Extending the Time for Thrombolysis in Emergency Neurological Deficits (EXTEND)
Source: View publication →
In patients with acute ischemic stroke presenting 4.5 to 9 hours after symptom onset or upon awakening, imaging-based selection of salvageable brain tissue using alteplase compared to placebo resulted in higher rates of functional independence at 90 days, albeit with an increased incidence of symptomatic intracranial hemorrhage.
Key Findings
Study Design
Study Limitations
Clinical Significance
The EXTEND trial provides evidence that the therapeutic window for intravenous thrombolysis in acute ischemic stroke can be safely and effectively extended beyond 4.5 hours when utilizing advanced perfusion imaging to identify patients with salvageable penumbral tissue. This finding supports the paradigm shift toward physiological imaging-guided patient selection rather than reliance on strict time-based thresholds, particularly in environments where endovascular thrombectomy is not readily available.
Historical Context
For decades, the standard of care for intravenous thrombolysis was strictly limited to 4.5 hours from symptom onset based on early trials using non-contrast CT. Following the success of endovascular thrombectomy in the late-window (DAWN, DEFUSE 3) and the emergence of perfusion-based patient selection, the EXTEND trial sought to determine if this physiologic selection could similarly expand the benefits of intravenous alteplase, a hypothesis confirmed by the results published in 2019.
Guided Discussion
High-yield insights from every perspective
The EXTEND trial utilizes the concept of the 'ischemic penumbra' to select patients. Pathophysiologically, what characterizes the penumbra compared to the ischemic core, and why is perfusion imaging necessary to identify it beyond the standard 4.5-hour window?
Key Response
The ischemic core represents irreversibly damaged tissue (often defined by a cerebral blood flow <30% of normal), while the penumbra is 'at-risk' tissue that is hypoperfused but still metabolically viable. Beyond 4.5 hours, the 'clock' is an unreliable proxy for tissue status due to individual variations in collateral circulation. Perfusion imaging (CTP or MRP) allows clinicians to move from a 'time-is-brain' to a 'tissue-is-brain' paradigm by identifying a mismatch—where the area of delayed perfusion (Tmax > 6.0s) is significantly larger than the core—indicating salvageable tissue.
A 72-year-old patient presents 6 hours after symptom onset with an NIHSS of 12. Imaging shows a 15ml core and a 100ml perfusion defect. Based on the EXTEND trial, how should you counsel the family regarding the risks and benefits of IV alteplase in this late window?
Key Response
According to EXTEND, treatment in the 4.5-9 hour window with a perfusion mismatch leads to a higher rate of functional independence (mRS 0-1) at 90 days (35.4% vs. 22.5% for placebo; adjusted risk ratio 1.44). However, the resident must also disclose a significantly higher risk of symptomatic intracranial hemorrhage (6.2% in the alteplase group vs. 0.9% in the placebo group). The 'number needed to treat' for a favorable outcome is approximately 8, while the 'number needed to harm' for sICH is approximately 19.
The EXTEND trial included patients with large-vessel occlusions (LVOs) as well as those without. How do the results of EXTEND integrate with the DAWN and DEFUSE 3 trials, and should a patient eligible for mechanical thrombectomy (EVT) in the 6-24 hour window also receive IV alteplase if they meet EXTEND criteria?
Key Response
DAWN and DEFUSE 3 established EVT as the standard for LVOs in late windows. EXTEND provides evidence for IV thrombolysis in the same window for those with mismatch. While many patients in EXTEND had LVOs (approx. 70%), the question of 'bridging' in the late window remains nuanced. Current guidelines (AHA/ASA 2019/2021) suggest that while EVT is the priority for LVOs, IV alteplase can be considered in the 4.5-9 hour window if mismatch criteria are met, though data specifically comparing 'bridging vs direct EVT' in the late window is less robust than in the early window.
EXTEND and the WAKE-UP trial both addressed patients with unknown time of onset, but used different imaging selection tools. How does the use of CT perfusion/MRI perfusion (EXTEND) vs. DWI-FLAIR mismatch (WAKE-UP) impact your clinical workflow and patient selection for late-window thrombolysis?
Key Response
WAKE-UP used DWI-FLAIR mismatch as a proxy for 'time' (suggesting the stroke is <4.5 hours old), whereas EXTEND used perfusion mismatch as a proxy for 'salvageable tissue' regardless of the biological clock. As an attending, the choice often depends on institutional resources. EXTEND allows for a broader window (up to 9 hours) and uses quantitative software (like RAPID) which may be more objective than a radiologist's visual assessment of FLAIR signal, though it requires more advanced post-processing capabilities.
Scholarly Review
Critical appraisal through the lens of expert reviewers and guideline development
The EXTEND trial was terminated early after the publication of the WAKE-UP trial results. From a statistical and methodological standpoint, what are the primary risks associated with early termination in this context, and how does it affect the trial's Fragility Index?
Key Response
Early termination for 'loss of equipoise' or 'external results' often leads to an overestimation of the treatment effect size and a decrease in statistical power. With only 225 of the planned 310 patients enrolled, the Fragility Index of the primary outcome is likely low, meaning only a few status changes from 'favorable' to 'unfavorable' in the treatment group would render the results non-significant (p > 0.05). This necessitates caution when generalizing the absolute benefit across more diverse populations.
As a reviewer, how would you evaluate the potential bias introduced by the use of the proprietary RAPID software for automated imaging analysis in the EXTEND trial, particularly concerning the reproducibility of the results in centers using different post-processing algorithms?
Key Response
A critical reviewer would flag 'software dependency' as a threat to external validity. Different automated platforms (e.g., RAPID vs. Viz.ai vs. Olea) utilize different thresholds and smoothing algorithms to calculate Tmax and core volumes. If the trial's success is predicated on specific RAPID thresholds (Tmax > 6s, core < 70ml), the results might not be reproducible using other software, potentially creating a 'black box' effect where the clinical decision is tied to a specific commercial product rather than a universal physiological principle.
Based on the EXTEND trial, should the AHA/ASA guidelines be updated to change the recommendation level for IV alteplase in the 4.5-9 hour window? Compare your recommendation to the 2018/2019 standards.
Key Response
The 2018 AHA/ASA guidelines initially gave IV alteplase in the 4.5-9h window a Class III (No Benefit) recommendation. Following EXTEND and a subsequent meta-analysis (EXTEND, ECASS-4, and EPITHET), the 2019 Update changed this to a Class IIa (Level of Evidence B-R) for patients with a CT/MRI perfusion mismatch. The committee must balance the moderate quality of evidence (due to early trial termination) against the significant benefit seen in functionally independent outcomes, while emphasizing that this is only for 'imaging-selected' patients, not a blanket time-extension.
Clinical Landscape
Noteworthy Related Trials
ECASS III Trial
Tested
Intravenous alteplase
Population
Patients with acute ischemic stroke
Comparator
Placebo
Endpoint
Disability at 90 days (mRS score 0-1)
DEFUSE 3 Trial
Tested
Endovascular thrombectomy
Population
Acute ischemic stroke patients with a perfusion deficit and small infarct core
Comparator
Standard medical therapy
Endpoint
Functional outcome at 90 days (mRS distribution)
DAWN Trial
Tested
Endovascular thrombectomy
Population
Patients with acute ischemic stroke and a mismatch between stroke severity and infarct volume
Comparator
Standard medical therapy
Endpoint
Functional outcome at 90 days (disability score)
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