Abstract
Objective
Obesity is a well‐established risk factor for obstructive sleep apnea (OSA), and weight loss is often recommended alongside hypoglossal nerve stimulation (HGNS) in continuous positive airway pressure‐intolerant patients. Our study assesses how long‐term changes in body mass index (BMI) following HGNS impact surgical success.
Study Design
A retrospective case review.
Methods
We performed a retrospective case review of OSA patients who underwent HGNS at our institution from 2017 to 2022 and multicenter prospective patient data from the ADHERE registry. Patients were included if they had pre‐ and postoperative sleep studies and BMI data 1 year following surgery, and postsurgical outcomes were evaluated. Patients were stratified into 3 cohorts based on their BMI change from baseline to final follow‐up following surgery: (1) remaining within 2 BMI points, (2) gaining at least 2 BMI points, or (3) losing at least 2 BMI points.
Results
In both institutional cohort (N = 222) and ADHERE cohort (N = 1949), most of the patients remained within 2 BMI points after HGNS (84.1% and 80.8%, respectively). The patients who lost at least 2 BMI points had the greatest reduction in AHI (−24.77 ± 13.84, −22.58 ± 19.76 events/h) compared to patients who maintained stable BMI (−15.10 ± 20.33, −18.85 ± 18.11 events/h) or gained at least 2+ BMI points (−6.39 ± 22.52, −15.82 ± 19.63 events/h) following HGNS (P = .002 and .006, respectively).
Conclusions
While HGNS effectively reduces AHI, this effect was most significant in patients who experienced weight loss compared to those who gained or remained in stable body weight. Our study underscores the importance of postoperative weight management and its adjunctive role in improving HGNS success in OSA patients.
Keywords: hypoglossal nerve stimulator, obesity, obstructive sleep apnea, weight loss, weight management
Obstructive sleep apnea (OSA) is characterized by episodes of complete or partial airway collapse with an associated oxygen desaturation or arousal from sleep. 1 Although the first‐line therapy for OSA is continuous positive airway pressure (CPAP), adherence rates are inconsistent, ranging from 46% to 80%. 2 , 3 Recently, hypoglossal nerve stimulation (HGNS) has emerged as an alternative for patients who cannot tolerate CPAP therapy. 4 , 5 The efficacy of HGNS has been demonstrated in various studies, including the multicenter Stimulation Therapy for Apnea Reduction (STAR) trial, which reported a success rate of 66% per Sher20 criteria, defined by a reduction of the apnea–hypopnea index (AHI) score by at least 50% from baseline to an AHI of ≤20 events/h. 5 , 6 These positive outcomes were sustained long term, with a therapy response rate of 64% at 18 months, 74% at 36 months, and 75% at 60 months. 7 , 8 , 9
The findings from the STAR trial and other phase II trials have suggested a higher likelihood of success with HGNS in patients with an AHI between 15 and 65 events/h, a body mass index (BMI) below 35 kg/m2, and no complete concentric collapse at the velopharynx during preoperative drug‐induced sleep endoscopy (DISE). 10 , 11 A post hoc analysis of the multicenter ADHERE registry also revealed that advancing age was associated with improved success, with a 4% increase in the odds of treatment success for every 1‐year increase in age. 10 Finally, increased preoperative BMI was associated with therapy failure, with each unit increase in BMI resulting in a 9% decrease in the odds of treatment success. 10
Although the findings from the trials and database demonstrate the effect of preoperative BMI on HGNS success, the potential influence of longitudinal weight changes on therapy outcomes remains uncertain. Of note, a recent single‐institution study by Renslo et al found BMI at the time of the second postoperative sleep study to be predictive of HGNS success in patients with higher baseline BMI, suggesting the importance of postoperative weight management in patients with higher adiposity. 12 Given the complex multifactorial interplay of weight and obesity on upper airway anatomy, physiology, and OSA, it is crucial to consider the role of weight management as a key component of treatment strategies for patients undergoing HGNS. We aim to build upon this existing literature by evaluating how weight changes from the time of surgery correlate with HGNS efficacy using a large cohort from our institution and a secondary corroborative analysis of the ADHERE registry. We hypothesize that long‐term changes in BMI will impact the success of HGNS therapy in patients with CPAP‐intolerant OSA.
Methods
Patient Selection
We performed a retrospective case review of electronic medical records of consecutive patients who underwent HGNS at our institution (Thomas Jefferson University Hospital) between 2017 and 2022. Patients were included if they had data available to calculate preoperative and postoperative AHI, baseline BMI, and BMI at 1 year postoperatively. Among 391 patients undergoing HGNS implant during the study period, 169 (43.2%) were excluded due to missing data needed to calculate changes in AHI or BMI. This study was ‐approved by the Institutional Review Board (IRB# 16D.238).
Additional cohort of patients using the same inclusion criteria was retrieved from the ADHERE database (updated as of August 2024), which collects observational data regarding outcomes following HGNS across multiple institutions. Patients in this registry have met clinical indications for HGNS: AHI 15 to 65 events/h, CPAP intolerance, and a lack of complete concentric collapse of the velum per DISE. Of 4950 patients in the ADHERE registry, 3001 were excluded due to missing data or incomplete follow‐up. AHI was determined using in‐lab titration HGNS titration study (tPSG), fullnight in‐lab polysomnograms (PSG), or type 3 home sleep apnea tests (HSAT) with HGNS. The ADHERE registry has been approved by ethics committees or institutional review boards at all participating centers.
Patient Cohorts
Patients in both analyses were stratified into 3 cohorts based on their BMI change from baseline to final follow‐up following surgery: (1) remaining within 2 BMI points, (2) gaining at least 2 BMI points, or (3) losing at least 2 BMI points. A BMI change of 2 kg/m2 was empirically selected as the cutoff because this represents a significant weight change of approximately 8 to 20 lbs or 5% to 10% of body weight for individuals with BMI ranging from 25 to 35 kg/m2, aligning with common weight loss guidelines. 13 , 14
Outcome Measures
Demographics (age, gender, and ethnicity) and OSA measures (AHI, SpO₂% nadir, Epworth Sleepiness Scale [ESS]) were assessed at baseline and final follow‐up. Postoperative AHI was collected from the sleep study closest to the 1‐year time point to align with the timing of the postoperative BMI measurement. For institutional patients, tPSG was used in cases where patients did not undergo a postop full‐night HSAT or PSG. In such cases, the overall AHI during the titration PSG was used, as opposed to the AHI at therapeutic amplitude (lowest AHI) to reduce any variance between AHI measurements on tPSG, which tend to underestimate AHI. Among the ADHERE cohort, tPSG AHI was determined from the portion of the sleep study that correlated with the patient's voltage and device setting for at‐home use before the study. Treatment response was measured according to Sher criteria, defined either as a ≥50% reduction in AHI in addition to final AHI ≤20 (Sher20) or ≤15 (Sher15). 6 HGNS adherence (hours/night) was recorded at the final visit for patients in the ADHERE registry. A subsequent subgroup analysis was performed to compare patients in the ADHERE registry who gained 2+ BMI points, stratified by those with a BMI ≤35 kg/m2 versus those with a BMI ≥35 kg/m2.
Statistical Analysis
Numeric variables are presented as mean ± standard deviation, and categorical variables are presented as proportion (frequency). One‐way ANOVA and Fisher's exact test were used to analyze numeric and categorical variables, respectively. Univariable linear regression models were conducted to evaluate the impact of BMI on reduction in AHI following treatment. A univariate logistic regression model was utilized to evaluate the association between BMI change and treatment responder per Sher criteria. All statistical analyses were performed using an alpha level of .05 and utilized R version 4.4.0 (R Foundation for Statistical Computing).
Results
Institutional Cohort Baseline Characteristics
A total of 222 patients from our institution were included in the analysis. Demographic and baseline characteristics are summarized in Table 1. The majority of the patients remained within 2 BMI points (N = 163, 84.1%). Thirty‐two patients (7.1%) gained 2+ BMI points, and 27 patients (8.9%) lost 2+ BMI points from baseline to the final postop visit at 12 months. Most patients were white (N = 163, 94%) and male (N = 152, 68%) with an average age of 61.36 ± 11.04. No significant differences were observed in gender, race, age, baseline AHI, or ESS among groups. Overall, AHI at baseline was 34.06 ± 17.32 events/h and ESS was 10.13 ± 5.30. Patients who lost 2+ BMI points had the highest baseline BMI (31.4 ± 4.0 kg/m2) compared to those who gained (27.8 ± 3.1 kg/m2) or remained stable (28.7 ± 3.6 kg/m2) (P < .001).
Table 1.
Summary of Baseline Patient Characteristics of Institutional Cohort Stratified by BMI Changes
| Characteristic | N | Overall, N = 222a | Gain 2+ BMI, N = 32a | Lost 2+ BMI, N = 27a | Within 2 BMI, N = 163a | P valueb |
|---|---|---|---|---|---|---|
| Gender | 222 | .053 | ||||
| Female | 70 (32%) | 9 (28%) | 14 (52%) | 47 (29%) | ||
| Male | 152 (68%) | 23 (72%) | 13 (48%) | 116 (71%) | ||
| Race | 174 | .2 | ||||
| Asian | 3 (1.7%) | 0 (0%) | 0 (0%) | 3 (2.5%) | ||
| Black or African American | 5 (2.9%) | 3 (10%) | 0 (0%) | 2 (1.6%) | ||
| Hispanic or Latino | 3 (1.7%) | 1 (3.3%) | 0 (0%) | 2 (1.6%) | ||
| White | 163 (94%) | 26 (87%) | 22 (100%) | 115 (94%) | ||
| Age | 222 | 61.36 (11.04) | 57.56 (15.16) | 62.44 (7.95) | 61.93 (10.42) | .4 |
| Preoperative BMI (kg/m2) | 222 | 28.88 (3.67) | 27.77 (3.05) | 31.44 (4.04) | 28.67 (3.56) | <.001 |
| Preoperative AHI (events/h) | 222 | 34.06 (17.32) | 31.23 (13.36) | 37.08 (13.19) | 34.11 (18.55) | .2 |
| Preoperative ESS | 202 | 10.13 (5.30) | 11.80 (5.69) | 10.48 (5.37) | 9.79 (5.20) | .3 |
| Months to postop sleep study (PSG or HSAT) | 208 | 11.37 (12.22) | 7.68 (7.94) | 10.00 (12.67) | 12.29 (12.69) | .2 |
Abbreviations: AHI, apnea–hypopnea index; BMI, body mass index; ESS, Epworth Sleepiness Scale; HSAT, home sleep apnea test; PSG, polysomnogram; SD, standard deviation.
N (%); mean (SD).
Pearson's Chi‐squared test; Fisher's exact test; Kruskal–Wallis rank sum test.
Institutional Cohort Treatment Outcomes
Postoperative BMI, AHI, ESS, and O2 saturation nadir of the institutional cohorts are summarized in Table 2. Overall, AHI decreased from 34.1 ± 17.32 to 19.0 ± 17.18 events/h, with an average reduction of −15.0 ± 20.47 events/h. Sher20 and Sher15 treatment responses were achieved by 53% and 50% of patients, respectively. Patients who lost 2+ BMI points demonstrated the greatest AHI reduction (−24.8 ± 13.8 events/h) compared to those who remained stable (−15.1 ± 20.3 events/h; post hoc P = .018) or gained 2+ BMI points (−6.4 ± 22.5 events/h; post hoc P < .001). Final AHI was lowest among those who lost 2+ BMI points (12.3 ± 8.8 events/h) compared to stable BMI (19.0 ± 16.5 events/h; post hoc P = .041) and BMI gainers (24.8 ± 23.3 events/h; post hoc P = .011). Patients who gained 2+ BMI points had the highest final BMI (31.1 ± 3.3 kg/m2), while those who lost 2+ BMI points had the lowest final BMI (27.0 ± 3.9 kg/m2). Those within 2 BMI points had a final BMI of 28.6 ± 3.6 kg/m2, and this difference was statistically significant (P < .001). The proportion of patients who gained, lost, and remained within 2 BMI meeting Sher15 response criteria was 44%, 59%, and 48%, respectively (P = .5). Similarly, the proportion of patients meeting Sher20 response criteria was 44%, 59%, and 51%, respectively (P = .5). Linear regression analysis demonstrated that each 1‐point increase in final BMI was associated with a 1.4 events/h decline in AHI reduction following therapy (β = 1.4, 95% confidence interval [CI]: 0.64, 2.1; P < .001). Likewise, each 1‐point increase in baseline BMI was associated with a 1.7 events/h decline in AHI reduction following therapy (β = 1.7, 95% CI: 0.53, 2.8; P = .004).
Table 2.
Summary of Postoperative Patient Characteristics of Institutional Cohort Stratified by BMI Changes
| Characteristic | Overall, N = 222a | Gain 2+ BMI, N = 32a | Lost 2+ BMI, N = 27a | Within 2 BMI, N = 163a | P valueb |
|---|---|---|---|---|---|
| Postop sleep study type | .4 | ||||
| HSAT | 141 (64%) | 17 (53%) | 20 (74%) | 104 (64%) | |
| PSG | 23 (10%) | 6 (19%) | 2 (7.4%) | 15 (9.2%) | |
| tPSG | 58 (26%) | 9 (28%) | 5 (19%) | 44 (27%) | |
| Postop BMI (kg/m2) | 28.77 (3.73) | 31.09 (3.25) | 27.01 (3.91) | 28.61 (3.59) | <.001c , d , e |
| BMI at time of postop sleep study (kg/m2) | 28.83 (3.81) | 29.32 (3.67) | 28.35 (3.92) | 28.81 (3.84) | .7 |
| BMI change at time of postop sleep study | −0.11 (2.53) | 1.33 (2.14) | −3.10 (3.02) | 0.13 (2.08) | <.001c , d , e |
| Preop AHI (events/h) | 34.06 (17.32) | 31.23 (13.36) | 37.08 (13.19) | 34.11 (18.55) | .4 |
| Postop AHI (events/h) | 19.04 (17.18) | 24.84 (23.28) | 12.31 (8.79) | 19.02 (16.50) | .020c , d , e |
| AHI Change (events/h) | −15.02 (20.47) | −6.39 (22.52) | −24.77 (13.84) | −15.10 (20.33) | .002c , d , e |
| Preop ESS | 10.13 (5.30) | 11.80 (5.69) | 10.48 (5.37) | 9.79 (5.20) | .2 |
| Postop ESS | 7.71 (5.09) | 7.80 (4.36) | 8.48 (6.40) | 7.55 (4.98) | .7 |
| ESS change | −2.59 (5.17) | −4.74 (4.04) | −2.28 (5.26) | −2.28 (5.27) | .10 |
| Preop SpO2 nadir (%) | 79.79 (7.01) | 80.35 (7.86) | 77.91 (7.55) | 80.03 (6.75) | .5 |
| Postop SpO2 nadir (%) | 82.47 (9.56) | 82.48 (7.91) | 84.30 (5.46) | 82.13 (10.42) | .6 |
| Change in SpO2 nadir (%) | 2.72 (10.31) | 1.70 (9.22) | 7.00 (9.19) | 2.14 (10.57) | .2 |
| Sher15 treatment response | 112 (50%) | 15 (47%) | 19 (70%) | 78 (48%) | .087 |
| Sher20 treatment response | 117 (53%) | 15 (47%) | 19 (70%) | 83 (51%) | .13 |
Abbreviations: AHI, apnea–hypopnea index; BMI, body mass index; ESS, Epworth Sleepiness Scale; HSAT, home sleep apnea test; PSG, polysomnogram; SD, standard deviation; tPSG, titration polysomnogram.
N (%); mean (SD).
Fisher's exact test; Kruskal–Wallis rank sum test; one‐way ANOVA; Pearson's Chi‐squared test.
Significant difference between gain 2+ BMI and lost 2+ BMI.
Significant difference between gain 2+ BMI and within 2 BMI.
Significant difference between lost 2+ BMI and within 2+ BMI.
ADHERE Cohort Baseline Characteristics
A total of 1949 ADHERE registry patients were included for analysis (Table 3). Most patients (N = 1575, 80.8%) remained within 2 BMI points, while 165 patients (8.5%) gained and 209 patients (10.7%) lost 2+ BMI points from baseline to final visit at 12 months postoperatively. Patients who lost 2+ BMI points had the highest baseline BMI (31.1 ± 3.5 kg/m2) compared to patients who gained (29.4 ± 4.2 kg/m2) or remained within 2 BMI points (28.7 ± 3.5 kg/m2) (P < .001). There were no statistical differences in age, baseline AHI, ESS, and ethnicity among the different BMI groups. However, there was a statistically significant difference in gender distribution among the groups, with patients who remained within 2 BMI points having the highest proportion of male patients (76.8%, P = .001).
Table 3.
Summary of Baseline Patient Characteristics of ADHERE Cohort Stratified by BMI Changes
| Variable | Gained 2+ BMI, N = 165 | Lost 2+ BMI, N = 209 | Within 2 BMI, N = 1575 | P valuea |
|---|---|---|---|---|
| Age | 59.85 ± 11.23, N = 164 | 62.02 ± 10.34, N = 209 | 60.6 ± 10.94, N = 1572 | .123 |
| Baseline BMI (kg/m2) | 29.41 ± 4.2, N = 165 | 31.08 ± 3.5, N = 209 | 28.74 ± 3.51, N = 1575 | <.001 |
| Baseline AHI (events/h) | 36.43 ± 16.36, N = 165 | 35.08 ± 15.6, N = 207 | 34.98 ± 14.93, N = 1573 | .501 |
| Baseline ESS | 10.8 ± 5.56, N = 148 | 10.46 ± 6.06, N = 180 | 11.24 ± 5.46, N = 1402 | .154 |
| Gender | .001 | |||
| Male | 65.6% (107) | 68.4% (143) | 76.8% (1205) | |
| Female | 34.4% (56) | 31.6% (66) | 23.2% (364) | |
| Ethnicity | .518 | |||
| Hispanic or Latino | 0.6% (1) | 1.5% (3) | 2% (31) | |
| Not Hispanic or Latino | 99.4% (161) | 98.5% (197) | 98% (1503) |
N (%); mean (SD).
Table includes patients who had a BMI entered at baseline and final visit.
Abbreviations: AHI, apnea–hypopnea index; BMI, body mass index; ESS, Epworth Sleepiness Scale; SD, standard deviation.
P values derived using ANOVA for numeric and Fisher's exact test for categorical variables.
ADHERE Cohort Treatment Outcomes
Postoperative AHI, ESS, Sher15, and Sher20 outcomes are summarized in Table 4. Similar to our institutional cohort, the final AHI was the lowest in patients who lost 2+ BMI points (13.26 ± 13.21 events/h), followed by patients who remained within 2 BMI points (15.86 ± 14.62 events/h), and highest in patients who gained 2+ BMI points (19.83 ± 17.97 events/h) (P = .003). Patients who lost 2+ BMI points demonstrated the greatest reduction in AHI (−22.58 ± 19.76 events/h) compared to those who remained within 2 BMI points (−18.85 ± 18.11 events/h; post hoc P = .0434) or gained 2+ BMI points (−15.82 ± 19.63 events/h; post hoc P = .0049), respectively.
Table 4.
Summary of Postoperative Patient Characteristics of ADHERE Cohort Stratified by BMI Changes
| Variable | Gained 2+ BMI, N = 165 | Lost 2+ BMI, N = 209 | Within 2 BMI, N = 1575 | P valuea |
|---|---|---|---|---|
| Final BMI (kg/m2) | 33.08 ± 4.82, N = 165 | 27.13 ± 3.76, N = 209 | 28.75 ± 3.62, N = 1575 | <.001c , d , e |
| Final AHI (events/h)b | 19.83 ± 17.97, N = 136 | 13.26 ± 13.21, N = 160 | 15.86 ± 14.62, N = 1250 | .003c , d , e |
| Decrease in AHI (events/h) | 15.82 ± 19.63, N = 136 | 22.58 ± 19.76, N = 158 | 18.85 ± 18.11, N = 1248 | .006c , e |
| Final ESSb | 7.09 ± 4.68, N = 134 | 6.08 ± 4.52, N = 165 | 6.87 ± 4.61, N = 1323 | .044e |
| Decrease in ESS | 3.57 ± 4.46, N = 121 | 4.76 ± 4.99, N = 143 | 4.29 ± 5.25, N = 1195 | .174 |
| Final therapy use (hour/night) | 5.34 ± 2.21, N = 137 | 5.71 ± 2.3, N = 162 | 5.78 ± 2.23, N = 1248 | .09 |
| Sher20 response | ||||
| Nonresponder | 45.6% (62) | 32.3% (51) | 39.3% (490) | .064 |
| Responder | 54.4% (74) | 67.7% (107) | 60.7% (758) | |
| Sher15 response | ||||
| Nonresponder | 51.5% (70) | 35.4% (56) | 44.7% (558) | .018 |
| Responder | 48.5% (66) | 64.6% (102) | 55.3% (690) | |
| Postop sleep study type | ||||
| HSAT | 50.3% (76) | 53% (97) | 55.8% (777) | .454 |
| tPSG | 43% (65) | 36.6% (67) | 38.2 (532) | |
N (%); mean (SD).
Not all patients indicated what type of sleep test was done at the final visit; thus, resulting in total percentages of <100%.
Abbreviations: AHI, apnea–hypopnea index; BMI, body mass index; ESS, Epworth Sleepiness Scale; HSAT, home sleep apnea test; SD, standard deviation; tPSG, titration polysomnogram.
P values derived using ANOVA for numeric and Fisher's exact test for categorical variables unless otherwise specified.
P values derived using Kruskal–Wallis test due to non‐normality.
Significant difference between gain 2+ BMI and lost 2+ BMI.
Significant difference between gain 2+ BMI and within 2 BMI.
Significant difference between lost 2+ BMI and within 2+ BMI.
Univariate linear regression revealed that patients who gained 2+ BMI points experienced a 3.03 events/h decrease in AHI reduction compared to patients who remained within 2 BMI points (P = .069), whereas patients who lost 2+ BMI points experienced a 3.74 events/h increase in reduction of AHI compared to patients who remained within 2 BMI points (P = .016). Linear regression analysis did not show a significant association between baseline BMI or final BMI and AHI reduction.
Final BMI was highest in patients who gained 2+ BMI points (33.1 ± 4.8 kg/m2) compared to those who lost 2+ BMI (27.1 ± 3.8 kg/m2) or remained within 2 BMI (28.8 ± 3.6 kg/m2) (P < .001). Patients who lost 2+ BMI points had significantly lower final ESS compared to other groups (P = .044). Therapy use (hours/night) did not differ among groups.
The proportion of patients meeting the Sher20 response threshold was 54.4%, 67.7%, and 60.7% among those who gained, lost, and remained within 2 BMI points, respectively (P = .006). The proportion of Sher15 responders who gained, lost, and remained within 2 BMI points was 48.5%, 64.6%, and 55.3%, respectively (P = .018). A logistic regression analysis of Sher20 treatment response at final visit did not identify any of the BMI changes as an independent predictor of Sher20 treatment response. However, same analysis of Sher15 response showed a 47% increased odds of treatment response in patients who lost 2+ BMI compared to those who stayed within 2 BMI points (P = .028).
Subgroup Analysis of 2+ BMI Cohort
Among patients who gained 2+ BMI points, there were no significant differences in final AHI, AHI changes, ESS, or therapy use between those with a final BMI > 35 kg/m2 (N = 51) and those with a final BMI ≤ 35 kg/m2 (N = 114). The proportion of patients meeting the Sher20 response threshold was 57% among those with a final BMI ≤ 35 kg/m² and 48.8% among those with a final BMI > 35 kg/m²; however, this difference was not statistically significant (Table 5).
Table 5.
Summary of Postoperative Patient Characteristics of ADHERE Patients Who Gained 2+ BMI Points, Stratified by Final BMI (BMI ≤ 35 vs BMI > 35)
| Variable | BMI ≤ 35 (N = 114) | BMI > 35 (N = 51) | P valuea |
|---|---|---|---|
| Final BMI | 30.7 ± 2.92, N = 114 | 38.42 ± 3.87, N = 51 | <.001 |
| Final AHI (events/h)b | 19.95 ± 18.8, N = 93 | 19.56 ± 16.24, N = 43 | .763 |
| Decrease in AHI (events/h) | 15.6 ± 18.6, N = 93 | 16.29 ± 21.91, N = 43 | .857 |
| Final ESSb | 6.99 ± 4.41, N = 93 | 7.32 ± 5.27, N = 41 | .936 |
| Decrease in ESS | 3.65 ± 4.68, N = 85 | 3.39 ± 3.97, N = 36 | .758 |
| Final therapy use (hour/night)b | 5.2 ± 2.28, N = 95 | 5.67 ± 2.02, N = 42 | .265 |
| Sher20 response | |||
| Nonresponder | 43% (40) | 51.2% (22) | .4594 |
| Responder | 57% (53) | 48.8% (21) | |
| Sher15 response | |||
| Nonresponder | 49.5% (46) | 55.8% (24) | .5807 |
| Responder | 50.5% (47) | 44.2% (19) | |
N (%); mean (SD).
Table includes patients who had a BMI entered at baseline and final visit.
Abbreviations: AHI, apnea–hypopnea index; BMI, body mass index; ESS, Epworth Sleepiness Scale.
P values derived using Student's t‐test for numeric and Fisher's exact test for categorical variables unless otherwise specified.
P values derived using Wilcoxon test due to non‐normality.
Discussion
This is the largest published study examining the impact of weight change on HGNS response. Our study shows that the patients who lost 2+ BMI had a greater AHI reduction and a lower final AHI with HGNS. Similarly, the greater surgical success was observed in ADHERE registry patients who lost 2+ BMI. These findings support that postimplantation weight changes significantly correlate with HGNS response, highlighting the importance of adjunctive weight management after the HGNS implantation.
It is a well‐accepted paradigm that weight changes directly correspond with OSA severity, which could explain the findings we see in our study. The Wisconsin cohort sleep study found that 10% of changes in body weight corresponded with approximately 30% of changes in the AHI irrespective of the patient's baseline body weight. 15 Likewise, Young et al showed a 4‐fold increase in the risk of OSA per 1 standard deviation increase in BMI. 16 Anatomical changes due to weight gain, such as increased airway collapsibility and tongue volume, could explain this relationship. A case‐control study by Schwab et al demonstrated that OSA patients had larger anatomical volumes of the lateral pharyngeal walls, tongue, and overall soft tissue compared to non‐OSA patients. 1 , 17 Other studies have also linked obesity to greater tongue volume, a critical factor in maintaining upper airway patency. 2 , 3 However, the influence of weight changes on HGNS response may be more nuanced, as weight change may directly influence on HGNS therapy efficacy in addition to affecting OSA severity. These anatomical changes, which correlate with weight changes, have also been identified as potential independent predictors of HGNS response. Notably, presence of complete lateral wall collapse and increased PAP pressure have been associated with lower odds of treatment response. 4 , 18
While many previous studies have demonstrated a predictive value of preoperative BMI on HGNS success, our findings emphasize the predictive value of postoperative BMI changes. 7 , 8 , 9 , 10 This aligns with Renslo et al, which showed that a lower BMI at the time of the second postoperative sleep study was predictor of better HGNS response. 12 In our study, regression analysis among the ADHERE cohort demonstrated a significantly greater reduction in AHI among patients who lost 2+ BMI, whereas baseline BMI had no significant correlation with AHI changes. Furthermore, the final BMI did not significantly predict AHI change on regression analysis (P = .07), and our subgroup analysis based on final BMI > 35 kg/m2 yielded inconclusive results. These findings suggest that HGNS response may be more influenced by weight change than preop or final BMI.
It is important to note that no differences in HGNS usage were observed among the different BMI groups. Furthermore, all BMI groups demonstrated significant symptom improvement, as evidenced by changes in the ESS. Some patients achieved therapeutic success with HGNS as monotherapy, even in the context of significant weight gain. While the therapy is generally well tolerated and can lead to significant reduction in symptom and AHI even in the patients with significant weight gain, the effectiveness is variable, demonstrated by the incomplete number of patients achieving treatment success in both institutional and adhere cohort. Therefore, it is essential for clinicians to identify patients with residual OSA burden and consider adjunctive tools to support their treatment with HGNS, given the potential long‐term cardiovascular risks of untreated or undertreated OSA. Potential adjunctive options include chin straps, surgical modifications of the pharyngeal airway, oral appliances, and weight loss. 19 , 20 , 21 Traditionally limited to lifestyle changes or surgery, weight management now include GLP‐1 agonists, which have shown efficacy in reducing both weight and OSA severity in a well‐tolerated and sustained manner. 22 Thus, these medications may represent an important adjunct to HGNS, particularly in the management of patients with subtherapeutic response to HGNS monotherapy given the findings of our study.
There are several limitations of this study. First, there is a lack of standardization in measuring the postoperative AHI. While the trend is shifting from titration studies toward full‐night PSG or HSAT to determine postoperative AHI, we cannot control for the differences in diagnostic equipment and protocols among different institutions to obtain standardized outcomes. Second, most patients in the analyses remained within 2 BMI points, limiting the sample size for those with significant BMI changes. Additionally, our study utilizes the BMI at the time of the final visit, which may not align with the BMI at the time of the sleep study, thus limiting the accuracy of our analysis. To address this issue, we aimed to collect postoperative AHI from the sleep study closest to the 1‐year follow‐up for the institutional cohort. However, many patients only had sleep study data available from a few months before or after this time point, introducing some variability in these measurements. The exact timing of sleep studies was not available for analysis for ADHERE patients. Thus, their final AHI could have been derived from postop titration study, which typically is performed 2 to 6 months from the implantation. For others, it may be derived from additional sleep studies conducted closer to the final 12 months visit once usage, tolerance, and symptom controls are optimized. Furthermore, the retrospective analysis of the study poses inherent selection bias, favoring patients with complete data. A large portion of our study period were excluded from this study due to missing data on pre‐ or postoperative AHI and/or BMI reflective of this issue. Finally, both the ADHERE and institutional cohort demographics predominantly consisted of healthy white middle‐aged males, which may restrict the generalizability of results.
Conclusion
In summary, our study underscores the significant impact of postoperative BMI changes on HGNS efficacy. Larger prospective studies are needed to clarify the relationship and mechanism between weight changes and HGNS response. Weight loss could play an important adjunctive role to HGNS therapy, especially with new options such as GLP‐1 agonists.
Future studies should undertake a prospective design to concurrently track postoperative weight change and AHI at fixed time points. It remains important to delineate an optimal weight, or weight reduction percentage, wherein HGNS provides the greatest efficacy. Additionally, research into how different methods of weight reduction, such as through decreased caloric intake, increased exercise, medications, and surgical interventions, may impact HGNS efficacy is of interest. Finally, long‐term HGNS efficacy data following sustained weight reduction is necessary to determine whether the beneficial effects of reduced AHI from weight loss are enduring.
Author Contributions
Chihun J. Han, conception, design, analysis, interpretation, manuscript writing; Praneet C. Kaki, data collection, analysis, interpretation, manuscript writing; Jennifer A. Goldfarb, data collection, manuscript writing; Thomas M. Kaffenberger, interpretation, manuscript writing; Nicole L. Molin, interpretation, manuscript writing; Anna Matzke, design, interpretation, manuscript writing; Maurits S. Boon, design, interpretation, manuscript writing, final approval; Colin T. Huntley, design, interpretation, manuscript writing, final approval.
Disclosures
Competing interests
Colin T. Huntley: research support from Nyxoah, Inspire Medical Systems, and Sommetrics; consultant for Nyxoah and Avivomed. Maurits S. Boon: chief medical officer for Nyxoah. Thomas M. Kaffenberger: consultant for Inspire Medical Systems and Nyxoah. Anna Matzke: supervisor, clinical data management for Inspire.
Funding source
None.
Acknowledgments
None.
This article was presented at the AAO‐HNSF 2024 Annual Meeting & OTO EXPO; September 28‐October 1, 2024; Miami Beach, Florida.
References
- 1. Gastaut H, Tassinari CA, Duron B. Polygraphic study of the episodic diurnal and nocturnal (hypnic and respiratory) manifestations of the pickwick syndrome. Brain Res. 1966;1(2):167‐186. 10.1016/0006-8993(66)90117-X [DOI] [PubMed] [Google Scholar]
- 2. Pépin JL, Krieger J, Rodenstein D, et al. Effective compliance during the first 3 months of continuous positive airway pressure. Am J Respir Crit Care Med. 1999;160(4):1124‐1129. 10.1164/ajrccm.160.4.9802027 [DOI] [PubMed] [Google Scholar]
- 3. Wolkove N, Baltzan M, Kamel H, Dabrusin R, Palayew M. Long‐term compliance with continuous positive airway pressure in patients with obstructive sleep apnea. Can Respir J. 2008;15(7):365‐369. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Costantino A, Rinaldi V, Moffa A, et al. Hypoglossal nerve stimulation long‐term clinical outcomes: a systematic review and meta‐analysis. Sleep Breath. 2020;24:399‐411. [DOI] [PubMed] [Google Scholar]
- 5. Strollo PJ, Soose RJ, Maurer JT, et al. Upper‐airway stimulation for obstructive sleep apnea. N Engl J Med. 2014;370(2):139‐149. 10.1056/NEJMoa1308659 [DOI] [PubMed] [Google Scholar]
- 6. Sher AE, Schechtman KB, Piccirillo JF. The efficacy of surgical modifications of the upper airway in adults with obstructive sleep apnea syndrome. Sleep. 1996;19(2):156‐177. 10.1093/sleep/19.2.156 [DOI] [PubMed] [Google Scholar]
- 7. Strollo PJ, Gillespie MB, Soose RJ, et al. Upper airway stimulation for obstructive sleep apnea: durability of the treatment effect at 18 months. Sleep. 2015;38(10):1593‐1598. 10.5665/sleep.5054 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Woodson BT, Soose RJ, Gillespie MB, et al. Three‐year outcomes of cranial nerve stimulation for obstructive sleep apnea. Otolaryngol Head Neck Surg. 2016;154(1):181‐188. 10.1177/0194599815616618 [DOI] [PubMed] [Google Scholar]
- 9. Woodson BT, Strohl KP, Soose RJ, et al. Upper airway stimulation for obstructive sleep apnea: 5‐year outcomes. Otolaryngol Head Neck Surg. 2018;159(1):194‐202. 10.1177/0194599818762383 [DOI] [PubMed] [Google Scholar]
- 10. Heiser C, Steffen A, Boon M, et al. Post‐approval upper airway stimulation predictors of treatment effectiveness in the ADHERE registry. Eur Respir J. 2019;53(1):1801405. 10.1183/13993003.01405-2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Steffen A, Hartmann JT, König IR, Ravesloot MJL, Hofauer B, Heiser C. Evaluation of body position in upper airway stimulation for obstructive sleep apnea—is continuous voltage sufficient enough. Sleep Breath. 2018;22(4):1207‐1212. 10.1007/s11325-018-1716-5 [DOI] [PubMed] [Google Scholar]
- 12. Renslo B, Virgen CG, Sawaf T, et al. Long‐term trends in body mass index throughout upper airway stimulation treatment: does body mass index matter? J Clin Sleep Med. 2023;19(6):1061‐1071. 10.5664/jcsm.10496 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.CDC. Healthy Weight. Diabetes. June 5, 2024. https://www.cdc.gov/diabetes/living-with/healthy-weight.html
- 14. Jensen MD, Ryan DH, Apovian CM, et al. 2013. AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. Circulation. 2014;129(25 Suppl 2):S102‐S138. 10.1161/01.cir.0000437739.71477.ee [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Peppard PE. Longitudinal study of moderate weight change and sleep‐disordered breathing. JAMA. 2000;284(23):3015‐3021. 10.1001/jama.284.23.3015 [DOI] [PubMed] [Google Scholar]
- 16. Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep‐disordered breathing among middle‐aged adults. N Engl J Med. 1993;328(17):1230‐1235. 10.1056/NEJM199304293281704 [DOI] [PubMed] [Google Scholar]
- 17. Schwab RJ, Pasirstein M, Pierson R, et al. Identification of upper airway anatomic risk factors for obstructive sleep apnea with volumetric magnetic resonance imaging. Am J Respir Crit Care Med. 2003;168(5):522‐530. 10.1164/rccm.200208-866OC [DOI] [PubMed] [Google Scholar]
- 18. Huyett P, Kent DT, D'Agostino MA, et al. Drug‐induced sleep endoscopy and hypoglossal nerve stimulation outcomes: a multicenter cohort study. Laryngoscope. Jul 2021;131(7):1676‐1682. 10.1002/lary.29396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Ramaswamy AT, Li C, Suurna MV. A case of hypoglossal nerve stimulator‐resistant obstructive sleep apnea cured with the addition of a chin strap. Laryngoscope. 2018;128:1727‐1729. [DOI] [PubMed] [Google Scholar]
- 20. Lee JJ, Sahu N, Rogers R, Soose RJ. Severe obstructive sleep apnea treated with combination hypoglossal nerve stimulation and oral appliance therapy. J Dent Sleep Med. 2015;2:185‐186. [Google Scholar]
- 21. Steffen A, Abrams N, Suurna MV, Wollenberg B, Hasselbacher K. Upper‐airway‐stimulation before, after or without uvulopalatopharyngoplasty ‐ a two‐year perspective. Laryngoscope. 2019;129:514‐518. [DOI] [PubMed] [Google Scholar]
- 22. Malhotra A, Grunstein RR, Fietze I, et al. Tirzepatide for the treatment of obstructive sleep apnea and obesity. N Engl J Med. Jun 2024;391:1193‐1205. 10.1056/NEJMoa2404881 [DOI] [PMC free article] [PubMed] [Google Scholar]