JAMA Internal Medicine August 05, 2019

Perioperative Management of Patients With Atrial Fibrillation Receiving a Direct Oral Anticoagulant (PAUSE)

James D. Douketis et al.

Bottom Line

A standardized protocol for interrupting and resuming direct oral anticoagulants without heparin bridging around the time of elective surgery was safe, resulting in very low rates of major bleeding and arterial thromboembolism.

Key Findings

1. The 30-day postoperative rate of major bleeding was exceptionally low across all cohorts: 1.35% (95% CI, 0-2.00%) for apixaban, 0.90% (95% CI, 0-1.73%) for dabigatran, and 1.85% (95% CI, 0-2.65%) for rivaroxaban.
2. Rates of arterial thromboembolism (ischemic stroke, systemic embolism, and transient ischemic attack) at 30 days were also extremely low: 0.16% (95% CI, 0-0.48%) for apixaban, 0.60% (95% CI, 0-1.33%) for dabigatran, and 0.37% (95% CI, 0-0.82%) for rivaroxaban.
3. A high proportion of patients had minimal or no residual anticoagulant levels (<50 ng/mL) at the time of the procedure (90.5% for apixaban, 95.1% for dabigatran, and 96.8% for rivaroxaban).

Study Design

Design
Prospective Cohort Study
N/A
Sample
3,007
Patients
Duration
30 days
Median
Setting
Multicenter, multinational
Population Adult patients (≥18 years) with atrial fibrillation receiving long-term apixaban, dabigatran, or rivaroxaban who were scheduled for an elective surgery or procedure requiring DOAC interruption.
Intervention A standardized DOAC interruption and resumption protocol without heparin bridging: DOAC was omitted 1 day before low-bleed-risk and 2 days before high-bleed-risk procedures (extended for dabigatran if CrCl <50 mL/min); resumed 1 day after low-bleed-risk and 2-3 days after high-bleed-risk procedures.
Comparator None (Single-arm cohort study with three parallel DOAC cohorts)
Outcome Major bleeding and arterial thromboembolism (ischemic stroke, systemic embolism, and transient ischemic attack) at 30 days post-procedure.

Study Limitations

The study utilized a prospective cohort design without a randomized control group, introducing potential selection bias.
Relatively few patients underwent neuraxial anesthesia, limiting the generalizability of safety conclusions to that specific high-risk subgroup.
Approximately 12.7% of patients deviated from the protocol, highlighting potential real-world adherence challenges to standardized interruption strategies.

Clinical Significance

The PAUSE trial fundamentally changed perioperative practice by demonstrating that patients on DOACs for atrial fibrillation can safely undergo surgery using a simple, standardized interruption schedule based on procedural bleeding risk and renal function, without the need for potentially harmful heparin bridging or routine preoperative coagulation testing.

Historical Context

Prior to the PAUSE trial, the perioperative management of DOACs was highly variable and heavily influenced by historical practices of 'heparin bridging' commonly used for warfarin. After the BRIDGE trial (2015) showed that bridging warfarin increased bleeding without proportionally reducing thromboembolism, there was an urgent need to establish an evidence-based protocol for DOACs, which have a much shorter half-life. PAUSE provided the definitive clinical framework for modern perioperative DOAC management.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

Why does the pharmacokinetic profile of direct oral anticoagulants (DOACs) make bridging therapy with low-molecular-weight heparin unnecessary and potentially harmful compared to traditional perioperative management with warfarin?

Key Response

DOACs have a rapid onset of action (1-4 hours) and a short half-life (typically 12-14 hours), meaning their anticoagulant effect predictably and quickly dissipates upon cessation and resumes rapidly upon restarting. Warfarin, by contrast, takes days to wash out and re-establish therapeutic levels. Bridging DOACs simply overlaps two rapid-acting anticoagulants, exponentially increasing the risk of major bleeding without providing additional protection against thrombosis.

Resident
Resident

When applying the PAUSE protocol to determine when to interrupt a DOAC prior to an elective procedure, what two primary patient and procedure-specific variables must you assess, and how does the choice of dabigatran alter this algorithm compared to apixaban or rivaroxaban?

Key Response

The two key variables are the procedure's bleeding risk (low vs. high) and the patient's creatinine clearance (CrCl). Dabigatran is predominantly renally cleared (80%), requiring a specifically extended interruption time (e.g., 4 days for high bleed risk) in patients with CrCl less than 50 mL/min compared to apixaban and rivaroxaban, which have significantly lower renal clearance.

Fellow
Fellow

The PAUSE study measured preoperative residual DOAC levels in a subset of patients, finding that a small percentage still had levels greater than 50 ng/mL despite strict protocol adherence. How should this finding influence your perioperative management of a patient scheduled for high-risk neuraxial anesthesia?

Key Response

While the vast majority of patients had negligible residual levels, the few with levels greater than 50 ng/mL pose a severe risk for epidural or spinal hematoma. For procedures where even microscopic bleeding is catastrophic (neuraxial anesthesia), fellows must recognize that clinical protocols are population-based. Unpredictable individual variations in clearance may necessitate checking specific anti-Xa levels, a dilute thrombin time, or further extending the interruption window to guarantee safe neuraxial puncture.

Attending
Attending

Despite robust evidence from the BRIDGE and PAUSE trials demonstrating the harms of heparin bridging, many surgical colleagues remain hesitant to leave patients 'unprotected' off anticoagulation. How can you use the PAUSE data to effectively reframe perioperative DOAC management during surgical risk discussions?

Key Response

The teaching point focuses on shifting the paradigm from 'preventing strokes' to 'doing no harm through bleeding.' PAUSE demonstrated a 30-day stroke risk of less than 1 percent without any heparin bridging. Attendings must emphasize that the perioperative hypercoagulable state is best managed by minimizing surgical bleeding, because post-operative bleeding forces a delayed resumption of the DOAC, which is the actual primary driver of perioperative stroke.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The PAUSE trial was designed as a prospective, single-arm management cohort study rather than a randomized controlled trial comparing standardized interruption against individualized 'usual care' or bridging. What are the methodological justifications and potential threats to internal validity inherent in choosing this study design?

Key Response

A randomized trial was deemed practically challenging and ethically questionable given the established harms of bridging from the BRIDGE trial for warfarin and the lack of clinical equipoise for DOAC bridging. However, the lack of a concurrent control group is a threat to internal validity; the exceptionally low event rates could be influenced by selection bias (enrolling lower-risk, elective surgery patients) or temporal trends, limiting definitive causal claims about the protocol's superiority over individualized care.

Journal Editor
Journal Editor

As a peer reviewer assessing the external validity of the PAUSE protocol, which specific patient populations were excluded or underrepresented in this cohort, and how should these exclusions impact the editorial decision regarding the manuscript's claims of real-world applicability?

Key Response

A critical reviewer would flag that PAUSE excluded patients with mechanical heart valves, recent (within 3 months) thromboembolic events, and severe renal impairment (CrCl less than 25-30 mL/min). Thus, the protocol cannot be universally applied to all anticoagulated patients. Highlighting these exclusions is crucial for an editor to ensure the authors explicitly define their study's boundaries and do not overstate the algorithm's safety in highly complex or inherently unstable patient populations.

Guideline Committee
Guideline Committee

How does the prospective validation of the PAUSE algorithm influence the level of evidence for current perioperative guidelines (such as the ACCP or ACC/AHA guidelines), and does this study provide sufficient evidence to support a strong recommendation against routine perioperative DOAC bridging?

Key Response

PAUSE elevates the evidence from expert consensus and retrospective data to high-quality prospective cohort data. Current guidelines (like the 2022 ACCP guidelines on perioperative management) strongly recommend against bridging DOACs. PAUSE solidifies this with a standardized, reproducible, pharmacokinetic-driven algorithm, transitioning the evidence base to Level B-R (or B-NR depending on the grading system) and serving as the foundational prospective evidence to firmly anchor Class 1 recommendations against DOAC bridging in elective procedures.

Clinical Landscape

Noteworthy Related Trials

2011

ARISTOTLE Trial

n = 18,201 · NEJM

Tested

Apixaban 5mg twice daily

Population

Patients with atrial fibrillation and at least one additional risk factor for stroke

Comparator

Warfarin (target INR 2.0-3.0)

Endpoint

Stroke or systemic embolism

Key result: Apixaban was superior to warfarin in preventing stroke or systemic embolism, caused less bleeding, and resulted in lower mortality.
2015

BRIDGE Trial

n = 1,884 · NEJM

Tested

No bridging with low-molecular-weight heparin

Population

Atrial fibrillation patients on warfarin undergoing elective surgery

Comparator

Bridging with dalteparin

Endpoint

Arterial thromboembolism and major bleeding

Key result: Forgoing bridging anticoagulation was noninferior to bridging for the prevention of arterial thromboembolism and significantly decreased the risk of major bleeding.
2015

RE-VERSE AD Trial

n = 503 · NEJM

Tested

Idarucizumab (5g intravenous)

Population

Dabigatran-treated patients with serious bleeding or requiring urgent surgery

Comparator

None (single-arm study)

Endpoint

Maximum percentage reversal of the anticoagulant effect of dabigatran

Key result: Idarucizumab completely and rapidly reversed the anticoagulant effect of dabigatran in almost all patients.

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