What's the Take Home?

A Woman on Warfarin With Atrial Fibrillation Who Needs a Colonoscopy

 Ronald Rubin, MD—Series Editor

Ronald N. Rubin, MD—Series Editor

Rubin RN. A Woman on warfarin with atrial fibrillation who needs a colonoscopy. Consultant. 2016;56(8):719-720.


A 71-year-old woman was due for colonoscopy. Three years ago, a routine colonoscopy had revealed 2 colonic polyps, which had been diagnosed as adenomas. She now was due for another colonoscopy for monitoring and removal of any new lesions. She had been asymptomatic as regards her gastrointestinal tract.

Her past medical history was significant for a decade-long history of atrial fibrillation (AF) related to lifelong hypertension and for a myocardial infarction 10 years ago. She was rate-controlled and had no overt symptoms of congestive heart failure (CHF), and she had never sustained an arterial thrombotic event. Her score on the 6-point CHADS2 (CHF, hypertension, age ≥ 75 years, diabetes, stroke) stroke risk grading system was 2.

Her medications of note included a β-blocker, warfarin, and aspirin at 81 mg daily. A review of her medical record from 3 years ago revealed that she had a perioperative anticoagulation bridging scheme, and that her anticoagulation needed to be interrupted due to some “minor” bleeding associated with her colonoscopy procedure and anticoagulation protocol.

Physical examination revealed an irregularly irregular rhythm of 76 beats/min. The chest was clear to auscultation. Heart sounds included no significant murmurs and no gallops. Rectal examination findings were negative for masses.

Diagnostic Testing

Stool guaiac test results were negative for blood. Results of a basic metabolic panel were normal, including the creatinine level. Results of a complete blood cell count showed a normal hemoglobin concentration and a ferritin level of 67 ng/mL.

An electrocardiogram showed a Q wave inferiorly, and AF with a rate of 75 beats/min. Echocardiography revealed an ejection fraction of 45%. Her most recent international normalized ratio (INR) values were 2.2, 2.4, and 2.4.


A case such as that of this woman, who has AF and requires a procedure, opens the issue of whether and how to safely continue or interrupt her chronic anticoagulation while at the same time balancing thrombotic vs hemorrhagic risk.

A copious literature exists describing a variety of schemes to switch from chronic, stable oral anticoagulation to perioperative parenteral forms, then restart oral anticoagulation again, a protocol commonly known as bridging.1-3 The most common protocol is to stop warfarin (in patients using that anticoagulant) 5 days prior to the procedure and use a low-molecular-weight heparin (LMWH) in the immediate periprocedural period; then, while still on LMWH, restart warfarin and maintain both until the therapeutic INR is reestablished.1-3

However, despite the ubiquity and detail of this bridging maneuver, data about whether such an exercise actually is required are scant, and practice guidelines covering this situation are weak and inconsistent.1,4 Thus, a very welcome study with a broad and deep array of expert coagulationists recently has been published to shed some data-driven light on this issue—the BRIDGE (Bridging Anticoagulation in Patients Who Require Temporary Interruption of Warfarin Therapy for an Elective Invasive Procedure or Surgery) clinical trial.1

In this extremely well-designed, randomized, double-blind, placebo-controlled trial, almost 2000 patients were randomly assigned to a literature-based bridging protocol (as described above) or to no bridging when undergoing a procedure. The primary efficacy endpoint was the rate of arterial thromboembolism, while the primary toxicity endpoint was major bleeding. Other monitored results included acute myocardial infarction (AMI), venous thromboembolism, and death, along with minor bleeding. Patients in certain very-high-risk situations for arterial thromboembolism (eg, mechanical valves) and for bleeding (eg, cardiac, intracranial, intraspinal procedures) were excluded. But a vast majority of common situations, such as colonoscopy and gastrointestinal and orthopedic procedures, were included.

The results of BRIDGE trial could not have been clearer. Using a noninferiority statistical model, in stable patients on warfarin undergoing operations, a strategy of forgoing anticoagulation bridging was found to be noninferior to bridging with LMWH in terms of the prevention of arterial thromboembolic events. Additionally and importantly, there was a very pronounced decrease in both major and minor bleeding events.

The absolute numbers were more impressive: When an anticoagulant bridge was used, at 30 days there were 3 events in 895 patients (0.3%) compared with 4 events in 918 patients (0.4%) in the nonbridging group, which clearly is nonsignificant (P = .73). On the toxicity side, a bridging scheme showed major bleeding in 29 of 895 (3.2%) patients compared with 12 of 918 (1.3%) in the nonbridging group (P = .005), and an even greater difference in minor bleeding in the bridging group—20.9% compared with 12.0% in nonbridging group (P = .001). Thus, the data strongly indicate that there is no inferiority to nonbridging and, in fact, a large benefit in the avoidance of bleeding complications. Note that rates of overall mortality, AMI, and venous thromboembolic events also were equivalent between the groups. Therefore, Answer A is correct here.

Similar data also support a nonbridging strategy when treating patients with the newer direct oral anticoagulants. A substudy of the RE-LY (Randomized Evaluation of Long-Term Anticoagulation Therapy) trial with dabigatran suggested a similar increased bleeding risk without significant effect on arterial thromboembolism,5 but these data are not as strong as the randomized, double-blind, placebo-controlled BRIDGE study and need to be confirmed. Answer B is thus an overstatement at this time and is not the optimal choice.

Answers C and D relate to other important issues about bridging. The etiology of AF does affect thrombotic risk in that mechanical valves and carotid endarterectomy increase the risk, and bridging likely is needed. Such patients were excluded from the BRIDGE trial and are in a separate category. Similarly, the nature of the procedure (eg, intracranial, intraspinal) also conveys a higher risk of bleeding, and patients in this situation were similarly excluded from BRIDGE. These situations require case-by-case risk-benefit analysis.


The decision was made to forego bridging. Warfarin was stopped 5 days prior to the woman’s colonoscopy and was resumed 24 hours after the procedure at her usual maintenance dose. She experienced neither bleeding issues nor thromboembolic events and was well on warfarin maintenance 3 months after her procedure.

Ronald Rubin, MD, is a professor of medicine at the Temple University School of Medicine and is chief of clinical hematology in the Department of Medicine at Temple University Hospital in Philadelphia, Pennsylvania.


  1. Douketis JD, Spyropoulos AC, Kaatz S, et al; BRIDGE Investigators. Perioperative bridging anticoagulation in patients with atrial fibrillation. N Engl J Med. 2015;373(9):823-833.
  2. Baron TH, Kamath PS, McBane RD. Management of antithrombotic therapy in patients undergoing invasive procedures. N Engl J Med. 2013;368(22):​2113-2124.
  3. Schulman S, Hwang H-G, Eikelboom JW, Kearon C, Pai M, Delaney J. Loading dose vs. maintenance dose for reinitiation after invasive procedures: a randomized trial. J Thromb Haemost. 2014;12(8):1254-1259.
  4. Douketis JD, Spyropoulos AC, Spencer FA, et al. Perioperative management of antithrombotic therapy: antithrombotic therapy and prevention of thrombosis. 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2012;141(2 suppl):e326S-e350S.
  5. Douketis JD, Healey JS, Brueckmann M, et al. Perioperative bridging anticoagulation during dabigatran or warfarin interruption among patients who had an elective surgery or procedure: substudy of the RE-LY trial. Thromb Haemost. 2015;113(3):625-632.