Mandibular Advancement vs Combined Airway and Positional Therapy for Snoring (2025)

Key Points

Question Are mandibular advancement devices more effective than airway and positional therapy for patients with primary snoring?

Findings In this pilot randomized clinical trial of 50 couples, partners reported that more people who snore who were randomized to use the mandibular advancement device experienced a clinically meaningful and statistically significant reduction in snoring and improvement in sleep compared with those randomized to combined airway and positional therapy.

Meaning The results of this randomized clinical trial indicate that mandibular advancement device and combined airway and positional therapy reduced snoring and improved sleep in people who snore and their sleeping partners, but dyads in the mandibular advancement device group experienced a larger clinically meaningful improvement than the airway and positional therapy group.

Snoring is a prevalent problem, with an estimated 150 million people who snore in the US.^1,2 Primary snoring is defined as snoring without the repetitive hypoxia that characterizes sleep apnea.^3,4 Currently available nonsurgical treatments consist of external nasal dilators (END), nasal rinses, lateral sleep positioning, and weight loss.^5,6 These treatments are targeted to either improve airflow through the nasal passage, decrease upper airway obstruction, or reduce mouth breathing during sleep.^6,7 ENDs open the nasal valve, reducing air resistance and improving flow. ENDs are associated with improved snoring in patients with obstructive sleep apnea (OSA) but not improved sleep-related metrics, such as apnea-hypopnea index scores.⁸ Nasal rinses reduce congestion, which in turn reduces mouth breathing and improves sleep quality.

The mandibular advancement device (MAD), a type of oral appliance therapy (OAT), secures the mandible anteriorly to open the airway and is a treatment for snoring and mild to moderate OSA. In 1995, the American Academy of Sleep Medicine (AASM) produced practice parameters that supported using OAT for primary snoring.⁹ Subsequent AASM practice parameter updates continued to support OAT for primary snoring.^10,11 A systematic review of 18 studies that included adults who snore with or without OSA concluded that MAD therapy could treat primary snoring.¹² However, only 1 of the 18 studies included primary snorers without OSA, and the sample size was small (N = 8) and did not provide a primary snoring definition. In 2023, the Australasian Sleep Association position statement on snoring identified MAD as first-line treatment for snoring.¹³

Snoring is an unusual clinical condition, as the person snoring may experience less bother than family members or friends. Loud snoring can be disruptive for the sleeping partner and other members of the house and can prevent people who snore from engaging in overnight trips and other social activities. Despite these consequences, to our knowledge little research has been conducted on the effect of snoring treatments on sleeping partners. One prospective cohort study stated that after continuous positive airway pressure therapy, partners have less sleep arousal and increased sleep efficiency.¹⁴ Thus, including the sleeping partner is important, and studies should be longer than 1 night.

The goal of this pilot study was to compare the efficacy of MAD vs combined airway and positional therapy (CAPT) in treating primary snoring. The results of this study may be used to inform discussions about treatment for primary snoring and improve shared decision-making.

This pilot randomized clinical trial was conducted at Washington University in St Louis, Missouri, between October 4, 2022, and July 3, 2023. The Washington University in St Louis institutional review board approved the study protocol (Supplement 1). All participants provided written informed consent. Study participants were dyads consisting of people who snore and their sleeping partners who were randomized to either MAD or CAPT. There was a 4-week pretreatment period and a 4-week treatment period. People who snore and partners completed multiple sleep questionnaires before the pretreatment period and at the conclusion of the study, except for the Clinical Global Impression of Improvement scale (CGI-I), which was only completed after treatment.

Due to the nature of the intervention, participants could not be masked. The senior author (J.F.P.) and statistician (D.K.) were masked throughout the recruitment, intervention, and data analysis phases of the trial and only unmasked after data analyses were completed.

Participants who snore and had a polysomnogram (PSG) within the previous 5 years who showed a lack of OSA (apnea-hypopnea index score of fewer than 5 events per hour of sleep) were recruited. Recruitment occurred at Washington University Department of Otolaryngology–Head and Neck Surgery clinics, the Washington University Sleep Medicine Center, and advertisements through Volunteer for Health (Washington University School of Medicine’s Research Participant Registry) (Ioerger). The electronic medical record system used by the Sleep Medicine Center to store PSGs was also used to identify potential participants.

To increase recruitment, the eligibility criteria were changed in January 2023 to allow enrollment of participants without a PSG who did snore and were deemed to not have sleep apnea based on a body mass index (calculated as weight in kilograms divided by height in meters squared) of less than 35, neck size of less than 16.5 inches, lack of uncontrolled hypertension, and no witnessed episodes of apnea. Alternatively, participants who had 2 or more PSGs more than 5 years prior, all showing lack of sleep apnea, were permitted to enroll. Exclusion criteria included a sleep apnea diagnosis, current use of hypnotic sedatives, having more than 2 alcoholic drinks per night, or being younger than 18 years.

Partners were invited to participate if they slept in the same bed or an adjacent room as the person who snored for more than 3 nights a week and were bothered by the snoring. Dyads were enrolled only if both the person who snored and partner agreed to participate. The participants were randomized according to computer-generated code and assigned using permuted block randomization in a 1:1 allocation between 2 arms: the MAD intervention and the standard therapy intervention (study statistician).

The experimental arm of the study consisted of an MAD (ProSomnusEVO) fitted by a maxillofacial prosthodontist (A.A.) at Barnes-Jewish Hospital in St Louis, Missouri. Devices were self-titrated to maximum tolerable advancement.

The comparator arm was the CAPT (external nasal dilator, nasal saline lavage with mometasone, mouth taping, and lateral positional therapy) (Figure 1). The nasal saline rinse consisted of 2-mg mometasone and a saline packet combined in a 240-mL sinus rinse bottle (NeilMed). The mixture was applied equally between both nostrils. The external nasal dilator was a standard size strip (Breathe-rite) placed horizontally over the nasal tip with both flanges attached to the ala. Mouth taping consisted of a 3 × 1–inch adhesive bandage placed vertically over the lips at the columella and did not prevent mouth breathing but instead gently brought the upper and lower lips together. The external nasal dilator and bandage were provided unless the participant chose to use a different brand of bandage or tape. For lateral positional therapy, participants were instructed to fall asleep on their side hugging a pillow, placing a pillow between their legs, or both. Participants were allowed to buy a commercial device to encourage side sleeping; however, this was not provided to improve generalizability.

In the first 4 weeks after randomization, CAPT arm participants received all parts of the therapy. Participants also practiced nasal saline lavage with mometasone to ensure adherence to saline rinse. After the 4-week pretreatment period, participants began the intervention.

The primary outcome measure was response to treatments determined from the sleeping partner rating on CGI-I for snoring. Response options included very much improved, much improved, minimally improved, no change, minimally worse, much worse, and very much worse to describe the magnitude of snoring change. The CGI-I was adapted from a previously validated measure commonly used in psychiatric studies and is a self-report scale.¹⁵ A responder to treatment was defined as a sleeping partner CGI-I rating of very much improved or much improved.

Secondary outcome measures included response to snoring treatment as reported by the person who snored. Other outcomes included the change from baseline to the end of the study in the scores of the Clinical Global Impression-Severity (CGI-S) scale, Epworth Sleepiness Scale (ESS), Pittsburgh Sleep Quality Index (PSQI), and Symptoms of Nocturnal Obstruction and Related Events (SNORE-25)^15-18 as assessed from the person who snored and their partner. Response options for the CGI-S were none, mild, moderate, severe, and problem as bad as it can be. Participants who rated the snoring as mild, moderate, severe, or problem as bad as it can be were rated as problematic snoring. The minimal clinically important difference (MCID) is the minimum change in total instrument score required for patients and physicians to consider a change as a meaningful improvement.¹⁹

To our knowledge, to date, no study has reported the rate of response to treatment in participants using MAD or CAPT. Due to lack of preliminary data, estimates of the sample size were determined based on feasibility. We targeted a final analyzable sample size of 40 dyads. To account for dropouts, we enrolled 52. A per protocol approach was used during data analysis.

Standard descriptive statistics were used to summarize baseline characteristics of study participants and distributions of the outcome measures. Histograms and the Shapiro-Wilk test were used to explore normal distribution of continuous-level variables. Median and range were used for summarizing distribution of non-normally distributed continuous variables. The precision of the effect size was defined with 95% CIs. The confidence intervals for proportion differences were estimated using the Jeffrey formula, and median differences were estimated using the Hodges-Lehman formula.

For all secondary outcomes, a change score was calculated between the end and start of study survey. All statistical analyses were performed using SPSS, version 28.0 (IBM).

In this study, most of the people who snored were self-reported White (41 [98%]) and female (26 [62%]) (Table 1). Overall, the percentage of sleeping partners who described their partner’s snoring as a problem (30 [71%]) was higher than the percentage of people who snore who described their snoring as a problem (24 [57%]) for a difference of 14 percentage points (95% CI, −6 to 35). Fifty dyads were randomized, 25 in each group (Figure 2). Of the 25 randomized to MAD, 23 (92%) completed the study and of the 25 randomized to CAPT, 19 (76%) completed the study. The percentage of people who snore who self-reported none or mild snoring severity was 18 percentage points higher (95% CI, −12 to 48) in the CAPT group (10 [52%]) than the MAD group (8 [35%]). There were no meaningful differences in distribution of baseline characteristics between the 2 groups of people who snore or partners. One person who snored in the CAPT group did not complete the end of study survey.

The percentage of sleeping partners reporting the snoring response to treatment as very much improved or much improved was higher in the MAD group (21 [91%]) than in the CAPT group (11 [58%]) for a difference of 33 percentage points (95% CI, 8-58) (Figure 3).

The median (minimum [min] to maximum [max]) pre- and posttreatment ESS score for the MAD group was 4 (0-16) and 2 (0-13), and for the CAPT group was 6 (2-14) and 4 (0-14), respectively (Table 2). The median posttreatment (min-max) difference was 1 (−1 to 4) between the MAD and CAPT groups. The change in ESS score pre- to posttreatment had a median (min-max) of 0 (−3 to 4) in the MAD group and 0 (−6 to 5) in the CAPT group. The median difference for ESS score change between groups was 1 (95% CI, −1 to 2), indicating that no difference statistically nor clinically important was observed between the 2 study groups.

The median (min-max) of the pre- and posttreatment PSQI score for the MAD group was 6 (2-14) and 4 (1-12), respectively. For the CAPT group, pre- and posttreatment PSQI scores were 6 (3-18) and 6 (1-17), respectively. The median (min-max) of pre- to posttreatment PSQI score change was 3 (−1 to 11) for the MAD group and 0 (−4 to 5) for the CAPT group for a median difference of −3 (95% CI, −6 to −1) favoring the MAD group.

The median (min-max) pre- and posttreatment SNORE-25 scores for the MAD group were 1 (0.08–2.16) and 0.2 (0.04-2.04), respectively, and for the CAPT were 1.04 (0.20-2.16) and 0.88 (0.04-2.48), respectively. The median (min-max) of pre- to posttreatment SNORE-25 score change was 0.36 (−0.72 to 1.76) for the MAD group and 0.16 (−1.48 to 1.72) for the CAPT group for a median difference of −0.44 (95% CI, −0.80 to 0) favoring the MAD group. The range of values plausible for differences in the secondary outcome measures in the population displayed by the confidence intervals indicated that clinically important differences were not likely to be observed between the 2 treatments.

The percentage of people who snore who self-reported response to treatment was higher in the MAD group (21 [91%]) than the CAPT group (8 [44%]) for a difference of 47 percentage points (95% CI, 21-73). Most of the partner-defined responders to treatment (20 of 21) in the MAD group were also identified as responders by the people who snore themselves. In the CAPT group, 6 of 11 people who snore (55%) agreed with their partners.

Median (min-max) pre- and posttreatment ESS scores for the MAD group were 4 (0-16) and 2 (0-13), respectively, and for the CAPT group were 6 (2-14) and 4 (0-14), respectively. The change in ESS score pre- to posttreatment had a median (min-max) of 1 (−3 to 6) for the MAD group and 0.5 (−9 to 9) in the CAPT group for a median difference of 0 (95% CI, −2 to 1).

Median (min-max) pre- and post-treatment PSQI scores for the MAD group were 6 (1-17) and 2 (1-10), respectively, and for the CAPT group were 7 (3-16) and 4 (1-10), respectively. The median (min-max) change pre- to posttreatment in PSQI scores for the MAD group was 2 (0-7) and for the CAPT group was 1.5 (−3 to 11) for a median difference of 0 (95% CI, −2 to 1).

Median (min-max) pre- and posttreatment SNORE-25 scores for the MAD group were 0.48 (0.08-1.8) and 0.2 (0.04-1.60), respectively, and for the CAPT group were 1.2 (0.28-2.08) and 0.54 (0.12-1.84), respectively. The median (min-max) change pre- to posttreatment in SNORE-25 scores for the MAD group was 0.28 (−0.16 to 1.56) and for the CAPT group was 0.14 (−0.52 to 1.36) for a median difference of −0.04 (95% CI, −0.24 to 0.24).

The range of values in the 95% CIs displayed that the plausible range of values for the differences in secondary outcomes in the population as reported by the people who snore was neither statistically nor clinically important.

Of the 22 participants to MAD, 10 (45%) reported self-resolving jaw pain/discomfort for a few days after beginning the device. Of the 19 CAPT participants, 5 (26%) experienced a brief sinus headache on initiating treatment with the nasal saline rinse. This was mild and persisted for 1 to 3 days before resolving without discontinuation. One participant experienced serious headaches from the saline rinse that required withdrawal. No serious adverse events were reported.

In this pilot randomized clinical trial, we evaluated the efficacy of 2 different treatments for snoring: an MAD and CAPT. After 4 weeks of the intervention, the sleeping partners in the MAD group reported clinically meaningful and statistically significant improvements in snoring; 91% of dyads randomized to MAD reported the snoring responded to treatment compared with 58% of the CAPT group. Based on the self-reported assessment of people who snore, 91% randomized to MAD reported that their snoring responded to treatment compared with 44% randomized to CAPT. While both therapies are safe and effective for snoring and either could be considered for treatment, most people who snore randomized to MAD experienced improvement of snoring. We believe these results are applicable for men and women as the study population was roughly equal between the 2 sexes.

Our results differ from prior literature. One study by Cooke and Battagel²⁰ showed a 64% improvement in snoring with MAD compared with the 91% from this study. Our study and the Cooke and Battagel²⁰ study followed up participants for similar periods: 4 weeks in this study vs 4 to 6 weeks. The higher rate of response observed in this study could be due to differences in the devices. In the Cooke and Battagel²⁰ study, a TheraSnore device was used compared with the Prosomnus EVO device used in this study. The TheraSnore device is a boil-and-bite device that includes 2 arches with limited adjustable options. Although it allows for protrusive titrations, it does not allow freedom of movement during wear. The Prosomnus EVO device is a custom fitted and milled oral device, making for a tighter and more comfortable fit. Our study provides further evidence for the use of MAD in patients with primary snoring, as recommended by the AASM^9-11 and Australasian Sleep Association Position Statement,¹³ which recommended that oral appliances should be considered for treating primary snoring without features of OSA or upper airway resistance syndrome.

The various sleep-related patient-reported outcome measures improved for partners and people who snore before and after treatment regardless of treatment assignment. For people who snore, frequent awakenings from the sleeping partner (eg, elbowing to stop snoring) or from interrupted sleep quality, which does not qualify sufficiently as sleep apnea, could have decreased. Likewise, improvement in sleep quality for the sleeping partner in both intervention arms was also observed. Anecdotes from participants in the study indicate that snoring substantially affects the quality of life and relationships between people who snore and partners and that relief from snoring dramatically improves their relationship.

There were several limitations of the study. First, the dyads were not masked to the intervention. The 2 interventions were quite different. The MAD is a sophisticated oral appliance that requires a follow-up visit with the maxillofacial prosthodontist to ensure correct fitting, while the CAPT is a simple and inexpensive intervention. Since the outcome measures used in this study were patient-reported measures, this inability to mask could have biased the person who snored and the sleeping partner’s evaluation of the intervention. A future trial could use video recordings and assessment of snoring by an independent rater masked to the intervention. Second, the analysis of results was done with a per-protocol analysis and not intention to treat. The per-protocol analysis was used because assessment of the outcome was only obtained from the participants who completed the study and not fore all the participants enrolled and randomized in this study. Third, the short 4-week duration of follow-up prevents detection of long-term adverse effects (eg, tooth movement with MAD, skin irritation with a nightly nasal dilator strip). Fourth, 30% of the participants did not have PSGs. Finally, the majority of study participants were White, limiting the generalizability of the results.

The cost difference between the 2 therapies is substantial. An MAD as described in this study can cost up to $3000 out of pocket. Often insurance will not cover this device for primary snoring. Conversely, the CAPT costs around $20 to $30 a month for all components. As such, there should be a patient-centered conversation between clinicians and patients on treatment of snoring in the absence of sleep apnea. A reasonable approach would be to start with CAPT and consider MAD if snoring persists or is bothersome.

Snoring affects the sleep quality and quantity for the person who snores and their sleeping partner. This pilot randomized clinical trial found that MAD had a higher rate of efficacy compared with CAPT. Short-term use of both treatments showed that they are well tolerated by people who snore and safe therapies to treat snoring. The cost of the MAD may be prohibitive to some patients; thus, CAPT is a reasonable, less expensive alternative. Individuals with primary snoring and their physicians should use the principles of shared decision-making to determine the management most appropriate for their situation.

Accepted for Publication: March 26, 2024.

Published Online: May 23, 2024. doi:10.1001/jamaoto.2024.1035

Corresponding Author: Jay F. Piccirillo, MD, Department of Otolaryngology–Head and Neck Surgery, Washington University School of Medicine in St Louis, 660 South Euclid Ave, Campus Box 8115, St Louis, MO 63110 ([emailprotected]).

Author Contributions: Drs Kallogjeri and Piccirillo had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Ioerger, Afshari, Hentati, Ju, Piccirillo.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Ioerger, Afshari, Piccirillo.

Critical review of the manuscript for important intellectual content: Ioerger, Hentati, Strober, Kallogjeri, Ju, Piccirillo.

Statistical analysis: Ioerger, Kallogjeri, Piccirillo.

Obtained funding: Piccirillo.

Administrative, technical, or material support: Ioerger, Afshari, Hentati, Strober, Ju, Piccirillo.

Supervision: Ioerger, Afshari, Piccirillo.

Conflict of Interest Disclosures: Dr Kallogjeri reported personal fees from JAMA Otolaryngology–Head & Neck Surgery outside the submitted work. Dr Ju reported a research agreement with GE outside the submitted work. Dr Piccirillo reported grants from the National Institute of Health during the conduct of the study as well as royalty payments from Washington University, compensation for serving as the Editor of JAMA Otolaryngology–Head & Neck Surgery, and honorarium for serving on the external advisory boards of the Georgia Clinical & Translational Science Alliance and Frontiers Clinical and Translational Science Institute, serving as chair and on the data and safety study monitoring board for the Endotypic Traits and Obstructive Sleep Apnea Surgery study, and for serving as a scientific consultant and cochair of the study advisory committee of the Nasal Steroids, Irrigation, Oral Antibiotics and Subgroup Targeting for Effective Management of Sinusitis study. No other disclosures were reported.

Funding/Support: Research reported in this publication was supported by the National Center For Advancing Translational Sciences of the National Institutes of Health (TL11TR00234; Dr Ioerger) and the National Institutes of Deafness and Other Communication Disorders (T32DC000022; Drs Hentati and Strober).

Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: We thank third author Dr Hentati in Figure 1 for granting permission to publish this information.

Disclaimer: Dr Piccirillo is Editor in Chief and Dr Kallogjeri Statistical Editor of JAMA Otolaryngology–Head & Neck Surgery but neither was involved in any of the decisions regarding review of the manuscript or its acceptance.

Additional Information: The full trial protocol can be accessed by contacting the corresponding author.