Early Bleeding Risk After DOAC Initiation With Concomitant Amiodarone in Older Adults With Atrial Fibrillation

Dr. Isadora D. GuarinoResident Research Spotlight

Journal of Cardiovascular Electrophysiology, 2026View Publication →

Study Overview

Amiodarone and direct oral anticoagulants (DOACs) are frequently co-prescribed in older adults with atrial fibrillation (AF), yet the safety of this combination during the critical early period after DOAC initiation has been incompletely characterized. Amiodarone inhibits P-glycoprotein and CYP3A4, two pathways central to DOAC metabolism, raising concern for elevated drug levels and bleeding risk, particularly in older patients who are already vulnerable.

In this retrospective propensity score-matched cohort study, Dr. Isadora Guarino and colleagues used the TriNetX US Research Network (2015–2025) to identify adults aged 75 years or older with nonvalvular AF newly initiating a DOAC. Amiodarone exposure was defined as a prescription or administration within 30 days before or on the DOAC start date. After 1:1 matching on demographics, comorbidities, and baseline cardiovascular medications, 136,806 patients were included (68,403 per cohort). Follow-up extended 90 days from DOAC initiation, capturing early bleeding events during the window of greatest pharmacokinetic vulnerability.

Key Findings

Concomitant amiodarone was associated with a significantly higher risk of GI bleeding (2.3% vs 1.9%; HR 1.30, 95% CI 1.20–1.40).
Risk of intracranial hemorrhage (ICH) was also elevated (0.4% vs 0.3%; HR 1.28, 95% CI 1.07–1.52).
There was no significant difference in ischemic stroke (1.3% vs 1.5%; HR 0.91, 95% CI 0.83–1.01), suggesting the drug interaction does not compromise anticoagulant efficacy.
The majority of index-date DOAC exposures were apixaban (82.7%) and rivaroxaban (16.5%); dabigatran and edoxaban were uncommon.
All standardized mean differences post-matching were <0.1, supporting well-balanced cohorts.

Clinical Significance

This study provides real-world evidence that the amiodarone-DOAC drug interaction carries clinically meaningful early bleeding risk, not just a theoretical pharmacokinetic concern. Even small absolute risk differences in ICH are consequential in adults aged 75 and older, where the morbidity of intracranial bleeding is particularly severe.

The findings support a structured approach at the time of DOAC initiation in amiodarone-treated patients: careful medication review, consideration of dose reduction where indicated (e.g., apixaban 2.5 mg BID criteria), and proactive counseling about early bleeding symptoms. Importantly, the absence of a difference in ischemic stroke suggests the combination remains protective against thromboembolism — the risk-benefit calculus favors continuing anticoagulation with heightened vigilance rather than avoidance.

⚖️ Methods Highlight

Propensity Score Matching in Pharmacoepidemiology

In observational studies of drug safety, patients who receive one treatment are often sicker or older than those who don't, creating confounding by indication. Propensity score matching (PSM) addresses this by balancing groups on observed covariates before comparing outcomes.

A propensity score is the probability of receiving treatment (here: amiodarone) given a set of baseline characteristics.
Patients are then matched 1:1 on this score, creating treated and untreated groups that look similar at baseline.
Standardized mean differences (SMDs) <0.1 after matching indicate adequate balance, a key quality check reported in this study.

PSM cannot account for unmeasured confounders (e.g., frailty, medication adherence), which is why residual confounding by indication remains a stated limitation of this and similar pharmacoepidemiology studies.

Tip: When evaluating a matched cohort study, always check whether SMDs are reported post-match, and whether the authors discuss what unmeasured confounders might still be at play.