What is the clinical evidence from this study?

Study design: Cohort. Population: Nonvalvular atrial fibrillation (NVAF) (n=116889). Intervention: Direct oral anticoagulants (DOACs). Primary outcome: Intracerebral hemorrhage (ICH) (OR 21.7, 95% CI 17.0-27.6, p=<0.001).

What question did this study set out to answer?

This study aims to evaluate the incidence and predictors of intracerebral hemorrhage (ICH) in nonvalvular atrial fibrillation patients treated with direct oral anticoagulants.

May 8, 2026Open Access

Abstract Number: Esoc2026a1508 Early and Late Intracerebral Hemorrhage in Direct Oral Anticoagulant–treated Nonvalvular Atrial Fibrillation: Incidence and Predictors

Q: What are the key findings of this study?

In DOAC-treated patients with nonvalvular atrial fibrillation, intracerebral hemorrhage occurred in 3.81%, with prior ICH strongly predicting early ICH (OR 21.7; 95% CI 17.0-27.6; p<0.001).

Q: What question did this study set out to answer?

This study aims to evaluate the incidence and predictors of intracerebral hemorrhage (ICH) in nonvalvular atrial fibrillation patients treated with direct oral anticoagulants.

Key Result

In DOAC-treated patients with nonvalvular atrial fibrillation, intracerebral hemorrhage occurred in 3.81%, with prior ICH strongly predicting early ICH (OR 21.7; 95% CI 17.0-27.6; p<0.001).

Key Points

This study aims to evaluate the incidence and predictors of intracerebral hemorrhage (ICH) in nonvalvular atrial fibrillation patients treated with direct oral anticoagulants.
Conducted a retrospective cohort study using electronic health records from a health-maintenance-organization.
Review involved adult nonvalvular atrial fibrillation patients initiating DOAC therapy between 2013-2024.
Predictors of Early-ICH and early subarachnoid hemorrhage were assessed using logistic regression, while Late-ICH predictors were analyzed using multivariable Cox regression.
ICH occurred in 4,455 patients (3.81%), with 514 classified as Early-ICH and 3,941 as Late-ICH.
Independent predictors of Early-ICH included prior ICH (OR 21.7, 95% CI 17.0–27.6) and prior ischemic stroke/TIA (OR 1.66, 95% CI 1.23–2.19).
For Late-ICH, higher baseline INR increased risk (HR 1.23, 95% CI 1.02–1.48, p=0.03) and older age at initiation also increased risk (HR 1.01 per year, 95% CI 1.01–1.04, p<0.001).

Study Design

Type

Cohort (n=116,889)

Structured PICO

Population

116,889 adult patients with nonvalvular atrial fibrillation (NVAF) initiating DOAC therapy, 51.1% female

Intervention

Direct oral anticoagulants (DOACs)

Outcome

Incidence and predictors of Early-ICH (all ICH types within 90 days of initiation) and Late-ICH (beyond 90 days)safety

Early and late intracerebral hemorrhage in DOAC-treated NVAF patients have distinct risk profiles, highlighting the importance of individualized risk stratification.

Main Result

Effect estimate: OR 21.7 (95% CI 17.0-27.6)

p-value: p=<0.001

Abstract

Abstract Background and aims Although uncommon, intracerebral haemorrhage (ICH) may occur in nonvalvular atrial fibrillation (NVAF) patients treated with direct oral anticoagulants (DOACs). This study aimed to assess the incidence and predictors of ICH in this population. Methods We conducted a retrospective cohort study using electronic health records from a health-maintenance-organization. Adult NVAF patients initiating DOAC therapy between 2013-2024 were reviewed for ICH occurrence during ongoing DOAC treatment. Predictors of Early-ICH (all ICH types within 90 days of initiation) and early subarachnoid haemorrhage (early-SAH) were assessed using logistic regression, while predictors of Late-ICH (beyond 90 days) were evaluated using multivariable Cox regression. Results A total of 116,889 patients (51.1% female) contributed 534,183 patient-years of DOAC exposure. ICH occurred in 4,455 patients (3.81%), with 514 (0.44%) classified as Early-ICH and 3,941 (3.37%) as Late-ICH. Independent predictors of Early-ICH included prior ICH (OR 21.7, 95% CI 17.0–27.6), prior ischemic stroke/TIA (OR 1.66, 95% CI 1.23–2.19), antiepileptic drug use (OR 1.46, 95% CI 1.14–1.85), and male sex (OR 1.45, 95% CI 1.19–1.77) (all p0.001). Predictors for early-SAH included previous ICH (OR 33, 95%CI 20-55) and baseline creatinine (OR 1.3, 95% CI 1-1.6). For Late-ICH, higher baseline INR (HR 1.23, 95% CI 1.02–1.48, p=0.03) and older age at initiation (HR 1.01 per year, 95% CI 1.01–1.04, p0.001) increased risk, whereas prior warfarin use was protective (HR 0.58, 95% CI 0.48–0.72, p0.001). Conclusions Early and Late ICH in DOAC-treated NVAF patients have distinct risk profiles, highlighting the importance of individualized risk stratification. Conflict of interest Asaf Honig,Ruth Smadar, Gal Ifergane, Jeremy Molad: Nothing to disclose