Senior Associate Editor Samuel K. Moore today reports from Washington, D.C., on a meeting of medical advisors to the Food and Drug Administration to consider the approval of a promising new therapy for patients with clinical depression who are unresponsive to conventional psychiatric treatments.
Samuel K. Moore
A panel of advisors to the U.S. Food and Drug Administration (FDA) today recommended against the government approving a new treatment for depression that uses powerful magnets to induce spikes of voltage inside the brain. The FDA makes the final decision, but the regulators have rarely gone against their advisors. The advisors do not actually vote for or against approval but were asked to give their consensus view on whether the device had met the FDA's main criteria: whether it was safe, effective, and that the ratio of risk to benefit for the new treatment was equivalent to treatment by a device already in use.
What Neuronetics Inc., the device's maker, needed was a slam dunk. What they delivered rolled around the rim a few times and then fell away. Under FDA rules, they had to prove that the risk-to-benefit ratio for their technology, repetitive transcranial magnetic stimulation (rTMS), was equivalent to that of a technology already in use, in this case electroconvulsive therapy. Electroshock carries a number of risks, not the least of which is memory loss, but it's also highly effective. In fact, it's the go-to treatment to snap patients out of serious depression when other treatments fail.
If Neuronetics ultimately succeeds with the FDA using this set of data, and it still can, though it seems unlikely, rTMS would join vagus nerve stimulation therapy (see our March 2006 cover feature "Psychiatry's Shocking New Tools") as one of only two medical devices approved by the regulators for a mental illness since it started ruling over such devices in the 1970s. (Electroconvulsive therapy was grandfathered into approval.)
The majority of the panel—made up of an engineer, several psychiatrists and neurologists, and a statistician—had no problem with rTMS's risks. There are almost none. The biggest worry with it is that it might accidentally spark a seizure, but that did not happen even once out of the 155 patients treated. The problem was that Neuronetics couldn't prove any benefit. Treated patients got a little better, but so did those patients that underwent a sham treatment.
The big problem boiled down to a decision the company made in its trial design years ago. There are three main tests used by clinicians to determine how depressed a patient is and whether they are getting better—MADRS, HAMD17, and HAMD24 (MADRS is the Montgomery Asberg Depression Rating Scale and HAMD is the Hamilton Depression Score). They're all questionnaires that the physician answers about a patient, such as what the patient's mood is or how often they fixate on suicide. The doctor answers each question with a number. All three tools are widely used in clinical trials of antidepressant drugs, as well as to determine whether a patient qualified—that they were depressed enough but not too depressed—for the trial. The doctors administering the trial had to fill out all three and didn't know which was most important to the study—the primary outcome.
Neuronetics used the HAMD24 to determine who should go in the study but the MADRS as the primary outcome. More specifically, they were looking for a significant improvement in treated patients versus those who got the sham. Wouldn't you know, the MADRS result was just shy of being statistically significant and the HAMD results were significant.
Neuronetics tried parsing the data this way and that, but ultimately to no avail. In what seemed a response to an impromptu plea by Columbia University psychiatrist Sarah H. Lisanby, several panel members admitted that they thought there was "a signal" in the data—that the device was having some effect. But even those panelists said they were not comfortable with the size of the signal. Their advice to Neuronetics: Try again.