours after surgery.23,47 Furthermore, the start of dalteparin was delayed or suspended in patients with inadequate postprocedural hemostasis, and in those undergoing pacemaker or cardiac defibrillator implantation or bowel polypectomy, who might be at increased risk of bleeding complications with early postprocedural initiation of anticoagulation.48,49 Third, our definition of major bleeding was narrower than that used in other studies involving postoperative anticoagulation,44,47 as it included only clinically overt bleeding. There is a need for a bleeding classification scheme specific for the postprocedural clinical setting that differentiates between clinically important bleeding that requires a change in clinical management (eg, reoperation) and increased wound-related blood loss that is expected and may require transfusion but is, otherwise, self-limiting and without clinical consequences.50
Thromboembolic complications were also infrequent. There was 1 episode of systemic embolism to the lower extremities, 1 transient ischemic attack and 2 deaths, possibly due to thromboembolism, yielding an overall incidence of 0.6%. Reliable estimates of periprocedural thromboembolic complications during temporary interruption of warfarin treatment are limited. It is postulated that the absolute risk of thromboembolic events during the 2- to 4-day period after warfarin interruption, when there is no anticoagulant effect present, is a small fraction of the 5% to 15% annual risk of thromboembolic events.8 However, thromboembolic risk may be higher than expected because after interruption of warfarin therapy a transient hypercoagulable state may develop owing to a rebound increase in thrombin generation or platelet activation.51,52 Furthermore, a surgical procedure may induce a hypercoagulable state through mechanisms that include vessel-wall injury and fibrinolysis inhibition.53- 55 In 3 prospective studies that have investigated periprocedural bridging anticoagulation with LMWH,25,26,56 thromboembolic events occurred in 1 (0.8%) of 128 of patients, an event rate that is consistent with our findings.
There are limitations of this study that should be addressed. First, we acknowledge that postprocedural patient follow-up was limited to about 1 week in most patients. Consequently, our findings may underestimate the risk of thromboembolic events because clinical manifestations of periprocedural thrombus formation, such as embolic stroke or valve thrombosis, may be delayed for several weeks after warfarin interruption.23 However, it is likely that most clinically overt bleeding episodes attributed to periprocedural bridging anticoagulation would be captured, as they are most likely to occur during the first week after a procedure.33,44- 46,56 Second, we acknowledge that our classification of patients according to the procedure-related bleeding risk, which determined whether postprocedural dalteparin was given, was subjective and has not been validated. Such a classification of patients was deemed necessary to minimize postprocedural bleeding and was supported by studies that have identified high-bleeding-risk procedures.36- 43 There are no clinical prediction rules, to our knowledge, that stratify patients according to the risk of postprocedural bleeding.57 Third, because this patient registry did not include a control group, we cannot comment on the efficacy or safety of our periprocedural anticoagulation regimen compared with other strategies that might involve intravenous heparin,11,12 or warfarin interruption without bridging anticoagulation.8,9 Finally, because this registry was undertaken at a single center that does not have a major focus on pulmonary or renal disease, there was underrepresentation of patients undergoing respiratory or nephrologic procedures. Consequently, our findings may not be generalizable to such patients, especially those with significant renal insufficiency in whom therapeutic-dose LMWH should be used with caution because of the potential for bioaccumulation.58 Additional studies are needed to confirm our findings in patients who have a longer duration of postprocedural follow-up and to identify high-bleeding-risk procedures in which postprocedural anticoagulation should be used with caution. Furthermore, there is a need for clinical trials to compare our periprocedural anticoagulation regimen with less intense strategies (eg, low-dose LMWH) or alternative oral anticoagulants such as direct thrombin inhibitors, which, because of a short half-life, may obviate the need for bridging therapy with LMWH.59
To summarize, in patients at increased risk for arterial thromboembolism who require temporary interruption of warfarin therapy, a standardized periprocedural anticoagulant regimen with LMWH is associated with a low risk of thromboembolic and major bleeding complications in the immediate periprocedural period, and is feasible for outpatient administration. Additional studies are needed to confirm these findings and investigate alternative periprocedural anticoagulation strategies.
Corresponding author: James D. Douketis, MD, FRCPC, Department of Medicine, St Joseph's Hospital, Room F-541, 50 Charlton Ave E, Hamilton, Ontario, Canada L8N 4A6 (e-mail: email@example.com).
Accepted for publication August 25, 2003.
There was no funding for this study. Dr Douketis is a recipient of a Research Scholarship from the Heart and Stroke Foundation of Canada, Ottawa, Canada.
We express sincere gratitude to Nicole Archer for establishing the patient registry database and to Nicole Zytaryk and Lauren Griffith for statistical assistance. We also thank Mark Crowther, MD, MSc, and Clive Kearon, MD, PhD, for their helpful reviews of the manuscript.