← Back
NeuroTrials.ai
Neurology Clinical Trial Database

TIMING

Timing of Oral Anticoagulant Therapy in Acute Ischemic Stroke With Atrial Fibrillation

Year of Publication: 2022

Authors: Jonas Oldgren, Signild Åsberg, Ziad Hijazi, ..., for the National TIMING Collaborators

Journal: Circulation

Citation: Circulation. 2022;146:1056-1066

Link: https://clinicaltrials.gov/ct2/show/NCT02961348


Clinical Question

Is early initiation (≤4 days) of NOAC noninferior to delayed initiation (5-10 days) after acute ischemic stroke in patients with atrial fibrillation, in terms of recurrent ischemic stroke, symptomatic intracerebral hemorrhage, or all-cause mortality?

Bottom Line

Early initiation of NOAC (≤4 days) was noninferior to delayed start (5-10 days) after acute ischemic stroke in patients with atrial fibrillation. No symptomatic intracerebral hemorrhage occurred in either group, and there were numerically lower rates of ischemic stroke and death with early initiation, supporting early NOAC initiation as safe for acute secondary stroke prevention.

Major Points

  • Primary composite outcome occurred in 6.89% (early) vs 8.68% (delayed) (risk difference -1.79%, 95% CI -5.31% to 1.74%, P_noninferiority=0.004)
  • No patient in either group experienced symptomatic intracerebral hemorrhage during 90-day follow-up
  • Ischemic stroke rates were numerically lower with early initiation: 3.11% vs 4.57% (HR 0.67, 95% CI 0.34-1.33)
  • All-cause mortality was numerically lower with early initiation: 4.67% vs 5.71% (HR 0.81, 95% CI 0.45-1.44)
  • Early start was not superior to delayed start (P_superiority=0.38)
  • NOAC was initiated on average at day 3 (66.8 hours) in early group vs day 5 (116.8 hours) in delayed group
  • Major bleeding at 28 days: 7 events (early) vs 3 events (delayed), including 3 asymptomatic hemorrhagic transformations
  • Results were consistent across all prespecified subgroups
  • First randomized controlled study evaluating clinical outcomes of NOAC timing within 10 days post-stroke

Design

Study Type: Registry-based randomized controlled noninferiority trial, open-label, blinded end-point

Randomization: 1

Blinding: Open-label for treatment allocation; blinded end-point evaluation. Patients and physicians aware of treatment group but outcome events verified through medical record review.

Enrollment Period: April 2, 2017 to December 30, 2020

Follow-up Duration: 90 days

Centers: 34

Countries: Sweden

Sample Size: 888

Analysis: Intention-to-treat primary analysis. Noninferiority tested by comparing upper limit of 2-sided 95% CI for absolute risk difference with predefined margin of 3%. Ordinary z test for proportions with Cox proportional hazards as sensitivity analysis. SAS version 9.4. Prespecified subgroup analyses using logistic regression with interaction terms.


Inclusion Criteria

  • Age ≥18 years
  • Atrial fibrillation (paroxysmal, persistent, or permanent; previously known or diagnosed during hospitalization)
  • Acute ischemic stroke within 72 hours of symptom onset
  • Eligible for and willing to start NOAC treatment
  • Patients on prior oral anticoagulation eligible if: NOAC interrupted ≥2 days at index stroke OR INR ≤1.7 if on prior warfarin
  • Control brain imaging performed before eligibility if patient received reperfusion therapy

Exclusion Criteria

  • Contraindication to NOAC therapy (e.g., ongoing bleeding, mechanical heart valve prosthesis)
  • Previous participation in TIMING study

Arms

FieldControlEarly
InterventionNOAC initiation 5-10 days from stroke onset. Choice of NOAC (apixaban, dabigatran, edoxaban, or rivaroxaban) and exact day within time window at treating physician discretion. Mean time to NOAC: 116.8 hours (day 5).NOAC initiation ≤4 days from stroke onset. Choice of NOAC (apixaban, dabigatran, edoxaban, or rivaroxaban) and exact day within time window at treating physician discretion. Mean time to NOAC: 66.8 hours (day 3).
Duration90-day follow-up90-day follow-up

Outcomes

OutcomeTypeControlInterventionHR / OR / RRP-value
Composite of recurrent ischemic stroke, symptomatic intracerebral hemorrhage (≥10mm with NIHSS increase ≥4 points or ≥2 points in one category), or all-cause mortality at 90 daysPrimary8.68% (38/438 patients)6.89% (31/450 patients)0.780.004 (noninferiority); 0.38 (superiority)
Recurrent ischemic stroke at 90 daysSecondary4.57% (20 patients)3.11% (14 patients)0.670.26
Symptomatic intracerebral hemorrhage at 90 daysSecondary0% (0 patients)0% (0 patients)
All-cause mortality at 90 daysSecondary5.71% (25 patients)4.67% (21 patients)0.810.47
Major bleeding at 28 daysSecondary3 events7 events
Per-protocol primary outcomeSecondary7.54% (30/398 patients)6.65% (28/421 patients)0.0384 (noninferiority)
Intracerebral hemorrhage (asymptomatic/hemorrhagic transformation) - Early groupAdverseN/A2 events (at 2 and 17 days; NIHSS change 1 and 0 points)
Intracerebral hemorrhage (asymptomatic/hemorrhagic transformation) - Delayed groupAdverse1 event (at 13 days; NIHSS change 0 points)N/A

Subgroup Analysis

Primary outcome was consistent across all prespecified subgroups including age (18-64, 65-74, 75-84, >84), sex, prior AF status, diabetes, prior stroke/TIA, drugs on admission (antihypertensives, antithrombotics), NIHSS on admission (0-3, 4-5, 6-10, 11-15, >15), and reperfusion therapy. Signs of potential harm with early initiation were implied only in small subgroups of thrombectomy patients (OR 3.79) and NIHSS >15 (OR 2.36), though interaction P-values were not significant. No interaction with NIHSS at time of NOAC initiation.


Criticisms

  • Smaller than preplanned sample size (888 vs target 3000) due to recruitment challenges and COVID-19 pandemic
  • Open-label design for treatment allocation, though end-point evaluation was blinded
  • Lack of brain imaging data (MRI with DWI) to assess lesion size
  • Stroke physicians' preference to start early may have introduced selection bias
  • Enrollment coincided with transition from VKA to NOAC era, with concurrent observational studies supporting early NOAC safety
  • Subgroup findings (potential harm in thrombectomy and NIHSS >15 patients) are hypothesis-generating only
  • Study not powered to demonstrate superiority
  • Registry-based design may have limitations despite strength in consecutive enrollment
  • 17 patients (1.9%) did not start NOAC treatment (4 early, 13 delayed)

Funding

Swedish Research Council (reference 2015-00881)

Based on: TIMING (Circulation, 2022)

Authors: Jonas Oldgren, Signild Åsberg, Ziad Hijazi, ..., for the National TIMING Collaborators

Citation: Circulation. 2022;146:1056-1066

Content summarized and formatted by NeuroTrials.ai.