FDA Panel Deems Stimulator for Sleep Apnea Safe, Effective

Pauline Anderson

February 21, 2014

The Anesthesiology and Respiratory Therapy Medical Devices Panel of the US Food and Drug Administration's (FDA's) Medical Devices Advisory Committee determined almost unanimously that the Inspire II Upper Airway Stimulation System (Inspire Medical Systems Inc) is safe and effective for treatment of sleep apnea and that the benefits of the system outweigh the risks.

However, during a day-long meeting, panel members expressed concern that device-related adverse events, such as tongue weakness, persist in some patients who have received the implant, that the device was tested almost exclusively in white patients, that it is incompatible with having MRI, and that physicians will need training in implanting the device.

The company is seeking premarket FDA approval for the device, which stimulates the hypoglossal nerve during sleep in sync with a patient's inspiration. The stimulation contracts the patient's upper airway muscles, pulls the base of the tongue forward, and helps keep the airway open.

The system consists of implanted components, including a pulse generator, and stimulation and sensing leads. It also has an external programmer to store therapy settings that physicians can configure according to patient comfort and treatment targets, as well as a programmer with buttons for patients to turn the system on at night and off in the morning.

Innovative Technique

Various panel members commented that the device "offers another tool," addresses "an unmet need," and is an "important and innovative technique" to treat obstructive sleep apnea (OSA). And members praised the company's data as "impressive," "compelling," and "robust" and noted its "astonishing" 98% retention rate in a 12-month study.

They were referring to the company's pivotal Stimulation Therapy for Apnea Reduction (STAR) trial. The study included 126 adult patients, mean age 54.5 years, with moderate to severe OSA (defined by a score of 15 or more on the apnea-hypopnea index [AHI]) at 15 sites in the United States and 7 in Europe. Almost all patients were white (there were no African American patients, a population that has a relatively high rate of OSA), and 83% were male.

All patients had not responded to or were intolerant of continuous positive airway pressure (CPAP) treatment, had a body mass index no greater than 32 kg/m2, and were confirmed not to have complete concentric collapse at the level of the soft palate, determined through drug-induced sleep endoscopy.

At 12 months, the median AHI score decreased 68%, from 29.3 events per hour to 9.0 events per hour (P < .001). The score for the other co-primary endpoint — oxygen desaturation index (ODI) — decreased 70%, from 25.4 events per hour to 7.4 events per hour (P < .001).

Also at 12 months, the responder rate for the AHI was 66% (P = .0002). (A responder was defined as having at least a 50% reduction in the AHI score compared with the mean of the preimplant screening and 1-month visit — after implantation but before therapy activation — and an AHI score less than 20 events per hour.)

The responder rate for ODI was 75% (P < .0001) at 12 months. (A responder was defined as having a 25% or greater reduction in ODI.)

The positive results persisted to 18 months. At that point, the AHI responder rate was 65% (P = .0008). And the ODI rate had improved to 80% (P < .0001).

There were also patient-reported improvements in sleep-related quality-of-life measurements (the Epworth Sleepiness Scale questionnaire, on the Functional Outcomes of Sleep Questionnaire) and improvements in the percentage of sleep time with oxygen saturation below 90%.

The company noted that while 34% of patients were considered nonresponders according to the AHI at 12 months, many of these were partial responders in that they did experience some improvement on both the primary and secondary outcomes.

The STAR trial also had a randomized controlled therapy withdrawal group that consisted of 43 consecutive responders. At 13 months, half of these patients had their stimulator turned off and half had it kept on. The 18-month analysis showed that those with the device turned off had regressed to having significant sleep apnea — with AHI scores almost triple those of the maintenance group — before levels dropped back to just above those of the maintenance group after the device was turned back on.

This, according to company officials, confirms that the improvements seen in the study were due to upper airway stimulation and not other factors, for example, weight loss.

Addresses Concerns

This withdrawal arm addressed some of issues raised by the FDA in its report provided to panel members. One of the FDA's initial concerns was that because STAR had no control group, a patient's condition could have changed even if he or she hadn't been treated with the device. Regression to the mean could be another issue because patients entering the study had a relatively high AHI score and may have been more likely to experience a decrease in AHI than patients with lower a AHI score, regardless of whether they received the device.

Because most study patients had prior surgical procedures (eg, uvulopalatopharyngoplasty, tonsillectomy/adenoidectomy), the panel considered whether such procedures should be a criterion for device implantation, but in the end agreed this should not be mandated.

Also, because 4 of the 5 cases of palatal collapse occurred at the European site that was the first to start the implants, panel members reiterated the importance of physician training in use of this new technology. Such training might involve a significant learning curve.

The implanted device did not affect sleep architecture — for example, the amount of rapid eye movement sleep — so sleep quality was unaffected, according to the results of the study, which have been published in the New England Journal of Medicine.

There were, however, some adverse events — a total of 681 events during the 18-month postimplantation period, with an additional 43 reported beyond this 18-month period.

The rate of procedure-related serious adverse events, however, was less than 2%. In 2 patients, the device was removed, in 1 case (in a patient who had developed an abscess) possibly related to the device. Three patients died; no deaths appeared related to the implant.

Although some of the device-related adverse events — including tongue weakness, dry mouth, pain, and numbness — did not resolve in some patients, the panel was generally not overly concerned about the adverse event profile because patients remained extremely compliant with the device.

But they agreed that patients should be forewarned. "In making the decision whether or not to have surgery, patients need to be educated about the possibility of these side effects and managing their expectations will help to improve compliance," said panel chair Steven D. Nathan, MD, Inova Fairfax Hospital, Falls Church, Virginia.

Dr. Nathan added that as part of "full disclosure about what the procedure entails," patients contemplating the implant should also know that the device is a second-line — not first-line — therapy, that it doesn't work in everyone, and that they may need to be closely monitored. Some patients in the study — 14 of the 126 — needed extra titrations (adjustments to the stimulation level) over and above the 2 titrations included in the protocol.

Robert Loeb, MD, University of Arizona, Tucson, emphasized that patient selection for the implant is "paramount," especially because some patients in the pivotal study "ended up worse with treatment," although this wasn't necessarily due to the device, he said.

Some panel members, including Nirjaleswar Nikhar, MD, The Neurology Clinic of Washington, Olney, Maryland, expressed concern that the device is incompatible with MRI. He pointed out that patients in the study were in their mid-50s, so the risk of further abnormalities — for example, myocardial infarction — will increase over time. "This might not be an issue now, but it could be 10 years down the road," he said.

More Inclusive

Although Dr. Loeb said that he "can't think of any mechanistic reason this device would work differently in different racial groups" and that the study results could be extrapolated to populations other than whites, he and the other panel members at the end of the day agreed to encourage the sponsor to be more inclusive in postmarket studies.

Going forward, the company is proposing to conduct 2 postapproval studies: an extended follow-up to 5 years of the premarket pivotal study cohort and a study of 127 new patients to evaluate "real world" effectiveness, device safety, and the learning curve among newly trained physicians. Panel members agreed that overall quality of life (in addition to sleep-related quality of life) would be important information to collect in future studies.

All but 1 of the 13 voting members agreed that the device was safe, with 1 abstention from Rosalie Dominik, DrPh, a biostatistician from the University of North Carolina at Chapel Hill, who was also the only abstainer on the benefit vs risk issue.

"Given that I'm not a clinician, it's hard to make a judgment about whether the device is safe and because I can't judge the safety that well, I feel I can't judge the risk benefit, so I limited my vote to efficacy," she said.

On that vote, again, all but 1 of the panel members determined that the company had demonstrated the device's efficacy. Here, the abstention was from Christopher John Lettieri, MD, Walter Reed National Military Medical Center, Bethesda, Maryland. He said he abstained because "there was residual sleep apnea and functional impairment," although he considered the treatment was "clearly better" than untreated sleep apnea.

The most common symptom of OSA is excessive daytime sleepiness, which could lead to accidents. Recent studies have linked OSA with the development of obesity, hypertension, diabetes mellitus, and congestive heart failure.

Traditional therapies for OSA have varying success rates. CPAP is an effective first-line therapy, but patient nonadherence due to discomfort, noise, or other factors limits its beneficial effect. Patients can also try oral appliances.

As for surgical therapies, success rates range from 40% and 60%. Most of these approaches address obstruction at a single anatomic location of the throat, while patients typically have multilevel airway obstructions. Other drawbacks are that while surgeons commonly perform procedures on the soft palate in the mouth, they have limited options to remove obstructions at the base of the tongue, and available surgical options may compromise tongue function or alter facial anatomy.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: