History and evolution of the HGNS concept
Animal studies have confirmed that the genioglossus muscle is a key protrusion muscle, as opposed to the styloglossus and hyoglossus muscles which retract the tongue [
13]. In 1989, Miki et al. found a relationship between the hypoglossal nerve and upper airway resistance during stimulation in six canines [
14]. In 1992, Schwarz et al. reported a correlation of V1max stimulation and decrease in critical closing pressure (Pcrit) in 18 decerebrate felines. Yoo et al. suggested that multi-contact nerve electrodes can be effective in achieving upper airway dilation and patency by selective activation of various branches of the hypoglossal nerve in eight beagles [
15]. Oliven et al. reported on the effect of airway modulation by selectively stimulating protrusor and retrusor muscles in anesthetized canines [
16]. The study confirmed that selective genioglossus muscle stimulation significantly stabilized the airway while stimulation of the styloglossus and hyoglossus muscles collapsed the airway.
The first attempt to improve upper airway patency in humans via neurostimulation was performed by Guilleminault et al. [
17]. Transcutaneous submental and intraoral electrical stimulation of the upper airway muscles was attempted with limited success. Preliminary successful human studies were first reported in 1996 by Schwartz et al. [
18]. By intramuscular stimulation of the lingual muscle in nine participants, the frequency of airway collapse decreased without causing sleep arousal. A pilot study in 2001 proposed that unilateral electrical stimulation of the hypoglossal nerve was a feasible and potential therapeutic option for OSA [
19]. Efficacy of OSA treatment was shown in seven of the eight hypoglossal nerve stimulation (HGNS) implanted patients (Inspire Medical systems, Maple Grove, MN). For the 6-month continuation of the study, the results were consistent; they were not successful in the long term due to technical defects that led to device dysfunction [
20].
HGNS devices and device concepts
Following the published technical limitations of the human pilot study in 2001, several investigators and medical device companies focused on improvements over a decade. In 2011, three clinical trials were sponsored by different and independent firms: Apnex (St. Paul, MN, USA), Inspire (Maple Grove, MN, USA), and ImThera Medical (San Diego, CA, USA).
Apnex and Inspire are based on unilateral inspiratory stimulation to the medial branch of the hypoglossal nerve. The two devices differed in inspiration sensors; Apnex was a double impedance sensor positioned over the lower ribs of both hemithoraces while Inspire was an effort sensor placed between the intercostal muscles.
Eastwood et al. [
21] were the first to report results from the Apnex trial. This clinical trial was performed in Australia enrolling 37 patients and ultimately implanting 21. Inclusion criteria consisted of patients with AHI between 20 and 100 events per hour (at least 80% hypopnea), BMI < 40 (kg/m2), and age between 21 and 70 years. The 6-month follow-up yielded successful results with more than 50% reduction in the AHI. But the larger scale phase 2–3 clinical trial did not deliver anticipated results with mean AHI decrease from 45 to 25 [
22]. Surgically unsuccessful results led to cessation of device development.
Van de Heyning et al. [
23] subsequently published the results from the 6-month clinical trial with the Inspire II device. It was a two-part study with the first consisting of a wide inclusion and exclusion criteria. The 20 patients enrolled in the study had BMI < 35 (kg/m2) and AHI > 25 events per hour. Fourteen patients showed the predicted outcome, and six patients showed greater than 50% reduced AHI that was also less than 20 events per hour after 6 months. The second part narrowed the criteria based on the findings from part one and included drug-induced sleep endoscopy (DISE) as a diagnostic modality. The inclusion criteria were BMI less than 32 (kg/m2), AHI between 20 and 50 events per hour, and absence of complete concentric collapse (CCC) of the velum as seen during DISE. Seven of the eight implanted patients assessed at 6 months post operatively showed successful results.
Succeeding the Inspire II trial, a larger multicenter 1-year phase 2–3 trial (STAR trial) was completed in 2014 by Strollo et al. [
12]. One hundred twenty-six patients were enrolled with inclusion criteria consisting of CPAP non-adherence, BMI less than 32 (kg/m2), AHI between 20 to 50 events per hour, absence of significant positional or central apneas, and absence of CCC on DISE at the velum. Sixty-six percent of the patients had achieved surgical success. The device was then FDA approved.
Mwenge et al. [
24] performed a 1-year clinical study treating 14 OSA patients with the ImThera Medical device. Stimulation was delivered at both inspiration and expiration and thereby excluding the need of an inspiratory sensor. The hypoglossal nerve electrode was cuffed on the main trunk and targeted all muscles in the hemilateral tongue. Inclusion criteria consisted of BMI between 25 and 40 (kg/m2), AHI below 20 events per hour, Mallampati score 1–3, and palatine tonsils grade 0–2. Ten of the 13 implanted patients achieved surgical success with mean AHI decrease from 45.2 to 21 events per hour. Currently, a larger phase 2–3 trial is underway.
Surgical technique
The HGNS system (Inspire, Maple Grove, MN, USA) is comprised of three parts: stimulation cuff electrode, pleural pressure sensing lead, and implantable pulse generator (IPG).
Advantages of HGNS
Even though CPAP is a treatment for OSA that has well-documented efficacy and low morbidity, the adherence rate is low [
7]. In patients with CPAP intolerance, HGNS can be an effective alternative. Daily use of the device at 12 months was 86% and daily use at 18 months was 84% [
26]. The advantage of HGNS being a titratable therapy like CPAP or MAD is especially beneficial as OSA is a chronic condition that needs monitoring and re-evaluation due to aging, weight gain, and decrease in elasticity of the soft tissues.
Another advantage of HGNS is that as a one-time procedure it improves airway collapsibility at multiple levels. DISE and fluoroscopy studies have shown retrolingual and retropalatal space enlargement with HGNS [
27,
28]. HGNS directly opens the hypopharyngeal airway with tongue protrusion, but it also impacts the retropalatal airway. Its mechanism restores muscle tone to avoid upper airway collapse during sleep.
Complications of HGNS
Strollo et al. announced the overall rate of serious adverse effects to be less than 2% [
12] with no serious IPG device infection requiring explantation and no permanent hypoglossal nerve damage. There is also less postoperative discomfort as compared to other soft tissue and skeletal operations [
25]. Discomfort in the IPG position, tongue stiffness, tongue abrasion, transient ipsilateral tongue paresis, and post-operative swelling were the reported side effects. Most were minor and can be mitigated by meticulous surgical techniques, adjustment of stimulation parameters, dental adjustments, and the use of mouth guards.
Limitations of HGNS
Woodson et al. reported the 36-month postoperative outcome data of the STAR trial patients [
29]. Continued follow-up supporting the use of HGNS is warranted. Another limitation is the narrow selection criteria. Patients over BMI 32 kg/m2 are excluded from HGNS as success rates become inconsistent. Nevertheless, the prevalence of obesity is high and continues to increase [
30]. The high cost of the HGNS implant is another limitation to wide application. Another drawback of HGNS is the current size of the IPG and MRI incompatibility.