Olson, Michael

Abstract Introduction Hypoglossal nerve stimulation (HGNS) is an effective therapy for moderate-to-severe obstructive sleep apnea (OSA) in patients intolerant of positive airway pressure. Comorbid insomnia (COMISA) is common and may shape their engagement. Despite reports of improved insomnia symptoms, real-world data suggest reduced adherence and increased support needs. This study evaluated the impact of COMISA on HGNS therapy patterns and outcomes. Methods We conducted a retrospective cohort study of 46 adults with OSA who underwent HGNS implantation at Mayo Clinic Rochester between July 2020 and August 2025. Patients were categorized as OSA alone or COMISA based on pre-implantation data, including documented insomnia diagnosis, insomnia severity index score ≥15, sleep initiation or maintenance complaints, hypnotic or melatonin use, participation in cognitive behavioral therapy for insomnia, or an insomnia care plan within one year prior to activation. We collected HGNS adherence metrics (30-day average nightly use, nightly pauses, percent nights used) at 1, 3, 6, and 12 months; non-visit HGNS-related clinical encounters at 6 and 12 months; and pre- and post-activation AHI from polysomnography or home sleep apnea testing. Longitudinal outcomes were analyzed using mixed-effects models with fixed effects for time and COMISA status and a random intercept for patient. Results The cohort was 54% male and 98% White, with median age 66 years and BMI 31 kg/m2 (24–34). Median pre-implant AHI was 27 (19–40), and 30 patients (65%) had COMISA. Median nightly HGNS use declined from 7.8 to 6.6 hours from 1 to 12 months; mixed-effects modeling confirmed significant decreases over time (p≤0.03). Median nightly therapy pauses increased from 0.1 to 0.5, and median non-visit clinical encounters increased cumulatively from 2.5 at 6 months to 3.0 by 12 months (both p≤0.01). Median AHI reduction was 18.5 (2–27) (p=0.047). There were no statistically significant differences between OSA-alone and COMISA groups in HGNS adherence trajectories, therapy pauses, support needs, or AHI improvement (all p>0.05). Conclusion Although HGNS use declined modestly over time, COMISA was not statistically associated with differences in adherence, clinical support needs, or treatment response. Directional trends suggest greater adaptation challenges among patients with COMISA, warranting evaluation in larger prospective studies. Support (if any)

Abstract Introduction Hypoglossal nerve stimulators (HGNS) are used in patients with moderate-to-severe obstructive sleep apnea (OSA) who are unable to tolerate or do not respond to continuous positive airway pressure (CPAP) therapy. However, a subgroup of patients fail to achieve resolution of OSA with HGNS alone. Previous studies have shown that lower BMI, lower nadir oxygenation, decreased sleep latency, and higher VOTE score on drug-induced sleep endoscopy (DISE) predict better treatment outcome. We hypothesize that baseline diagnostic polysomnography (PSG) may further predict which patients will achieve successful therapeutic outcome with HGNS therapy. Methods We conducted a retrospective analysis of all adult patients who underwent HGNS implantation at Mayo Clinic Arizona between 2019-2025 and had a pre-implantation diagnostic PSG available for review. Treatment success was defined by Sher Criteria: >50% reduction in post-treatment apnea hypopnea index (AHI) to a final AHI of < 20/hr. Responders met Sher criteria and non-responders did not meet Sher criteria. Demographics, sleep questionnaire results, PSG variables, and VOTE score on DISE were collected and analyzed using Spearman’s rank correlation of reduction in AHI. Results Twenty-three patients (11M/12F, median age: 69[IQR:64,75] years) met inclusion criteria. Median baseline AHI was 24[IQR:19,37]/hr. Median post-implant AHI was 8 [IQR:3, 14]/hr. Responders (n=15, 65%) and non-responders (n=8, 35%) demonstrated median post-implant AHI of 4.5[IQR:2.3,7.3]/hr and 15.7[IQR:12.3,28.5]/hr, respectively. Treatment response was associated with less time with SpO2 < 88% (T88%) (21min [IQR:3,64] vs. 73[IQR:36,105mins]; p=0.023) and higher central apnea index (CAI) (1[IQR:0.1,3] vs. 0[IQR:0,1], p=0.03) on pre-implantation PSG. Conclusion In this small cohort of OSA patients, less nocturnal hypoxemia as defined by T88% and higher baseline central apnea index were predictors of better treatment response to HGNS therapy as measured by Sher criteria. We hypothesize that greater nocturnal hypoxemia likely reflects more severe airway obstruction during sleep. Our analysis also showed that responders had higher baseline CAI than non-responders, however we believe this association is not clinically significant and likely driven by one outlier (CAI 13/hr). Further analysis in a larger HGNS cohort is needed to confirm these findings. Data enrichment efforts and more sophisticated analyses are underway. Support (if any)