A Novel Predictive Scoring System for Hypoglossal Nerve Stimulation Outcomes in Obstructive Sleep Apnea Patients

Introduction: Hypoglossal nerve stimulation (HGNS) is an effective treatment for select obstructive sleep apnea (OSA) patients, but identifying optimal candidates remains a challenge. We developed and internally validated a novel predictive scoring model based on key anatomical, physiological, and demographic factors to assess HGNS treatment outcomes.

Methods: A retrospective analysis of 461 HGNS patients was conducted. Predictive factors (age, sex, BMI, apnea-hypopnea index (AHI), DISE-PAP opening pressure (PhOP), oropharyngeal lateral wall collapse) were selected based on existing literature and validated for association with treatment outcomes using multivariate regression in Python. A scoring model was developed by assigning weighted points to each factor, and predictive accuracy was evaluated for three outcomes: post-op AHI, AHI change, and Sher15 success (≥50% AHI reduction to < 15 events/hr).

Results: A composite score included:
• Age: ≤34 (5), 35-44 (4), 45-54 (3), 55-64 (2), 65-74 (1), ≥75 (0)
• Sex: Female (5), Male (0)
• BMI: ≤25 (5), 25.1-27.5 (4), 27.6-30 (3), 30.1-32.5 (2) ,32.6-35 (1), >35 (0)
• Baseline AHI: ≥65 (5), 55-64.99 (4), 45-54.99 (3), 35-44.99 (2), 25-34.99 (1), 15-24.99 (0)
• PhOP: ≤8 (5), >8 (0)
• Lateral Wall Collapse: 0 (5), 1-2 (0)

The score was significantly predictive for all outcome measures with higher scores correlating with lower post-op AHI, greater AHI reduction, and higher Sher15 success rates. Score correlations: AHI change (β = -1.31, p < 0.001), post-op AHI (β = -0.0288, p < 0.001), Sher15 (β = 0.0471, p < 0.05). Each point corresponded to ~1.31 expected AHI reduction, ~2.84% decrease in expected post-op AHI, and ~4.82% increased odds of Sher15 success.

Conclusions: Our novel scoring system demonstrated significant predictive accuracy for three clinically meaningful outcomes: post-op AHI, AHI change, and Sher15 success. The model could serve as a valuable tool for preoperative planning, facilitating more personalized care while improving outcomes, and optimizing resource allocation in OSA management. It can also provide a foundation for other studies to externally validate or refine.