Josh Power Endovascular Therapy for Ischemic Stroke with Perfusion-Imaging Selection Campbell BCV et al N Engl J Med 2015; 372:1009-1018 March 12, 2015 DOI: 10.1056/NEJMoa1414792 |
I personally think that the "Thrombolysis Horse" has well-and-truly bolted unfortunately, and that therefore Endovascular treatment of stroke will become much more common-place now that multiple studies support it. And in fact, if the skilled emergency physician, radiologist and neurologist with a CT scanner...and experienced interventionalist were all available when i presented with a dense MCA clot, I'd want it sucked out!
Josh
- Multiple positive trials published simultaneously
- Strongly positive results, in contradiction to thrombolysis trials
- Still significant confounders present, let alone massive financial implications!
Context
Study
- 70 patients enrolled from Australian hospitals over 2 years, 35 in each arm
- Planned enrolment was 100 patients, but results of the MR CLEAN trial triggered a pre-planned interim analysis of results, at which point the study was terminated early
They enrolled patients who:
|
- Randomised to receive either “Standard care” which included aforementioned tPA, or endovascular clot retrieval as well
Endovascular approach involved:
|
- All patients had a repeat CT Angio and Perfusion within 24 hours, and were followed to 3 months.
- Coprimary outcome was restoration of perfusion (as defined by flow of 2b or 3 on the modified Treatment in cerebral ischaemia scale for those interested) and improvement of NIHSS within 24 hours. Secondary outcomes included ordinal analysis of modified Rankin Score change at 3 months, and harm.
Results
- 89% reperfusion rate for "Endovascular" arm vs 34% for the “Standard Care” arm
- Improved NIHSS scores at 24 hrs, and improved modified Rankin scores at 3 months for the intervention arm also
- Utility of these stats attract the same debate as per thrombolysis trials, so I won't detail/reiterate them here
Discussion
1. It seems like a plausible intervention - for that small group of patients with a proximal clot, early presentation, significant new neurological deficit and poor prognosis otherwise, it seems to make sense that fishing out tha clot early will help
2. They are tiny numbers:
- It’s hard to put much weight behind any result when each patient makes up 3 whole percentage points
- Why were so few patients enrolled over such a long period of time? It has to make you wonder, were they super selective to deliberately aim for a strong result? And regardless of the enrolment strategy, surely we can now expect this intervention to be generalised beyond this narrow population.
3. The intervention group had a 4-point higher NIHSS score to start, and this may have exaggerated the treatment effect
4. The crux of the matter as we saw it was really about resource allocation, which unfortunately still centres on thrombolysis. If you had to establish a pathway for this small group of patients from scratch ( prehospital notification > prioritise ED assessment > complex expensive invasive investigation soon after arrival > rapid interpretation by radiology and neurology), all with a view to initiating a very expensive intervention, in aa centre that may deploy this intervention once a week at most, you probably wouldn't be able to justify that. But as we’re already performing all those expensive complex elements of the assessment to determine if thrombolysis should be offered, the only new element is whether you have the resources to staff the cath lab with appropriately skilled people, and can afford the devices(s). Most emergency physicians would argue that there's still no evidence of benefit from thrombolysis, and so adding the option of a very expensive intervention to select subgroup of stroke patients would simply compound this inappropriate use of oru stretches resources.