Systemic anticoagulation is widely used to treat patients with or at risk for thromboembolic events. Vitamin K antagonists (VKAs) such as warfarin have been the standard therapy for long-term systemic anticoagulation since the mid-20th century. The discovery of the heparins followed shortly thereafter. Most recently, the development of direct clotting factor antagonists has radically changed the landscape of modern anticoagulation therapy. Although these new anticoagulants boast convenient dosing regimens, their use can result in potentially disastrous bleeding complications. While reversal agents are available for VKAs and certain heparins, no antidotes exist to rapidly reverse anticoagulation with the new direct factor antagonists.
Medical providers and patients must weigh the risks and benefits of systemic anticoagulation. Both spontaneous and traumatic bleeding are the most common and consequential complications of all anticoagulants. Intracranial hemorrhages and bleeding at noncompressible sites, such as the gastrointestinal tract, are examples of life-threatening bleeding events; smaller bleeds are also common and can occur at almost any location.
For the patient who presents with significant blood loss, restoration and maintenance of effective circulating volume is a priority. Although crystalloid infusion and blood transfusion with packed red blood cells (pRBCs) will replace volume, neither reverses medication-induced anticoagulation. Instead, both interventions potentially worsen coagulopathy by producing hypocalcemia from citrate toxicity, dilutional thrombocytopenia, and dilution of existing clotting factors. The use of targeted antidotal therapy depends on the particular anticoagulant involved. Operative or definitive management for the bleeding complications may be necessary in select cases (e.g., drainage of epidural hematomas).
VITAMIN K ANTAGONISTS
Warfarin is the most commonly prescribed oral anticoagulant. It inhibits vitamin K 2,3-epoxide reductase and vitamin K quinone reductase, causing anticoagulation from a depletion of activated factors II, VII, IX, and X.1,2 In addition to inhibition of these procoagulant clotting factors, VKAs also inhibit anticoagulant factors C and S, which can result in a transient procoagulant state at the initiation of VKA therapy.2
History and Physical Exam
Adverse drug events may occur even in VKA-anticoagulated patients who are within therapeutic ranges of anticoagulation. There are many factors that increase the risk for major bleeding in patients treated with the VKAs, including age, comorbid medical problems, labile international normalized ratios (INRs), concomitant ethanol or drug use, and genetic factors.3,4 Labile INRs are common in patients anticoagulated with the VKAs because of numerous dietary and drug interactions.2,5
Bleeding is the most prevalent complication from VKA use, but the emergency physician should also be familiar with a variety of associated nonhemorrhagic complications. Warfarin skin necrosis and purple toe syndrome are uncommon, and providers may misidentify these adverse drug reactions. Warfarin skin necrosis is believed to occur more frequently in patients with deficiencies of protein C, protein S, or antithrombin (AT) III.6–8 It is thought dermal thrombosis may lead to ischemia, although the mechanism of action is poorly understood.7,9 Patients typically develop painful and erythematous skin in areas with higher amounts of subcutaneous fat, such as the breasts, abdomen, thighs, and buttocks, within a week of warfarin initiation. These areas progress to necrotic patches that can extend up to 5 cm deep into the tissue.8,10 In addition, secondary infection is frequently reported as an additional source of morbidity.11
Purple toe syndrome is an embolic phenomenon that typically occurs 3 to 8 weeks after initiating treatment with warfarin. It is caused by VKA-induced bleeding into atherosclerotic plaques, with subsequent release of cholesterol emboli.11
Prothrombin time (PT) and INR are widely available and inexpensive laboratory tests used to determine warfarin’s anticoagulant effect. The PT measures the extrinsic pathway of the coagulation cascade. Because the PT varies due to individual laboratory and reagent variability, the INR is used to standardize results. Each lab calculates an INR based on a PT ratio raised to the lab’s unique international sensitivity factor.12,13 Target INRs typically range between 2 and 3.5, depending on indication for anticoagulation with warfarin.
The PT and INR, however, can be elevated in conditions not specific to warfarin-induced anticoagulation. Hepatic failure, inhibitors to clotting factors, disseminated intravascular coagulation, and a number of other conditions may cause a prolongation of PT and INR. This prolongation does not necessarily reflect anticoagulation.
A mixing study can assist in differentiating between a factor deficiency (such as that created by VKAs) and factor inhibitors (such as heparin). After combining equal volumes of pooled normal plasma and the patient’s plasma, failure to correct PT and INR confirms the presence of a clotting factor inhibitor.14
Reversal of VKA-induced coagulopathy can be achieved by restoration of activated clotting factors II, VII, IX, and X. In patients requiring rapid reversal, such as those with life-threatening bleeding, immediate factor replacement with fresh frozen plasma (FFP) or prothrombin complex concentrates (PCC) rapidly reverses coagulopathy. The American College of Chest Physicians (ACCP) 2012 guideline recommends 4-factor PCC for rapid reversal of VKA-induced coagulopathy in patients with bleeding.14 The 2012 version recommends PCC over FFP, though there are no large randomized controlled trials that directly compare these reversal strategies. The guideline is based on small, nonblinded, and unevenly matched studies.15,16 Theoretical advantages for PCC administration include smaller infusion volumes, which decrease time of administration and mitigate risks for volume overload. PCC also obviates the need for cross-matching of blood types, circumvents transfusion reactions, and decreases the risk of viral transmission.17 However, FFP is more frequently administered to reverse VKA-induced coagulopathy because of lower costs and because 4-factor PCC was unavailable in the United States until its approval in April 2013. Both 3- and 4-factor PCC contain non-activated factors II, VII, IX, and X. However, 3-factor PCC contains reduced amounts of factor VII to decrease thrombogenesis. Some have advocated the use of recombinant activated factor VII (rFVIIa), which is indicated only in patients with hemophilia or inhibitors to factors VIII or IX, for reversal of VKA-induced coagulopathy. However, the latest ACCP guideline explicitly recommends against using rFVIIa for this purpose. The risk of thrombosis following off-label rFVIIa administration may be higher in patients without hemophilia or inhibitors to factor VIII or IX.18
Because clotting factors have a definitive half-life, an essential step in reversing VKA-induced coagulopathy is the administration of vitamin K1 to promote reactivation of inactive vitamin K–dependent clotting factors. Oral administration results in peak plasma concentrations in 3 to 6 hours, whereas intravenous administration results in immediate peak plasma concentrations.19 Improvement in INR may lag, as clotting factor activation via hepatic gamma–glutamyl carboxylase is the rate-limiting step. In a single study of excessively anticoagulated patients, intravenous administration of vitamin K1 resulted in return to target INR at 6 hours, while oral administration required 12 hours.20 Maintaining a normal INR depends on the half-life of vitamin K1, plasma concentration of vitamin K1, and the duration of action of the specific VKA. Long-acting VKAs can require repeated doses of vitamin K1 to prevent excessive anticoagulation.
For adult patients with life-threatening bleeding, 10 mg of vitamin K1 should be administered intravenously if the benefits of INR normalization outweigh the risks. The rate of infusion should not exceed 1 mg/min, as intravenous administration is associated with anaphylactoid reaction. In nonbleeding patients with an INR >10, small doses of vitamin K1 (1 to 2.5 mg) should be administered orally. Patients with supratherapeutic INRs <10 can omit their next warfarin dose and recheck their INR if bleeding is not an issue.14 Subcutaneous administration of vitamin K1 results in unpredictable absorption kinetics and should be saved for rare occasions when patients are unable to take oral medications.
Patients who develop warfarin skin necrosis should discontinue warfarin therapy, and heparin should be used for systemic anticoagulation. Surgical debridement and amputation of limbs have been reported in severe cases.11
Unfractionated heparin is a mixture of glycosaminoglycans that causes a conformational change in AT, increasing its activity. This inhibits both thrombin and a number of clotting factors, including factors IXa, Xa, XIa, and XIIa.21 Low molecular weight heparins (LMWHs) are short fragments derived from unfractionated heparin. LMWHs cause distinct conformational changes in AT, which targets inhibition specifically at factor Xa.22 LMWHs have several potential clinical advantages compared to unfractionated heparin, including longer half-lives and the ability to use fixed-dosing regimens.
Complications of heparin therapy fall into two categories: bleeding complications (expected from any of the anticoagulants) and nonbleeding complications including heparin-induced thrombocytopenia (HIT) and heparin-induced thrombocytopenia and thrombosis syndrome (HIT(T)). It is important to distinguish between postoperative consumptive thrombocytopenia and HIT(T) derived from heparin-induced complications. Postoperative thrombocytopenia usually occurs on postoperative day 1 or 2, followed by an improvement 1 or 2 days later. HIT(T) usually manifests between 5 and 10 days after introduction of heparin. However, an earlier fall in platelet count may occur in patients previously treated with heparin, and this early appearance of thrombocytopenia may confuse providers.23
HIT(T) develops as a result of antibodies that recognize a heparin–platelet factor 4 complex. When these antibody–antigen complexes bind to platelets, the consequences are either platelet destruction or platelet activation. HIT occurs when platelets are destroyed without any thrombotic sequelae. HITT occurs when platelets are activated and cause thrombosis. If HIT goes untreated, up to 55% of patients will develop HITT. HIT(T) can also occur with the LMWHs.23,24
Activated partial thromboplastin time (aPTT) is the test of choice for monitoring unfractionated heparin’s anticoagulant effect. Nomograms can help providers alter heparin dosing based on aPTT results. In patients requiring high-dose heparin, such as those undergoing cardiovascular procedures, activated clotting time can be used instead of aPTT.25 A subset of patients may manifest heparin resistance, where high doses of heparin cannot achieve aPTTs in the therapeutic range. In these cases, anti-Xa levels may be measured instead of aPTT, which permits lower dosing of heparin while providing similar therapeutic effect and safety profiles.26
Patients treated with LMWHs typically do not receive laboratory monitoring.27 VTE prophylaxis dosing is fixed, while VTE treatment is weight based. Laboratory monitoring with anti-Xa activity measurements should be performed in the case of pregnant patients, obese patients, or those with chronic kidney disease. Therapeutic anti-Xa activity ranges in these populations vary based on indication for anticoagulation.25
HIT(T) must be suspected if the platelet count drops below 100 × 109/L or if there is a 40% drop in the platelet count after heparin initiation.28 Patients who are at highest risk of developing HIT(T) are postoperative patients on either prophylactic or therapeutic heparin. The incidence of HIT(T) in these patients is between 1% and 5%. Cardiac surgery patients also have a higher risk of developing HIT(T) with an incidence between 1 and 3%.23 In patients with a risk of >1% for developing HIT(T), platelet counts should be monitored every 2 to 3 days starting on day 4 of heparin therapy. If a patient does not develop HIT(T) by day 14 of therapy, then further monitoring is not necessary.23 HIT antibody assays should be sent to confirm the diagnosis of HIT(T) if the platelet counts drop by the aforementioned amount.
For nonsignificant bleeding with an elevated aPTT, cessation of anticoagulation therapy may be sufficient, as unfractionated heparin has a short duration of action between 1 and 2.5 hours, depending on amount administered.29,30 Significant bleeding, by contrast, may necessitate antidotal therapy. Protamine sulfate binds to heparin and effectively neutralizes its anticoagulant capabilities. One milligram of intravenous protamine neutralizes 100 U of unfractionated heparin, and dosing should reflect the amount of heparin calculated to be present at the time of antidote administration (assume that the half-life of heparin is 60 to 90 minutes).29 Adverse effects of protamine are numerous. Paradoxical anticoagulation can occur if excessive protamine is administered. Hypotension and bradycardia can occur with correct dosing, but slow infusion helps decrease the risk of these events. Anaphylaxis is also possible, and patients with a history of receiving protamine, fish allergy, or history of vasectomy have a higher risk of developing anaphylaxis to protamine. Because of the risk of anaphylaxis, only patients with life-threatening bleeding should receive protamine.25
Protamine is sometimes used to treat patients anticoagulated with LMWHs if they suffer from life-threatening bleeding. There are no proven antidotes for LMWH, but partial reversal may be possible with protamine. One milligram of protamine should be administered per 100 anti–factor Xa units (1 mg enoxaparin is the equivalent to 100 anti–factor Xa units) if LMWH was administered within 8 hours. If bleeding continues, a second dose of protamine at 0.5 mg per 100 anti–factor Xa units can be considered. Smaller doses of protamine should be administered if more than 8 hours has elapsed since LMWH administration.25
If HIT(T) is suspected or confirmed in a patient, the most important intervention is cessation of heparin or LMWH therapy. Alternative anticoagulants such as the direct thrombin inhibitors (DTI) argatroban and bivalirudin, or a Xa inhibitor such as danaparoid should be used until a therapeutic INR is achieved with warfarin. Novel oral anticoagulants, such as dabigatran, rivaroxaban, and apixaban, have not been studied for this indication.
DIRECT THROMBIN INHIBITORS
To circumvent the problems associated with the VKAs and the heparins, DTIs have been utilized parenterally, and more recent developments have led to the introduction of oral DTIs. These medications are derived from hirudin, a peptide secreted by the medicinal leech, and they are used to treat acute coronary syndrome and VTE.31 The most commonly used parenterally administered DTIs are bivalirudin and argatroban. Providers typically use these agents when patients have a contraindication to heparin, such as HIT(T).
Dabigatran, an oral DTI, is currently approved only for VTE and stroke prophylaxis in patients with nonvalvular atrial fibrillation. Initial studies showed favorable results for VTE prophylaxis in patients with nonvalvular atrial fibrillation with lower bleeding and mortality rates.32 Unfortunately, a higher-than-anticipated bleeding rate associated with dabigatran has been identified in postmarketing analyses and studies.33,34 A U.S. FDA advisory statement cited dabigatran as the leader in reported adverse drug events in 2012.35 However, further data in this advisory suggest that risk factors for bleeding include inappropriate renal dosing, age older than 75 years,33 and use for the wrong indication.34 In addition, the risk of myocardial infarctions and acute coronary syndrome appeared to be higher in groups treated with dabigatran when compared to patients treated with other anticoagulants.36
To estimate the degree of anticoagulation with the DTIs bivalirudin and argatroban, serial aPTT measurements are used most often.25 Unfortunately, estimates made with aPTT are frequently inaccurate, since the relationship between aPTT and degree of anticoagulation with DTIs is not linear. With dabigatran, the aPTT plateaus when concentrations are over 200 ng/mL, and PT and INR also follow a nonlinear relationship with the serum dabigatran concentration.37 The use of thrombin time (TT) and ecarin clotting time (ECT), which follow a more linear relationship to DTI concentration, has been proposed as better tests to estimate DTI concentration and anticoagulation.25,37 With the increased use of DTIs, some laboratories have increased the availability of the TT; ECT still remains largely unavailable for use in real time.
The most concerning issue with the DTIs is the absence of a proven reversal agent or strategy. Because the half-lives of the parenteral medications are relatively short, the most critical action is to stop the infusion or administration of the DTI when bleeding is suspected. A single case report describes the use of FFP in a patient who received a 13-fold overdose of argatroban. He was treated with FFP for a prolonged aPTT, which did not normalize. Fortunately, the patient did not suffer any bleeding consequences.38 Orally administered dabigatran has a half-life of 8 to 12 hours, whereas the parenterally administered DTIs have half-lives of approximately 0.5 to 2 hours.39 Therefore, bleeding while anticoagulated with dabigatran can be particularly difficult to manage. The manufacturers of dabigatran suggest using supportive care and transfusion of pRBCs and FFP. They also mention that rfVIIa, PCC, and hemodialysis may be considered, although there are no prospective, randomized controlled human studies that show better outcomes with any of these interventions. In a murine study, mice with intracranial bleeds were given increasing doses of 4-factor PCC, which resulted in a dose-dependent response in minimizing hematoma expansion; the mice given PCC at doses of 100 U/kg showed the best response, however, in none of the mice were bleeding times normalized. In the same study, neither rfVIIa nor FFP reduced hematoma expansion.40 Healthy human volunteers anticoagulated with 2.5 days of dabigatran continued to have abnormal coagulation studies despite treatment with 50 U/kg of 4-factor PCC.41
Hemodialysis has been proposed as a possible lifesaving intervention in bleeding patients who are anticoagulated with dabigatran. A single study shows that hemodialysis affords extraction ratios of 62% to 68% at hours 2 and 4 during hemodialysis, respectively, but this study was performed in dialysis-dependent patients who were given a single dose of dabigatran.42 Unfortunately, further pharmacokinetic characterization published in multiple case reports demonstrates that significant drug rebound, up to 87%, following hemodialysis may limit the effectiveness of hemodialysis. Furthermore, while hemodialysis may decrease serum dabigatran concentrations, it does not necessarily normalize aPTT or TT.43,44 In addition, providers may be reluctant to place a large-bore hemodialysis catheter in an excessively anticoagulated patient. Despite aggressive intervention in some of these patients, including massive transfusion and hemodialysis, deaths from exsanguination in dabigatran-anticoagulated patients may occur.43,45
A monoclonal antibody that neutralizes dabigatran is currently undergoing evaluation for use in patients requiring rapid reversal of anticoagulation.46,47 Until it or another antidote is approved, providers must make use of imperfectly effective options. From the limited available data, the best choice may be to attempt reversal with PCC in aliquots of 25 U/kg up to a maximum of 100 U/kg. The risk of thrombosis remains and should be weighed against the benefits of treatment in a bleeding patient. 4-factor PCC was utilized in studies that evaluated PCC’s ability to reverse dabigatran-induced anticoagulation and is, therefore, preferred over 3-factor PCC if available.40,41 Hemodialysis may be helpful, particularly in patients with suspected supratherapeutic dabigatran concentrations, assuming that the degree of anticoagulation is directly related to dabigatran concentration. Initiation of hemodialysis should not delay required definitive treatments such as operative intervention to control bleeding.
DIRECT FACTOR XA INHIBITORS
Rivaroxaban and apixaban are orally active, direct factor Xa inhibitors. They are approved for VTE prophylaxis in patients with atrial fibrillation. Rivaroxaban holds an additional indication for VTE prophylaxis in some postsurgical patients. The factor Xa inhibitors may be safer than the DTIs because they prevent thrombin activation upstream from thrombin itself.48,49 This indirect inactivation allows downstream administration of clotting factors in bleeding patients. While rivaroxaban is both renally and hepatically eliminated, a large portion of apixaban elimination is fecal.50,51
Rivaroxaban inhibits factor Xa activity and prolongs PT and PTT via a dose-dependent relationship.52–56 The HepTest is a nonapproved assay that measures anti-Xa and anti-IIa activity. It is not widely available, and it has not yet gone through the FDA approval process. Studies show that this assay correlates well with the anticoagulant effect of apixaban.52,53
Similar to the DTIs, the factor Xa inhibitors have no definitive antidotes. In healthy human volunteers given rivaroxaban for 2.5 days, PT and endogenous thrombin potential, a thrombin generation assay, normalized after 50 U/kg of 4-factor PCC.41 In a rabbit study, bleeding animals continued to bleed despite treatment with 4-factor PCC and rfVIIa, while a battery of coagulation parameters such as bleeding time, aPTT, anti-Xa activity, and clotting time improved.57 There are currently no studies that evaluate clinical outcomes associated with antidotal treatment in apixaban-treated patients. Unlike dabigatran, rivaroxaban and apixaban are not amenable to hemodialysis because they are highly protein bound.58
Because the anticoagulation effect of these factor Xa inhibitors can last more than 24 hours under certain circumstances, specific interventions to reverse anticoagulation may be necessary if supportive measures with volume resuscitation are not helpful. While PCC may help improve coagulation parameters, there are no randomized controlled outcome studies evaluating this intervention. Under life-threatening circumstances, it is reasonable to administer 4-factor PCC in doses of 25 U/kg to patients anticoagulated with direct factor Xa inhibitors. If necessary, repeated doses up to a total of 100 U/kg may be considered, with the knowledge that an unknown risk of thrombosis is present. If 4-factor PCC is not available, 3-factor PCC may be substituted, but its efficacy is not well studied.