Introduction
The optimal anticoagulant during the treatment of chronic limb ischemia (CLI) is largely unknown. For the past several years, unfractionated heparin (UFH) has been the main anticoagulant utilized by most endovascular specialists in treating CLI. Multiple trials in the coronary literature have shown that UFH is less effective or safe than the direct thrombin inhibitor bivalirudin^1–8 or the low-molecular weight heparin enoxaparin.⁹ In this study, we compare the in-hospital complications of various intraprocedural anticoagulants utilized in the treatment of CLI patients by retrospectively reviewing a series of consecutive patients treated at our institution.

Methods
We retrospectively reviewed 65 consecutive patients with CLI treated percutaneously by 3 interventionalists at our institution (March 1998 to November 2004, with 51/65 [78%] procedures performed from 2002 to 2004). All patients were in the Rutherford-Baker classifications IV (rest limb pain) and V (limb ulceration). Patients were divided into three groups: UFH (adjusted per operator to keep activated clotting time [ACT] > 250 seconds), bivalirudin (REPLACE-2 dosing⁷) and enoxaparin (0.75 to 1 mg/kg intravenous bolus). Adjunctive use of glycoprotein (GP) IIb/IIIa inhibitors (integrilin and abciximab) and/or fibrinolytics (tissue plasminogen activator and reteplase) was recorded for all patients. No ACT was measured for bivalirudin, which typically provides a very reliable and predictable anticoagulant response within a few minutes of the bolus.¹³ Demographic and clinical differences among the various anticoagulants were determined. The use of clopidogrel prior to the procedure was recorded. In-hospital complications were defined as death, stroke, major bleeding (loss of > 3 gm of hemoglobin with a source of bleed, bleed including puncture site hematoma, retroperitoneal bleeding and/or intracranial bleeding), vascular access site complication (pseudo-aneurysm or arteriovenous fistula), unplanned major amputation and renal failure (rise of creatinine by > 0.5 from preprocedure levels). Descriptive analysis was performed on all variables (Table 1). A complication-free procedure was determined for the different anticoagulants and compared using Fisher’s exact test.

Results
The three anticoagulant groups were as follows: unfractionated heparin (UFH) (with 3 patients receiving adjunctive GP IIb/IIIa inhibition) (n = 17), enoxaparin (with 2 patients receiving adjunctive GP IIb/IIIa inhibition) (n = 22) and bivalirudin (with 8 patients receiving adjunctive GP IIb/IIIa inhibition and 3 patients receiving adjunctive fibrinolytic therapy plus GP IIb/IIIa inhibition) (n = 26). Clopidogrel 300 mg po load was administered immediately prior to or during the procedure in 73.4%, 76.2% and 62.5% of the patients in the UFH, enoxaparin and bivalirudin groups, respectively, followed by 75 mg po daily. The remainder of the patients received their clopidogrel immediately post procedure.

A complication-free procedure occurred in 76.4%, 90.9% and 84.6% of patients in the UFH, enoxaparin and bivalirudin groups, respectively (p = NS). Major bleeding occurred in 5.9%, 0% and 3.8% of patients in the UFH, enoxaparin and bivalirudin groups, respectively (p = NS). Bivalirudin bleeding rates remained very low despite 30.8% of patients receiving adjunctive GP IIb/IIIa inhibition and/or 11.5% receiving adjunctive fibrinolytic treatment.

Discussion
UFH has been the anticoagulant of choice in PPI. UFH, however, is an unpredictable anticoagulant, activates platelets and continues to result in thrombin activation during PPI.¹⁰ Data suggest that higher doses of this drug are associated with an increase in major adverse events in patients undergoing PPI.¹¹ Numerous studies in percutaneous coronary interventions (PCI) have shown that UFH is inferior to the direct thrombin inhibitor bivalirudin^1–8 or possibly the low-molecular weight enoxaparin.⁹ The role of GP IIb/IIIa inhibitors in PPI is still poorly defined, but its role along with bivalirudin in chronic limb ischemia appears promising.¹²

Bivalirudin has been shown to offer predictable anticoagulation in percutaneous interventions, including PPI, in contrast to UFH.^13–18 Data from the the Angiomax Peripheral Procedure Registry Of Vascular Events (APPROVE) trial has shown a low rate of ischemic events (1.4%) or TIMI major hemorrhage (0.4%) with bivalirudin use as a base anticoagulant.¹³ The APPROVE registry, however, included low-risk peripheral vascular patients. Allie et al¹² presented data on the use of bivalirudin and GP IIb/IIIa inhibitors in PPI in patients with chronic limb ischemia, as compared to a historic control with heparin alone. In this nonrandomized study, there was a trend toward lower 6-month reintervention and limb salvage rates (10.7% versus 18.8%; p = 0.0501 and 93.9% versus 88.5%; p = 0.053) and less distal embolization (1.3% versus 5.4%) in the bivalirudin-GP IIb/IIIa inhibitor group compared to the UFH group, respectively. In our study, a complication-free procedure in CLI patients was achieved in 84.6% of patients receiving bivalirudin and 90.9% of patients receiving enoxaparin versus 76.4% in the UFH group (p = NS). Bivalirudin and enoxaparin appear safe and did not result in an apparent increase in complication rates in CLI patients. In contrast, an apparent reduction in complications was noted with these drugs compared to UFH, and this reduction, although not statistically significant, is probably clinically important. Whether the use of bivalirudin is cost effective in PPI is unknown at this time and can only be ascertained from randomized trials, particularly in the high-risk population, such as the limb ischemia patients.

In conclusion, this study suggests that UFH carries an increased rate of complications in the percutaneous intervention of CLI patients. Bivalirudin and enoxaparin appear safe and effective in this population group, with a trend toward less complications than with UFH. Whether the use of bivalirudin or enoxaparin truly reduces the incidence of procedure-related complications in CLI patients is unknown and needs to be proven in a multicenter, randomized trial in this high-risk group.

Limitations of the study
This study is limited by its small number of patients and its retrospective design. The three anticoagulant groups were, however, reasonably balanced. This study is not powered to compare outcomes between the various anticoagulants in the treatment of CLI patients, but to demonstrate the feasibility of these anticoagulants in the treatment of this high-risk population.