Predictors of disease alleviation with mandibular advancement devices in obstructive sleep apnea: a retrospective cohort study.

Background: Obstructive sleep apnea (OSA) can be treated with mandibular advancement devices (MAD), preventing collapse of the upper airway and decreasing apnea-hypopnea index (AHI)/ h. Disease alleviation is expected to vary depending on specific predictors including OSA-severity and cephalometric parameters. This retrospective cohort study aimed to identify predictors of disease alleviation with MAD in adult patients with mild to moderate and severe OSA. Secondary outcomes included assessing the necessity of lateral cephalograms and the therapeutic success in severe OSA-cases.

Methods: OSA-patients, treated with MAD at the orthodontic department of the University Hospital Regensburg, Germany, were allocated to mild to moderate (AHI ≤ 30/ h) and severe OSA groups (AHI > 30/ h). BMI, poly(somno)graphic, demographic and cephalometric variables were evaluated before (T0) and after 3 to 6 months of MAD-treatment (T1). Applying linear regression analyses, predictors were identified, following an assessment of their effect on disease alleviation by independent two-tailed t-tests for continuous, and absolute and relative frequencies for categorical variables. Then, the need for cephalometric analysis and the disease alleviation in severe OSA-patients were evaluated.

Results: Sixty-six predominantly male patients (mean age 55 ± 11 years; male:female = 52:14) were stratified to mild to moderate (n = 45) and severe (n = 21) OSA-groups. Regression analysis revealed baseline-AHI as a significant and relevant predictor, whereas few cephalometric parameters proved significance with small effect sizes (absolute AHI/ h-reduction, univariate model: -0.64 (95% CI: -0.75; -0.53), p < 0.001, R² = 0.666). Compared to mild to moderate OSA-cases, severe OSA-patients had a significantly higher AHI (19.1 ± 11.7 vs. 6.0 ± 4.0, p < 0.001) at T1, but also a higher disease alleviation according to absolute AHI-reduction (-26.1 ± 16.0 vs. -9.6 ± 6.4, p < 0.001), indicating comparable treatment success in all OSA-degrees.

Conclusions: Disease alleviation with MAD in adult OSA-patients can be predicted with specific poly(somno)graphic parameters (especially baseline-AHI), whereas cephalometric variables appeared inappropriate. Still, lateral cephalograms are helpful in the diagnostics and follow-up of MAD-treatment, e.g. for evaluating side effects. Providing good patient selection, therapeutic success can be achieved in both severe and mild to moderate OSA.