Association Between Loop Gain and Arousal Threshold with Hypoglossal Nerve Stimulation Therapy Use

Tuesday, October 20, 2026

8:18 AM - 8:24 AM PT

Location: 403B

Presenting Author(s)

AR

Andrea S. Rivera Diaz, MSc

Medical Student
Columbia University
New York, New York, United States

Co-Author(s)

AL

Allison Lin, BS

Medical Student
Columbia University Irving Medical Center
New York, New York, United States
RK

Rachel B. Kutler, MD, MS

Resident Physician
University of California, San Francisco
New York, NY, United States
Maria V. Suurna, MD

Professor
University of Miami
Miami, Florida, United States
AK

Ashutosh Kacker, MBBS, MD

Professor
Weill Cornell Medicine
New York, NY, United States

Senior Author(s)

Yi Cai, MD

Director of Sleep Surgery
Columbia University
New York, New York, United States

Disclosure(s):

Rachel B. Kutler, MD, MS:

faculty for this accredited education activity has no relationship(s) with ineligible companies to disclose.

Ashutosh Kacker, MBBS, MD:

faculty for this accredited education activity has no relationship(s) with ineligible companies to disclose.

Introduction: Hypoglossal nerve stimulation (HGNS) patients have variable therapy usage. Arousal threshold (ArTH) and loop gain (LG) are two endotypic features that have been underexplored as potential predictors of HGNS use. This study evaluated whether sleep study-derived estimates of ArTH and LG are associated with nightly HGNS use at 3, 6, and 12 months.

Methods: Patients at an academic hospital who underwent HGNS between 3/2022-6/2025 with available usage data and detailed pre-HGNS sleep study metrics were analyzed. ArTH and LG were assessed through validated clinical prediction algorithms. Nightly use was summarized and within prespecified windows centered at 3 months (days 60–120), 6 months (days 150–210), and 12 months (days 335–395) for participants with sufficient data (≥50% of nights) in each window. Monthly change in HGNS usage and associations of ArTH and LG with usage were assessed using linear mixed-effects growth-curve models with patient-specific random intercepts and random slopes for time.

Results: Of 32 patients included (mean age 63.5±9.6 years, BMI 29.1±3.8, 72% male), 18 (56.3%) had high ArTH and 8 (25%), high LG. HGNS usage declined during the first 12 months (-4.1 min/night per month, p = 0.005). In mixed-effects models, high ArTH was associated with lower nightly HGNS use at 12 months (difference -1.09 h/night, p=0.044), but not at 3 nor 6 months (p=0.16, 0.056). LG was not associated with usage at 3, 6, or 12 months (all p>0.66). Neither ArTH nor LG predicted rate of change in use from 3 to 6 to 12 months (interaction p≥0.44).

Conclusions: HGNS use modestly declined over the first 12 months. High ArTH but not LG was associated with lower use at 12 months. Larger studies are needed to validate these findings and assess other predictors of HGNS use.