Background <p>Many patients have positional obstructive sleep apnoea (OSA) with an apnoea-hypopnoea index (AHI) that is two-fold higher in supine sleep than in non-supine sleep. In these patients, the proportion of time spent in supine sleep influences the overall AHI. With the example of OSA patients treated with a mandibular advancement device (MAD), the aim of this study was to test the hypothesis that between-measurement differences in the proportion of supine sleep time are a significant factor affecting changes in AHI.</p> Methods <p>One hundred sixty-five adult OSA patients treated with MAD were included in the study. Data on AHI, supine sleep time, age, sex, body mass index (BMI), Epworth Sleepiness Scale (ESS) scores, mandibular protrusion, patient-reported use, and adverse effects of the MAD were retrospectively collected from medical records or the Swedish Sleep Apnoea Register.</p> Results <p>Among included patients, 27.3% (45/165) had both positional OSA at baseline and a ≥ 50% difference in the proportions of supine sleep time between baseline and follow-up. A generalized linear model showed that changes in the proportion of supine sleep time had a statistically significant impact on the change in overall AHI from baseline to follow-up of similar size as the effect the MAD.</p> Conclusions <p>Changes in the proportion of supine sleep time are an important contributor to between -measurements changes in overall AHI in many patients treated with MAD. Positional OSA must be acknowledged not only in OSA diagnostics but also in MAD treatment follow-up.</p>