Blog Viewer

NEWS: DEFUSE 3 Non-DAWN Patients: A Closer Look at Late Window Thrombectomy Selection

By Currents Editor posted 04-04-2019 12:47


Authors: Leslie-Mazwi TM, Hamilton S, et al. DEFUSE 3 Non-DAWN Patients: A Closer Look at Late Window Thrombectomy Selection. Stroke. 2019 Feb 6:STROKEAHA-118.

Reviewed by: Wazim Mohamed, MD, Assistant Professor of Neurology, Division of Neurotrauma and Critical Care Neurology, Wayne State University

Read the article.*

*You will need to log in to US National Library of Medicine National Institutes of Health to read this article. 


The 2018 American Heart Association guidelines for the management of acute ischemic stroke recommend mechanical thrombectomy for up to 24 hours based on the recently published DEFUSE 3 and DAWN trials.

These trials demonstrated good functional outcomes in carefully selected patients beyond the six-hour time window. The DEFUSE 3 trial time window was from six to 16 hours and included patients with a NIHSS > 6 baseline-modified Rankin scale (mRS) ≤ 2, pre-treatment core infarct volume ≤ 70 ml and substantial penumbra volumes (mismatch ratio is >1.8 and mismatch volume is >15 ml).

The DAWN trial included three groups of patients in the six- to 24-hour time window, but the study was less inclusive with a pre-treatment core infarct threshold of 50 ml based on patient age and NIHSS (A: ³80 y/o, NIHSS ³10 + core < 21 ml; B: <80 y/o, NIHSS ³10 + core < 31 ml; C: <80 y/o, NIHSS  ³ 20 + core < 51 ml). In addition, DAWN only included patients with baseline mRS 0-1.

The authors of the current paper hypothesized that a favorable treatment effect would remain for DEFUSE 3 patients who would have been excluded from the DAWN trial (DEFUSE 3 non-DAWN patients). Three subgroups of patients were identified: infarct core too large (CTL), NIHSS too low (presentation scores of 6–9) or baseline mRS of 2 at presentation.

There were 13 patients in the DEFUSE 3 trial with a baseline mRS of 2, which was inadequate to elicit a treatment effect, hence these patients were categorized based on their NIHSS. Of these, three patients had an NIHSS 6-9 and were included in the NIHSS too low group. Analysis was done for CTL group versus core not too large (CNTL) group and for NIHSS 6-9 group versus NIHSS ³ 10 group.

These subgroups were compared with the remainder of the DAWN ineligible DEFUSE 3 cohort, as well as the entire DEFUSE 3 cohort. The primary outcome for this study was a good functional outcome (mRS 0-2) at 90 days.


DEFUSE 3 was terminated early for efficacy by the DSMB after the randomization of 182 patients. Seventy-one patients met criteria as DEFUSE 3 non-DAWN eligible patients. The reasons for ineligibility were as follows: 31 patients with NIHSS too low, 33 patients with CTL and 13 patients with baseline mRS of 2 at presentation (this group was categorized based on their NIHSS as explained in methods, with three patients included in the NIHSS 6-9 group). 

CTL patients: 42% of patients in the CTL group were > 79 years compared to 21% in the CNTL group (p=0.016). Median ASPECTS for the CTL group was 7 compared to 8 for the CNTL group (p=0.02) with no difference in the level of vascular occlusion. As expected, the ischemic core volume was higher in the CTL group (45 ml [IQR 37.6-60.4] vs. 7.3 ml [IQR 0-13.9]; p<0.001), as well as the measured volume of the perfusion lesion (145.8 mL [IQR, 127.3–184.7] vs. 98.7 mL [IQR, 67.5–144.4]; p<0.001). A similar proportion of patients from each group were randomized for intervention and achieved target reperfusion. There were no differences in achieving the primary end point (mRS 0-2 at 90 days) between both groups: 24% in the CTL group and 32% in the CTNL group. The rates of death and symptomatic hemorrhage were also similar among both groups, even with higher infarct growth (89.3 ml [IQR 39.4-135.8] vs. 23.8 ml [IQR 11.7-51.1]; p<0.001) and median infarct volume at 24 hours (119 ml [IQR 74.6-180] vs. 31.5 ml [IQR 17.6-64.3]; p<0.001) in the CTL group. Within the CTL group, when compared to medical management, endovascular therapy conferred a better functional outcome (OR 20.9 [95% CI 1.3-337.8]) after correcting for age and NIHSS.

NIHSS too low (6-9) patients: This group was compared with the entire DEFUSE 3 cohort with NIHSS ³10. The median NIHSS was 8 (IQR 7-9) in this group compared to 18 (IQR 14-21) in the high NIHSS cohort. There were no differences in age, sex, administration of tPA or randomization to endovascular therapy between the groups. There were no differences between the groups with respect to timing and procedures. The reperfusion rates were also similar between the groups. The primary end point was more likely in the low NIHSS group (74% vs. 22%; p<0.001). There were also fewer deaths in the low NIHSS group (6% vs. 23%; p=0.02). For the low NIHSS group, endovascular therapy offered a better chance at functional outcome compared to medical management (OR 1.88 [95% CI: 0.36-9.53]), although the results were not statistically significant in this small subgroup.


The authors of the current paper conclude that core infarct volumes up to 70 ml (but too large to be included under the DAWN criteria) may still benefit from thrombectomy. There was also a trend toward better outcomes in the thrombectomy group for patients with a low NIHSS (6-9).

This study is of importance in further understanding thrombectomy treatment effects beyond the six-hour time window and helps to expand candidacy for endovascular treatment. While previous trials have demonstrated poor functional outcome with larger pretreatment core infarct size, a subgroup analysis of the MR CLEAN study suggested that a functional benefit may still be obtained in patients with ASPECTS 5-7. The current analysis demonstrates functional benefit from thrombectomy for CTL patients (more than 40% with age >80 years), despite this group possibly having less robust collaterals leading to significantly higher infarct growth size. In addition, there were no additional complications with reperfusion therapy when compared to the CNTL group. This suggests a potential for including a higher pre-treatment core infarct threshold in future studies.

The study also showed trends toward a benefit for patients with a low NIHSS at later timepoints as well. The limitation of the current study encompasses those that are usually related to subgroup analyses. The sample size of the current analyses was limited by the early termination of DEFUSE 3. Most importantly, the current study is unable to comment on the treatment effect for patients within the 16- to 24-hour time window. The findings from this study support efforts to expand thrombectomy candidacy in future studies.


By Kristi Tucker, MD   Trainees are taught how to communicate with patients and how to behave professionally during patient interactions, but sadly there is much less emphasis placed on effective communication with medical colleagues.  In residency programs, there is little, if any, formal instruction ...
To most, Sunday, April 7, 2019, does not stir up strong emotions or even spark a distant memory. The most popular tweet of the day was “Keep your emotions off the internet” from NFL player JuJu Smith-Schuster; Old Town Road by Lil Nas X jumped 14 spots to No. 1 on the Billboard Hot ...
Amarenco P, Kim JS, Labreuche H, et al.  A comparison of two LDL cholesterol targets after ischemic stroke.  NEJM 2020; 382:9-19.   Reviewed by Kyle Hobbs, MD Read the article .   Summary :  Intensive lipid lowering therapy is recommended after ischemic stroke or TIA, but the exact LDL ...