Abstract
Although there is an association between nasal obstruction, oral breathing and obstructive sleep apnoea syndrome (OSAS), it remains unknown whether increased oral breathing occurs in patients with OSAS who are free of nasal obstruction. The present study evaluated the relationship between breathing route and OSAS in patients without nasal obstruction.
The breathing route of 41 snorers (25 male; aged 26–77 yrs) with normal nasal resistance was examined during overnight polysomnography using a nasal cannula/pressure transducer and an oral thermistor.
In total, 28 patients had OSAS (apnoeics) and 13 patients were simple snorers. Apnoeics had a higher percentage of oral and oro-nasal breathing epochs. Oral and oro-nasal breathing epochs were positively related with apnoea/hypopnoea index (AHI) and duration of apnoeas/hypopnoeas and inversely related to oxygen saturation. Additionally, oro-nasal breathing epochs correlated with body mass index (BMI). In multiple linear regression analysis, oral breathing epochs were independently related only to AHI (r2 = 0.443), and oro-nasal breathing epochs were independently related to AHI (r2 = 0.736) and BMI (r2 = 0.036).
In conclusion, apnoeics spent more time breathing orally and oro-nasally than simple snorers, and the apnoea/hypopnoea index is a major determinant of the time spent breathing orally and oro-nasally.
Humans preferentially breathe via the nasal route, the purpose being to filter, warm and humidify the inspired air 1. During sleep, healthy subjects free of nasal disease are estimated to inhale via the oral route only ∼4% of the total ventilation, irrespective of sleep stage or body position 2. Only ageing influences breathing route and thus, older individuals have decreased nasal breathing during sleep 3.
In healthy subjects, it is not surprising that nasal obstruction results in an increase of the proportion of mouth breathing 4. However, oral breathing during sleep has been associated with breathing disorders 5. Indeed, experimental nasal occlusion disturbed sleep and triggered the induction of obstructive apnoeas 6. Similarly, allergic rhinitis provoked both sleep fragmentation and obstructive sleep apnoea syndrome (OSAS) 7. OSAS appeared to be reversible during remissions of the allergic symptoms 7.
The association between nasal obstruction and OSAS has been thoroughly evaluated in several studies 8–10. For instance, Lofaso et al. 9 demonstrated that patients with OSAS tended to have higher nasal resistance than snorers without OSAS, and that nasal resistance was an independent risk factor for OSAS. McLean et al. 10 demonstrated a marked reduction in mouth breathing when nasal resistance was reduced with a decongestant. However, in the latter study the impact of nasal resistance on sleep apnoea severity was very modest indeed. Thus, it remains uncertain whether increased oral breathing occurs in patients with OSAS in the absence of increased nasal resistance. Data linking mouth breathing and sleep apnoea in the absence of increased nasal resistance is sparse and indirect 3. An evaluation of the breathing route during sleep in OSAS could be an important step in understanding the upper airway physiology in these patients.
Therefore, the present study aimed to document the breathing route in patients with OSAS but without nasal obstruction. It was hypothesised that these patients would have increased oral or oro-nasal breathing, which may be associated with the severity of the disease. This may be the result of nocturnal hypoxaemia, along with sleep fragmentation, leading to further instability of the ventilatory control mechanisms 11, thus affecting the breathing route that patients use during sleep.
MATERIALS AND METHODS
Study subjects
The study comprised 62 subjects referred to the Centre of Sleep Disorders of Evangelismos General Hospital (Athens, Greece), for suspected sleep disordered breathing during a 3-month period. The enrolment criteria were: 1) snoring and either excessive daytime sleepiness or observed apnoeas; 2) no upper or lower respiratory tract disease, including a history of nasal allergy; 3) no recent surgery involving the upper airway; 4) no history of evaluation or treatment for sleep apnoea; 5) no use of medications known to influence nasal resistance (antihistamine, decongestants, etc.); 6) no history of any neuromuscular disorder; and 7) no history of cardiovascular disease. Enrolment criteria were chosen to provide a spectrum of disease ranging from clinically significant OSAS to nonapnoeic snoring. Exclusion criteria were: 1) nasal resistance values exceeding the normal values (see below); 2) central apnoeas more than three incidences·h-1 or 5% of total apnoeas; and 3) total sleep time <3 h. The local hospital ethics committee approved the study and all subjects gave their written informed consent prior to enrolment in the study.
Study design
Each subject reported to the sleep laboratory between 21:00 and 22:00 h. Nasal resistance was measured in upright seated and supine positions. A full-night diagnostic polysomnography with concomitant monitoring of the breathing route during sleep was then performed, usually from 00:00–07:00 h.
Rhinomanometry
For each subject, nasal resistance to airflow was measured during wakefulness without decongestion, first in the upright seated position and then in a supine position after lying down for 10 min. Active anterior rhinomanometry (PDD-301/r; Piston, Budapest, Hungary) was performed following international recommendations 12. In brief, patients wore a closely fitting face mask, which did not distort the nostrils or the nasal valve, and breathed through one nostril only (first the left and then the right) with the mouth closed. The pressure probe was placed at the opening of the contralateral occluded nostril not being tested. Total resistance was then automatically calculated from the two unilateral measurements. Nasal resistance was given at a pressure difference of 150 Pa across the nasal passage. Active posterior rhinomanometry (with the pressure sensor placed transorally into the posterior pharynx) was also performed in all subjects who had normal nasal resistance in the active anterior rhinomanometry. However, owing to a lack of cooperation, these measurements were not reliable or reproducible in six subjects and were omitted from further analysis. The nasal resistance values of the remaining subjects, as measured by active posterior rhinomanometry, were similar to those measured using active anterior rhinomanometry. Nasal resistance values <3.0 cmH20·L-1·s were considered within normal limits 13.
Polysomnography
A full-night diagnostic polysomnography (EMBLA S7000; Medcare Flaga, Reykjavik, Iceland) was performed in each subject. To determine the stages of sleep, an electroencephalogram (with four channels, C4-A1, C3-A2, O2-A1, O1-A2), electro-oculogram (with two channels) and electromyogram of the submentalis muscle (with one channel) were obtained. Arterial blood oxyhaemoglobin was recorded with the use of a finger pulse oximeter. Thoracoabdominal excursions were measured qualitatively using respiratory effort sensors (XactTrace belts featuring Respiratory Inductive Plethysmography; Medcare Flaga) placed over the ribcage and abdomen (two channels). Snoring was detected with a vibration snore sensor (one channel) and body posture with a body position sensor (one channel). Airflow was monitored using an oral thermistor (oral flow sensor; Medcare Flaga) placed in front of the mouth and a nasal cannula/pressure transducer (21in/53cm; Medcare Flaga) inserted in the opening of the nostrils and linked to independent channels; both were supported by hooking their leads behind the ears and taping them to the sides of the face.
Thermally sensitive devices such as thermistors cause a minimum amount of disturbance to the patient during sleep. However, the flow signal they provide is not a direct measure of actual flow 14 and thus, the detection of airflow provides only qualitative information 15. The signal obtained by a nasal cannula/pressure transducer has, however, been shown to be comparable in both shape and amplitude to that of a conventional pneumotachograph 16. Consequently, this nonobtrusive device is ideal not only for qualitative measurements but also for quantitative monitoring of respiration during sleep 17. Cross-contamination between the oral and nasal channel was meticulously excluded by regular testing during polysomnographic calibration. Thus, subjects were asked to breathe normally and exclusively through the nose for 30 s and subsequently through the mouth for another 30 s in both supine and right lateral postures to allow verification that each sensor was activated exclusively. Sensors were continually checked during the recording to avoid dislodgement. All variables were recorded with a digital acquisition system (Somnologica 3.3; Medcare Flaga).
Analysis
Sleep stage was scored manually in 30-s epochs 18. Obstructive respiratory events were scored using standard criteria 15, 19 by an experienced technician. Thus, apnoea was defined as the absence of airflow for >10 s in the presence of continued respiratory efforts 15. Hypopnoea was defined as a reduction in chest wall movement to an amplitude that was smaller than ∼70% of the baseline level, lasting >10 s, and leading to a decrease in haemoglobin saturation of 4% 19. The number of episodes of apnoeas and hypopnoeas per hour of sleep is referred to as the apnoea/hypopnoea index (AHI). Obstructive sleep apnoea was diagnosed if the AHI was >5 15.
The route of breathing was evaluated using the oral and nasal sensors signals to classify each 30-s epoch as nasal, oral or oro-nasal breathing 20. A nasal breathing epoch was defined as an epoch containing at least three consecutive phasic signals on the nasal sensor only, and an oral breathing epoch defined as an epoch containing at least three consecutive phasic signals on the oral sensor only. An oro-nasal breathing epoch contained at least three consecutive phasic linked signals on both the nasal and oral sensors. Epochs containing apnoeas, hypopnoeas, arousals, movements and signal artefacts were excluded from analysis. The occurrence of nasal, oral and oro-nasal breathing epochs was expressed as a percentage of the total sleep epochs (TSE) analysed 20. Body posture was expressed as a percentage of total sleep time.
Quantitive data are reported as mean±sd. The normality of data distribution was assessed by the Kolmogorov–Smirnov test. Comparison of data between apnoeics and snorers was carried out using an unpaired t-test. Relationships between oral or oro-nasal breathing epochs and various variables were investigated by performing simple linear regression analysis for each variable separately. Multiple linear regression analysis was performed to identify the variables independently related to oral or oro-nasal breathing epochs. All variables that correlated significantly with oral or oro-nasal breathing epochs in simple linear regression analysis were the independent variables included in the model. The stepwise procedure was used to select the best model. A p-value <0.05 was considered to indicate statistical significance.
RESULTS
Of 62 snorers (38 male) initially enrolled in the study, 41 were considered eligible for further analysis and 21 were excluded (nine with nasal resistance values exceeding normal limits in seated and/or in supine position, three with more than three central apnoeas·h-1, two with central apnoeas >5% of total apnoeas detected, and seven who slept for <3 h).
The remaining 41 subjects (25 males, mean age 51.3±14.5 yrs, mean body mass index (BMI) 31.5±5.7 kg·m-2) had nasal resistance within the normal range in erect seated (mean±sd (range) 1.86±0.45 (1.10–2.60) cmH2O·L−1·s) and in supine position (2.3±0.45 (1.30–2.90) cmH2O·L-1·s). The mean total sleep time was 301.0±55.5 min and the mean sleep efficiency was 89.3±10.9%. The total sleep epochs analysed were 434.5±173.0 (102–731) epochs. Total sleep time was divided into the following sleep stages: stage 1: 4.3±4.0%; stage 2: 76.9±10.4%; stage 3: 2.8±4.6%; stage 4: 1.5±3.6%; and rapid eye movement: 14.4±7.5%. The subjects spent 50.5±29.3% of total sleep time in a supine position and the rest in lateral positions. Prone position was seldom detected.
OSAS was diagnosed in 28 patients (apnoeics). Their AHI was 27.6±5.2 (8.3–103.1) events·h-1. The remaining 13 subjects (snorers) had an AHI of 1.3±0.3 (0.5–4.7) events·h-1. Table 1⇓ illustrates anthropometric data, sleep parameters and nasal resistance in snorers and apnoeics.
Anthropometric data, nasal resistance values and sleep parameters in snorers and apnoeics
Nasal breathing epochs were more frequent in snorers than in apnoeics (p<0.001; fig. 1a⇓), whereas the inverse was true for oral (p = 0.004; fig. 1b⇓) and oro-nasal breathing (p<0.001; fig. 1c⇓) epochs. Oral breathing epochs were rare in snorers (0.02±0.08% of TSE) and apnoeics (2.3±3.9% of TSE).
Occurrence of a) nasal, b) oral and c) oro-nasal breathing epochs in snorers and apnoeics during sleep. Data are presented as mean±sd. TSE: total sleep epochs. #: p = 0.004 versus snorers; ***: p<0.001 versus snorers.
In apnoeics, 41.4% of the oral or oro-nasal breathing epochs occurred in the first quarter of the time between two events (i.e. after apnoeas/hypopnoeas), and 32.2% in the last quarter (i.e. immediately before apnoeas/hypopnoeas); 12.8% and 13.6% of the oral or oro-nasal breathing epochs occurred in the second and third quarter of the time between events, respectively.
Relationships between the occurrence of oral or oro-nasal breathing epochs and polysomnographic or anthropometric variables are shown in tables 2⇓ and 3⇓, respectively. As can be seen, oral breathing epochs were positively related with AHI (fig. 2a⇓), mean and longest duration of apnoea/hypopnoea, and were inversely related to average and lowest oxygen saturation. In addition, the relationship between oro-nasal breathing epochs and AHI (fig. 2b⇓), mean and longest duration of apnoea/hypopnoea and BMI was positive, whereas average and lowest oxygen saturation were inversely related to oro-nasal breathing.
Relationship between apnoea/hypopnoea index and a) oral (r2 = 0.443; p<0.0001) and b) oro-nasal (r2 = 0.736; p<0.0001) breathing epochs in snorers (○) and apnoeics (•). TSE: total sleep epochs.
Simple linear regression analysis models for oral breathing
Simple linear regression analysis models for oro-nasal breathing
The results of the forward-stepwise multiple linear regression analysis are summarised in tables 4⇓ and 5⇓. All variables that significantly related to oral (table 2⇑) and oro-nasal breathing epochs (table 3⇑) in the simple linear regression analysis were the independent variables included in the models. Oral breathing epochs were independently related only to AHI (higher in patients with increased AHI). AHI explained 44.3% of the variance of oral breathing epochs (r2 = 0.443; table 4⇓). Oro-nasal breathing epochs were independently related to AHI (higher in patients with increased AHI) and BMI (higher in patients with increased BMI). AHI accounted for almost all of the variance (r2 = 0.736), whereas BMI accounted for only a small amount of the variance (r2 = 0.036) explained by the model (r2 = 0.772; table 5⇓).
Multiple linear regression analysis for oral breathing
Multiple linear regression analysis for oro-nasal breathing
DISCUSSION
The main findings of the present study are as follows. 1) Patients with OSAS demonstrate an increased proportion of sleep-time breathing orally or oro-nasally in comparison with snorers. 2) OSAS severity, expressed by the AHI, is a major determinant of time spent breathing orally and oro-nasally. 3) BMI is also a small but independent contributor to the time spent breathing oro-nasally.
In the present study, AHI determines a significant part of oral breathing epochs (r2 = 0.443; table 4⇑) and simultaneously constitutes a powerful predictor of oro-nasal breathing epochs, accounting for more than two-thirds of their variance (r2 = 0.736; table 5⇑). A pathophysiological mechanism is inevitably implied. It has been noted that apnoeas tend to be followed by large mouth breaths 3, as if the subjects is trying to compensate for the lack of air. These breaths were excluded from analysis in the majority of cases as they were part of epochs containing apnoeas/hypopnoeas. During intervals of undisturbed sleep between apnoeas/hypopnoeas, episodes of oral and oro-nasal breathing epochs reappeared. Interestingly, these episodes were more frequent soon after and before an apnoea/hypopnoea event, as if there were an affinity between oral or oro-nasal breathing epochs and apnoeas/hypopnoeas. Accordingly, it is plausible to consider that the appearance of oral or oro-nasal breathing triggers a vicious cycle in which a number of other factors contribute to further increase of apnoeas/hypopnoeas and thus, to more frequent mouth breathing. These factors are associated either with mouth opening/breathing or absence of nasal breathing. Indeed, it is well documented that mouth opening increases the propensity to airway collapse 21–23. However, mouth opening does not necessarily imply mouth breathing, because it has been shown that the tongue and soft palate can be in close apposition with concomitant preservation of nasal breathing 21. On the contrary, the reverse condition is intuitively obvious 24. When the oral thermistor detects the presence of oral airflow, opening the mouth is an evident prerequisite given that the possibility of cross-contamination by nasal airflow is minimal. Therefore, in this study, mouth breathing was considered to coincide with different degrees of mouth opening. In addition, there is evidence suggesting that irritation of nasal airflow-sensitive receptors during nasal breathing may be important in maintaining upper airway patency by increasing oro-pharyngeal muscle activity. In fact, White et al.25 blocked these receptors using 4% lidocaine local anaesthesia and provoked a four-fold increase in the sleep disordered breathing events 25, whereas Basner et al. 26 measured increased genioglossal and alae nasi electromyographic activity in awake humans breathing through the nose. The latter finding was not corroborated by authors who, subsequently, confirmed only the increase in alae nasi activity 27, 28. Therefore, it is possible that the above mentioned vicious cycle, along with nocturnal hypoxaemia and sleep fragmentation, may progressively influence the ventilatory control mechanisms of patients with OSAS 11, finally affecting the breathing route that patients use during sleep and establishing frequent oral or oro-nasal breathing.
Interestingly, BMI was a determinant, albeit weak, of oro-nasal breathing. It is supposed that as BMI increases, adipose tissue deposition under the jaw rises, rendering jaw opening easier and favouring oro-nasal breathing 28.
Three other studies have examined the correlation between apnoeas and mouth breathing or mouth opening. Gleeson et al. 3 investigated breathing route during sleep in 14 healthy individuals. Although, nasal resistance was not measured in all the participants (eight out of 14 subjects) and the measurements took place several months after the sleep studies, the authors concluded that the proportion of mouth breathing was significantly higher in subjects with apnoeic episodes (four males who had three to nine apnoeas per hour). In another study, Hollowell and Surratt 24 demonstrated that during sleep, the jaws of patients with OSAS were open more than those of normal subjects and opened further at the termination of apnoeas. The authors pointed out that the greater jaw opening in patients with OSAS could destabilise and compromise the upper airway. Moreover, Fitzpatrick et al. 29 studied 12 healthy subjects who underwent overnight polysomnography with a face mask with independent oral and nasal compartments, and forced them alternatively to breathe either orally or nasally. The occurrence of upper airway obstruction episodes was significantly higher in oral ventilation (AHI 43±6) than in nasal ventilation (AHI 1.5±0.5).
Some possible weaknesses of the current study must be acknowledged and deserve consideration. First, the use of the nasal cannula/pressure transducer and oral thermistor to detect airflow presents some drawbacks that have been thoroughly discussed previously 20. Although these devices are nonobtrusive and easily tolerated, they cannot quantify ventilation, partly because their signal–flow relationship is nonlinear. Indeed, the nasal cannula/pressure transducer has a quadratic pressure–flow relationship that may result in underestimation or absence of ventilation at low flows 15. In addition, the signal–flow relationship of the oral thermistor is logarithmic, which may result in overestimation of ventilation at low flows 15. Accordingly, it would be possible that ora- only breathing may still have a nasal component, and any detection of oral-only breathing might actually be scarce. Therefore, the frequency of oral-only breathing epochs, although rarely encountered, could be overestimated in the present study. Moreover, with the instrumentation used, assessment of the heterogeneity of oro-nasal breathing pattern is unfeasible since it is impossible to distinguish different percentages of oral and nasal components. Secondly, sensor dislodgement from the nares or from the mouth could be a potential problem. The incidence of signal loss has been reported to reach ∼5% of cases 17. In the present study, only three subjects had a brief loss of signal in some part of the recording, which never exceeded 5 min. The loss of signal was easily recognised and corrected by the technician on duty. However, it is likely that slight deviations in thermistor position may not have been avoided and this may have resulted in nasal airflow contamination of the oral signal. Finally, reliable posterior rhinomanometry measurements were not available in all patients. In the present study, nasal resistance was measured in seated and supine positions using both posterior and anterior rhinomanometry. A key point in posterior rhinomanometry is to keep the soft palate elevated and the upper surface of the tongue away from the palate so as to allow free communication between the oropharynx and the oral cavity. Insufficient palatal control and cooperation were the reasons for poor reproducibility encountered in six subjects of the present study. However, anterior rhinomanometry gave reliable and reproducible findings in all patients. Indeed, anterior rhinomanometry requires minimal cooperation, thus increasing reproducibility and minimising failure rate, although posterior nasal malformations cannot be determined 12.
Mouth breathing, as indicated with an oral thermistor, appears to be one of the signs of OSAS, which probably ought to be taken into account when assessing the clinical presentation of a subject referred to a sleep disorders centre. However, interpretation of this sign warrants attention, because self-reports of mouth breathing during sleep 10 and even during wakefulness 4 are not frequently reliable, as they may diverge from sleep study data 4, 10.
In summary, the present study illustrates that apnoeics and snorers without nasal obstruction differ in the breathing route used during sleep. Apnoeics tend to spend more time than snorers breathing either orally or oro-nasally. Additionally, as the severity of obstructive sleep apnoea syndrome increases, the proportion of oral and oro-nasal breathing epochs appears to rise.
- Received May 2, 2006.
- Accepted September 5, 2006.
- © ERS Journals Ltd
References
