What We Don’t Know

Aaron Carroll of The Incidental Economist recently blogged about an announcement from the NIH that the Look AHEAD trial has been halted early for futility regarding its primary endpoint. This is a large, long-term study of whether an intensive lifestyle modification program reduces the risk of cardiovascular events in type 2 diabetics.

Not surprisingly, many salutary effects of this intervention have been published over the years, including improvements in sleep apnea, reduction in diabetes medication requirements, and quality of life. Quite surprisingly, the intervention did not reduce the number of cardiovascular events.

Now, the full study results aren’t published yet and maybe a critical analysis will raise the possibility of a type II error. But what I’d like to pivot off of today is this passage from Dr. Carroll:

A larger point is that we don’t actually know the stuff we think we know. I would have assumed that weight loss would work here. It doesn’t. It took millions of dollars, and years of work to find that out. Research is hard and expensive. We need to keep funding it, and doing it.

Emphasis mine. This point comes up so often on the stroke service that I sometimes feel like a broken record when restating it. (Luckily, each July brings a fresh crop of PG2s for us  to indoctrinate 🙂 ) The list of things we don’t know in stroke is long, but here’s a partial one:

  • Is IV tPA beneficial in the 3-4.5 hour window for those excluded from the ECASS 3 study? (Age > 80, diabetics with prior stroke, any warfarin use, etc.)
  • After a patient has received IV tPA, does endovascular therapy for persistent large vessel occlusion confer benefit?
  • Does perfusion imaging select patients likely to benefit from acute endovascular recanalization?
  • Should hemicraniectomy really be restricted to patients under age 60? (Journal club next week).
  • Which is better for dissection–anticoagulation or antiplatelet therapy?
  • Is dual antiplatelet therapy the preferred regimen for secondary prevention in patients with intracranial atherosclerosis?
  • When a patient has had a stroke due to atrial fibrillation, how soon should anticoagulation be started or restarted? Does it matter how large the infarct is?

The point that bears emphasis isn’t just that there are lots of unanswered questions–life is full of unanswered questions–it’s that we sometimes claim to know the answers to these questions when in fact we don’t. Flecainide was thought to be be beneficial for preventing malignant dysrhythmias in acute MI, but actually causes increased mortality (in part from arrhythmia!)  Carotid bypass surgery used to be widely performed to prevent stroke in those with carotid occlusion. Then a randomized study showed it to be ineffective. Then the hypothesis arose that perhaps selecting only those with abnormal cerebral perfusion would show the surgery to be beneficial, but the COSS study failed to show that either.

Here’s a contemporary example: The MR RESCUE study is designed to answer two of the pesky questions above: 1) Is acute endovascular intervention beneficial for ischemic stroke? and 2) Does perfusion imaging select those patients likely to benefit? It’s a very good design whereby perfusion software installed on the CT or MR console automatically determines whether a “mismatch” pattern is present. The subject is then randomized to intervention or non-intervention in a stratified manner–those with mismatch have a 50:50 chance of intervention and those without mismatch also have a 50:50 chance.

Regrettably, the study didn’t recruit as briskly as hoped, in part because there are many in the neuroendovascular community and even some vascular neurologists who believe that they already know the answers. They insist that it is unethical to randomize some of these patients to non-intervention (and unethical to perform the intervention on others who have what is currently believed to be an unfavorable perfusion pattern).

Sometimes the parachute study is invoked. After all, just as we don’t need an RCT to show that parachutes prevent death when jumping out of airplanes, so too we don’t need an RCT to show that in patients presenting within 9 hours of ischemic stroke onset and who have a perfusion mismatch pattern defined as an area of MTT prolongation ≥ 50% larger than the corresponding DWI lesion, endovascular recanalization with any combination of IA tPA or the Merci, Penumbra, or Solitaire devices is superior to non-interventional care as measured by modified Rankin scale score at 90 days.

Um, no, we don’t know that. If it’s not even clear that exercise reduces cardiovascular events in diabetics then we definitely don’t know that . . .

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.