Your patient has atrial fibrillation, chest pain and a positive stress test with risk factors for coronary disease, and is planning to undergo cardiac catheterization. The international normalized ratio (INR) is 2.2. When is it safe to perform coronary angiography on this individual?

It is estimated that 5% of patients undergoing coronary angiography are on long-term warfarin therapy for a variety of clinical indications. Cardiac catheterization in a patient on warfarin is thought to be associated with an increased risk of bleeding and access site complications, especially after percutaneous coronary intervention (PCI). Generally, it is taught that arterial puncture for cardiac catheterization should not proceed until the INR is < 1.6, or in some laboratories, below 1.8. Postponing invasive procedures until the INR is < 1.8 may extend the hospitalization and/or delay important diagnostic testing. Bridging therapy with intravenous heparin in the hospital has been recommended in this setting, but there are scant data to indicate whether this is truly helpful. This traditional approach has been challenged in recent years. Cessation of warfarin therapy must balance an increased hospitalization against bleeding risks with bridging heparin. It is worth revisiting the question to see when it is safe to perform cardiac catheterization on a patient with uninterrupted warfarin therapy.

Recently, Annala et al reported on the “Safety of Diagnostic Coronary Angiography During Uninterrupted Therapeutic Warfarin Treatment.”1 This study reviewed 258 patients on warfarin therapy referred for diagnostic coronary angiography in two centers with extensive experience in uninterrupted warfarin and in one center with a policy of preprocedural warfarin pause. In this study, radial artery access was used 50% of the time in the uninterrupted warfarin group, and 60% in the controlled warfarin pause group. The authors found there was no difference between the two groups in access site bleeding complications (1.9% versus 1.6%) or major cerebrovascular or cardiovascular events (0.4% versus 0.8%). Regarding bridging therapy using heparin in 30% of patients, the bridged patients had higher INR levels than the uninterrupted warfarin group (2.3 versus 1.9), and had higher access site bleeding complications (1.7% versus bridging therapy of 8.3%). Nonetheless, only two patients needed a blood transfusion in the bridging therapy group. It is important to note that access site complications were more common in the supratherapeutic anticoagulant group with INRs > 3 than in patients with therapeutic periprocedural INRs (9.1% versus 1.5%; p < 0.05). These data suggest that a simple strategy of performing coronary angiography during uninterrupted therapeutic warfarin anticoagulation can be performed safely. Furthermore, this approach may be an acceptable, if not superior, alternative to bridging therapy, a strategy that could lead to considerable cost savings.

This timely paper by Annala et al focuses our attention on the data needed to answer the question of how to manage patients on anticoagulation coming to the cath lab. Who are the patients that require long-term warfarin anticoagulation? Clinical indications for warfarin include atrial fibrillation, valvular heart disease (with or without prosthetic heart valves), cardiomyopathy, deep venous thrombosis, recent pulmonary embolism, or intracardiac thrombus. The risk of recurrent or de novo thromboembolism with subtherapeutic anticoagulation is unknown but potentially catastrophic. Annala et al suggest warfarin cessation is associated with a favorable outcome in the period surrounding the cardiac catheterization.

Some of our electrophysiologic colleagues tell us that automatic implantable cardioverter defibrillator (AICD) or pacemaker implantations often occur in the setting of atrial fibrillation and/or severe cardiomyopathy. It is felt that there is more risk to interrupting warfarin therapy than there is to performing the implantation procedures off anticoagulation. Nonetheless, we worry about potential bleeding, not only at the access site, but also related to other potential complications such as cardiac or arterial perforation, venous or arterial vascular injury, or spontaneous retroperitoneal bleeding. Catheterization using the radial artery access is an excellent approach to the patient on warfarin and is associated with less access site bleeding, despite a small risk of retroperitoneal bleeding (as could occur in any anticoagulated patient). Some operators who advocate the “cath on warfarin” approach indicate that these rare anticoagulation-related complications can be managed with the administration of vitamin K and clotting factors, and if the femoral artery was used, apply a vascular closure device (VCD). VCDs can achieve good hemostasis in anticoagulated patients with a marked reduction of bleeding complications.

There is no large-scale trial that convincingly shows us whether the outcomes of catheterization on continued treatment with warfarin are better than the warfarin pause or bridging approach. Annala et al had 5 access site complications in the warfarin group, with the INRs ranging between 1.7 and 4.2. The one patient who received a blood transfusion had an INR of 4.2. Bridging therapy has been associated with an increased risk of bleeding events, as reported by MacDonald et al2 using low-molecular heparins with an incidence of bleeding complications of approximately 8%. Another concern of warfarin cessation is the transient prothrombotic state that may result due to protein C and S suppression after re-initiation of warfarin.3

Mixing different anticoagulants in the same patient also increases the bleeding risk. For example, in the SYNERGY trial, the patients who had the highest bleeding events were those who received more than one anticoagulant (heparin, low-molecular-weight heparin, or bivalurudin). As a matter of tradition, especially for PCI, the fear of unopposed bleeding had mandated the cessation of warfarin before interventional cardiac procedures. The argument against discontinuation of warfarin now is that the bleeding can be overcome by infusion of blood clotting factors II, VII, IX, X and use of VCDs. This short answer certainly does not address the rare but critical situations of coronary perforation with tamponade or retroperitoneal bleeding from sites remote from the femoral access.

What should we do at this time with the data we have? I can propose an initial strategy: for patients who must remain on warfarin for strong clinical indications, perform catheterization with the INR between 1.5 and 2, thus achieving continued therapy without excessive risk. If the INR must be maintained between 2 and 3, consider the radial artery access as a first approach, or perform femoral access with accurate (lower rather than higher) puncture and anticipate using a secure VCD. Of course, we still employ the traditional approach of heparin as a bridging therapy after stopping warfarin until the INR is < 1.8, then proceeding to catheterization in a timely manner. This empiric approach to patients on warfarin anticoagulation needs to be carefully considered; hopefully, one day, a large trial will help us find the optimal strategy.