Chapter 34: Anticoagulation

Anticoagulation is the mainstay of therapy for VTE and is a key therapeutic component for a number of other clinical settings including atrial fibrillation (AF), mechanical heart valves, and idiopathic pulmonary arterial hypertension (IPAH). In the critical care setting, the use of anticoagulation is frequently warranted, but is often associated with an increased risk for complications based on underlying disease states, necessary invasive procedures, trauma, thrombocytopenia, coagulopathy, renal failure and hepatic failure, and the potential for major surgery. Reversing anticoagulation must be considered in certain clinical scenarios. The focus of this chapter will be anticoagulation in the intensive care unit (ICU), with a particular focus on therapeutic anticoagulation.

Anticoagulation in the critically ill patient requires therapeutic levels when treating an acute condition such as acute VTE. Given the high risk of morbidity and mortality caused by VTE, patients in the ICU, almost without exception, require VTE prophylaxis. Mechanical prophylaxis is used when pharmacologic prophylaxis is contraindicated or in certain lower-risk settings.

The ideal anticoagulant would be effective, easily administered with a predictable anticoagulant effect (oral, if the patient is capable) and rapid in onset, require no monitoring, and be easily reversible. Ideally, it would cause no significant adverse effects, including thrombocytopenia, and would not cause bleeding. The latter is a virtual impossibility, based on the clinical effect required. The direct oral anticoagulants (DOACs) are a significant advance, reminiscent of how LMWH dramatically changed the anticoagulation world nearly 2 decades ago.

Standard, Unfractionated Heparin

In spite of its limitations, intravenous (IV) standard UFH remains the most commonly utilized parenteral therapy when therapeutic doses of anticoagulants are required in the ICU. Monitoring with the aPTT is required. Heparin is a “promiscuous” molecule. It binds almost indiscriminately to various plasma proteins, monocytes, and endothelial cells. Because it binds to a number of circulating proteins and cell types, different clinical conditions can significantly affect the heparin levels. The response can be thus, difficult to predict. Weight-basing heparin is crucial to hasten the achievement of therapeutic levels but monitoring is still essential (Figure 34–1). The aPTT is a clot-based, in vitro assay using citrated, platelet-depleted plasma. While the aPTT is sensitive to the inactivation of thrombin and factor Xa by UFH, it is a nonspecific assay. The therapeutic range should be a heparin concentration of 0.3 to 0.7 anti-Xa units/mL, using plasma samples from patients being treated with UFH for VTE. This aPTT range is specific for the reagent used by a given manufacturer lot. Alternatively, heparin levels (anti–factor Xa assay) can be monitored.

Decades ago, IV UFH dosing was routinely a 5000 unit bolus followed by an infusion of 1000 units/h followed by aPTT monitoring. In 1993, Raschke and colleagues¹ demonstrated that dosing heparin based on actual body weight (bolus of 80 units/kg followed by 18 units/kg/h) achieved a therapeutic aPTT at 24 hours more often than the standard dosing regimen. Furthermore, there was no difference between the 2 regimens with regard to bleeding events, although the weight-based subjects had a higher percentage of supratherapeutic aPTTs. The rate of recurrent VTE was statistically significantly higher in the non-weight-based regimen.

In this study, only 9 patients in the weight-based arm weighed more than 100 kg, and the largest patient was 131 kg making it difficult to extrapolate these findings to obese patients in general.¹ The volume of distribution of heparin is about the same as blood volume. Blood volume is increased in obesity, but adipose tissue has less blood volume compared with lean body tissue. This suggests that using actual body weight for the morbidly obese could increase the incidence of supratherapeutic aPTTs as well as the bleeding risk.

Several studies and reviews have attempted to address the dilemma of heparinization in the obese. One concluded that actual body weight in both nonobese and obese patients should be used to calculate heparin infusion rates, as long as a dosing cap of 10,000 units and infusion rate limit of 15,000 units/h was utilized.² A smaller study also found that actual body weight in both nonobese and obese patients was appropriate.³ Still another study concluded that morbidly obese patients dosed according to actual body weight on a weight-based heparin nomogram were more likely to experience supratherapeutic aPTT values compared to nonmorbidly obese patients.⁴ They suggested a dosing cap but believed more data were needed. In summary, the available evidence supports the use of actual body weight for calculating heparin bolus doses and initial infusion rates for obese patients. Experience with morbidly obese patients is still limited. The concept of a dose cap remains controversial. Careful monitoring remains crucial.

Resistance to heparin may be encountered. Approximately 25% of patients with acute VTE on IV standard UFH require more than 35,000 units of heparin per day to achieve a therapeutic aPTT. Resistance may occur because of the nonspecific binding of the drug to various plasma proteins, altered intravascular volume, and/or because of increased heparin clearance.⁵ Other reasons for alterations in the aPTT dose-response include increased concentrations of clotting factors, including thrombin or factor VIII, or a reduction in anticoagulation factors such as antithrombin which may occur in the critically ill setting. In the case of apparent heparin resistance, the anti-Xa level should be utilized.⁶ This may help differentiate pharmacokinetic from biochemical heparin resistance in patients requiring high infusion rates.

Another option is to simply monitor anti-Xa levels in any ICU patient on heparin since such patients appear more susceptible to variable aPTTs.⁷ No data thus far, however, have proven better outcomes utilizing this method. In any patient found to have prolonged baseline clotting times, the heparin level by anti-Xa assay should be determined prior to initiating anticoagulant therapy.

Low-Molecular-Weight Heparins

LMWHs have distinct advantages over standard UFH as well as potential disadvantages. They are much more bioavailable than standard heparin, and are thus, more predictable. They are at least partially reversible with protamine. They appear to cause less thrombocytopenia than standard UFH. They can be given by the subcutaneous route (SC) and in most settings do not require monitoring. However, in certain populations, particularly the morbidly obese, those with renal insufficiency, and pregnant patients, these drugs may be less predictable. A creatinine clearance less than 30 mL/min requires a dose reduction or discontinuation.

While in most prophylactic and therapeutic anticoagulation settings, no monitoring of LMWH is necessary, critically ill patients who have marked metabolic derangements may be less predictable. A number of factors including hypotension, particularly requiring vasopressor therapy, renal insufficiency, and variable absorption may affect drug efficacy. Furthermore, such conditions may rapidly change in severity, making prediction of appropriate levels more difficult.⁸ In such scenarios, when therapeutic anticoagulation is required, the goal should be a chromogenic anti-Xa assay level of 0.5 to 1.0 units/mL 4 hours after SC dosing, if the patient is on an every 12-hour dosing regimen. A higher level of 2 units/mL is recommended for a once-daily regimen.⁹ Trough or even random anti-Xa levels may also be useful when there are concerns, for example, that a LMWH may be accumulating in renal failure. The anti-Xa assay should be calibrated to the specific LMWH being utilized. Finally, while levels are generally not checked in patients on deep venous thrombosis (DVT) prophylaxis, a level of approximately 0.1 to 0.3 units/mL measured 3 to 4 hours after a SC injection would appear reasonable.¹⁰ Despite some limitations of antifactor Xa monitoring for LMWH therapy, the unreliability of SC drug administration in critically ill patients warrants the routine use of monitoring to assure adequate drug exposure.

Fondaparinux

Fondaparinux, a selective factor Xa inhibitor, has a longer half-life than the LMWHs and thus, is less useful in the ICU. For prophylaxis in total joint arthroplasty, it is prescribed as 2.5 mg SC once-daily and is as effective as LMWH. It does not appear to be associated with HIT.

Argatroban

Argatroban, a small molecule synthetic direct thrombin inhibitor (DTI) administered IV, is Food and Drug Administration (FDA) approved for prevention and treatment of thrombosis in patients with HIT. Plasma concentrations reach steady state in 1 to 3 hours. This drug is hepatically metabolized which must be considered. However, this also allows its use in renal insufficiency. Its half-life is about 50 minutes. The recommended range for the aPTT to monitor argatroban is 1.5 to 3 times control, not to exceed 100 seconds. The ecarin clotting time may have certain advantages in monitoring DTI therapy, but it is not routinely available. It has not yet been adequately evaluated to be used in place of the aPTT. Lepirudin, a recombinant hirudin, is no longer available.

Warfarin

Warfarin is impractical in the critical care setting. It is very long acting and associated with innumerable drug interactions. Although reversible with vitamin K, it is not immediately reversible. It is still used by some orthopedic surgeons as prophylaxis for total knee or hip arthroplasty is used less commonly since the advent of LMWH and DOACs (see “Anticoagulant Reversal”).

Aspirin

Aspirin has been studied for long-term secondary prevention of DVT/pulmonary embolism (PE). In this setting, it appears more effective than placebo but less effective than all other available anticoagulants drugs. While it is used in the setting of acute coronary syndromes and AF, it has no real role in the ICU patient with acute VTE.

Direct Oral Anticoagulants

DOACs inhibit either thrombin (factor IIa) or factor Xa. Thrombin is a serine protease and activates factors V, VIII, and XI and catalyzes the conversion of fibrinogen to fibrin, as well as stimulating platelet aggregation. Factor Xa is also a serine protease that represents the intersection of the extrinsic and the intrinsic coagulation pathways. It catalyzes the conversion of prothrombin to thrombin (see Figure 34–1).

Four DOACs are approved for use in the United States, including the Xa inhibitors rivaroxaban (Xarelto) and apixaban (Eliquis), and the direct thrombin inhibitor, dabigatran etexilate (Pradaxa). All 4 are approved for stroke prevention in nonvalvular AF, as well as for treatment of established DVT and/or PE. Rivaroxaban and apixaban are also approved for prophylaxis for total hip and knee arthroplasty (Table 34–1). Others, including betrixaban could follow. The safety and efficacy of these drugs have not been studied for mechanical prosthetic heart valves and thus are not recommended in this setting.

A major advantage of these DOACs includes the lack of need for monitoring although this could be useful in emergencies. There is a marked reduction in drug interactions compared with warfarin, although cytochrome P450 (CYP) metabolism is important for factor Xa inhibitor metabolism and P-glycoprotein (P-gp) metabolism affects the metabolism of all 4 of these agents. Renal metabolism also must be considered. Dosing for dabigatran, rivaroxaban, and apixaban, for treatment of established VTE and AF is outlined in Table 34–2 and the potential for drug interactions is further outlined in Table 34–3.

Large, randomized clinical trials in thousands of patients with AF and acute DVT and/or PE have been completed with these agents, indicating that they are at least as effective as warfarin. They appear to be at least as safe, and in some instances safer, with regard to bleeding.

Concomitant use of other drugs affecting hemostasis increases the risk of bleeding, including other anticoagulants, aspirin, other antiplatelet agents, thrombolytic agents, and nonsteroidal anti-inflammatory drugs.

Edoxaban

Edoxaban is the most recently approved direct factor Xa inhibitor in the United States. It is indicated to reduce risk of stroke and systemic embolism in patients with nonvalvular AF and for the treatment of DVT and PE following 5 to 10 days of initial therapy with a parenteral anticoagulant (references 1 and 2 below). For AF, the recommended dose is 60 mg PO once daily. It should not be prescribed in patients when the CrCL is > 95 mL/min because of an increased risk of stroke compared with warfarin. The dose must be reduced to 30 mg once daily in patients with CrCl of 15 to 50 mL/min. In the nonvalvular AF trial, major bleeding occurred with a significantly lower incidence with edoxaban compared with warfarin. The endpoint of death or ICH also occurred in significantly fewer patients receiving edoxaban. Importantly, fatal bleeding and life-threatening bleeding occurred significantly less often with edoxaban, as did gastrointestinal bleeding with the lower dose. In contrast, the higher edoxaban dose led to more gastrointestinal bleeding than warfarin. For DVT and PE, the recommended dose is 60 mg once daily. The dose should be similarly reduced to 30 mg once daily for patients with CrCL 15 to 50 mL/min, body weight ≤ 60 kg, or in patients on certain P-gp inhibitors.

Dabigatran

Thrombin is a logical target for an anticoagulant. When it is activated from prothrombin, it converts soluble fibrinogen to insoluble fibrin, activates coagulation factors V, VIII, and XI, thus, generating more thrombin, and activates platelets. Dabigatran etexilate is a synthetic, orally available prodrug that is rapidly absorbed and converted by esterases to its active form, dabigatran, a potent direct inhibitor of both free thrombin and clot-bound thrombin. This drug has a rapid onset of action, very few drug interactions, no reported food interactions, and does not require routine coagulation monitoring.

Dabigatran is 35% bound to plasma proteins and is renally excreted, with 80% of the drug entering the urine unchanged. The anticoagulant effect accumulates in the setting of renal insufficiency, and such bioaccumulation correlates well with the degree of renal dysfunction.¹¹ In cases of moderate hepatic impairment, dabigatran can be administered safely and no dose adjustment is necessary.¹²

P-gp inhibition and renal insufficiency are major independent factors in increased exposure to dabigatran. There is no significant cytochrome P450 metabolism. Concomitant use of P-gp inhibitors in patients with renal impairment is expected to produce increased exposure of dabigatran compared to that seen with either factor alone (see Table 34–2). Absorption of dabigatran etexilate is mediated by P-gp; dabigatran etexilate is a substrate for P-gp, but active dabigatran is not. Thus, P-gp inhibitors can increase dabigatran absorption, increasing both area under curve (AUC) and C_max and P-gp inducers can reduce its absorption, resulting in inadequate levels. Concomitant P-gp inducers (eg, rifampin) should not be used in patients on dabigatran (see Table 34–3).

This drug is FDA approved for stroke prevention in patients with nonvalvular AF and for acute VTE at a dose of 150 mg twice daily. It is not approved, at present, for prophylaxis following total hip or knee replacement in the United States.

For AF, dabigatran was shown to be superior to dose-adjusted warfarin with a similar rate of major bleeding.¹³ A dose of 75 mg twice daily, is approved in the United States for patients with nonvalvular AF and severe renal insufficiency; that is, creatinine clearance (CrCl) of 15 to 30 mL/min.

Over 9000 patients were evaluated in studies involving dabigatran leading to the approval of this drug for the treatment of acute VTE, including RE-COVER, RE-COVER II, RE-MEDY, and RE-SONATE.^14,15,16 The first 2 studies were initial VTE therapy studies utilizing a parenteral anticoagulant bridge. Dabigatran proved effective standard therapy, and while major bleeding rates were similar, there were fewer episodes of nonmajor bleeding with dabigatran than with warfarin. There was, however, a higher rate of gastrointestinal (GI) bleeding.^14,15 The extension studies in patients who had completed a course of anticoagulation demonstrated not surprisingly, that extended duration of anticoagulation with dabigatran was noninferior to warfarin but superior to placebo in reducing recurrent VTE. Based on the way these clinical trials were conducted, for acute VTE, the parenteral anticoagulant bridge is required for 5 to 10 days prior to oral dabigatran alone. Dabigatran is rated pregnancy category C.

Rivaroxaban

Factor Xa represents the rate-limiting factor in thrombin generation and amplification, generating the Xa complex that converts prothrombin to thrombin. The direct factor Xa inhibitors inhibit free factor Xa, factor Xa in the prothrombinase complex, and factor Xa found in clots, and this is independent of antithrombin. In contrast, LMWHs, UFH, and fondaparinux, are dependent on antithrombin to inhibit factor Xa.

Clearance of this drug is decreased to some extent in patients with renal impairment, but two-thirds of its primary mode of clearance is nonrenal. While approximately 67% of rivaroxaban is eliminated by the kidney, half of this is clearance of active drug and half of it is clearance of inactive rivaroxaban, which is not clinically important. CYP450, P-gp and breast cancer–related protein (BCRP) are all involved with metabolism. It is 93% protein bound and thus, not dialyzable. Rivaroxaban is a substrate of CYP3A4/5, CYP2J2, and the P-gp and ATP-binding cassette G2 (ABCG2) transporters. Inhibitors and inducers of these CYP450 enzymes or transporters, such as P-gp, may result in changes in rivaroxaban exposure.

Concomitant use of rivaroxaban with combined P-gp and strong CYP3A4 inhibitors (eg, ketoconazole, itraconazole, lopinavir/ritonavir, ritonavir, indinavir, and conivaptan) should be avoided. In addition, concomitant use of rivaroxaban with drugs that are combined P-gp and strong CYP3A4 inducers (eg, carbamazepine, phenytoin, rifampin, St. John's wort) should also be avoided (see Tables 34–2 and 34–3).

For stroke prevention, the randomized, double-blinded ROCKET AF trial found that rivaroxaban 20 mg daily (15 mg daily if CrCl is 15-50 mL/min) was noninferior to warfarin in efficacy, with no significant difference in major bleeding events.¹⁷ Due to high plasma protein binding (> 90%), rivaroxaban is not eliminated during hemodialysis.

The Einstein DVT and PE studies (> 8000 patients) led to FDA approval for rivaroxaban for the treatment of established DVT and/or PE.^18,19 These large prospective, randomized trials demonstrated noninferiority to warfarin with regard to recurrent VTE, and in the PE study, the risk of major bleeding was significantly lower. For acute VTE, the drug is dosed as 15 mg q12h for 3 weeks and then 20 mg once daily. Unlike in AF, where the dose can be reduced with a CrCl of 15 to 50 mL/min, this has not been recommended in VTE. However, in acute VTE with CrCl less than 30 mL/min, the drug should not be used (see Table 34–2).

Rivaroxaban is also approved in the United States for VTE prophylaxis after hip or knee replacement surgery based on superiority over enoxaparin, and at least comparable safety. At present, however, no NOAC is approved specifically for VTE prophylaxis in the ICU.

When changing from warfarin, rivaroxaban should be initiated when the INR is less than or equal to 3. Finally, tablets may be crushed and either mixed in applesauce or suspended in water and administered via an nasogastric (NG) tube to appropriate patients who have difficulty swallowing a whole tablet. This drug is pregnancy category B and should be used during pregnancy only if the potential benefit outweighs the potential risk to the mother and the fetus.

Apixaban

Like rivaroxaban, apixaban is FDA approved for stroke prevention in AF, acute VTE, and for prophylaxis in total knee and hip replacement. Apixaban is an oral, direct factor Xa inhibitor that is highly bioavailable (80%), is highly protein bound, and reaches peak plasma concentration within 2 to 3 hours after intake. It is 75% hepatically metabolized with the rest renally excreted. Dosing including renal dosing is outlined in Table 34–2. No dose adjustment is required in patients with mild hepatic impairment.

It has the least renal dependence of the three FDA-approved NOACs and has limited drug interactions. While there is minimal CYP3A4 metabolism, the drug does have potential interactions with potent CYP3A4 inhibitors. For patients on doses of apixaban greater than 2.5 mg twice daily, the dose should be decreased by 50% when it is administered with drugs that are strong dual inhibitors of CYP3A4 and P-gp (eg, ketoconazole, itraconazole, clarithromycin, or ritonavir). For patients receiving a dose of 2.5 mg twice daily, strong dual inhibitors of CYP3A4 and P-gp should be avoided (see Table 34–3).

In the setting of nonvalvular AF, the ARISTOTLE trial demonstrated that apixaban was superior to dose-adjusted warfarin in preventing stroke and systemic embolism. There was a lower rate of bleeding complications, and a lower mortality.²⁰ In the AVERROES trial, apixaban 5 mg twice daily was compared with aspirin (81-325 mg) for stroke prevention.²¹ There was a lower risk of stroke with apixaban but, interestingly, no difference in the bleeding rate compared with aspirin.

FDA approval of apixaban for acute DVT/PE was based on 2 large, prospective, phase 3 trials, AMPLIFY²² and AMPLIFY-EXTENSION.²³ There was no 5 to 10 day heparin bridge required. Apixaban proved noninferior to standard therapy with LMWH followed by warfarin with regard to recurrent VTE events, and was associated with significantly less major bleeding, and less clinically relevant nonmajor bleeding than standard therapy. The FDA-approved dose of apixaban for the treatment of acute DVT and/or PE is 10 mg twice daily for 1 week followed by 5 mg twice daily. Based on the extension study, a dose of 2.5 mg twice daily is indicated to reduce the risk of recurrent DVT and PE following initial 6 months treatment for DVT and/or PE.

In the setting of total joint replacement, apixaban is administered as 2.5 mg orally 12 to 24 hours after surgery. For hip replacement, the dose is 2.5 mg twice daily for 35 days and for total knee replacement (TKR), the dose is 2.5 mg po twice daily for 12 days. As noted, none of the NOACs are approved for use as prophylaxis other than TKR and total hip replacement (THR). Importantly for ICU patients, when patients cannot swallow whole tablets, 5 mg and 2.5 mg apixaban tablets may be crushed and suspended in 60 mL D₅W and immediately administered through a nasogastric tube. There are no data regarding crushed and suspended apixaban tablets swallowed by mouth. Like rivaroxaban, apixaban is pregnancy category B.

Anticoagulant Monitoring for Novel Oral Anticoagulants

The NOACs do not require routine monitoring in the setting of AF, acute VTE, or in the setting of VTE prophylaxis. However, it is important to check coagulation tests when a patient is admitted to the ICU on a NOAC, particularly if it is unclear whether or not the patient has been taking one. This may also be useful if there is bleeding or a high risk of bleeding. Thus, coagulation assays are not helpful in adjusting doses of NOACs but may help determine whether there is significant ongoing anticoagulant effect. With acute bleeding, or with emergency invasive surgical procedure, this is particularly important.

Dabigatran catalyzes the conversion of fibrinogen to fibrin, and thus, results in prolongation of most routine coagulation assays except the prothrombin time (PT). The ecarin clotting time is useful but less widely available. Because the thrombin time increases linearly with increasing dabigatran concentration, it seems intuitive that this measurement would provide the best direct assessment of thrombin activity. However, the thrombin time is overly sensitive to dabigatran levels and may be prolonged in the setting of a clinically insignificant dabigatran effect. The dilute thrombin time (Hemoclot assay) has very good linear correlation to plasma levels of dabigatran and is probably the most reliable method to measure the anticoagulant effect of dabigatran.²⁴ The aPTT can also be used; however, the relationship between dabigatran concentration and the aPTT is nonlinear and so is less reliable. A normal aPTT will likely indicate the absence of a clinically important anticoagulant effect. It is crucial to develop additional laboratory studies to correlate coagulation assay results with plasma levels of dabigatran.

Rivaroxaban and apixaban directly inhibit factor Xa, complexing with factor Va, independent of antithrombin. An increased PT tells us it is likely there is rivaroxaban in the bloodstream. There is less effect on the aPTT. A specific assay has been developed for direct Xa inhibitors that differs from the anti-Xa assay used to monitor LMWH, and this could potentially provide an effective method to determine the effect of rivaroxaban.²⁵ More data are needed, however. At present, there is no precise means by which to accurately assess the degree of anticoagulation in a patient on an Xa inhibitor.

Surgery and Procedures in the Critically Ill Patient

When urgent or emergent procedures arise, there are a number of potential anticoagulant considerations. The risk/benefit of anticoagulant discontinuation depends on the reason the patient is anticoagulated, the bleeding risk imparted by the procedure, and concomitant comorbidities. A patient with submassive or massive acute PE with another life-threatening condition requiring immediate surgery should simply have an inferior vena cava filter (IVCF) placed and have anticoagulation discontinued. Patients should be individualized.

As described earlier, neither the aPTT nor PT can determine the anticoagulant effects of dabigatran, rivaroxaban, or apixaban at any given time. A normal aPTT suggests that hemostatic function is not impaired by dabigatran, and a normal PT or lack of antifactor Xa activity would indicate that hemostatic function is not impaired on rivaroxaban or apixaban. Thus, these tests are useful when upcoming surgical procedures impart a significant bleeding risk. A normal thrombin time excludes a significant dabigatran effect; this measurement may be useful when a high-risk procedure is emergent.

Dabigatran is primarily renally eliminated so the timing of discontinuation should be based on the CrCl and the bleeding risk associated with the procedure. The pharmacodynamic effect of dabigatran declines in parallel with its plasma concentration, so surgery may only need to be delayed for about 12 hours after the last dabigatran dose.

If the CrCl is 31 to 50 mL/min, the last dose of dabigatran should be at least 48 hours before the procedure for low-risk surgery, and even longer (consider 4 days) before a procedure that poses a high risk of bleeding.¹¹ Mild or moderate renal impairment would appear to be of less concern in patients on rivaroxaban, in whom a decreased CrCl appears to have a more limited effect on the half-life of the drug, but it should still be considered when stopping the drug preoperatively.

A large nonrandomized study examined perioperative bleeding rates from 7 days prior until 30 days after invasive procedures for patients on warfarin or dabigatran.²⁶ The procedures included pacemaker/defibrillator insertion, dental procedures, diagnostic procedures, cataract removal, colonoscopy, and joint replacement. The last dose of dabigatran was given a mean of 49 hours (range 35-85) before the procedure, compared to 114 hours (range 87-144) for the last preprocedure dose of warfarin (P < 0.001). There was no significant difference in the rates of periprocedural major bleeding between the drugs. Among patients having urgent surgery, major bleeding occurred in 21.6% with warfarin, 17.7% with dabigatran at 150 mg, and 17.8% with dabigatran at 110 mg.²⁶

While there are no large studies examining perioperative bleeding rates in patients receiving rivaroxaban or apixaban, the same general principles should apply with regard to procedure bleeding risk and drug elimination based on the characteristics of these drugs. It should be recognized that these guidelines are approximate and that patients should be carefully scrutinized based on the specific procedure, perceived bleeding risk, and comorbidities including renal function.

For rivaroxaban and apixaban, it appears acceptable to delay low-risk surgery for approximately 24 hours after the previous dose. When the bleeding risk is higher, 48 hours would appear to be safer. With a more significant decrease in CrCl (ie, < 30 mL/min), an approximately 4 day delay would be appropriate. Resumption of a NOAC in a low bleeding risk scenario should be at least 24 hours after the procedure. The European Society of Anaesthesiology and the French Study Group on Thrombosis and Hemostasis have published recommendations about perioperative management of NOACs.^27,28

The general principles described above also apply to neuraxial anesthesia but it should be realized that bleeding in this setting has tremendous implications. Catheter placement should be considered when the anticoagulant level is at its trough and while removal may be less critical; it should also be timed as carefully as possible. At least 2 half-lives should be allowed to pass before catheter removal, at which point only 25% of the drug remains active. Patients should be monitored carefully for bleeding after catheter removal. Recommendations for the use of the NOACs in the setting of neuraxial anesthesia have been proposed by Llau and colleagues based on existing guidelines and the pharmacokinetics of each drug.²⁹

Anticoagulant Reversal

With the much shorter half-lives of the NOACs, discontinuing a NOAC and providing supportive care may be all that is required depending on the type and severity of the bleed, or whether an invasive procedure can be delayed. The half-life of dabigatran after multiple doses is approximately 14 to 17 hours and is not dose dependent; if there is no active bleeding after an overdose, stopping the drug may be sufficient. The shorter half-life proven in younger healthy patients on rivaroxaban or apixaban compared with elderly patients may also be favorable with regard to simply stopping the drug.

When a bleeding event occurs, initial measures include control of the bleeding site, volume resuscitation with fluids and/or packed red blood cells, as well as determination of the source of bleeding. Minor bleeding such as epistaxis, or other mucosal or superficial bleeding can often be managed symptomatically with compression/nasal packing and drug discontinuation. Gastrointestinal bleeding is managed by anticoagulant discontinuation, blood transfusion as needed, and aggressive endoscopic and other specific therapy. Life-threatening bleeding including intracerebral hemorrhage (ICH) requires not only withdrawal of the anticoagulant and supportive measures but also ICU transfer and potentially interventional procedures as well as consideration for reversal. Nonspecific reversal agents can be considered in patients with major or life-threatening bleeding.^30,31

Warfarin and other VKAs may be reversed with vitamin K and/or FFP. Reversal is not instantaneous. Four-component PCCs are recommended in recent guidelines.²⁸ Four-factor PCC is FDA approved for use in the United States for warfarin reversal in the setting of severe bleeding.

In the case of overdose with an anticoagulant, activated charcoal may prevent additional oral drug absorption when administered within 1 to 2 hours of ingestion. The minimal data available suggest that activated charcoal may be useful in dabigatran and apixaban overdose or accidental ingestion, and probably this applies to rivaroxaban as well.^32,33 Hemodialysis may reverse the anticoagulant effects of dabigatran overdose in severe bleeding because only about 35% of dabigatran is bound to plasma proteins.¹¹ Rivaroxaban and apixaban are highly protein bound (95% and 87%, respectively) precluding removal with dialysis.

While FFP is frequently administered for initial control of bleeding in anticoagulated patients, its use as a NOAC reversal agent has not undergone detailed evaluation in humans. The 2011 American College of Cardiology Foundation/American Heart Association guidelines recommended that severe bleeding from dabigatran merits transfusion of FFP, packed red cells, and surgical intervention as appropriate.³⁴ Studies in mice suggest that FFP may help limit ICH hematoma expansion.³⁵ In humans, data are inadequate to support the use of FFP in ICH caused by dabigatran.

While FFP may be useful in cases of coagulation factor depletion, it is not generally effective in reversing bleeding resulting from inhibition of coagulation factors.³⁰ However, it may be that delivery of clotting factors may overwhelm an ongoing inhibitory effect (see “Nonspecific Reversal Agents”). Still, overall, FFP remains controversial based on a lack of clear supportive data.

Specific NOAC Reversal Agents

Idarucizumab is a monoclonal antibody fragment that binds dabigatran with an affinity that is 350 times higher than that observed with thrombin. (REF 1 BELOW) Thus, this drug binds free and thrombin-bound dabigatran and neutralizes its activity (SAME REF 1 AS BELOW). In a clinical trial leading to FDA approval, idarucizumab completely reversed the anticoagulant effect of dabigatran within minutes. Andexanet alfa is a novel recombinant, modified factor Xa molecule that acts as a factor Xa decoy that binds and sequesters direct Xa inhibitors in the blood (REF 2 BELOW). The native factor Xa is then be available to participate in the coagulation process and restore hemostasis. In a preliminary report of an ongoing cohort study in patients with acute major bleeding associated with the use of factor Xa inhibitors, andexanet rapidly reversed anti–factor Xa activity and was not associated with serious side effects (SAME REF 2 AS BELOW). Effective hemostasis was achieved 12 hours after an infusion of andexanet in 79% of the patients. Thrombotic events occurred in 18% of the patients in the safety population. Additional data are pending as is consideration for FDA approval. Finally, another agent, ciraparantag (PER977), is being evaluated as a more universal anticoagulant reversal agent (REF 3 BELOW). This small, synthetic, water-soluble, cationic molecule is designed to bind specifically to unfractionated heparin an LMWH through noncovalent hydrogen bonding and charge–charge interactions. This drug binds in a similar way to the new oral factor Xa inhibitors and to dabigatran.

Nonspecific Reversal Agents

Nonspecific agents have been used for reversal of major bleeding in this setting^37,38; however, more data are needed. Recombinant factor VIIa (NovoSeven) initiates thrombin generation by activating factor X. Prothrombin complex concentrates were originally utilized for treating patients with hemophilia B. More recently, data have been published related to the treatment of VKA-related bleeding for their ability to rapidly and effectively correct the INR. These preparations are concentrated solutions derived from human plasma containing coagulation factors II, IX, X, and/or VII. They exist as either 3-factor or 4-factor PCCs depending on their factor VII content. Three-factor PCCs contain the inactivated vitamin K–dependent coagulation factors II, IX, and X, with minimal to no factor VII. Four-factor PCCs contain the same 3 coagulation factors with inactivated factor VII concentrations similar to their factor IX content. Thus, thrombin formation is stimulated. Both 3-factor and 4-factor PCCs have been studied in major bleeding with VKAs, although none have been compared head-to-head.

There is only one 4-factor PCC available in the United States, marketed under the name Kcentra (marketed as Beriplex P/N in 25 other counties) and the only PCC in the United States that is FDA-approved indication for the urgent reversal of acquired coagulation factor deficiency in the setting of vitamin K–related acute major bleeding. In addition to clotting factors, it also contains proteins C and S, antithrombin III, human albumin, and heparin. In view of the latter, it is contraindicated in HIT.³⁹

There are two 3-factor PCCs available in the United States—Profilnine SD and Bebulin. Another PCC, an activated PCC (aPCC, also referred to as anti-inhibitor coagulant complex) is available in the United States under the name FEIBA NF. It differs from other PCCs in that it contains inactivated factors II, IX, and X and small amounts of activated factor VII; this combines the effect of both recombinant factor VIIa and 4-factor PCC.

van Ryn and colleagues found that recombinant factor VIIa and activated prothrombin complex concentrate may be potential antidotes for dabigatran-induced severe bleeding in humans.³⁸ Marlu and associates determined that activated prothrombin complex concentrate as well as 4-factor PCC could be reasonable antidotes to dabigatran and rivaroxaban.³⁷

Thus far, it appears that 3-factor PCC products may be less effective than 4-factor PCCs in reversing elevated INRs in patients with warfarin overdose, but more data are needed for the NOACs. Finally, more data are needed with regard to the thrombotic risk associated with the use of these nonspecific prohemostatic agents.

Critically ill patients continue to require careful consideration with regard to anticoagulation and bleeding risk. Parenteral anticoagulants are favored in the ICU and we have substantial experience with these agents with regard to monitoring, perioperative dosing, and reversal. The NOAC era is upon us. The learning curve remains steep. Renal dosing and drug interactions must be considered. As with LMWHs, antifactor Xa activity monitoring may become a more available validated means of testing for exposure to rivaroxaban and apixaban. Newer assays derived from INR testing may become more useful methods to monitor NOACs. More studies are needed to examine reversal.