All patients in this study were diagnosed as having OSA by ambulatory polygraphic recording (APR) and were referred by sleep medicine specialists for oral device treatment to the Oral and Maxillofacial Department, Kuopio University Hospital (KUH). Patients were consecutively recruited from the autumn of 2016 to the end of the year 2017. Subjects were enrolled if they were 18 years or older, their AHI was at least 10 events/h, their body mass index (BMI) was less than 35 kg/m2, and they had at least 5 teeth/jaw. This study protocol was approved by the Research Ethics Committee of the Hospital District of Northern Savo, Kuopio, Finland (7 February 2017; 80/2017). All patients provided written informed consent before participating in the study. The study flowchart is presented in Fig. 1.
Five indicators of signs and symptoms of OSA (AHI, arterial oxyhemoglobin saturation (SaO2), snoring symptoms inventory (SSI), Epworth sleepiness scale (ESS), 15 D health-related quality of life (15D)) were evaluated in this study. Besides, changes in nasal resistance with the MAD in situ and upper airway volume and minimal cross-sectional areas with the MAD in situ were recorded. Data were also collected on age, gender, marital status, education, BMI, smoking, alcohol consumption, comorbidity, prescribed medication, and previous treatment of OSA.
Ambulatory polygraphic recording
Nocturnal single-night ambulatory polygraphic recordings were conducted to diagnose OSA prior to this study and after 6 months with the MAD in situ. Home APRs were conducted by using either the Embletta (Embletta, Embla, Broomfield, CO, USA) or Nox A1 ambulatory polygraphic device (Nox Medical, Reykjavík, Iceland) with accepted guidelines for diagnosing OSA. Apneas and hypopneas were automatically scored (Remlogic, version 3.2, and Somnologica, version 3.2 software, Embla Co., Broomfield, CO, USA) and manually verified and edited. Apnea was defined as a cessation (≥ 90%) of airflow for ≥ 10 s. Determination of hypopneas was based on a decrease (≥ 30%) in airflow for ≥ 10 s and associated with oxygen desaturation ≥ 4. Trained physicians evaluated the recordings. The following parameters were assessed: AHI, supine AHI, SaO2, percentage of SaO2 below 90%, and proportion of sleep time spent snoring. Positional OSA was defined as a difference of 50% or more in apnea index between supine and non-supine sleep positions . Based on the second APRs, patients were classified according to the treatment success of MAD therapy as follows: complete responders (residual AHI < 5 events/h), partial responders (residual AHI > 5 events/h, decrease in baseline AHI > 50%), and weak responders (residual AHI > 5 events/h, decrease in baseline AHI < 50%).
Daytime sleepiness and quality-of-life measurements
In order to evaluate the discomfort in life due to OSA and snoring, the subjects were asked to fulfill ESS, SSI, and 15D questionnaires at the baseline and 6 and 12 months after MAD therapy. Daytime sleepiness was assessed using ESS score (range 0–24), with a score of 11 or more being indicative of excessive daytime sleepiness . The SSI score was composed of two components: the social-work and the physical-embarrassment component including 25 symptoms (range 0–100), with a score of 0 corresponding to no problems with snoring . Health-related quality of life among patients was assessed with the 15D questionnaire . It contains 15 dimensions of life (questions), each having five different levels of functional status. These dimensions can be presented as a 15-dimensional profile or as a single index score. The score can vary between 0 and 1, a low score indicating more pathology than a high score.
Clinical data included occlusal findings (molar occlusion, overjet, overbite, crossbite, open bite, scissor bite, crowding, palatal morphology) assessed according to the modified method of Björk , examination of temporomandibular joints (TMJs), bite muscles, mandibular movements , craniofacial morphology (visual evaluation of the facial profile based on the angle formed by the soft tissue glabella, subnasale, and soft tissue pogonion, anteroposterior jaw relationships, and vertical proportions) [13, 14], and the Mallampati score . The cephalometric measurements were not included in the original study protocol. Measurements were repeated at 3, 6, and 12 months after MAD therapy. All the examinations were based on routine orthodontic clinical examinations done by the same experienced orthodontist (RP). In our previous study , the accuracy of the evaluation of the facial profile by this method was good (kappa value 0.920 with the agreement of 93.3%).
Rhinomanometry was performed 3 months after MAD therapy. An NR6-rhinomanometer (GM Instruments Ltd., Kilwinning, UK) was used to conduct the rhinomanometric recordings, in which total inspiratory nasal resistance was recorded at a radius of 150 Pa. Recordings were made in the supine position without MAD and MAD in situ. Nasal decongestion was not used.
Upper airway imaging
Images were taken about 6 months after MAD therapy on a CBCT machine with Romexis software (Promax 3D Max: Planmeca, Helsinki, Finland). CBCT imaging was done in an upright position with the Frankfort horizontal line parallel to the floor. The patients were advised to keep their head in an upright position, hold light intercuspidation, and breathe normally but not to swallow or do any conscious movements during the imaging and measurement (13 s).
The airway analysis tool was used to define the portion of interest and to calculate the volume and cross-sectional areas of each section. The sections were determined using the boundaries described elsewhere . The image orientation and selection of sensitive threshold values were conducted manually. There were two experienced oral radiologists who measured the parameters.
Mandibular advancement device therapy
Patients were treated with the SomnoDent Flex (SomnoMed Ltd., Sydney, Australia) custom-made acrylic duo block titratable oral device. In this device, the upper and lower splints are connected by adjustable interlocking acrylic buccal extensions. In order to determine the desired amount of mandibular advancement, the inter-occlusal bite was registered with the SOMGauge bite registration device to obtain 60% of the maximal protrusion of the mandible.
Statistical analyses were performed using the IBM SPSS statistics, version 22.0 (IBM Corp., Armonk, NY, USA). For the statistical analysis, partial and non-complete responders were combined. The chi-square test was used to analyze the differences in categorical variables between males and females and between complete and partial/non-complete responders. Fisher’s exact test was used when the numbers of subjects in some cells were small. The differences in continuous variables were analyzed using Student’s t test for normally distributed variables and the Mann–Whitney U test for variables with skewed distributions. Multivariate linear regression analysis was used to investigate the associations of age, gender, baseline and positional AHI, convex profile, extreme overjet, increased lower facial height, decreased palatal width, and vertically restricted throat with an increase in total airway volume. Furthermore, logistic regression analysis was used to study the association between treatment response dichotomized complete response vs. partial/non-complete response as a dependent variable and age, gender, BMI, convex profile, overjet, overbite, cross bite, increase in airway volume, and vertically restricted throat. The independent variables were added simultaneously in regression analyses and their choices were based on previous literature which suggests that they are related to sleep-disordered breathing and upper airway volume. Also, some of them contributed to the outcomes in bivariate analyses. Associations with p values of < 0.05 were considered statistically significant.
The intra- and inter-examiner reproducibility of CBCT measurements for airway volume and minimum cross-sectional areas were investigated by repeating measurements (n = 20) of those variables 2 weeks apart by two experienced oral radiologist using intra-class correlation coefficients (ICC) and its 95% confidence interval (CI). The intra-examiner repeatability of CBCT measurements varied from 0.945 (hypopharynx volume) to 0.996 (minimal cross-sectional area of the oropharynx), but it was only fair for the minimal cross-sectional area of the nasopharynx (0.387). The inter-examiner repeatability varied from 0.774 (minimal cross-sectional area of nasopharynx) to 0.995 (volume of oropharynx).