The New England Journal of Medicine August 27, 2015

Perioperative Bridging Anticoagulation in Patients with Atrial Fibrillation

James D. Douketis, Alex C. Spyropoulos, Scott Kaatz, et al.

Bottom Line

In patients with atrial fibrillation requiring temporary interruption of warfarin for an elective procedure, omitting bridging anticoagulation was noninferior to bridging with LMWH for preventing arterial thromboembolism and significantly reduced the risk of major bleeding.

Key Findings

1. Arterial thromboembolism occurred in 0.4% of patients in the no-bridging group and 0.3% of patients in the bridging group (risk difference, 0.1 percentage points; 95% CI, -0.6 to 0.8; P=0.01 for noninferiority) [1.2.8].
2. Major bleeding occurred in 1.3% of the no-bridging group compared to 3.2% of the bridging group (relative risk, 0.41; 95% CI, 0.20 to 0.78; P=0.005 for superiority) [1.2.8].

Study Design

Design
Randomized Controlled Trial
Double-Blind
Sample
1,884
Patients
Duration
30 days
Median
Setting
Multicenter, North America
Population Adults with atrial fibrillation or flutter receiving warfarin for ≥3 months (CHADS2 score ≥1) who required temporary interruption of anticoagulation for an elective invasive procedure or surgery.
Intervention No bridging (placebo): Warfarin interrupted 5 days prior to the procedure; subcutaneous placebo administered 3 days before until 24 hours prior, and for 5 to 10 days post-procedure.
Comparator Bridging therapy: Warfarin interrupted 5 days prior; subcutaneous dalteparin (100 IU/kg twice daily) administered 3 days before until 24 hours prior, and for 5 to 10 days post-procedure.
Outcome Composite of arterial thromboembolism (stroke, systemic embolism, or transient ischemic attack) and major bleeding at 30 days.

Study Limitations

The study population had a relatively low mean CHADS2 score (2.3), meaning the findings may not generalize to patients at the highest risk of thromboembolism, such as those with a CHADS2 score of 5 or 6 [3.1.6].
Patients with mechanical heart valves were specifically excluded, so these results cannot be applied to that population.
The vast majority of enrolled patients (89%) underwent minor procedures with a low intrinsic risk of bleeding.

Clinical Significance

The BRIDGE trial established a monumental paradigm shift in perioperative management by providing definitive evidence against the routine use of bridging anticoagulation for most patients with atrial fibrillation. It proved that forgoing LMWH bridging prevents unnecessary and potentially harmful major bleeding without exposing patients to a higher risk of stroke or systemic embolism.

Historical Context

Historically, perioperative bridging with heparin or LMWH was standard practice for patients on chronic warfarin therapy due to the fear of devastating thromboembolic events, despite limited rigorous data supporting its safety or efficacy. Prior to the BRIDGE trial, retrospective data and meta-analyses raised concerns about increased bleeding risks. As the first large-scale, randomized, double-blind, placebo-controlled trial on the subject, BRIDGE definitively challenged long-held clinical dogma, directly informing major societal guidelines to de-escalate or abandon routine bridging in patients with atrial fibrillation undergoing elective procedures.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What are the pharmacokinetic differences between warfarin and low-molecular-weight heparin (LMWH) that historically formed the rationale for perioperative bridging, and how does the BRIDGE trial challenge this traditional practice?

Key Response

Warfarin has a long half-life and delayed onset/offset, leaving patients unprotected for days during interruption. LMWH has a shorter half-life and predictable kinetics, theoretically protecting the patient. The BRIDGE trial shows that the actual clinical risk of stroke during this window is much lower than feared, while the bleeding risk from LMWH is substantial.

Resident
Resident

Based on the BRIDGE trial, how should you manage a patient with a CHADS2 score of 2 taking warfarin for atrial fibrillation who is scheduled for an elective colonoscopy with polypectomy, and what specific post-procedural risk does LMWH introduce?

Key Response

The patient should stop warfarin 5 days prior and resume post-op without bridging. Bridging provides no significant reduction in stroke risk for a CHADS2 score of 2 but more than doubles the risk of major bleeding. Resuming therapeutic LMWH post-polypectomy introduces a high risk of post-procedural hemorrhage.

Fellow
Fellow

The BRIDGE trial excluded patients with mechanical heart valves and had a low mean CHADS2 score (2.3). How should these exclusions influence the decision to bridge a patient with a CHADS2 score of 6 or a patient with a mechanical mitral valve?

Key Response

The trial applies specifically to non-valvular AF with low-to-moderate stroke risk. Mechanical valves (especially mitral) have a distinct, highly thrombogenic pathophysiology where bridging is still recommended. Patients with CHADS2 scores of 5-6 were underrepresented, meaning the noninferiority of 'no bridging' might not confidently extend to the highest-risk AF cohorts.

Attending
Attending

For decades, bridging was the standard of care driven by the fear of periprocedural stroke. How does the BRIDGE trial illustrate the cognitive bias of 'commission bias' in medical practice?

Key Response

Commission bias is the tendency to favor action over inaction, even if the action causes harm. BRIDGE is a classic example where acting on theoretical physiologic principles (preventing a rare stroke) caused measurable iatrogenic harm (major bleeding). It teaches trainees that 'doing something' is not always safer.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The BRIDGE trial utilized a noninferiority design for the primary efficacy endpoint with a predefined margin. What are the statistical implications of the chosen noninferiority margin in the context of the unexpectedly low overall event rate, and how does this affect robustness?

Key Response

If the margin is wide relative to the actual event rate (which was under 0.5%), a treatment could be deemed noninferior even if it allows a higher relative risk of the outcome. Proving noninferiority robustly with such low event rates is statistically challenging, though the overwhelming superiority in the safety endpoint (bleeding) ultimately drove the clinical conclusion.

Journal Editor
Journal Editor

As an editor reviewing the BRIDGE manuscript, what concerns would you raise regarding the external validity of the study given the high proportion of low-risk patients and the exclusive focus on warfarin rather than DOACs?

Key Response

A reviewer would flag that the cohort's low mean CHADS2 score dilutes the power to detect differences in arterial thromboembolism in high-risk patients. Furthermore, contemporary practice heavily relies on DOACs, which have short half-lives and do not require bridging, potentially limiting the trial's relevance to modern antithrombotic management.

Guideline Committee
Guideline Committee

Considering current perioperative antithrombotic guidelines, how does the BRIDGE trial shift the Class of Recommendation for bridging in non-valvular AF, and what gray areas must guidelines still address?

Key Response

BRIDGE provides Level of Evidence B-R supporting a Class III (Harm) recommendation against bridging in low-to-moderate risk non-valvular AF. Guidelines must explicitly address 'gray areas' not covered by BRIDGE, such as patients with recent stroke, very high CHA2DS2-VASc scores, or mechanical valves, where bridging (Class IIa or IIb) might still be considered based on expert consensus.

Clinical Landscape

Noteworthy Related Trials

2009

RE-LY Trial

n = 18,113 · NEJM

Tested

Dabigatran 110mg or 150mg twice daily

Population

Patients with nonvalvular atrial fibrillation

Comparator

Warfarin

Endpoint

Stroke or systemic embolism

Key result: Dabigatran 150mg reduced stroke and systemic embolism compared to warfarin with similar rates of major bleeding.
2019

PAUSE Trial

n = 3,007 · JAMA Intern Med

Tested

Standardized perioperative DOAC interruption without bridging

Population

Atrial fibrillation patients on DOACs undergoing elective surgery

Comparator

Predefined safety boundaries

Endpoint

Major bleeding and arterial thromboembolism at 30 days

Key result: A simple, standardized DOAC interruption strategy without bridging resulted in low rates of major bleeding and arterial thromboembolism.
2021

PERIOP2 Trial

n = 1,471 · BMJ

Tested

Postoperative dalteparin bridging

Population

Patients with mechanical heart valves or AF on warfarin undergoing surgery

Comparator

Placebo

Endpoint

Composite of major thromboembolism or major bleeding

Key result: Postoperative bridging with LMWH did not reduce thromboembolic events but significantly increased major bleeding.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis