Comparative efficacy of mandibular advancement devices in obstructive sleep apnea: a network meta-analysis.

PURPOSE: To analyze relative efficacies of mandibular advancement devices (MAD) in sleep apnea treatment. METHODS: From eligible randomized controlled trials (RCT), MADs were classified based on their mechanistic designs. Data on apnea-hypopnea index (AHI), Epworth sleepiness scale (ESS), nadir oxygen saturation (minSaO2), and sleep efficiency (SE%) from RCTs were then analyzed in network meta-analyses, and relative ranking of different MADs was computed based on P scores (a method of ranking similar to SUCRA). Similar analyses were conducted based on the different brands of MADs. RESULTS: There were no statistically significant differences between MADs in any of the outcomes analyzed. However, the P-scores, based on the point estimates and standard errors of the network estimates, ranked some MADs higher than others in some of the outcomes. Of the different mechanistic designs, the highest P scores were achieved for attached midline traction (P score = 0.84) and unattached bilateral interlocking (P score = 0.78) devices for AHI reduction, attached bilateral traction (P score = 0.78) and unattached bilateral interlocking (P score = 0.76) for ESS, monobloc (P score = 0.91) and unattached bilateral interlocking (P score = 0.64) for minSaO2, and unattached bilateral interlocking (P score = 0.82) and attached bilateral traction (P score = 0.77) for SE%. Notable findings in the network meta-analyses based on MAD brands, of the limited number of studies that specified them were the effects of SomnoDent Flex™, TAP™, and IST® in their effects on AHI reduction, with P scores of 0.94, 0.83, and 0.82, respectively. Monobloc decreased supine-AHI the most (- 44.46 [- 62.55; - 26.36], P score = 0.99), and unattached bilateral interlocking had the greatest effect on REM-AHI (- 11.10 [- 17.10; - 5.10], P score = 0.87). CONCLUSIONS: Findings from this study show clinically (but not statistically) significant differences between MADs in terms of their relative efficacy when analyzed for different sleep apnea treatment outcomes and sleep apnea phenotypes.