Team Endovascular Extends Winning Streak to 4

Age Range≥ 18≥ 18≥ 1818 - 8018 - 85
NIHSS Range≥ 2> 5None stated8 - 29≥ 6
Time Window6 hours12 hours6 hours6 hours8 hours
IV tPA Treatment89%76%100%100%73%
Vascular LesionsDistal ICA, M1, M2, A1, or A2Carotid T, Carotid L, M1, or all M2sICA, M1, or M2Intracranial ICA, carotid terminus, or M1ICA or M1
Penumbral AssessmentNoneASPECTS ≥6

Moderate-good collaterals
on Multiphase CTA
CT or MR perfusion mismatch
(see paper for definition)
Hypodensity or DWI
< 1/3 of MCA territory

Primary Outcome(s)Shift in mRS at 90 daysShift in mRS at 90 daysReduction in perfusion lesion volume

NIHSS reduction ≥ 8 points
or to score of 0-1 at day 3
Distribution of mRS at 90 days

% mRS 0-2 at 90 days
Shift in mRS at 90 days
Primary Result(s)Odds ratio 1.67Odds ratio 2.6100% vs. 37% reperfusion

80% vs. 37% NIHSS
Cochran-Mantel-Haenszel p-value = 0.0002Odds ratio 1.7
mRS 0-2 at 90 days33% vs. 19%53% vs. 29%71% vs. 40%60% vs. 36%
(co-primary outcome)
44% vs. 28%
sICH Risk7.7% vs. 6.4%3.6% vs. 2.7%5.7% vs. 5.7%
(NINDS definition)
1% vs. 3.1%
4.9% vs. 1.9%
Mortality18.9% vs. 18.4%10.4% vs. 19.0%8.6% vs. 20%
Absolute values not reported
3.7% absolute difference in favor
of endovascular
18.4% vs. 15.5%

Lots of exciting news coming from Nashville, as expected. In addition to MR CLEAN, we now have positive results from ESCAPE, EXTEND-IA, and SWIFT PRIME. I thought it might be useful to line up the key clinical and radiographic inclusion criteria and outcomes in a table. I’m new to table-making, so I’ll try to adjust the column widths if I get time later; I think the basic idea comes through. Update: Here’s a link to a full-screen picture of the table.

A few observations: First, all three of the published papers (the SWIFT PRIME study isn’t published yet–the link above is to the slides from the oral abstract session) included links to their protocols and supplementary data sets. Kudos!

Next, I thought it interesting that MR CLEAN included subjects with NIHSS down to 2. Looking through the paper and the supplementary information, I found subgroup analyses at the higher NIHSS ranges, but nothing about the very low range (e.g., how many subjects actually had NIHSS of 2, or 2-5? Was endovascular treatment beneficial for them?) This is important because from time to time we do see patients with seemingly minor strokes and large vessel occlusions. We’re then faced with the difficult decision whether to offer them revascularization right away, or wait to see if they deteriorate. Note that struggling with this second order question is quite a luxury, born of the fact that the basic principles of endovascular treatment are quickly becoming established.

I think the most heterogenous (and therefore interesting) aspect of these studies is their assessment of tissue viability (penumbra). As far as I can tell, MR CLEAN didn’t require such. However, the manuscript does state that they employed the uncertainty principle, meaning that the local investigators enrolled subjects if they felt uncertain as to whether endovascular treatment would benefit them. Presumably, these investigators would have excluded potential subjects who had low ASPECT scores, or who had perfusion imaging suggesting large infarct cores. Employing the uncertainty principle is not ideal, as it assumes that investigators know who clearly will and who clearly won’t benefit from treatment, which is, in fact, the very question under study. My guess is that having some wiggle room regarding subject selection was necessary in order to get everyone on board with the randomized trial.

ESCAPE employed two means of penumbral assessment: ASPECTS and multiphase CT perfusion. The former assesses whether the parenchyma is irreversibly damaged using CT hypodensity. The latter is relatively new to the literature and assesses the degree of collateral blood flow by visual inspection of the CTA at several time points.

EXTEND-IA required CT- or MR perfusion imaging, using prolonged mean transit time to identify the ischemic tissue and reduced cerebral blood volume to identify the irreversibly damaged tissue.

SWIFT PRIME started out by excluding potential subjects with large areas of hypodensity on CT or high signal intensity on DWI. Later on, they added a requirement that the ASPECT score be at least 6.

So, on the one hand the studies used varying methods to exclude subjects unlikely to benefit from revascularization. On the other hand, all of the studies were positive, so perhaps the specific method doesn’t matter so long as those with predicted large infarct cores are excluded. If it is true that there are several ways to ascertain the extent of salvagable vs. non-salvagable brain, then perhaps the simplest one should prevail. Assigning an ASPECT score and assessing collateral flow with multiphase CT are pretty straightforward (this website has training materials), whereas post-processing of CT perfusion data takes some time, and MRI is always inconvenient.

Anyway, it’s been a great week for stroke research and will be very interesting to see how systems of care develop in response to these results. Here’s a link to the International Stroke Conference’s “Late-Breaking Science” web page, which in turn has links to many other interesting studies presented this year.

Update 2/13/15: Table edited to reflect ESCAPE’s inclusion of subjects with NIHSS > 5.

Update 5/14/15: Table updated with REVASCAT data.

About Justin A. Sattin

I'm a vascular neurologist and residency program director. I created this blog in order to share some thoughts with my resident and other colleagues, and to foster my own learning as well.
This entry was posted in Medical Knowledge and tagged , . Bookmark the permalink.

6 Responses to Team Endovascular Extends Winning Streak to 4

  1. Dave Weisman says:

    This is a terrific summary. Covidien reps are popping bottles this weekend. The question, I think, is if this will change anything? IMSIII failed because: 1. Everyone already thought they knew the answer (the most classic last line in all of medicine), so put bad cases into the trial. 2. Few people rushed and the onset to recanalization times were slow. 3. Many people didn’t have clots on angio, their disease was not affected, and if they had been assigned to placebo then they enriched the placebo groups with better outcomes.

    And what changed after IMSIII? Some of the more crazy cases, at 5 hours say, already showing early ischemia in a 86 year old no longer got taken to the cath lab. But I bet the overall numbers barely budged and in retrospect, centers were probably taking the cases that would benefit – prospectively and not enrolling into a trial.

    With these + studies, I think we’ve established boundaries. The question: who is going to benefit from surgical recanalization has been answered. The question: who will NOT benefit had already been answered. So now we have a field of play.

    The admins who believed that a sizable number of stroke patients would flood their cath labs with cash flow are wrong. The cases suitable for stroke treatment is stenotic.

    And once again, it shows that fools rush in. It shows that when you make a specific prediction in medicine, “I know the people who will benefit, you have to do it like this,” you’re likely wrong. You should also be shamed and laughed at, but if you say such things, you’re already a full professor or chairman and developed a cult of personality around you.

    • Justin A. Sattin says:

      Thanks, David. I agree that a big problem for years was that we “knew” that endovascular treatment was beneficial and so had a hard time getting high-quality evidence. I too continued to offer patients endovascular therapy despite the negative IMS III and MR RESCUE trials, using the same rationale that I used before those were published: 1) The treatment makes biological sense; 2) I’ve seen some patients derive great benefit; 3) I’ve seen many patients neither helped nor harmed; 4) I’ve seen relatively few harmed; 4) To me and to the families, it seems absurd to do nothing when a 52 year old man has a left MCA syndrome, is still early on, and hasn’t responded to IV tPA.

      The key thing, in my view, was always to present endovascular therapy to patients and families as a promising, but unproven, therapy. I was very explicit that adding IA therapy to IV tPA was experimental. This was a controversial stance in some quarters, but I stand by it. The fact that we (as a community) randomized subjects to IV alone vs. IV + IA therapy is irrefutable evidence that equipoise existed and that the combination treatment was appropriately presented at that time as experimental, regardless of whether it was being offered in the research context or not. Now that we have a developing avalanche of positive results, we can now offer IA and IV+IA therapies as scientifically proven. I’m sure that the next iteration of the relevant clinical practice guidelines will all upgrade the level of evidence for such treatments.

      An interesting ethical dimension to this is that I was very hesitant to send patients to endovascular therapy, especially after IV tPA, if consent couldn’t be obtained. Whereas the AAN has a position statement that written consent isn’t required for on-label IV tPA treatment, that doesn’t apply to IA. I think with the improved evidence base, one can now ethically support the practice of sending appropriate patients for endovascular therapy on the basis that most people would want to avail themselves of a proven treatment if they found themselves in a life-threatening situation and weren’t able to consent for themselves. This is similar to the situation of a trauma patient being taken to the ED and being found to need emergency surgery. It isn’t withheld when consent can’t be obtained from patient or surrogate–life saving intervention is offered regardless.

    • Justin A. Sattin says:

      A few additional comments regarding “boundaries”. I think there are two boundary issues that need to be explored. As I mentioned in the original post, one is the situation where a patient has a large vessel occlusion but mild deficits. Just as the question of “who is too mild for IV tPA” remains a thorny one, so too is this likely to be a challenge. I predict that with these positive results, many more people will be getting CT angiograms and we’ll be finding more of these patients. That’s why I’m very interested to know how many subjects with low NIHSS scores were actually enrolled in MR CLEAN.

      Another boundary issue concerns the time windows and use of penumbral imaging. Whereas the other trials used 6 hour windows, ESCAPE used a 12 hour window and thus provides some support for the proposition that, as some of our colleagues put it, the “tissue clock” is more relevant than the “ticking clock”. On the other hand, the median time from stroke onset to reperfusion in ESCAPE was 4 hours and other analyses have shown that reperfusion beyond ~ 6 hours is unlikely to be beneficial. It would be very interesting to conduct a trial of endovascular therapy vs. conservative care for subjects who are 6-24 hours out and whose ASPECT scores, CT perfusion, etc. still look favorable. I would stratify the enrollments such that those 6-12, 12-18, and 18-24 hours out were balanced between IA and conservative care and could be analyzed separately.

  2. dave weisman says:

    When you say “a promising, but unproven, therapy,” the families hear “a promising therapy which will cure them.”

    We had very few wake up stroke with ESCAPE favorable imaging. But then, we had very few ESCAPE favorable patients. I think we had about 8 in about a year, because of the inclusion/exclusion boundaries. ESCAPE’s mantra: regardless of timing, small core ischemia, fast CT to recanalization times, means win. They put a w up.

    We were able to put ESCAPE into place within our clinical scheme very easily. Mod to severe stroke: let CTA sort them into medical vs. ESCAPE (medical plus +/- surgical).

    • Ahsan Sattar says:

      Thanks Dr. Sattin for this outstanding summary. Regarding the question about no use of penumbral imaging for MR Clean trial; In the discussion section of Extend-IA, the authors mentioned that 65% of MR Clean patients did get CT perfusion. However it was not part of the study protocol. Hence it might have affected the decision making of the neurointerventionalists. Hence some type of penumbral tissue assessment is warranted at this point. DAWN trial is currently enrolling patients for wake up strokes and will hopefully answer the question regarding IA therapy beyond 6-8 hours. At hour centre we follow the tissue clock mantra. Everybody gets CTP including wake up strokes and if the core is small and there is a clinical-perfusion mismatch they are taken for IA therapy.

      • Justin A. Sattin says:

        I think your comment proves my point, Ahsan. Many subjects in MR CLEAN did undergo CTP. Some did not. I’m willing to bet that different centers used slightly different post-processing algorithms. Thus, we just don’t know yet whether we should be selecting endovascular patients on the basis of CTP, and if so, using what post-processing algorithm. Maybe we should be using ASPECT scores and multiphase CTA, as was done in ESCAPE. Maybe we should be using DWI / MRP.

        I agree that eventually, we’ll likely be using some combination of time and physiologic imaging. We simply need more clinical trial data to settle this important issue.

Comments are closed.