A stepwise titration protocol for oral appliance therapy in positional obstructive sleep apnea patients: proof of concept

This study is part of a randomized controlled trial in which OAT was compared with a position trainer in patients with POSA [12, 13]. Participants were eligible for enrollment if they had mild or moderate POSA, were 18 years or above, and were able to provide informed consent. Exclusion criteria were inadequate dentition for wearing an OA, subjective snoring in lateral position, diagnosed with central sleep apnea, night or rotating shift work, severe chronic heart disease, active psychiatric disease, seizure disorder, medication usage for sleeping disorders, muscular or joint injuries in head/neck or back area, previous OAT usage, simultaneous other treatments for OSA, reversible morphological upper airway abnormalities (e.g., enlarged tonsils), pregnancy, and coexisting non-respiratory sleep disorders (e.g., insomnia, periodic limb movement disorder, narcolepsy) that would compromise functional sleep assessment.

This study is a sub-assessment of the OAT therapy group in the POSA trial [12, 13], focusing on the standardized stepwise titration protocol used. The POSA trial is a multicenter, prospective randomized controlled trial. The randomization sequence was generated by an independent clinical research unit. Allocation of treatment of 1:1 was performed with random block sizes of maximum 6 and stratified for smoking and body mass index (BMI). The protocol of this study was approved by the medical ethics committee (Amsterdam UMC: METC2012_208) and was registered before the start in clinicaltrials.​gov (NCT02045576). All participants provided written informed consent before enrollment. Independent monitors performed verification of the source data and documentation. The study investigators had full access to the data and had the right to submit the manuscript for publication without input from the sponsor.

The OA was a custom-made duobloc (SomnoDent flex, SomnoMed, Sidney, Australia). The OA was adjusted individually, and advancement was titrated using a standardized stepwise titration protocol, developed by one of the authors (GA). After adequate assessment of the central relation and maximum protrusion using the George Gauge system with a standard 5-mm vertical dimension (Great Lakes Orthodontics, Tonawanda, NY), the OA was set at 60% advancement of maximum protrusion at baseline. During the first 3 months, at each consecutive visit, the OA was evaluated and advanced to 75% or 90% if subjective improvement (e.g., perceived reduction of snoring or apneic events) of OSA was not reached. On the other hand, if side effects were not acceptable for the patient (e.g., tooth pain or signs of temporomandibular disorders), the advancement was adjusted backward to 75%, 60%, or 45%. No adjustments were made when the patient reported a sufficient efficacy without side effects. The patient returned to clinical practice at 6, 10, and 14 weeks after placement of the OA for this standardized titration protocol. After the titration procedure was completed for each patient, they only returned when further adjustments were necessary, based on subjective impairments, viz., recurrent snoring or increasing excessive daytime sleepiness. The vertical dimension was not controlled in our patients using frontal elastics. Objective compliance was measured using a temperature-sensitive microsensor with on-chip integrated read-out electronics (Theramon, Handels- und Entwicklungsgeselschaft, Handelsagentur Gschladt, Hargelsberg, Austria). The temperature was recorded at a sampling rate of 1 measurement per 15 min, allowing data acquisition on usage for a consecutive 100-day period. A recorded temperature of > 30 °C indicated that the OA was worn. This microsensor was embedded in the OA at the lower right side. Data was extracted at 3 months (± 2 weeks) using a dedicated reading station. Missing results in the adherence data are due to problems with the reading station of the microsensors. Our titration protocol is enclosed as supplementary material 1.

The titration protocol was analyzed using different outcome measures. Polysomnographic response is presented by the AHI and ODI, and these parameters were used for assessment of therapy failure/success for OAT. Other measures were defined as patient’s adherence to OA treatment and self-reported side effects or discomfort resulting in termination of treatment. Adherence was defined as the percentage of daily use of ≥ 4 h per night, during ≥ 5 nights per week. Adherence failure was defined as an inability of the patient to continue treatment for any reason mentioned by the patient. These outcomes were recorded during the adaptation period for each new position of the mandible. Adverse events were reported in accordance with the International Conference of Harmonization ICH E2A guidelines (Good Clinical Practices) by the principal investigators and evaluated by clinical data monitors.

Data were assessed on normality, both graphically using histogram plots and by the Shapiro-Wilk test and were analyzed and expressed as median (interquartile range) or mean ± SD for descriptive purposes. The presented variables were tested for differences using the Fisher exact test for categorical variables and the Wilcoxon signed-rank test for continuous variables with a non-normal distribution. When comparing groups between specific titration positions, the Kruskal-Wallis test was used in case of a non-normal distribution as an alternative for the ANOVA for normally distributed variables. Associations between continuous variables were described using Spearman’s Rho correlation. A P value < 0.05 was considered statistically significant. The test used for the statistical analysis is mentioned in the manuscript when a P value is provided (T, t test; W, Wilcoxon signed-rank test; and H, Kruskal-Wallis test).

All patients completed the titration protocol in the first 3 months. None of the patients ended treatment in the first 3 months because of discomfort or side effects due to OAT. Two patients encountered severe problems with OAT, for which titration was set back to 45% in the beginning, but at the evaluation period at 3 months with PSG, the titration was set again at 60%, after symptoms improved. One of these two patients had no decrease of AHI together with severe discomfort while wearing OA and terminated the study directly after the 3-month evaluation. The other patient could be motivated to continue therapy and eventually ended the study at 12 months successfully on a titration of 75% with good efficacy. After the 3-month evaluation with PSG, two patients terminated OAT because of severe discomfort in wearing the appliance, and three patients ended the study preliminary because of lack in efficacy with no desire to continue the protocol (no decrease AHI and persistent excessive daytime sleepiness). These patients did not report any discomfort. In total, six patients (17%) terminated OAT because of treatment-related reasons. One patient was lost to follow-up at 9 months because of emigration.

Median adherence (> 4 h, 5 nights a week) was 97.4% (61.4–100.00) after 3 months of OAT (N = 33) and 100.0% (90.0–100.0) after 12 months (N = 23). No differences were seen after 3 months between the different advancement levels (H, 2.567; P = 0.109). After 9 and 12 months of OA use, the median adherence was excellent (100%) in the 75% and the 90% titration groups. The few patients (n = 2 and n = 3) in the 60% titration group after 9 and 12 months showed lower adherence when compared with the 75% and the 90% groups (H, 1.122; P = 0.571) (Fig. 2 and Table 4).