Ending the Sham
At first glance, Mainstay Medical’s announcement that its Reactiv8 B study for back pain failed to meet its primary endpoint [see article, p5] seems all too familiar. A promising neurotech startup with a novel therapy achieves good results in preliminary trials, raises significant funds from investors, goes public, then crashes and burns after a failed pivotal trial. As was the case with Northstar Neuroscience a decade earlier, Mainstay’s share price dropped precipitously on news of the trial results, falling to about €2.50 from about €15. The company plans to continue marketing its CE Marked therapy in Europe while it pursues FDA approval, and to be sure, there have been other neurotech devices approved despite failing to meet primary endpoint in a pivotal trial.
But in our view, there’s more at issue here than the financial health of a young neuromodulation company. The responder rate in the trial’s treatment group was 56 percent, compared to 47 percent in the control group, a result that is not statistically significant. But the performance of the control group in this trial makes us continue to question the wisdom of requiring implantable neuromodulation device trial sponsors to perform “sham” stimulation on a control group. Aside from the ethical dubiousness of performing needless invasive procedures on a group of patients, there is no clinical value in eliminating a so called “placebo” effect from an otherwise viable therapy, particularly if there is a differential “placebo” effect between therapeutic alternatives. (Can you think of a sugar pill that has achieved 47 percent response in a clinical trial?)
Make no mistake, neurotech firms that have the funds to undertake a clinical trial should have a better understanding at the outset of who responds and who doesn’t respond to a new therapy. But trial sponsors should rightly prioritize making their target population get better, as opposed to preventing a different population from getting better. One way to achieve this might be to perform a non-inferiority trial of an implanted device compared to a pharmaceutical therapy and to do away with the sham stimulation control group altogether. This type of trial design would preserve, rather than discard, any differential “placebo” effect between drug and device therapies. Particularly when it comes to subjective disorders like pain and mood disorders, the “placebo” effect could well be a significant adjunct in the therapeutic regimen, and as Howard Fields from UCSF demonstrated at NANS earlier this year [NBR Jan18 p7], there is a neural basis for the placebo effect with identified pain-modulating circuits.
If the FDA doesn’t like this, maybe another approach is a hybrid neurostimulation/placebo therapy that incorporates the pathways that Fields has identified. In this case, the only logical “control” group would be patients who scored high on the pessimism assessment test.
Editor and Publisher