Hypoglossal Nerve Stimulation: Targeted Neurostimulation Device for Treatment of Obstructive Sleep Apnea

1

Introduction

Obstructive sleep apnea (OSA) is a multi-factorial disorder where state-dependent loss of upper airway muscle tone and neuromuscular dysfunction result in cyclical upper airway obstruction during sleep. Hypoglossal neurostimulation is a new, functional approach to treating OSA that specifically addresses neuromuscular dysfunction, as opposed to conventional therapies such as positive airway pressure, oral appliance therapy, and reconstructive upper airway surgery, that act via direct, mechanical upper airway stabilization. Hypoglossal neurostimulation is an important new avenue for treatment, as conventional therapies are not successful in many patients due to insufficient effectiveness or efficacy, often related to low adherence.

The rationale for targeted hypoglossal neurostimulation (THN) is based on the lingual hydrostat model. The tongue is a muscular hydrostat, being a complex array of intertwined muscles that are relatively incompressible. To generate a myriad of shapes and various movements, tongue muscles attach to each other at various angles and support each other for various functions of the tongue in the hydrostat model. Complex co-activation is required for various physiologic activities and is supported as being needed for airway patency, as shown in animal studies where airway patency was greatest via truncal hypoglossal stimulation and lower with lesions that denervated the intrinsic tongue muscles. In human studies, when hypoglossal stimulation was performed via the nerve trunk or via surface stimulation of protrusors and retrusors, airflow was better or similar to just protrusor stimulation. The concept of hypoglossal nerve activation with co-activation of tongue muscles via nerve trunk stimulation was adopted by ImThera Medical, Inc., whose system ( Fig. 49.1 ) is the focus of this chapter (see Chapter 46 for Inspire Medical's approach). ImThera Medical's system at the time of publication is CE mark approved for use outside the United States but is still experimental in the United States and presently in Food and Drug Administration (FDA) pivotal trial testing.

The hypoglossal neuromuscular system has particular properties that are favorable for targeted neurostimulation. Tongue muscles—in particular, the posterior and tongue base muscles—have properties that make them more fatigue resistant and more favorable to exogenous stimulation, mitigating the risk of fatigue and dysfunction. In addition, the hypoglossal nerve trunk's structure in the submandibular region facilitates targeted stimulation, as it is non-fasciculated , and thus surface stimulation at particular zones results in current spread to nearby neuronal bundles, which results in different tongue activities. Thus a system was designed where a six-contact stimulation cuff is placed on the trunk, and stimulation is near continuous, but cycled via multiple contacts to reduce the risk of muscle fatigue. The implantable pulse generator (IPG) has six independent constant-current channels that correspond to the six nerve-cuff contacts ( Fig. 49.2 ). Stimulation does not rely on timing via a sensing/triggering lead for respiratory synchronization, as used in another system, but rather via rotation of stimulation among the titrated contacts along the nerve periphery. The result of stimulation is reshaping and stiffening of the tongue, with mild protrusion, rather than mostly protrusion. With tongue activation, the pharyngeal walls become less compliant and dilate laterally (personal observations).

2

Clinical Trials

The THN pilot study (THN1) included patients with a body mass index (BMI) <40; Apnea/Hypopnea Index (AHI) ≥20; without hypertrophic tonsils, craniofacial anomalies, macroglossia, or predominant central sleep apnea, implanted at a single center. Patients were all positive airway pressure (PAP) intolerant, and some had previously undergone upper airway surgery. Thirteen patients were followed for a year post-implantation. The mean BMI was 30.8 ± 3.4 and the mean age was 50.0 ± 10.2. The baseline AHI was 45.2 ± 17.8 and Oxygen Desaturation Index (ODI) was 29.2 ± 19.6. After 1 year, the mean AHI was 21.0 ± 16.5 ( P < .001) and the ODI was 15.3 ± 16.2 ( P <.001). Response was defined as AHI reduction of ≥50%, and for the 10 responders the mean AHI declined from 41.5 ± 13.1 to 13.2 ± 5.5; the ODI declined from 23.1 ± 10.2 to 7.8 ± 5.3 ( P <.001); the Epworth Sleepiness Score for all declined from 10.8 ± 6.2 to 7.8 ± 4.2 ( P = .09). Complications were technical: a device defective connector noted during surgery, three lead dislodgements, and one pulse generator needing replacement.

After 1 year seven responders underwent polysomnography with stimulation on one night and off the other, in a randomized fashion. Their mean baseline AHI was 41 ± 13, and ODI was 23 ± 11. After 1 year, with THN, the AHI was 15 ± 5 and the ODI was 9 ± 3. During the night of stimulation withdrawal, the AHI and ODI were not significantly different, at 16 ± 9 and 11 ± 10, respectively. Thus THN resulted in persistent benefit even with short-term therapy withdrawal.

An extended safety and feasibility study (THN2 FDA IDE) was then performed in multiple sites in the United States and Europe. Forty-six patients who met inclusion criteria were implanted. The criteria were similar to THN1 except that the BMI was limited to <37, central sleep apnea was limited to <10%, and positional OSA or the presence of other implanted medical devices were exclusions. The baseline AHI was 35 ± 23, and ODI was 32 ± 22 and at 6 months were 25 ± 23 and ODI 24 ± 22 (both P < .01). The Epworth Sleepiness Scale was 12 ± 5 at baseline and 8 ± 4 at 6 months ( P <.001). About 35% of the patients were responders, having ≥50% AHI decrease, and for this group the AHI declined from 36 ± 19 to 9 ± 6 and the ODI from 33 ± 19 to 8 ± 6 (both P <.0001). There were only three serious adverse events that were device or procedure related, consisting of device malfunction needing replacement, a hematoma, and device repositioning surgery.

An analysis was then performed to refine patient selection based on outcome from THN1, THN2, and patients implanted outside of the trial under CE mark approval since 2012. The predictors of success were determined to be BMI <35, AHI <65, Apnea Index ≤30, and <15 events per hour of oxygen desaturations of >10%. Theoretically this may exclude patients with extreme collapsibility or those with greater neuromuscular dysfunction due to hypoxemia.

The current THN3 pivotal FDA trial incorporated the previous predictors of success criteria for patient selection. The trial is a multi-center, prospective, parallel, two-armed, randomized controlled study of about 141 patients with moderate to severe OSA who have failed or refuse continuous positive airway pressure (CPAP). After implantation, patients are randomized 2 : 1 to stimulation start at 1 month or 4 months post-operatively (control). The control group thus crosses over to treatment after a 3-month delay, and all are assessed by polysomnography and quality-of-life questionnaires after 12 to 15 months from the date of implantation surgery.

3