Subscribe to eTOC

Coverage Denied
CMS Unlikely to Pay for Brain Stimulation in Depression

Will vagus nerve stimulation (VNS) for depression be a reimbursable therapy for people for whom all other treatments have failed? Possibly not, according to a proposed decision by the Centers for Medicare and Medicaid Services (CMS).

On Feb. 5, the CMS recommended in a memo that Medicare should not provide coverage for VNS as a therapy for treatment-resistant depression. Pointing to the fact that the treatment failed to meet its primary outcome measure in a randomized trial, the memorandum concluded, “CMS does not believe there is a treatment effect directly attributable to VNS therapy based on the current evidence.”

In VNS therapy, a small, battery-operated pulse generator — implanted in the patient's upper chest — sends mild, short, periodic pulses to the left vagus nerve via connecting leads. The exact mechanism for therapeutic action is unknown, but VNS has been shown to affect blood flow to specific areas of the brain, as well as levels of neurotransmitters such as serotonin and neurepinephrine.

The CMS based its decision on a 2005 randomized trial by A. John Rush, MD, of the department of psychiatry at the University of Texas Southwestern Medical Center, and co-investigators (Biol Psychiatry 2005;58:347–354). The 10-week, randomized controlled trial compared adjunctive VNS with sham treatment in 235 outpatients with nonpsychotic major depressive disorder. A two-week, single-blind recovery period (in which there was no stimulation) and then 10 weeks of masked active or sham VNS followed implantation.

On the Hamilton Rating Scale for Depression — in which physicians assess low mood, insomnia, agitation, anxiety, and weight loss — the response rates were 15.2 percent for the active group and 10 percent for the sham group (p=.251).VNS was well-tolerated, the study author concluded, but there was no definitive evidence of short-term efficacy for adjunctive VNS in treatment-resistant depression.

The final CMS decision regarding VNS for treatment-resistant depression will be issued 60 days following the conclusion of a 30-day comment period, but it seems unlikely that CMS will reverse itself.

“It's not often that we switch from deciding not to cover to extending coverage, but it has happened, such as with bariatric surgery a few years ago,” said Steve Phurrough, MD, director of the CMS Coverage and Analysis Group. “In this case, with a primary trial that did not reach its endpoint, we were just not comfortable saying that we ought to be paying for it.”


The FDA approved VNS for treatment-resistant depression in July of 2005, but the decision was controversial. One year before, it had issued a non-approvable letter to the device's manufacturer, Cyberonics, and a Senate Finance Committee investigation into the approval process concluded that the director of the FDA Center for Devices and Radiological Health, Daniel Schultz, MD, had overruled more than twenty FDA officials who had reviewed the data on VNS and recommended against approval.

FDA approval does not always equal Medicare coverage, said Dr. Phurrough. “It's not too uncommon for the FDA and CMS to disagree. The FDA has a different decision to make: theirs is whether a particular technology has enough evidence to demonstrate that it should be available on the market. Ours is not whether it should be sold, but whether we should pay for it. That's a different evidentiary question.”

The CMS decision leaves many patients with treatment-resistant depression — and the physicians who treat them — with a conundrum: VNS is approved and available, but at a price that's beyond the reach of many if Medicare or private insurance will not pay. Surgical and hospital costs for implantation may vary from $12,000 to $25,000.

“As a researcher and as a clinician, I would like my patients to have access to VNS. I think it can be useful for some patients with treatment-resistant depression; several of my patients have responded to VNS,” said Sarah Lisanby, MD, chief of the Brain Stimulation and Therapeutic Modulation Division at the New York State Psychiatric Institute. “But it has been exceedingly difficult if not impossible for most of my patients who might benefit from VNS to access it because of the cost, and that is dismaying. You can imagine, from the patient's perspective, how it must feel to hear the news that the FDA has approved a treatment specifically for treatment-resistant depression, but then to hear, ‘Oh, by the way, you can never have it.’”


Dr. Lisanby and other experts in the field concede the CMS point that the studies evaluating VNS for treatment-resistant depression were not optimally designed. “The data are not as strong as they might be, and there are a couple of reasons for that,” said Andrew Leuchter, MD, professor and vice chair of the department of psychiatry and biobehavioral sciences at the David Geffen School of Medicine at the University of California-Los Angeles, where he directs the Laboratory of Brain, Behavior, and Pharmacology.

“First, in retrospect we can see that the studies weren't really designed to capture the information needed. It turns out that the patients who benefit most from VNS don't really start to show improvement until after about four months, and some not until six or eight months — but the double-blinded aspect of the studies didn't go out that far. Knowing what we now know, should the double-blinded study have gone on longer? Yes, but we didn't know that then.”

It's also particularly difficult to study therapeutic effects in treatment-resistant depression, given that the population is, by definition, treatment-refractory. “The response rates are not going to be overwhelming. These people have not responded to multiple medications, many medications used in combination, and frequently electroconvulsive therapy, as well,” said Dr. Leuchter.

He suggested that one strong piece of evidence pointing to VNS' efficacy is the lateness of the peak response. “Usually, if you're going to see antidepressant effectiveness, you get it within the first three months, and then over time it loses effectiveness,” he explained. That was the case with the major Sequenced Treatment Alternatives to Relieve Depression (STAR-D) trials (Am J Psychiatry 2006;163:1905–1917), on which Dr. Leuchter was a principal investigator.

“But it seems to be the opposite with VNS; while some patients had early and significant improvements in depressive symptoms, it was a near miss in showing significant short-term benefits from the treatment. But we tend to find that more and more patients are responding in these trials over time. More had benefited at nine months than at three months. As opposed to medication, which is tapering at that point, VNS appears to be building.”

The trials as designed failed to persuade CMS of that point. In fact, CMS even expressed a surprising level of skepticism about treatment-resistant depression in general, writing, “The concept of treatment-resistant depression is vaguely defined, subject to varying determination, and until a scientifically valid definition exists, is of little help in treatment selection for individual patients.”

This disturbs many experts. “At one level, the CMS memo is appropriate given the level of data submitted. I've been underwhelmed by the scientific approach taken to the study design, and the data are full of signal that's hard to interpret,” said Paul Holtzheimer, MD, an assistant professor in the department of psychiatry and behavioral sciences at Emory University School of Medicine, who has studied focal brain stimulation for treatment-resistant depression.

“But the memo is troubling in how it treats the concept of treatment-resistant depression and the need for treatment. They imply that the placebo rate in this population is high, which it's not. And while it's true that in the early stages of treatment-resistant depression, it's hard to know what counts — is it failing one, two, or three medications? These patients have often failed every pure antidepressant there is, and multiple other modalities, as well. Everybody agrees that these patients are unlikely to respond to treatments approved by the FDA and unlikely to have a placebo-response rate.”


In 1999, CMS agreed to extend coverage for VNS to treat pharmaco-resistant epilepsy — based, Dr. Holtzheimer said, on data that were not much stronger than that for treatment-resistant depression. “There are more data overall, but there are similar weaknesses,” he said. “In the studies looking at long-term response, mean seizure rate decreased 26 percent in one year and overall by 30 percent at five years. I'm sure that depression severity decreased by at least that much among the treatment-resistant patients. I worry, based on how CMS has handled this, that they're establishing a two-tiered system for dealing with devices in depression.”

That could have implications beyond just the VNS decision, since other brain stimulation approaches are now being investigated in the treatment of depression, including transcranial magnetic stimulation and deep brain stimulation. “This chain of events suggests that it's important, in the early stages of planning clinical trials that are designed for FDA approval, to ensure they are also designed to yield the quality of evidence that's required to pass CMS muster,” said Dr. Lisanby. “It's useless if something is FDA-approved, but patients can't pay for it. The shortest path to regulatory approval may not be the best path to getting safe and effective therapies to our patients.”

“These are not easy trials to do,” said Dr. Leuchter. “I think that all of the brain stimulation kinds of treatments are going to run into similar barriers. First, the people who are the best candidates for those treatments are those who have failed multiple treatments in the past. You're not going to look at brain stimulation as a first-line treatment in the near future. Because of their very low response rates, it will be hard to detect the signal. But as in the case of VNS, I think most people believe there is a signal there, that it is an efficacious treatment, and that it ought to be available.”

In the meantime, what will happen next with VNS? The only thing that might convince CMS to change its coverage decision, said Dr. Holtzheimer, is a successful sham-controlled study going out to 12 months, which is unlikely to happen. But Cyberonics is currently engaged in a dose-finding study that was required as a condition of FDA approval. Dr. Lisanby said a dose-response rate might provide further evidence of efficacy.

“I specialize in treatment-resistant depression; that's all I do,” Dr. Lisanby said. “Most of my patients have had electroconvulsive therapy and it didn't work, or they relapsed. That's a distressingly common story. Some of these patients have responded to VNS, and nothing else is available that's appropriate to help them.”


Although the FDA has approved vagus nerve stimulation for treatment-resistant depression, the Centers for Medicare and Medicaid Services have proposed that the therapy should not be reimbursed. Experts on depression weigh in on that proposed decision.


• Rush AJ, Marangell LB, Cooke RG, et al. Vagus nerve stimulation for treatment-resistant depression: a randomized, controlled acute phase trial. Biol Psychiatry 2005;58:347–354.
• Rush AJ, Trivedi MH, Fava M, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry 2006;163:1905–1917.