Key Points
Question
What are the upper airway anatomical factors determining the outcome of mouth closure on airflow in patients with obstructive sleep apnea (OSA)?
Findings
This nonrandomized clinical trial of 54 patients with OSA demonstrated that mouth closure improves airflow during sleep in patients with OSA, but the outcomes were heterogeneous. In patients who breathe through their mouth and have severe obstruction at the soft palate, airflow worsens with mouth closure.
Meaning
These results suggest that personalized approaches to recommending therapies aimed at preventing mouth breathing should be considered.
Abstract
Importance
Mouth breathing is associated with increased airway resistance, pharyngeal collapsibility, and obstructive sleep apnea (OSA) severity. The common belief is that closing the mouth can mitigate the negative effects of mouth breathing during sleep. However, mouth breathing may serve as an essential route to bypassing obstruction along the nasal route (eg, the velopharynx).
Objective
To investigate the role of mouth breathing as an essential route in some patients with OSA and its association with upper airway anatomical factors.
Design, Setting, and Participants
This nonrandomized clinical trial included participants diagnosed with OSA who underwent drug-induced sleep endoscopy. Patients were stratified into 3 quantiles based on oral-breathing level (quantile 1: oral airflow < 0.05 L/min; quantile 2: oral airflow 0.05-2.2 L/min; quantile 3: oral airflow > 2.2 L/min).
Interventions
Closing the mouth during sleep during alternating breaths by applying pressure to the mentum until teeth are in occlusion.
Main Outcomes and Measures
The primary outcome was total inspiratory flow defined as the change in airflow in the transition from mouth relaxed to mouth closed, analyzed overall and by 3 oral-breathing quantiles. The association of velopharyngeal obstruction on the change in total inspiratory airflow was also investigated.
Results
Of 66 enrolled patients with OSA, 12 were excluded due to insufficient baseline airflow. The analytic cohort consisted of 54 patients (39 [72%] male; median [IQR] age, 55 [46-64] years; apnea-hypopnea index, 26.9 [17.6-39.9] events/h; and body mass index calculated as weight in kilograms divided by height in meters squared, 28.9 [27.1-31.6]). Mouth closure increased total inspiratory flow by 27.8 percentage points overall (β, 1.0 [95% CI, 0.4-1.9] L/min). However, outcomes varied based on the degree of baseline oral breathing. No association was found for 10 patients with near-zero mouth breathing (0.9 [95% CI, −0.2 to 2.1] L/min). Airflow improved with mouth closure in 32 patients with moderate levels of mouth breathing (2.0 [95% CI, 1.3-2.7] L/min), whereas it worsened in patients with high levels of mouth breathing (−1.9 [95% CI, −3.1 to −0.6] L/min). Velopharyngeal obstruction was associated with increased mouth breathing (0.6 [95% CI, 0.1-3.0] L/min) and reduced airflow with mouth closure (−1.9 [95% CI, −3.1 to −0.7] L/min).
Conclusion and Relevance
Although mouth closure increased inspiratory airflow in the overall cohort of this nonrandomized clinical trial, the outcomes were heterogeneous. In patients who breathe primarily through their mouth during sleep and have velopharyngeal obstruction, airflow worsens with mouth closure. Hence, personalized approaches to treating mouth breathing should be considered.
Trial Registration
ClinicalTrials.gov Identifier: NCT06547658
This nonrandomized clinical trial evaluates the overall association of mouth closure with inspiratory airflow and the change in inspiratory airflow with mouth closure across 3 oral-breathing quantiles.
Introduction
Mouth breathing is considered an abnormal form of breathing associated with increased airway resistance, pharyngeal collapsibility, and obstructive sleep apnea (OSA) severity.1 This was notably described by Fitzpatrick et al,2,3 who measured nasal and mouth breathing separately in healthy participants. They concluded that during mouth breathing, total upper airway resistance was 2.5 times higher than when nasal breathing and increased the mean (SD) apnea-hypopnea index (AHI) from 1.5 (0.5) to 43.0 (6.0) episodes per hour. Suzuki et al4 followed up with a study of patients with OSA who underwent computed tomography scans during nasal and oral breathing while awake. Using computational fluid dynamics, they found that, compared to nasal breathing, oral breathing was associated with higher resistance to airflow. These findings promote the general belief that preventing mouth breathing during sleep is beneficial to sleep and breathing. This has been a particularly popular concept in the media in recent years.
These studies are of limited application to individuals who breathe through their mouths spontaneously in response to obstruction along the nasopalatal route. In this case, mouth breathing becomes the essential route that bypasses the obstruction and effectively lowers upper airway resistance. We hypothesized that preventing mouth breathing in these patients would likely increase upper airway resistance and be detrimental to breathing during sleep; however, this has not been formally studied. As such, there is a clinical need to characterize the mechanism by which patients breathe through their mouth to bypass palatal obstruction, and whether closing the mouth in these patients is detrimental to breathing during sleep.
This nonrandomized clinical trial addresses this need by investigating (1) the overall outcomes of mouth closure on inspiratory airflow in patients with OSA; (2) whether the outcomes are modified by the magnitude of mouth breathing as an essential route; and (3) whether outcomes vary based on upper airway anatomic factors. We hypothesized that eliminating mouth breathing would have heterogeneous outcomes, depending on the amount of baseline mouth breathing and upstream palatal obstruction. To test this hypothesis, we used a partitioned oronasal mask to measure the nasal and oral airflow separately during drug-induced sleep endoscopy (DISE). DISE was selected as the human model of pharyngeal collapse during sleep due to the ability to manipulate mouth opening acutely with minimal concern for confounding due to arousals.
Methods
Study Design and Participants
This nonrandomized clinical trial was conducted at the Massachusetts Eye and Ear Infirmary in Boston, Massachusetts, between December 17, 2021, and July 10, 2022. The Brigham and Women's Hospital institutional review board approved the study protocol (Supplement 1). All participants provided written informed consent before undergoing DISE.
A total of 120 patients with diagnosed OSA scheduled to undergo a DISE were enrolled in the study. Exclusion criteria included pregnancy, age younger than 18 years, poor general health, allergy to propofol or dexmedetomidine, or history of surgical treatment for sleep apnea, such as palate, tongue base, or epiglottis surgery. Among the 120 patients, we could not perform the mouth closure maneuver without arousing the patient from sleep in some patients, so that 66 patients then remained. The baseline characteristics of all 120 patients are summarized in eTable 1 in Supplement 2. Among the 66 remaining patients, 12 patients without sufficient airflow (inspiratory airflow < 0.01 L/s) under baseline conditions were excluded from the analysis so that the intervention could only demonstrate a beneficial outcome or no response in these patients but not a deleterious outcome. As such, data from 54 patients were analyzed.
Trial Protocol
All patients underwent DISE with simultaneous airflow measurements separately through the nose and mouth using a tight-fitting partitioned oronasal mask and 2 pneumotachometers. All patients were studied while supine. To evaluate the outcome of eliminating mouth breathing on airflow, we performed a mouth closure maneuver that we found to significantly limit, if not eliminate, mouth breathing while preventing arousal from sleep. The maneuver involved applying pressure to the mentum until the teeth were in occlusion, without altering the head position. The maneuver was applied on alternating breaths with the goal of minimizing the outcome of ventilatory effort (ie, drive) on airflow, based on the assumption that effort changes between adjacent breaths will be minimal, similar to an approach used to study the acute outcome of upper airway stimulation on ventilation.5 Additionally, it facilitates the detection of arousal caused by mouth closure because the pattern of airflow changes from arousals are distinct from the alternating mouth closure maneuver. Figure 1 illustrates the changes in airflow with alternating mouth closure. In contrast, arousal caused by mouth closure will result in a sustained increase in airflow, consistent with a ventilatory response to arousal. If mouth closure caused arousal, the maneuver was abandoned until flow-limited breathing resumed. Lastly, the mouth closure maneuver was performed during periods of hypopnea or stable flow-limited breathing, avoiding periods of apnea. The senior author (D.V.) performed the mouth closure maneuver on all participants.
Figure 1. Example Airflow Traces Illustrating Differences in Airflow Response to Mouth Closure.
A, A negative responder (quantile 3: oral airflow > 2.2 L/min) whose airflow deteriorates on mouth closure. B, A responder (quantile 2: oral airflow 0.05-2.2 L/min) whose airflow improved on mouth closure.
Drug-Induced Sleep Endoscopy
DISE was performed in the operating room as part of routine clinical care. Patient vital signs were monitored, and the targeted bispectral index (Medtronic) values were 50 to 70. Propofol was used as the sedative agent with a starting dose of 200 µg/kg/min, increasing by 25 µg/kg/min until the bispectral index level reached 50 to 70 and clinical signs of obstructive breathing developed (snoring, paradoxical breathing, mean SpO2 percentage in the mid-90s with intermittent desaturations and resaturation).
Site of Upper Airway Collapse
The site of pharyngeal collapse during DISE was evaluated using the established VOTE classification, which classifies the anatomic sites as velum, oropharynx, tongue base, and epiglottis.6 The degree of obstruction is scored at inspiration, in which 0% to 50% obstruction compared to the awake status was scored as no obstruction; 50% to 75%, partial obstruction; and 75% or more, complete obstruction. We also labeled 2 additional features of the airway that in preliminary studies were found to be an important determinant of airflow response to mouth closure. First, we differentiate velum concentric collapses (VC) into VC with dominant lateral salpingopharyngeus muscle component (VCL) and concentric collapse without lateral salpingopharyngeal muscle enlargement (VCC).7 Second, we identified patients who exhibited consistent palatal prolapse (PP),8 wherein the palate would prolapse or balloon into the nasopharynx during expiration, thereby blocking expiration out of the nose and causing expiration through the mouth. We considered complete obstruction only as an anatomical risk factor. Epiglottic collapse was only scored when more than 50% of obstructed breaths exhibited epiglottic collapse, based on our previous study.9 Lastly, we defined obstruction upstream from the oral route as the presence of any of the velopharyngeal collapse categories (VAP, VCC, and VCL).
Breath-by-Breath Airflow
To evaluate the airflow of individual breaths, the start and stop times of breaths were detected automatically and manually adjusted as necessary. Breath-by-breath inspiratory airflow was calculated as the mean airflow within the inspiratory period and expressed in units of L/min. Airflow is also presented as a percentage of the local mean ventilation (eupneic ventilation, expressed in units of %eupnea), calculated from all breaths recorded during the DISE study when the maneuver was not performed. Total, oral, and nasal inspiratory airflow, as well as the oral fraction (oral tidal volume divided by total tidal volume), were calculated for each breath. Individual breaths were manually annotated as “mouth closed,” if the mouth closure maneuver had been applied, or “relaxed,” if it had not. We considered airflow during the relaxed position as the patient’s baseline breathing pattern. Participant-level averages for all airflow metrics during baseline and mouth-closed positions, as well as the differences in airflow between mouth positions, were computed for statistical analysis.
Statistical Analysis
To evaluate differences in patient characteristics and sites of collapse categories between included and excluded patients, bivariate linear or logistic regression assessed the association of exclusion (binary independent variable) with each characteristic (ie, age, sex, body mass index, and AHI) and site of collapse category.
Primary evaluation of the outcome of mouth closure on total inspiratory flow was performed using mixed-model regression with total inspiratory flow as the continuous dependent outcome variable, the mouth closure intervention as a fixed effect, and participant as a random effect. Flow was logarithmically transformed to achieve normally distributed model residuals, and results were back-transformed for presentation. Additional analyses examined the associations of the intervention with oral and nasal airflow separately.
To assess whether the intervention is less efficacious at raising total inspiratory flow in patients with higher levels of baseline mouth breathing, we modeled the absolute change in total inspiratory flow (intervention − baseline) as a function of baseline oral breathing. Preliminary analysis revealed a nonlinear association, such that our primary analysis used quadratic regression (Δflow ≈ ax2 + bx + c, where x denotes baseline oral breathing in L/min). Simplified secondary analysis described the change in total inspiratory flow within 3 quantiles of roughly equal sample sizes based on oral breathing. Quantile 1 was defined as oral airflow of less than 0.05 L/min; quantile 2, oral airflow between 0.05 and 2.2 L/min; and quantile 3, oral airflow greater than 2.2 L/min. Adjusted analysis for age is presented in eAppendix 3 and eTable 5 in Supplement 2.
To assess whether the intervention is less efficacious at raising total inspiratory flow in patients with obstruction upstream from the oral route, we modeled Δflow as a function of upstream collapse (any of the following: VAP, VCC, or VCL, as single dichotomous independent variables) using linear regression. A similar descriptive model examined mouth breathing (oral airflow, as a continuous variable in L/min) as a function of upstream collapse. To facilitate clinical interpretation, we examined the potential value of upstream collapse to determine its association with potentially beneficial outcomes of the mouth closure intervention, defined as any increase in inspiratory airflow greater than 0.0 L/min. Finally, exploratory analysis evaluated the outcome of the individual site of collapse (VAP, VCC, and VCL; oropharyngeal collapse, tongue collapse, epiglottic collapse) on ∆flow (as well as on baseline mouth breathing per oral airflow). Adjusted analysis for age is presented in eAppendix 3 and eTable 5 in Supplement 2. Results from statistical analyses on transformed variables were back-transformed for interpretation. Statistical analyses were performed using MATLAB software, version 2023b (MathWorks).
Results
Participant Characteristics
Of 66 enrolled patients with OSA, 12 were excluded due to insufficient baseline airflow. The analytic cohort consisted of 54 patients (39 [72%] male; median [IQR] age, 55 [46-64] years; apnea-hypopnea index, 26.9 [17.6-39.9] events/h; and body mass index calculated as weight in kilograms divided by height in meters squared, 28.9 [27.1-31.6]; Table 1). Median (IQR) eupneic ventilation outside the experimental period was 4.2 L/min (3.0-5.1 L/min). The total number of breaths analyzed from included patients was 678 breaths (mean [SD] per patient, 12.6 [4.4] breaths).
Table 1. Characteristics of the Study Population.
| Characteristic | No. (%) | Effect size of included − excluded (95% CI)a | ||
|---|---|---|---|---|
| Study population (N = 66) | Included (n = 54) | Excluded (n = 12) | ||
| Age, median (IQR), y | 55 (46 to 64) | 55 (46 to 64) | 55 (44 to 62) | –2.9 (–11.1 to 5.4) |
| Sex | ||||
| Female | 18 (27.3) | 15 (27.8) | 3 (25) | 0.9 (0.2 to 3.7) |
| Male | 48 (72.7) | 39 (72.2) | 9 (75) | |
| BMI, median (IQR) | 29.0 (26.7 to 31.8) | 28.9 (27.1 to 31.6) | 28.3 (27.4 to 29.3) | –0.5 (–2.7 to 1.7) |
| AHI, median (IQR), events/h | 27.5 (18.6 to 41.7) | 26.9 (17.6 to 39.9) | 35.9 (22.4 to 58.2) | 9.8 (–3.7 to 28.9) |
| Anatomic site of collapseb | ||||
| Rostral | 28 (42.4) | 23 (42.6) | 5 (41.7) | 1.0 (0.3 to 3.5) |
| VCC | 3 (4.5) | 2 (3.7) | 1 (8.3) | 2.4 (0.2 to 29.8) |
| VAP or PP | 19 (28.8) | 15 (27.8) | 4 (33.3) | 1.3 (0.3 to 5.1) |
| VCL | 6 (9.1) | 6 (11.1) | 0 | NA |
| Oropharyngeal | 12 (18.2) | 11 (20.4) | 1 (8.3) | 0.4 (0.04 to 3.2) |
| Tongue | 18 (27.3) | 12 (22.2) | 6 (50) | 3.5 (0.9 to 13.2) |
| Epiglottis | 7 (10.6) | 5 (9.3) | 2 (16.7) | 2.0 (0.3 to 12.0) |
Abbreviations: AHI, apnea-hypopnea index; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NA, not applicable; PP, palatal prolapse; VAP, anteroposteriorly direct velum collapse; VCC, concentric velum collapse; VCL, concentric velum collapse with dominant lateral salpingopharyngeus muscle component.
Effect size of differences between included and excluded patients were determined from regression analyses and are represented by the β (95% CI) coefficient for continuous variables (eg, age, sex, BMI, and AHI) or odds ratio (95% CI) for binary variables (eg, sex and anatomic factors).
Individual patients could have belong to more than 1 category of anatomic site of collapse.
Overall Inspiratory Airflow Outcome
As hypothesized, the mouth closure maneuver had overall beneficial outcomes (eTable 2 in Supplement 2), increasing total inspiratory flow (β, 1.0 [95% CI, 0.4-1.9] L/min; Δ%eupnea, 27.8 pp [95% CI, 11.1-48.1 pp]). This overall increase was consequent to a rise in inspiratory nasal flow (1.4 [95% CI, 0.8-2.2] L/min; Δ%eupnea, 38.1 pp [95% CI, 23.8-55.4 pp]), which markedly outweighed the decrease in inspiratory oral flow (−0.14 [95% CI, −0.17 to −0.06] L/min; Δ%eupnea, −4.7 pp [95% CI, −4.3 to −2.6 pp]). We also note that the mouth closure maneuver virtually eliminated mouth breathing as intended (baseline oral fraction of total ventilation ≈ 0.2; −0.2 [95% CI, −0.3 to −0.1] change in fraction of oral breathing). Mixed-effects model on breath-level data replicated these findings.
Patients With Highest Levels of Mouth Breathing
The quadratic regression model confirmed that change in airflow with mouth closure (Δflow) was nonlinearly associated with the level of baseline mouth breathing (Figure 2; eFigure 2 in Supplement 2). To illustrate this, patients were separated into 3 quantiles based on the level of airflow through the mouth (Figure 2). In 10 patients with near-zero mouth breathing (quantile 1: oral airflow < 0.05 L/min), mouth closure had an insignificant association with change in inspiratory airflow (0.9 [95% CI, −0.2 to 2.1] L/min; Δ%eupnea, 20.6 pp [95% CI, −5.4 to 46.6 pp]). In 32 patients with moderate levels of mouth breathing (quantile 2: oral airflow = 0.05-2.20 L/min), mouth closure was beneficial, increasing airflow by 2.0 L/min (95% CI, 1.3-2.7 L/min), or a 53.1 pp (95% CI, 37.2-69.1 pp) change in %eupnea. In 12 patients with the highest levels of mouth breathing (quantile 3: oral airflow > 2.2 L/min), mouth closure was detrimental, decreasing airflow by −1.86 L/min (95% CI, −3.10 to −0.62), a −40.4 pp (95% CI, −68.9 to −11.9 pp) change in %eupnea (Figure 3). Mixed-effects model on breath-level data replicated that the outcomes of mouth closure were beneficial to airflow patients in quantile 2 and detrimental to patients in quantile 3. Adjustment for age had no impact on the association between oral breathing and change in airflow with mouth closure (eAppendix 3 and eTable 5 in Supplement 2).
Figure 2. Illustrated Differences in the Association of Mouth Closure With Airflow.
In responders in the top row, some combination of oral and nasal airflow (top left) gets converted to majority nasal airflow and overall airflow increase (top right). The negative responding patient in the bottom row is oral breathing to circumvent velopharyngeal obstruction (bottom left), such that forcing airflow through the nasal route via mouth closure (bottom right) obstructs and ultimately reduces airflow. The arrows under the chin indicate the mouth closure maneuver, in which the investigator applied pressure to the mentum until the teeth were in occlusion.
Figure 3. Patients With Substantial Mouth Breathing Respond Negatively to Mouth Closure Maneuvers.
Nasal breathers did not experience a change with mouth closure; however, it was beneficial to partial mouth breathers and detrimental to patients who breathe primarily through their mouth. The graph represents the average modeled change in airflow in each quantile representing nasal breathers, partial mouth breathers, and mouth breathers. The solid orange line represents the modeled change in airflow with mouth closure, and the shaded region represents the 95% CI.
Site of Airway Collapse
As hypothesized, upstream collapse was associated with greater baseline mouth breathing and a negative airflow response to mouth closure (Table 2). Based on β coefficients, these associations appeared to be driven primarily by 2 of the 3 obstruction subcategories of upstream collapse, namely VAP and VCC (Table 2). Conversely, oropharyngeal lateral wall collapse was associated with a positive airflow response to mouth closure but had no association with mouth breathing (Table 2). Lastly, exploratory analysis showed that the presence of upstream obstruction was a potential predictor of airflow response to mouth closure (accuracy, 77% [95% CI, 65%-91%]; specificity, 81% [95% CI, 69%-94%]; sensitivity, 74% [95% CI, 61%-86%]; eTable 3 in Supplement 2). Adjustment for age had no impact on the association between site of collapse and change in airflow with mouth closure (eTable 5 in Supplement 2).
Table 2. Effect Sizes of the Associations of Site of Collapse Categories With Oral Airflow and Change in Airflow With Mouth Closurea.
| Anatomic site of collapse | Baseline mouth breathing | Flow change with mouth closure | ||
|---|---|---|---|---|
| Oral airflow, L/min | Oral fraction of total ventilation | ∆flow, L/min | ∆%eupnea | |
| Upstream collapse | 0.6 (0.10 to 3.0) | 0.3 (0.04 to 0.8) | −1.9 (−3.1 to −0.7) | −45.2 (−74.3 to −16.1) |
| VCC | 3.0 (0.0 to 68.2) | 0.7 (−0.03 to 3.5) | −3.2 (−6.5 to 0.2) | −82.1 (−161.7 to −2.5) |
| VAP or PP | 0.5 (0.03 to 3.2) | 0.2 (−0.02 to 0.7) | −2.1 (−3.5 to −0.8) | −51.1 (−83.1 to −19.2) |
| VCL | 0.3 (−0.1 to 3.9) | 0.2 (−0.04 to 1.1) | 0.7 (−1.4 to 2.8) | 21.6 (−27.7 to 70.9) |
| Oropharyngeal | −0.01 (−0.1 to 1.1) | 0.1 (−0.05 to 0.6) | 2.3 (0.8 to 3.8) | 66.2 (31.9 to 100.5) |
| Tongue | −0.2 (−0.2 to 0.4) | −0.03 (−0.1 to 0.3) | −1.3 (−2.8 to 0.3) | −25.2 (−62.1 to 11.7) |
| Epiglottis | −0.1 (−0.2 to 1.1) | 0.02 (−0.1 to 0.7) | 0.3 (−1.9 to 2.5) | 8.5 (−45.3 to 62.3) |
Abbreviations: PP, palatal prolapse; VAP, anteroposteriorly direct velum collapse; VCC, concentric velum collapse; VCL, concentric velum collapse with dominant lateral salpingopharyngeus muscle component.
Effect sizes (ie, β coefficients [95% CI] from the regression analysis) for analysis evaluating baseline mouth breathing and flow change with mouth closure across different sites of collapse are expressed as oral airflow (L/min) and oral fraction of total ventilation for the respective site of collapse category, relative to all other site of collapse categories.
Discussion
The results of this nonrandomized clinical trial demonstrated that mouth closure acutely increases overall airflow in patients with OSA by a physiologically meaningful magnitude of approximately 25%. Thus, mouth closure may be a promising strategy for improving pharyngeal obstruction in OSA. However, changes in airflow with mouth closure were highly heterogeneous. Notably, we identified a subgroup in whom mouth closure significantly worsens airflow (Figure 3). This subgroup was characterized by high levels of baseline mouth breathing during sleep, as well as pharyngeal collapse that was upstream from the oral route (ie, palatal obstruction; Table 2). Overall, these findings support the concept that patients with OSA commonly breathe through the mouth to bypass significant palatal obstruction. Blocking this oral bypass via mouth closure ultimately reduces overall airflow in these individuals. Hence, potential OSA therapies that prevent mouth breathing may result in deleterious outcomes in individuals with palatal collapse who rely on the oral route for adequate ventilation. We emphasize that these findings should temper the general belief that preventing mouth breathing during sleep will be universally beneficial.
Mouth-Breathing Outcomes
Mouth breathing during sleep is generally detrimental to airflow compared to nasal breathing.2,10,11,12,13,14 Specifically, in healthy controls, forced oral breathing (nasal route blocked) vs forced nasal breathing (oral route blocked) during sleep resulted in 2.5 times greater upper airway resistance and AHI increasing from 1.5 events/h to 43 events/h.2 Thus, experimentally opening the mouth, at least in the presence of a patent nasopharyngeal breathing route (healthy individuals), is deleterious for respiratory mechanics. Although the mechanism remains unclear, mandibular rotation associated with mouth opening has been shown to shift the tongue and soft palate toward the posterior pharyngeal wall, thereby narrowing the airway,15,16 and likely contributing to increased airway resistance.2 Yet, whether actively closing the mouth improves pharyngeal mechanics in the presence of, or in response to, obstruction remains uncertain.
The present study addresses whether actively closing the mouth improves respiratory mechanics during flow-limited breathing in patients with OSA. Indeed, we showed that on average, forcing nasal breathing in patients with OSA acutely improves airflow, supporting further development of mouth-closure therapies (eg, chin strap, lip taping) to enhance respiratory mechanics in OSA.
Heterogeneity in the Airflow Response to Mouth Closure
This general finding from our study comes with a major caveat: The response to mouth closure in patients with OSA was highly heterogeneous. In 12 of 54 patients (22%), airflow worsened with mouth closure. To our knowledge, no study has described this deleterious response to mouth closure before. As hypothesized, an unfavorable response to mouth closure was associated with greater oral airflow at baseline (ie, mouth breathing; Figure 3); approximately one-quarter of patients with the greatest oral airflow exhibited a marked reduction in airflow of approximately 2.0 L/min with mouth closure. These findings imply that substantial mouth breathing in patients with OSA occurs as a compensatory response to obstruction (ie, this subgroup exhibits a reliance on the oral route for breathing; Figure 2). This finding is supported by prior literature, indicating that OSA severity can increase in some patients with mouth taping and chin strap interventions for mouth breathing prevention.17,18,19 Our study provides mechanistic evidence to suggest that unfavorable responses to mouth closure are physiologically plausible and the observed heterogeneity may not be readily explained away by night-to-night variability. Therapies aimed at preventing mouth breathing may require judicious patient selection to mitigate the risk associated with including patients who rely on their oral breathing route during sleep.
Anatomic Endotypes of Mouth Breathing
The finding that oral breathing was greatest in patients exhibiting obstruction upstream from the oral cavity (ie, velopharyngeal obstruction) supported the concept that patients with greater mouth breathing rely on the oral route as an essential airway, as hypothesized. Patients with velopharyngeal obstruction exhibited not only greater oral breathing but also significantly worse airflow responses to mouth closure. In fact, upstream obstruction strongly predicted response to mouth closure with an estimated accuracy of 77% (eTable 3 in Supplement 2). These data further support the concept that severe velopharyngeal obstruction, which blocks airflow along the nasal route, may result in the presence of oral breathing and reliance on it to bypass the obstruction.
One discovery from our work was that patients with oropharyngeal lateral wall collapse responded most favorably to mouth closure with increased airflow. This suggests a major positive outcome of mouth closure: the dilation of the airway laterally, as visually observed with endoscopy (eFigure 1 in Supplement 2). Mechanistically, mouth closure might generate traction on the lateral oropharyngeal walls, similar to the caudal pharyngeal traction produced by increased lung volume,20 or ansa cervicalis stimulation, the latter of which demonstrates benefit to patients with oropharyngeal lateral wall collapse.21
Clinical Implications
The results of our study suggested that caution should be used in universally applying therapies to prevent mouth breathing in patients with OSA, as approximately one-quarter may experience a deleterious response. Further clinical studies are required; meanwhile, clinicians should closely monitor the outcome of therapies for preventing mouth breathing to ensure it is not worsening their condition.
Conversely, patients with OSA and oropharyngeal lateral wall collapse preferentially benefit from mouth closure interventions. This subgroup is notable due to its low success rate from common non–positive airway pressure therapies, including oral appliances and hypoglossal nerve stimulation therapy.22,23,24 Combining mouth closure interventions with these therapies could potentially enhance treatment efficacy in this subgroup, which warrants further investigation.
Limitations
This study is not without limitations. First, the use of propofol to simulate sleep may not accurately replicate natural sleep. Propofol-induced sedation produces patterns on electroencephalography that are observed during stage N2 sleep.25,26 Although this is the preferred stage of sleep for physiological assessments,27 propofol-induced sedation may not represent the complexities of all sleep stages. However, this approach was necessary to achieve our study objectives, allowing for manual mouth closure on alternating breaths without patient arousal. The approach also facilitated the comparison of airflow between adjacent breaths, such that the assumption can be made that all other aspects of sleep physiology (eg, ventilatory drive, sleep stage, and pharyngeal collapsibility) are highly similar and, therefore, controlled for in the study.
In addition, we investigated the acute outcome of mouth closure in single breaths, which does not reflect the longer-term outcome of mouth closure. Our approach was intentional because we specifically evaluated the direct outcome of the intervention on pharyngeal anatomic factors. Prolonged mouth closure may not be as effective due to feedback mechanisms. Specifically, if you improve the airway and ventilation, feedback neuromuscular control and neuromuscular effects will likely impact the overall outcome of mouth closure. Future studies should investigate the prolonged outcome of mouth closure across different pharyngeal anatomical subgroups.
Furthermore, compared to natural sleep, DISE may produce more passive mouth breathing via mandibular muscle relaxation from propofol, as opposed to mouth breathing as an active consequence of bypassing nasopalatal obstruction. If true, we would expect a greater improvement in total airflow with mouth closure in an individual who breathes primarily through their mouth. However, our findings show the opposite: increased oral airflow is associated with reduced total airflow with mouth closure. Given our findings, we suspect that propofol-induced passive mouth breathing is not the dominant mechanism explaining oral breathing in patients with obstruction along the nasopalatal route.
Additionally, OSA in excluded patients was more severe compared to the study sample, which may impact the generalizability of the results to a population with more severe OSA. However, we repeated our analyses in the entire sample of 66 patients and replicated outcomes across all study aims (eAppendix 1 and eTable 4 in Supplement 2).
Moreover, the present study addresses the physiological efficacy of mouth closure under controlled conditions, rather than efficacy and tolerability in natural sleep. Whether mouth closure interventions are tolerable and improve breathing during sleep in patients absent nasopalatal obstruction remains to be determined in future studies.
Lastly, we did not measure the contribution of nasal obstruction to open-mouth breathing, which may be important because previous studies report that high nasal resistance correlates with mouth breathing during sleep.28 However, we did examine patient-reported symptoms of nasal obstruction, mouth breathing, nasal obstruction symptom evaluation (NOSE) score values, and physical examination findings of nasal obstruction, which did not impact the outcomes in our study (eTable 6 in Supplement 2). Further investigations are needed to determine whether the adverse effects of mouth closure extend to patients who engage in oral breathing to bypass nasal obstruction.
Conclusions
In this nonrandomized clinical trial, mouth closure was associated with improved airflow during sleep in patients with OSA. However, in a subgroup of patients with severe palatal obstruction who primarily breathe through their mouths, airflow worsened when the mouth was closed. These findings support the concept that patients with OSA who mouth breathe rely on the oral route as an essential breathing route. This insight is in contrast to the popular belief that nasal breathing is a universally advantageous form of breathing. Thus, a personalized approach to recommending therapies aimed at preventing mouth breathing should be considered.
Trial Protocol
eResults
eTable 1. Characteristics of All Patients Enrolled During the Study Period
eTable 2. Airflow Parameters of the Mouth Closure Maneuvers (Median [IQR])
eTable 3. Upstream Obstruction Predicts a Negative Response to Mouth Closure: Sensitivity, Specificity, PPV, NPV, and Accuracy [95% CI] of the Proximal Obstruction
eFigure 1. Endoscopic Images of the Pharynx in an Example Patient With Complete Oropharyngeal Lateral Walls Collapse
eFigure 2. Change in Ventilation With Mouth Closure Varies Non-linearly With Amount of Oral Breathing
eAppendix 1. Analysis on All Enrolled Patients (n = 66) Without Exclusion
eTable 4. The Effect Sizes of the Associations of Site of Collapse Categories With Oral Airflow and Change in Airflow With Mouth Closure Using Data From All Participants (n = 66)
eAppendix 2. Breath-Level Analysis
eAppendix 3. Adjusting for Age
eTable 5. The Effect Sizes of the Associations of Site of Collapse Categories With Oral Airflow and Change in Airflow With Mouth Closure With and Without Adjustment for Age
eAppendix 4. Adjusting for Nasal Breathing Parameters
eTable 6. The Effect Sizes of the Associations of Upstream Collapse With Oral Airflow and Change in Airflow With Mouth Closure With and Without Adjustment for Measures of Nasal Obstruction and Mouth Breathing
Data Sharing Statement
References
- 1.Medical subject headings (MeSH): mouth breathing. Accessed February 6, 2024. https://www.ncbi.nlm.nih.gov/mesh/68009058.
- 2.Fitzpatrick MF, McLean H, Urton AM, Tan A, O’Donnell D, Driver HS. Effect of nasal or oral breathing route on upper airway resistance during sleep. Eur Respir J. 2003;22(5):827-832. doi: 10.1183/09031936.03.00047903 [DOI] [PubMed] [Google Scholar]
- 3.Fitzpatrick MF, Driver HS, Chatha N, Voduc N, Girard AM. Partitioning of inhaled ventilation between the nasal and oral routes during sleep in normal subjects. J Appl Physiol (1985). 2003;94(3):883-890. doi: 10.1152/japplphysiol.00658.2002 [DOI] [PubMed] [Google Scholar]
- 4.Suzuki M, Tanuma T. The effect of nasal and oral breathing on airway collapsibility in patients with obstructive sleep apnea: computational fluid dynamics analyses. PLoS One. 2020;15(4):e0231262. doi: 10.1371/journal.pone.0231262 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Schwartz AR, Barnes M, Hillman D, et al. Acute upper airway responses to hypoglossal nerve stimulation during sleep in obstructive sleep apnea. Am J Respir Crit Care Med. 2012;185(4):420-426. doi: 10.1164/rccm.201109-1614OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kezirian EJ, Hohenhorst W, de Vries N. Drug-induced sleep endoscopy: the VOTE classification. Eur Arch Otorhinolaryngol. 2011;268(8):1233-1236. doi: 10.1007/s00405-011-1633-8 [DOI] [PubMed] [Google Scholar]
- 7.Van de Perck E, Heiser C, Vanderveken OM. Concentric vs anteroposterior-laterolateral collapse of the soft palate in patients with obstructive sleep apnea. Otolaryngol Head Neck Surg. 2022;166(4):782-785. doi: 10.1177/01945998211026844 [DOI] [PubMed] [Google Scholar]
- 8.Azarbarzin A, Sands SA, Marques M, et al. Palatal prolapse as a signature of expiratory flow limitation and inspiratory palatal collapse in patients with obstructive sleep apnoea. Eur Respir J. 2018;51(2):1701419. doi: 10.1183/13993003.01419-2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kim HY, Sung CM, Jang HB, Kim HC, Lim SC, Yang HC. Patients with epiglottic collapse showed less severe obstructive sleep apnea and good response to treatment other than continuous positive airway pressure: a case-control study of 224 patients. J Clin Sleep Med. 2021;17(3):413-419. doi: 10.5664/jcsm.8904 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Carskadon MA, Bearpark HM, Sharkey KM, et al. Effects of menopause and nasal occlusion on breathing during sleep. Am J Respir Crit Care Med. 1997;155(1):205-210. doi: 10.1164/ajrccm.155.1.9001313 [DOI] [PubMed] [Google Scholar]
- 11.Lavie P, Fischel N, Zomer J, Eliaschar I. The effects of partial and complete mechanical occlusion of the nasal passages on sleep structure and breathing in sleep. Acta Otolaryngol. 1983;95(1-2):161-166. doi: 10.3109/00016488309130930 [DOI] [PubMed] [Google Scholar]
- 12.Suratt PM, Turner BL, Wilhoit SC. Effect of intranasal obstruction on breathing during sleep. Chest. 1986;90(3):324-329. doi: 10.1378/chest.90.3.324 [DOI] [PubMed] [Google Scholar]
- 13.Young T, Finn L, Kim H; The University of Wisconsin Sleep and Respiratory Research Group . Nasal obstruction as a risk factor for sleep-disordered breathing. J Allergy Clin Immunol. 1997;99(2):S757-S762. doi: 10.1016/S0091-6749(97)70124-6 [DOI] [PubMed] [Google Scholar]
- 14.Zwillich CW, Pickett C, Hanson FN, Weil JV. Disturbed sleep and prolonged apnea during nasal obstruction in normal men. Am Rev Respir Dis. 1981;124(2):158-160. [DOI] [PubMed] [Google Scholar]
- 15.Burger CD, Stanson AW, Daniels BK, Sheedy PF II, Shepard JW Jr. Fast-computed tomographic evaluation of the effect of route of breathing on upper airway size and function in normal men. Chest. 1993;103(4):1032-1037. doi: 10.1378/chest.103.4.1032 [DOI] [PubMed] [Google Scholar]
- 16.Lee SH, Choi JH, Shin C, Lee HM, Kwon SY, Lee SH. How does open-mouth breathing influence upper airway anatomy. Laryngoscope. 2007;117(6):1102-1106. doi: 10.1097/MLG.0b013e318042aef7 [DOI] [PubMed] [Google Scholar]
- 17.Bhat S, Gushway-Henry N, Polos PG, et al. The efficacy of a chinstrap in treating sleep disordered breathing and snoring. J Clin Sleep Med. 2014;10(8):887-892. doi: 10.5664/jcsm.3962 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huang TW, Young TH. Novel porous oral patches for patients with mild obstructive sleep apnea and mouth breathing: a pilot study. Otolaryngol Head Neck Surg. 2015;152(2):369-373. doi: 10.1177/0194599814559383 [DOI] [PubMed] [Google Scholar]
- 19.Labarca G, Sands SA, Cohn V, et al. Mouth closing to improve the efficacy of mandibular advancement devices in sleep apnea. Ann Am Thorac Soc. 2022;19(7):1185-1192. doi: 10.1513/AnnalsATS.202109-1050OC [DOI] [PubMed] [Google Scholar]
- 20.Owens RL, Malhotra A, Eckert DJ, White DP, Jordan AS. The influence of end-expiratory lung volume on measurements of pharyngeal collapsibility. J Appl Physiol (1985). 2010;108(2):445-451. doi: 10.1152/japplphysiol.00755.2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kent DT, Scott WC, Zealear D, Schwartz AR. Ansa cervicalis stimulation increases pharyngeal patency in patients with obstructive sleep apnea. J Appl Physiol (1985). 2021;131(2):487-495. doi: 10.1152/japplphysiol.00076.2021 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Op de Beeck S, Dieltjens M, Verbruggen AE, et al. Phenotypic labelling using drug-induced sleep endoscopy improves patient selection for mandibular advancement device outcome: a prospective study. J Clin Sleep Med. 2019;15(8):1089-1099. doi: 10.5664/jcsm.7796 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Vena D, Huyett P, Op De Beeck S, et al. Flow-shape-derived Site of pharyngeal collapse predicts response to hypoglossal nerve stimulation therapy (abstract). Am J Respir Crit Care Med. 2023;207:A6271. doi: 10.1164/ajrccm-conference.2023.207.1_MeetingAbstracts.A6271 [DOI] [Google Scholar]
- 24.Huyett P, Kent DT, D’Agostino MA, et al. Drug-induced sleep endoscopy and hypoglossal nerve stimulation outcomes: a multicenter cohort study. Laryngoscope. 2021;131(7):1676-1682. doi: 10.1002/lary.29396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Murphy M, Bruno MA, Riedner BA, et al. Propofol anesthesia and sleep: a high-density EEG study. Sleep. 2011;34(3):283-91A. doi: 10.1093/sleep/34.3.283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Stamatakis EA, Adapa RM, Absalom AR, Menon DK. Changes in resting neural connectivity during propofol sedation. PLoS One. 2010;5(12):e14224. doi: 10.1371/journal.pone.0014224 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wellman A, Edwards BA, Sands SA, et al. A simplified method for determining phenotypic traits in patients with obstructive sleep apnea. J Appl Physiol (1985). 2013;114(7):911-922. doi: 10.1152/japplphysiol.00747.2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.McLean HA, Urton AM, Driver HS, et al. Effect of treating severe nasal obstruction on the severity of obstructive sleep apnoea. Eur Respir J. 2005;25(3):521-527. doi: 10.1183/09031936.05.00045004 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
eResults
eTable 1. Characteristics of All Patients Enrolled During the Study Period
eTable 2. Airflow Parameters of the Mouth Closure Maneuvers (Median [IQR])
eTable 3. Upstream Obstruction Predicts a Negative Response to Mouth Closure: Sensitivity, Specificity, PPV, NPV, and Accuracy [95% CI] of the Proximal Obstruction
eFigure 1. Endoscopic Images of the Pharynx in an Example Patient With Complete Oropharyngeal Lateral Walls Collapse
eFigure 2. Change in Ventilation With Mouth Closure Varies Non-linearly With Amount of Oral Breathing
eAppendix 1. Analysis on All Enrolled Patients (n = 66) Without Exclusion
eTable 4. The Effect Sizes of the Associations of Site of Collapse Categories With Oral Airflow and Change in Airflow With Mouth Closure Using Data From All Participants (n = 66)
eAppendix 2. Breath-Level Analysis
eAppendix 3. Adjusting for Age
eTable 5. The Effect Sizes of the Associations of Site of Collapse Categories With Oral Airflow and Change in Airflow With Mouth Closure With and Without Adjustment for Age
eAppendix 4. Adjusting for Nasal Breathing Parameters
eTable 6. The Effect Sizes of the Associations of Upstream Collapse With Oral Airflow and Change in Airflow With Mouth Closure With and Without Adjustment for Measures of Nasal Obstruction and Mouth Breathing
Data Sharing Statement
