Direct Versus Indirect Thrombin Inhibition in Percutaneous Coronary Intervention (Part II)

Continued from previous page Randomized trials. A retrospective analysis of randomized studies supports the findings of the case-control studies. In 1,863 heparin-treated patients, there was an inverse relation between the ACT values measured five minutes after an initial heparin bolus and the risk of abrupt closure. For every 10-second increase in ACT, the probability of abrupt closure decreased by 1.3%. The relationship between the maximal ACT and the risk of abrupt vessel closure was most striking in patients undergoing brief procedures (odds ratio, 1.6% for every decrease of 10 seconds in maximal ACT).⁵⁴ A pooled analysis of recent major clinical trials in patients treated with heparin also provides evidence that higher ACT values ranging from 350–375 seconds result in greater protection against ischemic events.²⁷ In this analysis, data from 6 contemporary randomized controlled trials of adjunctive antithrombotic regimens for PCI where UFH and aspirin constituted the control arm were analyzed using similar definitions for ischemic endpoints. In a total of 5,216 patients, the incidence of death, myocardial infarction, or any revascularization at 7 days was calculated for each ACT group. An ACT in the range of 350–375 seconds provided the lowest composite ischemic event rate of 6.6%, or a 34% relative risk reduction in 7-day ischemic events, compared to a rate of 11.6% observed between 171–295 seconds (p = 0.001) (Figure 3). The current trend toward lower heparin dosing to achieve lower ACT values of approximately 250–275 seconds may not provide optimal effectiveness in the absence of GP IIb/IIIa receptor inhibitors. However, in the presence of a GP IIb/IIIa inhibitor (abciximab), this pooled analysis did demonstrate that lower ACT values are beneficial, resulting in less bleeding risk and no increased risk in ischemic complications (Figure 4). This same pooled analysis demonstrated that in order to achieve the lowest rate of bleeding, in the absence of a GP IIb/IIIa inhibitor, ACTs must be kept in a lower range (325–350 seconds) than the range to prevent ischemic complications (350–375 seconds), underscoring the fact that reductions in ischemic effects when targeting for maximal efficacy with heparin is a strategy that is associated with a cost of increased bleeding. This inability to achieve a low rate of both ischemic and hemorrhagic complications underscores an important limitation in the management of patients undergoing PCI. The excess in thrombotic risk at higher ACTs observed in the analysis by Chew et al. may represent a consequence of platelet activation that has been reported in connection with high doses of unfractionated heparin.^24,55,56 Heparin dosing during PCI in conjunction with a GP IIb/IIIa antagonist. The trend in recent years has been to reduce the dose of heparin during PCI, particularly in patients treated with GP IIb/IIIa antagonists. Data from the EPIC trial57 suggested an unfavorable relationship between the prevailing heparin dose strategy (fixed dose of 10,000–12,000 U bolus IV) and bleeding in patients who were treated with abciximab. The odds ratio for major bleeding risk was 1.22 (95% CI, 0.28–5.38) for patients receiving 120 U/kg. Patients who received relatively high doses of heparin did not experience a reduction in the rate of peri-procedural ischemic events.58 This observation prompted the initiation of the PROLOG⁵⁹ and EPILOG⁶⁰ studies, which established that reduction of the heparin dose from 100 U/kg to 70 U/kg in patients receiving abciximab lowered the rate of hemorrhage [non-CABG related major bleeding, 1.9% versus 1.1% (p = 0.7); minor bleeding, 7.6% versus 4.0% (p 61 Findings from EPILOG suggested that patients with diabetes may represent an exception to this finding; event rates were lowest in diabetics receiving abciximab combined with standard dose heparin (100 U/kg).⁶² However, significant event rate reductions among diabetics assigned to abciximab with low-dose, weight-adjusted heparin in the subsequent EPISTENT trial⁶³ cast doubt on the applicability of these observations. In the ESPRIT trial,⁶⁴ the heparin dose used in patients receiving eptifibatide was 60 U/kg. No relationship between ACT and the frequency of ischemic events was observed in patients receiving the GP IIb/IIIa antagonist.⁶⁴ Therefore, experience from a number of studies using a variety of GP IIb/IIIa inhibitors suggests that among patients who are treated with a GP IIb/IIIa antagonist, reduction of the heparin dose to achieve an ACT of just over 200 seconds is necessary to avoid an increase in hemorrhagic events.65 However, the preponderance of evidence suggests that this is not the case when PCI is performed in the absence of a GP IIb/IIIa antagonist. Low molecular weight heparins Low molecular weight heparins have become increasingly popular within the last several years, in large part as a consequence of two trials suggesting superior outcomes with enoxaparin compared to UFH in the medical management of patients with unstable angina and non-Q wave myocardial infarction.^66,67 Pooled analysis of the two studies for major and minor bleeding events, however, showed no reduction of bleeding for enoxaparin versus heparin (major, 1.3% enoxaparin versus 1.1% heparin; minor, 10% enoxaparin versus 4.3% heparin); if a PCI was required during either of these trials, enoxaparin was discontinued prior to the procedure. Consequently, the use of LMWH during PCI was not extensively studied. To assess whether LMWH can be used in PCI, subsequent registry studies have investigated the use of enoxaparin in patients undergoing PCI. In NICE-1 (enoxaparin, IV bolus of 1.0 mg/kg)68 and NICE-4 (enoxaparin, 0.75 mg/kg plus abciximab),69 levels of anti Xa activity were similar to levels reported in patients treated with a 10,000 U UFH bolus.⁶⁹ The rates of ischemic and bleeding complications were lower than historical controls from the EPILOG low-dose population, leading to the conclusion that enoxaparin, with or without abciximab, appeared to provide safe and effective anticoagulation during PCI. Similar results have been reported for the NICE 3 registry, in which ACS patients receiving enoxaparin were treated with one of the three available GP IIb/IIIa antagonists. Although the study suggests that it is feasible to extend the use of enoxaparin to the cath lab during PCI, no conclusions can be drawn about the benefits of using enoxaparin during the PCI procedure compared to UFH.⁷⁰ Since the data in these open-label registries were compared to historical controls and comparative trials have not been performed, additional prospective studies are needed to establish the relative safety and efficacy of enoxaparin in PCI. In one randomized study where LMWH (reviparin) was compared to UFH, reviparin did not reduce the composite of ischemic events at 6 months.⁷¹ Reviparin administered at 7,000 U bolus followed by infusion and subcutaneous injection was compared with UFH at 10,500 U bolus followed by infusion in 612 patients without the use of GP IIb/IIIa antagonists and with stent implantation performed only for salvage purposes. The rate of rescue stent implantation on the day of PCI was reduced from 6.9% to 2.0% with reviparin.⁷¹ However, because of the lack of benefit at 6 months, the product was not further developed for this indication. Direct thrombin inhibitors as a potential alternative to heparin in PCI Currently available direct thrombin inhibitors include bivalirudin, hirudin and argatroban. Each of these inhibitors binds directly to thrombin without the need for the cofactor antithrombin and are effective at inhibiting both fibrin-bound and free circulating thrombin.⁹ Hirudin and bivalirudin are bivalent inhibitors that bind to thrombin at exosite 1 (the substrate recognition site) and at the active site.^72–74 In contrast, argatroban binds thrombin only at its active site in a univalent fashion. Hirudin and argatroban have received clinical approval for treatment and management of patients with heparin-induced thrombocytopenia. Bivalirudin is the only direct thrombin inhibitor that is indicated for use in patients with unstable angina undergoing percutaneous coronary angioplasty. Lepirudin. Hirudin, a 65-amino acid polypeptide originally isolated from the salivary glands of the medicinal leech Hirudo medicinalis, is now available through recombinant DNA technology. Recombinant hirudin (lepirudin) binds tightly to thrombin to form a slowly reversible complex.⁷³ Lepirudin is cleared via renal mechanisms and has a half-life of approximately 50–60 minutes (increasing up to 3 hours depending on patient characteristics and disease states).⁷⁵ There have been no randomized trials of lepirudin in HIT patients undergoing PCI, although several lepirudin-treated patients in the clinical studies HAT-1 and HAT-2 also underwent successful PCI. In 2 trials of ACS (without ST-elevation myocardial infarction),^76,77 clinical thrombotic outcomes were better with lepirudin, but the risk of major bleeding was substantially increased. A recent analysis of the larger of the two studies has shown that the increased risk of bleeding may relate to the development of thrombocytopenia with hirudin (0.9%), which was similar to the incidence reported for heparin (1.1%).⁷⁵ Desirudin has been studied in 3 large trials of acute coronary syndromes (ACS), including a PCI trial,^78,79 but at present, desirudin is not approved for any use in the United States. Results from 1 trial demonstrated significant reductions in the incidence of acute ischemic episodes following PCI.⁷⁸ This improvement over heparin was particularly evident among patients considered to be at high risk for ischemic complications, but bleeding complication rates were not reduced. The finding of improved efficacy of hirudin over heparin (but with an increased risk of bleeding) is consistent across many studies, including those in acute myocardial infarction^79–81 and ACS without ST elevation.^79,82,83 Argatroban. Argatroban is a synthetic derivative of arginine. The drug binds reversibly to the catalytic site of thrombin with a terminal elimination half-life of 39–51 minutes. The primary route of clearance is liver metabolism. Unlike the selective bivalent direct thrombin inhibitors lepirudin and bivalirudin, 54% of the argatroban dose given binds to human serum proteins, with binding to albumin and a 1-acid glycoprotein at 20% and 34%, respectively.⁸⁴ In a prospective study with historical controls, the composite endpoint of all-cause death, all-cause amputation, or new thrombosis was reduced from 38.8% in the historical control group to 25.6% in argatroban-treated patients. Bleeding was similar in both groups (~ 2%) and hematuria occurred in ~ 12% of argatroban patients compared to ~ 1% of the historical control group.⁸⁵ Argatroban has been evaluated for PCI in 2 trials of patients with HIT (total, n = 50).^86,87 Of the patients who underwent PCI, > 98% had a successful procedure. The significant complications in patients given argatroban included 1 retroperitoneal hematoma and 1 abrupt vessel closure that required bypass surgery. The anticoagulant effect of the drug appeared to be adequate for all but 1 patient.^86–88 Based on historical controls, argatroban appears to provide adequate anticoagulation but little bleeding advantage over UFH in clinical trials to date. Bivalirudin. Bivalirudin is a synthetic, 20-amino acid polypeptide modeled after hirudin and comprised of an active site-directed peptide, D-Phe-Pro-Arg-Pro, linked via a tetraglycine spacer to a dodecapeptide analogue of the carboxy-terminal of hirudin (Figure 5). Bivalirudin interacts with thrombin at both the active site and exosite 1, forming a 1:1 stoichiometric complex. Once bound, however, thrombin slowly cleaves the Arg3-Pro4 bond of bivalirudin, resulting in recovery of thrombin’s active site function.⁹ Bivalirudin’s carboxy-terminal dodecapeptide remains bound to exosite 1, but with low affinity, such that bivalirudin initially acts as a non-competitive inhibitor of thrombin, then becomes a competitive inhibitor enabling thrombin to subsequently participate in hemostatic reactions.⁹ Thus, the biochemical properties of bivalirudin confer to the molecule a unique combination of features that appear favorable for PCI patients including: 1) the ability to inhibit thrombin regardless of its location (fibrin-bound versus soluble thrombin), a property of potential importance to patients with visible thrombus on angiography prior to stent placement; 2) rapid return to hemostatic competence upon cessation of infusion (i.e., ultra-short plasma and biologic half-life) due to vulnerability of bivalirudin at active site; 3) the ability to profoundly inhibit thrombin-mediated platelet aggregation. In vitro, thrombin-mediated platelet aggregation is fully inhibited at 1/500th of the plasma concentrations typically achieved in PCI;89 and 4) the ability to achieve high levels of anticoagulation without causing platelet activation. On the contrary, given thrombin’s central importance to platelet activation,90,91 it is likely that by virtue of its antithrombin effect bivalirudin attenuates platelet activation during PCI. The safety and efficacy of bivalirudin in PCI were evaluated in 2 trials. In the first, bivalirudin was used as an alternative to heparin in a dose-finding trial of 291 patients undergoing angioplasty.⁹² Bivalirudin produced dose-dependent prolongation of the ACT and aPTT. In the second study, the Bivalirudin Angioplasty Trial (BAT), a total of 4,312 patients with unstable angina or MI requiring PCI were randomized in a double-blind fashion to receive bivalirudin or heparin during the procedure.⁹³ Evaluation of the adjudicated data demonstrated a 22% relative risk reduction with bivalirudin in the rate of death, MI, or repeat revascularization at 7 days compared to heparin (7.9% versus 9.3%, respectively; p = 0.038) and significantly less major bleeding (3.5% versus 9.3%, respectively, a 62% relative reduction; p 94 This unique reduction of both ischemic and hemorrhagic events was even more pronounced in a prespecified and randomized subgroup of post-MI patients where those treated with bivalirudin had a 51% relative risk reduction in death, MI or revascularization versus heparin (4.9% versus 9.9%, respectively; p = 0.009) and a relative risk reduction of 73% for major bleeding (2.4% versus 11.8%, respectively; p 94 Clinically significant hemorrhage was defined as overt bleeding with a fall in hemoglobin level of 3 g/dl or more, or transfusion of >= 2 units of blood, or retroperitoneal or intracranial bleeding. The reduced rate of hemorrhage in patients treated with bivalirudin compared to heparin-treated patients was found at all ACT levels (Figure 6A). This suggests that bivalirudin may reduce bleeding even in those settings where heparin doses are titrated to lower levels of ACT. Similarly, the incidence of the combined endpoints of death, MI, or repeat revascularization was lower in bivalirudin-treated patients compared to those treated with heparin across the ACT levels (Figure 6B). To further substantiate the claim that bivalirudin results in a reduction in bleeding complications in comparison to heparin independent of the precise dose of heparin administered, the incidence of major bleeding utilizing TIMI criteria was calculated for both the BAT and the heparin arm of a series of contemporary PCI trials. Double-blind, randomized clinical trials of anticoagulant agents used in PTCA performed predominantly in North America and coordinated by The Cleveland Clinic Foundation and/or Duke Clinical Research Institute between 1993 and 1998 were included in the pooled analysis alongside the Bivalirudin Angioplasty Trial. The selection criteria excluded one US-based clinical trial (RESTORE) and a small number of clinical trials performed predominantly outside the US (HELVETICA, REDUCE) as well as trials conducted after 1998. The TIMI major hemorrhage rate in the heparin arm of the BAT trial was well within the range of the rate observed in the heparin-only control arm of the more recent PCI trials in this pooled analysis (Figure 7). A reduced rate of bleeding risk among patients treated with bivalirudin has been reported in other trials as well. In a systematic overview (meta-analysis) of clinical outcomes that included three additional trials comparing bivalirudin to heparin, Kong et al. reported a significant reduction in major hemorrhage (odds ratio, 0.41; 95% CI, 0.32–0.52; p 95 Finally, bivalirudin may provide an attractive alternative to UFH for the treatment of coronary patients referred for catheterization after receiving subcutaneous LMWH. In a small, open-labeled, randomized trial of 40 patients treated with LMWH for ACS, patients were switched to either UFH or bivalirudin prior to angioplasty. Coagulation parameters and adverse effects were assessed. In these patients, the last dose of LMWH was administered at least 8 hours prior to randomization. The increases in ACT values were more predictable, and the return to normal ACT levels was more rapid in bivalirudin-treated patients than in heparin-treated patients. The median aPTT was higher with UFH than with bivalirudin and remained elevated at sheath removal time (89 seconds versus 44 seconds). Consistent with these observations, anti-Xa levels remained persistently elevated up to the time of sheath removal in UFH-treated patients. In contrast, anti-Xa levels in patients treated with bivalirudin declined steadily throughout the study period. Adverse events were comparable between groups. One patient in the UFH group had a post-procedure MI; no deaths or major hemorrhages occurred in either group. A complete report of this study awaits publication. Summary. Heparin remains the most widely used anticoagulant in interventional cardiology, but is limited by its inability to inhibit clot-bound thrombin and by its propensity to activate platelets. Furthermore, when attempts are made to overcome these limitations by increasing the dose of heparin, bleeding complications ensue. LMWH represents a theoretic advance in antithrombotic therapy by virtue of its greater pharmacokinetic predictability and reduced propensity to stimulate platelet aggregation. However, with respect to PCI, the extended half-life of LMWH (particularly when administered subcutaneously) relative to UFH represents a major theoretic disadvantage. Direct thrombin inhibitors, such as bivalirudin, that possess shorter half-lives and potentially more potent antithrombotic activity than UFH in a broad range of coronary lesion morphologies may provide interventionalists agents with more desirable pharmacokinetic profiles, particularly in an era characterized by shorter procedure times due to improvements in stent technology.