Future trials on endovascular stroke treatment: the not-so-easy-to-pluck fruits

Abstract

The EXTEND-IA trial showed a positive result in 71 patients (31% absolute effect size for mRS 0-2) [1].Data from the HERMES collaboration showed an NNT of 2.5 for a one point improvement in mRS in patients undergoing endovascular thrombectomy [2]. The DAWN trial showed an absolute difference of 35% for mRS 0-2 between the thrombectomy and control arms [3].These massive effect sizes are very uncommon in any medical field. How did these treatment effects occur? It is likely a combination of factors. First, we did not really expect IMS3 [4] that studied evolving endovascular stroke treatment techniques to show a negative result. It was a neutral study due to a combination of factors. Cherry picking (taking younger, earlier, healthier patients outside of the trial direct to endovascular thrombectomy) may have been a factor. Other factors commonly associated with IMS-3 that contributed to the negative results are as follows: absence of vascular imaging to definitively document the presence of LVO; lack of additional imaging markers such as collaterals or perfusion imaging to exclude patients with poor collaterals or a very large ischemic core; not using modern thrombectomy devices such as stent retrievers, an equal reperfusion rate in both treatment arms; and finally, insufficient focus on efficiency and workflow of the thrombectomy procedure. The results of IMS3 as well as SYNTHESIS Expansion [5] and MR RESCUE [6]) galvanized the stroke community into action, and many trials were then designed and successfully executed. These trials were for the most part designed to overcome the limitations of IMS3 as described above and furthermore, knowingly or unknowingly, selected inclusion/ exclusion criteria so as to conclusively demonstrate a large benefit of endovascular treatment. After the first group of trials, a second group of trials led by DAWN and DEFUSE-3 were conducted to show benefit in a highly focused group of patients (selected by imaging) beyond 6 h from stroke onset. Once again, the effect size of treatment was staggering.With the success of all these trials (MR CLEAN, ESCAPE, EXTEND IA, SWIFT PRIME, REVASCAT, THRACE, DAWN [1, 7-11]), there is the obvious question: are there other sub-populations of patients who would benefit from endovascular thrombectomy that were not sufficiently included in the prior trials? If so, should we use the success and momentum from the successful trials to show definitive proof for these sub-populations as well?

Stroke patients treated by thrombectomy in real life differ from cohorts of the clinical trials: a prospective observational study

“…This is very likely explained by the fact that the suggested ASA/AHA-criteria, in the same way as the inclusion criteria of the RCTs, select those patients with the highest likelihood of treatment benefit. This approach was previously referred to as "cherry picking" of patients of the large thrombectomy trials [20]. While trials in highly selected populations are well suited to provide a clear demonstration of treatment benefit as a proof-of-concept, in clinical practice these results must be transferred to treatment decisions in individual patients with characteristics different from the trial cohorts.…”

What neurointerventionists think about the treatment of unruptured brain arteriovenous malformations: the complexity of moving towards evidence-based treatment

“…2. Conducting a trial focused on the "low-hanging-fruit" [18] in order to minimize the risk of an uninformative trial. This would mean to randomize only patients at low risk of treatment complications and would require a strategy for operators to identify these patients.…”

Trials on ischemic stroke treatment: mission accomplished?

“…In this issue, the editorial by Goyal et al [1] addresses consequences of a recent paradigm shift in ischemic stroke treatment.…”

“…The effect of thrombectomycombined with or without IV thrombolytics-on clinical outcomes is so convincing that Goyal et al [1] question whether similar effects might be possible in patient populations so far excluded from thrombectomy trials. They question whether the definition of Blarge vessel occlusion^should include other arteries in addition to the internal carotid artery and the first segments of the anterior and middle cerebral arteries, whether the indication for thrombectomy can be extended to the posterior circulation, to patients with mild strokes or patients with extended ischemic edema (Alberta Stroke Program Early CT Score, ASPECTS < 6), to the pediatric population, and to patients with late admittance.…”

“…They are concerned, however, that randomized trials on these subpopulations may not show a significant effect, because the effect sizes will be smaller compared to already established indications of thrombectomy. Furthermore, Goyal et al [1] are afraid that study sites having already widened the indication for thrombectomy may only randomize selected patients with questionable prognosis for such trials thus heavily biasing the results. The slow recruitment for the Basilar Artery International Cooperation Study (BASICS)-145 patients since 2011-begs the question how many basilar artery stroke patients have not been randomized, but instead treated with thrombectomy at the 16 study sites [http://www.basicstrial.com/Newsletters_files/newsletter BASICS 2017 july.pdf].…”