Thrombolysis Guided by Perfusion Imaging up to 9 Hours after Onset of Stroke (EXTEND)
Source: View publication →
In patients with acute ischemic stroke and salvageable brain tissue identified on perfusion imaging, extending the intravenous alteplase window to 4.5–9 hours after onset (or wake-up stroke) resulted in a higher rate of excellent functional outcome compared to placebo, though with an increased risk of symptomatic intracranial hemorrhage.
Key Findings
Study Design
Study Limitations
Clinical Significance
EXTEND provided Level 1 evidence that systemic intravenous thrombolysis can improve functional outcomes beyond the standard 4.5-hour window—specifically up to 9 hours or in wake-up strokes—when patients are appropriately selected using advanced perfusion imaging to identify a salvageable ischemic penumbra. Along with the WAKE-UP trial, it radically shifted the acute stroke paradigm from strict 'time-is-brain' constraints to a personalized 'tissue-is-brain' approach.
Historical Context
Historically, the therapeutic window for intravenous alteplase in acute ischemic stroke was strictly bounded at 3 hours (NINDS, 1995) and later extended to 4.5 hours (ECASS III, 2008) based primarily on non-contrast CT imaging. Beyond 4.5 hours, trials consistently showed an unacceptable hemorrhagic risk without significant clinical benefit. However, the introduction of endovascular therapy trials such as DAWN and DEFUSE 3 (2018) demonstrated that multimodal perfusion imaging could identify salvageable penumbra up to 24 hours out. EXTEND successfully applied this image-guided, tissue-mismatch concept to the systemic administration of intravenous alteplase for late-presenting and wake-up strokes.
Guided Discussion
High-yield insights from every perspective
How does perfusion imaging differentiate between the ischemic core and the ischemic penumbra, and why is identifying this mismatch crucial for extending the thrombolysis window beyond the traditional 4.5 hours?
Key Response
This question tests the understanding of basic stroke pathophysiology. The ischemic core represents irreversibly infarcted tissue, while the penumbra is hypoperfused but salvageable tissue. Extending the window without advanced imaging risks treating patients who only have a large core, offering them no potential benefit while exposing them to a high risk of fatal hemorrhage.
A patient wakes up with right-sided weakness 8 hours after last known well. According to the EXTEND trial criteria, what specific automated perfusion imaging parameters must be met to consider this patient for IV alteplase, and what primary risk must you discuss with the family?
Key Response
Residents must know the practical inclusion criteria for extended window thrombolysis: a target mismatch profile typically defined as an ischemic core of less than 70 ml, a mismatch ratio greater than 1.2, and an absolute mismatch volume greater than 10 ml. They must also be prepared to counsel on the significantly increased risk of symptomatic intracranial hemorrhage compared to placebo.
The EXTEND trial was terminated early following the publication of the WAKE-UP trial. How does the reliance on perfusion mismatch in EXTEND fundamentally differ from the MRI DWI-FLAIR mismatch used in WAKE-UP, and how does this affect patient eligibility?
Key Response
Fellows need to deeply understand and compare advanced imaging methodologies. WAKE-UP utilized a 'tissue clock' approach where a DWI-positive but FLAIR-negative lesion implies the stroke occurred within 4.5 hours. In contrast, EXTEND utilized a 'tissue viability' approach based on core-penumbra mismatch, which allows treatment up to 9 hours even if the exact onset time is known to be beyond 4.5 hours, thereby expanding eligibility.
Given that the EXTEND trial demonstrated both an increased rate of excellent functional outcomes and a higher rate of symptomatic intracranial hemorrhage, how do you frame the risk-benefit discussion in shared decision-making for a patient in the 4.5-9 hour window?
Key Response
Attendings must synthesize the complex trade-offs of delayed reperfusion into actionable clinical advice. The number needed to treat (NNT) for an excellent functional outcome was around 17, but this is counterbalanced by a tangible number needed to harm (NNH) for symptomatic hemorrhage. This requires highly nuanced communication of competing probabilities compared to the safer, standard early window.
Scholarly Review
Critical appraisal through the lens of expert reviewers and guideline development
EXTEND was stopped prematurely due to loss of equipoise after the WAKE-UP trial results were released. How does early termination for external reasons affect the statistical power and the point estimates of the primary outcome, and what advanced analytical methods can be used to adjust for potential overestimation?
Key Response
Truncated trials often overestimate treatment effects due to capturing random highs in the data. PhD researchers should critically evaluate the implications of early stopping rules, assess the fragility index of the results, and understand the application of Bayesian methods or penalized regressions to adjust for the bias introduced by premature trial termination.
A notable critique of EXTEND is the potential for baseline imbalances due to the relatively small sample size enrolled before the trial was halted. As an editor evaluating this manuscript, what specific supplementary analyses would you demand to ensure the observed functional benefit of alteplase was not driven by baseline confounding?
Key Response
Editors must aggressively look for flaws in prematurely stopped RCTs. Requesting multivariable logistic regression adjusted for baseline NIHSS, initial ischemic core volume, and site of large vessel occlusion is essential to validate that the primary endpoint was truly driven by the alteplase intervention and not by a favorable baseline imbalance in the treatment arm.
Based on the findings of EXTEND and similar delayed-window trials, how should current AHA/ASA guidelines for acute ischemic stroke be updated regarding the recommendation class and level of evidence for administering IV alteplase between 4.5 and 9 hours in patients who are not candidates for mechanical thrombectomy?
Key Response
Guidelines previously drew a strict temporal line at 4.5 hours for IV tPA. The committee must evaluate whether EXTEND provides sufficient Level B-R (Randomized) evidence to issue a new Class IIa recommendation for tPA in the extended window using automated perfusion imaging, carefully weighing the functional benefits against the increased bleeding risks.
Clinical Landscape
Noteworthy Related Trials
ECASS III Trial
Tested
Intravenous alteplase
Population
Acute ischemic stroke patients presenting 3 to 4.5 hours after onset
Comparator
Placebo
Endpoint
Favorable outcome (mRS 0-1) at 90 days
WAKE-UP Trial
Tested
Intravenous alteplase
Population
Acute ischemic stroke of unknown onset time with MRI DWI-FLAIR mismatch
Comparator
Placebo
Endpoint
Favorable outcome (mRS 0-1) at 90 days
DEFUSE 3 Trial
Tested
Endovascular thrombectomy plus standard medical therapy
Population
Acute ischemic stroke 6 to 16 hours after onset with salvageable tissue on perfusion imaging
Comparator
Standard medical therapy alone
Endpoint
Functional outcome (ordinal shift in mRS score) at 90 days
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