Controls Out of Control
The report by EnteroMedics CEO Mark Knudson that an unanticipated therapeutic effect in the control arm of that company’s recent Empower study may have been the reason the trial failed to reach its efficacy endpoints [see article, p5] is just the latest example of how failure to fully consider the control arm of a clinical trial has led to serious financial harm. In our view, the technology behind EnteroMedics’ vagus nerve block therapy is still very promising, though the failed trial decimated the company’s market capitalization and put its future in jeopardy.
As we all recall, Northstar Neuroscience experienced a similar financial ruin when its Everest trial for stroke recovery failed to meet its primary endpoints. In that trial, the results would have been profoundly better had the subjects in the treatment arm who failed to receive a threshold dose of cortical stimulation been counted in the control arm. In both cases, very promising neurotechnology therapies—not to mention millions of dollars in investor and shareholder financing—were sacrificed to the inadequate consideration of who’s actually receiving the therapy versus who’s not.
This loss of control has affected other neurotechnology companies struggling to design a randomized, double-blind clinical trial. And for manufacturers of implanted neurotechnology devices, this process is far more complicated than for drug trials that use a sugar pill for control. Make no mistake, the management of both Northstar and EnteroMedics bear the responsibility for proceeding with their clinical trials without fully understanding the therapeutic regimen for their devices. But it is worth exploring the concept of a randomized control trial for an implanted device in the overall context of evidence-based medicine.
Particularly when it comes to disorders such as pain or depression, where subjective reports from the patient are all we have to go on, the notion of placebo effect becomes somewhat cloudy. To some researchers, the placebo effect is itself a form of neuromodulation—and we understand its mechanism of action about as well as we understand the mechanism of certain drugs or device interventions such as ECT. If a new therapy can reliably and consistently elicit the placebo effect—whatever that is—why should that be discarded? All the patient cares about is feeling better. If that feeling emanates in part from some subconscious recognition that an internal device may be pumping out juice, should the individual be deprived of an effective therapy?
Whether we’re talking about neurological disorders with objectively measurable deficits like stroke or epilepsy, or disorders like pain and depression with more fuzzy indices, the bottom line is alleviating suffering and restoring function. The goal of true evidence-based medicine should be to identify therapies that work for a given population, not to discard new effective treatments because the control arm could not be controlled.
Editor and Publisher