Abstract Rationale Randomized and prospective trials support hypoglossal nerve stimulation (HGNS) for obstructive sleep apnea (OSA), but real-world data on early response predictors remain limited. This study identifies baseline characteristics, adjunctive therapies, and oximetry patterns associated with achieving apnea-hypopnea index by 4% criteria (AHI4%)15 events/hr on first post-implant polysomnography. Methods Retrospective cohort of adults who underwent HGNS implantation between 2019-2025 at a tertiary sleep center. Primary outcome was treatment response (AHI4%15 events/hr) on first post-implant polysomnography. Sensitivity analyses examined alternative response thresholds (AHI4%10/hr and 5/hr; AHI3%15/hr and10/hr). Wilcoxon rank-sum and chi-square tests were used for group comparisons. Results Among 56 patients who underwent HGNS implantation, 35 had complete first post-implant polysomnography data available (median age 61.0 years IQR 55.2-66.3; 31% female; median time to first post-implant study 243 days IQR 199-316). The cohort was predominantly non-Hispanic (92%) and White (83%). Response rates varied by threshold: 80.0% (28/35) achieved AHI4%15/hr, 65.7% (23/35) achieved AHI4%10/hr, and 28.6% (10/35) achieved AHI4%5/hr. Using AHI3% criteria, response rates were lower: 30.6% (11/36) achieved15/hr and 11.1% (4/36) achieved10/hr. At the primary threshold (AHI4%15/hr), responders had lower body mass index (28.6 vs 31.1 kg/m², p = 0.035), and baseline AHI4% (20.3 vs 37.5 events/hr, p = 0.021). All non-responders (100%) demonstrated oximetry “banding” patterns, a characteristic of increased loop gain, compared to 77% of responders (p = 0.16). Non-responders also had higher rates of adjunct medical therapy use (71% vs 57%, p = 0.49), including hypnotics (57% vs 43%, p = 0.50) and acetazolamide (43% vs 25%, p = 0.35). HGNS adherence trended higher in responders but was not significant: median nightly hours (5.95 vs 5.20, p = 0.31), monthly nights used (29.0 vs 26.0, p = 0.64), and monthly nights≥4 hours (25.0 vs 19.0, p = 0.23). Similar patterns were observed across alternative response thresholds. Conclusions In this real-world HGNS cohort, lower baseline body mass index was associated with successful early response. Response rates declined substantially with more stringent criteria, particularly for AHI3%, suggesting HGNS is more effective in addressing obstructive events associated with more significant oxygen desaturation (4% criteria) than those with milder desaturation (3% criteria). Non-significant trends toward greater adjunctive pharmacotherapy use and oximetry banding patterns were associated with inadequate treatment response. These findings underscore the importance of characterizing OSA pathophysiological traits beyond anatomic obstruction to guide patient selection and suggest opportunities for combination therapeutic approaches to optimize HGNS outcomes. Larger multicenter studies are needed to validate these predictors and inform clinical decision-making. This abstract is funded by: None
Tsou et al. (Fri,) studied this question.