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, Richard C. Becker, et al.

Bottom Line

In patients with atrial fibrillation requiring temporary interruption of warfarin for elective surgery or procedures, a no-bridging strategy was noninferior to low-molecular-weight heparin (dalteparin) bridging for arterial thromboembolism and resulted in significantly lower rates of major bleeding.

Key Findings

1. The primary efficacy endpoint of arterial thromboembolism (stroke, TIA, or systemic embolism) occurred in 0.4% of patients in the no-bridging group compared to 0.3% in the dalteparin bridging group (P=0.01 for noninferiority).
2. The primary safety endpoint of major bleeding was significantly lower in the no-bridging group, occurring in 1.3% of patients, compared to 3.2% in the dalteparin bridging group (P=0.005).
3. Minor bleeding rates were also significantly lower in the no-bridging group (12.0%) compared to the dalteparin group (20.9%, P<0.001).

Study Design

Design
RCT
Double-Blind
Sample
1,884
Patients
Duration
30 days
Median
Setting
Multicenter, US/Canada
Population Patients with permanent or paroxysmal atrial fibrillation or flutter receiving long-term warfarin therapy who required temporary interruption for an elective procedure or surgery and had at least one CHADS2 risk factor.
Intervention Subcutaneous dalteparin (100 IU/kg) administered twice daily, starting 3 days before the procedure, stopped 24 hours pre-procedure, and resumed post-procedure.
Comparator Matching subcutaneous placebo administered twice daily according to the same schedule as the intervention group.
Outcome Composite of arterial thromboembolism (stroke, transient ischemic attack, or systemic embolism) and major bleeding.

Study Limitations

The trial did not include patients with mechanical heart valves, who are at a much higher thrombotic risk and may still require bridging.
Patients at very high risk for stroke (e.g., CHADS2 scores of 5 or 6) were under-represented.
The study focused specifically on warfarin-treated patients and cannot be directly generalized to those on direct oral anticoagulants (DOACs).

Clinical Significance

This landmark study fundamentally shifted clinical practice by demonstrating that routine 'bridging' with low-molecular-weight heparin is unnecessary for the majority of patients with atrial fibrillation undergoing elective procedures, as it provides no incremental protection against thromboembolism while significantly increasing the risk of major and minor bleeding.

Historical Context

Prior to this trial, clinical guidelines and practice patterns were highly variable and often leaned toward universal bridging with heparin to mitigate the risk of perioperative stroke during warfarin interruption, despite a lack of high-quality randomized evidence and potential for bleeding complications.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

What is the physiological and pharmacological rationale behind 'bridging' a patient on warfarin with a short-acting anticoagulant like dalteparin before a surgical procedure?

Key Response

Warfarin inhibits the synthesis of vitamin K-dependent clotting factors (II, VII, IX, X) and has a long half-life (approx. 40 hours). When warfarin is stopped for surgery, the patient remains in a sub-therapeutic state for several days, increasing the risk of thromboembolism. Bridging uses a short-acting agent (like LMWH) to minimize the 'anticoagulation gap' because its effects can be quickly started and stopped, allowing for precise control near the time of incision.

Resident
Resident

According to the BRIDGE trial results, which specific patient outcomes were most significantly affected by the decision to withhold bridging, and how should this influence your perioperative management of a patient with a CHADS2 score of 2?

Key Response

The trial found that no-bridging was noninferior to bridging for preventing arterial thromboembolism (0.3% vs 0.4%) but significantly reduced major bleeding (1.3% vs 3.2%). For a patient with a moderate risk (CHADS2 of 2), the evidence suggests that the risk of bleeding from bridging outweighs the potential benefit of stroke prevention, supporting a strategy of simply interrupting warfarin without LMWH.

Fellow
Fellow

The BRIDGE trial predominantly enrolled patients with a mean CHADS2 score of 2.3. Discuss the limitations of applying these findings to 'high-risk' patients, specifically those with mechanical heart valves or a CHADS2 score greater than 4.

Key Response

Patients with mechanical heart valves, recent stroke (within 12 weeks), or very high CHADS2 scores (>4) were excluded from the trial. Because these populations have a significantly higher baseline risk of thromboembolism, the noninferiority demonstrated in BRIDGE cannot be confidently generalized to them, and bridging may still be clinically indicated in these specific high-risk cohorts.

Attending
Attending

How does the BRIDGE trial's demonstration of 'noninferiority' for efficacy and 'superiority' for safety challenge the traditional 'risk-stratification' paradigm that previously favored bridging for most atrial fibrillation patients?

Key Response

Historically, the fear of perioperative stroke led clinicians to favor bridging. BRIDGE proved that the thromboembolic risk during a brief warfarin interruption is lower than previously estimated (under 0.5%), while the risk of procedure-related bleeding is nearly tripled by bridging. This shifts the default practice toward a 'no-bridging' strategy for the vast majority of elective procedures in AF patients.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The BRIDGE trial utilized a noninferiority margin of 1.0% for the primary efficacy outcome. Given that the observed event rate was only 0.3-0.4%, critique the statistical power and the clinical relevance of the chosen margin.

Key Response

A noninferiority margin of 1.0% was relatively large given the actual low incidence of thromboembolic events. If the true event rate is significantly lower than the margin, the study might be 'underpowered' to detect a small but clinically meaningful difference in stroke rates, or conversely, the margin might be too wide to satisfy a conservative clinician. However, the dramatic reduction in bleeding (the safety endpoint) provides a compelling 'net clinical benefit' argument.

Journal Editor
Journal Editor

As a reviewer, how would you evaluate the threat to external validity posed by the trial's exclusion of patients undergoing major surgeries with extremely high bleeding risks, such as neurosurgery or cardiac surgery?

Key Response

The trial focused on elective procedures where warfarin interruption is standard. By excluding surgeries with the highest bleeding risk or those requiring very specific anticoagulation protocols (like bypass), the trial maintains high internal validity but may leave a gap in evidence for the most complex surgical cases. A reviewer would flag this to ensure the authors do not over-extrapolate the safety profile to major 'high-stakes' surgeries.

Guideline Committee
Guideline Committee

Based on the BRIDGE trial evidence, how should the strength of recommendation for perioperative bridging in AF be updated in the CHEST or ACC/AHA guidelines compared to previous iterations?

Key Response

Prior to BRIDGE, guidelines often recommended bridging based on CHADS2 risk tiers (e.g., CHEST 2012). Following this trial, guidelines (such as the 2017 ACC Expert Consensus and 2019 AHA/ACC/HRS updates) moved toward a Class IIa or even Class III (Harm) recommendation for bridging in low-to-moderate risk AF, emphasizing that for most patients with a CHADS2 score ≤4, bridging should be avoided to minimize major bleeding.

Clinical Landscape

Noteworthy Related Trials

2009

RE-LY Trial

n = 18,113 · NEJM

Tested

Dabigatran 110 mg or 150 mg twice daily

Population

Patients with atrial fibrillation and risk of stroke

Comparator

Adjusted-dose warfarin

Endpoint

Stroke or systemic embolism

Key result: Dabigatran 150 mg was superior to warfarin in reducing stroke, while 110 mg was noninferior with lower rates of major hemorrhage.
2011

ROCKET AF Trial

n = 14,264 · NEJM

Tested

Rivaroxaban 20 mg once daily

Population

Patients with nonvalvular atrial fibrillation at increased risk of stroke

Comparator

Adjusted-dose warfarin

Endpoint

Composite of stroke or systemic embolism

Key result: Rivaroxaban was noninferior to warfarin for the prevention of stroke or systemic embolism in patients with nonvalvular atrial fibrillation.
2011

ARISTOTLE Trial

n = 18,201 · NEJM

Tested

Apixaban 5 mg twice daily

Population

Patients with atrial fibrillation at risk of stroke

Comparator

Adjusted-dose warfarin

Endpoint

Composite of stroke or systemic embolism

Key result: Apixaban was superior to warfarin in preventing stroke, caused less bleeding, and resulted in lower mortality.

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