Long-term changes of stimulation intensities in hypoglossal nerve stimulation

Abstract

Study Objectives:

Hypoglossal nerve stimulation (HNS) is a novel therapy in the treatment of obstructive sleep apnea. Previous studies have focused on the effectiveness of HNS, but there are no studies specifically investigating the long-term changes of the stimulation intensities in HNS. Increasing stimulation intensity requirements have been reported in the past in other peripheral nerve stimulation therapies. The aim of this study was to investigate the development of stimulation intensities over the observation period of 4 years.

Methods:

All patients who were implanted with an HNS system since December 2013 and maintained a bipolar configuration over the observation period were included. Sensation threshold (ST), functional threshold, the titrated stimulation intensity (SI), and the apnea-hypopnea index (AHI) were recorded.

Results:

A total of 82 patients were enrolled (sex: 69 men, 13 women, age: 60 ± 11 years, body mass index: 29. 8 ± 4.0 kg/m²). Two months after surgery, the median ST was 0.8 ± 0.5 V. During the observation period of 48 months, no significant change of ST was observed. The median ST was 1.0 ± 0.4 V (P = 0.93) at 48 months. Similar results were found for functional threshold and the titrated stimulation intensity. There was a significant reduction of the baseline median AHI when compared with the median AHI at 1, 12, 24, 36 and 48 months after surgery (P < 0.05).

Conclusions:

The stimulation intensities in HNS show no significant changes over 4 years. Despite the constant stimulation intensity, AHI was significantly reduced. This indicates that the stimulation threshold of the hypoglossal nerve does not change over time with this therapy.

Citation:

Zhu Z, Hofauer B, Wirth M, Heiser C. Long-term changes of stimulation intensities in hypoglossal nerve stimulation. J Clin Sleep Med. 2020;16(10):1775–1780.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Previous studies have focused on the effectiveness of hypoglossal nerve stimulation, but so far there is no study that investigates the development of stimulation parameters. The aim of this study was to investigate the long-term changes of stimulation intensities of hypoglossal nerve stimulation.

Study Impact: So far, this is the first study that analyzes the development of stimulation intensities of hypoglossal nerve stimulation over the observation period of 4 years. Our study also shows that there are no sex differences in stimulation parameters.

INTRODUCTION

Obstructive sleep apnea (OSA) is a subset of sleep-disordered breathing that is characterized by recurrent episodes of upper airway narrowing and closure resulting in apneas and hypopneas.¹ OSA has been shown to lead to excessive daytime sleepiness, increased likelihood of errors and accidents, and be associated with behavioral, cognitive, and cardiovascular morbidities.^2,3 Therefore, successful treatment is important. Positive airway pressure (PAP) is the standard treatment for patients with moderate to severe OSA, but suboptimal adherence is a limiting factor of PAP therapy effectiveness.⁴

Selective hypoglossal nerve stimulation (sHNS) is an emerging treatment option for OSA. The uniqueness of this therapy lies in the fact that sHNS does not alter the anatomy of the upper airway. In the past, prospective multicenter trials such as the STAR (Stimulation Therapy for Apnea Reduction) trial have already shown the effectiveness of sHNS.⁵ More recently, the long-term safety and effectiveness was evaluated by the ADHERE (Adherence and Outcome of Upper Airway Stimulation for OSA International Registry) registry.⁶ The sHNS system (Inspire Medical Systems, Maple Grove, MN) consists of a stimulation lead with a cuff electrode that is placed around branches of the hypoglossal nerve, an implantable pulse generator, and a respiratory sensing lead, resulting in a unilateral stimulation of the hypoglossal nerve.⁷ By including only the medial branches of the hypoglossal nerve, a selective stimulation is achieved for the genioglossus, geniohyoid, and the intrinsic tongue muscles. This results in unrestricted protrusion of a stiffened tongue, which dilates the airway.^7,8 Stimulation parameters can be adjusted and titrated individually. These parameters include the electrode configuration and the stimulation amplitude. The most common electrode configuration is the bipolar setting “+-+”.⁹ When the stimulation device is first activated, the sensation threshold (ST) and the functional threshold (FT) are determined. ST is the intensity at which the patient perceives the stimulation, while FT represents the stimulation intensity at which tongue protrusion over the lower teeth is first observed. The patient then begins a customization phase of therapy to allow for acclimatization to different amplitudes of stimulation. The history of modern peripheral nerve stimulation goes back to Wall and Sweet in 1967,¹⁰ who described pain relief following brief electrical stimulation. Since then, peripheral electrical stimulation has been used in various medical specialties, eg, phrenic nerve stimulation for diaphragmatic palsy, stimulation of the extremities in patients with spinal cord injury, and transcutaneous nerve stimulation for neuropathic pain.¹¹ In the past, some studies reported of postsurgical fibrosis, especially perilead fibrosis, which increases resistance, resulting in higher stimulation intensity requirements.¹² So far there are no studies that show long-term changes of the stimulation intensity in sHNS. The aim of this study was to investigate whether there is a significant change in stimulation parameters during HNS therapy over the observation period of 4 years.

METHODS

Patient selection

For this study, all patients with moderate to severe OSA (apnea-hypoxia index [AHI] between 15 and 65 events/h) and with a body mass index (BMI) lower than 35 kg/m² who received an implant for sHNS since December 2013 and maintained a bipolar configuration over the observation period were enrolled. Preoperative screening included in-lab polysomnography (PSG) according to the American Academy of Sleep Medicine guidelines from 2012, clinical examination and drug-induced sleep endoscopy (DISE) to rule out complete concentric collapse at the level of the soft palate. Patients were excluded if pronounced anatomical abnormalities preventing the effective use of sHNS were identified during clinical examination (eg, enlarged tonsils). Informed consent was obtained for each patient.

Upper airway stimulation system and electrode configuration

Qualified patients underwent surgical implantation of the sHNS system (Inspire II Upper Airway Stimulation System, Inspire Medical Systems, Maple Grove, MN) as previously described.¹³ Proper functioning of the complete system was ascertained intraoperatively prior to wound closure. One month after implantation (M1), the stimulation system was activated with the standard settings (bipolar electrode configuration “+-+,” pulse width 90 µs, frequency 33 Hz). After several weeks of acclimatization, a polysomnography with therapy optimization was done for appropriate voltage titration and sensing settings. For consistent comparison of the stimulation intensities, only patients who maintained a bipolar configuration over the observation period of 48 months were included. In general, the electrodes can be configured to provide either monopolar or bipolar stimulation. The three electrodes embedded in the stimulation cuff and the implantable pulse generator (IPG) can be configured as anode (+), cathode (−), or inactive (off). This leads to 2 possible bipolar configurations (“+-+”; “−+−”) and 3 monopolar configuration options (“—,” “off-off,” and “-off-”). The most commonly used configuration is the bipolar configuration “+-+”.⁹ In this configuration the stimulation current loop is only inside the stimulation cuff, resulting in a high current loop that is restricted to the perilead area, avoiding stimulation of adjacent nerves. In the monopolar configuration, the IPG functions as the anode, which causes a wider electrical field with a stimulation current loop between the cuff and the IPG.

Follow-up

Follow-up visits were scheduled at months (M) 1, 2, 6, 12, 24, 36, and 48. At M1 the ST, FT, and SI were collected via telemetry. Over the acclimatization period, patients were instructed to increase stimulation strength gradually from the initially programmed amplitude. Baseline ST, FT, and SI at M2 were compared to corresponding voltage thresholds at follow-up visits M12, M24, M36, and M48. Furthermore, the baseline AHI before sHNS implantation and the AHI at follow-up visits were recorded. Two months after surgery (M2) a therapy optimization of the stimulation during an in-lab PSG was performed. The SI was titrated according to the AHI and patient’s comfort. The aim was to reach the lowest AHI possible, preferably lower than 10 events/h. Patients could modify the titrated SI at home at a certain range, which was adjusted during their sleep lab visit (usually ± 0.4 V) according to their needs (sleep quality, self-reported effectiveness). For M12, M24, M36, and M48, a home sleep test (HST) was performed at the titrated SI that the patient was using as a home therapy level. The PSG and HST were performed according to the American Academy of Sleep Medicine guidelines from 2012. The same scoring criteria for apneas and hypopneas were used for all sleep studies: The alternative definition was used to score a hypopnea as a reduction of airflow by > 30% for at least 10 seconds with a corresponding oxygen desaturation of ≥ 4%. Apneas were scored based on a ≥ 90% reduction in airflow for at least 10 seconds.

Statistical analysis

Data sets were tested for normality, then the appropriate statistical tests were performed based on the distribution of the data. Results were summarized as means ± standard deviation. Statistical analysis is done by Wilcoxon test and Mann-Whitney U test for nonparametric statistics and Student’s t test for approximately normally distributed data. Paired tests were used for comparing paired observations within individuals. For comparing the ST, FT, and SI, only 56 data pairs were analyzed for M12, and 26, 10, 4 data pairs were analyzed for M24, M36, and M48 respectively. Because M1 data are not representative for the actual stimulation parameters, paired t tests were used to calculate statistical significance between M2 and M12, M24, M36, and M48. The alpha-level of significance was set at 0.05 (2-tailed). Analyses were performed using SPSS version 22.0 statistical software (SPSS Inc., Chicago, IL).

RESULTS

A total number of 82 patients were enrolled. Table 1 shows the demographic and baseline characteristics. For statistical analysis, data for 82, 56, 26, 10, and 4 patients were available at M2, M12, M24, M36, and M48, respectively. Two months (M2) after device implantation, which included 1 month of acclimatization the median ST was 0.8 ± 0.5 V. The median ST did not change significantly at M48 (1.0 ± 0.4 V, P = 0.93) (Figure 1). The median FT at M2 was 1.4 ± 0.5 V and showed no significant difference compared with M48 (1.9 ± 0.6, P = 0.18). Similar results were observed for the titrated SI. The SI was 1.8 ± 0.6 V at M2 and did not change significantly over the period of time. At M48, the SI was 2.1 ± 0.4 V (P = 0.68). To have a better comparison, paired sets of data were analyzed for each follow-up (Table 2). Table 2 shows M2 data for each follow-up group separately. Except for the FT at M12, there was no significant difference for ST, FT, and SI at any other follow-up compared with M2. The AHI reduction was also analyzed. The analyzed AHI values represent the entire night value. There was a significant reduction of the baseline median AHI compared with the median AHI at M2 (n = 82), M12 (n = 56), M24 (n = 26), M36 (n = 10), and M48 (n = 4) (P < 0.05) (Figure 2).

Table 1.

Patients’ baseline characteristics.

Parameter (Median ± SD)
Number	82
Age, y	60 ± 11
Sex (male/female)	69/13
Body mass-index, kg/m2	29.8 ± 4.0
Apnea-hypopnea index, events/h	31.4 ± 13.1

Median values and standard deviation (SD) including the range of age, sex, body mass index, and baseline apnea-hypopnea index.

Figure 1. Development of median values of sensation threshold, functional threshold and stimulation intensity (in volts) of all patients with a bipolar electrode configuration over the observation period of 48 months.

Table 2.

Median values of sensation threshold, functional threshold, and stimulation intensity (in volts) paired up at each follow-up.

Stimulation parameter in V	ST	FT	SI
M2 vs M12 (n = 56)	0.8 ± 0.47	1.4 ± 0.51	1.8 ± 0.55
	vs	vs	vs
	0.95 ± 0.42	1.6 ± 0.53*	1.9 ± 0.53
M2 vs M24 (n = 26)	0.7 ± 0.44	1.4 ± 0.48	1.7 ± 0.60
	vs	vs	vs
	0.8 ± 0.33	1.5 ± 0.51	1.9 ± 0.53
M2 vs M36 (n = 10)	0.8 ± 0.35	1.6 ± 0.39	2.0 ± 0.43
	vs	vs	vs
	0.95 ± 0.44	1.7 ± 0.51	2.0 ± 0.47
M2 vs M48 (n = 4)	1.1 ± 0.45	1.7 ± 0.37	2.0 ± 0.17
	vs	vs	vs
	1.0 ± 0.39	1.9 ± 0.56	2.1 ± 0.36

Statistical analysis was performed for each follow-up group with M2 by using paired t tests. *P < 0.05. FT = functional threshold, SI = stimulation intensity, ST = sensation threshold.

Figure 2. Boxplot with maximum and minimum values showing median AHI reduction over 48 months.

Median baseline apnea-hypopnea index (AHI) was compared with AHI values at 2 months after treatment (M2), M12, M24, M36, and M48 and significant difference is marked with *P < 0.05.

To further investigate influences on the stimulation parameters, the study group was divided by sex and age, with an older patient being defined by age > 65 years. Our results show a significant difference of the FT between younger and older patients for the follow-up visits M2, M12, and M24 (Table 3). Statistical significance for further follow-ups was not analyzed due to small sample sizes and missing data. Baseline AHI and baseline BMI were compared between male and female patients and older and younger patients, respectively. No significant difference was found for baseline AHI (P = 0.27; P = 0.29) or for baseline BMI of those groups (P = 0.15; P = 0.74).

Table 3.

Median values of sensation threshold, functional threshold, and stimulation intensity (in volts) of different patient groups.

Patient group/in V	ST at M2	ST at M12	ST at M24	FT at M2	FT at M12	FT at M24	SI at M2	SI at M12	SI at M24
Female patients (n = 13)	0.8 ± 0.51	0.8 ± 0.28	1.0 ± 0.35	1.6 ± 0.25	1.6 ± 0.53	1.8 ± 0.44	2.0 ± 0.49	1.9 ± 0.46	2.2 ± 0.50
Male patients (n = 69)	0.8 ± 0.47	0.9 ± 0.44	0.8 ± 0.51	1.3 ± 0.51	1.6 ± 0.52	1.4 ± 0.51	1.7 ± 0.53	1.8 ± 0.53	1.7 ± 0.49
Patients < 65 y (n = 56)	0.8 ± 0.44	0.9 ± 0.40	0.8 ± 0.29	1.2 ± 0.51	1.4 ± 0.51	1.2 ± 0.46	1.7 ± 0.49	1.8 ± 0.45	1.7 ± 0.49
Patients ≥ 65 y (n = 26)	0.9 ± 0.54	1.1 ± 0.46	1.0 ± 0.34	1.6 ± 0.53*	1.8 ± 0.49*	1.8 ± 0.48*	1.9 ± 0.66	2.0 ± 0.61	2.1 ± 0.47

No significant difference between female and male patients was observed for ST, FT and SI. Older patients (≥ 65 y) showed a significantly higher FT than younger patients (< 65 y). *P < 0.05. FT = functional threshold, SI = stimulation intensity, ST = sensation threshold.

DISCUSSION

Previous studies focused on the effectiveness, safety, and adherence of sHNS,^14–16 but so far there is no study that investigates the development of stimulation parameters. In other peripheral nerve stimulation therapies there have been reports of perineural fibrosis after implantation of the stimulation lead, which resulted in higher stimulation intensities over time.¹² Since sHNS is a novel stimulation therapy, it is essential to analyze the long-term changes of stimulation intensity.

Sometimes it is necessary to change the electrode configuration during advanced titration of sHNS therapy to optimize tongue motion and outcome for the patient. Although the majority of patients only require 1 titration night, some patients need a second titration to further optimize and individualize therapy. During this second in-lab PSG (advanced titration), testing of specific electrode configurations, stimulation timing, and impulse settings are performed. Different electrode configurations can lead to varying nerve stimulation and muscle activation, resulting in variable tongue movements.¹⁷ In monopolar stimulation, the stimulation intensity is lower than in the bipolar configuration, which makes it difficult to compare the development of stimulation intensities in patients with changing configurations over time. Since this study is a first approach to investigate the long-time development of stimulation intensities, only bipolar stimulation intensities were analyzed.

Our study shows that there is no significant increase of stimulation parameters over 48 months. The ST, FT, and SI stay almost constant over the observed time period. The SI was initially titrated during in-patient polysomnography, with the aim of reaching the lowest possible AHI. The SI is the stimulation intensity, which lead to the lowest AHI during the titration night. The SI might either have changed during time if the patient modified it (eg, due to increasing sleepiness) or by the physician due to results of the follow-up home sleep test or patient´s symptoms. The FT was always above the ST, whereas the installed SI was slightly above the FT. In Figure 1, it appears that the FT increases insignificantly between the follow-ups M24 and M48, but this may be due to the few data points that are available. Of 10 patients, 6 actually showed a decreasing tendency when comparing the FT at M24, M36, and M48, whereas 4 patients had a higher FT in the subsequent follow-ups. Similar results were observed for ST and SI. When comparing paired data of M2 and M12 for FT, a significant increase is shown (Table 2). For future studies, more data samples over a longer period are necessary to interpret the results for the increasing tendency of FT. Recent investigations published by Strollo et al¹⁸ showed comparable findings. The prospective multicenter study analyzes the durability of the sHNS treatment effect at 18 months and shows a significant AHI reduction and stable values for the FT over time. When comparing the ST of the follow-ups at M12 and M18, a significant reduction of the ST was found.¹⁸ This was also mentioned by another study, which described the 5-year outcome of sHNS therapy. Apparently, ST, FT, and subdiscomfort thresholds decreased over time, although data were not shown.¹⁹

To further differentiate and study possible effects of age and sex on stimulation parameters, the study group was subdivided. Since sex differences in electrical stimulation threshold has been a well-studied issue in other electrical stimulations, we investigated whether these differences could be also found for sHNS. Previous studies reported that women show a lower sensory threshold than men and have a significantly higher functional threshold in surface electrical stimulation of skeletal muscle.²⁰ However, we found no significant difference when comparing stimulation parameters of the male group with the female group. This could be due to the fact that sHNS is a direct nerve stimulation instead of a surface electrical stimulation. Nevertheless some of the patients do also sense the stimulation during sHNS. There are previously published results of patient-reported outcomes of sHNS, showing that 49% of patients sense the stimulation during the night. However, with more time and use of the therapy, fewer patients found the sensation disruptive.²¹ Although not fully understood, this phenomenon has been observed already with neuromuscular electrical stimulation. Improved tolerance and conditioning were reported after repeated sessions of stimulation.²² Interestingly, our data show a significantly higher FT for patients older than 64 years compared with the younger group, whereas the SI did not differ significantly. This leads to the conclusion that the motor threshold of the tongue increases with age. This was already observed by other groups previously.²³ However, it seems that once the motor threshold is reached, the muscle response of the tongue in older patients is comparable with the muscle response of the younger patients, explaining the insignificant difference of the SI. In addition, our previous study also underlines this fact by showing an equally effective AHI reduction in both older and younger patients.²⁴

In concordance with previous studies, the AHI was reduced significantly, and although the SI remained constant over the time, the AHI still continued to decrease.¹⁵ This implies that the stimulation threshold of the hypoglossal nerve remains the same. Although cuff electrodes have been reported to have higher risks for perineural fibrosis,¹² our data do not support this hypothesis. The consistent SI indicates that there is no increasing impedance that may be expected if perineural fibrosis had occurred.

A possible disadvantage of this study is the fact that the ST and FT are subjective parameters that depend on the patient’s and investigator’s perspective. Another limitation that needs to be noted is the generalizability of this data, as we only included patients that maintained a bipolar configuration “+-+” for their stimulation treatment. For further studies, some methods need to be developed to compare all possible configurations. It also needs to be pointed out that AHI was collected by different methods (PSG and HST). The baseline AHI and the titration of the therapy at month 2 were done by PSG, and HST was performed during the later follow-up. This conforms to the best clinical practice. Only patients who show an AHI > 10 events/h in the HST get an additional PSG. To reduce possible errors made by the automated scoring system, all HSTs were scored manually by experienced technologists, and the same definition for apnea and hypopnea were applied. However, due to the lack of an EEG-signal during the HST, no information on the sleep stages were available, which might influence the evaluation of sleep-related breathing disorders to a certain degree.

In conclusion, this study shows that there is no significant change of stimulation intensities in sHNS over 4 years, once again confirming the safety of sHNS therapy and arguing against the theory of the development of perineural fibrosis. Similar to previous results, the AHI was reduced significantly. There were no significant differences of stimulation parameters between female and male patients. Older patients show a significantly higher FT, suggesting possible age-related changes of the motor threshold of the tongue muscle.

DISCLOSURE STATEMENT

All authors have seen and approved this manuscript. Work for this study was performed at the Otorhinolaryngology, Head and Neck Surgery Department of the Klinikum rechts der Isar, Technical University Munich in Germany. Benedikt Hofauer and Markus Wirth received compensation for travel costs and congress fees from Inspire Medical Systems. Clemens Heiser is a study investigator and received honoraria, travel, and research support from Inspire Medical Systems. Zhaojun Zhu reports no conflicts of interest.

ABBREVIATIONS

AHI

apnea-hypopnea index

BMI

body mass index

DISE

drug-induced sleep endoscopy

FT

functional threshold

HNS

hypoglossal nerve stimulation

HST

home sleep test

IPG

implantable pulse generator

M1

1 month after surgery

M2

2 months after surgery

M12

12 months after surgery

M24

24 months after surgery

M36

36 months after surgery

M48

48 months after surgery

OSA

obstructive sleep apnea

PSG

polysomnography

sHNS

selective hypoglossal nerve stimulation

SI

titrated stimulation intensity

ST

sensation threshold