Nasal obstruction and male gender contribute to the persistence of mouth opening during sleep in CPAP-treated obstructive sleep apnoea

Introduction

Continuous positive airway pressure (CPAP) is the first-line treatment for moderate to severe obstructive sleep apnoea (OSA).1 CPAP adherence is crucial for improvement of symptoms2 and to reduce cardio-metabolic consequences.3 Adherence depends on several mechanisms such as the self-perceived benefit/inconvenience ratio, social status,4 depressive status and technological aspects. Among this latter, the type of mask may impact on CPAP adherence.5 The choice of the mask is essential for CPAP treatment to be successful.

Nasal masks are most commonly used.6 Oronasal masks, covering the nose and mouth, are usually offered either as a second choice when patients report mouth leaks during CPAP treatment with a nasal mask or immediately when patients have severe nasal obstruction or report sleeping with their mouth open.7

There is a close link between obstructive sleep apnoea–hypopnoea syndrome and mouth breathing during sleep. During an obstructive respiratory event, respiratory effort increases and recruitment of supra-hyoid muscles, mylohyoid, geniohyoid and genioglossus cause mouth opening, which is not compensated by the jaw closing muscles, masseter, internal pterygoid and temporalis.8, 9 Mouth opening in turn induces mandibular retrusion, decreases pharyngeal diameter,10 reduces the mechanical efficiency of the pharyngeal dilator muscles11 and increases pharyngeal resistance and collapsibility.12

CPAP treatment with a nasal mask, by alleviating pharyngeal obstructions, reduces sleep time spent in mouth breathing13, 14 as well as the amplitude of mouth opening movements.15 However, some patients report mouth leaks under nasal CPAP suggesting the persistence of mouth opening during sleep and require an oronasal mask. In clinical practice, oronasal masks is offered about 25%5 of patients treated with CPAP and are often suggested after a few days or weeks of nasal CPAP treatment. The proportion of sleep time spent with the mouth open is usually not documented. The choice of an oronasal mask is based on patient's self-reported symptoms particularly dry mouth, the level of leaks observed with CPAP built-in software and the clinical impression. When an oronasal mask is prescribed, a switch back to a nasal mask is rarely envisaged, implying that mouth opening and mouth leaks may be a long-term condition. However, it has been recently shown that symptoms such as dry mouth and leaks can evolve during treatment with CPAP.16 Moreover, factors contributing to persistent mouth opening under CPAP remain relatively unknown.

In this context, the main objective of this cross-sectional study was to compare mean mouth opening during sleep in patients with obstructive sleep apnoea syndrome treated with CPAP for several months by nasal and oronasal masks. The secondary objective was to identify independent characteristics of patients with persistent mouth opening during CPAP therapy.

Methods

This was a cross-sectional prospective study. Our institutional review board (IRB-6705) approved this study, and all patients included signed written informed consent.

Determination of mouth opening and mandible movements

A mid-sagittal mandibular movement magnetic sensor (Brizzy Nomics, Liege, Belgium) measured the distance between two parallel, coupled, resonant circuits placed on the forehead and on the chin (Fig. 1). The transmitter generated a pulsed magnetic wave of low energy at 10 Hz. The change in the magnetic field recorded at the receiver is inversely related to the distance between the chin and forehead probe. The resolution of the measurement was 0.1 mm. The signal was interpreted as follows: the greater the value of the signal, the lower the mandible position, reflecting wider opening of the mouth. Mandible behaviour has been shown as a reliable marker of respiratory effort during sleep.19 An automatic analysis of mandible movement has been previously developed for the scoring of apnoea-hypopnoea events,9, 20, 21 the scoring of sleep/wake phases22, 23 as well as micro-arousals.21 In the present study, the automatic analysis of this signal was used to assess (i) the mean value of mouth opening over the entire sleep period and (ii) the mean amplitude of the mandible oscillation. The mandibular movement signal was expressed as an absolute value (mm).

Results

Features of sleep recording under CPAP according to the type of mask

Figure 2 shows features of the sleep recording according to the type of mask. Patients treated with oronasal mask had greater mean mouth opening during the night (a). Additionally, the oxygen desaturation index (at 3%) was higher in patients with oronasal masks than in patients with nasal masks (c) although the 95th percentile pressure was not different in the two groups (e). Non-intentional leaks tended to be more important in patients with oronasal masks (f).

Patients' characteristics linked with mouth opening

Table 3 and Figure 3 show the characteristics that were associated with mean mouth opening. Male gender, active smoking, nasal obstruction, neck circumference and lower FEV₁ were significantly associated with mouth opening during sleep.

Table 3. Patients' characteristics according to mouth opening (univariate linear regression analysis)

	Regression coefficient	95% CI	P-value
Age	0.00	(−0.13; 0.14)	0.997
Gender (male vs female)	4.53	(2.14; 6.92)	<0.001
BMI	0.02	(−0.26; 0.31)	0.800
Active smoking (yes vs no)	5.56	(0.03; 11.1)	0.049
Neck circumference	0.41	(0.13; 0.13)	0.006
Waist circumference	0.07	(−0.03; 0.18)	0.146
Mallampati score (3;4 vs 1;2)	−2.63	(−5.62; 0.37)	0.083
FEV1	−0.09	(−0.16; −0.01)	0.022
FEV1/FVC	−0.11	(−0.23; 0.002)	0.055
PaO2	−0.05	(−0.18; 0.07)	0.394
PaCO2	−0.03	(−0.32; 0.26)	0.831
NOSE Score (≥50 vs < 50)	4.02	(1.24; 6.81)	0.006
AHI	−0.02	(−0.07; 0.04)	0.576
Epworth Sleepiness Scale	−0.002	(−0.34; 0.33)	0.989

• AHI, Apnoea–Hypopnea Index; BMI, body mass index; CI, confidence interval; FEV₁ (% predicted value), forced expiratory volume in 1 s expressed in percentage of predicted value; FVC, forced vital capacity; NOSE Score, Nasal Obstruction Symptom Evaluation score.

In multivariate analysis (Table 4), male gender and nasal obstruction remained independently associated with mouth opening during sleep.

Table 4. Patients' characteristics according to mouth opening (stepwise selection procedure in multivariate linear regression model

Variables	Regression coefficient	95% CI	P-value
Gender (male vs female)	4.442	(2.394; 6.489)	0.0001
NOSE Score ≥50 versus < 50	3.904	(1.686; 6.122)	0.0011
Adjusted R2	0.4865

• NOSE score, Nasal Obstruction Symptom Evaluation score.

Discussion

The main findings of this study are: (i) patients treated with an oronasal mask had more mean mouth opening during the night than patients treated with nasal masks; (ii) patients using oronasal mask had a greater oxygen desaturation index than patients on nasal masks; (iii) active smoking, nasal obstruction, male gender, larger neck circumference and lower FEV₁were linked with mouth opening in univariate analysis; and (iv) in a multivariate analysis, male gender and nasal obstruction remained independently associated with mouth opening.

Previous randomized cross-over control studies comparing oronasal versus nasal masks25-28 have included patients who were either CPAP naïve27, 28 or already on effective long-term treatment with nasal CPAP without a chinstrap,25 therefore they did not necessarily need an oronasal mask. In the present study was different since an oronasal mask was mainly provided in second intention to counteract mouth leaks with a nasal mask. Our results show that even after several months of CPAP, patients using ONM persist in keeping their mouths open at night. Moreover, the oxygen desaturation index was also greater in patients with oronasal masks suggesting a lower degree of control of sleep-related respiratory events. This is in accordance with previous published studies25-28 that showed a higher residual AHI under oronasal compared to nasal masks. The greater residual AHI could in part be related to the weight of the mask on the chin and the headgear straps of the oronasal mask that put pressure on the mandible, displacing it backwards and therefore inducing an increase in pharyngeal resistance.7, 29 Interestingly, while the oxygen desaturation index was greater with oronasal masks, the 95th percentile pressure delivered by the auto-PAP was not different between nasal and oronasal masks. This finding is in agreement with Teo et al.28 who also used auto-titrating PAP. In contrast, using polysomnography (PSG)-guided PAP titration, Ebben et al. have shown that oronasal masks required higher pressure levels than nasal masks.27 Thus, the ability of auto-titrating PAP to set appropriate pressures in patients who keep their mouths open while using an oronasal mask could be questioned. It might be hypothesized that the pressure has not been adequately titrated by using auto-titrating CPAP and that persistent mouth opening is the consequence of an elevated upper airway resistance and obstructive events. However, the design of the present study did not allow to delineate the respective importance of mechanisms contributing to mouth opening and oxygen desaturation index between patients treated with nasal mask versus oronasal masks. Moreover, it is possible that the oronasal mask by itself promotes mouth opening by pressing on the jaw. In further studies, it would be interesting to explore if mouth opening persists in patients with a nasal mask and an oronasal mask, respectively, after PSG titration.

Beyond these technical considerations, patients treated with oronasal masks were more often active smokers, tended to have a lower respiratory function (FEV₁) and greater nasal obstruction. Nasal obstruction that can be aggravated by active smoking is a well-known key factor in the persistence of mouth breathing and mouth opening under CPAP.30 Treating nasal obstruction reduced mouth breathing during sleep.31-33

Anthropometric variables such as male gender, neck circumference and respiratory function (FEV₁) were also linked with mouth opening in univariate analysis. Fernandez et al.34 have assessed patients' preference when prescribing the mask for non-invasive home ventilation; they observed that patients who preferred an oronasal mask were more often male, and non-invasive ventilation settings with the oronasal mask required around 2 cm H₂O higher pressure support than settings with nasal masks.34 This study was mainly performed with patients with obesity hypoventilation syndrome who demonstrate an increased work of breathing. This suggests that effort breathing may favour mouth opening and mouth breathing. In our study, male gender was strongly correlated with neck circumference (V′ Cramer > 0.5), thus this latter variable could not be proposed to the multivariate analysis. However, neck circumference also reflects central fat distribution. Centrally distributed fat has detrimental effects on thoraco-pulmonary mechanics35 and upper airway collapsibility,36 particularly in recumbent positions, increasing the effort breathing,35 and may favor mouth opening during sleep.

This study has some limitations. First, the small number of patients treated with an oronasal mask. The objective was to recruit at least 36 patients from consecutive patients attending our sleep clinic. The ratio of patients treated with oronasal mask versus nasal mask was equivalent to that observed in a large cohort study.5 Multiple comparisons should have led to decrease in study power and inflation of the type I error. In order to counteract this problem, we have reported P-values (Holm–Bonferroni correction); as shown with these corrected P-values, comparisons between patients with nasal mask versus oronasal mask have to be taken with caution. These results refer to a first exploratory study that needs to be confirmed by further trials. However, regarding our main objective, even with Bonferroni correction, patients with oronasal mask exhibited a significantly greater mouth opening during sleep than patients with nasal mask.

The second main limitation is that we did not use polysomnography to document sleep duration and residual respiratory events. However, the oxygen desaturation index remains a robust marker of sleep-related respiratory disturbance. Finally, non-intentional leaks tended to be more important in patients with oronasal masks. It was difficult in such a sample to compare several oronasal masks from several brands. Moreover, we cannot formally exclude that the wire of the chin sensor has not interfered with the oronasal mask seal and thus minimally contributed to the leaks; however this wire is lightweight (<1 mm of diameter), and we did not identify leaks related to this wire during the sensor placement.

In conclusion, the present study shows that apnoeic patients using oronasal masks with CPAP keep their mouths open during the night even after several months of treatment. Nasal obstruction is determinant; anatomical and functional abnormalities that increase nasal resistance should be evaluated and if possible treated prior to the initiation of oronasal mask use. The NOSE questionnaire may be a simple tool in this decision process. Others factors such as antropometric features which contribute to mouth opening need to be better understood.

Acknowledgements

We thank Professor Frédéric Sériès (CRIUCPQ, Laval University, Quebec, Canada) for the fruitful discussion and Dr Alison Foote (Grenoble Clinical Research Centre) for the English editing. The study was funded by the ‘AGIR à dom. Association’ and ‘Fonds AGIR pour les Maladies Chroniques’.