Find synonyms Find exact match
INTRODUCTION — The role for anticoagulants in many cardiovascular disorders is well established, and their use as prophylaxis against stroke or thromboembolism is increasing. As a result, many patients undergoing elective surgery or an invasive procedure may be taking one of these agents. The management of anticoagulation in such patients both before and after such procedures will be reviewed here. Other issues related to anticoagulation in the perioperative period are discussed separately:
●Rapid reversal of excess warfarin anticoagulation (see "Correcting excess anticoagulation after warfarin", section on 'Surgery/invasive procedure')
●Management of antiplatelet agents (see "Perioperative medication management", section on 'Medications affecting hemostasis')
●Use of warfarin or antiplatelet agents during eye surgery (see "Cataract in adults", section on 'Antithrombotic agents')
PROBLEM OVERVIEW — Although continuation of anticoagulation increases the risk of bleeding following invasive procedures, interruption of such therapy may increase the risk of thromboembolism in patients taking anticoagulants to prevent thrombosis [1-5].
Several factors are important in managing anticoagulation perioperatively, including the type the underlying indication for anticoagulation, the specific agent used, and the type (and urgency) of the surgical procedure. (See 'Type of surgery or procedure' below.)
Many patients receiving continuous anticoagulation are taking a vitamin K antagonist, most frequently warfarin. Warfarin has a biological half-life of 36 to 42 hours, considerably longer than that of the vitamin K antagonist acenocoumarol (8 to 11 hours) and shorter than that of phenprocoumon (3 to 5 days) and fluindione (69 hours). (See "Therapeutic use of warfarin and other vitamin K antagonists", section on 'Biological properties'.)
Increasingly, target-specific oral anticoagulants are being used for a variety of indications. These agents include direct thrombin inhibitors (eg, dabigatran) and factor Xa inhibitors (eg, rivaroxaban, apixaban, edoxaban). Important features that differ from vitamin K antagonists include the shorter half-lives of these agents and the lack of a specific antidote or reversal strategy. (See "Anticoagulation with direct thrombin inhibitors and factor Xa inhibitors".)
Accordingly, individual circumstances should be carefully reviewed before an informed decision on modifying anticoagulation therapy is made in the patient undergoing surgery or an invasive procedure.
There is also concern about the following issues in anticoagulated patients:
●Rebound hypercoagulability may occur following the abrupt cessation of anticoagulation.
●Several days may be required after warfarin therapy is resumed to reestablish a therapeutic and adequate level of anticoagulation.
The importance of these issues varies in part with the indication for anticoagulation (eg, prophylaxis for thromboembolism versus treatment for an acute thrombotic episode). Accordingly, there is no general recommendation that can be applied to all patients undergoing elective surgery who are taking long-term anticoagulation.
THROMBOTIC RISK IF ANTICOAGULATION IS STOPPED — Patients may be taking an anticoagulant as prophylaxis against the development of new thrombi or embolism (eg, atrial fibrillation, severe myocardial dysfunction, prosthetic heart valve), or as treatment of acute thrombus-related problems (eg, deep venous thrombosis [DVT] or pulmonary embolism [PE]).
Cessation of anticoagulation used to treat an acute thrombotic event may exacerbate the condition, which may itself be life threatening. The risk of under-anticoagulation varies with the type of thromboembolic event. While recurrent DVT or PE is associated with an approximately 5 percent risk of fatal pulmonary embolism [6,7], the consequences of arterial thromboembolism from atrial fibrillation or prosthetic heart valves are much more serious, with 20 percent of episodes being fatal and 40 percent causing permanent disability [8-11]. The accompanying table has been formulated in order to estimate this risk (table 1) [2,12].
Appropriate use of alternative strategies, such as intravenous heparin or subcutaneous low molecular weight (LMW) heparin to provide antithrombotic coverage (ie, "bridging" anticoagulation) during the period when warfarin is withdrawn or reintroduced has been utilized in an attempt to minimize the risks involved. (See 'Overview' below.)
Patients at risk for recurrent venous thromboembolism — Long-term anticoagulation is the recommended treatment for patients at high risk of recurrence of venous thromboembolism. Discontinuation of warfarin in such patients is associated with a significant risk of thromboembolism as high as 15 percent per year; warfarin probably reduces this risk by approximately 90 percent [13-15].
After an acute episode of venous thromboembolism, the recurrence risk is highest within the initial three to four weeks, and diminishes over the following two months [16,17]. Without anticoagulation, the early risk of recurrent venous thromboembolism is approximately 50 percent, but treatment with warfarin for one month reduces this risk to 8 to 10 percent, and to 4 to 5 percent after three months of warfarin therapy [18,19]. Thus, discontinuing warfarin anticoagulation within the first month after an acute venous thromboembolic episode is associated with a high risk of recurrent venous thromboembolism ; this risk is reduced if surgery is delayed and there is a longer period of warfarin treatment.
Patients at risk for arterial thromboembolism — The risk of recurrent arterial embolism from any cardiac source is approximately 0.5 percent per day in the first month after an acute event . The two most common clinical settings in which this is a major concern (ie, chronic atrial fibrillation, prosthetic heart valves) are discussed below. This risk is reduced by two-thirds with warfarin.
Chronic atrial fibrillation — Arterial thromboembolism is most commonly associated with atrial fibrillation; embolic stroke is fatal or associated with a severe neurologic deficit in over 60 percent of these patients [8,9]. Patients with atrial fibrillation not due to valvular heart disease have an overall risk of systemic embolism of 4 to 5 percent per year in the absence of warfarin therapy; anticoagulation reduces the risk of embolization by approximately two-thirds in this setting [10,22].
However, patients with atrial fibrillation are not a homogeneous group, and the risk of stroke and thromboembolism varies. Management should therefore be tailored to an individual patient's risk of thromboembolism as compared with the risk of bleeding during surgery. It is possible to risk stratify such patients based upon clinical and echocardiographic criteria. (See "Risk of embolization in atrial fibrillation".)
The balance of risks in patients with chronic atrial fibrillation (ie, thrombosis if anticoagulation is stopped for an invasive procedure versus bleeding if bridging anticoagulation is employed) was reviewed in a retrospective analysis of results from the randomized RE-LY trial of warfarin versus dabigatran for atrial fibrillation. While enrolled in this study, 4591 individuals underwent at least one invasive procedure during which anticoagulation with warfarin or dabigatran was temporarily stopped; bridging anticoagulation with intravenous heparin or LMW heparin was employed in 15 to 28 percent of these procedures . The incidence of ischemic stroke or systemic embolism for these 4591 individuals was 0.5 percent, while the incidence of major periprocedural bleeding varied from 3.8 to 5.1 percent. Thus, using the interruption approach taken in RE-LY, there were eight times more major bleeding events than strokes in the periprocedural period for both agents . These results call into question whether bridging anticoagulation benefits patients with chronic atrial fibrillation, but definitive conclusions await the results of ongoing randomized trials comparing bridging and no bridging perioperative strategies.
Prosthetic heart valves — Systemic embolization (predominantly cerebrovascular events) occurs at a frequency of approximately 0.7 to 1.0 percent per patient per year in patients with mechanical valves who are treated with warfarin, 2.2 percent per patient per year with aspirin, and 4.0 percent with no anticoagulation. A major advantage of the bioprosthetic valve is freedom from anticoagulation after three months of treatment and low risk for systemic embolism thereafter .
The management of anticoagulation in such patients, both in general and in the perioperative period, is discussed in depth separately. (See "Antithrombotic therapy in patients with prosthetic heart valves", section on 'Interruption of warfarin for surgical procedures'.)
Other high risk settings — Other causes of thromboembolism include a dilated and poorly contractile left ventricle or a left ventricular aneurysm in which intraventricular thrombi may form and embolize [26-28].
●Among patients with left ventricular dysfunction, one report found an 18 percent increase in stroke risk for every 5 percent reduction in left ventricular ejection fraction, although the absolute short-term risk was low . Anticoagulation with warfarin was associated with an 81 percent reduction in total stroke risk while aspirin therapy reduced the risk by 56 percent. (See "Antithrombotic therapy in patients with heart failure".)
Among patients with left ventricular aneurysm, the frequency of left ventricular thrombi in aneurysms reported by postmortem studies can range between 14 and 68 percent, a value consistent with findings at the time of surgical aneurysmectomy (50 to 95 percent) .
BLEEDING RISK IF ANTICOAGULATION IS CONTINUED — The risk of bleeding occurring with surgery in patients taking anticoagulant therapy is dependent upon patient age, the presence of other disease states (eg, chronic renal disease), the type of surgery or procedure the patient is undergoing , the anticoagulant regimen, intensity, and duration, the use of other drugs that affect hemostasis (eg, heparin, aspirin, antiplatelet agents), the stability of anticoagulation, and the degree of anticoagulation [3,30,31].
Type of surgery or procedure — Prolonged, complex, and major surgery is much more likely to cause significant bleeding problems than short, simple, and minor surgical procedures. As examples:
●Low bleeding risk procedures — Most patients can undergo low risk surgical procedures (eg, cataract surgery, coronary arteriography, venography, joint aspiration, dental procedures such as tooth extraction and root canal, minor skin procedures, arthrocentesis, bone marrow biopsy) without alteration of their anticoagulation regimen [32-34]. In such patients, oral anticoagulation with a vitamin K antagonist can be continued at or below the low end of the therapeutic range (eg, INR 1.7 to 2.3). (See "Cataract in adults", section on 'Antithrombotic agents'.)
●High bleeding risk procedures — More complex or high risk surgical procedures (eg, open-heart surgery, abdominal vascular surgery, intracranial or spinal surgery, major cancer surgery, urologic procedures) require discontinuation of a vitamin K antagonist (eg, warfarin), followed by temporary perioperative coverage with unfractionated heparin or LMW heparin in those patients who are at high risk of thromboembolism . (See 'Perioperative interruption and resumption of anticoagulation' below.)
Gastroenterologic procedures — Management of anticoagulation in patients undergoing gastroenterologic procedures (eg, endoscopy with or without mucosal biopsy), as with any other surgery or procedure, is anchored on the estimated risk for bleeding associated with the procedure and the estimated thromboembolic risk if the patient temporarily stops anticoagulation (table 3 and table 4 and table 5) . This subject is discussed in detail separately. (See "Management of anticoagulants in patients undergoing endoscopic procedures", section on 'Elective procedures in anticoagulated patients'.)
Dental or excisional cutaneous procedures — In patients undergoing dental extraction, warfarin anticoagulation is associated with a minimal risk of serious bleeding if the INR is within the therapeutic range just prior to the contemplated surgery [37-42]. Tranexamic acid or aminocaproic acid mouthwash, if available (eg, 4.8 to 5 percent aqueous solutions used four times per day for at least two days), can be used in anticoagulated patients to limit gingival bleeding after dental procedures [12,42-46]. The use of aspirin, NSAIDS, or Cox-2 selective inhibitors for analgesia should be avoided.
There are documented cases of rare but serious embolic events when warfarin has been withdrawn prior to dental procedures. In a literature review of 542 documented cases in 493 patients in whom warfarin was withdrawn for a dental procedure (without heparin bridging), there were five serious embolic complications (0.9 percent of cases) .
Vitamin K antagonists are generally safe in patients undergoing minor skin procedures (eg, skin cancer removal) if the INR is maintained within the therapeutic range [29,47-49]. As with dental procedures, cessation of prophylactic anticoagulation (warfarin or antiplatelet therapy) has been associated with a small risk of thromboembolic events [47-49].
Invasive cardiac procedures — A prospective, observational registry (BNK Online bRiDging REgistRy, BORDER) of patients treated by cardiologists evaluated current practice of perioperative management of oral anticoagulant therapy with a vitamin K antagonist (eg, warfarin) in a large outpatient cohort of patients undergoing 1000 invasive procedures (eg, cardiac catheterization, pacemaker implantation, surgery) .
●Ninety-four percent of these patients received bridging therapy (LMW heparin in all but two cases) during interruption of the vitamin K antagonist. LMW heparin was given in prophylactic, half-therapeutic, or full therapeutic doses in 2, 78, and 20 percent, respectively.
●Four thromboembolic complications were observed during 30 days of follow-up (0.4 percent).
●One major (0.1 percent) and 35 clinically relevant bleeding episodes (3.5 percent) occurred. Independent predictors of bleeding included a history of mechanical heart valve replacement and the HAS-BLED bleeding risk score (table 6). A HAS-BLED risk score ≥3 was the most predictive variable for hemorrhage (HR 12).
Use of heparin — The risk of bleeding after the use of heparin is variable. A two- or three-day course of intravenous heparin before surgery, with cessation four hours before the procedure, is unlikely to result in pre- or intra-operative bleeding. In the non-perioperative setting, the risk of bleeding associated with intravenous heparin therapy is less than 5 percent in patients with acute venous thromboembolism. However, in patients with deep vein thrombosis who are judged to be at high risk for bleeding, the incidence of major bleeding is approximately 11 percent during the first five days of intravenous heparin therapy [31,51].
PERIOPERATIVE INTERRUPTION AND RESUMPTION OF ANTICOAGULATION
Interruption of warfarin — After stopping warfarin, it usually takes two to three days for the INR to fall to below 2.0, and four to six days for the INR to normalize. One study prospectively evaluated 22 patients with a baseline INR of 2.6 in whom it was deemed safe to discontinue warfarin . In these patients the mean INR was 1.6 and 1.2 at 2.7 and 4.7 days after discontinuation of warfarin, respectively. The time required for the INR to normalize after stopping warfarin may be longer in patients receiving higher-intensity anticoagulation (INR: 2.5 to 3.5), and in elderly patients .
Once the INR is 1.5 or below, surgery can be performed with relative safety in most cases, although a normalized INR is typically required in patients undergoing surgery associated with a high bleeding risk (eg, intracranial, spinal, urologic) or if spinal anesthesia is to be used. Following surgery and after warfarin is restarted, it takes approximately five days for the INR to rise above 2.0. It is therefore estimated that if warfarin is withheld for five days before surgery and is restarted as soon as possible afterwards, patients would have a subtherapeutic INR for approximately four days before surgery and four days after surgery .
A slight elevation of the INR to approximately 1.5 around the time of surgery, in theory, should provide partial protection against venous thromboembolism [54,55]. However, there is no evidence that such low-intensity perioperative anticoagulation, for example that which is used for prevention of postoperative DVT, effectively prevents arterial thromboembolism. (See "Drug-induced thrombosis and vascular disease in patients with malignancy", section on 'Prophylactic anticoagulation'.)
If the patient has been adequately anticoagulated for some time prior to stopping warfarin, it is generally assumed that almost any preexisting thrombus would have either resolved or be endothelialized, thereby minimizing the risk of embolism . Among patients with nonvalvular atrial fibrillation, for example, over 85 percent of thrombi resolve after four weeks of warfarin therapy as determined by transesophageal echocardiography .
Nevertheless, although the INR itself may not be a good guide to a reduced risk of thromboembolism, some patients have a significant reduction in their usual anticoagulant intensity during surgery and a minor increase in the risk of thromboembolism is probably unavoidable . Among patients with atrial fibrillation, chronic low dose warfarin plus aspirin is much less effective than adjusted dose warfarin in preventing embolic events (figure 1 and figure 2), demonstrating that such lesser degrees of anticoagulation do not provide optimal protection . (See "Antithrombotic therapy to prevent embolization in atrial fibrillation".)
Timing of interruption and reversal — Reversing the anticoagulant activity of warfarin and other vitamin K antagonists depends upon the amount of time available before the surgical or invasive procedure, the elimination half-life of the vitamin K antagonist (ie, 36 to 42 hours for warfarin; 8 to 12 hours for acenocoumarol; and 96 to 140 hours for phenprocoumon), as well as the estimated bleeding and thrombotic risk. The following guidelines are most appropriate (algorithm 1) :
Fully elective surgery — In patients with an INR between 2.0 and 3.0 who are undergoing elective surgery that requires temporary cessation of anticoagulation, warfarin should be withheld for approximately five days to allow the INR to be normal (INR <1.3) or near normal (INR 1.3 to 1.4) before surgery [20,52,58].
Semi-urgent surgery — If more rapid reversal of warfarin anticoagulation is required (eg, over one to two days), warfarin should be withheld and a small dose of vitamin K administered, either intravenously (eg, 1.0 to 2.5 mg) or orally (eg, 2.5 to 5.0 mg). (See "Correcting excess anticoagulation after warfarin", section on 'Surgery/invasive procedure'.)
Urgent surgery — If urgent reversal of warfarin anticoagulation is required (eg, less than one day), warfarin should be withheld and a higher dose (eg, 2.5 to 5.0 mg) of intravenous vitamin K administered. If more immediate correction is required (eg, minutes to hours), this can be achieved via the use of those prothrombin complex concentrates (PCCs) that contain adequate amounts of factor VII (eg, 4-factor PCC) or fresh frozen plasma, in addition to vitamin K . Details concerning how such rapid reversal of the warfarin effect can be accomplished are presented separately. (See "Correcting excess anticoagulation after warfarin", section on 'Significant or life-threatening bleeding'.)
Timing of warfarin resumption — Warfarin therapy should be restarted 12 to 24 hours after surgery, typically the evening after surgery, provided that surgical hemostasis has been achieved . If warfarin is resumed alone, without heparin bridging, a full anticoagulant effect will take four to six days to occur, thereby allowing a more gradual re-anticoagulation, which may be appealing in patients undergoing surgery associated with substantial expected blood loss.
Dabigatran — Dabigatran is an oral direct thrombin inhibitor that is currently approved for use as stroke prevention in atrial fibrillation and, in some countries, for DVT prevention after hip and knee replacement surgery. It has a time to peak anticoagulant activity of two to three hours after ingestion and an elimination half-life of 12 to 14 hours in patients with normal renal function and approximately 28 hours in those with severe renal impairment . (See "Anticoagulation with direct thrombin inhibitors and factor Xa inhibitors", section on 'Dabigatran'.)
These pharmacokinetic parameters affect the timing of cessation and resumption of treatment with this agent, as follows [2,61,62]:
●Cessation prior to invasive or surgical procedures — For those with a creatinine clearance ≥50 mL/minute, discontinue dabigatran one to two days before the procedure. For those with a creatinine clearance <50 mL/minute, discontinue this agent three to five days before the procedure. Longer periods than the above should be considered for those undergoing major surgery, spinal puncture, placement of a spinal or epidural catheter or port, in whom complete hemostasis may be required. It has been recommended that a normal or near-normal aPTT or thrombin clotting time be documented to ensure that dabigatran has been adequately cleared from the circulation prior to surgery [34,61,63].
●Resumption following procedure — Dabigatran should be resumed postoperatively when hemostasis has been achieved. Since dabigatran has a rapid onset of action, with peak effects occurring two to three hours after intake, caution should be used when resuming dabigatran, especially in patients who have had major surgery or other procedures associated with a high bleeding risk. In some respects, resuming dabigatran at its typical treatment dose of 150 mg twice daily is similar to resuming LMW heparin bridging and, in both instances, caution should be exercised after surgery. Thus, in patients having high bleeding risk surgery or procedures, it is sensible to delay resumption of dabigatran for two to three days after such procedures and, if needed, to administer a lower dabigatran dose for the initial two to three postoperative days (eg, 110 mg once daily) or use a low-dose LMW heparin for this period.
●Need for bridging anticoagulation — In general, the rapid offset and onset of dabigatran activity obviates the need for bridging anticoagulation with heparin or LMW heparin. In one study of dabigatran-treated patients who required an elective surgery or procedure and who, in most cases, did not receive bridging anticoagulation, the incidence of postoperative bleeding was similar to that in warfarin-treated patients who needed surgery and the incidence of thromboembolic events was low (eg, <1 percent) .
●Resumption following use of heparin or LMW heparin — If patients are receiving heparin or LMW heparin after surgery and there is an intent to resume dabigatran, it should be done ≤2 hours prior to the time of the next scheduled dose of the LMW heparin or at the time of discontinuation of intravenous heparin.
Rivaroxaban and apixaban — Rivaroxaban and apixaban are oral direct factor Xa inhibitors. As with dabigatran, they have a rapid onset of action (peak activity after two to three hours) and a similar elimination half-life of 9 to 12 hours, but, unlike dabigatran, are less dependent on renal clearance (eg, 25 to 33 percent instead of 80 percent). In countries where these drugs are approved for long-term use and patients may require perioperative management, the same approach should be used as with dabigatran-treated patients . (See "Anticoagulation with direct thrombin inhibitors and factor Xa inhibitors", section on 'Factor Xa inhibitors'.)
Risk of rebound hypercoagulability — As mentioned above, rebound hypercoagulability may occur following the abrupt cessation of anticoagulation. Accordingly, alternative preoperative and/or postoperative prophylaxis against thromboembolism with unfractionated heparin (UFH) or LMW heparin should be considered in high risk patients (eg, prosthetic valve in the mitral position, venous thromboembolism within the previous four weeks, or active malignancy) for the period during which the INR is less than 2.0 [64-66].
The clinical effects of rebound hypercoagulability after stopping warfarin are unclear. There is some biochemical evidence for this phenomenon, as well as the recommendation by some investigators that warfarin should be withdrawn gradually [67-69].
●In another report, however, 32 patients were randomly assigned to receive abrupt or gradual withdrawal of warfarin . Very high levels of thrombin activation were seen in a few patients treated with abrupt withdrawal, two of whom developed a thrombotic event (one recurrent deep vein thrombosis and one thrombosis in a varicose vein).
Surgery itself increases the risk of thromboembolism as documented by changes in hemostatic markers, which are also part of the acute phase response and wound healing process . High levels of hemostatic markers, such as fibrin D-dimer, an index of intravascular thrombogenesis and fibrin turnover, are predictive of postoperative thrombosis [71,72]. Although there is evidence that surgery increases the risk of venous thromboembolism, there is no evidence that surgery itself increases the risk of arterial thromboembolism, apart from risks associated with particular procedures, such as carotid surgery .
Overview — Bridging anticoagulation can be defined as the administration of a short-acting anticoagulant, typically a LMW heparin, during the perioperative interruption of warfarin. Bridging can be used as an alternative to warfarin interruption (ie, a period of no anticoagulation). The intent of bridging is to minimize the time patients are not being anticoagulated, thereby minimizing patients’ risk for perioperative thromboembolism.
The therapeutic benefit of bridging anticoagulation is not fully established in every patient population. Because of the lack of evidence-based information indicating whether bridging anticoagulation is or is not warranted, there is considerable variation in the use of this modality [74-76].
Given this uncertainty, the decision about such bridging anticoagulation should be based upon individual patient and surgery-related factors. In general, such bridging anticoagulation may be considered in patients with the following :
●Prior stroke or systemic embolic event
●Mechanical mitral valve
●Mechanical aortic valve and additional stroke risk factors
●Atrial fibrillation and multiple stroke risk factors (eg, CHADS2 ≥4, CHA2DS2-VASc ≥5)
●Recent (within three months) venous thromboembolism
●Active coronary or peripheral vascular disease
●Previous thromboembolism during interruption of warfarin therapy
●Major cardiac or vascular surgery
Of note, the BRUISE CONTROL trial, which randomized patients undergoing implantation of a cardiac implantable electronic device to continuation of warfarin or bridging anticoagulation with heparin, found greater postoperative bleeding in those who received bridging than in those who continued warfarin . A potential explanation is that continuation of warfarin allowed identification of sites of surgical bleeding that could be addressed during the procedure, whereas bridging temporarily reduced bleeding from these sites, causing them to be overlooked. Those who continued on warfarin may thus have had better surgical hemostasis. (See "Implantable cardioverter-defibrillators: Complications", section on 'Bleeding'.)
Of the available non-randomized studies assessing perioperative anticoagulant management during warfarin interruption [35,64,78-86], the following prospective observational cohort studies are instructive concerning the thrombotic and hemorrhagic risks attendant to the use or non-use of bridging anticoagulation in this setting.
Two uncontrolled observational studies assessed patients, most of whom did not receive bridging anticoagulation, who had warfarin interruption. The first study involved a cohort of 1024 patients, most of whom had atrial fibrillation, whose warfarin therapy was temporarily withheld on 1293 different occasions for an outpatient invasive procedure (colonoscopy; oral, dental, or ophthalmic surgery; epidural injection; prostate or breast biopsy; dermatologic procedure). The following observations were made :
●The duration of warfarin therapy interruption was variable, although >80 percent had warfarin therapy withheld for ≤5 days. Bridging therapy with heparin or LMW heparin was given in only 8.3 percent of the procedures.
●Six patients (0.6 percent; 95% CI 0.2-1.3) experienced major bleeding and an additional 17 patients (1.7 percent; 95% CI 1.0-2.6) experienced a clinically significant, nonmajor bleeding episode. Four of the six patients with major bleeding and 10 of the 17 patients with non-major clinically significant bleeding had received periprocedural bridging therapy, for an overall bleeding rate in bridged patients of 13 percent.
●Postprocedural thromboembolism within 30 days occurred in seven patients (0.7 percent; 95% CI 0.3-1.4). None of the seven patients had received periprocedural bridging therapy; two of these patients would have been considered to be at high risk for thromboembolism (ie, recent venous thromboembolism or active malignancy).
The authors concluded that perioperative anticoagulation may be unnecessary for a significant proportion of low to intermediate risk outpatients who have undergone long-term anticoagulation, whose warfarin therapy must be interrupted for a brief period (ie, ≤5 days), and that bridging therapy with unfractionated or LMW heparin may result in significant and potentially avoidable perioperative hemorrhage.
Similar conclusions were reached in a second study involving a cohort of 345 individuals with atrial fibrillation undergoing an invasive procedure. Warfarin therapy was temporarily withheld on 342 different occasions for a mean period of 6.6 days. The decision to use bridging anticoagulation with therapeutic doses of either intravenous unfractionated heparin or LMW heparin was individualized, but was generally given only to those patients deemed to be at high risk for stroke. The following observations were made :
●The three-month cumulative incidence of thromboembolism was 1.1 percent and did not differ significantly between those who did or did not receive bridging therapy.
●Of interest, despite not receiving bridging therapy after brief warfarin cessation, none of the 43 patients with a prior thromboembolic event, and none of the 51 patients with a CHADS2 score ≥3 developed postoperative thromboembolism during the three-month follow-up period.
●The three-month cumulative incidence of major bleeding was 2.7 percent for those receiving bridging therapy. Of the 10 episodes of major bleeding, six occurred in five patients given bridging therapy with LMW heparin.
The authors indicated that they currently utilize bridging with LMW heparin only for those patients at highest thrombotic risk (eg, prior stroke, CHADS2 score ≥4) while taking into account the procedure-associated risk of bleeding. (See 'Type of surgery or procedure' above.)
A meta-analysis that assessed the use of bridging anticoagulation in warfarin-treated patients who had elective surgery found that in patients who received therapeutic-dose LMW heparin bridging there was no significant difference in thromboembolism when compared with patients who did not receive bridging (0.9 versus 0.6 percent, odds ratio = 0.80; 95% CI: 0.42-1.54) . However, the use of bridging anticoagulation was associated with a significant increase in overall bleeding (odds ratio 5.40; 95% CI 3.00-9.74).
These findings from nonrandomized trials (ie, no decrease in thromboembolic events coupled with an increase in bleeding events when bridging anticoagulation is employed) should be interpreted with caution because of the potential for bias among patients who may have preferentially received (or not received) bridging anticoagulation. More definitive conclusions about the therapeutic benefits of bridging will be based on the results of ongoing randomized trials.
The following two nonrandomized studies employed sub-therapeutic doses of LMW heparin as bridging therapy.
In the first study, subtherapeutic doses of LMW heparin (eg, 3800 international units of nadroparin or 4000 international units [40 mg] of enoxaparin once daily the night before the procedure for those at low risk of thrombosis, or 3800 or 4000 international units of these agents twice daily for those at high thrombotic risk) were employed in 103 patients undergoing surgery and in 225 non-major invasive procedures. Results included :
●The overall incidence of a thromboembolic event was 1.8 percent and was not significantly different between those in the low risk (0.54 percent) or high risk (3.4 percent) groups.
●The overall incidence of major bleeding was 2.1 percent and was not significantly different between those in the low risk (0.5 percent) or high risk (4.1 percent) groups. All of the major bleeding events occurred in those undergoing major surgery.
The second study was a report from a prospective registry of 198 consecutive patients receiving phenprocoumon undergoing planned surgery . The majority of patients (88 percent) were judged to be at intermediate thromboembolic risk. Phenprocoumon was stopped seven days before surgery. All patients received enoxaparin in a half-therapeutic dose (1 mg/kg per day) starting when the INR was <2.0 and continuing until the day before surgery. Enoxaparin was restarted after the procedure at the same total daily dose (in two divided doses) and phenprocoumon was resumed within the first 14 postoperative days, depending upon the bleeding risk. Only one patient (0.5 percent) developed a postoperative arterial thromboembolic event, and one patient (0.5 percent) required a second surgical intervention due to severe bleeding.
Since bridging anticoagulation with subtherapeutic doses of LMW heparin was safe and effective in groups with low, intermediate, and high risk groups, this approach appears to be a reasonable compromise between the risks of bleeding and thrombosis posed by either lower or higher doses of bridging anticoagulation.
Anticoagulant drugs and doses used for bridging anticoagulation
Unfractionated heparin — The biologic half-life of intravenous unfractionated heparin (UFH) is approximately 45 minutes [12,89]. Thus, most bridging anticoagulation studies have suggested that intravenous UFH should be stopped four to five hours before the planned surgery or procedure, a time interval that is approximately five elimination half-lives of UFH. If UFH is used for bridging, it is administered by an infusion to attain an activated partial thromboplastin time (aPTT) that is 1.5 to 2 times the control aPTT, in a manner similar to that used to treat patients with acute venous thromboembolism.
LMW heparin — The biologic half-life of subcutaneous LMW heparin is approximately three to five hours [12,89]. Thus, most bridging anticoagulation guidelines have suggested that the last dose of subcutaneous LMW heparin should be given 24 hours before the planned surgery or procedure, a time interval that is approximately five elimination half-lives of LMW heparin. If LMW heparin is used for bridging, it can be given as a therapeutic-dose regimen, either twice-daily (eg, enoxaparin, 1 mg/kg twice daily) or once-daily (eg, dalteparin, 200 international units/kg once daily). LMW heparin can also be given at a lower dose, referred to as an intermediate-dose regimen (eg, enoxaparin, 40 mg twice daily), which is higher than the dose typically administered for postoperative prevention of DVT (eg, enoxaparin, 40 mg once daily). There is no established dosing regimen, although therapeutic-dose regimens tend to be used more often in North America.
However, because some studies have shown residual anticoagulant effect at 24 hours after stopping therapeutic-dose LMW heparin [16,90], the 2012 ACCP Guidelines recommend that on the day before surgery or the procedure, therapeutic-dose LMW heparin should be administered at one-half of the usual total daily dose .
Bridging with LMW heparin is typically commenced three days before a planned surgery or procedure, with the last preoperative dose given on the morning (or at least 24 hours) before surgery. If a twice-daily LMW heparin regimen is given, the evening dose the night before surgery is omitted, whereas if a once-daily regimen is given (eg, dalteparin 200 international units/kg), one-half of the total daily dose is given in the morning of the day before surgery. This dosing schedule ensures that no significant residual anticoagulant will be present at the time of surgery [90,91].
Postoperative resumption of bridging anticoagulation — The onset of anticoagulation after both UFH and LMW heparin is similar, at approximately one hour after administration, with peak anticoagulant activity at approximately three to five hours. Thus, if post-operative or post-procedural anticoagulation is contemplated, these agents should not be employed too early following the operation in order to insure that hemostasis has been secured at the operative or procedural site . This requires both a pre-procedural estimate of the anticipated risk of bleeding as well as a post-procedural determination of the adequacy of hemostasis .
However, for most minor procedures associated with a low bleeding risk, therapy with LMW heparin or UFH can usually be resumed at 24 hours post-procedure, whereas for those undergoing major surgery or those with a high bleeding risk procedure, such treatment should be delayed for 48 to 72 hours after hemostasis has been secured .
Importance of assessing postoperative wound hemostasis — Irrespective of the bridging anticoagulation regimen used, it is important to assess patients for postoperative hemostasis, which consists of a subjective assessment of wound drainage, bleeding into bandages and, where appropriate, hemoglobin levels. Ideally, bridging anticoagulation should be resumed when the wound bed is considered to be dry, that is when there is no ongoing wound bleeding. This determination will vary depending on the surgery type and individual patient considerations, and may be difficult for surgery (eg, cardiac, intracranial) where ongoing bleeding is not readily apparent. Nonetheless, attention to postoperative hemostasis is clinically important since too early resumption of bridging anticoagulation, especially within 24 hours after surgery, is associated with a two- to fourfold increased risk for major bleeding [74,87,93]. If bleeding occurs, this will prolong the interruption of warfarin, which in turn will place patients at higher risk for thromboembolism.
SPECIFIC SCENARIOS — A number of clinical situations require special consideration. These are discussed below.
Venous thromboembolism — The management of anticoagulation in patients with previous venous thromboembolic disease depends upon the temporal relationship to surgery or an invasive procedure (table 7).
●First month — Within the first month after an acute episode of venous thromboembolism (VTE), the incidence of recurrence without anticoagulation is approximately 1 percent per day. While postoperative intravenous heparin doubles the rate of bleeding, there is a net reduction in serious morbidity in such patients, since the risk of postoperative recurrent VTE is high. Thus, heparin therapy is recommended both before and after surgery .
●Two to three months — By two to three months after an acute episode of VTE, the risk of recurrence is significantly reduced so that preoperative heparin therapy is probably not justified unless there are other risk factors for thromboembolism (eg, prolonged hospitalization and confinement to bed) . However, because of an expected increase in the risk of VTE after surgery, these patients should be treated postoperatively with heparin.
●More than three months — At more than three months after an episode of VTE, preoperative anticoagulation is not needed and postoperative intravenous heparin is also probably not necessary. In this setting, the bleeding associated with postoperative intravenous heparin offsets any beneficial effect from the prevention of major thromboembolic events . Prophylactic measures that reduce the thrombotic risk, such as subcutaneous LMW heparin or compression stockings, are associated with a lower risk of bleeding than intravenous heparin and are safer alternatives [4,94]. (See "Treatment of lower extremity deep vein thrombosis", section on 'Length of treatment'.)
Arterial thromboembolism including atrial fibrillation — The approach is somewhat different in patients at risk for arterial thromboembolism, because the thromboembolic risk is similar both before and after surgery. Furthermore, the risk of bleeding is much higher in such patients after surgery if heparin is continued or the INR is maintained at around 2.0. Thus, preoperative treatment with intravenous heparin is advised if anticoagulation with a vitamin K antagonist is stopped. Postoperative therapy with intravenous heparin is probably useful for patients undergoing minor surgery where the risk of bleeding is low. In contrast, the net benefit of using intravenous heparin in reducing long-term disability after major surgery is small because of the high risk of serious bleeding.
If the risk of acute arterial thromboembolism is low, as in patients with nonvalvular atrial fibrillation receiving warfarin for thromboprophylaxis, postoperative intravenous heparin therapy probably increases, rather than decreases, serious morbidity. Restarting warfarin on the second postoperative day and the use of low risk regimens, such as subcutaneous heparin, if warfarin was discontinued preoperatively, are preferable to minimize the risk of bleeding.
●Unfractionated heparin should be stopped four hours before surgery with the expectation that the anticoagulation effect will have worn off at the time of surgery . If LMW heparin has been used, it should be stopped, preferably 24 hours before surgery , with the same expectation.
These recommendations may not apply to patients undergoing minor procedures such as skin biopsy or dental extractions [38,44,47]. However, it is important to confirm that the INR does not exceed the therapeutic range. Although uninterrupted anticoagulation may be continued in most patients, we and many other clinicians withhold warfarin for two to four days prior to the procedure, and reinstitute therapy after the procedure .
●Elective surgery should be avoided in the first month after arterial thromboembolism. If surgery is essential, preoperative and postoperative heparin therapy is recommended as described above for VTE, but only if the risk of postoperative bleeding is low.
●In patients receiving warfarin as prophylaxis against arterial embolization, such as low risk patients with a prosthetic heart valve or nonvalvular atrial fibrillation, the risk of thromboembolism is not high enough to warrant routine preoperative or postoperative therapy with intravenous heparin, especially in view of the bleeding risk [10,96].
If the surgical intervention itself is associated with a high risk of postoperative VTE, the brief use of subcutaneous low-dose heparin or LMW heparin in doses used for prophylaxis against VTE has been suggested . (See "Antithrombotic therapy to prevent embolization in atrial fibrillation", section on 'Temporary cessation of anticoagulation' and "Prevention of venous thromboembolic disease in surgical patients", section on 'Risk factors for VTE'.)
The approach is different in high risk patients with atrial fibrillation (eg, prior thromboembolism, rheumatic heart disease, left ventricular dysfunction) or prosthetic heart valves (eg, those with older generation mechanical valves or atrial fibrillation), in whom there is a delicate balance between the risks of bleeding and thromboembolism [10,96]. In this setting, a number of options exist and the appropriate treatment of such patients is unclear and controversial [95,96].
However, it is our preference in this setting to administer intravenous heparin until five to six hours before the procedure, to be restarted as soon as surgical hemostasis has been assured ; the dose is adjusted to achieve an activated PTT that is 2.0 times control. Warfarin is then reinstituted prior to discharge from the hospital; the INR should be in the therapeutic range for at least 48 hours before heparin is discontinued.
Cardiac procedures — The perioperative management of anticoagulation for cardiac procedures depends on the type of procedure, the indication for anticoagulation, and the specific agent used. Discussions of the perioperative management of anticoagulation and antiplatelet therapy in patients undergoing specific cardiac procedures are presented separately:
●Cardiac implantable electronic device (eg, permanent pacemaker, implantable cardioverter-defibrillator) (see "Implantable cardioverter-defibrillators: Complications", section on 'Bleeding')
●Mechanical prosthetic heart valve (see "Antithrombotic therapy in patients with prosthetic heart valves", section on 'Interruption of warfarin for surgical procedures')
●Percutaneous coronary intervention (see "Antithrombotic therapy for percutaneous coronary intervention: General use", section on 'Patients who require warfarin')
Insertion and removal of a catheter for neuraxial anesthesia — The risk of bleeding from an epidural site exists both at the time of insertion and removal of an indwelling epidural catheter. Evidence-based guidelines from the American Society of Regional Anesthesia suggest the following when low molecular weight (LMW) heparin is used for anticoagulation in the following two scenarios:
●Wait at least 24 hours after the last dose of LMW heparin is administered before a spinal/epidural catheter is placed.
●Wait at least 6 to 8 hours after catheter removal before resuming treatment with LMW heparin, and when there is adequate surgical site hemostasis.
●Before surgery, wait at least 24 hours after the last dose of LMW heparin is administered before a spinal/epidural catheter is placed.
●After surgery, wait at least 24 hours after catheter removal before resuming therapeutic-dose LMW heparin (applies to once daily and twice daily LMW heparin dosing), and when there is adequate surgical site hemostasis.
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, “The Basics” and “Beyond the Basics.” The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on “patient info” and the keyword(s) of interest.)
SUMMARY AND RECOMMENDATIONS — The risk of thromboembolism in patients who discontinue anticoagulation before an invasive procedure must be weighed against the risk of bleeding if these agents are continued or bridging anticoagulation is employed. Accordingly, the thrombotic and bleeding risks of the patient should be carefully assessed before an informed decision on modifying anticoagulation therapy can be made. (See 'Problem overview' above.)
Estimating thrombotic and bleeding risks
●Estimated perioperative thromboembolic risks for patients anticoagulated for various indications (ie, mechanical heart valves, atrial fibrillation, venous thromboembolism) are shown in the table (table 1). (See 'Thrombotic risk if anticoagulation is stopped' above.)
Use or nonuse of bridging anticoagulation
●Elective surgery should be avoided, if at all possible, in the first month after an acute episode of venous thromboembolism as well as in the first month after arterial thromboembolism. Such patients are at extremely high risk of thromboembolism if anticoagulation is discontinued prior to surgery. If surgery is essential, we suggest the use of bridging anticoagulation rather than perioperative discontinuation of anticoagulation (Grade 2C). (See 'Arterial thromboembolism including atrial fibrillation' above and 'Overview' above.)
●If acute venous thromboembolism has occurred within the past two to four weeks prior to the planned surgery/procedure and the risk of bleeding precludes perioperative continuation of anticoagulant therapy, we suggest a three-step approach: 1) interrupt anticoagulation with a vitamin K antagonist and administer therapeutic-dose LMW heparin bridging (eg, enoxaparin 1 mg/kg subcutaneously twice daily) before surgery; 2) temporary use of a vena caval filter that can be removed after surgery; 3) resumption of heparin bridging and anticoagulation after surgery (Grade 2C). (See "Placement of inferior vena cava filters and their complications".)
●For patients at high risk for thrombosis, we suggest the use of bridging anticoagulation with intravenous heparin or LMW heparin over no bridging during temporary interruption of anticoagulation with warfarin or other vitamin K antagonist (Grade 2C). Clinical judgement is required to determine the dose of these agents (ie, therapeutic, sub-therapeutic, or prophylactic dosing) as the quality of evidence for making this choice is poor. We suggest using a therapeutic-dose bridging regimen (eg, enoxaparin 1 mg/kg subcutaneously twice daily) or an intermediate-dose bridging regimen (eg, enoxaparin 40 mg subcutaneously twice daily) in patients with atrial fibrillation, or a mechanical heart valve and a prophylactic-dose bridging regimen (eg, enoxaparin, 40 mg once daily) in patients with an episode of VTE that has occurred more than four weeks prior to the planned surgery/procedure. (See 'Patients at risk for recurrent venous thromboembolism' above and 'Overview' above.)
●For patients at intermediate risk for thrombosis, the decision as to whether bridging anticoagulation should be used is unclear and depends upon the patient’s estimated bleeding risk. Bridging anticoagulation is a reasonable choice if the overall bleeding risk is estimated to be low.
●For patients at low risk for thrombosis, we suggest that bridging anticoagulation not be employed (Grade 2C). (See 'Thrombotic risk if anticoagulation is stopped' above and 'Bleeding risk if anticoagulation is continued' above.)
●For patients undergoing low bleeding risk procedures, we suggest no alteration of their anticoagulation regimen, provided that their INR is within the therapeutic range (Grade 2C). (See 'Type of surgery or procedure' above and 'Bleeding risk if anticoagulation is continued' above.)
●Patients undergoing placement of cardiac electronic devices who are continued on warfarin rather than managed with bridging anticoagulation had a lower risk of postoperative device-pocket hematoma, perhaps because full-dose warfarin exposed potential sites of surgical bleeding that could be addressed during the procedure, leading to better surgical hemostasis. (See 'Overview' above and 'Cardiac procedures' above and "Implantable cardioverter-defibrillators: Complications", section on 'Bleeding'.)
●Guidelines for stopping and restarting direct thrombin inhibitors (eg, dabigatran) and factor Xa inhibitors (eg, rivaroxaban, apixaban) have been formulated on the known pharmacokinetics of these agents as well as the patient’s renal function. These are described in the text. (See 'Dabigatran' above and 'Rivaroxaban and apixaban' above.)
•In general, the rapid offset and onset of anticoagulation activity with these three agents, particularly in patients with normal or mildly-impaired renal function, obviates the need for bridging anticoagulation with heparin or LMW heparin, and studies are ongoing to assess the efficacy and safety of interruption of the direct thrombin inhibitor and factor Xa inhibitor anticoagulants without the need for heparin or LMW heparin bridging. Until such studies are available, bridging may be considered for selected patient groups, which include: 1) pre-operative bridging with heparin in patients with impaired renal function who should stop dabigatran four to five days before surgery; 2) post-operative bridging with low-dose heparin or LMW heparin in patients who have had major surgery and/or cannot take medications by mouth.
Stopping and starting bridging anticoagulation
●For patients with a high risk of procedural bleeding who are taking a vitamin K antagonist, the pre-procedural INR should be ≤1.5. We recommend that patients at low risk for thrombosis stop warfarin five days preoperatively rather than for a shorter interval (Grade 1C). Once surgical hemostasis has been achieved, we suggest that warfarin therapy be resumed 12 to 24 hours post-surgery, rather than for a shorter or longer interval (Grade 2C).
●Bridging anticoagulation with unfractionated heparin should be stopped four to five hours before surgery. If LMW heparin has been used, it should be stopped 24 hours before surgery. (See 'Anticoagulant drugs and doses used for bridging anticoagulation' above.)
●We suggest that heparin or LMW heparin in therapeutic doses should not be restarted postoperatively until at least 24 hours after surgery and delayed longer if there is any evidence of bleeding (Grade 2C). (See 'Anticoagulant drugs and doses used for bridging anticoagulation' above.)
●We suggest resuming therapeutic-dose LMW heparin 48 to 72 hours after surgery instead of resuming LMW heparin within 24 hours after surgery in patients undergoing major surgery or a surgery/procedure associated with a high risk for bleeding (Grade 2C).
All topics are updated as new information becomes available. Our peer review process typically takes one to six weeks depending on the issue.