The Impact of Upper Airway Stimulation on Swallowing Function

Introduction

Obstructive sleep apnea (OSA) is increasingly common and may potentiate adverse health and psychological effects when left untreated. Although continuous positive airway pressure (CPAP) remains first-line therapy, a large proportion of patients on CPAP are noncompliant with treatment, with many unable to tolerate CPAP at all. ¹ Traditional surgical alternatives to CPAP consist of anatomy modifying procedures that alter the maxillofacial skeleton, palate, or tongue base. These approaches have focused on altering upper airway structure rather than function, to alleviate obstruction leading to apnea and hypopnea. One of the known complications of any procedure altering the native anatomy of the pharynx is swallowing dysfunction, as has been shown after uvulopalatopharyngoplasty (UPPP) and tongue base reduction. ^{2 –4} These procedures have been evaluated to determine the mechanism of postoperative dysphagia. Palate procedures such as UPPP and expansion sphincteroplasty have been demonstrated to increase hyoid movement time, increase duration of food stasis in the hypopharynx, and decrease pharyngeal constriction time. ⁵ Patients who underwent transoral robotic tongue base reduction surgery for OSA have exhibited findings of vallecular residue and early spillage, presumably from postoperative changes leading to decreased base of tongue sensitivity from neural damage as well as postoperative scarring. Postoperative dysphagia can present both in the immediate postoperative period and in late stage follow-up. ⁶

Upper airway stimulation (UAS) is a novel surgical alternative for patients with moderate to severe OSA who are unable to tolerate CPAP therapy. The UAS device (Inspire Medical Systems, Minneapolis, Minnesota) consists of 3 parts: a stimulation lead, an internal pulse generator, and the respiratory sensing lead. The stimulation lead is placed anteriorly along the hypoglossal nerve to selectively stimulate the protrusor muscles of the tongue. Stimulation is activated at night, is linked to respiration, and allows for relief of upper airway obstruction in response to each breath. Since UAS takes advantage of the native structure of the hypoglossal nerve to generate muscle tone and anteriorly displace the tongue, the patient’s native pharyngeal and tongue base anatomy remains unaltered. However, the impact of UAS on swallowing has not been directly assessed.

With this study, we utilized the Eating Assessment Tool (EAT-10), a self-administered 10-item measure of symptoms of dysphagia, which has previously demonstrated excellent criterion-based validity. ⁷ We hypothesize that UAS, despite chronically stimulating the tongue musculature throughout the course of the night, does not potentiate swallowing dysfunction as there is no change in the normal maxillofacial or pharyngeal anatomy.

Ethical Considerations

Prior to initiation of this study, approval was obtained from the institutional review board of Thomas Jefferson University. We abided by all policies outlined by this oversite committee.

Methods

After approval by the institutional review board of Thomas Jefferson University, we designed a prospective cohort study to assess the impact of UAS therapy on swallowing function in the postoperative setting. In addition to our standard protocol, we administered an EAT-10 survey preoperatively and postoperatively. The postoperative survey data were collected at least 3 months after UAS activation. This was done in order to assess both surgical and device-related impact on swallowing function.

Our institutional protocol for UAS implantation is to offer therapy to all patients with moderate to severe OSA, who are unable to tolerate positive pressure therapy, whose central apnea index encompasses less than 25% of the overall Apnea Hypopnea Index (AHI), and who have favorable examination findings on drug-induced sleep endoscopy (absence of concentric collapse at the velum). We do not consider BMI >32 to be an absolute contraindication, but rather base our surgical decision-making on body habitus and anatomy. All procedures were performed by the senior author. After UAS implantation, patients are seen routinely at 1 week postoperatively for a wound check and suture removal. They are then seen at 1 month postoperatively for device activation. After a period of 4 to 8 weeks of self-titration, the patient undergoes a titration polysomnogram (PSG) in which the device is optimally titrated.

With this study, we included all patients undergoing UAS implantation between April 1, 2016, and May 1, 2017, who underwent postoperative device activation, and who had completed a pre- and postoperative EAT-10 survey, at least 3 months after activation. Preoperative survey results were collected at the preoperative office visit. Postoperative results were collected at postoperative office visits or via patient phone call. We excluded all patients who did not complete a preoperative survey, had evidence of preoperative dysphagia (EAT-10 ≥ 3), and who did not have postoperative survey or PSG results. We also collected demographic, preoperative PSG, and postoperative PSG data.

Statistical analysis was performed using SPSS software version 24. We used a paired samples t test to compare preoperative PSG and EAT-10 data to the postoperative values.

Results

During the study period, we performed 33 UAS implantations on patients who completed the preoperative EAT10 survey. Twenty-seven of the 33 underwent device activation, postoperative titration PSG, and met inclusion/exclusion criteria. This cohort consisted of 16 men and 11 women who were included in the study. The mean ± standard deviation age and BMI were 63.63 ± 12.08 years and 29.37 ± 4.44 kg/m², respectively. Thirteen patients in total met other inclusion/exclusion criteria but did not complete the preoperative EAT-10 survey and were excluded from analysis.

The mean ± standard deviation values of the pre- and postoperative PSG and EAT-10 data are shown in Table 1. We found a significant difference between the pre- and postoperative AHI, O₂ desaturation nadir, and Epworth Sleepiness Scale (ESS) (P < .001, P < .001, and P = .026). The mean ± standard deviation pre- and postoperative EAT-10 values were 0.37 ± 0.79 and 0.22 ± 0.69, respectively (P = .461; Table 1). The postoperative survey was administered at an average of 164.29 ± 117.44 days after device activation.

OPEN IN VIEWER

Table 1.

Pre- and Postoperative PSG, ESS, and EAT-10 Data.^a

	Preoperative	Postoperative	P Value
AHI	34.90 ± 15.59	7.59 ± 10.27	<.001
O2 desaturation nadir	78.76 ± 9.04	89.94 ± 1.75	<.001
ESS	10.33 ± 5.69	5.25 ± 4.03	.026
EAT-10	0.37 ± 0.79	0.22 ± 0.70	.461

Abbreviations: EAT-10, Eating Assessment Tool; ESS, Epworth Sleepiness Score.

^aValues are mean ± standard deviation.

Postoperative titration PSG data showed 82.6%, 87%, and 56.5% patients reaching surgical success, postoperative AHI less than 15 and less than 5, respectively. Surgical success was defined as a decline in postoperative AHI by 50% compared to baseline value and to a level less than 20 (Figure 1).

OPEN IN VIEWER

Figure 1.

Rate of surgical success, postoperative Apnea Hypopnea Index (AHI) less than 15 and postoperative AHI less than 5 with upper airway stimulation therapy.

We analyzed the demographic and pre- and postoperative sleep study data of the group of patients who did not fill out the pre- and/or postoperative EAT-10 survey and compared the data to those who were included in our primary analysis. The excluded cohort consisted of 10 men and 3 women. The mean ± standard deviation age, BMI, preoperative ESS, AHI, O₂ desaturation nadir, and postoperative ESS, AHI, and O₂ desaturation nadir were 59.15 ± 15.02, 30.75 ± 3.78, 11.88 ± 4.02, 42.42 ± 23.84, 83.44 ± 8.23, 4.50 ± 0.71, 6.81 ± 8.34, and 89.38 ± 3.93, respectively. We saw a significantly higher preoperative AHI and significantly lower postoperative O₂ desaturation nadir in the group of patients excluded from primary analysis (P = .008 and P = .027).

Discussion

Untreated or undertreated OSA poses significant risk to clinical health and well-being. Demonstrated and potential adverse health consequences include hypertension, coronary artery disease, stroke risk, and predisposition to insulin resistance and diabetes mellitus. Moreover, untreated OSA lowers perceived quality of life, increases daytime sleepiness, and significantly increases the risk of motor vehicle accident.7 First-line therapy for OSA remains CPAP. However, using a definition of adherence of at least 4 hours of nightly use, 46% to 83% of patients are reported to be nonadherent. ¹ Moreover, CPAP adherence has been shown to decrease over a year of use. ⁸

Traditional surgical approaches for the treatment of OSA have aimed to alter either the boney facial skeleton or pharyngeal soft tissue in an effort to increase the size of the upper airway and eliminate obstruction. Among traditional surgical alternatives, Maxillomandibular advancement (MMA) has previously demonstrated the highest rates of AHI reduction (87%), although common drawbacks to the procedure include change in physical appearance of the face and jawline, risk of dental malocclusion, and facial sensory disturbances. ² Other surgical alternatives have demonstrated considerably lower percentage reductions in AHI (UPPP 33%, Laser Assisted Uvulopalatoplasty 18%, Radiofrequency ablation 34%). ² Both UPPP and modified UPPP, in particular, have shown significant rates of foreign body sensation (40.4%), velopharyngeal insufficiency (9.1%), and swallowing dysfunction (17.6%) that persist beyond 1 year after treatment. ³ Since the development of transoral robotic surgery base of tongue reduction, transient dysphagia is also a commonly encountered adverse event, occurring in 7.2% of cases. ⁴

Upper airway stimulation offers a definitive approach to treatment of OSA without altering native airway anatomy. The initial UAS clinical trial, Stimulation Therapy for Apnea Reduction (STAR trial), showed improvement in both objective PSG variables and subjective quality of life measures using Epworth Sleepiness Scale (ESS) and Functional Outcomes of Sleep Questionnaire scores. These data have been followed at 18, 24, 36, and 48 months and shown endurance of effect. ^{8 –10} Clinical trial data have been corroborated by single and multi-institutional case series showing similar improvements in PSG and quality of life measures. ^8,11,12 In addition to improved measures of apnea burden, rates of procedural or device-related adverse events have been low with most resolving with little to no need for intervention. ^9,13

With repeated stimulation of the medial branch of the hypoglossal nerve and the protrusor muscles of the tongue, it is feasible that muscle fatigue or tongue hypertrophy could be produced with UAS leading to dysphagia. Additionally, iatrogenic hypoglossal nerve weakness could lead to postoperative swallowing difficulty. However, UAS has not shown increased risk for postoperative dysphagia. This is likely due to the low incidence of iatrogenic hypoglossal nerve weakness and the lack of alteration of the boney or soft tissue anatomy of the facial skeleton and pharynx. ¹⁴

The EAT-10 is a self-administered 10-item measure of symptoms of dysphagia. This instrument, scored on a 0- to 30-point scale, has previously been used to assess presence and severity of swallowing dysfunction and monitor treatment response, with superb test–retest reproducibility and criterion-based validity. ¹⁵ An EAT-10 score of 3 or higher is considered abnormal. In our 13-month prospective cohort, 27 patients, without evidence of preoperative dysphagia, who underwent UAS, completed both preoperative and postoperative EAT-10. We found no difference between pre- and postoperative EAT-10 scores.

We recognize the limitations inherent in the study. Specifically, it suffers from a limited sample size and the possibility of selection bias. Since we were unable to collect postoperative EAT-10 data on all patients undergoing UAS, we may have been selecting a subset of the cohort and excluding those with differing results. In addition, no objective measure of dysphagia, such as modified barium swallow or functional endoscopic evaluation of swallowing, was performed. As dysphagia was not the primary preoperative complaint of this cohort, these studies would not have been part of our normal workup. Our results could be confirmed and verified if a prospective trial was designed comparing the outcomes of UAS to more traditional sleep surgery, such as MMA or UPPP, using both subjective and objective assessments.

Conclusion

Upper airway stimulation is a promising new surgical therapy to address OSA in patients who cannot tolerate CPAP. Unlike other surgical options, UAS does not alter native airway anatomy. As such, it does not seem to lead to postoperative dysphagia.