Library › Pharmacology › Anticoagulant bleeding and reversal

Anticoagulant bleeding, reversal, and periprocedural management

Management of bleeding in patients taking anticoagulants, the reversal agents and their doses, interruption and bridging around procedures, and the timing of resumption after bleeding.

Classifying bleeding severity and site in anticoagulated patients

The management of bleeding in a patient taking an anticoagulant starts with a severity grade, because the grade decides whether the drug is held, whether a reversal agent is given, and how soon anticoagulation returns. The 2020 American College of Cardiology (ACC) expert consensus decision pathway (ECDP) on bleeding in patients on oral anticoagulants defines a major bleed by at least one of three features. These are bleeding at a critical site, hemodynamic instability, and clinically overt bleeding with a hemoglobin fall of 2 g/dL or more or transfusion of 2 or more units of red blood cells. The definition is adapted from the International Society on Thrombosis and Haemostasis (ISTH) definition of major bleeding used in anticoagulant trials, which adds fatal bleeding. Bleeding that meets none of these criteria is nonmajor. A nonmajor bleed that requires hospitalization, a procedure, or transfusion is managed with temporary interruption of the anticoagulant and source control. Nonmajor bleeding without these features is usually managed with local measures while anticoagulation continues.

Adapted from the 2020 ACC ECDP on bleeding in patients on oral anticoagulants (Figure 2) and the ISTH major bleeding definition.

Critical sites. The ECDP lists the sites at which a small volume of blood can cause death or permanent disability. Each has its own presenting features and its own specialist route to hemostasis.

Critical-site bleeds from the 2020 ACC ECDP (Table 1). Gastrointestinal bleeding is not a critical site; it becomes major through hemoglobin fall, transfusion, or instability.

Gastrointestinal bleeding is the most common major bleed on DOACs and warfarin. Intracranial hemorrhage is the most lethal. About one fifth of spontaneous intracerebral hemorrhages occur in patients taking anticoagulants, and anticoagulant-associated hematomas are larger and expand for longer. Both observations shape management: speed matters most in the brain, and the decision to restart matters most after gastrointestinal bleeding, where resumption is usually appropriate.

Initial assessment: the drug, the last dose, kidney function, and the bleeding source

Four facts determine whether a reversal agent can help: which anticoagulant the patient takes, the dose, the time of the last dose, and kidney function. A patient who took apixaban 30 hours ago with a creatinine clearance of 80 mL/min has little drug left, and the reversal agent adds thrombotic risk without hemostatic benefit. A patient with a creatinine clearance of 25 mL/min who took dabigatran 30 hours ago may still be fully anticoagulated. The same assessment records interacting drugs that raise DOAC exposure (strong P-glycoprotein or CYP3A4 inhibitors such as ketoconazole and, for dabigatran, dronedarone and verapamil). It also records concomitant antiplatelet drugs and nonsteroidal anti-inflammatory drugs, liver disease, thrombocytopenia, and uremia.

Approximate pharmacokinetic values from US prescribing information. Half-lives lengthen with kidney impairment for dabigatran, edoxaban, enoxaparin, and fondaparinux. Parenteral agents are described in detail on the parenteral anticoagulants and heparin-induced thrombocytopenia page.

Kidney function is estimated by Cockcroft-Gault creatinine clearance with actual body weight, the method used in the DOAC trials and US labeling. Acute kidney injury from hypovolemia or contrast is common in bleeding patients and prolongs DOAC clearance during the hours when the decision is made. The most recent creatinine is therefore used, and a single admission value is not assumed to be stable.

The bleeding source is identified in parallel. Endoscopy, angiography with embolization, surgery, packing, and compression stop more bleeding than any drug. A reversal agent shortens the time during which the coagulopathy adds to bleeding, but it does not seal a visible vessel, an ulcer, or a ruptured aneurysm.

Laboratory assessment of anticoagulant effect

Routine coagulation tests measure warfarin and unfractionated heparin well and DOACs poorly. The prothrombin time (PT) and its standardized ratio, the international normalized ratio (INR), quantify the warfarin effect. The activated partial thromboplastin time (aPTT) and the anti–factor Xa activity calibrated for heparin quantify unfractionated heparin; low-molecular-weight heparin is measured by heparin-calibrated anti-Xa activity. For the DOACs, the ECDP separates two purposes: excluding a clinically relevant drug level and quantifying the level.

Interpretation of coagulation tests in bleeding patients, from the 2020 ACC ECDP (Tables 2 and 3). Liquid chromatography–tandem mass spectrometry is the reference method for DOAC concentrations and is rarely available in time for an emergency decision.

Drug-level thresholds. The ISTH Scientific and Standardization Committee suggests that a reversal agent be considered for serious bleeding when the DOAC level exceeds 50 ng/mL. Before an invasive procedure with high bleeding risk, it suggests the same when the level exceeds 30 ng/mL. ANNEXA-4 analyzed efficacy only in patients with a baseline anti-Xa level of at least 75 ng/mL, because lower levels were judged too small to need reversal. The 30-ng/mL threshold reappears in the fifth edition of the American Society of Regional Anesthesia and Pain Medicine (ASRA) guidelines as the residual level at which neuraxial block can proceed when time-based intervals cannot be met. These thresholds are consensus values; no trial has tested treatment decisions based on them.

Viscoelastic tests. Thromboelastography and rotational thromboelastometry measure whole-blood clot formation and are used to guide transfusion in trauma, liver disease, and cardiac surgery. They are not specific for any anticoagulant and do not replace a drug level, but they identify hypofibrinogenemia, hyperfibrinolysis, and platelet dysfunction that add to anticoagulant bleeding.

When no assay result can be obtained in time, treatment proceeds on the clinical history. A patient with life-threatening bleeding who took a DOAC within the last dosing interval, or whose kidney function is impaired, is treated as anticoagulated.

General measures: stopping the drug, source control, and resuscitation

Every major bleed is managed with the same steps regardless of the anticoagulant. The anticoagulant and any antiplatelet drug are stopped. Local hemostasis is applied: compression, packing, tamponade devices, or endoscopic therapy. Hemodynamic instability is treated with isotonic crystalloid and blood products, with correction of hypothermia, acidosis, and hypocalcemia, which impair coagulation independently of the drug. The appropriate service (surgery, interventional radiology, gastroenterology, neurosurgery) is involved early, and the decision to give a reversal agent is made at the same time and not after these measures fail. Coexisting causes of bleeding are treated: thrombocytopenia, uremic platelet dysfunction, liver disease, and drug interactions.

Adjuncts that support hemostasis have specific indications:

• Activated charcoal. Oral activated charcoal, 50 g, is suggested by the ECDP when a DOAC was ingested within the previous 2 to 4 hours. The 2022 American Heart Association/American Stroke Association (AHA/ASA) intracerebral hemorrhage guideline rates charcoal within the previous few hours as possibly reasonable (Class 2b). It is avoided when the airway is unprotected or when endoscopy is planned for upper gastrointestinal bleeding.
• Hemodialysis. Dabigatran is the only oral anticoagulant removed by dialysis, because of its low protein binding. A 4-hour session removes about half of circulating drug, with redistribution afterward. Dialysis is used when idarucizumab is unavailable, in severe kidney failure, or in overdose. Factor Xa inhibitors are highly protein bound and are not dialyzable.
• Desmopressin. Desmopressin 0.3 µg/kg intravenously raises von Willebrand factor and factor VIII and shortens the bleeding time in uremic platelet dysfunction. The ECDP lists it with cryoprecipitate and dialysis for bleeding patients with kidney failure. In intracerebral hemorrhage on antiplatelet drugs, its effect on hematoma growth is uncertain (2022 AHA/ASA, Class 2b); the Neurocritical Care Society guideline suggests a single dose of 0.4 µg/kg. Hyponatremia follows repeated doses.
• Tranexamic acid. The antifibrinolytic reduces death from bleeding in trauma when given within 3 hours of injury, and the ECDP endorses that use. In HALT-IT, among 12,009 patients with significant gastrointestinal bleeding, a high-dose 24-hour infusion did not reduce death from bleeding within 5 days (4% in both groups; risk ratio 0.99) and increased venous thromboembolism (0.8% versus 0.4%; risk ratio 1.85). Tranexamic acid is therefore not given routinely for gastrointestinal bleeding. It is used topically as a mouthwash after dental extraction in patients who continue warfarin.

Transfusion thresholds in anticoagulant-associated bleeding

Red cells. In patients with acute upper gastrointestinal bleeding, a randomized trial found that a restrictive strategy (transfusion when hemoglobin fell below 7 g/dL) improved survival and reduced rebleeding compared with a liberal threshold of 9 g/dL. The trial excluded patients with massive exsanguinating bleeding and acute coronary syndromes. The ECDP adopts a hemoglobin of 7 g/dL as the threshold for patients with symptomatic anemia or active bleeding and 8 g/dL for patients with coronary artery disease, particularly those with an acute coronary syndrome. In MINT, 3,504 patients with acute myocardial infarction and a hemoglobin below 10 g/dL were assigned to a liberal strategy (transfusion to 10 g/dL or more) or a restrictive strategy (threshold 7 or 8 g/dL). The liberal strategy did not significantly reduce 30-day death or recurrent infarction. The primary outcome occurred in 14.5% with the liberal and 16.9% with the restrictive strategy, which leaves open a benefit of liberal transfusion in this population.

Massive transfusion. The ECDP suggests that activation of a massive transfusion protocol be considered when 3 or more units of red cells are needed within 1 hour. Balanced ratios of plasma and platelets to red cells, or goal-directed replacement guided by viscoelastic testing, are used. Ionized calcium falls with citrate load and is replaced.

Platelets and fibrinogen. The ECDP suggests platelet transfusion to keep the count at 50 × 10⁹/L or more in active bleeding, and cryoprecipitate or fibrinogen concentrate to keep fibrinogen above 100 mg/dL. Platelet transfusion does not reverse antiplatelet drugs reliably. In PATCH, among 190 patients with intracerebral hemorrhage taking antiplatelet therapy, platelet transfusion increased the odds of death or dependence at 3 months (adjusted common odds ratio 2.05). The 2022 AHA/ASA guideline classes platelet transfusion as potentially harmful in patients taking aspirin who are not undergoing emergency neurosurgery, and as a possible option (Class 2b) before emergency neurosurgery. Transfused platelets are inhibited while ticagrelor or its active metabolite circulates.

Plasma. Fresh frozen plasma is used for warfarin reversal only when four-factor prothrombin complex concentrate is not available, at 10 to 15 mL/kg. It requires thawing and ABO matching, takes hours to infuse, corrects the INR incompletely at feasible volumes, and causes transfusion-associated circulatory overload in older patients with heart disease. Plasma has no role in DOAC reversal.

Reversal and hemostatic agents by anticoagulant

Reversal agents are of two kinds. Specific antidotes bind the anticoagulant: protamine for heparin, idarucizumab for dabigatran, and andexanet alfa for factor Xa inhibitors. Nonspecific hemostatic agents supply clotting factors in amounts that overcome the drug. They include prothrombin complex concentrates (PCCs), which contain factors II, IX, and X with (four-factor, 4F-PCC) or without (three-factor) factor VII. Activated PCC (aPCC; factor eight inhibitor bypassing activity, FEIBA) contains activated factor VII and other partly activated factors. Recombinant activated factor VII and plasma are also used. The ECDP suggests a reversal or hemostatic agent for life-threatening bleeding and for major bleeding that does not resolve with initial management, and not for most nonmajor bleeding.

Doses from the 2020 ACC ECDP (Figure 3 and Table 5), US prescribing information, and the 2016 Neurocritical Care Society/Society of Critical Care Medicine reversal guideline. 4F-PCC, four-factor prothrombin complex concentrate; aPCC, activated prothrombin complex concentrate; ACT, activated clotting time.

The intensity of reversal follows the bleed. For a critical-site or life-threatening bleed, the agent is given as soon as the drug and the last dose are known, in parallel with imaging and source control. For a major bleed that is being controlled, supportive measures and the passage of time are often enough with DOACs, whose effect falls by half within about 12 hours with normal kidney function. Every reversal shifts the patient toward thrombosis, because the underlying indication for anticoagulation remains and some agents are prothrombotic in their own right.

Reversal of warfarin: vitamin K and four-factor prothrombin complex concentrate

Warfarin lowers functional factors II, VII, IX, and X. Vitamin K restores their synthesis, but the INR begins to fall only several hours after intravenous vitamin K and takes 12 to 24 hours to normalize. Four-factor PCC supplies the factors and corrects the INR within minutes of the end of the infusion. The two are given together for major bleeding: the PCC for immediate correction and the vitamin K so that the correction persists after the infused factors are cleared, since the half-life of factor VII is about 6 hours.

Four-factor PCC dosing. Kcentra, the first four-factor PCC approved in the United States, and Balfaxar, approved later, are dosed in units of factor IX by the pretreatment INR and body weight. Actual weight is used up to 100 kg; above 100 kg the maximum dose is not exceeded. Repeat dosing has not been studied and is not recommended in the Kcentra prescribing information. Kcentra contains heparin and is contraindicated in heparin-induced thrombocytopenia and in disseminated intravascular coagulation.

Kcentra prescribing information and the 2020 ACC ECDP (Figure 3). The infusion rate is 0.12 mL/kg/min (about 3 units/kg/min), up to 8.4 mL/min. The INR is measured about 30 minutes after the infusion ends.

Vitamin K. For major bleeding, 5 to 10 mg of vitamin K (phytonadione) is given by slow intravenous infusion, which acts faster and more reliably than oral or subcutaneous dosing. Anaphylactoid reactions to intravenous vitamin K are rare and are associated with rapid injection. For bleeding in patients with a mechanical valve, the 2020 ACC/AHA valve guideline rates intravenous vitamin K as reasonable when warfarin will not be resumed for about 7 days (Class 2a). Larger doses of vitamin K make the patient resistant to warfarin when it is restarted. Management of a high INR without bleeding, where oral vitamin K at low doses is used, is described on the warfarin page.

Trial evidence. Four-factor PCC was compared with plasma in two randomized trials. In the trial in patients with major bleeding (202 patients), effective hemostasis at 24 hours was 72.4% with 4F-PCC and 65.4% with plasma, meeting noninferiority. Rapid INR correction to 1.3 or less at 30 minutes after the infusion occurred in 62.2% versus 9.6%. In the trial in patients needing urgent surgery or an invasive procedure (181 patients), effective hemostasis was 90% versus 75% and rapid INR correction 55% versus 10%, both superior with 4F-PCC. Fluid overload occurred in 3% versus 13%. In INCH, among 50 patients with warfarin-associated intracranial hemorrhage, an INR of 1.2 or less within 3 hours was reached in 67% with PCC 30 units/kg and 9% with plasma 20 mL/kg. The trial was stopped early for safety. Hematoma expansion caused five deaths in the plasma group and one in the PCC group.

Guideline positions. The 2023 ACC/AHA/American College of Chest Physicians (ACCP)/Heart Rhythm Society (HRS) atrial fibrillation (AF) guideline recommends 4F-PCC over plasma for life-threatening bleeding on warfarin (Class 1, Level A). The 2022 AHA/ASA intracerebral hemorrhage guideline recommends 4F-PCC over plasma in warfarin-associated hemorrhage with an INR of 2.0 or more (Class 1) and intravenous vitamin K immediately after factor replacement (Class 1). With an INR of 1.3 to 1.9, PCC may be reasonable (Class 2b). The American Society of Hematology (ASH) 2018 guideline also prefers 4F-PCC to plasma. Three-factor PCC, which contains little factor VII, corrects the INR less completely and is not used when four-factor PCC is available. Recombinant factor VIIa normalizes the INR by supplying factor VII alone without replacing prothrombin or factor X, and it is not used for warfarin reversal.

Gastrointestinal bleeding on warfarin. The 2022 American College of Gastroenterology–Canadian Association of Gastroenterology (ACG-CAG) guideline takes a more restrictive position for gastrointestinal bleeding, in which hemostasis is usually achieved endoscopically. It suggests against plasma and against vitamin K and suggests PCC over plasma if reversal is needed. It found the evidence insufficient to recommend for or against PCC compared with no PCC. The difference reflects the bleeding site. In variceal or ulcer bleeding that can be treated at endoscopy, the cost of reversal (thrombosis and later warfarin resistance) is weighed against a bleeding risk that endoscopic therapy can control.

Reversal of dabigatran: idarucizumab

Idarucizumab is a humanized monoclonal antibody fragment that binds free and thrombin-bound dabigatran with an affinity about 350 times that of dabigatran for thrombin. The complex is cleared by the kidney. The dose is 5 g intravenously, given as two consecutive 2.5-g infusions or bolus injections. Reversal is complete within minutes and does not depend on kidney function.

In the full cohort of RE-VERSE AD, 503 patients received idarucizumab: 301 with uncontrolled bleeding (group A; 45.5% gastrointestinal and 32.6% intracranial) and 202 needing an urgent procedure (group B). The median maximum reversal of dabigatran, measured by dilute thrombin time or ecarin clotting time, was 100%. In group A, the median time to cessation of bleeding was 2.5 hours. In group B, the median time to the procedure was 1.6 hours, and periprocedural hemostasis was normal in 93.4%. At 90 days, thrombotic events had occurred in 6.3% and 7.4% and death in 18.8% and 18.9% of the two groups, figures that reflect the age, comorbidity, and interrupted anticoagulation of the population. The study had no control group.

A minority of patients have a rise in dabigatran concentration 12 to 24 hours after idarucizumab, as drug redistributes from the extravascular space, particularly when kidney function is poor. The ECDP states that a second 5-g dose may be reasonable when bleeding persists or recurs with laboratory evidence of persistent dabigatran effect, or before a second invasive procedure. Dabigatran can be restarted 24 hours after idarucizumab when the clinical situation allows.

When idarucizumab is not available, PCC or activated PCC 50 units/kg (maximum 4,000 units) is suggested by the ECDP, with hemodialysis when drug levels are very high or kidney function is impaired. The 2023 AF guideline recommends idarucizumab for life-threatening bleeding on dabigatran (Class 1) and rates activated PCC as reasonable when it is unavailable (Class 2a). The 2022 AHA/ASA guideline rates idarucizumab as reasonable in dabigatran-associated intracerebral hemorrhage (Class 2a). The ACG-CAG guideline suggests against idarucizumab in gastrointestinal bleeding, which usually stops with endoscopic therapy and drug clearance.

Reversal of factor Xa inhibitors: andexanet alfa and prothrombin complex concentrate

Andexanet alfa. Andexanet alfa is a recombinant, catalytically inactive form of human factor Xa that binds apixaban, rivaroxaban, and edoxaban (and the antithrombin–enoxaparin complex) and frees native factor Xa. It received accelerated US approval in 2018 for life-threatening or uncontrolled bleeding on apixaban or rivaroxaban, on the basis of ANNEXA-4. The labeled regimen was a bolus followed by a 2-hour infusion. The high dose (800-mg bolus at 30 mg/min, then 8 mg/min for 120 minutes) was used when two conditions were met. The last dose was more than 5 mg of apixaban, more than 10 mg of rivaroxaban, or of unknown size. It had been taken less than 8 hours earlier or at an unknown time. The low dose (400-mg bolus at 30 mg/min, then 4 mg/min for 120 minutes) was used when the last dose was 5 mg or less of apixaban or 10 mg or less of rivaroxaban. It was also used when the last dose had been taken 8 hours or more earlier. Its effect ends within about 2 hours of the end of the infusion, as anti-Xa activity returns toward the untreated level.

ANNEXA-4 enrolled 352 patients with acute major bleeding within 18 hours of a factor Xa inhibitor dose; 64% had intracranial and 26% gastrointestinal bleeding. Anti-Xa activity fell by 92% with apixaban and with rivaroxaban, and adjudicated hemostatic efficacy at 12 hours was excellent or good in 82% of evaluable patients. Within 30 days, 14% died and 10% had a thrombotic event. The reduction in anti-Xa activity did not predict hemostatic efficacy overall.

ANNEXA-I, the randomized trial required for full approval, enrolled 530 patients with acute intracerebral hemorrhage within 15 hours of a factor Xa inhibitor dose. Andexanet was compared with usual care, which was PCC in 85.5% of the control group. Hemostatic efficacy was a composite of hematoma expansion of 35% or less, an increase in the National Institutes of Health Stroke Scale score of less than 7 points at 12 hours, and no rescue therapy. It was achieved in 67.0% with andexanet and 53.1% with usual care (adjusted difference 13.4 percentage points). Anti-Xa activity fell by a median of 94.5% versus 26.9%. Thrombotic events occurred in 10.3% versus 5.6%, ischemic stroke in 6.5% versus 1.5%, and there was no difference in functional outcome or 30-day mortality. On the basis of these data and postmarketing safety reports, the US Food and Drug Administration concluded that the risks of andexanet outweighed its benefits. The manufacturer withdrew the product from the US market in December 2025. Andexanet remains licensed in Europe, the United Kingdom, and Japan.

Prothrombin complex concentrate. Four-factor PCC does not bind factor Xa inhibitors. It raises the concentrations of prothrombin and factor X so that thrombin generation recovers despite the inhibitor. Evidence for its use comes from ex vivo studies, volunteer bleeding models, and observational cohorts. In one cohort, a fixed 2,000 units achieved hemostatic efficacy in 85% of 66 patients; in another, weight-based doses (1,500 units below 65 kg and 2,000 units above) achieved efficacy in 69% of 84 patients. On this basis the ECDP rates a fixed dose of 2,000 units as reasonable for severe or life-threatening bleeding on a factor Xa inhibitor. Doses of 25 to 50 units/kg are used in other protocols, and no dose has been compared with another or with placebo. The usual-care arm of ANNEXA-I, which mostly received PCC, had a lower thrombotic event rate than the andexanet arm. In the United States, 4F-PCC is now the agent used for life-threatening bleeding on apixaban, rivaroxaban, and edoxaban. Activated PCC 50 units/kg is an alternative in some institutional protocols, at the cost of greater thrombotic potential.

Guideline positions before and after the withdrawal. The 2020 ECDP and the 2022 AHA/ASA guideline (andexanet Class 2a; PCC or activated PCC Class 2b) preferred andexanet when it was available. The 2023 AF guideline listed andexanet or 4F-PCC. These recommendations predate both ANNEXA-I and the US withdrawal, and current US practice follows the PCC alternatives that the same documents list. The ACG-CAG guideline suggests against PCC and against andexanet in gastrointestinal bleeding on DOACs, since most such bleeding is controlled by endoscopy and drug clearance.

Andexanet and heparin. Andexanet binds the antithrombin–heparin complex and was associated with failure of heparin anticoagulation in patients who went on to cardiopulmonary bypass or extracorporeal circuits. Where andexanet is still used, it is avoided before surgery requiring systemic heparinization.

Reversal agents in development. Ciraparantag, a small molecule that binds heparins and direct oral anticoagulants, has been studied in volunteers. Abelacimab and other factor XI inhibitors, in development as anticoagulants with lower bleeding rates, would need their own reversal strategies.

Reversal of heparin, low-molecular-weight heparin, fondaparinux, and parenteral direct thrombin inhibitors

Unfractionated heparin. Protamine sulfate, a basic protein, combines with heparin to form an inactive salt. One milligram neutralizes about 100 units of heparin. Because heparin has a short half-life, the dose is calculated from the heparin given in the preceding 2 to 3 hours. A patient receiving 1,250 units/h needs about 25 to 38 mg for the 2,500 to 3,750 units infused in the last 2 to 3 hours, plus any bolus within that time. A single dose does not exceed 50 mg, and protamine is infused at no more than 5 mg/min. The aPTT or activated clotting time is measured after the dose. The 2022 AHA/ASA guideline rates protamine as reasonable for heparin-associated intracerebral hemorrhage (Class 2a).

Low-molecular-weight heparin. Protamine neutralizes the anti-IIa activity of low-molecular-weight heparin but at most about 60% of its anti-Xa activity, because the shorter chains bind protamine less well. For enoxaparin given within 8 hours, the dose is 1 mg of protamine per 1 mg of enoxaparin; between 8 and 12 hours, 0.5 mg per 1 mg; beyond 12 hours, protamine adds little. A second dose of 0.5 mg per 1 mg can be given if bleeding continues. For dalteparin and tinzaparin, 1 mg neutralizes 100 anti-Xa units given in the preceding 8 hours. The 2022 AHA/ASA guideline rates protamine for low-molecular-weight heparin as possibly reasonable (Class 2b).

Adverse reactions to protamine. Hypotension, bradycardia, pulmonary vasoconstriction with acute right ventricular failure, and anaphylactoid reactions occur, most often with rapid infusion. Prior protamine exposure (including protamine-containing insulin such as neutral protamine Hagedorn (NPH) insulin), fish allergy, and vasectomy are associated with antibodies to protamine. Excess protamine has a weak anticoagulant effect of its own.

Fondaparinux. No antidote exists, and protamine does not bind fondaparinux. The Neurocritical Care Society guideline suggests activated PCC in fondaparinux-associated intracranial hemorrhage, with recombinant factor VIIa as an alternative, both on low-quality evidence. Its long half-life (17 to 21 hours, longer in kidney impairment) makes fondaparinux a poor choice when bleeding or an urgent procedure is likely.

Bivalirudin and argatroban. Both direct thrombin inhibitors have short half-lives (bivalirudin about 25 minutes, prolonged in kidney failure; argatroban about 45 minutes, prolonged in liver disease), and stopping the infusion is the main treatment. Bivalirudin is partly removed by hemodialysis. Their use in heparin-induced thrombocytopenia and during percutaneous coronary intervention (PCI) is described on the parenteral anticoagulants page.

Four-factor prothrombin complex concentrate dose calculator

The calculator gives the labeled INR-based 4F-PCC dose for warfarin reversal, the ECDP fixed-dose alternatives, and the weight-based range used off-label for factor Xa inhibitor bleeding.

Warfarin: 25 units/kg for INR 2.0 to below 4.0 (maximum 2,500 units), 35 units/kg for INR 4.0 to 6.0 (maximum 3,500), and 50 units/kg above 6.0 (maximum 5,000), with weight capped at 100 kg (Kcentra prescribing information). Fixed doses: 1,000 units for non-intracranial and 1,500 units for intracranial bleeding on warfarin, and 2,000 units for factor Xa inhibitor bleeding (2020 ACC ECDP).

Bleeding on antiplatelet drugs and combined antithrombotic therapy

No routinely available drug reverses aspirin, clopidogrel, prasugrel, or ticagrelor. Aspirin and the thienopyridines inhibit platelets irreversibly, so their effect ends only as new platelets enter the circulation (about 10% of the platelet mass per day). Ticagrelor binds reversibly, but while it and its active metabolite circulate they inhibit transfused platelets as well, so platelet transfusion within about 24 hours of the last dose is largely ineffective. Bentracimab, a monoclonal antibody fragment that binds ticagrelor, reversed its effect within minutes in studies of urgent surgery and major bleeding but is not routinely available in the United States.

Management of bleeding on antiplatelet drugs therefore rests on source control, stopping the drug when the ischemic risk allows, and correcting other defects. Platelet transfusion is reserved for life-threatening bleeding or emergency neurosurgery in a patient taking an irreversible inhibitor; in intracerebral hemorrhage without surgery it is harmful (PATCH). Desmopressin is an option for antiplatelet-associated intracranial hemorrhage, with uncertain benefit. In acute gastrointestinal bleeding, the ACG-CAG guideline suggests continuing aspirin taken for secondary prevention, or resuming it on the day hemostasis is confirmed endoscopically, and suggests against platelet transfusion.

The decision to stop an antiplatelet drug in a recently stented patient weighs bleeding against stent thrombosis. Stopping both drugs of dual antiplatelet therapy within the first month after stenting carries a high risk of stent thrombosis and myocardial infarction. In patients taking an anticoagulant with one or two antiplatelet drugs, the antiplatelet drug is usually the first agent stopped, and the anticoagulant is reversed only for life-threatening bleeding. Duration and de-escalation of antiplatelet therapy after bleeding are covered on the dual antiplatelet therapy and antithrombotic therapy in atrial fibrillation with PCI pages.

Intracranial hemorrhage on anticoagulants: reversal, blood pressure, and neurosurgery

Intracranial hemorrhage is the bleed in which minutes matter. Hematoma expansion occurs mostly in the first hours, is more frequent and more prolonged in anticoagulated patients, and predicts death and disability. Every anticoagulated patient with a new headache, focal deficit, or depressed consciousness after a fall is imaged with noncontrast CT without delay. When hemorrhage is confirmed, reversal begins before neurosurgical consultation is complete, and the two proceed in parallel.

Selected recommendations of the 2022 AHA/ASA guideline for spontaneous intracerebral hemorrhage. The factor Xa recommendations predate ANNEXA-I and the US withdrawal of andexanet.

Targets of reversal. For warfarin-associated hemorrhage, the INR is measured after PCC and again over the following hours, with a target of 1.3 or less in most protocols (the INCH trial used 1.2). For DOACs, the drug level is not available fast enough to guide the first dose, and reversal is given when the history indicates likely drug effect. Repeat CT at about 6 to 24 hours identifies hematoma expansion and guides further reversal or surgery.

Neurosurgical care. Large cerebellar hemorrhage with brainstem compression or hydrocephalus is treated surgically. Supratentorial evacuation, external ventricular drainage, and minimally invasive techniques are decided by the neurosurgical team; each requires correction of the coagulopathy first. Subdural hematoma, common after falls in older anticoagulated patients, may need burr-hole drainage or middle meningeal artery embolization. Traumatic and spontaneous hemorrhages differ in prognosis and in the risk of recurrence when anticoagulation resumes.

Venous thromboembolism prophylaxis. Intermittent pneumatic compression begins on admission. Low-dose heparin or low-molecular-weight heparin prophylaxis is commonly started 24 to 48 hours after the hemorrhage once it is shown to be stable on repeat imaging.

Thrombotic risk after reversal

Reversal returns a patient to the thrombotic state that justified anticoagulation in the first place, and some agents add to that risk. Thrombotic events after reversal arise from three sources. The first is the untreated indication: atrial fibrillation, a mechanical valve, or recent venous thromboembolism. The second is the prothrombotic state of acute bleeding, surgery, immobility, and transfusion. The third is the agent itself: PCCs supply clotting factors, activated PCC and recombinant factor VIIa contain activated factors, and andexanet binds tissue factor pathway inhibitor as well as the target drug.

The event rates in the reversal studies show the size of the problem. In the two randomized trials of 4F-PCC against plasma, the Kcentra label reports possible thromboembolic events in 6.8% (13 of 191) with 4F-PCC and 7.1% (14 of 197) with plasma. In the trial of acute major bleeding, thromboembolic events were more frequent with 4F-PCC (8.7% versus 5.5%). The difference was larger in patients with a thromboembolic event more than 3 months before entry (11.6% versus 3.8%). In the trial of urgent procedures, they were less frequent with 4F-PCC (4.5% versus 9.1%). None of these differences was statistically significant. A pooled post hoc analysis of the same 388 patients, which counted every event that investigators identified with predefined thromboembolic terms, found thromboembolic events in 7.3% (14 of 191) with 4F-PCC and 7.1% (14 of 197) with plasma, with no relation to factor levels above the normal range. The authors attribute the difference from the label to the Food and Drug Administration's different classification of some events; by either count, the rates with 4F-PCC and plasma were similar. The Kcentra label carries a boxed warning for arterial and venous thromboembolism and notes that patients with a thromboembolic event in the prior 3 months were not studied. Thrombotic events occurred in 6.3% to 7.4% within 90 days after idarucizumab, which has no prothrombotic activity of its own. That rate reflects interrupted anticoagulation. Thrombotic events occurred in 10% within 30 days in ANNEXA-4 and in 10.3% versus 5.6% in ANNEXA-I, where the comparison with usual care isolated an effect of andexanet.

The same studies show that resumption reduces the risk. In observational analyses of patients reversed with PCC or andexanet, thrombotic events were concentrated in those not restarted on anticoagulation. Reversal is therefore paired with a plan for resumption from the first day: mechanical prophylaxis, early pharmacologic prophylaxis when hemostasis allows, and a documented decision on full-dose anticoagulation.

Restarting anticoagulation after gastrointestinal bleeding

Most patients whose gastrointestinal bleeding is controlled benefit from restarting anticoagulation. The ECDP summarizes a systematic review of 12 observational studies (3,098 patients). Resumption after anticoagulant-associated gastrointestinal bleeding was associated with less thromboembolism (relative risk 0.30) and lower mortality (relative risk 0.51), at the cost of more recurrent bleeding (relative risk 1.91). The timing has not been tested in randomized trials. In a cohort of 1,329 warfarin-treated patients with atrial fibrillation and major gastrointestinal bleeding, restarting warfarin after 7 days was associated with lower mortality and thromboembolism than restarting after 30 days, without more recurrent bleeding than later resumption. Restarting within the first 7 days was associated with more rebleeding.

The ECDP concludes that for most gastrointestinal bleeding it is reasonable to restart oral anticoagulation once hemostasis has been achieved. The ASH 2018 guideline suggests resuming oral anticoagulation within 90 days of a major bleed in patients with a continuing indication. The 2023 AF guideline rates resumption as possibly reasonable after correction of reversible causes and multidisciplinary reassessment (Class 2b). A bleed from a treatable source (an ulcer, an angiodysplasia that has been ablated, a polypectomy site) supports resumption; occult bleeding from an unidentified source, recurrent bleeding from diffuse angiodysplasia, and bleeding with an unchangeable cause weigh against it.

Several measures reduce the risk of recurrence when anticoagulation resumes:

• A proton pump inhibitor in patients at high risk of gastrointestinal bleeding who take oral anticoagulants, antiplatelet drugs, or both (2025 ACC/AHA acute coronary syndrome guideline, Class 1), with eradication of Helicobacter pylori after ulcer bleeding. The evidence is reviewed on the dual antiplatelet therapy page.
• Withdrawal of antiplatelet drugs that lack a current indication and of nonsteroidal anti-inflammatory drugs.
• Correct dosing for kidney function, age, and weight, and review of interacting drugs.
• Choice of anticoagulant. Apixaban had gastrointestinal bleeding rates similar to warfarin in ARISTOTLE, whereas rivaroxaban, edoxaban 60 mg, and dabigatran 150 mg had more gastrointestinal bleeding than warfarin in their pivotal trials. Low-dose DOAC regimens below the labeled dose are not used to reduce bleeding, because they reduce stroke prevention.

Restarting anticoagulation after intracranial hemorrhage

After intracranial hemorrhage, the balance is closer and depends on the cause of the hemorrhage and the strength of the thrombotic indication. Spontaneous lobar hemorrhage, especially with cerebral amyloid angiopathy (multiple cortical microbleeds, cortical superficial siderosis, or convexity subarachnoid hemorrhage on MRI), carries the highest risk of recurrence. Deep hemorrhage from hypertensive arteriopathy, once blood pressure is controlled, and traumatic hemorrhage carry a lower risk. Spontaneous subdural hematoma also recurs often. MRI with susceptibility-weighted imaging is obtained before the decision because it identifies these markers.

Observational studies of warfarin-associated hemorrhage reported 50% to 70% lower risks of thrombosis and death with resumption, without a significant increase in recurrent hemorrhage, but these studies select patients who were judged fit to restart. The ECDP favors delaying resumption for at least 4 weeks in patients without high thrombotic risk, and the median time to restart in the largest supportive cohort was about 1 month. Randomized trials have tested the decision.

Randomized evidence on anticoagulation after intracranial hemorrhage in atrial fibrillation. HR, hazard ratio.

The trials agree that anticoagulation after intracranial hemorrhage nearly abolishes ischemic stroke and that part of this benefit is lost to recurrent hemorrhage, concentrated in lobar and amyloid-related bleeding. Resumption is favored after traumatic or deep hemorrhage with controlled blood pressure, in patients with a strong thrombotic indication, and with a DOAC in preference to warfarin, since DOACs cause about half as much intracranial hemorrhage. Avoidance is favored after lobar hemorrhage with cerebral amyloid angiopathy markers, recurrent hemorrhage, or uncontrolled hypertension. The 2023 AF guideline states that delayed resumption, 4 to 8 weeks after the hemorrhage, may be considered (Class 2b). In patients with a very high thrombotic risk, such as a mechanical valve or rheumatic heart disease, early resumption at 1 to 2 weeks is reasonable (Class 2a), with warfarin for a mechanical valve. The same guideline also states that an anticoagulation-sparing strategy such as left atrial appendage occlusion may be considered in patients at high risk of recurrent hemorrhage, such as those with cerebral amyloid angiopathy (Class 2b). Further aspects of stroke and hemorrhage in atrial fibrillation are covered on the atrial fibrillation page.

Timing of resumption by bleeding site and thrombotic risk

The ECDP frames resumption as a sequence of five questions:

• Whether the bleed was at a critical site.
• Whether an urgent procedure is planned.
• Whether the patient is at high risk of rebleeding.
• Whether thrombotic risk is low enough that anticoagulation can stop.
• Whether the patient wishes to restart.

Patients at low thrombotic risk (for example, atrial fibrillation with a low CHA₂DS₂-VASc score, or venous thromboembolism provoked more than 3 months earlier) can discontinue anticoagulation. Patients with a critical-site bleed or an unidentified source have resumption delayed. Others restart as soon as it is medically safe.

Synthesized from the 2020 ACC ECDP (Section 5.7 and Figures 4 to 6), the 2023 AF guideline, and the trials above. Timing in each case is individualized with the specialists who managed the bleed.

The decision is shared with the patient. The ECDP recommends a discussion before resumption. It covers the warning signs of recurrent bleeding at the original site and the personal risk of thrombosis, estimated by a score such as CHA₂DS₂-VASc. It also covers the consequences of cardioembolic stroke, which is more often disabling and fatal than other ischemic strokes. When a patient declines resumption, the decision and a date for review are recorded.

Left atrial appendage occlusion after major bleeding. In patients with atrial fibrillation in whom long-term anticoagulation is contraindicated because of a nonreversible bleeding cause, percutaneous left atrial appendage occlusion is an alternative. It still requires a short period of antithrombotic therapy after implantation. Device selection, trials, and recommendations are described on the atrial fibrillation page.

Bleeding risk scores and modifiable bleeding risk

Bleeding risk scores identify patients in whom bleeding is more likely and the factors that can be changed. They predict major bleeding modestly, with C-statistics near 0.6 to 0.7, and several of their elements (age, hypertension, kidney disease, prior stroke) also raise thrombotic risk. The 2023 AF guideline states that bleeding scores are not used in isolation to withhold anticoagulation from patients at high stroke risk. HAS-BLED, its calculator, and the full discussion of modifiable bleeding risk are on the atrial fibrillation page.

The ORBIT bleeding score (derived in the ORBIT-AF registry). Scores of 0 to 2 are low risk, 3 medium, and 4 to 7 high. In its derivation it predicted major bleeding similarly to a 12-variable model (C-index 0.67) and was better calibrated than HAS-BLED and ATRIA in an external trial population.

ORBIT was derived in patients already taking oral anticoagulants and weights the two strongest predictors of bleeding, anemia and prior bleeding. After a major bleed, each patient scores at least 2 points for bleeding history. The score is therefore used to identify the remaining modifiable elements (anemia that can be treated, an antiplatelet drug that can be stopped) and not to decide whether anticoagulation resumes.

Falls, frailty, and anticoagulation decisions

Fear of falls is among the most common reasons older patients with atrial fibrillation are not anticoagulated, and it rarely justifies that decision. A fall produces anticoagulant-attributable intracranial hemorrhage only through a chain of events: the fall must involve the head, the impact must cause hemorrhage, and the hemorrhage must be one that would not have occurred without the anticoagulant. In a published decision analysis, the optimal antithrombotic choice for older patients with atrial fibrillation did not depend on their risk of falling. A patient taking warfarin would have to fall about 295 times in a year for the risk of subdural hematoma to outweigh the benefit. Absolute stroke risk rises with CHA₂DS₂-VASc score, and the absolute benefit of anticoagulation is largest in the oldest and most frail patients, who are also the most likely to be denied it.

The choice of drug changes the balance. DOACs caused about half as much intracranial hemorrhage as warfarin in the atrial fibrillation trials, so a DOAC at the labeled dose is preferred in an older patient who falls. In frail patients already stable on warfarin, however, FRAIL-AF found that switching to a DOAC increased bleeding without reducing thromboembolism. The 2023 AF guideline lists serious bleeding related to recurrent falls whose cause cannot be treated among the contraindications to long-term anticoagulation.

Management of a patient who falls includes a fall-risk assessment and treatment of its causes (orthostatic hypotension, sedative and antihypertensive drugs, visual impairment, neuropathy, arrhythmia), physical therapy, home safety measures, and education that every head injury on an anticoagulant requires urgent evaluation. Noncontrast CT is obtained after head trauma in anticoagulated patients even when the examination is reassuring, because delayed intracranial hemorrhage occurs. The older-adult aspects of anticoagulation, including ELDERCARE-AF, are discussed on the atrial fibrillation page.

Procedural bleeding risk classification

Periprocedural management depends on two risks: the bleeding risk of the procedure and the thrombotic risk of the patient. The 2022 American College of Chest Physicians (CHEST) guideline classifies procedures into three groups by the expected 30-day risk of major bleeding.

Empiric procedural bleeding risk classification of the 2022 CHEST perioperative antithrombotic guideline. Coronary angiography is classed by access: radial access lowers access-site bleeding.

Patient factors raise the procedural risk within each class: prior bleeding with a similar procedure, kidney or liver disease, thrombocytopenia, concomitant antiplatelet drugs, and an inherited bleeding disorder. Bleeding at certain sites carries consequences out of proportion to its volume (intracranial, spinal, intraocular, pericardial), and procedures at those sites are treated as high risk even when the expected bleeding is small.

Thromboembolic risk during interruption

The same guideline stratifies the risk of thromboembolism during interruption of a vitamin K antagonist. The strata determine whether heparin bridging is considered and how short the interruption should be.

Empiric thromboembolic risk classification of the 2022 CHEST guideline (Table 1). Major stroke risk factors for valve patients are atrial fibrillation, prior stroke or transient ischemic attack, prior valve thrombosis, rheumatic heart disease, hypertension, diabetes, heart failure, and age 75 years or older. The classification is consensus based; perioperative thrombotic event rates in trials were far lower than these annualized estimates suggest, because the period of interruption is short.

A prior thromboembolic event during an earlier interruption of anticoagulation places a patient in the high-risk group whatever the underlying indication. Left ventricular thrombus within 3 months is treated in the same way in the 2024 AHA/ACC perioperative guideline.

Interruption schedules for direct oral anticoagulants: the PAUSE protocol

DOACs have predictable half-lives, so their interruption is timed by pharmacokinetics and procedural bleeding risk, without heparin bridging and without routine coagulation testing. PAUSE tested a standardized protocol in 3,007 patients with atrial fibrillation taking apixaban (1,257), dabigatran (668), or rivaroxaban (1,082); one third had a high-bleeding-risk procedure. The DOAC was omitted for 1 day before a low-bleeding-risk procedure and 2 days before a high-bleeding-risk procedure, with longer omission for dabigatran when creatinine clearance was below 50 mL/min. It was resumed 1 day after a low-bleeding-risk procedure and 2 to 3 days after a high-bleeding-risk procedure.

Thirty-day major bleeding was 1.35% with apixaban, 0.90% with dabigatran, and 1.85% with rivaroxaban; in high-bleeding-risk procedures it was 2.96% with apixaban and 2.95% with rivaroxaban. Arterial thromboembolism occurred in 0.16%, 0.60%, and 0.37%. Most patients had a residual DOAC level below 50 ng/mL at the time of surgery. The 2022 CHEST guideline adopts the PAUSE schedule for apixaban, rivaroxaban, edoxaban, and dabigatran, suggests against heparin bridging when a DOAC is interrupted, and suggests against routine preoperative DOAC level testing (all conditional recommendations). The 2023 AF guideline recommends interruption without bridging in most patients (Class 1) and resumption after hemostasis is secure (Class 2a).

PAUSE schedule as adopted by the 2022 CHEST guideline and the 2023 AF guideline (Table 18). No dose is taken on the day of the procedure. For lower clearance than listed, few data exist, and the AF guideline suggests holding for an additional 1 to 3 days, especially before high-bleeding-risk procedures. Neuraxial anesthesia follows the longer ASRA intervals.

After high-bleeding-risk surgery in patients at high risk of venous thromboembolism, prophylactic-dose low-molecular-weight heparin (for example, enoxaparin 40 mg daily) or unfractionated heparin is given for the first 48 to 72 hours, until the full-dose DOAC is resumed. A DOAC reaches full anticoagulant effect within hours, so resuming it is equivalent to resuming therapeutic heparin and is timed with the same caution. For patients unable to take oral drugs after surgery, therapeutic or prophylactic parenteral anticoagulation substitutes until oral intake returns.

Day 0 is the day of the procedure. Creatinine clearance = (140 − age) × weight ÷ (72 × serum creatinine), × 0.85 for women, with actual body weight. Schedules: PAUSE (CHEST 2022; 2023 AF guideline Table 18); warfarin, CHEST 2022; neuraxial, ASRA fifth edition (treatment doses).

Interruption of warfarin and the preoperative INR

Warfarin is stopped 5 days or more before an elective procedure (CHEST 2022, conditional recommendation), which allows the INR to fall below 1.5 in most patients. Longer interruption is needed in older patients, those taking low maintenance doses, and those with a higher target INR. Acenocoumarol requires 2 to 3 days and phenprocoumon 10 to 12 days. The INR is measured the day before surgery. When it remains above 1.5, the 2022 CHEST guideline suggests against routine preoperative vitamin K (conditional recommendation), citing uncertainty about dose and later resistance to warfarin. Low-dose oral vitamin K (1 mg) remains an option in selected patients, and the procedure may be deferred. Warfarin is resumed the evening of the procedure or the next day at the usual dose, since its anticoagulant effect takes several days to return and early resumption adds little bleeding.

For minimal-bleeding-risk procedures, warfarin is not interrupted. The CHEST guideline suggests continuing warfarin for dental procedures and recommends continuing it for pacemaker and defibrillator implantation (strong recommendation), with an INR below 3.0 on the day of surgery. Cataract surgery and minor skin procedures are performed on uninterrupted warfarin with the INR in the therapeutic range. The ACG-CAG guideline suggests continuing warfarin for elective endoscopy, with interruption reserved for procedures at high risk of bleeding. The full warfarin pharmacology, including the management of a supratherapeutic INR, is on the warfarin page.

When bridging with heparin is indicated: BRIDGE, PERIOP2, and guideline positions

Bridging replaces warfarin with a short-acting parenteral anticoagulant, usually low-molecular-weight heparin, during the days when the INR is below the therapeutic range. Its purpose is to prevent thromboembolism; its cost is bleeding, because therapeutic heparin given close to surgery overlaps with the period of greatest surgical bleeding risk.

In BRIDGE, 1,884 patients with atrial fibrillation whose warfarin was interrupted for an elective procedure were randomized to dalteparin 100 IU/kg twice daily or placebo. Treatment ran from 3 days before until 24 hours before the procedure and for 5 to 10 days afterward. Arterial thromboembolism occurred in 0.4% without bridging and 0.3% with bridging, meeting noninferiority, and major bleeding in 1.3% and 3.2%. The trial included few patients with CHADS₂ scores of 5 or 6 and excluded mechanical valves and recent stroke. In PERIOP2, 1,471 patients with atrial fibrillation or a mechanical valve all received preoperative dalteparin and were randomized after surgery to dalteparin or placebo. Major thromboembolism within 90 days occurred in 1.0% with dalteparin and 1.2% with placebo, and major bleeding in 1.3% and 2.0%, with consistent results in the valve subgroup. Patients with more than one mechanical valve, a caged-ball valve, or recent stroke were excluded.

Bridging decisions from the 2022 CHEST, 2023 AF, 2020 ACC/AHA valve, 2024 AHA/ACC perioperative, and ASH 2018 guidelines.

The difference between the valve guideline and CHEST for mechanical valves reflects the absence of randomized evidence in the highest-risk valve patients. In practice both documents converge on the same patients: bridging for a mechanical mitral valve, a caged-ball or tilting-disc valve, multiple mechanical valves, or recent thromboembolism, and no bridging for an uncomplicated bileaflet aortic valve. Valve-specific management is detailed on the prosthetic heart valves page.

Bridging regimen. When bridging is used, therapeutic low-molecular-weight heparin starts when the INR falls below the therapeutic range, usually 3 days before surgery. The regimens are enoxaparin 1 mg/kg twice daily or 1.5 mg/kg daily, or dalteparin 100 IU/kg twice daily or 200 IU/kg daily. The last preoperative dose is given on the morning of the day before surgery, at half the total daily dose, about 24 hours before the procedure. Intravenous unfractionated heparin, used in kidney failure, is stopped 4 to 6 hours before surgery. After surgery, therapeutic heparin resumes about 24 hours after a low-to-moderate-bleeding-risk procedure. After a high-bleeding-risk procedure, prophylactic doses are given for 24 to 72 hours and therapeutic doses resume 2 to 3 days after surgery. Warfarin restarts the evening of surgery or the next day, and heparin stops when the INR is therapeutic.

Neuraxial anesthesia and deep blocks in anticoagulated patients

Spinal epidural hematoma after neuraxial anesthesia, epidural catheter removal, or lumbar puncture is rare but causes permanent paraplegia unless decompressed within hours. Because its rarity precludes trials, the ASRA guidelines set conservative, consensus-based intervals. The fifth edition (2025) replaces the terms "prophylactic" and "therapeutic" with "low dose" and "high dose," applies the same intervals to deep plexus and deep peripheral blocks, and adds drug-level thresholds.

Selected intervals from the ASRA Pain Medicine evidence-based guidelines, fifth edition (2025), as summarized in published guideline digests. The full guideline gives additional recommendations for fondaparinux, combinations of drugs, and indwelling catheters.

The ASRA intervals for treatment-dose DOACs (72 hours) are longer than the PAUSE interval for high-bleeding-risk procedures (2 days off, about 60 to 68 hours between the last dose and surgery). In PAUSE, 2 days off left most patients having high-bleeding-risk procedures with residual levels below 50 ng/mL, and the CHEST guideline applies its 2-day schedule to neuraxial anesthesia. The ASRA interval remains the standard for neuraxial block, and a randomized comparison of the two intervals (PAUSE-2) is designed to settle the difference. Heparin-specific neuraxial guidance is also described on the parenteral anticoagulants page.

Anticoagulant management for specific procedures

Common procedures have their own evidence, and for several of them uninterrupted anticoagulation is safer than interruption.

Procedure-specific management. PEG, percutaneous endoscopic gastrostomy; ERCP, endoscopic retrograde cholangiopancreatography.

Cardiac device surgery. Pocket hematoma after pacemaker or defibrillator implantation prolongs hospitalization, interrupts anticoagulation, and raises the risk of device infection. In BRUISE CONTROL, among 681 patients taking warfarin with an annual thromboembolic risk of 5% or more, clinically significant hematoma occurred in 3.5% with continued warfarin and 16.0% with heparin bridging. The trial stopped early. In BRUISE CONTROL-2, among 662 patients with atrial fibrillation taking dabigatran, rivaroxaban, or apixaban, clinically significant hematoma occurred in 2.1% with continued and 2.1% with interrupted DOAC. The 2023 AF guideline recommends continued warfarin over bridging (Class 1) and rates continued or interrupted DOAC as reasonable (Class 2a).

Endoscopy. The ACG-CAG guideline suggests continuing warfarin for elective endoscopy and, if warfarin is held for a high-risk procedure, suggests against bridging except in patients with a mechanical valve. It suggests interrupting DOACs, with the interval set by the procedure's bleeding risk. It found the evidence insufficient to recommend same-day resumption over resumption 1 to 7 days later, so the timing follows the endoscopist's assessment of the resection site. Delayed bleeding after polypectomy can occur up to 2 weeks after the procedure, and the size and number of polyps resected determine the risk.

Cardiac catheterization. In elective angiography, a vitamin K antagonist is held until the INR is 2.0 or lower (lower thresholds are used for femoral access), and DOACs are held according to kidney function and access. In acute coronary syndromes, angiography proceeds on anticoagulation through radial access, and additional parenteral anticoagulation is given during PCI according to the anticoagulant level. The hold tables and procedural heparin dosing are given on the AF and PCI page.

Antiplatelet drugs around surgery in patients who also take anticoagulants

Aspirin taken for secondary prevention is usually continued through noncardiac surgery. In POISE-2, starting or continuing aspirin around noncardiac surgery did not reduce death or myocardial infarction at 30 days (7.0% versus 7.1%) and increased major bleeding (4.6% versus 3.8%). The trial excluded patients with recent coronary stents, and in the continuation stratum the excess bleeding was smaller. The 2022 CHEST guideline suggests continuing aspirin over interrupting it for elective noncardiac surgery (conditional recommendation), with interruption, when needed, for 7 days or less before surgery. Aspirin is stopped before procedures where small amounts of bleeding are catastrophic: intracranial, spinal canal, and posterior eye chamber surgery.

P2Y12 inhibitors are stopped 5 days (clopidogrel), 3 to 5 days (ticagrelor), and 7 days (prasugrel) before elective surgery and resumed within about 24 hours after surgery when hemostasis allows. In patients with coronary stents who still need dual antiplatelet therapy, elective surgery is delayed. The timing of elective surgery after stenting and bridging with cangrelor are covered on the dual antiplatelet therapy page. In a patient taking an anticoagulant and an antiplatelet drug, the two are interrupted on separate schedules. The antiplatelet drug is omitted more often when its indication is remote.

Emergency surgery and urgent procedures in anticoagulated patients

The approach to an unplanned procedure depends on how long it can wait.

Warfarin. When surgery can wait 12 to 24 hours or more, intravenous vitamin K 5 to 10 mg lowers the INR without exposing the patient to clotting factor concentrates. When surgery cannot wait, 4F-PCC is given with vitamin K; Kcentra is approved for urgent reversal before surgery or invasive procedures as well as for bleeding. In the randomized trial of urgent procedures, effective hemostasis was achieved in 90% with 4F-PCC and 75% with plasma. In patients with a mechanical valve, the 2020 ACC/AHA valve guideline rates 4F-PCC as reasonable for emergency surgery (Class 2a).

Dabigatran. Idarucizumab allows surgery within minutes. In RE-VERSE AD, the median time from idarucizumab to the urgent procedure was 1.6 hours, and hemostasis was normal in 93.4%. When the procedure can wait, the dabigatran effect is estimated from the time since the last dose and creatinine clearance, and a thrombin time or dilute thrombin time confirms clearance.

Factor Xa inhibitors. No agent is approved for reversal of apixaban, rivaroxaban, or edoxaban before surgery; andexanet was approved only for bleeding. Surgery is delayed when possible: 24 hours from the last dose for low-to-moderate-bleeding-risk procedures and 48 hours for high-bleeding-risk procedures in patients with normal kidney function, longer with impaired clearance. When a drug-calibrated anti-Xa level is available, a level above 30 ng/mL before a high-bleeding-risk procedure is the ISTH threshold for considering reversal, and below that level surgery proceeds. When surgery cannot wait and the level is high or unknown, 4F-PCC is used off-label, and surgical hemostasis is planned for continued anticoagulant effect.

Cardiac surgery. For emergency cardiac surgery on cardiopulmonary bypass, the high doses of heparin dominate intraoperative anticoagulation. Residual DOAC effect adds to postoperative bleeding and is managed after protamine with factor concentrates, guided by viscoelastic testing. Andexanet is avoided before bypass because of heparin resistance.

Thrombolysis for acute ischemic stroke. Intravenous thrombolysis is contraindicated in patients taking warfarin with an INR above 1.7 and in patients who took a DOAC within the previous 48 hours unless laboratory tests show no drug effect. Endovascular thrombectomy remains available. Idarucizumab before thrombolysis in dabigatran-treated patients has been reported in case series.

Trials that define periprocedural anticoagulant management

Randomized and prospective management studies of periprocedural antithrombotic therapy.

Reversal and periprocedural management in pregnancy

Low-molecular-weight heparin is the anticoagulant of choice in most pregnant patients, because it does not cross the placenta; warfarin is used in selected patients with mechanical valves, and DOACs are avoided. Protamine reverses unfractionated heparin fully and low-molecular-weight heparin partially in pregnancy as outside it. Vitamin K and 4F-PCC reverse warfarin in the mother; warfarin also anticoagulates the fetus, and fetal coagulopathy is a reason that warfarin is replaced by heparin before planned delivery. Idarucizumab and 4F-PCC have been used in pregnancy in case reports only.

Around delivery, therapeutic low-molecular-weight heparin is converted to unfractionated heparin or stopped so that neuraxial anesthesia is possible, following the ASRA intervals (12 hours after low-dose and 24 hours after high-dose low-molecular-weight heparin). Anticoagulation is restarted after delivery once hemostasis is secure, with prophylactic doses within hours and therapeutic doses later. Management of pregnancy with a mechanical valve is described on the prosthetic heart valves page.

Institutional protocols, stewardship, and availability of reversal agents

Reversal is time sensitive, uses expensive products with thrombotic risk, and is needed unpredictably in many parts of a hospital. Institutions manage it through protocols that specify the indications for each agent, fixed or weight-based doses, who may order them, the laboratory tests that accompany them, and how resumption of anticoagulation is planned. Anticoagulation stewardship programs, often led by pharmacists, review reversal orders to prevent use for nonmajor bleeding or after the drug has cleared, and they track thrombotic complications.

Availability shapes the protocol. 4F-PCC is the agent around which US reversal protocols are now built. Idarucizumab is widely, but not universally, stocked. Activated PCC is stocked mainly for hemophilia with inhibitors. Andexanet alfa is no longer available in the United States. Drug-calibrated anti-Xa and dilute thrombin time assays are not available with rapid turnaround in many hospitals, so protocols state how decisions are made without them. A protocol also specifies the INR, aPTT, or anti-Xa measurement after reversal and the timing of repeat imaging.

Differences in European guidance

European documents agree on the principles: classification by severity, 4F-PCC with vitamin K for warfarin, idarucizumab for dabigatran, interruption of DOACs without bridging, and bridging limited to very high thrombotic risk. The 2021 European Heart Rhythm Association (EHRA) practical guide times DOAC interruption in hours by creatinine clearance. For factor Xa inhibitors with clearance of 30 mL/min or more, the interval is 24 hours before low-risk and 48 hours before high-risk procedures, and it is longer for dabigatran and for lower clearance. The 2022 European Society of Cardiology noncardiac surgery guideline adopts similar intervals. The main difference is andexanet alfa, which remains licensed in Europe and is recommended there for life-threatening bleeding on apixaban or rivaroxaban, whereas US practice now relies on 4F-PCC.

Sources of error in anticoagulant bleeding and periprocedural management

• Reversing nonmajor bleeding. Reversal agents add thrombotic risk and cost; most nonmajor bleeding needs only local measures and a brief interruption.
• Reversing a drug that has already cleared. A DOAC taken 36 hours earlier with normal kidney function has little effect left; the time of the last dose and creatinine clearance are established first.
• Relying on PT and aPTT to exclude factor Xa inhibitor effect. A normal PT or aPTT does not exclude a clinically relevant apixaban level; a calibrated anti-Xa assay or the dosing history decides.
• Giving 4F-PCC without vitamin K. The INR rebounds within hours as the infused factor VII is cleared while warfarin persists.
• Using plasma when 4F-PCC is available. Plasma corrects the INR slowly and incompletely and causes volume overload; in INCH it was associated with hematoma expansion and death.
• Giving large-dose vitamin K when warfarin will soon be needed. Ten milligrams of vitamin K makes a patient with a mechanical valve resistant to warfarin for days.
• Transfusing platelets for antiplatelet-associated intracerebral hemorrhage without surgery. PATCH showed harm.
• Expecting protamine to reverse low-molecular-weight heparin or fondaparinux fully. Protamine neutralizes at most about 60% of anti-Xa activity of enoxaparin and none of fondaparinux.
• Bridging DOACs or bridging warfarin in uncomplicated atrial fibrillation. BRIDGE showed more bleeding without less thromboembolism; DOACs need no bridge.
• Interrupting warfarin and bridging for device implantation. Bridging raised pocket hematoma from 3.5% to 16.0% in BRUISE CONTROL.
• Applying the PAUSE interval to neuraxial anesthesia without regard to ASRA. The ASRA intervals for treatment-dose DOACs are longer.
• Using the dabigatran schedule for normal kidney function in a patient with clearance below 50 mL/min. Dabigatran clearance is about 80% renal, and its interruption doubles.
• Discontinuing anticoagulation permanently after a treatable gastrointestinal bleed. Resumption after hemostasis is associated with lower mortality and less thromboembolism.
• Resuming full-dose anticoagulation too early after high-bleeding-risk surgery. A DOAC is fully active within hours; therapeutic dosing waits 48 to 72 hours after high-risk procedures.
• Withholding anticoagulation because of falls alone. The stroke risk prevented almost always exceeds the risk of fall-related intracranial hemorrhage.
• Citing andexanet as the US standard for factor Xa inhibitor bleeding. It was withdrawn from the US market; 4F-PCC is used.

References

1. Tomaselli GF, Mahaffey KW, Cuker A, et al. 2020 ACC expert consensus decision pathway on management of bleeding in patients on oral anticoagulants: a report of the American College of Cardiology Solution Set Oversight Committee. J Am Coll Cardiol. 2020;76:594–622. doi:10.1016/j.jacc.2020.04.053
2. Douketis JD, Spyropoulos AC, Murad MH, et al. Perioperative management of antithrombotic therapy: an American College of Chest Physicians clinical practice guideline. Chest. 2022;162:e207–e243. doi:10.1016/j.chest.2022.07.025
3. Doherty JU, Gluckman TJ, Hucker WJ, et al. 2017 ACC expert consensus decision pathway for periprocedural management of anticoagulation in patients with nonvalvular atrial fibrillation: a report of the American College of Cardiology Clinical Expert Consensus Document Task Force. J Am Coll Cardiol. 2017;69:871–898. doi:10.1016/j.jacc.2016.11.024
4. Joglar JA, Chung MK, Armbruster AL, et al. 2023 ACC/AHA/ACCP/HRS guideline for the diagnosis and management of atrial fibrillation: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2024;149:e1–e156. doi:10.1161/CIR.0000000000001193 (corrections: Circulation. 2024;149:e167, e936, e1413)
5. Thompson A, Fleischmann KE, Smilowitz NR, et al. 2024 AHA/ACC/ACS/ASNC/HRS/SCA/SCCT/SCMR/SVM guideline for perioperative cardiovascular management for noncardiac surgery: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2024;150:e351–e442. doi:10.1161/CIR.0000000000001285 (correction: Circulation. 2024;150:e466)
6. Greenberg SM, Ziai WC, Cordonnier C, et al. 2022 guideline for the management of patients with spontaneous intracerebral hemorrhage: a guideline from the American Heart Association/American Stroke Association. Stroke. 2022;53:e282–e361. doi:10.1161/STR.0000000000000407
7. Frontera JA, Lewin JJ 3rd, Rabinstein AA, et al. Guideline for reversal of antithrombotics in intracranial hemorrhage: a statement for healthcare professionals from the Neurocritical Care Society and Society of Critical Care Medicine. Neurocrit Care. 2016;24:6–46. doi:10.1007/s12028-015-0222-x
8. Witt DM, Nieuwlaat R, Clark NP, et al. American Society of Hematology 2018 guidelines for management of venous thromboembolism: optimal management of anticoagulation therapy. Blood Adv. 2018;2:3257–3291. doi:10.1182/bloodadvances.2018024893
9. Kopp SL, Vandermeulen E, McBane RD, et al. Regional anesthesia in the patient receiving antithrombotic or thrombolytic therapy: American Society of Regional Anesthesia and Pain Medicine evidence-based guidelines (fifth edition). Reg Anesth Pain Med. 2025. doi:10.1136/rapm-2024-105766
10. Abraham NS, Barkun AN, Sauer BG, et al. American College of Gastroenterology–Canadian Association of Gastroenterology clinical practice guideline: management of anticoagulants and antiplatelets during acute gastrointestinal bleeding and the periendoscopic period. Am J Gastroenterol. 2022;117:542–558. doi:10.14309/ajg.0000000000001627
11. Levy JH, Ageno W, Chan NC, et al. When and how to use antidotes for the reversal of direct oral anticoagulants: guidance from the SSC of the ISTH. J Thromb Haemost. 2016;14:623–627. doi:10.1111/jth.13227
12. Schulman S, Kearon C; Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. 2005;3:692–694. doi:10.1111/j.1538-7836.2005.01204.x
13. Sarode R, Milling TJ Jr, Refaai MA, et al. Efficacy and safety of a 4-factor prothrombin complex concentrate in patients on vitamin K antagonists presenting with major bleeding: a randomized, plasma-controlled, phase IIIb study. Circulation. 2013;128:1234–1243. doi:10.1161/CIRCULATIONAHA.113.002283
14. Goldstein JN, Refaai MA, Milling TJ Jr, et al. Four-factor prothrombin complex concentrate versus plasma for rapid vitamin K antagonist reversal in patients needing urgent surgical or invasive interventions: a phase 3b, open-label, non-inferiority, randomised trial. Lancet. 2015;385:2077–2087. doi:10.1016/S0140-6736(14)61685-8
15. Steiner T, Poli S, Griebe M, et al. Fresh frozen plasma versus prothrombin complex concentrate in patients with intracranial haemorrhage related to vitamin K antagonists (INCH): a randomised trial. Lancet Neurol. 2016;15:566–573. doi:10.1016/S1474-4422(16)00110-1
16. Pollack CV Jr, Reilly PA, van Ryn J, et al. Idarucizumab for dabigatran reversal: full cohort analysis. N Engl J Med. 2017;377:431–441. doi:10.1056/NEJMoa1707278
17. Connolly SJ, Crowther M, Eikelboom JW, et al. Full study report of andexanet alfa for bleeding associated with factor Xa inhibitors. N Engl J Med. 2019;380:1326–1335. doi:10.1056/NEJMoa1814051
18. Connolly SJ, Sharma M, Cohen AT, et al. Andexanet for factor Xa inhibitor–associated acute intracerebral hemorrhage. N Engl J Med. 2024;390:1745–1755. doi:10.1056/NEJMoa2313040
19. Douketis JD, Spyropoulos AC, Kaatz S, et al. Perioperative bridging anticoagulation in patients with atrial fibrillation. N Engl J Med. 2015;373:823–833. doi:10.1056/NEJMoa1501035
20. Douketis JD, Spyropoulos AC, Duncan J, et al. Perioperative management of patients with atrial fibrillation receiving a direct oral anticoagulant. JAMA Intern Med. 2019;179:1469–1478. doi:10.1001/jamainternmed.2019.2431
21. Kovacs MJ, Wells PS, Anderson DR, et al. Postoperative low molecular weight heparin bridging treatment for patients at high risk of arterial thromboembolism (PERIOP2): double blind randomised controlled trial. BMJ. 2021;373:n1205. doi:10.1136/bmj.n1205
22. Birnie DH, Healey JS, Wells GA, et al. Pacemaker or defibrillator surgery without interruption of anticoagulation. N Engl J Med. 2013;368:2084–2093. doi:10.1056/NEJMoa1302946
23. Birnie DH, Healey JS, Wells GA, et al. Continued vs. interrupted direct oral anticoagulants at the time of device surgery, in patients with moderate to high risk of arterial thrombo-embolic events (BRUISE CONTROL-2). Eur Heart J. 2018;39:3973–3979. doi:10.1093/eurheartj/ehy413
24. Devereaux PJ, Mrkobrada M, Sessler DI, et al. Aspirin in patients undergoing noncardiac surgery. N Engl J Med. 2014;370:1494–1503. doi:10.1056/NEJMoa1401105
25. Baharoglu MI, Cordonnier C, Al-Shahi Salman R, et al. Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial. Lancet. 2016;387:2605–2613. doi:10.1016/S0140-6736(16)30392-0
26. HALT-IT Trial Collaborators. Effects of a high-dose 24-h infusion of tranexamic acid on death and thromboembolic events in patients with acute gastrointestinal bleeding (HALT-IT): an international randomised, double-blind, placebo-controlled trial. Lancet. 2020;395:1927–1936. doi:10.1016/S0140-6736(20)30848-5
27. Villanueva C, Colomo A, Bosch A, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med. 2013;368:11–21. doi:10.1056/NEJMoa1211801
28. Carson JL, Brooks MM, Hébert PC, et al. Restrictive or liberal transfusion strategy in myocardial infarction and anemia. N Engl J Med. 2023;389:2446–2456. doi:10.1056/NEJMoa2307983
29. Qureshi W, Mittal C, Patsias I, et al. Restarting anticoagulation and outcomes after major gastrointestinal bleeding in atrial fibrillation. Am J Cardiol. 2014;113:662–668. doi:10.1016/j.amjcard.2013.10.044
30. SoSTART Collaboration. Effects of oral anticoagulation for atrial fibrillation after spontaneous intracranial haemorrhage in the UK: a randomised, open-label, assessor-masked, pilot-phase, non-inferiority trial. Lancet Neurol. 2021;20:842–853. doi:10.1016/S1474-4422(21)00264-7
31. Schreuder FHBM, van Nieuwenhuizen KM, Hofmeijer J, et al. Apixaban versus no anticoagulation after anticoagulation-associated intracerebral haemorrhage in patients with atrial fibrillation in the Netherlands (APACHE-AF): a randomised, open-label, phase 2 trial. Lancet Neurol. 2021;20:907–916. doi:10.1016/S1474-4422(21)00298-2
32. Veltkamp R, Korompoki E, Harvey KH, et al. Direct oral anticoagulants versus no anticoagulation for the prevention of stroke in survivors of intracerebral haemorrhage with atrial fibrillation (PRESTIGE-AF): a multicentre, open-label, randomised, phase 3 trial. Lancet. 2025;405:927–936. doi:10.1016/S0140-6736(25)00333-2
33. Al-Shahi Salman R, Stephen J, Tierney JF, et al. Effects of oral anticoagulation in people with atrial fibrillation after spontaneous intracranial haemorrhage (COCROACH): prospective, individual participant data meta-analysis of randomised trials. Lancet Neurol. 2023;22:1140–1149. doi:10.1016/S1474-4422(23)00315-0
34. Man-Son-Hing M, Nichol G, Lau A, Laupacis A. Choosing antithrombotic therapy for elderly patients with atrial fibrillation who are at risk for falls. Arch Intern Med. 1999;159:677–685. doi:10.1001/archinte.159.7.677
35. O'Brien EC, Simon DN, Thomas LE, et al. The ORBIT bleeding score: a simple bedside score to assess bleeding risk in atrial fibrillation. Eur Heart J. 2015;36:3258–3264. doi:10.1093/eurheartj/ehv476
36. Steffel J, Collins R, Antz M, et al. 2021 European Heart Rhythm Association practical guide on the use of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation. Europace. 2021;23:1612–1676. doi:10.1093/europace/euab065
37. Halvorsen S, Mehilli J, Cassese S, et al. 2022 ESC guidelines on cardiovascular assessment and management of patients undergoing non-cardiac surgery. Eur Heart J. 2022;43:3826–3924. doi:10.1093/eurheartj/ehac270
38. CSL Behring. Kcentra (prothrombin complex concentrate, human) prescribing information. DailyMed, US National Library of Medicine. dailymed.nlm.nih.gov
39. Andexanet alfa pulled from US market. TCTMD. December 22, 2025. tctmd.com/news/andexanet-alfa-pulled-us-market
40. Milling TJ Jr, Refaai MA, Goldstein JN, et al. Thromboembolic events after vitamin K antagonist reversal with 4-factor prothrombin complex concentrate: exploratory analyses of two randomized, plasma-controlled studies. Ann Emerg Med. 2016;67:96–105.e5. doi:10.1016/j.annemergmed.2015.04.036