Chapter 25: Principles of Antithrombotic Therapy

Antithrombotic agents are highly effective and are among the most commonly used drugs in medicine because thrombotic diseases are the leading cause of mortality and morbidity in Western countries. Antithrombotic agents are characterized separately as anticoagulants, antiplatelet agents, or fibrinolytic drugs, depending on their primary mechanism, although there is overlap in their activities and clinical indications (Table 25–1). Their greatest use is in prevention of thrombosis in patients at high risk, but they also have important applications for treating acute thrombosis. For many agents, the risk-to-benefit ratio is narrow, with the result that bleeding complications occur, and bleeding is the most common adverse effect. Consequently, the clinician should carefully weigh the risks and benefits for each patient when selecting treatment. Generally, these drugs do not cause bleeding by themselves; instead, they exacerbate preexisting bleeding or predispose to bleeding from pathologic lesions that may be found in the gastrointestinal or genitourinary tracts or central nervous system. A careful review of comorbid conditions that may increase bleeding risk is important when deciding on therapy.

Anticoagulant therapy acts to decrease fibrin formation by inhibiting the formation and action of thrombin, and its most common uses are in preventing systemic embolization in patients with atrial fibrillation and for secondary prevention of venous thromboembolism. Anticoagulant therapy is sometimes monitored using coagulation testing because of marked biologic variation in effect with some agents. Antiplatelet agents act to inhibit platelet function, and their primary uses are in preventing thrombotic complications of cerebrovascular and coronary artery disease. They also have a role in treatment of acute myocardial infarction and some effect in preventing venous thrombosis. Fibrinolytic agents accelerate lysis of thrombi by increasing conversion of plasminogen to plasmin, and are primarily used in the acute management of myocardial infarction, in clearing occluded catheters and also in selected patients with stroke or venous thromboembolism. Fibrinolytic therapy is associated with a higher risk of bleeding complications than treatment with either anticoagulants or antiplatelet agents. Treatment of acute thrombosis often involves combinations of agents with multiple actions for maximum effect.

Antithrombotic therapy has changed considerably with the introduction of several novel target-specific oral anticoagulants. These drugs offer more predictable pharmacokinetics and fewer drug interactions than warfarin, thus eliminating the need for frequent monitoring. These agents have been studied in numerous phase III randomized clinical trials that have compared these agents to standard anticoagulation therapies. In most cases, the new anticoagulants are equivalent to conventional anticoagulation in reducing thrombotic events, while in many cases they are associated with lower rates of major bleeding complications.

The development of vitamin K antagonists as oral anticoagulants began in the 1920s with investigation of a hemorrhagic disease in cattle, the cause of which was eventually traced to ingestion of moldy hay leading to hypoprothrombinemia.¹ A coumarin that inhibited vitamin K was purified and eventually introduced into clinical practice in the 1940s. Several coumarin derivatives with differing pharmacologic properties are now available as anticoagulants worldwide, and are collectively referred to as vitamin K antagonists, but warfarin is nearly universally used in North America. These agents are widely used to prevent or treat common thrombotic diseases and represent the most commonly used oral anticoagulant currently available.²

PHARMACOLOGY

The coumarins are competitive inhibitors of vitamin K. They inhibit γ-carboxylation reactions required for synthesis of several coagulation proteins, including factors II, VII, IX, and X, as well as proteins C and S, which are involved in inhibitory regulation of hemostasis. The synthesis of these proteins requires a posttranslational modification of several glutamic acid residues, converting them to γ-carboxylated glutamic acid (Gla), which is required for proper membrane interaction and biologic activity (Chap. 115).^3,4 The carboxylation reaction requires reduced vitamin K, which is converted to vitamin K epoxide in the reaction. Vitamin K epoxide subsequently undergoes reduction by an enzyme that is inhibited by warfarin.^5,6,7 Therefore, treatment with warfarin causes reduced γ-carboxylation, leading to synthesis of molecules with impaired activity.^8,9,10

Warfarin preparations consist of a racemic mixture of S and R enantiomers in approximately equal proportion in an oral formulation with high bioavailability. Warfarin is water soluble and rapidly absorbed after oral administration, reaching a peak concentration after 60 to 90 minutes. An intravenous preparation is also available for patients who cannot take oral medications or who have malabsorption. It is tightly bound to plasma proteins with a half-life of 35 to 45 hours, with only the free, nonbound form having biologic activity. Warfarin is metabolized through the cytochrome P450 (CYP) system, the activity of which is influenced by environmental factors and also by genetic polymorphisms that alter the structure of common enzymes. Other vitamin K antagonists have similar activities but exhibit differences in absorption and elimination.

Because warfarin is a vitamin K antagonist, its action is influenced by the vitamin K content of the diet. Naturally occurring vitamin K is found in a variety of vegetables, and changes in diet can affect the vitamin K availability and warfarin effect.¹¹ This may be seen particularly in patients receiving warfarin who are on strict weight-reduction diets or in those with little oral intake because of illness. Also, diarrhea can affect vitamin K availability, as can administration of broad-spectrum antibiotics leading to marked warfarin sensitivity in hospitalized patients. Ingestion of vitamin K in dietary supplements or vitamins also affects sensitivity to warfarin. Liver disease can increase sensitivity to warfarin because of impaired synthesis of coagulation factors, and hyper- or hypometabolic states may also alter sensitivity. Hereditary resistance to warfarin has been described and related to specific mutations in vitamin K epoxide reductase.¹² Many drug interactions can influence the pharmacodynamics of warfarin by altering synthesis or clearance of vitamin K–dependent coagulation factors or interfering with warfarin metabolism, and patients should be advised to consult their physician or pharmacist about effects on anticoagulation when changing drug therapy or starting new medications (Table 25–2).¹¹ Other commonly used drugs affecting hemostasis, such as aspirin, nonsteroidal antiinflammatory agents, heparins, and other anticoagulants can add to the antihemostatic effects of warfarin and can lead to bleeding.

ADMINISTRATION AND MONITORING

The anticoagulant effect of warfarin is the result of decreased levels of vitamin K–dependent coagulation factors, and their concentration represents a balance of synthesis and metabolism. Warfarin administration impairs synthesis, and levels of vitamin K–dependent factors fall in relation to their metabolism. This is short for factor VII, with a half-life of approximately 5 hours, but longer for factors X and IX (t_1/2 = 24 hours) and longest for factor II (prothrombin) with a half-life of approximately 72 hours. The desired anticoagulant effect results from a balanced reduction of all factors and requires several days to achieve. Imbalances in reduction of coagulation factors may occur during initiation of therapy as factor VII level falls rapidly, whereas others, especially factor II, decline more slowly. The initial rapid fall in factor VII level may lead to an early elevation in the prothrombin time (PT) expressed as international normalized ratio (INR) without reflecting the desired anticoagulant effect. Because protein C is a natural inhibitor of coagulation with a short half-life (approximately 8 hours), its level may fall rapidly, theoretically inducing a procoagulant state during initiation of therapy.

As a result of the delayed anticoagulant effect of warfarin, therapy is initiated with a rapidly acting agent, such as heparin or low-molecular-weight heparin (LMWH), if immediate anticoagulation is needed. For example, patients with venous thromboembolism are typically given heparin or an LMWH for rapid effect, and warfarin is also administered within the first 24 hours. After a period of 5 or more days, the necessary anticoagulant effect of warfarin is achieved, and the parenteral anticoagulant can be stopped. Anticoagulation is initiated with a dose close to the expected daily maintenance requirement, which is usually between 5 and 10 mg.^13,14,15 There is, however, great variability in the doses required, and smaller amounts should be used for frail, elderly, or poorly nourished patients, or those with an increased bleeding risk. In patients with a low level of protein C or protein S as a result of an inherited deficiency, initiation of warfarin therapy without concomitant heparin or other immediately acting anticoagulant can lead to very low levels of these natural anticoagulants with ensuing thrombosis such as skin necrosis.

The anticoagulant effect of the vitamin K antagonists is monitored using the PT, which is sensitive to decreases in vitamin K–dependent factors and is progressively lengthened as the vitamin K–dependent factors reach lower levels. A critical component of the PT is the thromboplastin reagent that is used. Variability in thromboplastin composition leads to variation in results. The widespread introduction of the INR has improved standardization of results.¹⁶ Manufacturers determine the potency of thromboplastins by measuring the international sensitivity index (ISI), and this is used as a correction factor for the responsiveness of the thromboplastin in the PT. The INR represents the ratio of the patient PT to control PT corrected by the ISI. By this method, INR values obtained in different laboratories can be reliably compared for therapeutic monitoring.

During initiation of therapy, the INR is checked every 2 to 3 days for 1 to 2 weeks until a stable therapeutic effect is achieved. The target INR for most indications is 2.5 with a desirable therapeutic range from 2 to 3. A higher INR is recommended for patients with mechanical heart valve replacement and for those who failed anticoagulant therapy despite well-documented INR values in the 2 to 3 range. During chronic therapy, the INR should be monitored regularly, depending on stability of the response, and minor dose adjustments are frequently needed. A recent randomized trial in patients on chronic warfarin therapy with a stable dose for 6 months demonstrated that INR monitoring every 12 weeks was noninferior to monitoring every 4 weeks with regards to time in therapeutic range (74.1 vs. 71.6 percent).¹⁷ Monitoring can also be performed using portable instruments that are suitable for home use, enabling selected patients to learn to modify their warfarin doses in response to their INR value.¹⁸ A nonblinded randomized study compared home monitoring to INR monitoring in a specialized coagulation clinic in a cohort of patients on chronic warfarin and capable of conducting home monitoring.¹⁹ In this trial, rates of the primary composite end point (thrombosis, major bleeding or death) were equivalent, but home monitoring was associated with significant improvement in patient satisfaction, quality of life and time in therapeutic range. Specialized clinics devoted to monitoring warfarin typically achieve better results in maintaining patients within the therapeutic range, resulting in fewer bleeding complications.^20,21 Problems with keeping patients within the therapeutic range often result from failure of compliance, changes in diet, medication or alcohol intake, or intercurrent illnesses.

Warfarin sensitivity is affected by polymorphisms in CYP and vitamin K epoxide reductase complex (VKORC), and pharmacogenomics may become important in dosing. The clearance of warfarin is the result of hepatic metabolism and CYP2C9 is the most important enzyme mediating its clearance.²² A number of polymorphisms have been identified, but the most important are CYP2C9*2 and CYP2C9*3, which are found in approximately 11 percent and 7 percent of patients and result in reductions of enzymatic activity of approximately 30 percent and 80 percent, respectively.^23,24,25 This reduced metabolic clearance leads to increased drug levels and an increased anticoagulant effect. VKORC1 converts oxidized vitamin K to the active reduced form as required for posttranslational carboxylation. VKORC1 is the target of warfarin, which functions as a competitive inhibitor. Numerous coding polymorphisms have been identified that can affect the response to warfarin.^26,27 Common haplotypes can be separated into low-dose (A) and high-dose (B) groups with different sensitivities to warfarin.

Evidence is clear that polymorphisms in either CYP2C9 or VKORC1 affect warfarin sensitivity. Retrospective studies have shown that patients with at least one variant allele have an increased risk of INRs over the desired range and the variant groups also require more time to achieve stable dosing compared to patient with wild type allele.^28,29 Additional prospective observational studies indicate that CYP2C9 and VKORC1 genotype affect warfarin sensitivity.^30,31,32 Several prospective studies have incorporated genotyping for CYP2C9 and VKORC1 into algorithms that typically include clinical variables such as age, gender, and drug interactions in an attempt to use genetic information to improve warfarin dosing. The results of these trials are conflicting possibly as a consequence of different trial design and patient populations. Incorporation of CYP2C9 and VKORC1 polymorphisms into a dosing algorithm in one randomized study more accurately predicted warfarin dose and reduced dose modifications, but did not improve time in therapeutic range.³³ Another randomized study compared dosing based on genotype and clinical variables to clinical variable alone in almost 1000 patients starting warfarin. This showed no difference in time in therapeutic range, bleeding, or thromboembolism.³⁴ Genotype-based dosing resulted in significantly decreased time in the therapeutic range for black patients in this study suggesting such an approach is detrimental for this subgroup. A third study conducted in Europe reported a significant improvement in time in therapeutic range (67.4 vs. 60.3 percent), episodes of excessive anticoagulation (INR >4), and median time to reach a therapeutic INR (21 vs. 29 days) with genotype-based dosing.³⁵ A similarly designed study by the same European group in patients starting acenocoumarol or phenprocoumon showed no improvement in these same measures with genotype based dosing.³⁶ Further research is needed before genetic testing becomes part of standard care for patients receiving warfarin.

COMPLICATIONS

The most serious and common complication of warfarin is bleeding, and its risk is related primarily to patient characteristics, the degree of the anticoagulation, and the duration of therapy. The rate of major bleeding with 3 to 6 months of warfarin treatment in recent clinical trials of patients with venous thromboembolism was 1.2 to 2.2 percent.^37,38,39 Several risk models have been developed to predict bleeding risk in patients on warfarin. The HAS-BLED score was derived and validated in a prospective cohort of more than 6000 patients on warfarin for atrial fibrillation (Table 25–3).⁴⁰ Variables in this model include hypertension, abnormal renal or liver function, stroke, prior bleeding complications, labile INR, age older than 65 years, and drug or alcohol abuse. The intensity of anticoagulation as reflected by the INR is the most important predictor of bleeding risk, which is low in the therapeutic range but increases as the INR prolongs further. The cumulative risk of bleeding increases with a longer duration of treatment, whereas the absolute risk is greatest early, possibly caused by pathologic lesions present at the time therapy is started.

A rare complication of warfarin therapy is skin necrosis, which if it occurs usually presents early in the course of anticoagulation.⁴¹ Thrombosis in dermal and subdermal venules is the underlying cause, and this may be caused by disproportionately rapid reduction in proteins C and S. Skin necrosis typically begins with burning and tingling at the affected site, which usually involves a region with a large amount of subcutaneous tissue, such as the breast, buttock, or thigh. Painful hemorrhagic full-thickness skin infarction may develop, and frequently will require skin grafting. Other complications from warfarin are rare. Occasional patients report alopecia; hypersensitivity reactions are rare and are almost uniformly caused by the dye used in the pill rather than by the warfarin itself.

Warfarin use in patients with heparin-induced thrombocytopenia (HIT) may be complicated by limb gangrene caused by occlusive venous thrombosis; this effect is likely a result of a combination of inadequate parenteral anticoagulant effect and reduced levels of protein C, in concert with relatively preserved levels of factors II and X that are seen early after the initiation of warfarin administration.⁴² This complication highlights the need for parenteral anticoagulants to be continued in patients with HIT until the coagulopathy has largely resolved, as indicated by a return of the platelet count to normal or near normal levels.

Oral anticoagulation should be avoided in pregnancy because warfarin crosses the placenta, and exposure during organogenesis in the first trimester can lead to fetal embryopathy with significant cranial bone malformations.⁴³ Anticoagulation during pregnancy increases bleeding complications, especially later in pregnancy. Warfarin may be considered during the second trimester, but heparin or an LMWH is a preferable alternative in most situations. Vitamin K antagonists are safe during lactation.⁴⁴

REVERSAL OF ANTICOAGULATION

Anticoagulation must be reversed for episodes of bleeding, surgery, trauma, or overdosage. Appropriate interventions for patients with excessively prolonged INRs without bleeding include holding warfarin doses, administering low doses of vitamin K (0.5 to 1.0 mg), and increasing the frequency of monitoring (Table 25–4).⁴⁵ Serious bleeding and major warfarin overdosage requires factor replacement and larger vitamin K doses that may need to be given intravenously. Four-factor concentrates have been developed and tested for reversal of warfarin. A large, open-label, prospective phase IIIB noninferiority study in nonsurgical patients on warfarin requiring reversal for major bleeding compared fresh-frozen plasma (FFP) to four-factor prothrombin complex concentrate (PCC) (factors II, VII, IX, X, protein C, and protein S).⁴⁶ Four-factor PCC was as effective in achieving hemostasis as plasma (72.4 percent with four-factor PCC vs. 65.4 percent with plasma; absolute difference, 7.1 percent [95 percent confidence interval, −5.8 to 19.9]) and superior to FFP with regards to INR correction within 0.5 hours of infusion (62.2 percent receiving four-factor PCC vs. 9.6 percent receiving FFP; absolute difference, 52.6 percent [95 percent confidence interval, 39.4 to 65.9]). The rate of adverse events, such as death and thromboembolism, was similar between groups.

Anticoagulated patients who need invasive procedures represent management problems, and decisions about periprocedural anticoagulation should be based on balancing the risk of thromboembolism with that of bleeding from the procedure. Evidence-based guidelines for bridging therapy are available,⁴⁵ and the topic has been reviewed.⁴⁷ The goal is to reduce the intensity of anticoagulation during and immediately after surgery, while avoiding thromboembolism caused by the underlying disease. Generally, the risk of recurrence is greatest in the period shortly after an episode of acute thrombosis and declines progressively over time. If possible, elective surgery and other invasive procedures associated with a high bleeding risk should be postponed during the first several months following acute thrombosis. The bleeding risk is usually highest during surgery and decreases rapidly to baseline after approximately 7 to 10 days. Most surgery can be done with a minimal bleeding risk in patients receiving warfarin and an INR of 1.5 or less. A recent cohort study of 1496 patients on chronic anticoagulation undergoing periprocedural bridging therapy showed that bleeding occurred in 5.1 percent of patients, and risk factors for bleeding included mitral mechanical valve, active cancer, prior bleeding history, and reinitiation of anticoagulation within 24 hours of procedure.⁴⁸ A randomized study of anticoagulated patients undergoing defibrillator placement showed that bridging therapy was associated with a significant increase in device-pocket hematoma (16 percent vs. 3.5 percent) without a reducing thromboembolic complications.⁴⁹ Patients at moderate or high risk of thrombotic recurrence should receive heparin or LMWH “bridging therapy” when their INR becomes subtherapeutic. Postprocedural bridging therapy should be undertaken only in patients in whom the risks of this therapy (principally bleeding) are less than the perceived benefits (a reduced risk of thromboembolism).

PHARMACOLOGY

Heparin (and the related LMWHs) is the most widely used, rapidly acting, parenteral anticoagulant. Heparin is a heterogeneous mixture of sulfated glycosaminoglycan molecules with varying chain lengths (molecular mass [Mr] 5000 to 30,000 daltons). Heparin has no direct anticoagulant effect, but serves to activate plasma antithrombin (AT), a serine protease inhibitor. Only about one-third of heparin molecules contain the necessary unique pentasaccharide sequence required to interact with AT and thus display anticoagulant activity.⁵⁰ AT inhibits thrombin, factor Xa, and other coagulation serine proteases in a reaction that is slow by itself, but is accelerated approximately 1000-fold in the presence of heparin.⁵¹ To inhibit thrombin, heparin binds to both the active enzyme and AT, forming a ternary complex. The inhibition of factor Xa, however, occurs through binding to heparin–AT complex without the requirement for heparin binding directly also to factor Xa. The requirement for a ternary heparin–AT–thrombin complex requires heparin molecules with 19 or more saccharide units, whereas smaller heparin molecules are effective in promoting factor Xa inactivation. Thrombin and factor Xa are relatively protected from inhibition by the heparin–AT complex when they are surface immobilized within thrombi or on cells.⁵²

ADMINISTRATION AND MONITORING

Heparin is not absorbed after oral ingestion, so must be given either subcutaneously or intravenously. The pharmacokinetics of unfractionated heparin is complex, a result of extensive protein binding, with a dose-dependent half-life in the range of 1 to 2.5 hours.⁵³ A common protocol uses an initial intravenous bolus of 5000 units or 75 U/kg, followed by a maintenance infusion of 1250 to 1660 U/h or 18 U/kg per hour. The anticoagulant effect is immediate, but laboratory monitoring is needed because of the variability in response among patients. Monitoring is most convenient with the activated partial thromboplastin time (aPTT), which is sensitive to plasma heparin concentrations of 0.1 U/mL or higher. Because different reagents and measuring systems have differing sensitivities to heparin, it is recommended that the therapeutic range be established for each laboratory by calibrating the aPTT to a plasma heparin concentration of 0.2 to 0.4 units by protamine sulfate titration, or 0.3 to 0.7 U/mL using an anti–factor Xa assay.⁵⁴ The usual aPTT range for heparin therapy is between 1.5 and 2.5 times the mean of the normal range. Clinically useful nomograms are available for adjusting the heparin dose using either fixed- or weight-based dosing.⁵⁵ Alternatively, monitoring can be performed using anti-Xa levels, which is a useful approach when the aPTT is unreliable, as in patients with baseline prolongation of the aPTT as a consequence of expressing a lupus anticoagulant. Rapid achievement of a therapeutic level, as assessed by the aPTT or anti-Xa level, is important in ensuring an adequate anticoagulant effect.

Some patients fail to display an adequately prolonged aPTT following treatment with unfractionated heparin, despite apparently adequate or even supratherapeutic doses of the drug. This phenomenon is termed heparin resistance and is usually caused by an acute-phase response that results in high levels of procoagulant proteins, including factor VIII. The antithrombotic effect of heparin correlates best with plasma heparin levels, which may be adequate in these circumstances despite a subtherapeutic aPTT.⁵⁶ For patients who require heparin doses of greater than 35,000 U/day to increase the aPTT into the therapeutic range, consideration should be given to using heparin levels determined by an anti-Xa assay. Substitution of LMWH for unfractionated heparin is another consideration. Although AT deficiency may cause heparin resistance, most AT-deficient patients can be adequately anticoagulated with heparin in usual doses. No monitoring is recommended when low doses of heparin are used for prophylaxis of venous thromboembolic disease, although minimal prolongation of the aPTT may occur. Care is required in very light weight and frail elderly who may be anticoagulated with usual “prophylactic” doses of unfractionated heparin; aPTT monitoring might be considered in such patients.

A large study demonstrated that patients with acute venous thromboembolism can be safely treated with fixed, weight-adjusted heparin doses without aPTT monitoring.⁵⁷ In this study, 708 patients were allocated randomly to receive either unfractionated heparin with a subcutaneous bolus dose of 333 U/kg followed by a twice-daily dose of 250 U/kg, or LMWH over 3 months of follow up. Recurrent venous thromboembolism occurred in 13 patients who were allocated to unfractionated heparin and in 12 who were allocated to LMWH. Major bleeding occurred in four and five patients, respectively. This study calls into question the need for routine aPTT monitoring of therapeutic dose, weight-adjusted unfractionated heparin and warrants validation.

REVERSAL

Heparin has a short half-life, and its anticoagulant effect disappears several hours after discontinuation of an intravenous infusion. Therefore, stopping the infusion and local measures are usually adequate to control bleeding. However, in major or life-threatening bleeding, the anticoagulant effect can be neutralized with protamine sulfate, which is a basic polypeptide that binds tightly to the acidic heparin molecule. The usual dose of protamine required is 1 mg to neutralize 100 units of heparin. The dose to be administered is based on the amount of heparin remaining in the circulation. Protamine is routinely used to neutralize heparin after cardiopulmonary bypass using standard formulas and activated clotting time monitoring.

ADVERSE EFFECTS

The most frequent complication of heparin administration is bleeding, which is related to the dose and intensity of treatment, as well as to patient characteristics. HIT is an immune-mediated platelet consumption caused by an antibody directed against a complex of heparin and platelet factor 4 (Chap. 118). Despite thrombocytopenia, HIT is more commonly associated with thrombotic complications than bleeding, and it occurs in approximately 3 percent of patients, depending on the type of heparin, the dose and route of administration, and the indication for heparin anticoagulation (e.g., more common in “surgical patients” than in “medical patients”). The “4T’ score is a clinical prediction tool with good negative predictive value.⁵⁸ Platelet counts should be monitored during treatment and heparin discontinued if thrombocytopenia occurs, and an alternative anticoagulant that does not interact with the heparin-platelet factor 4 complexes should be administered. Vitamin K antagonists should be given only after the platelet count has risen to greater than 150,000/μL. Long-term heparin therapy can also cause osteoporosis, and radiographic evidence of bone loss occurs in approximately 15 percent of women who receive prolonged treatment during pregnancy, with symptomatic vertebrae fractures in approximately 2 percent. The bone loss may resolve after heparin is discontinued.

LOW-MOLECULAR-WEIGHT HEPARIN

Limitations of unfractionated heparin led to studies correlating structural and functional relationships of heparin, and this eventually resulted in the development of LMWHs, several of which are available. LMWH preparations are produced by treating heparin chemically or enzymatically to decrease the size of the polysaccharide chains, yielding products with restricted molecular weight distributions with a mean of approximately 4000 to 5000 daltons.⁵⁹ Like heparin, LMWHs exert antithrombotic effects through interaction with AT. In the presence of LMWH, AT inactivates factor Xa in the same way as unfractionated heparin, but it is less able to inactivate thrombin because the shorter polysaccharide length does not allow formation of the necessary ternary complex. Consequently, LMWHs have a greater proportion of anti–factor Xa than AT activity.

LMWHs also have different pharmacokinetic properties than unfractionated heparin.⁵⁴ Following subcutaneous administration, LMWHs are nearly completely absorbed, a clear benefit over unfractionated heparin, which exhibits variable and dose-dependent absorption. LMWHs also exhibit less binding to plasma proteins and cells than unfractionated heparin, resulting in more predictable blood levels and anticoagulant effects. LMWHs have a longer plasma half-life than unfractionated heparins, allowing once- or twice-daily subcutaneous administration.

LMWHs have significant renal clearance, and high levels can accumulate in patients with renal insufficiency. Care must be taken in dosing LMWHs in patients with reduced renal function as bleeding risks are increased,⁶⁰ and monitoring with anti–factor Xa levels may be needed. Similarly, monitoring may be necessary to achieve appropriate levels in very obese patients, although weight-based dosing probably achieves better anticoagulation. Protamine sulfate does not completely reverse the anticoagulant effect of LMWH but is partially effective and can be useful in patients with serious hemorrhage.⁶¹ Several LMWH preparations are available and approved for both prophylaxis and treatment of venous and arterial thrombotic diseases. Each preparation differs slightly and is pharmacologically unique, although the agents are likely similarly effective for the treatment and prevention of venous thrombosis.

Similar to unfractionated heparin, the most common adverse effect is bleeding, which occurs at approximately the same frequency and severity when used in similar patient groups for the same indication. HIT is much less common than with unfractionated heparin,⁶² occurring only in 0.3 to 0.45 percent of patients. However, cross-reactivity of the antibody occurs, and LMWH is not an acceptable choice for continued anticoagulation in patients with HIT induced by treatment with unfractionated heparin. Animal studies suggest that osteoporosis may be less common with LMWH, and this is reported by several small clinical trials.⁶³

The factors governing the choice of heparin or LMWH concern effectiveness, safety, convenience, and cost. For treatment of venous thromboembolic disease, the safety and efficacy of heparin and LMWH are comparable. A meta-analysis of 14 studies comparing unfractionated heparin to LMWH in 4754 patients showed that LMWH-treated patients had less recurrent venous thromboembolic complications (4.3 percent vs. 5.6 percent, odds ratio [OR] 0.76, 95 percent confidence interval 0.57 to 1.01) and fewer major bleeding events (1.3 percent vs. 2.1 percent, OR 0.6, 95 percent confidence interval 0.39 to 0.93).⁶⁴ LMWH also offers better convenience because subcutaneous administration permits outpatient treatment. However, LMWHs may be difficult to use in patients with renal insufficiency because of decreased clearance, and intravenous heparin may offer advantages in such patients. LMWHs are incompletely reversed by protamine sulfate, making them more difficult to use for cardiac bypass surgery. Unfractionated heparin may be preferable in patients who require an invasive procedure on an urgent basis because of its shorter half-life. LMWHs may offer some advantages for patients with acute coronary syndromes.

DANAPAROID

Danaparoid is a mixture of glycosaminoglycans and is composed of approximately 84 percent heparan sulfate, 12 percent dermatan sulfate, and 4 percent chondroitin sulfate. It is an AT-dependent anticoagulant with predominant anti–factor Xa activity. The plasma half-life is approximately 24 hours with predominant renal clearance. Danaparoid is not reversed by protamine sulfate. Danaparoid differs structurally from heparin and it has been used successfully to treat patients with HIT. Although there is in vitro cross-reactivity of 10 to 20 percent of heparin antibodies with danaparoid, this is of uncertain clinical relevance. Danaparoid is administered subcutaneously, and levels may be monitored with anti–factor Xa assays performed using a danaparoid standard curve. At the time of writing danaparoid has not been approved in the United States, and availability elsewhere was limited.

FONDAPARINUX

Fondaparinux is a unique heparin-like anticoagulant with highly selective AT-dependent anti–factor Xa activity.⁵⁴ It is a completely synthetic pentasaccharide whose structure is based on the heparin sequence that interacts with AT. It binds reversibly and with high affinity to AT, resulting in a conformational change that renders it effective in inhibiting factor Xa, but because of its small size, does not inhibit thrombin. Whereas unfractionated heparin and all LMWHs are derived from animal sources, fondaparinux is synthesized in a structurally homogenous form containing no animal products. Consequently, fondaparinux does not induce allergic responses. Because it inhibits factor Xa but has no direct action on thrombin, its mechanism of action depends on reducing thrombin generation.

Pharmacologic studies show that maximum plasma levels are reached approximately 2 hours after subcutaneous administration with an elimination half-life of approximately 17 hours independent of the dose.⁶⁵ Bioavailability is nearly complete after subcutaneous or intravenous administration. There is a low intra- and intersubject variability with little accumulation after multiple daily doses. Because elimination is primarily renal and the agent is excreted unchanged in the urine, fondaparinux is contraindicated in patients with severe renal impairment. Fondaparinux plasma levels can be measured with the anti–factor Xa assay, but there is no effect on other coagulation assays including the activated clotting time (ACT), aPTT, or thrombin clotting time.

Clinical studies have evaluated the use of fondaparinux in several conditions, and it is approved by the FDA for prevention of venous thromboembolism (VTE) in patients undergoing major orthopedic surgery or with hip fracture, prophylaxis after abdominal surgery, and treatment of deep venous thrombosis (DVT) or pulmonary embolism (PE). For prophylaxis it is administered subcutaneously in a dose of 2.5 mg once daily, whereas a weight-adjusted dose is used for treatment of VTE. The principal adverse effect is bleeding, and its frequency and severity have been comparable to those observed with LMWH. Elevated levels may occur in patients with renal insufficiency, and caution should be exercised in using fondaparinux in patients with renal compromise. Fondaparinux is a good choice for an anticoagulant in patients with HIT as there is no cross-reactivity.⁶⁶

BIVALIRUDIN

Bivalirudin, a direct thrombin inhibitor, is a recombinant protein based on the structure of hirudin, is composed of a dodecapeptide analogue of the carboxyterminal region of hirudin linked by a four-glycine residue to a structure directed to the active site of thrombin.⁶⁷ Pharmacokinetic studies show that plasma clearance is rapid (4.6 mL/min/kg) in patients with normal renal function with a volume of distribution of 0.2 L/kg and elimination half-life of approximately 30 minutes.⁵⁴ There is dose-dependent prolongation of the ACT and aPTT that correlates with plasma concentrations. the drug is eliminated by both renal and hepatic clearance, and consequently, dose modification is recommended for patients with moderate-to-severe functional liver or kidney disease.

Bivalirudin is effective when used with aspirin in patients with unstable angina or postinfarction angina undergoing angioplasty, and it is approved for this use.⁶⁸ It is also approved for patients with HIT undergoing percutaneous coronary intervention. In some clinical trials, bivalirudin also shows efficacy in preventing restenosis after coronary angioplasty, as an adjunct to streptokinase in acute myocardial infarction, and in preventing venous thrombosis after orthopedic surgery and in patients with HIT. However, these indications have not been approved by the FDA. The most common adverse effect is bleeding, and no specific antidote is available. The infusion should be discontinued in patients with bleeding complications and blood levels monitored with the aPTT or other coagulation parameters. Antibivalirudin antibodies have not been detected following therapy.

ARGATROBAN

Argatroban is a small-molecule arginine derivative that reversibly and directly inhibits thrombin by binding to the active catalytic site of the enzyme with a K_i of 3.9 × 10⁻⁸ mol/L.⁵⁴ Because of its small size, argatroban is an effective inhibitor of thrombin, both bound to surfaces and in solution.⁶⁹ The anticoagulant effect can be assessed with either the aPTT or ACT, and both correlate with plasma concentrations of the drug. Argatroban is approximately 50 percent protein bound and has a volume of distribution of 0.2 L/kg and an elimination half-life of 39 to 51 minutes.⁷⁰ Metabolism is primarily hepatic, and the clearance and half-life are prolonged in patients with hepatic functional abnormalities requiring dose reduction. Renal function has less effect on argatroban pharmacokinetics.

Argatroban is approved for treatment and prophylaxis of HIT and for percutaneous interventions in patients with HIT. It also shows some benefit in patients with thrombotic stroke in clinical trials. For treatment of HIT, argatroban is administered at 2 mcg/kg per hour and adjusted to maintain the aPTT at 1.5 to 3 times baseline. For patients with HIT who are undergoing percutaneous coronary interventions, the drug is administered as a bolus of 350 mcg/kg followed by a continuous infusion of 15 to 400 mcg/kg per minute for a target ACT of 300 to 450 seconds. As with other direct thrombin inhibitors, the main side effect is bleeding, and no specific agent is available to reverse its action. The anticoagulant effect may be prolonged in patients with hepatic impairment. If overdosage or excess bleeding occurs, the infusion should be discontinued and the aPTT and other coagulation parameters monitored.

The transition from argatroban to warfarin in patients requiring long-term anticoagulation is complicated because argatroban has a significant effect on both the PT and the aPTT. In patients transitioning to warfarin an INR should be measured; if it is greater than 4.0, the argatroban should be stopped for several hours and the INR remeasured. If the INR is still greater than 2.0, the argatroban can be discontinued; if it is less than 2.0, the argatroban should be reinstituted and the same procedure followed on the next day.

LEPIRUDIN

Lepirudin is closely related to hirudin, a natural anticoagulant found in the salivary glands of the leech. The half-life is 1 to 3 hours in normal volunteers with predominantly renal catabolism, but it may be as long as 2 days in dialysis-dependent patients.⁷¹ Lepirudin prolongs the aPTT in a concentration-dependent manner.⁷² The manufacturer discontinued marketing of this product and it is no longer available.

DABIGATRAN ETEXILATE

Dabigatran etexilate, one of the novel oral anticoagulants, is a direct thrombin inhibitor prodrug with a bioavailability of approximately 6 percent after oral administration. The absorbed drug is rapidly converted by esterases to dabigatran. Peak levels occur 1 to 2 hours after an oral dose; the half-life is approximately 12 hours. Dabigatran does not require a cofactor and reversibly inhibits the active site of thrombin. Dabigatran does not interfere with drugs that are metabolized by the CYP enzyme system and it produces a predictable anticoagulant response, which allows therapy without the need for monitoring.⁷³ Dabigatran prolongs several coagulation assays. The PT, aPTT, and ACT lack sensitivity for therapeutic drug levels, while the dilute thrombin time and ecarin clotting time appear to correlate well across a broad range of drug levels.⁷⁴ None of these tests have been shown to predict clinical outcomes such as thrombosis or bleeding.

The major side effect of dabigatran is hemorrhage. No specific antidote is currently available, although such antidotes are in development.⁷⁵ Consequently, bleeding complications are managed symptomatically. Although not well studied, there are published case reports showing that dialysis or hemoperfusion likely removes this compound from the circulation.⁷⁶ In animal models the administration of various coagulation factor concentrates or recombinant activated factor VII (factor VIIa) improved prolonged bleeding times, although these agents did not cause correlative changes in anticoagulation tests.⁷⁷

Dabigatran etexilate, and the other novel oral anticoagulants, are effective in various clinical settings (Table 25–5). Dabigatran is now FDA approved for management of VTE and prevention of stroke and systemic embolism in nonvalvular atrial fibrillation. In the RE-LY trial, patients with nonvalvular atrial fibrillation and a risk of stroke were randomized to dabigatran etexilate (110 mg BID or 150 mg BID) or dose-adjusted warfarin.⁷⁸ In this noninferiority trial, rates of stroke or systemic embolism were 1.69 percent per year with warfarin, 1.53 percent per year with dabigatran etexilate 110 mg (relative risk [RR] 0.91, p<0.001 for noninferiority) and 1.11 percent per year with dabigatran etexilate 150 mg (RR 0.66, p<0.001 for superiority). The rates of major bleeding in the RE-LY trial were similar between higher dose dabigatran and warfarin, but significantly lower for low-dose dabigatran compared to warfarin. The rate of intracranial bleeding was lower for both doses of dabigatran compared to warfarin. The RE-COVER trial was a randomized double-blind noninferiority trial comparing dabigatran etexilate (150 mg BID) to dose-adjusted warfarin in patients with acute VTE who received initial parenteral anticoagulation for a median of 9 days.³⁷ The rates of recurrent VTE were similar (2.4 vs. 2.1 percent, P<0.001 for noninferiority). In a meta-analysis of three large randomized trials, dabigatran was as effective as LMWH for prevention of VTE and VTE-related mortality after hip and knee replacement, and had a similar bleeding risk.⁷⁹ Dabigatran does not appear to be safe or effective in prevention of thromboembolic complications following heart valve replacement. The RE-ALIGN study randomized patients to dabigatran or warfarin following valve replacement and was terminated early because of excess thrombotic and bleeding complications in dabigatran-treated patients.⁸⁰

RIVAROXABAN

Rivaroxaban is a direct factor Xa inhibitor that is administered orally and produces its anticoagulant effect through reversible binding with factor Xa. Rivaroxaban can inhibit both free and thrombus-associated factor Xa. Like dabigatran, it is dependent on renal excretion, and bioaccumulation may occur in patients with renal insufficiency.

Rivaroxaban produces its peak anticoagulant effect within 4 hours of oral administration and has a terminal elimination half-life of 5.7 to 9.2 hours. PT is more accurate than the aPTT in measuring rivaroxaban effects; however, the sensitivity varies by specific reagent and is not good at lower therapeutic drug levels.⁸¹ A rivaroxaban calibrated anti-Xa assay appears to be the most accurate test as it correlates well with drug level across a broad therapeutic range.⁸²

Rivaroxaban is effective in a number of clinical settings. In the EINSTEIN-PE³⁹ and EINSTEIN-DVT³⁸ randomized clinical trials treatment with rivaroxaban (15 mg BID for 3 weeks followed by 20 mg once daily) was noninferior to standard anticoagulation (LMWH followed by warfarin) in preventing recurrent thrombosis in patients with acute VTE. In a large randomized trial of patients with nonvalvular atrial fibrillation, rivaroxaban was noninferior to warfarin for prevention of stroke or systemic embolism.⁸³ Prophylactic rivaroxaban (10 mg once daily) significantly reduced rates of postoperative VTE compared to the LMWH Lovenox (40 mg once daily) following hip arthroplasty (1.1 percent vs, 3.7 percent, p<0.001)⁸⁴ and knee arthroplasty (9.6 percent vs. 18.9 percent),⁸⁵ and in hospitalized acutely medically ill patients was noninferior to Lovenox for standard 10-day thromboprohylaxis.⁸⁶

The principal side effect of rivaroxaban therapy is bleeding. In a pooled analysis of the EINSTEIN studies, major bleeding was less common in patients receiving rivaroxaban compared to standard anticoagulation (1.0 percent vs. 1.7 percent, p = 0.002), and this includes a significant reduction in intracranial bleeding.⁸⁷ In another meta-analysis of phase III trials comparing rivaroxaban to standard anticoagulation, rivaroxaban was associated with a reduced risk of fatal bleeding complications.⁸⁸ There is no antidote for the anticoagulant effect of rivaroxaban. Prothrombin complex concentrates partially reverse prolonged coagulation times in healthy volunteers receiving rivaroxaban,⁸⁹ and PCC and recombinant activated factor VIIa improved bleeding times after rivaroxaban treatment in animal studies.⁹⁰ However, these agents have not been proven effective in managing rivaroxaban treated patients with bleeding complications.

APIXABAN

Apixaban is an orally administered direct inhibitor of factor Xa which is metabolized primarily by the hepatic cytochrome CYP3A4 enzyme. The onset of action and half-life are 3 hours and 12 hours respectively, and approximately 25 percent is excreted by the kidney. Apixaban affects standard anticoagulation assays less than rivaroxaban, and the effect is variable for different reagents within an assay group. In general the PT is more sensitive than aPTT, and chromogenic apixaban specific anti-Xa assays display the most accurate linear correlation across a therapeutic dose range.⁹¹ There is no antidote for reversal of apixaban. In animal studies recombinant activated factor VIIa but not PCCs reduced bleeding times in apixaban-treated animals.⁹²

Apixaban, like rivaroxaban and dabigatran, has been studied in several clinical settings. Two large phase III randomized clinical trials with slightly different designs led to the approval of apixaban for management of nonvalvular atrial fibrillation. In the ARISTOTLE trial, patients with nonvalvular atrial fibrillation and a risk for stroke were randomized to apixaban (5 mg BID) or dose-adjusted warfarin. Apixaban was superior to warfarin in reducing the rate of embolic or hemorrhagic stroke and systemic embolism (1.27 percent vs. 1.6 percent per year, p = 0.01), and fewer patients experienced major bleeding complications (2.13 percent vs. 3.09 percent, p<0.001).⁹³ The AVERROES trial was terminated early because of a clear reduction in stroke and systemic embolism in patients with nonvalvular atrial fibrillation deemed unsuitable or unwilling to take vitamin K antagonists randomized to apixaban versus aspirin (1.6 percent vs. 3.7 percent, p<0.001).⁹⁴ Interestingly, there was no difference in major bleeding or intracranial bleeding between apixaban and aspirin. Apixaban (10 mg BID for 7 days then 5 mg BID) has also been shown to be noninferior to standard anticoagulation (LMWH then dose-adjusted warfarin) in initial management of acute VTE,⁹⁵ and effective for secondary VTE prevention in patients treated for an extended time following acute VTE.⁹⁶ Major bleeding rates were significantly lower with apixaban compared to standard anticoagulation in the initial period after acute VTE (0.6 percent vs. 1.8 percent, p<0.001), and rates of major bleeding were equivalent in patients receiving extended thromboprophylactic dosing of apixaban compared to placebo (0.1 percent in the 5-mg apixaban group, 0.2 percent in the 2.5-mg apixaban group, 0.5 percent in the placebo group). Prophylactic apixaban (2.5 mg BID) is better than Lovenox (40 mg daily) at preventing postoperative VTE following hip replacement.⁹⁷ Following total knee replacement apixaban did not meet prespecified noninferiority criteria in ADVANCE-1, where VTE rates were similar in patients receiving apixaban (2.5 mg BID) and twice daily Lovenox (30 mg BID) (9.0 percent vs. 8.8 percent),⁹⁸ however in ADVANCE-2 apixaban was proven to be noninferior to once daily Lovenox in prevention of postoperative VTE following knee replacement.⁹⁹

EDOXABAN

Edoxaban is a direct oral factor Xa inhibitor with peak onset approximately 1 to 2 hours after administration and a half-life of approximately 8 hours.¹⁰⁰ Like other direct Xa inhibitors, edoxaban is renally excreted. Multiple coagulation tests can be affected by edoxaban, but partial thromboplastin time (PTT) and anti-Xa assay show the best correlation with drug levels.

Edoxaban is effective in management of atrial fibrillation and acute venous thromboembolism, and in the prevention of VTE following hip or knee arthroplasty. In the ENGAGE AF-TIMI trial, the annualized rate of stroke or systemic embolism in patients with nonvalvular atrial fibrillation and moderate to high risk of thrombosis was 1.50 percent with warfarin, 1.18 percent with high-dose edoxaban (P<0.001 for noninferiority), and 1.61 percent with low-dose edoxaban (P = 0.005 for noninferiority).¹⁰¹ Edoxaban-treated patients experienced significantly fewer major bleeding events (3.43 percent with warfarin vs. 2.75 percent with high-dose edoxaban and 1.61 percent with low-dose edoxaban). In the Hokusai-VTE study, edoxaban was noninferior to warfarin in preventing recurrent VTE in patients with acute PE or DVT (3.2 percent vs. 3.4 percent), and bleeding complications were reduced in the edoxaban treated group (8.5 percent vs. 10.3 percent).¹⁰²

Fibrinolytic therapy is administered by infusing high doses of a plasminogen activator to accelerate the conversion of plasminogen to the active fibrinolytic enzyme plasmin, which proteolytically degrades fibrin (Chap. 135). The specific biochemical and pharmacologic properties of different agents are important determinants of the administration regimen, the efficacy of clot lysis, and the nature of adverse effects. For example, some fibrinolytic drugs are bacterial products that are antigenic and can cause allergic responses, whereas others are recombinant human proteins. Some agents activate plasminogen prominently, both in blood and at the clot surface, and induce a systemic fibrinolytic state in addition to accelerating clot lysis. In contrast, the activity of other agents is more specifically limited to the clot surface with fewer systemic effects. Fibrinolytic therapy is used for treatment of both venous and arterial thrombosis and represents standard treatment for many patients presenting with acute myocardial infarction because it accelerates reperfusion, decreases mortality, and reduces morbidity (Chap. 135). Thrombolytic therapy has also become standard for many patients presenting with thrombosis of peripheral arteries, bypass grafts, and catheters.¹⁰³ It is used for treatment of selected patients with thrombotic stroke. Fibrinolytic therapy improves outcome in selected patients with large pulmonary emboli (Chap. 135).

STREPTOKINASE

Streptokinase was the first plasminogen activator used clinically. It is derived from β-hemolytic streptococci and has a unique indirect mechanism of action. By itself, streptokinase has no enzymatic activity, but it combines with plasminogen to form an equimolar streptokinase–plasminogen complex that can then convert other plasminogen molecules to plasmin. Additionally, the streptokinase–plasminogen complex can undergo proteolytic cleavage itself, resulting in activation. When administered in therapeutic doses, streptokinase is an effective thrombolytic agent. The streptokinase–plasmin(ogen) complex can bind to fibrin through the “kringle” domains of plasmin and activate clot-bound plasminogen to accelerate clot lysis (Chap. 135), but can also act on plasminogen in the blood to produce plasmin, giving rise to systemic proteolysis termed the lytic state. This results in consumption of plasminogen and α₂-antiplasmin, degradation of fibrinogen, factor V, and factor VIII, proteolysis of platelet membrane proteins by plasmin, and platelet activation. Streptokinase has a rapid plasma clearance with a half-life of approximately 20 minutes, but the duration of the proteolytic effect is more prolonged.¹⁰⁴

Streptokinase can be used to treat either venous or arterial thrombosis. Higher doses given over a shorter time are typically used for arterial disease. For either venous or arterial thrombosis, a sufficient dose must be administered to overcome circulating neutralizing antibodies, which are common because of the frequency of streptococcal infections in the population. Occasionally, individuals have a high titer of antibodies that neutralize this amount of streptokinase, resulting in resistance. Streptokinase is antigenic, and high-titer antibodies develop 1 to 2 weeks after use, precluding retreatment until the titer declines. High titers can also cause febrile or hypotensive reactions. The first large study to demonstrate the utility of coronary reperfusion employed streptokinase.¹⁰⁵ Although not widely used in North America, streptokinase is still extensively used given its low cost, widespread availability, and familiarity.

TISSUE-TYPE PLASMINOGEN ACTIVATOR AND RECOMBINANT TISSUE PLASMINOGEN ACTIVATOR (ALTEPLASE)

Tissue-type plasminogen activator (t-PA) is a naturally occurring plasminogen activator that is structurally and immunologically distinct from urokinase. t-PA is synthesized by endothelial cells as a single-chain polypeptide and was originally produced from cell culture for pharmacologic use, but is now synthesized by recombinant techniques (alteplase). t-PA directly converts plasminogen to plasmin in a reaction that is accelerated several-hundred-fold in the presence of fibrin. In the absence of fibrin, t-PA has much less activity, and this property accounts for the relative “fibrin specificity” of t-PA observed physiologically. However, when administered pharmacologically in a high dose, significant proteolysis of plasma fibrinogen often occurs, but this is typically less prominent than observed with treatment using either streptokinase or urokinase. The half-life of t-PA following intravenous administration is approximately 5 minutes, which requires a constant infusion to maintain therapeutic plasma levels. t-PA is not antigenic because it is a physiologic enzyme.¹⁰⁴

t-PA has been evaluated in treatment of VTE, myocardial infarction, stroke, catheter thrombosis, and peripheral arterial occlusion. In patients with PE, a regimen of 100 mg intravenously over 2 hours results in a high rate of clot lysis and hemodynamic improvement. t-PA has been evaluated in many large studies for acute myocardial infarction and administration results in improved mortality and morbidity. t-PA has also been evaluated in treatment of stroke and results in significant benefit in highly selected patients who are treated within hours of symptom onset (Chap. 135). In the CaVenT study,¹⁰⁶ use of alteplase in catheter-directed thrombolysis of ileofemoral DVT resulted in an 18 percent reduction in postthrombotic syndrome.

RETEPLASE

Recombinant technology has been used to engineer many t-PA mutants in an attempt to improve pharmacologic properties. The structural modifications in reteplase include removal of the finger, kringle 1, and EGF (epidermal growth factor) receptor domains. These changes result in enhanced fibrin specificity and a significantly longer half-life of 15 minutes compared to 4 minutes with t-PA, so it can be administered as an intravenous bolus rather than a continuous infusion. Its mode of administration (two 10-U IV boluses given over 2 minutes, 30 minutes apart) make it particularly useful for prehospital administration in remote areas, or areas with limited access to primary percutaneous coronary interventions.¹²⁹

TENECTEPLASE

Tenecteplase (TNK) is another bioengineered variant of t-PA with a longer half-life, increased resistance to inactivation by plasminogen activator inhibitor-1, and improved fibrin specificity. Advantages include a longer half-life, greater fibrin specificity, ease and rapidity of administration, and similar clinical efficacy as t-PA for treatment of acute myocardial infarction. It has a half-life of more than 30 minutes and can be administered as a single IV bolus. Large studies in patients with ST elevation myocardial infarction have shown it to be equivalently effective as other t-PA derivatives.¹⁰⁴ The PEITHO study examined the use of heparin with TNK or placebo in hemodynamically stable patients with large pulmonary emboli associated with right ventricular dysfunction. Thrombolysis resulted in a reduction in the primary end point of death or hemodynamic decompensation at 7 days following randomization (6 percent vs. 3 percent). The administration of thrombolytic agents was associated with increased bleeding complications, which was more prominent in elderly patients.

Platelets play an important role in hemostasis and thrombosis and inhibitors of platelet function are important therapeutic agents (Chap. 134). Platelets adhere to exposed subendothelium, become activated, release contents of their dense and α granules, and form aggregates. Additional platelets from the circulating blood are then recruited by adenosine diphosphate (ADP), which is released from dense granules, and also by thromboxane A₂ synthesized by activated platelets in the aggregate. Simultaneous with the initial platelet adhesion and aggregation, thrombin generation is initiated. The activated platelet phospholipid membrane is an effective surface for binding of coagulation factors to enhance the rate of thrombin generation. As thrombin is formed it activates additional platelets and also cleaves fibrinopeptides from fibrinogen to form fibrin in and around the platelet plug, consolidating it. The role of platelets in initiating thrombosis is greater in the arterial circulation than in the venous circulation because higher shear forces present in arteries activate platelets. Consequently, antiplatelet drugs are more effective in arterial than in venous thrombosis. Table 25–6 summarizes the types of drug, their use in clinical settings, their mechanism of action, and their dosages.

CYCLOOXYGENASE-1 INHIBITORS

Cyclooxygenase (COX)-1 is an enzyme that is present in most cells. It converts arachidonic acid released from membrane phospholipids by phospholipase A₂ or phospholipase C and diacylglycerol to prostaglandin (PG) G₂ (Chap. 112). A peroxidase converts PGG₂ to PGH₂, which is then converted by thromboxane synthase in platelets to thromboxane A₂. Thromboxane A₂ is a potent activator of platelets. In endothelial cells, PGH₂ is converted to prostacyclin, a potent inhibitor of platelet function, through an increase in intraplatelet cyclic adenosine monophosphate (cAMP).

Aspirin (acetylsalicylic acid) was recognized as an inhibitor of platelet function in the 1960s, although the mechanism of its action was unknown at that time. It prolonged the bleeding time in normal subjects slightly, although usually not out of the normal range, and its effect lasted for several days. It was demonstrated that acetylation of COX is important in platelet inhibition by aspirin. Because platelets cannot synthesize new COX, irreversible enzyme inhibition by aspirin means that inhibition persists for the life span of the platelet. Most cells have two forms of COX, known as COX-1 and COX-2. COX-1 is synthesized constitutively, whereas COX-2 is only synthesized under stress conditions. Both COX-1 and COX-2 are inhibited by aspirin and most nonsteroidal antiinflammatory drugs (NSAIDs), with aspirin acetylating both forms. The nonaspirin COX inhibitors are reversible inhibitors, so they are active only while in the circulation. It was thought initially that only COX-1 is found in platelets, but COX-2 has been detected in platelets and its effect is particularly apparent when there was a rapid platelet turnover. Because COX-1 is the major COX in platelets, COX-2–specific inhibitors have minimal effect on platelet function.

Aspirin and several of the commonly used NSAIDs (e.g., indomethacin, ibuprofen, and naproxen) have similar in vitro effects on platelet function. Platelet aggregometry demonstrates that the second wave of aggregation induced by ADP or epinephrine in citrated platelet-rich plasma (PRP) is abolished after aspirin ingestion and that aggregation induced by low concentrations of collagen is markedly decreased. Arachidonic acid–induced aggregation is abolished after aspirin ingestion. Additionally, secretion of dense granule components (ADP, ATP, and serotonin) and of α-granule proteins by ADP, epinephrine, collagen, and arachidonic acid is inhibited in PRP after aspirin ingestion or with addition of indomethacin to PRP. Because of these in vitro effects of aspirin, the drug has been used extensively as an inhibitor of platelet function in vivo, with beneficial effects in primary and secondary prevention and in treatment of myocardial infarction (Chap. 135). Aspirin is also beneficial in stroke prevention with carotid artery disease and embolic stroke, although anticoagulation with warfarin or its analogues is generally more effective than aspirin in embolic stroke in most patients with a cardiac embolic source.¹⁰⁷ Aspirin is less often used to prevent venous thrombosis, although two recent large randomized studies showed significant benefit for secondary prevention of recurrent VTE compared to placebo.^108,109 Other drugs that inhibit COX-1 are not used to prevent either arterial or venous thrombosis.

A daily dose of 81 to 325 mg is recommended for most indications, as lower-dose aspirin appears as effective and may be associated with a lower risk of gastrointestinal bleeding than higher doses.^110,111 Broadly, aspirin is currently recommended for primary and secondary prevention of a wide variety of atherosclerotic outcomes including stroke, myocardial infarction, and peripheral vascular disease.

DRUGS THAT MODULATE CYCLIC ADENOSINE MONOPHOSPHATE LEVELS

cAMP in platelets is formed from ATP by the action of adenylate cyclase and degraded by cAMP phosphodiesterase, and basal levels of cAMP in platelets are low. Elevated levels of intraplatelet cAMP are induced by inhibition of cAMP phosphodiesterase, or by stimulation of adenylate cyclase activity, resulting in inhibition of platelet activation through several pathways: (1) modulation of phosphorylation of specific proteins; (2) inhibition of several steps in metabolism of phosphoinositol phosphates; and (3) lowering of intracellular Ca²⁺, and accumulation of Ca²⁺ by platelet microsomes. Agents that inhibit the cAMP phosphodiesterase include theophylline, papaverine, and dipyridamole, as well as pentoxifylline and cilostazol. Several prostaglandins stimulate adenylate cyclase, including PGE₁, PGD₂, and PGI₂ (prostacyclin). Drugs that elevate cAMP levels are dipyridamole, pentoxifylline, and cilostazol. Dipyridamole can be used alone or in combination with aspirin. A very large study of dipyridamole in combination with low-dose aspirin (25 mg) found the combination equivalently effective to clopidogrel for the secondary prevention of noncardioembolic stroke.¹¹² Recent systematic reviews have also suggested that the combination of dipyridamole and aspirin is superior to aspirin alone for the prevention of cerebrovascular events.¹¹³

The other two phosphodiesterase inhibitors (pentoxifylline and cilostazol) are used primarily in patients with peripheral vascular disease. In addition to their inhibitory effect on platelets they may exert a beneficial effect on blood rheology and the microcirculation by increasing red cell deformability, thereby reducing blood viscosity. Cilostazol increases vascular endothelial growth factor levels, which may lead to an increase in collateral circulation. It has been shown to reduce risk of stroke in Asian populations,^114,115 and increases walking distance in patients with peripheral vascular disease.¹¹⁶ Pentoxifylline inhibits vascular smooth-muscle cell proliferation and collagen synthesis, which may enhance vasodilation. Pentoxifylline probably is an effective treatment for ulcers associated with peripheral vascular disease; however, it is only modestly effective for treatment of peripheral vascular disease.¹¹⁷

ADENOSINE DIPHOSPHATE RECEPTOR BLOCKERS

The third class of platelet inhibitors is the ADP receptor blockers, which include thienopyridines (ticlopidine, clopidogrel, and prasugrel) and nonthienopyridines (ticagrelor and cangrelor). There are three ADP receptors on platelet membranes (Chap. 112), with the thienopyridines inhibiting one of them, the P2Y12 receptor. The inhibition of binding of ADP to the P2Y12 receptor results in inhibition of adenylate cyclase.

Ticlopidine was available for clinical use before clopidogrel. The Canadian American Ticlopidine Study was a randomized, placebo-controlled, double-blind study showing a 30 percent risk reduction for recurrent cardiovascular events with ticlopidine.¹¹⁸ A review of four trials of ticlopidine plus aspirin versus oral anticoagulants for coronary stenting showed benefit to the combination in terms of reduced risk of nonfatal myocardial infarction and revascularization at 30 days, combined negative events (mortality, myocardial infarction, revascularization at 30 days), and major bleeding, but increased the risk of thrombocytopenia and neutropenia.¹¹⁹

In clinical practice fear of toxicity has largely led ticlopidine to be abandoned in favor of clopidogrel. In a direct comparison of ticlopidine and clopidogrel in patients undergoing coronary stenting (CLASSICS trial), clopidogrel was associated with a significantly lower rate of major adverse events (4.6 percent vs. 9.1 percent).¹²⁰ The first clinical trial of clopidogrel was the Clopidogrel Versus Aspirin in Patients at Risk of Ischemic Events (CAPRIE) trial, a large randomized, blinded trial of clopidogrel versus aspirin in 19,000 patients at risk of ischemic events.¹²¹ Patients were enrolled after recent myocardial infarction (MI) or stroke, or if they had symptomatic peripheral arterial disease. The primary outcome was the occurrence of ischemic stroke, MI, or vascular death. With a mean followup of 1.91 years, there was a relative risk reduction of 8.7 percent in the clopidogrel group (p = 0.043). No major differences were noted in terms of safety. Clopidogrel is also used in acute coronary syndromes, based on studies like the Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE) study, which showed a significant reduction in the combined end point of cardiovascular death, nonfatal MI, or stroke with clopidogrel and aspirin versus aspirin alone.¹²² Several studies have demonstrated the effectiveness of clopidogrel in patients undergoing percutaneous coronary intervention, and a recent meta-analysis of several large studies showed that clopidogrel treatment prior to intervention is associated with decreased incidence of major cardiac events (MI, stroke, urgent revascularization) compared to treatment after intervention.¹²³ The degree of platelet inhibition after clopidogrel therapy varies. Larger loading doses of clopidogrel appear to reduce variability in response; the safety and efficacy of such doses are being compared in ongoing studies.¹²⁴

Prasugrel is a “third-generation” P2Y12 blocking agent. Unlike clopidogrel it can be converted to its active metabolite via esterases present in either the liver or the gut. Like clopidogrel it irreversibly blocks the P2Y12 receptor.¹⁴⁴ Evidence for the use of prasugrel comes predominately from one large study of prasugrel compared with clopidogrel. This study randomized 13,608 patients with moderate-to-high-risk acute coronary syndromes and who were scheduled to undergo percutaneous coronary intervention to receive prasugrel (a 60-mg loading dose and a 10-mg daily maintenance dose) or clopidogrel (a 300-mg loading dose and a 75-mg daily maintenance dose) for up to 15 months.¹²⁵ Although prasugrel significantly reduced death from cardiovascular causes, nonfatal MI, and nonfatal stroke, it significantly increased all forms of bleeding, including major and fatal hemorrhage.

Ticagrelor is a member of the cyclopentyltriazolopyrimidines chemical class, a group of agents that reversibly inhibit the P2Y12 platelet receptor. Ticagrelor does not require hepatic activation and has a more rapid onset of action than clopidogrel. The PLATO trial was a large phase III randomized study comparing ticagrelor and aspirin (180 mg loading dose followed by 90 mg twice daily) to clopidogrel (300 to 600 mg dose followed by 75 mg daily) and aspirin for treatment of acute coronary syndrome.¹²⁶ Patients receiving ticagrelor experienced the composite primary end point (death from vascular causes, MI, or stroke) less often than patients receiving clopidogrel (9.8 percent vs. 11.7 percent, hazard ratio [HR] 0.84) without a significant difference in the rates of major bleeding (11.6 percent vs. 11.2 percent).

Cangrelor, the first parenteral ADP receptor blocker, is not currently available for clinical use, but has been tested in three large randomized clinical trials of patients undergoing intracoronary stent implantation procedures. Advantages of this agent include parenteral administration, very rapid onset, and short half-life. The CHAMPION studies (PLATFORM,¹²⁷ PCI,¹²⁸ and PHOENIX¹²⁹) compared cangrelor to clopidogrel or placebo in patients undergoing coronary interventions. Although only the PHOENIX trial showed a significant reduction (4.7 percent vs. 5.9 percent) in the 48-hour composite end point (death, MI, ischemia-driven revascularization, or stent thrombosis), in a pooled analysis of patient level data from the three CHAMPION trials cangrelor use was associated with a significant reduction in the primary efficacy end point compared to control (3.8 percent vs. 4.7 percent, OR 0.81).¹³⁰

α_IIbβ₃ BLOCKERS

Fibrinogen binds specifically and saturably to the surface of activated platelets, and the α_IIbβ₃ complex is the fibrinogen receptor. This complex mediates platelet aggregation induced by all physiologic agonists. Fibrinogen binds only to the activated conformation of the receptor α_IIbβ₃. Monoclonal antibodies have been developed against α_IIbβ₃ complex on resting or activated platelets, which prevent aggregation by blocking ligand binding. Abciximab, is the Fab′₂ fragment of a chimeric mouse–human antibody. Initial human pharmacodynamic studies were performed in patients with unstable angina and in patients undergoing high-risk coronary angioplasty, and dose-related inhibition of platelet function was found. No spontaneous bleeding was observed, despite prolongation of the template bleeding time. Because of the mouse component of abciximab, it may induce antimouse antibodies, preventing repeated use in patients.

The first large clinical trial of abciximab was the Evaluation of c7E3 for the Prevention of Ischemic Complications (EPIC) trial,¹³¹ published in 1994, in which the drug was used in patients with high-risk coronary angioplasty. Abciximab reduced ischemic events after angioplasty when given together with heparin and aspirin, but it also increased the risk of bleeding. Subsequent studies of patients undergoing percutaneous coronary intervention, the Evaluation in PTCA to Improve Long-term Outcome with Abciximab GP IIb/IIIa (α_IIbβ₃) Blockade (EPILOG) study¹³² and EPISTENT trial,¹³³ demonstrated efficacy in both low-risk and high-risk patients without any increase in major bleeding. Abciximab has also been tested in patients with acute ischemic stroke, but resulted in significant increase intracranial hemorrhage without any clinical efficacy so it has not been further pursued for this clinical indication.¹³⁴

Other types of inhibitors of fibrinogen binding to platelets have also been developed. Those in clinical use are eptifibatide, a cyclic heptapeptide based on a rattlesnake venom peptide, and tirofiban, a nonpeptide derivative of tyrosine. Pharmacokinetic and pharmacodynamic studies in animals and humans showed a rapid onset of action, short plasma half-life, and rapid reversibility of action. The pharmacodynamics of eptifibatide are substantially altered by anticoagulants that chelate calcium, and pharmacokinetic modeling suggests that optimal dosing is obtained by giving a second bolus 10 minutes after the first bolus. Eptifibatide is not immunogenic.

The first major clinical trial of eptifibatide was the Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis (IMPACT) II trial in patients undergoing any kind of coronary intervention.¹³⁵ There was a highly significant reduction in the composite end point of death, MI, coronary artery bypass grafting, repeat urgent or emergent coronary intervention, or stent placement for abrupt closure at 24 hours with both eptifibatide dosing arms. There was no increase in major bleeding. The effect was no longer significant at 30 days on intention-to-treat analysis.

Animal studies with tirofiban were performed in dogs. Dose-dependent inhibition of ex vivo platelet aggregation was achieved, with rapid reversibility at the end of the infusion. Electrically induced coronary artery thrombosis was markedly reduced by tirofiban infusion, without significant extension of the bleeding time. Pharmacokinetic and pharmacodynamic studies in humans showed that tirofiban provided a well-tolerated reversible means of inhibiting platelet function. Bleeding time was prolonged, and ADP-induced aggregation was blocked by at least 80 percent in normal volunteers. The plasma half-life was 1.6 hours. ADP- and collagen-induced platelet aggregation in normal volunteers returned to 55 percent and 89 percent of baseline, respectively, by 3 hours after the end of infusion. Similar results were found in a dose-ranging study in patients undergoing coronary angioplasty. Based on these results tirofiban has been extensively studied as an adjunct to therapies for patients with, or at risk of, acute coronary syndromes. Clinical results for patients treated with tirofiban have been reviewed.¹³⁶

THROMBIN RECEPTOR BLOCKERS

Platelet activation occurs through a variety of cell surface receptors including the thrombin receptor, a potent platelet activator mediated by its binding to and cleaving the protease-activated receptor (PAR)-1. Vorapaxar is an irreversible PAR-1 thrombin receptor antagonist, which has been studied in two large phase III randomized studies. In the TRACER study 12,944 patients with non–ST elevation acute coronary syndromes were randomized to vorapaxar or placebo plus standard care.¹³⁷ Vorapaxar use was associated with a significant reduction in death from cardiovascular causes (14.7 percent vs. 16.4 percent), but was also associated with a significant increase in major bleeding events, including intracranial hemorrhage (1.1 percent vs. 0.2 percent). The TRA 2P-TIMI 50 study randomized 26,449 patients with prior MI, stroke, or peripheral artery disease to vorapaxar versus placebo.¹³⁸ This phase III study also demonstrated clinical efficacy with a reduction in death from cardiovascular cause (11.2 percent vs. 12.4 percent), but it was at the expense of increased major bleeding complications. Efforts are underway to identify subgroups that may have the greatest net benefit from vorapaxar. Atopaxar is a low-molecular-weight inhibitor of PAR-1, which has been studied in phase II trials of patients with acute coronary syndrome or high-risk coronary artery disease.¹³⁹ These studies showed an increase in minimal bleeding events (16.4 percent vs. 4.5 percent) but no difference in clinically significant bleeding.