Perioperative management of patients receiving anticoagulants

INTRODUCTION — The management of anticoagulation in patients undergoing surgical procedures is a common and challenging clinical problem.

●Interrupting anticoagulation for a procedure may transiently increase the risk of thromboembolism, while at the same time, surgery and invasive procedures have associated bleeding risks that are increased by the anticoagulant(s) administered for thromboembolism prevention.

●If the patient bleeds from the procedure, their anticoagulant may need to be discontinued for a longer period, resulting in a longer period of increased thromboembolic risk.

Overall, the aim of periprocedural anticoagulant management is to minimize the risk of both thromboembolism and bleeding. Standardized management protocols have been developed for many patient groups, depending on the anticoagulant they are receiving, but such protocols may be modified to allow for a patient-centric approach that accounts for individual patient characteristics that affect perioperative thromboembolic and bleeding risks.

Other perioperative management issues are anticoagulant specific. For those taking a vitamin K antagonist (VKA), it takes several days until the anticoagulant effect is reduced and then re-established perioperatively; during the period of vitamin K interruption, a short-acting anticoagulant such as low-molecular-weight heparin may be administered in a selected minority of patients to minimize the time patients are not fully anticoagulated during the perioperative period. The direct oral anticoagulants (DOACs), comprising the factor IIa inhibitor dabigatran and the factor Xa inhibitors apixaban, edoxaban and rivaroxaban have a more rapid onset and offset of action than VKA, making them easier to discontinue and resume perioperatively and, thereby, obviate the need for heparin bridging.

Our approach to managing ongoing anticoagulation in patients undergoing surgery or an invasive procedure is based on the type of anticoagulant a patient is receiving. Additional details regarding the use of specific anticoagulants and antiplatelet agents are presented separately.

●Vitamin K antagonists – (See "Warfarin and other VKAs: Dosing and adverse effects".)

●Heparins – (See "Heparin and LMW heparin: Dosing and adverse effects".)

●Direct thrombin inhibitors and direct factor Xa inhibitors – (See "Direct oral anticoagulants (DOACs) and parenteral direct-acting anticoagulants: Dosing and adverse effects".)

●Antiplatelet agents – (See "Perioperative medication management", section on 'Medications affecting hemostasis'.)

Specific recommendations for individuals with prosthetic heart valves are discussed separately. (See "Anticoagulation for prosthetic heart valves: Management of bleeding and invasive procedures".)

Perioperative venous thromboembolism prevention in patients not receiving ongoing anticoagulation is also discussed separately. (See "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients" and "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement" and "Prevention of venous thromboembolism (VTE) in adults undergoing non-major extremity orthopedic surgery".)

OVERVIEW OF OUR APPROACH

General approach — Interruption of anticoagulation temporarily increases thromboembolic risk, and continuing anticoagulation increases the risk of bleeding associated with invasive procedures; both of these outcomes can increase mortality rates [1-6]. Our approach to perioperative management of anticoagulation aims to minimize these risks and determines management according to anticoagulant-specific characteristics such as key pharmacokinetic properties (elimination half-life, time to peak effect) [7].

Our approach is based (whenever possible) on evidence from randomized controlled trials and well-designed observational studies, although evidence from observational studies is also considered secondarily. In addition, thrombotic and bleeding risks may vary depending on individual circumstances, and data from randomized trials or well-designed observational studies are limited to guide practice for certain patient groups. Thus, our approach discussed herein should be used as clinical guidance and should not substitute for clinician judgment in decisions about perioperative anticoagulant management for individual patients.

Our approach to decision-making follows a step-wise approach outlined below; this is presented in an interactive format at Thrombosis Canada; a similar approach is provided on the IPRO-MAPPP website in the United States:

●Step 1: Estimate bleeding risk – A higher bleeding risk confers a greater need for perioperative hemostasis and hence a longer period of anticoagulant interruption. Bleeding risk is dominated by the type and urgency of surgery; some patient comorbidities also contribute. On the other hand, a minimal bleeding risk may allow a procedure to proceed without the need for anticoagulant interruption. However, some minimal bleed risk procedure (eg, dental, skin) may require anticoagulant interruption based on individual patient characteristics. (See 'Estimating procedural bleeding risk' below and 'Deciding whether to interrupt anticoagulation' below.)

●Step 2: Estimate thromboembolic risk – The higher the thromboembolic risk, the greater the importance of minimizing the interval without anticoagulation (table 1). We estimate thromboembolic risk for patients with atrial fibrillation based on age and comorbidities. For those with a recent deep vein thrombosis (DVT) or pulmonary embolism (PE), we estimate the risk based on the interval since diagnosis. Estimating thromboembolic risk is important particularly in patients who are receiving a vitamin K antagonist (VKA) such as warfarin, as this will inform the need for heparin bridging during VKA interruption.

In addition, for patients with a recent (within the previous one to three months) stroke, PE or DVT, in whom the risk of recurrent thromboembolism is markedly increased, we prefer to delay surgery when possible until the risk lessens and is close to baseline. (See 'Estimating thromboembolic risk' below.)

●Step 3: Determine the timing of anticoagulant interruption – The timing of anticoagulant interruption depends on the specific agent the patient is receiving. Warfarin requires earlier discontinuation than the direct oral anticoagulants (DOACs; apixaban, dabigatran, edoxaban, rivaroxaban). Additional considerations may be required in individuals with reduced kidney and/or liver function. (See 'Timing of anticoagulant interruption' below.)

●Step 4: Determine whether to use bridging anticoagulation – For most patients, we do not use bridging anticoagulation (use of a short-acting parenteral agent to reduce the interval without anticoagulation) because it increases bleeding risk without reducing the rate of thromboembolism. However, some patients on warfarin with an especially high thromboembolic risk (eg, mechanical heart valve, recent stroke) may benefit from bridging. (See 'Bridging anticoagulation' below.)

Example cases — The following examples illustrate our decision-making process using this approach in general terms; management of every case must be individualized based on the judgment of the treating clinicians:

●A 76-year-old female with nonvalvular atrial fibrillation, hypertension, and prior stroke three months ago (CHA₂DS₂-VASc score = 5), receiving warfarin, requires elective hip replacement with neuraxial anesthesia; kidney function is normal, and weight is 75 kg. This patient has a high thromboembolic risk (table 1) and a high bleeding risk (table 2).

•Omit warfarin for five days before the procedure (last dose on preoperative day minus 6).

•Preoperative bridging with therapeutic-dose low molecular weight (LMW) heparin (eg, dalteparin, 100 units/kg [7500 units] subcutaneously twice daily) starting on preoperative day minus 3, with last dose on the morning of day minus 1.

•Resume warfarin within 24 hours after surgery (usual dose).

•Postoperative low-dose LMW heparin for venous thromboembolism (VTE) prevention (eg, dalteparin, 5000 units subcutaneously once daily) within 24 hours after surgery until postoperative bridging is started.

•Postoperative bridging on postoperative day 2 or 3, when hemostasis is secured (eg, dalteparin, 100 units/kg [7500 units] subcutaneously twice daily); continue for at least four to five days, until the international normalized ratio (INR) is therapeutic.

●A 70-year-old male with nonvalvular atrial fibrillation, diabetes, and hypertension (CHA₂DS₂-VASc score = 3) receiving dabigatran who requires a colon resection for cancer; kidney function is normal. This patient has a moderate thrombotic risk (table 1) and a high bleeding risk (table 2).

•Omit dabigatran for two days before the procedure (last dose of dabigatran on day minus 3).

•No bridging.

•Resume dabigatran on postoperative day 2 or 3, when patient is able to take medication by mouth.

•Use prophylactic-dose LMW heparin for VTE prophylaxis for the first two to three postoperative days.

●A 55-year-old male with an unprovoked DVT four months ago, receiving apixaban 5 mg twice daily, who requires a colonoscopy because of a history of multiple premalignant colorectal polyps; kidney function is normal. This patient will undergo a colonoscopy and likely repeat polypectomy and is characterized as having a high thrombotic risk (table 1). Although most colonoscopy procedures would be classified as low bleeding risk, in this specific case a high bleeding risk is assumed because of the likely need for polypectomy (table 2).

•Omit apixaban for two days before the procedure (last dose of apixaban on day minus 3).

•No bridging.

•Resume apixaban two days after the polypectomy, after at least 48 hours have elapsed and when hemostasis is secured.

Additional information about endoscopy and polypectomy in patients receiving anticoagulants is presented separately. (See "Management of anticoagulants in patients undergoing endoscopic procedures".)

●A 68-year-old female with nonvalvular atrial fibrillation, hypertension, and congestive heart failure (CHA₂DS₂-VASc score = 4), receiving rivaroxaban 15 mg daily in the morning, requires a dental cleaning and two dental extractions; creatinine clearance (CrCl) is 35 mL/min. This patient has a moderate thrombotic risk (table 1) and a low bleeding risk (table 2).

•Omit rivaroxaban on the day of the procedure.

•Use oral tranexamic acid mouthwash just before the procedure and two to three times that day after the procedure.

•Resume rivaroxaban the day after the procedure, after at least 24 hours have elapsed (assuming the dental extractions were uneventful).

ESTIMATING THROMBOEMBOLIC RISK — The major factors that increase thromboembolic risk are atrial fibrillation, prosthetic heart valves, and recent venous or arterial thromboembolism (eg, within the preceding three months).

Atrial fibrillation — Atrial fibrillation accounts for the highest percentage of patients for whom perioperative anticoagulation questions arise. Patients with atrial fibrillation are a heterogeneous group; risk can be further classified according to clinical variables such as age, hypertension, congestive heart failure, diabetes, prior stroke, and other vascular disease (table 1) [2,8]. The CHA₂DS₂-VASc score (table 3) (calculator 1), which incorporates these variables, is discussed in detail separately, but use of risk scores has not been prospectively validated in the perioperative setting.

The magnitude of this issue was illustrated in three large trials: RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy), ROCKET AF (Rivaroxaban Once daily, oral direct factor Xa inhibition Compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation), and ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) [9-11]. In these trials, a total of 15,000 to 20,000 patients were randomly assigned to warfarin versus one of the direct oral anticoagulants (dabigatran, rivaroxaban, or apixaban, respectively). Surgical or other invasive procedures were required in one-fourth of patients in RE-LY and one-third of patients in ROCKET AF and ARISTOTLE.

●RE-LY (dabigatran versus warfarin) – Of the 4591 patients who underwent elective procedures in RE-LY, the perioperative thromboembolic risk was 1.2 percent, based on a composite endpoint of stroke, cardiovascular death, and pulmonary embolus (PE) [9]. There were no differences in thromboembolic risk with dabigatran versus warfarin, or with the high versus the low dabigatran dose. However, urgent surgery was associated with a higher risk of ischemic stroke or systemic embolism than elective surgery (warfarin: 1.8 versus 0.4 percent; dabigatran 150 mg twice daily: 1.4 versus 0.4 percent; dabigatran 110 mg twice daily: 2.8 versus 0.3 percent).

●ROCKET AF (rivaroxaban versus warfarin) – Of the 4692 anticoagulant interruptions in this trial, 40 percent were for surgery or invasive procedures [11]. The thromboembolic risk during anticoagulant interruption was similar for rivaroxaban and warfarin (0.3 and 0.4 percent).

●ARISTOTLE (apixaban versus warfarin) – During 9260 procedures performed on patients in the ARISTOTLE trial, the perioperative thromboembolic risk was 0.57 percent for warfarin and 0.35 percent for apixaban [10].

Bleeding risk in these trials and registries are presented below. (See 'Overview of whether to interrupt' below.)

Prosthetic heart valve — The risks of thromboembolism and perioperative management of patients with prosthetic heart valves are discussed separately. (See "Anticoagulation for prosthetic heart valves: Management of bleeding and invasive procedures", section on 'Planning for invasive procedures' and "Mechanical prosthetic valve thrombosis or obstruction: Clinical manifestations and diagnosis".)

Recent thromboembolism — Thromboembolic risk is greater in the immediate period following a thromboembolic event and declines over time. Individuals with a recent thromboembolic event (eg, within the previous three months) are likely to benefit from delaying surgery, if possible. There are no high-quality data to determine when risk declines to baseline. If emergency surgery is required (eg, acute cholecystectomy), bridging anticoagulation may be used to reduce the interval without an anticoagulant. (See 'Bridging anticoagulation' below.)

Venous — The perioperative risk of venous thromboembolism (VTE) is greatest in individuals with an event (eg, deep vein thrombosis [DVT], PE) within the prior three months and those with a history of VTE associated with a high-risk inherited thrombophilia (table 1). However, many patients with VTE do not require thrombophilia testing, and we do not perform this testing specifically to evaluate perioperative thrombotic risk in patients who otherwise do not warrant screening. Appropriate use of thrombophilia screening is discussed separately. (See "Clinical presentation and diagnosis of the nonpregnant adult with suspected deep vein thrombosis of the lower extremity".)

Individuals with cancer have a moderate risk, and those with a thromboembolic event more than one year ago have a low risk of VTE complications.

Thus, patients who require surgery within the first three months following an episode of VTE are likely to benefit from delaying elective surgery, even if the delay is only for a few weeks. This approach is supported by data showing that the recurrence risk for individuals with a recent VTE is highest within the initial three to four weeks and diminishes over the following two months [12-14]. Without anticoagulation, the early risk of recurrent VTE was approximately 50 percent; treatment with warfarin for one month reduced this risk to 8 to 10 percent, and after three months of warfarin therapy the risk declined to 4 to 5 percent [14-16].

Arterial — 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 [17]. Thus, patients with a recent arterial embolism are likely to benefit from delaying elective surgery, if such a delay is possible.

The vast majority of cases are due to atrial fibrillation; other less common cardiac sources include paradoxical embolism, nonbacterial thrombotic endocarditis in a patient with malignancy, dilated or poorly contractile left ventricle, or left ventricular aneurysm [18-20].

ESTIMATING PROCEDURAL BLEEDING RISK — The risk of bleeding is dominated by the type of surgery or procedure. Comorbidities (eg, older age, reduced kidney function) and medications that affect hemostasis (eg, aspirin) may also contribute [3,21,22].

As a general guideline, we divide procedures into high and low bleeding risk (two-day risk of major bleeding 2 to 4 percent or 0 to 2 percent, respectively); examples of high bleeding risk procedures include coronary artery bypass surgery, kidney biopsy, and any procedure lasting >45 minutes; low bleeding risk procedures include cholecystectomy, carpal tunnel repair, and abdominal hysterectomy (table 2) [2]. These categories do not substitute for clinical judgment or consultation between the surgeon and other treating clinicians. Neuraxial, intracranial, and cardiac procedures are especially concerning because the location of potential bleeding increases the risk of serious complications. (See "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication".)

Major bleeding is generally defined as bleeding that is fatal, involves a critical anatomic site (eg, intracranial, pericardial), requires surgery to correct, lowers the hemoglobin by ≥2 g/dL, or requires transfusion of ≥2 units packed red cells; however, there is heterogeneity in definitions used by different clinicians [23].

The risks of some specific types of procedures are also discussed in detail separately in the following topic reviews, along with management issues specific to those procedures:

●Neuraxial anesthesia – (See "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication".)

●Gastrointestinal procedures – (See "Management of anticoagulants in patients undergoing endoscopic procedures", section on 'Elective procedures'.)

●Percutaneous coronary intervention (eg, angioplasty, atherectomy, stenting) – (See "Periprocedural management of antithrombotic therapy in patients receiving long-term oral anticoagulation and undergoing percutaneous coronary intervention", section on 'Elective patients' and "Coronary artery disease patients requiring combined anticoagulant and antiplatelet therapy", section on 'Prevention'.)

●Ophthalmologic procedures – (See "Diabetic retinopathy: Prevention and treatment", section on 'Patients taking antiplatelet or anticoagulant medication' and "Cataract in adults", section on 'Management of antithrombotic agents'.)

Dental and cutaneous procedures are generally associated with a low risk of bleeding. (See 'Settings in which continuing the anticoagulant may be preferable' below.)

Patient factors can also contribute to bleeding risk; these patient-related risks can be quantified using bleeding risk scores. An example is the HAS-BLED score (calculator 2), which was used in the BNK Online Bridging Registry (BORDER), an observational registry that assessed perioperative outcomes in outpatients undergoing invasive cardiac procedures (cardiac catheterization, pacemaker implantation, cardiac surgery) [24]. The HAS-BLED score assigns one point each for hypertension, abnormal kidney or liver function (two points for both), stroke, bleeding tendency, labile international normalized ratios (INRs; only applies if on warfarin), elderly age, and drugs (anti-platelet agents, nonsteroidal antiinflammatory drugs [NSAIDs]) or excess alcohol (table 4).

Nearly all of the patients in the BORDER study were receiving a vitamin K antagonist, which was interrupted for the procedure and replaced with a bridging agent, usually a low molecular weight (LMW) heparin. There were 35 clinically relevant bleeding episodes during 1000 procedures (3.5 percent). A HAS-BLED bleeding risk score ≥3 was the most predictive variable for bleeding (hazard ration [HR] 11.8, 95% CI 5.6-24.9) [24].

DECIDING WHETHER TO INTERRUPT ANTICOAGULATION

Overview of whether to interrupt — Once the thromboembolic and bleeding risks have been estimated, a decision can be made about whether the anticoagulant should be interrupted or continued [8,25,26]. Data comparing the relative benefits of continuing anticoagulation versus interrupting an anticoagulant are limited, and decisions that balance thromboembolic and bleeding risks must be made on a case-by-case basis. No scoring system can substitute for clinical judgment in this decision-making.

Our approach is illustrated in the algorithm (algorithm 1) and summarized as follows:

●In general, the anticoagulant must be discontinued if the surgical bleeding risk is high. Those at very high or high thromboembolic risk should limit the period without anticoagulation to the shortest possible interval; in some cases, this involves the use of a bridging agent. (See 'Settings requiring anticoagulant interruption' below.)

●In contrast, individuals undergoing selected minimal bleeding risk procedures often can continue their anticoagulant; in certain cases, continuation of the anticoagulant may be preferable. (See 'Settings in which continuing the anticoagulant may be preferable' below.)

Practices to reduce bleeding and thromboembolic risks should be employed regardless of whether the patient's anticoagulant is interrupted or continued. Examples include the following:

●Agents that interfere with platelet function (nonsteroidal antiinflammatory drugs [NSAIDs], aspirin) should be avoided for routine analgesia unless the benefit outweighs the increased risk of bleeding; routine perioperative use of aspirin should be avoided due to an increased risk of bleeding and lack of benefit. By contrast, if these agents are administered for a separate indication (eg, recent stroke, acute coronary syndromes, implanted coronary stent), they can (and generally should) be continued [27]. Perioperative aspirin use is discussed in detail separately. (See "Perioperative medication management", section on 'Medications affecting hemostasis' and "Periprocedural management of antithrombotic therapy in patients receiving long-term oral anticoagulation and undergoing percutaneous coronary intervention".)

●For those not receiving an anticoagulant in the immediate postoperative period, thromboprophylaxis to reduce the risk of venous thromboembolism (VTE) should be used when appropriate. (See "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients".)

Settings requiring anticoagulant interruption — Individuals undergoing surgery with a high risk of bleeding will require interruption of their usual anticoagulant perioperatively, putting them at higher risk of thromboembolic complications related to their underlying condition.

●If the very high risk of thromboembolism is transient (eg, ischemic stroke within the previous three months), attempts should be made to delay elective surgery, if possible, until the thromboembolic risk has returned to baseline.

It may also be advisable to delay elective surgery in a patient with atrial fibrillation who has had inadequate anticoagulation in the preceding month. This is based on the observation that among patients with nonvalvular atrial fibrillation, over 85 percent of thrombi resolve after four weeks of warfarin therapy [28].

●Individuals with a temporarily very high or high thromboembolic risk in whom surgery cannot be delayed (eg, potentially curative cancer surgery in a patient who just had an acute VTE) should limit the interval without an anticoagulant to minimize the risk of VTE recurrence. This generally involves stopping the usual anticoagulant as close to surgery as possible, restarting it as soon as possible, and, for those on warfarin, using a bridging agent before and/or after surgery while the usual anticoagulant is not therapeutic. A temporary inferior vena cava (IVC) filter may also be appropriate in selected individuals. (See 'Timing of anticoagulant interruption' below and 'Bridging anticoagulation' below and 'Temporary IVC filters' below.)

●For individuals with a chronically elevated thromboembolic risk who are receiving warfarin, we often use bridging anticoagulation to minimize the period when anticoagulation is not being used. (See 'Limited indications for bridging' below.)

●Individuals with a moderate thromboembolic risk generally can interrupt their anticoagulant for surgery without bridging. The bleeding risk from bridging may outweigh any potential benefit, especially in those with low-risk nonvalvular atrial fibrillation [29,30].

Temporary IVC filters — Placement of a temporary inferior vena cava (IVC) filter is indicated in patients with a very recent (within the prior three to four weeks) acute VTE who require interruption of anticoagulation for a surgery or major procedure in which it is anticipated that therapeutic-dose anticoagulation will need to be delayed for more than 24 hours postoperatively. As an example, most patients who require surgery using general or neuraxial anesthesia that must be performed within three to four weeks of an acute VTE would require placement of an IVC filter. Further information about the placement of IVC filters is presented separately. (See "Placement of vena cava filters and their complications".)

In contrast, patients who require temporary interruption of anticoagulation for a minor procedure such as central venous catheter placement, which may be performed with omission of one dose of an anticoagulant, would not require an IVC filter. Individuals with a VTE more than four weeks prior to the intended surgery do not require placement of an IVC filter, and other clinical situations such as prior perioperative VTE or high-risk thrombophilia are not routine indications for perioperative placement of an IVC filter.

Settings in which continuing the anticoagulant may be preferable — For individuals undergoing selected procedures associated with a minimal risk of bleeding (dental extraction, skin biopsy, cataract surgery) it may be preferable for them to continue their anticoagulant, depending on patient factors and the judgment of the treating clinician. Continuing the anticoagulant likely reduces the risk of thromboembolism, and in some settings (eg, cardiac implantable electronic device) it actually reduces the risk of bleeding as well. For those receiving warfarin or another vitamin K antagonist, it is important to confirm that the international normalized ratio (INR) is not above the therapeutic range at the time of the procedure.

●Dental procedures – Dental procedures are generally considered to confer a low risk of bleeding, and anticoagulation can be continued in most patients during these procedures. If the number of extractions is unclear, it is reasonable to delay the morning dose and decide whether to administer the dose after the procedure (administer for extractions of one to three teeth; hold for extractions of more than three teeth) [26]. The evidence for the safety of continuing anticoagulation comes from patients receiving warfarin with an INR in the therapeutic range [31-37]. In the ARISTOTLE trial, which included patients anticoagulated with warfarin versus apixaban for atrial fibrillation, perioperative bleeding rates were approximately 1 percent in patients undergoing dental and other low bleeding risk procedures. Bleeding can be further reduced with the use of topical hemostatic agents (eg, tranexamic acid mouthwash, used three to four times daily for one to two days) [8,36,38-41].

An exception is multiple tooth extractions, which we consider high bleeding risk. (See 'Settings requiring anticoagulant interruption' above.)

●Cutaneous procedures – Cutaneous procedures (eg, skin biopsy, tumor excision) are also generally considered to confer a low risk of bleeding; the potential for local control measures further reduces concerns about bleeding risk.

●Selected cardiac procedures – For certain cardiac procedures, there is evidence that continuing anticoagulation is safe (and in some cases associated with better outcomes) compared with stopping and restarting the anticoagulant.

•Cardiac implantable devices – We agree with a position document from the European Heart Rhythm Association (EHRA) that states the majority of patients undergoing implantation of a cardiac electronic device (pacemaker, cardioverter-defibrillator) should continue their anticoagulant perioperatively [42]. This is based on data from the BRUISE CONTROL trial, which randomly assigned patients on warfarin undergoing implantation of a cardiac implantable electronic device to continuation of warfarin or heparin bridging, as well as other smaller trials [43]. This trial found a lower risk of bleeding in patients who continued warfarin. Potential explanations for the increased bleeding in the heparin-bridging arm include initiation of post-procedure bridging too early (eg, within 24 hours after the procedure) or better identification of surgical bleeding sites that could be addressed during the procedure in patients receiving continued warfarin.

An exception is a patient with a low risk of thromboembolic events, in whom warfarin may be discontinued, or a patient receiving a direct oral anticoagulant (DOAC), for whom temporary discontinuation is likely to be appropriate; bridging anticoagulation is not recommended in such individuals [42]. In an analysis of 611 patients from the RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy) trial (dabigatran versus warfarin in atrial fibrillation) who underwent implantable device surgery, pocket hematomas occurred at a similar frequency in those who had interruption of dabigatran and those who had interruption of warfarin without bridging [44]. It is not clear whether uninterrupted dabigatran would be associated with lower risks of bleeding and/or thrombosis. (See "Cardiac implantable electronic devices: Periprocedural complications".)

•Endovascular procedures and catheter ablation – Endovascular procedures include a variety of venous and arterial interventions, such as angioplasty, catheter ablation, and atherectomy. In a meta-analysis of randomized trials involving over 20,000 patients undergoing these procedures, uninterrupted warfarin therapy was associated with equivalent or lower rates of complications compared with interruption of warfarin [45]. As an example, a benefit of warfarin continuation rather than discontinuation with bridging was reported in the COMPARE trial, which randomly assigned patients with atrial fibrillation undergoing catheter ablation to continued warfarin or discontinuation of warfarin with bridging [46]. In this trial, patients randomized to continue warfarin had a lower risk of stroke and less bleeding.

We also agree with the EHRA position document statement that all patients undergoing catheter ablation for atrial fibrillation should receive full anticoagulation with heparin in addition to continuing their oral anticoagulant [42].

If a decision is made to discontinue the anticoagulant (eg, patient with impaired kidney function), bridging is reserved for those with a high to very high thromboembolic risk and not used for those with a moderate thromboembolic risk. (See 'Limited indications for bridging' below.)

TIMING OF ANTICOAGULANT INTERRUPTION — If a decision has been made to interrupt the anticoagulant for surgery with high or moderate bleeding risk, the agent should be stopped in sufficient time to allow anticoagulation to resolve. For some agents such as warfarin, laboratory testing is a reliable indicator that the anticoagulant effect has resolved after discontinuation; for direct oral anticoagulants (DOACs), well-validated and easily accessible testing is not always available. Data to guide the timing of anticoagulant interruption are evolving, especially for DOACs, and much of our practice is based on expert opinion and observational studies as we await results from ongoing trials [47].

●Validated approaches have been developed to guide the timing of DOAC interruption. (See 'DOAC interruptions (overview)' below.)

●If a low-to-moderate or high bleeding risk surgery is required urgently or immediately, reversal of the anticoagulant may also be required. (See 'Urgent/emergency invasive procedure' below.)

●Risks of bleeding with neuraxial anesthesia and risks of thrombosis in patients with prosthetic heart valves are especially concerning; these issues are discussed in detail separately. (See "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication" and "Antithrombotic therapy for mechanical heart valves".)

Typical durations of anticoagulant interruption are illustrated by the RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy) trial, which randomly assigned individuals with nonvalvular atrial fibrillation to warfarin or dabigatran for prevention of thromboembolism [9]. In this trial, nearly one-half of patients treated with dabigatran had surgery within 48 hours of stopping the drug, whereas only approximately 1 in 10 patients treated with warfarin had surgery within 48 hours of drug discontinuation. The incidence of thromboembolism was low (<1 percent), and bleeding rates were similar for those receiving warfarin or either dabigatran dose. Similar findings were reported in the PAUSE study, which focused on individuals with atrial fibrillation receiving a DOAC. (See 'Atrial fibrillation' above and 'DOAC interruptions (overview)' below.)

Warfarin interruption — Warfarin blocks a vitamin K-dependent step in clotting factor production; it impairs coagulation by interfering with the functions of factors II (prothrombin), VII, IX, and X. Resolution of warfarin effect is determined by measurement of the prothrombin time (PT), which is standardized across institutions using an international normalized ratio (INR).

The BRIDGE trial assessed the need for LMW heparin in patients with atrial fibrillation who were receiving warfarin and required anticoagulant interruption for an elective surgery or invasive procedure [48]. This trial suggested that LMW heparin bridging is not needed in most patients with atrial fibrillation (see 'Evidence for avoiding bridging in most individuals with atrial fibrillation' below); it also provided a protocol of perioperative LMW heparin bridging that was associated with a relatively low risk for bleeding and could be applied to other patient groups in whom the intent would be to administer bridging, such as those with a mechanical mitral valve or prior perioperative stroke.

●Discontinuation – If warfarin discontinuation is appropriate, we typically omit warfarin for five days before an elective surgery (the last dose of warfarin is given on day minus 6); this duration of warfarin interruption should lead to normalization of the INR by the time of surgery and obviates the need for routine INR testing one to two days before the surgery [8,14,49,50]. This approach also appears to be safe, without exposing patients to an increased risk for perioperative bleeding [48,51]. When INR testing is not routinely done, routine use of vitamin K is avoided.

There may be circumstances when preoperative INR testing is warranted, such as in patients who have a <5 day warfarin interruption or a recent high INR (>4.5). In such cases, we check the INR one to two days before the surgery, and, if the INR is >1.5, a low dose of oral vitamin K (eg, 1 to 2 mg) can be given for selected patients and/or procedures in which a normalized INR is required; this can be followed by re-checking an INR the following day. An INR in the normal (<1.3) or near-normal (1.3 to 1.4) range is important in patients undergoing surgery associated with a high bleeding risk (eg, intracranial, spinal, urologic) or if neuraxial anesthesia is to be used. (See 'Estimating procedural bleeding risk' above and 'Neuraxial anesthesia' below.)

This timing of warfarin discontinuation is based on the biologic half-life of warfarin (36 to 42 hours) and the observed time for the PT/INR to return to normal after stopping warfarin (two to three days for the INR to fall to below 2; four to six days to normalize) [49]. Normalization of the INR may take longer in patients receiving higher-intensity anticoagulation (INR 2.5 to 3.5) and in older individuals [52]. Half-lives of other vitamin K antagonists also differ (eg, 8 to 11 hours for acenocoumarol; approximately four to six days for phenprocoumon [53]; approximately three days for fluindione). (See "Warfarin and other VKAs: Dosing and adverse effects", section on 'Warfarin administration'.)

For a procedure that requires more rapid normalization of the INR, additional interventions may be needed to actively reverse the anticoagulant. (See 'Urgent/emergency invasive procedure' below.)

This discontinuation schedule will produce a period of several days with subtherapeutic anticoagulation. As an example, it is 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 eight days (four days before and four days after surgery) [14]. Thus, for patients at very high or high thromboembolic risk, bridging may be appropriate.

●Use of bridging preoperatively – We generally reserve bridging for selected individuals considered at very high or high risk of thromboembolism (eg, recent [within the prior three months] stroke, mechanical mitral heart valve, atrial fibrillation and CHA₂DS₂-VASc score of 7 or 8 (table 3) (calculator 1), CHADS₂ score of 5 or 6) if they require interruption of warfarin. In these cases, the bridging agent (eg, therapeutic-dose subcutaneous low molecular weight [LMW] heparin) is started three days before surgery. (See 'Bridging anticoagulation' below.)

A bridging agent may also be appropriate if there is a prolonged period during which the patient cannot take oral medications (eg, postoperative ileus) and in patients who have had a thromboembolic event during a prior episode of perioperative anticoagulant interruption. This practice is based on expert opinion and has not been formally evaluated in a clinical trial.

●Restarting warfarin and postoperative bridging – We resume warfarin 12 to 24 hours after surgery, typically the evening of the day of surgery or the evening of the day after surgery, assuming there were no unexpected surgical issues that would increase bleeding risk and the patient is taking adequate oral fluids [54]. We use the same dose the patient was receiving preoperatively.

After warfarin is restarted in the perioperative setting, it takes 5 to 10 days to attain a full anticoagulant effect as measured by an INR above 2. Thus, we generally treat individuals at very high risk and some individuals with a high risk of thromboembolism with a heparin bridging agent during this period. (See 'Bridging anticoagulation' below.)

DOAC interruptions (overview) — Individuals who interrupt therapy with a direct oral anticoagulant (DOAC) will have a shorter period without anticoagulation than those who interrupt therapy with warfarin, due to the rapid resolution of anticoagulant effect when a DOAC is discontinued preoperatively and the rapid resumption of effect when a DOAC is restarted postoperatively.

The PAUSE (Perioperative Anticoagulation Use for Surgery Evaluation) study, which prospectively evaluated outcomes in 3007 individuals who were taking a DOAC for atrial fibrillation and underwent an elective surgery or procedure and followed a simple, standardized management approach for interruption of their anticoagulant [55]. There was no preoperative coagulation testing and no heparin bridging.

Rates of thromboembolic and hemorrhagic complications associated with this management were low (major bleeding in <2 percent; ischemic stroke in <0.5 percent), thereby supporting the safety of this approach. In addition, there was approximately 94 percent adherence to the preoperative and postoperative DOAC interruption and resumption protocols, thereby supporting the generalizability of this approach.

The PAUSE approach is illustrated in the figure (figure 1) and summarized as follows:

●Low/moderate bleed risk – For low/moderate bleeding risk surgery, omit the DOAC one day before and resume one day (approximately 24 hours) after the procedure, provided hemostasis is secure. The total duration of interruption is two days.

●High bleed risk – For high bleeding risk surgery, omit the DOAC two days before and resume two days (approximately 48 hours) after the procedure, provided hemostasis is secure. The total duration of interruption is four days. Waiting an additional one day before resumption may be appropriate in some cases.

●Impaired kidney function – For individuals with impaired kidney function (creatinine clearance [CrCl] <30 to 50 mL/min) who are taking dabigatran, there is an additional one-day interruption before low/moderate bleeding risk procedures and an additional two-day interruption before high bleeding risk procedures. Direct factor Xa inhibitors (apixaban, edoxaban, rivaroxaban) do not require adjustments for kidney function.

For individuals with severely impaired kidney function (CrCl 15 to 30 mL/min or on dialysis) who are taking a DOAC, data on perioperative DOAC management are lacking. We suggest following the PAUSE perioperative management protocol in such patients, but we interrupt the DOAC for an additional one to two days before an elective surgery or invasive procedure [55].

Using this approach, residual DOAC levels (not required in routine practice) were in a low range (<50 ng/mL) in most patients overall, and in 99 percent of patients having a high bleeding risk surgery [55]. This approach is outlined in more detail in the following sections.

The population in the PAUSE study was exclusively individuals with atrial fibrillation. The perioperative management from PAUSE can also be applied to patients who are receiving DOAC therapy for venous thromboembolism (VTE) and require treatment interruption for an elective procedure (algorithm 1). For individuals in whom the VTE was >30 days prior, management can be followed in the same manner as individuals with atrial fibrillation. For those who had a recent VTE (within the prior 30 days), perioperative management should be individualized and may include placement of a temporary inferior vena cava (IVC) filter or shorter periods of DOAC interruption. (See 'Temporary IVC filters' above.)

Dabigatran — Dabigatran is a direct thrombin inhibitor; it reversibly blocks the enzymatic function of thrombin in converting fibrinogen to fibrin.

●Discontinuation – Dabigatran can be omitted for one day before a low/moderate bleeding risk surgical procedure and for two days before a high bleeding risk procedure, in individuals with normal or mildly impaired kidney function (CrCl >50 mL/min) (figure 1). For those with impaired kidney function (CrCl 30 to 50 mL/min), dabigatran can be omitted for two days before a low/moderate bleeding risk procedure and four days before a high bleeding risk procedure.

As an example, in a patient on dabigatran with a CrCl >50 mL/min undergoing a high bleeding risk procedure, the patient will skip four doses of dabigatran (no drug on surgical day minus 2 and day minus 1) and no drug on the day of surgery (table 5). In dabigatran-treated patients with a CrCl 30 to 50 mL/min undergoing a high bleeding risk procedure such as neuraxial anesthesia, a longer interval for interruption is required. (See 'Neuraxial anesthesia' below.)

The last preoperative day on which dabigatran is administered can be more closely estimated based on the elimination half-life of dabigatran, which varies according to kidney function (eg, 12 to 14 hours in patients with normal kidney function) [2,56-59]. A protocol incorporating bleeding risk and CrCl was tested in a prospective cohort of 541 dabigatran-treated patients undergoing surgery and was associated with low rates of bleeding and thrombotic complications [59].

Routine coagulation tests have not been validated for ensuring that dabigatran effect has resolved. A normal or near-normal activated partial thromboplastin time (aPTT) may be used in selected patients to evaluate whether dabigatran has been adequately cleared from the circulation prior to surgery (eg, patients at high risk of surgical bleeding) (table 6). Importantly, the reliability of aPTT testing may depend on the specific assay used; if available, a diluted plasma thrombin time may be preferable [57,60,61]. (See "Direct oral anticoagulants (DOACs) and parenteral direct-acting anticoagulants: Dosing and adverse effects", section on 'Dabigatran' and "Clinical use of coagulation tests".)

●Use of bridging – In general, the rapid offset and onset of dabigatran activity makes bridging anticoagulation unnecessary. We reserve bridging anticoagulation for selected individuals who are at very high risk for postoperative thromboembolism and require extended interruption of dabigatran. Examples include postoperative bridging in patients who are unable to take oral medications postoperatively due to intestinal ileus from gastrointestinal surgery. (See 'Bridging anticoagulation' below.)

●Restarting dabigatran – Dabigatran should be resumed postoperatively when hemostasis has been achieved, at the same dose the patient was receiving preoperatively. In general, dabigatran can be restarted one day (approximately 24 hours) after a low/moderate bleeding risk surgery/procedure and two days (approximately 48 hours) after a high bleeding risk surgery/procedure. Since dabigatran has a rapid onset of action, with peak effects occurring two to three hours after intake, caution should be used in patients who have had major surgery or other procedures associated with a high bleeding risk. In cases where the resumption of dabigatran is delayed for two to three days and there is concern about patients being exposed to an increased risk for VTE, we usually administer a low-dose low molecular weight (LMW) heparin regimen (eg, enoxaparin 40 mg daily) until the DOAC is resumed.

If this schedule is used, most patients will omit dabigatran for two total days for a low/moderate bleeding risk procedure and four total days for a high bleeding risk procedure. An additional one to two days of dabigatran interruption is done in patients with moderately impaired kidney function (CrCl 30 to 50 mL/min).

Rivaroxaban — Rivaroxaban is a direct factor Xa inhibitor; it reversibly blocks the enzymatic function of factor Xa in converting prothrombin to thrombin.

●Discontinuation – Rivaroxaban can be omitted for one day in patients who are having a low/moderate bleeding risk surgical procedure and for two days before a high bleeding risk procedure (figure 1). Thus, for low/moderate bleeding risk procedures, the patient will omit one dose of rivaroxaban on the one day before the procedure; for high bleeding risk procedures, the patient will omit two doses of rivaroxaban on the two days before the procedure. These intervals are based on an elimination half-life of 7 to 11 hours and apply to individuals with normal kidney function or mildly impaired kidney function (CrCl >50 mL/min), who are likely to be receiving the 20 mg once daily dose; and to those with moderately reduced kidney function (CrCl 30 to 50 mL/min), who are likely to be receiving the 15 mg once daily dose.

Longer intervals for interruption may be required for situations in which the bleeding risk is very high, such as neuraxial anesthesia. (See 'Neuraxial anesthesia' below.)

Rivaroxaban interacts with dual inhibitors of CYP-3A4 and P-glycoprotein (eg, systemic ketoconazole, ritonavir); dose adjustment or substitution of heparin may be appropriate if these dual CYP-3A4 and P-glycoprotein inhibitors are used perioperatively. Interactions with drugs that inhibit only one of these enzymes do not seem to alter rivaroxaban anticoagulant effect.

Routine coagulation tests have not been validated for ensuring that the rivaroxaban anticoagulant effect has resolved. A normal or near-normal anti-factor Xa activity level may be used in selected patients to evaluate whether rivaroxaban has been adequately cleared from the circulation prior to surgery (eg, patients at high risk of surgical bleeding) (table 6) [2]. The reliability of anti-factor Xa activity testing may depend on the specific assay used, and clinicians are advised to speak with their clinical laboratory to determine whether this assay is available at their institution and whether it has been validated for direct factor Xa inhibitors.

●Use of bridging – In general, the rapid offset and onset of rivaroxaban makes bridging anticoagulation unnecessary. In rare cases bridging may be required, such as the use of postoperative bridging in individuals who have a very high thromboembolic risk and are unable to take oral medications postoperatively due to intestinal ileus from gastrointestinal surgery. (See 'Bridging anticoagulation' below.)

●Restarting rivaroxaban – Rivaroxaban can be resumed postoperatively when hemostasis has been achieved, at the same dose the patient was receiving preoperatively. In general, rivaroxaban can be restarted one day after a low/moderate bleeding risk procedure and two days after a high bleeding risk procedure. Since rivaroxaban has a rapid onset of action, caution should be used in patients who have had major surgery or other procedures associated with a high bleeding risk. We generally restart rivaroxaban one day after low bleeding risk surgery and two to three days after high bleeding risk surgery. In cases where the resumption of rivaroxaban is delayed for two to three days and there is concern about patients being exposed to an increased risk for VTE, we usually administer a low-dose LMW heparin regimen (eg, enoxaparin 40 mg daily) until the DOAC is resumed.

If this schedule is used, most patients will omit rivaroxaban for two total days for a low/moderate bleeding risk procedure and four total days for a high bleeding risk procedure. Adjustments are not made routinely for patients with moderately impaired kidney function (CrCl 30 to 50 mL/min).

Apixaban — Apixaban is a direct factor Xa inhibitor; it reversibly blocks the enzymatic function of factor Xa in converting prothrombin to thrombin.

●Discontinuation – Apixaban can be omitted for one day before a low/moderate bleeding risk procedure and for two days before a high bleeding risk procedure (figure 1). Thus, for low/moderate bleeding risk procedures, the patient will omit two doses of apixaban on the one day before the procedure; for high bleeding risk procedures, the patient will omit four doses of apixaban on the two days before the procedure. These intervals are based on an apixaban elimination half-life of 8 to 12 hours. These intervals apply to individuals with normal kidney function or mildly impaired kidney function (CrCl >50 mL/min), who are likely to be receiving the 5 mg twice daily dose; and to those with moderate to severe kidney insufficiency (CrCl 30 to 50 mL/min), who are likely to be receiving the 2.5 mg twice daily dose.

Longer intervals for interruption may be required for situations in which the bleeding risk is very high, such as neuraxial anesthesia. (See 'Neuraxial anesthesia' below.)

Routine coagulation tests have not been validated for ensuring that apixaban effect has resolved. A normal or near-normal anti-factor Xa activity level may be used in selected patients to evaluate whether apixaban has been adequately cleared from the circulation prior to surgery (eg, patients at high risk of surgical bleeding) (table 6). The reliability of anti-factor Xa activity testing may depend on the specific assay used, and clinicians are advised to speak with their clinical laboratory to determine whether this assay is available at their institution and whether it has been validated for direct factor Xa inhibitors.

●Use of bridging – In general, the rapid offset and onset of apixaban makes bridging anticoagulation unnecessary. In rare cases, bridging may be required, such as the use of postoperative bridging in individuals who have a very high thromboembolic risk and are unable to take oral medications postoperatively due to intestinal ileus from gastrointestinal surgery. (See 'Bridging anticoagulation' below.)

●Restarting apixaban – Apixaban can be resumed postoperatively when hemostasis has been achieved, at the same dose the patient was receiving preoperatively. In general, apixaban can be restarted one day after a low/moderate bleeding risk procedure and two days after a high bleeding risk procedure. Since apixaban has a rapid onset of action, caution should be used in patients who have had major surgery or other procedures associated with a high bleeding risk. We generally restart apixaban one day after low/moderate bleeding risk surgery and two days after a high bleeding risk surgery. In cases where the resumption of apixaban is delayed for two to three days and there is concern about patients being exposed to an increased risk for VTE, we usually administer a low-dose LMW heparin regimen (eg, enoxaparin 40 mg daily) until the DOAC is resumed.

If this schedule is used, most patients will omit apixaban for two total days for a low/moderate bleeding risk procedure and four total days for a high bleeding risk procedure. Adjustments are not made routinely for patients with moderately impaired kidney function (CrCl 30 to 50 mL/min).

Edoxaban — Edoxaban is a direct factor Xa inhibitor; it reversibly blocks the enzymatic function of factor Xa in converting prothrombin to thrombin.

●Discontinuation – Edoxaban can be omitted for one day before a low/moderate bleeding risk procedure and for two days before a high bleeding risk procedure (figure 1). Thus, for low/moderate bleeding risk procedures, the patient will omit one dose of edoxaban on the one day before the procedure; for high bleeding risk procedures, the patient will omit two doses of edoxaban on the two days before the procedure. These intervals are based on an edoxaban elimination half-life of 10 to 14 hours. These intervals apply to individuals with normal kidney function or mildly impaired kidney function (CrCl >50 mL/min) and those with moderately impaired kidney function (CrCl 30 to 50 mL/min), who are likely to be receiving the 60 mg once daily or the 30 mg once daily doses, respectively.

Longer intervals for interruption may be considered for those undergoing major surgery, neuraxial anesthesia or manipulation, or other situations in which complete hemostatic function may be required. (See 'Neuraxial anesthesia' below.)

Routine coagulation tests have not been validated for ensuring that edoxaban effect has resolved. A normal or near-normal anti-factor Xa activity level may be used in selected patients to evaluate whether edoxaban has been adequately cleared from the circulation prior to surgery (eg, patients at high risk of surgical bleeding) (table 6). The reliability of anti-factor Xa activity testing may depend on the specific assay used, and clinicians are advised to speak with their clinical laboratory to determine whether this assay is available at their institution and whether it has been validated for direct factor Xa inhibitors.

●Use of bridging – In general, the rapid offset and onset of edoxaban makes bridging anticoagulation unnecessary. In rare cases, bridging may be required, such as the use of postoperative bridging in individuals who have a very high thromboembolic risk and are unable to take oral medications postoperatively due to intestinal ileus from gastrointestinal surgery. (See 'Bridging anticoagulation' below.)

●Restarting edoxaban – Edoxaban can be resumed postoperatively when hemostasis has been achieved, at the same dose the patient was receiving preoperatively. In general, edoxaban can be restarted one day after a low/moderate bleeding risk procedure and two days after a high bleeding risk procedure. Since edoxaban has a rapid onset of action, caution should be used in patients who have had major surgery or other procedures associated with a high bleeding risk. We generally restart edoxaban one day after low bleeding risk surgery and two to three days after high bleeding risk surgery. In cases where the resumption of edoxaban is delayed for two to three days and there is concern about patients being exposed to an increased risk for VTE, we usually administer a low-dose LMW heparin regimen (eg, enoxaparin 40 mg daily) until the DOAC is resumed.

If this schedule is used, most patients will omit edoxaban for two total days for a low/moderate bleeding risk procedure and four total days for a high bleeding risk procedure.

BRIDGING ANTICOAGULATION — Bridging anticoagulation involves the administration of a short-acting anticoagulant, typically a low molecular weight (LMW) heparin, during the interruption of a longer-acting agent, typically warfarin.

Limited indications for bridging — The intent of bridging is to minimize the time the patient is not anticoagulated, thereby minimizing the risk for perioperative thromboembolism. However, this needs to be balanced with the importance of mitigating the risk of postoperative bleeding. Accumulating evidence suggests that in the vast majority of patients, bridging does not provide a benefit in lowering thromboembolic risk, whereas most data show a consistent increase in bleeding risk. Our approach to the use of bridging is depicted in the algorithm (algorithm 1) and summarized as follows:

●Avoid bridging in individuals who have low thromboembolic risk with anticoagulant interruption [25]:

•Individuals receiving a direct oral anticoagulant (DOAC), unless they have a high thromboembolic risk and a prolonged period during which they cannot take the DOAC postoperatively (eg, due to intestinal ileus)

•Routine prophylactic anticoagulation in atrial fibrillation

•Secondary prophylaxis following venous thromboembolism (VTE) (more than three months prior)

●Bridging may be appropriate during warfarin discontinuation in the following individuals, shown below; however, it may also be considered in other individuals in whom the clinician deems bridging appropriate based on individual patient characteristics and the type of surgery or procedure [1,25]:

•Mechanical mitral valve (exceptions may include those with newer-generation On-X valves or without any additional stroke risk factors) (see "Antithrombotic therapy for mechanical heart valves", section on 'Lower INR target for On-X aortic valve')

•Mechanical aortic valve with major additional stroke risk factors (eg, prior stroke or TIA)

•Embolic stroke within the previous three months or very high stroke risk (eg, CHADS₂ score of 5 or 6)

•VTE within the previous three months (except those with a calf deep vein thrombosis (DVT) and no evidence of DVT on repeat ultrasound, who may not require bridging)

•Possibly in selected individuals with recent coronary stenting (eg, within the previous three months) (see "Periprocedural management of antithrombotic therapy in patients receiving long-term oral anticoagulation and undergoing percutaneous coronary intervention")

•Previous thromboembolism during interruption of chronic anticoagulation (based on presumed increased risk; not addressed in clinical trials)

Individuals with valvular atrial fibrillation (or associated with mitral valvular heart disease) have not been included in large trials, and there may be cases in which the clinician who best knows the patient may have greater concerns about thrombosis and may decide that bridging is appropriate. We feel more strongly about avoiding bridging the lower the patient's baseline thromboembolic risk (eg, lower CHADS₂ or CHA₂DS₂-VASc score (table 3)) and the higher the risk of bleeding.

Evidence for avoiding bridging in most individuals with atrial fibrillation — A 2020 meta-analysis that included six randomized trials and 12 cohort studies found that bridging was associated with an increased risk of bleeding (relative risk [RR] 2.83, 95% CI 2.00-4.01) with no statistical reduction in thromboembolic risk (RR 1.26, 95% CI 0.61-2.58) [62].

Supporting evidence for the practice of avoiding bridging in most individuals with atrial fibrillation includes the following:

●In the BRIDGE trial (Bridging Anticoagulation in Patients who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery), which randomly assigned 1884 patients with atrial fibrillation who required interruption of warfarin for an invasive procedure to receive bridging anticoagulation with the LMW heparin dalteparin versus placebo, the incidence of arterial thromboembolic events 30 days after the procedure was similar in those who received dalteparin or placebo (0.3 versus 0.4 percent) [48]. The incidence of major bleeding (a secondary outcome) was higher in those who received dalteparin (3.2 versus 1.3 percent), although none of the bleeds were fatal. Patients were excluded from the trial if they had a mechanical heart valve or a recent (within previous 12 weeks) stroke, embolism, or transient ischemic attack.

●The PERIOP-2 trial assessed 1471 patients with atrial fibrillation or a mechanical heart valve who required warfarin interruption for an elective surgery/procedure [51]. A mechanical heart valve was present in 21 percent (mitral in 9 percent; aortic in 12 percent). Before the surgery/procedure, all patients received bridging with the LMW heparin dalteparin, 200 IU/kg daily (100 IU/kg daily on the day before the surgery/procedure). After the surgery/procedure, patients were randomly assigned to receive dalteparin 200 IU/kg daily (fixed-dose 5,000 IU daily in patients at high-bleed-risk) or placebo until the INR was ≥2.0 and were followed for 12 weeks. There was no significant difference in the risk of major thromboembolism (for atrial fibrillation, 1.41 versus 0.75 percent, p = 0.27; for mechanical valves, 0 versus 0.67 percent, p = 0.49) or for major bleeding (for atrial fibrillation, 2.62 versus 1.64 percent, p = 0.25; for mechanical valves, 1.96 versus 0.67 percent, p = 0.62).

●Bridging versus no bridging did not affect major outcomes in patients who required a major procedure during participation in large anticoagulation trials for atrial fibrillation, including the RE-LY (warfarin versus dabigatran), ROCKET-AF (warfarin versus rivaroxaban), and ARISTOTLE (warfarin versus apixaban) trials [9-11]. In the RE-LY trial, patients receiving warfarin had more thromboembolic events associated with bridging than with non-use of bridging (1.8 versus 0.3 percent); patients who received bridging also had a higher risk of major bleeding (warfarin: 6.8 percent with bridging, 1.6 percent without; dabigatran: 6.5 percent with bridging, 1.8 percent without) [63].

●Additional real-world data come from the ORBIT-AF and Dresden registries.

•ORBIT-AF (Outcomes Registry for Better Informed Treatment of Atrial Fibrillation) is a community-based registry of outpatients with atrial fibrillation receiving any oral anticoagulant; in this study, 2200 of 7372 individuals (30 percent) had interruption of anticoagulation for a procedure [64]. Bridging was used in 24 percent of these interruptions, especially in patients with a history of stroke or a mechanical heart valve and/or receiving warfarin; bleeding events were more common in individuals who received bridging compared with those who did not receive bridging (5 versus 1.3 percent). A composite endpoint that included major bleeding, myocardial infarction, stroke, systemic embolism, hospitalization, or death within 30 days was also higher in those who received bridging (13 versus 6.3 percent).

•In the Dresden NOAC registry, over 800 patients who were receiving dabigatran, rivaroxaban, or apixaban for any indication and underwent an invasive procedure had similar rates of major cardiovascular events if they received bridging, no bridging, or no anticoagulant discontinuation [65]. Bridging was not an independent risk factor for major bleeding; however, individuals undergoing major procedures were more likely to receive bridging and to have major bleeding.

Evidence for avoiding bridging in individuals with VTE and other populations — Supporting evidence for the practice of avoiding bridging in other groups includes the following:

●VTE – In a systematic review from 2019 that evaluated patients who were receiving a vitamin K antagonist to treat venous thromboembolism (VTE; 28 cohort studies that included nearly 7000 procedures), the pooled incidence of recurrent VTE was similar with and without bridging [66]. The risk of bleeding was generally higher when bridging was used.

●Any indication for anticoagulation – In a large systematic review and meta-analysis involving 34 studies (33 observational and one randomized) in patients receiving a vitamin K antagonist for any indication who were undergoing elective surgery or procedures. there was no significant difference in the rate of thromboembolism in patients who received bridging compared with patients who did not (odds ratio [OR] 0.80, 95% CI 0.42-1.54) [67]. Bridging was associated with a threefold increase in major bleeding compared with no bridging (OR 3.60, 95% CI 1.52-8.50); full-dose heparin was associated with an increase in overall bleeding compared with lower heparin doses (OR 2.28, 95% CI 1.27-4.08). (See 'Heparin product and dose' below.)

A potential role for bridging in reducing the risk of "rebound hypercoagulability" has also been proposed; however, this premise is not supported by data from the BRIDGE trial discussed above [48]. (See 'Evidence for avoiding bridging in most individuals with atrial fibrillation' above.)

Heparin product and dose — Typically, LMW heparins are used for bridging, as they have similar efficacy compared with unfractionated heparin, are more convenient to use, and generally do not require monitoring. Intravenous unfractionated heparin is less costly and can be reversed more rapidly than subcutaneous LMW heparin; it may be a reasonable alternative in some individuals.

●We prefer LMW heparin for bridging anticoagulation in individuals with a very high risk of arterial thromboembolism (eg, rheumatic heart disease, atrial fibrillation with recent embolic stroke, high-risk mechanical heart valve) and those with a moderate risk of thromboembolism (eg, active cancer) [29,30,68].

Perioperative anticoagulation management in individuals with prosthetic heart valves is discussed in detail separately. (See "Anticoagulation for prosthetic heart valves: Management of bleeding and invasive procedures", section on 'Planning for invasive procedures'.)

●For individuals with impaired kidney function and/or those requiring hemodialysis, intravenous or subcutaneous unfractionated heparin can be used more easily because dosing is unaffected by kidney function [69]. (See "Heparin and LMW heparin: Dosing and adverse effects", section on 'LMW heparin standard dosing'.)

There are no data on using the direct oral anticoagulants (DOACs) as bridging agents. We do not use any of the DOACs for bridging.

Heparins can be dosed at prophylactic doses, therapeutic doses, or doses intermediate between the two. The term "therapeutic dose" refers to doses typically used for treatment of thromboembolic disease, despite the fact that in this case it is being used prophylactically (to prevent thromboembolism). There are no clinical trial data or practice standards to guide dosing, and clinical judgment is required to determine the appropriate dose for each patient [67,70,71].

●Therapeutic dosing – Therapeutic dosing (also called "full dose") is appropriate for bridging anticoagulation for individuals with a potential arterial thromboembolic source (eg, atrial fibrillation, mechanical heart valve) or VTE within the preceding month. Typical regimens include enoxaparin, 1 mg/kg subcutaneously twice daily or dalteparin, 100 units/kg subcutaneously twice daily.

●Intermediate dosing – Intermediate-dose anticoagulation may be appropriate for individuals with atrial fibrillation or VTE within the preceding month when bridging is needed but concerns about bleeding are greater. Typical regimens include enoxaparin, 40 mg twice daily, or dalteparin, 5000 units subcutaneously twice daily.

●Prophylactic dosing – Prophylactic-dose anticoagulation (also called "low dose") generally is not used for bridging in patients with atrial fibrillation because there is no evidence that prophylactic-dose heparin prevents stroke in this setting. This dose level may be reasonable in patients who have had a VTE event within the preceding 3 to 12 months. Typical prophylactic regimens include enoxaparin, 40 mg once daily, or dalteparin, 5000 units subcutaneously once daily.

The use of prophylactic-dose heparin for postoperative VTE prevention in patients not receiving ongoing anticoagulation is discussed separately. (See "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients".)

Additional details regarding heparin products, including dose adjustments for obesity and impaired kidney function, are provided separately. (See "Heparin and LMW heparin: Dosing and adverse effects".)

Timing of bridging — Once a decision to use bridging has been made, the next decision is whether to use bridging before the procedure, after the procedure, or both (table 7). Sample bridging protocols are provided on the Thrombosis Canada website.

●Atrial fibrillation – As noted above, we suggest not using bridging for most patients with atrial fibrillation. (See 'Limited indications for bridging' above.)

However, for individuals for whom bridging is used due to a very high risk of thromboembolism, we use bridging both preoperatively and postoperatively [14]. Warfarin is usually resumed 12 to 24 hours after surgery, typically the evening of the day of surgery or the evening of the day after surgery, as long as adequate hemostasis has been achieved. (See 'Warfarin interruption' above.)

●Venous thromboembolism

•First three months – For individuals within the first three months after an acute episode of venous thromboembolism (VTE), we use bridging both preoperatively and postoperatively, typically with therapeutic-dose LMW heparin (eg, enoxaparin 1 mg/kg twice daily) [14]. This practice is based on the high incidence of recurrence without anticoagulation. While postoperative intravenous heparin doubles the rate of bleeding, there is a net reduction in serious morbidity in such patients because the risk of postoperative recurrent VTE is high. (See 'Preoperative timing of bridging' below and 'Postoperative timing of bridging' below.)

In selected patients in whom surgery cannot be delayed beyond the first month after the diagnosis of an acute VTE, it may be appropriate to use a temporary inferior vena cava (IVC) filter, especially if bridging anticoagulation cannot be used postoperatively due to high bleeding risk. (See "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)", section on 'Patients at high risk of bleeding' and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults", section on 'Inferior vena cava filters'.)

•Greater than three months – For individuals greater than three months after an acute episode of VTE, we generally use postoperative bridging, typically with a low-dose LMW heparin regimen (eg, enoxaparin 40 mg daily), but not preoperative bridging [72]. For patients who are undergoing a minor procedure or day surgery, bridging is probably not justified. This practice is based on the significantly reduced risk of VTE recurrence after the first month [73,74]. (See 'Postoperative timing of bridging' below.)

Preoperative timing of bridging — We generally initiate heparin bridging three days before a planned procedure (ie, two days after stopping warfarin), when the prothrombin time/international normalized ratio (PT/INR) has started to drop below the therapeutic range.

●LMW heparin – We discontinue low molecular weight (LMW) heparin 24 hours before the planned surgery or procedure, based on a biologic half-life of most subcutaneous LMW heparins of approximately three to five hours [8,68,75]. 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 on the morning of the day before surgery. This ensures that no significant residual anticoagulant will be present at the time of surgery, based on studies that have shown a residual anticoagulant effect at 24 hours after stopping therapeutic-dose LMW heparin, and it is consistent with the 2022 American College of Chest Physicians (ACCP) guidelines [8,12,76,77].

●Unfractionated heparin – For therapeutic-dose unfractionated heparin, we continue the intravenous infusion until four to five hours before the procedure, based on the biologic half-life of intravenous unfractionated heparin of approximately 45 minutes [8,75,76]. If subcutaneous unfractionated heparin is used, typically with a dose of approximately 250 international units/kg twice daily, the last dose can be given the evening before the procedure.

Postoperative timing of bridging — Postoperative resumption of unfractionated heparin and LMW heparin is similar, based on the onset of anticoagulation at approximately one hour after administration for both forms of heparin, and peak anticoagulant activity at approximately three to five hours.

●The resumption of bridging, especially when given as a therapeutic-dose regimen, should be delayed until there is adequate hemostasis based on a clinical assessment of the wound site, drainage fluid amount, and expected postoperative bleeding; coupled, where appropriate, with hemoglobin levels [78]. This assessment will vary depending on the surgery type and individual patient considerations, and it may be difficult for surgery where ongoing bleeding is not readily apparent (eg, cardiac, intracranial).

●A slight delay in resumption of postoperative anticoagulation is preferable to premature initiation of postoperative bridging that results in bleeding, which ultimately will lengthen the period without an anticoagulant and thus increase thromboembolic risk.

●For those undergoing major surgery or those with a high bleeding risk procedure, therapeutic-dose unfractionated heparin or LMW heparin should be delayed for 48 to 72 hours after hemostasis has been secured [8].

●For most minor procedures associated with a low bleeding risk in which bridging is used (eg, laparoscopic hernia repair), therapeutic-dose unfractionated heparin or LMW heparin can usually be resumed 24 hours after the procedure.

Resumption of bridging anticoagulation too early, especially the use of therapeutic-dose heparin within 24 hours after surgery, is associated with a two- to fourfold increased risk for major bleeding compared with no bridging or prophylactic-dose heparin. The increased bleeding risk was demonstrated in the Prospective peri-operative enoxaparin cohort trial (PROSPECT), which evaluated bleeding risk in a cohort of 260 patients undergoing major surgery whose treating clinicians used bridging anticoagulation [79]. In this trial, nine patients had major postoperative bleeding (3.5 percent), most on postoperative day 0, and 19 (7.3 percent) had minor bleeding.

Postoperatively, warfarin is generally resumed on the same postoperative day as the heparin. Heparin can be discontinued when the INR reaches the therapeutic range for individuals at moderate thromboembolism risk.

Individuals with heparin-induced thrombocytopenia — Heparin-induced thrombocytopenia (HIT) is a potentially life-threatening condition in which heparin-induced antibodies to platelets can cause thrombocytopenia and/or venous or arterial thrombosis. (See "Clinical presentation and diagnosis of heparin-induced thrombocytopenia".)

Patients with HIT should not receive any heparin (eg, they should not receive heparin flushes, unfractionated heparin, or LMW heparin). Non-heparin anticoagulants that can be used in patients with HIT are discussed separately. (See "Management of heparin-induced thrombocytopenia", section on 'Anticoagulation'.)

SPECIAL SITUATIONS

Urgent/emergency invasive procedure — Reversal of the patient's usual anticoagulant may be required for more urgent or emergency surgery or procedures or to treat perioperative bleeding. Agents with a potential prothrombotic effect (eg, prothrombin complex concentrates [PCCs], plasma products, immediate reversal agents) should be reserved for the treatment of severe, life-threatening bleeding or anticipated severe bleeding (eg, intracranial hemorrhage, emergency major surgery with elevated prothrombin time/international normalized ratio [PT/INR]). Agent-specific strategies include the following:

●Warfarin – For individuals who require reversal of warfarin or other vitamin K antagonists, the appropriate reversal strategy is determined by the degree of anticoagulation (eg, prothrombin time/international normalized ratio [PT/INR], clinical bleeding), urgency of the procedure, and degree of bleeding risk (table 8).

•If semi-urgent reversal of warfarin is required (eg, within one to two days), warfarin should be withheld and vitamin K administered (eg, 2.5 to 5 mg of oral or intravenous vitamin K). (See "Management of warfarin-associated bleeding or supratherapeutic INR", section on 'Urgent surgery/procedure'.)

•If immediate reversal is required (eg, for emergency surgery or active bleeding), this can be achieved via the use of PCCs or plasma products (eg, Fresh Frozen Plasma [FFP], Plasma Frozen Within 24 Hours After Phlebotomy [PF24]) along with vitamin K (table 9) [80,81]. The 4-factor PCCs contain adequate amounts of all vitamin K-dependent clotting factors, whereas 3-factor PCCs may require supplementation with FFP for adequate factor VII (table 10). Of note, there is a thrombotic risk associated with these products, and they should be used only if there is life-threatening bleeding and prolongation of the INR by a vitamin K antagonist [81]. (See "Management of warfarin-associated bleeding or supratherapeutic INR", section on 'Serious/life-threatening bleeding'.)

●Dabigatran – Dabigatran is an oral direct thrombin inhibitor; it can be reversed by idarucizumab (table 11). In an open-label study involving 503 patients who had bleeding or required emergency surgery, idarucizumab effectively reversed the anticoagulant effect of dabigatran [82]. The use of this agent as well as other potential strategies for individuals receiving dabigatran who are at great risk of serious bleeding with an emergency procedure are presented separately. (See "Management of bleeding in patients receiving direct oral anticoagulants".)

●Apixaban, edoxaban, and rivaroxaban – Apixaban, edoxaban, and rivaroxaban are oral direct factor Xa inhibitors; in patients receiving these drugs, nonspecific prohemostatic agents (PCC) or DOAC-specific reversal agents (andexanet alpha) can be given with the aim of minimizing perioperative bleeding (table 11). These and other potential strategies for individuals receiving these agents who are at great risk of serious bleeding with an urgent/emergency procedure are presented separately. (See "Management of bleeding in patients receiving direct oral anticoagulants".)

Algorithms for anticoagulant reversal depending on the severity of bleeding are provided by various societies and groups such as Thrombosis Canada.

Additional discussions of postoperative bleeding are presented separately. (See "Postoperative complications among patients undergoing cardiac surgery", section on 'Hematologic dysfunction'.)

Neuraxial anesthesia — Neuraxial (ie, spinal or epidural) anesthesia should not be used in anticoagulated individuals, due to the risk of potentially catastrophic bleeding into the epidural space. The increased risk of bleeding applies both at the time of catheter placement and the time of removal.

If neuraxial anesthesia is considered for surgical anesthesia or postoperative pain control, the timing of anesthesia and anticoagulant administration should be coordinated to optimize the safe use of both. Early consultation with the anesthesiologist is advised. This subject is discussed in detail separately. (See "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication".)

The timing of anticoagulant use in patients receiving neuraxial anesthesia is illustrated by evidence-based guidelines from the American Society of Regional Anesthesia (ASRA), which suggest the following [83,84]:

●Prophylactic-dose low molecular weight (LMW) heparin (eg, enoxaparin, 40 mg once daily):

•Before surgery, wait at least 10 to 12 hours after the last dose of LMW heparin is administered before a spinal/epidural catheter is placed.

•After surgery, when there is adequate surgical site hemostasis, wait at least six to eight hours after catheter removal before resuming treatment with LMW heparins.

●Therapeutic-dose LMW heparin (eg, enoxaparin, 1 mg/kg twice daily):

•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, when there is adequate surgical site hemostasis, for twice daily dosing, wait at least 24 hours after catheter removal before resuming therapeutic-dose LMW heparin. For once daily dosing, wait at least six to eight hours after catheter removal before the first dose; the second postoperative dose should occur no sooner than 24 hours after the first dose.

Patients on VTE prophylaxis — In patients who are receiving prophylactic-dose low molecular weight (LMW) heparin for venous thromboembolism (VTE) prophylaxis and who require a planned procedure or surgery, the last dose is typically given the day before or evening before the planned intervention.

For selected patients in whom the timing of a surgery is not defined, such as a patient with a hip or wrist fracture awaiting surgery, use of unfractionated heparin, 5000 international units twice daily can be given instead of LMW heparin, as the short half-life of unfractionated heparin can allow surgery to proceed when needed, even after recent (within the previous two to three hours) administration of the last unfractionated heparin dose.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Anticoagulation".)

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 5^th to 6^th 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 10^th to 12^th 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.)

●Basics topics (see "Patient education: Taking oral medicines for blood clots (The Basics)" and "Patient education: Choosing an oral medicine for blood clots (The Basics)")

●Beyond the Basics topic (see "Patient education: Warfarin (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Balancing thromboembolic and bleeding risk – Interruption of anticoagulation temporarily increases thromboembolic risk, and continuing anticoagulation increases the risk of bleeding. We take into account these risks, along with specific features of the anticoagulant the patient is taking. Internet-based resources and case examples are provided above. (See 'Overview of our approach' above.)

•Thromboembolic risk – Those at very high or high thromboembolic risk should limit the period without anticoagulation to the shortest possible interval (table 1). If thromboembolic risk is transiently increased (eg, recent stroke, recent pulmonary embolism), we prefer to delay surgery until the risk returns to baseline, if possible. (See 'Estimating thromboembolic risk' above.)

-Atrial fibrillation – We estimate thromboembolic risk for patients with atrial fibrillation based on clinical variables including age and comorbidities. In the RE-LY (Randomized Evaluation of Long-Term Anticoagulant Therapy) trial, the perioperative thromboembolic risk was 1.2 percent based on a composite endpoint of stroke, cardiovascular death, and pulmonary embolus. 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. (See 'Atrial fibrillation' above and "Atrial fibrillation in adults: Use of oral anticoagulants".)

-Prosthetic heart valve – The risks of thromboembolism and perioperative management of patients with bioprosthetic and mechanical heart valves are discussed separately. (See "Anticoagulation for prosthetic heart valves: Management of bleeding and invasive procedures", section on 'Planning for invasive procedures' and "Mechanical prosthetic valve thrombosis or obstruction: Clinical manifestations and diagnosis".)

-Recent thromboembolism – The perioperative risk of venous thromboembolism (VTE) is greatest in individuals with an event within the prior three months, and those with a history of VTE associated with a high-risk inherited thrombophilia. Patients who require surgery within the first three months following an episode of VTE are likely to benefit from delaying elective surgery, even if the delay is only for a few weeks. (See 'Recent thromboembolism' above.)

•Bleeding risk – A higher bleeding risk confers a greater need for perioperative hemostasis, and hence a longer period of anticoagulant interruption. Bleeding risk is dominated by the type and urgency of surgery (table 2); some patient comorbidities (eg, older age, decreased kidney function) and medications that affect hemostasis also contribute. (See 'Estimating procedural bleeding risk' above and 'Deciding whether to interrupt anticoagulation' above.)

-High risk – High bleeding risk procedures include coronary artery bypass surgery, kidney biopsy, and any procedure lasting >45 minutes. In general, the anticoagulant must be discontinued if the surgical bleeding risk is high. (See 'Settings requiring anticoagulant interruption' above.)

-Low risk – Low bleeding risk procedures include dental extractions, minor skin surgery, cholecystectomy, carpal tunnel repair, and abdominal hysterectomy. Individuals undergoing selected low bleeding risk surgery often can continue their anticoagulant. (See 'Settings in which continuing the anticoagulant may be preferable' above.)

-Cardiac implantable device or catheter ablation for atrial fibrillation – Continuing warfarin was associated with a lower risk of bleeding in patients on the BRUISE CONTROL trial who were undergoing implantation of a cardiac implantable electronic device (eg, pacemaker, implantable cardioverter-defibrillator) and patients on the COMPARE trial who were undergoing catheter ablation for atrial fibrillation. (See 'Overview of whether to interrupt' above and "Cardiac implantable electronic devices: Periprocedural complications".)

●Timing of interruption – The timing of interruption depends on the periprocedural bleeding risk and the specific anticoagulant. The algorithm summarizes our approach (algorithm 1). The figure illustrates the timing of direct oral anticoagulant (DOAC) interruption (figure 1). (See 'Timing of anticoagulant interruption' above.)

●Bridging – Bridging anticoagulation involves the administration of a short-acting anticoagulant, typically a low molecular weight (LMW) heparin, during the interruption of a longer-acting agent, typically warfarin. The intent is to minimize the risk of perioperative thromboembolism. For most individuals, bridging increases bleeding risk without lowering thromboembolic risk, and we suggest not using bridging (Grade 2B). We feel more strongly about avoiding bridging the lower the baseline thromboembolic risk and the higher the bleeding risk. We generally do not use bridging for DOACs. (See 'Limited indications for bridging' above.)

For selected patients on warfarin (eg, stroke, systemic embolism, or transient ischemic attack within the previous three months; atrial fibrillation and very high risk of stroke [eg, CHADS₂ score of 5 or 6]; VTE within the previous three months; recent coronary stenting; previous thromboembolism during interruption of chronic anticoagulation), we suggest the use of bridging (Grade 2C).

•Agent – When bridging is used, we prefer LMW heparin for most patients. An exception is an individual with impaired kidney function and/or hemodialysis requirement, for whom intravenous or subcutaneous unfractionated heparin can be used more easily. We do not use DOACs for bridging. Non-heparin anticoagulants that can be used in patients with heparin-induced thrombocytopenia are discussed separately. (See 'Heparin product and dose' above.)

•Timing – Bridging can be used preoperatively, postoperatively, or both, depending on the underlying condition for which the patient is being anticoagulated (table 7). The timing depends on the heparin product used and the procedural bleeding risk. Importantly, resumption of bridging anticoagulation too early is associated with an increased risk for major bleeding. (See 'Timing of bridging' above.)

The role of bridging in individuals with mechanical heart valves is discussed separately. (See "Anticoagulation for prosthetic heart valves: Management of bleeding and invasive procedures".)

●Urgent/emergency procedure – Reversal of the patient's usual anticoagulant may be required for more urgent or emergency procedures or to treat perioperative bleeding. Agents with a potential prothrombotic effect (eg, immediate reversal agents, prothrombin complex concentrates, plasma products) should be reserved for the treatment of life-threatening, severe bleeding or anticipated severe bleeding (intracranial hemorrhage, emergency major surgery). (See 'Urgent/emergency invasive procedure' above and "Management of bleeding in patients receiving direct oral anticoagulants".)