The role of functional respiratory imaging in the prediction of treatment outcome with fixed mandibular advancement in OSA patients

Abstract

Introduction: Previous work by our group showed that the severity of OSA correlates with the morphology of the patient9s upper airway (UA) lumen and skull. Treatment response is affected by gender, baseline OSA severity and UA shape and morphologic changes do correlate with changes in OSA severity. In spite of such correlations no single parameter was reported that is able to predict the outcome of a given treatment. This work aims to develop and test a binomial linear model to predict treatment response to mandibular advancement (MA) in a large population of OSA patients. Material and Methods: 77 OSA patients were (83%M; 47.4±11.5years; AHI=21.0±11.2) treated with MA (protrusion=75% of the maximal MA). The following data was used for the analysis: from PSG: OSA severity (AHI-ODI) from Functional Respiratory Imaging: UA collapsibility (without or with MA), change in UA volume and resistance, movement of mandible and hyoid bone Treatment was considered positive if ΔAHI≥50%, or if post-MA AHI A binomial generalized linear (bgl) model was made to predict response by aforementioned parameters. Fitted response values ≥ 0.5 were considered to result in a positive response. Results: The success rate in terms of AHI of the MA treatment was 47%. The fitted bgl model was significantly (p Conclusion: A combination of OSA severity, gender, UA collapsibility, and the lumen and skeletal response to mandibular advancement can be used to predict the outcome of MA treatment with an accuracy above 80%.