Evaluation of Oral Appliance Therapy in Patients with Obstructive Sleep Apnea Syndrome: A Comparative Analysis by Age, Severity, and BMI

ABSTRACT

Background

Obstructive sleep apnea syndrome is a common disorder characterized by repeated upper airway obstruction during sleep. Oral appliance therapy is a noninvasive treatment option commonly recommended for patients who are intolerant to continuous positive airway pressure. This study aimed to evaluate the treatment outcomes of oral appliance therapy in patients with obstructive sleep apnea syndrome and to examine the relationship between therapeutic effectiveness and patient characteristics, including age, disease severity, and body mass index.

Methods

A total of 42 patients diagnosed with obstructive sleep apnea syndrome and treated with oral appliance therapy were retrospectively analyzed. The treatment effect was assessed using the apnea–hypopnea index measured before and after therapy (by polysomnography or portable monitoring). Effective treatment was defined as a post-treatment apnea–hypopnea index of <15 events per hour and a ≥ 50% reduction from baseline.

Results

The mean apnea–hypopnea index decreased significantly from 27.6 ± 15.7 to 10.3 ± 8.7 events per hour following oral appliance therapy, with a mean improvement in AHI of 59.6%. Effectiveness was achieved in 66.7% of patients. A significant overall improvement was observed. Reductions in the apnea–hypopnea index were seen across age, severity, and body mass index (BMI) strata. However, only three patients had BMI ≥30 kg/m²; hence, BMI-stratified findings for this subgroup are descriptive and no inferential testing was conducted.

Conclusion

Oral appliance therapy was effective in reducing the apnea–hypopnea index in a broad range of patients, including those traditionally considered less responsive because of advanced age, obesity, or severe disease. These findings suggest a potential role for oral appliance therapy as an alternative to continuous positive airway pressure treatment within the examined strata (age, severity, and BMI), particularly when continuous positive airway pressure therapy is not feasible. Larger prospective studies are warranted.

Obstructive sleep apnea syndrome (OSAS) is the most prevalent type of sleep-disordered breathing, and is characterized by repeated episodes of upper airway obstruction during sleep, often accompanied by snoring and respiratory effort. OSAS is strongly associated with male sex, aging, and obesity.^1,2,3 Craniofacial morphology, particularly mandibular retrusion, is also recognized as a contributing factor, especially among Asians.⁴ The apnea–hypopnea index (AHI) is the standard metric for diagnosing OSAS, representing the number of apnea and hypopnea episodes per hour of sleep. Severity is classified as mild (5–15 events/h), moderate (15–30 events/h), and severe (≥ 30 events/h). OSAS is associated with increased cardiovascular morbidity and mortality due to intermittent hypoxia, hypercapnia, and sympathetic activation. Continuous positive airway pressure (CPAP) therapy remains the gold standard treatment, especially for moderate-to-severe cases. However, oral appliances (OAs) are recommended for patients with an AHI < 20 events/h or those who are intolerant to CPAP. Although OA therapy is considered less effective in older and obese patients, a clinical consensus in Japan remains unclear. This study aimed to evaluate the outcomes of OA therapy by stratifying the patients according to age, OSAS severity, and body mass index (BMI).

MATERIALS AND METHODS

We retrospectively analyzed patients diagnosed with OSAS at the Department of Oral and Maxillofacial Surgery, Tottori University Hospital, between January 1, 2004, and December 31, 2020, who were treated with an OA. Of these, 21 patients were diagnosed using a portable monitor, and 14 were diagnosed using full-night polysomnography (PSG). In six cases, full-night polysomnography was used for the initial diagnosis and a portable monitor was used for re-evaluation. The diagnostic method was unclear in one case. The reasons for referral for OA therapy were as follows: 16 patients had an AHI < 20 events/h; seven patients showed poor compliance with CPAP; 12 patients refused CPAP and initially preferred OA therapy; six patients were judged by the referring physician to be more suitable for OA therapy because of mandibular morphology or other anatomical considerations; and one patient wished to use OA as an alternative to CPAP during travel. After the OA treatment and follow-up evaluation, the results were returned to the referral department. The referrals originated from the Departments of Respiratory Medicine (n = 22), Otorhinolaryngology (n = 13), Cardiology (n = 3), and external institutions (n = 4). The inclusion criteria were patients aged 20–90 years with available AHI data before and after treatment. Patients who opted out of the study were excluded. Candidates for OA therapy were assessed for exclusion criteria such as extensive tooth loss, severe periodontal disease, or temporomandibular joint disorders. The OAs were fabricated using a thermoplastic acrylic resin, with mandibular advancement set to approximately 70% of the maximum protrusion. Adjustments were made based on subjective symptoms. Post-treatment reassessment was performed by PSG or a portable monitor; if results were insufficient, further mandibular advancement and retesting were conducted. Improvement in AHI was defined as ((AHI before treatment−AHI after treatment)/AHI before treatment) × 100. Positive values indicate reduction from baseline. Effectiveness was defined as a post-treatment AHI <15 events/h and an improvement in AHI ≥ 50%. Subjective outcomes (e.g., daytime sleepiness and quality of life) were not systematically collected in this retrospective review. The durations of OA use and adherence were not uniformly recorded in the source charts; therefore, formal adherence analyses were not performed. Statistical analyses were performed using IBM SPSS Statistics version 21. Subgroups were compared using the chi-squared and Kruskal–Wallis tests. Analysis of covariance (ANCOVA) was performed using BMI as a covariate. Pearson’s correlation coefficient was used for correlation analysis. Within-group pre–post comparisons were performed using paired t-tests or Wilcoxon signed-rank tests, as appropriate, based on distributional assumptions. The study protocol was approved by the Ethics Committee of the Tottori University Hospital (approval number: 21A182).

RESULTS

Comparison with OA-prescribed patients without post-treatment evaluation

The evaluated group comprised 42 patients [29 men (69.0%), 13 women (31.0%); mean age: 57.6 ± 10.2 years; mean BMI: 24.7 ± 3.4 kg/m²]. The mean pre-treatment AHI was 27.6 ± 15.7/h. Severity classification included 9 mild, 18 moderate, and 15 severe cases. Compared with the 61 patients who were prescribed OAs without post-treatment evaluation, no significant differences were observed in sex, age, BMI, AHI, or severity (Table 1). Baseline characteristics were similar between the evaluated cohort and patients prescribed OA but without post-treatment evaluation; however, because post-treatment outcomes were unavailable in the latter group, this group did not constitute a control, and residual selection bias cannot be excluded.

Table 1.  Clinical characteristics of patients with and without effectiveness assessment.

	Evaluated cohort (OA-prescribed with post-treatment AHI assessment)	OA-prescribed without post-treatment AHI assessment
Sex
Male (n)	29	37
Female (n)	13	24
Average age (years)	57.6 ± 10.2	55.3 ± 13.9
Average BMI (kg/m2)	24.7 ± 3.4	23.8 ± 4.0
Average AHI (events/h)	27.6 ± 15.7	23.1 ± 14.4
Severity Classification
Mild (n)	9	20
Moderate (n)	18	30
Severe (n)	15	11

Overall effectiveness

After OA therapy, the mean AHI decreased significantly from 27.6 ± 15.7 events/h to 10.3 ± 8.7 events/h (P < 0.001), with a mean improvement in AHI of 59.6 ± 34.0%. Among the 42 patients, 11 (26.2%) achieved an AHI < 5 events/h. Effectiveness (AHI < 15 events/h and ≥ 50% improvement in AHI) was observed in 28 patients (66.7%). The results are summarized in Table 2.

Table 2.  Effect of OA therapy on AHI in the evaluated group (n = 42).

	Mean ± SD
AHI before treatment (events/h)	27.6 ± 15.7
AHI after treatment (events/h)	10.3 ± 8.7
Improvement in AHI (%)	59.6 ± 34.0
Effectiveness rate (%)	66.7
AHI < 5/h (%)	26.2

Age-based analysis

The patients were categorized into three age groups: < 50 years (n = 9), 50–64 years (n = 23), and ≥ 65 years (n = 10). All groups showed significant reductions in AHI. Improvement in AHI was highest in the 50–64 years group (66.5 ± 22.3%), followed by the ≥ 65 years (57.9 ± 26.2%) and < 50 years (43.8 ± 54.6%) groups. No significant differences in effectiveness were found between the groups (Table 3, Fig. 1). Between-group comparisons were conducted using the Kruskal–Wallis test; detailed p-values are provided in the corresponding tables.

Table 3.  Comparison of OA therapy effectiveness by age group.

	< 50 years (n = 9)	50–64 years (n = 23)	≥ 65 years (n = 10)	P-Value
AHI before treatment (events/h)	31.8 ± 18.9	27.0 ± 15.5	25.5 ± 11.7	0.835
AHI after treatment (events/h)	11.9 ± 7.3	9.4 ± 8.8	11.1 ± 9.2	0.553
Improvement in AHI (%)	43.8 ± 54.6	66.5 ± 22.3	57.9 ± 26.2	0.624
Effectiveness rate (%)	44.4	78.3	60	0.166
AHI < 5/h (%)	22.2	30.4	40	0.702

Fig. 1.

 Comparison of the effectiveness of OA therapy on AHI by age group. A significant reduction in AHI was observed across all age groups following OA therapy.

Severity-based analysis

All severity groups exhibited significant reductions in AHI. Improvement rates were the highest in the severe group (68.0 ± 24.2%), followed by moderate (58.8 ± 23.2%) and mild (47.3 ± 55.3%) groups. The rate of achieving an AHI < 5 events/h was higher in the mild group (55.6%) (Table 4 and Fig. 2). Under our effectiveness definition, the highest baseline AHI among effective cases was 53.7 events/h, which decreased to 14.2 events/h post-treatment. By contrast, two ultra-severe cases (69.8 events/h and 75.0 events/h) decreased to 15.5 events/h and 34.0 events/h, respectively, and thus did not meet the effectiveness criterion; nevertheless, both showed marked improvement in AHI. These observations are descriptive and hypothesis-generating. Between-group comparisons were conducted using the Kruskal–Wallis test; detailed p-values are provided in the corresponding tables.

Table 4.  Comparison of OA therapy effectiveness by severity classification.

	Mild	Moderate	Severe	P-Value
	(n = 9)	(n = 18)	(n = 15)
AHI before treatment (events/h)	11.0 ± 3.0	27.0 ± 15.5	44.9 ± 12.5	Kruskal–Wallis P < 0.001†
AHI after treatment (events/h)	6.4 ± 6.6	9.4 ± 8.8	14.5 ± 10.9	0.083
Improvement in AHI (%)	47.3 ± 55.3	58.8 ± 23.2	68.0 ± 24.2	0.497
Effectiveness rate (%)	55.6	66.7	73.3	0.67
AHI < 5/h (%)	55.6	22.2	20	0.128

†Pairwise comparisons were Holm-adjusted: Mild vs Moderate P < 0.01, Mild vs Severe P < 0.001, Moderate vs Severe P < 0.001.

Fig. 2.

 Comparison of the effectiveness of OA therapy on AHI by OSAS severity. A significant reduction in AHI was observed across all severity groups. Although the effectiveness rate was higher in moderate and severe cases, the proportion of patients achieving an AHI < 5/h was lower in the severe group.

BMI-based analysis

Effectiveness rates were 68.4% for BMI < 25, 56.3% for 25–< 30, and 66.7% for ≥ 30 (Table 5). Among the 38 patients with available BMI data, BMI < 25 kg/m² and 25–< 30 kg/m² groups showed a significant reduction in AHI after OA therapy. In the BMI ≥30 kg/m² subgroup (n = 3), AHI changed for each patient as follows: 7.4→0.0, 43.0→16.7, and 35.4→14.2. The median paired reduction (before minus after) was 21.2 (range 7.4–26.3). Notably, 2 of 3 patients in this subgroup met the effectiveness criterion. Given the small sample size, these observations were descriptive, and no formal hypothesis testing was conducted. The individual improvement in AHI was 100.0%, 61.2%, and 59.9%, respectively. Improvement in AHI was 58.2 ± 26.9% for BMI < 25 kg/m², 57.7 ± 44.8% for 25–< 30 kg/m², and 73.7 ± 18.6% for ≥ 30 kg/m². (Table 5, Fig. 3). Between-group comparisons were conducted using the Kruskal–Wallis test; detailed p-values are provided in the corresponding tables.

Table 5.  Comparison of OA therapy effectiveness by BMI category.

	BMI < 25.0 kg/m2 (n = 19)	25.0 ≤ BMI < 30.0 kg/m2 (n = 16)	BMI ≥ 30 kg/m2 (n = 3)
AHI before treatment (events/h)	27.3 ± 12.9	29.5 ± 19.7	28.6 ± 15.3
AHI after treatment (events/h)	10.3 ± 7.7	11.0 ± 10.5	10.3 ± 7.4
Improvement in AHI (%)	58.2 ± 26.9	57.7 ± 44.8	73.7 ± 18.6
Effectiveness rate (%)	68.4	56.3	66.7
AHI < 5/h (%)	21.1	37.5	33.3

Fig. 3.

 Comparison of AHI before and after OA therapy by BMI category. A reduction in AHI was observed across BMI categories. Given that the BMI ≥ 30 kg/m² subgroup included only three patients (n = 3), no inferential testing was performed and the results are descriptive.

In ANCOVA adjusted for continuous BMI (and baseline AHI), no significant association with improvement in AHI was detected.

DISCUSSION

Japanese clinical guidelines recommend OA therapy for OSAS patients with AHI <20 events/h and CPAP for those with AHI ≥ 20 events/h.³ CPAP is widely considered the gold standard treatment, particularly for moderate-to-severe OSAS, owing to its well-established benefits in reducing cardiovascular morbidity and mortality.^5,6,7,8,9 However, adherence to CPAP remains a major clinical challenge, and OA has emerged as a viable alternative, especially when CPAP intolerance occurs.¹⁰

Although OA therapy is generally thought to be more effective in patients with mild OSAS, several recent studies, including multicenter analyses, have reported effectiveness rates of approximately 67%.^{11, 12} Our study demonstrated a similar effectiveness rate of 66.7%, although the sample was skewed toward moderate and severe cases (78.6%). This suggests that even in more advanced disease states, OA therapy may retain its clinical effectiveness when tailored appropriately. In our study, effectiveness was defined as achieving both a post-treatment AHI <15 events/h and a ≥ 50% reduction, whereas Okuno et al.^{11, 12} defined it as either an AHI < 5 events/h or a ≥ 50% reduction. Although the definitions differ, none of the cases deemed effective in our study would have been classified as ineffective according to the criteria of Okuno et al.

Age has traditionally been considered a limiting factor for OA effectiveness, with concerns over age-related anatomical changes such as pharyngeal narrowing, reduced muscle tone, and soft tissue laxity.¹³ However, our results demonstrated that the reduction in AHI was statistically significant across all age groups, including patients ≥ 65 years. This finding aligns with recent evidence indicating that age alone may not be a reliable predictor of OA response and supports its use in older populations, provided careful monitoring and device titration are performed.^{12, 14}

Regarding disease severity, OA therapy has been considered less effective in patients with severe OSAS.^{15, 16} However, in our study, although the proportion of patients who achieved an AHI < 5 events/h was lower in both the moderate and severe groups, the overall effectiveness rate was highest in the severe group (73.3%). This finding highlights an important clinical implication: even when OA therapy does not fully normalize AHI in severe OSA, it can still produce substantial improvement in AHI from baseline. Moreover, as OA therapy generally demonstrates higher adherence rates than that for CPAP, the sustained use of an OA may lead to meaningful clinical benefits, including symptom relief, improved quality of life, and potentially reduced cardiovascular risk.^{9, 15, 17}

Obesity is another critical factor in OSAS pathogenesis, with high BMI contributing to upper airway collapse via fat deposition in the tongue and pharyngeal tissues.¹⁸ Previous studies have reported diminished OA responsiveness in obese patients.¹⁹ In our cohort, only three patients had a BMI ≥ 30.0 kg/m², which precluded meaningful statistical comparisons for this category. Notably, two of the three patients in this subgroup met the effectiveness criterion, and thus the results of this subgroup should be interpreted as descriptive and exploratory. Larger studies including a sufficient number of patients with severe obesity (≥ 35.0 kg/m² or ≥ 40.0 kg/m²) are needed to determine the effectiveness of OA therapy in this population.

Although several studies have reported attenuated OA responses in older, more obese, or more severe OSA populations, our null findings across strata may primarily reflect limited power, a narrower BMI distribution, and cohort-specific clinical pathways and titration practices; thus, the absence of evidence should not be interpreted as evidence of absence.^{12, 15, 16, 18,19,20}

A notable limitation of our study was the relatively small sample size (n = 42) for treatment effect analysis. In particular, only 38 patients had available BMI data, and fewer than half underwent post-treatment PSG, partly because of subjective symptom improvement or logistical challenges. Clinically, OA therapy was ineffective in our two ultra-severe cases (baseline AHI 69.8 and 75.0 events/h), whereas the highest baseline AHI among effective cases was 53.7 events/h, suggesting a potential ceiling effect at very high disease burden beyond which mandibular advancement alone may be insufficient; given the small numbers, this observation is hypothesis-generating and requires confirmation in larger cohorts. This introduced potential selection bias and limited the generalizability of our findings. In addition, long-term follow-up data are unavailable, precluding the assessment of sustained effectiveness or adverse effects, such as dental misalignment or temporomandibular joint dysfunction.^{3, 10} A further limitation is the absence of validated patient-reported outcomes (e.g., Epworth Sleepiness Scale and Functional Outcomes of Sleep Questionnaire measures), which precludes the assessment of symptomatic benefit beyond AHI. Objective adherence metrics (e.g., nightly wear time and nights/week) were not systematically recorded, precluding the analysis of dose–response (use vs. AHI reduction) and limiting inferences regarding durability beyond the initial reassessment. Pre–post AHI comparisons mixed PSG and portable monitors, introducing measurement heterogeneity that may bias effect estimates and limit generalizability. Future studies should focus on prospective, large-scale studies that incorporate long-term outcomes and objective adherence monitoring. Furthermore, the use of imaging or airflow dynamics analysis may help identify phenotypes that are more likely to respond to OA.^{14, 20} Combining OA with adjunctive therapies, such as positional therapy or weight loss interventions, may also enhance outcomes.

Accordingly, our findings indicate that OA therapy may be considered a viable alternative to CPAP when CPAP is not feasible, across the age, severity, and BMI strata examined. Nevertheless, larger prospective studies with standardized adherence assessments and extended follow-up periods are warranted to validate these results.