higher effective oronasal versus nasal continuous positive airway pressure in obstructive sleep apnea : Effect of mandibular stabilization

1Respiratory Epidemiology and Clinical Research Unit, McGill University, and Respiratory Division, McGill University Health Centre; 2Clinique d’Orthodontie, Université de Montréal; 3Sleep Laboratory, Centre Hospitalier Universitaire de Montreal – Hotel-Dieu, Montreal; 46900 Bd Cousineau, St-Hubert, Quebec Correspondence: Dr Marta Kaminska, McGill University Health Centre, 687 Pine Avenue West, Room L4.08, Montreal, Quebec H3A 1A1. Telephone 514-934-1934 ext 35650/36117, fax 514-843-1695, e-mail marta.kaminska@mcgill.ca Obstructive sleep apnea (OSA) is a highly prevalent condition (1) that is associated with obesity and specific craniofacial features such as retrognathia, primarily in nonobese patients (2-4). OSA is best treated using continuous positive airway pressure (CPAP) (5). Mandibular advancement devices (MADs) are an alternative treatment in patients with mild-to-moderate OSA (6). A variety of interfaces are available for CPAP (7). Although nasal masks are most frequently used, oronasal masks, covering the nose and mouth, may be chosen due to patient preference, nasal obstruction or air leak through the mouth. In practice, nasal and oronasal masks are often used interchangeably, and are generally believed to be equally effective (7-9). There is, however, growing evidence suggesting that CPAP by oronasal mask is less effective at establishing airway patency in some individuals (10-13). We hypothesized that, in individuals in whom oronasal CPAP is less effective than nasal CPAP at correcting upper airway obstruction, this results from posterior displacement of the mandible by the oronasal mask, exacerbating upper airway obstruction. The primary objective of the present study was to determine whether mandibular stabilization in a neutral position using a MAD can reduce the oronasal CPAP level required for effective OSA correction in such individuals. The secondary objective was to evaluate whether any anatomical features are associated with poor response to oronasal CPAP by analyzing cephalometric characteristics.


Procedures
All subjects underwent overnight PSG twice.Each time, CPAP was titrated twice successively, once with a nasal mask and once with an oronasal mask.The order in which the masks were used was determined randomly for the first subject, alternating for subsequent subjects.The order of the first night was maintained for the second night.During the second PSG, CPAP titration was repeated using the MAD in the neutral position.All subjects also underwent cephalometry.

PSG and CPAP titration
Recording included standard electroencephalogram and electrooculogram leads, chin electromyography, electrocardiogram, snoring and body position (Sandman system, Covidien, USA).Oxygen saturation was measured using pulse oximetry (Oximax, Nellcor Puritan Bennett Ltd, USA).Respiratory efforts were measured via thoracic and abdominal piezoelectric belts.Respiratory flow was measured using a pneumotachograph (Hans Rudolph Inc, USA) directly connected to the mask.Pressure at the mask was measured via a pressure transducer (Validyne Engineering Corp, USA) connected through a catheter inserted into a tight opening in the mask shell.With the nasal mask, a thermistor was used to detect any airflow through the mouth.Intrathoracic pressure was recorded using an esophageal balloon (Cardinal Heath, USA) connected to a pressure transducer (Validyne, USA).Sleep scoring was performed according to Rechtschaffen and Kales (14).
For CPAP titration, the best-fitting size of nasal (Profile Lite, Philips-Respironics, USA) and oronasal (Disposable Non-vented Full Face Mask, ResMed, USA) masks were chosen for each subject by the same experienced respiratory therapist attending the PSG.Masks themselves were devoid of intentional leaks.The exhaust was distal to the pneumotachograph through a fenestrated connector piece.A BiPAP Synchrony connected to an Omnilab system (Philips-Respironics, USA) was used for titration.Titration was performed only during supine stage 2 sleep, by increasing the pressure by 1 cmH 2 O every 5 min until the effective pressure (Peff) was achieved for each type of mask or a maximal pressure of 20 cmH 2 O was reached.Peff was defined by achievement of stable tidal breathing with absence of snoring and inspiratory esophageal pressure stability.Total air leak was continuously measured by the Omnilab system.Higher-than-expected leak led to mask readjustment before any further titration.

MAD
Subjects had a set of upper and lower dental impressions made in alginate (Plastalgin ortho, Septodont, USA) with a bite (Imprint Bite, 3M ESPE, USA) taken in maximum intercuspidation.A MAD (orthèse O.R.M, Laboratoires Narval, Res Med, USA) was fabricated in a neutral position (without any advancement of the mandible).The MAD was adjusted in the clinic by the same orthodontist for proper fit and comfort with and without the CPAP masks in place.

Cephalometry
Each subject underwent a lateral cephalometric radiograph (Orthophos CD orthopantomograph, Siemens, Canada) taken in natural head position to perform measurements using the landmarks shown in Appendix 1.The tracings and measurements (15) were performed manually from the radiograph by the same orthodontist on two different occasions for all subjects to ensure reproducibility of measurements (Appendix 2).

RESULTS
Six subjects were identified between July 2008 and February 2009.Subject characteristics are presented in Table 1.None of the subjects had undergone any oropharyngeal surgery.
Nasal Peff significantly correlated with the respiratory disturbance index (r=0.843;P=0.04).Using the oronasal mask, Peff was not attained in four patients, with obstructive events persisting at the maximal CPAP of 20 cmH 2 O (Figure 1).In the other two patients, oronasal Peff was higher than nasal Peff.Total air leak was similar with the two types of mask (Figure 2).
With mandibular stabilization, Peff by nasal mask remained within 1 cmH 2 O of the nasal Peff without MAD for five of the six subjects,   3).Cephalometric measurements were compared with published norms (16,17) by calculating Z scores.The group mean Z score for each measurement was >−1.65, suggesting no significant differences compared with the general population.

DISCUSSION
We identified and studied individuals with OSA whose Peff by oronasal mask was considerably higher compared with nasal Peff.In four patients, airway patency could not be established by oronasal mask even with CPAP 20 cmH 2 O, but CPAP by nasal mask resulted in stable breathing at ≤15 cmH 2 O.We used esophageal manometry to confirm the obstructive nature of the persistent respiratory events.Excessive air leak was not responsible for the difference between masks.Mandibular stabilization resulted in partial reduction of the higher oronasal Peff.
While oronasal masks were previously found to be effective for OSA treatment (8,9,11,18), several reports suggest they may not be equivalent, at least in some patients, to nasal masks.In one study, oronasal Peff was at least 2 cmH 2 O greater than nasal Peff in 46% of patients.In other studies, patient satisfaction and average nightly compliance were lower (18), and sleep was slightly more perturbed (8) with oronasal compared with nasal masks.It has been suggested that the higher Peff by oronasal mask is primarily due to excessive air leak (12).Our results suggest otherwise: any excessive leak led to immediate mask readjustment and leak was equivalent for both types of masks (Figure 2).Interestingly, in a physiological study, Smith et al (19) were unable to obtain airway patency by oronasal mask in their group of subjects.The discrepancy between the results from the study by Smith et al (19) and the prevalent use of oronasal CPAP has not been explained.A recent case report (10) described a patient in whom nasal CPAP was effective at a relatively low pressure whereas oronasal CPAP resulted in persistent airway obstruction with elevated pressures; endoscopy confirmed oropharyngeal airway obstruction.
Oronasal masks may cause or exacerbate upper airway obstruction directly by displacing the mandible posteriorly.We used a MAD in the neutral position in conjunction with oronasal CPAP to verify this hypothesis.MADs are a treatment option for mild-to-moderate OSA, particularly in nonobese patients with positional OSA (6).A response (significant partial or complete OSA correction) has been associated with an increase in the size of the pharyngeal lumen (20).We expected the MAD in neutral position to lower oronasal Peff to close to nasal Peff by preventing any posterior manibular displacement caused by the oronasal mask.Nasal Peff showed no or minimal change, confirming that there was no significant direct effect of the MAD on upper airway obstruction.With the oronasal mask, the MAD appeared to show partial benefit in three of the six patients.Hence, it appears that mandibular displacement accounted in part for the unfavourable effect of the oronasal mask, and that the mechanisms underlying poor oronasal mask effectiveness may differ among individuals.This may be related to different sites of obstruction.In the single case described by Schorr et al (10), endoscopy demonstrated posterior displacement of the base of the tongue with the oronasal mask.However, the level of obstruction may differ among individuals.Also, the MAD may not achieve the same effect in all cases.The literature suggests that patients with oropharyngeal collapse at baseline ( 21) or a smaller oropharynx (22) are more likely to respond to a MAD.However, a magnetic resonance imaging study showed that response to the MAD was primarily explained by expansion of the velopharyngeal volume in its lateral dimension (20).It should also be noted that the MAD itself may create a change in the position of the mandible by opening the occlusion, as well as crowding in the mouth with posterior displacement of the tongue.Both of these factors may exacerbate airway obstruction.These effects may explain the small increase in both nasal and oronasal Peff in one of our cases.However, because nasal Peff did not change significantly with the MAD for the remaining subjects, this effect is unlikely to be relevant in most cases.
With an oronasal mask, patients may mouth breathe and mouth opening may be responsible for mandibular position change, which can in turn affect upper airway patency because the position of the mandible (23) and tongue (24) are significant elements in OSA pathogenesis.Additionally, CPAP, when applied at the nose and mouth simultaneously, may fail to achieve a sufficient airway transmural pressure figure 2) Total air leak during continuous positive airway pressure titration via nasal and oronasal mask (without mandibular advancement device).While leak was continuously available during polysomnography (PSG) titration on the Omnilab (Philips-Respironics, USA) system, it was not recorded as part of the PSG.Rather, it was documented intermittently as a comment; hence, it is not available for all subjects for all pressure levels difference to establish airway patency (19).Mouth breathing can also increase upper airway surface tension, in turn worsening OSA (25).The MAD may then work by promoting mouth closure and tongue apposition to the palate.We did not ascertain mouth opening with the oronasal mask in our study.However, Sanders et al (9) showed that even with a mouthpiece in place to maintain the mouth open, oronasal CPAP remained effective at a level comparable with the nasal mask in their group of subjects, suggesting that mouth opening does not explain oronasal CPAP ineffectiveness, although the effect, once again, may vary among individuals and the mouthpiece itself may alter the physiology.
Concomitant use of the MAD in neutral position with CPAP was well tolerated in our subjects, although more long-term data will be needed to confirm our single-night findings.Nevertheless, concomitant use of CPAP with a MAD is feasible and could be a treatment option in patients who do not tolerate nasal CPAP but have very high or unattainable oronasal Peff.
While not assessed in the present study, the MAD with mandibular advancement may further reduce oronasal Peff, and likely also nasal Peff.This may be useful for treatment of severe OSA.It is conceivable that both a MAD and CPAP, used concomitantly, could be titrated for optimal OSA correction while minimizing the discomfort and adverse effects of excessive mandibular advancement and high pressure, respectively.This may lead to improved overall tolerance and, hence, adherence to OSA treatment.
A limitation of our study was that two titrations (one with each mask) were performed consecutively during a single night, once without and once with the MAD.To avoid any potential effect of differences in OSA from early to late night, we have alternated the type of mask used first (Table 1), and standardized our analyses to supine stage 2 sleep only.Additionally, while the technologist was aware of the results of the first titration while performing the second titration of each night, bias was minimized by using esophageal pressure monitoring to reduce the risk of overtitration.
We did not find any cephalometric variables in our subjects that could predict ineffectiveness of oronasal CPAP or its improvement with the MAD.Changes in airway dimensions may occur in the supine position (26), which we did not assess.Cephalometric data should be interpreted with caution given our limited sample size.

SUMMARY
Although the prevalence of the difference in nasal versus oronasal CPAP effectiveness and impact on CPAP adherence are unknown, it is important to be aware of this potential source of CPAP failure.Given the increasingly recognized adverse consequences of OSA, optimizing treatment effectiveness and adherence is an important goal for each patient.Our data suggest that it may be feasible to combine therapy using a MAD with oronasal CPAP in individuals intolerant to CPAP by nasal mask.

figure 1 )
figure 1) Example of polysomnographic recordings for subject 3 -oronasal versus nasal mask.Recordings with an oronasal mask at continuous positive airway pressure (CPAP) 20 cmH 2 O (A), and with a nasal mask at CPAP 15 cmH 2 O (b), both in supine stage 2 sleep.The epoch depicted is 180 s.Inspiration on the nasal flow signal is an upward deflection

figure 3 )
figure 3) Example of polysomnographic recordings for subject 2 -oronasal mask with and without mandibular retention.Recordings with an oronasal mask alone at continuous positive airway pressure (CPAP) 20 cmH 2 O (A), and with an oronasal mask AND mandibular advancement device at CPAP 18 cmH 2 O (b), both in supine stage 2 sleep.The epoch depicted is 180 s.Inspiration on the nasal flow signal is an upward deflection

:
The authors acknowledge Dr Francois Bellemare for his help in setting up the study protocol.They extend their thanks to Magalie Valières who performed the titration studies, and to Anne-Marie Laurin, Chief Technician of the Sleep Laboratory of the Centre-Hospitalier Universitaire de Montreal.fUNDING: This study was funded by the Foundation of the Centre Hospitalier Universitaire de Montreal.DISCLOSURES: The authors have no financial disclosures or conflicts of interest to declare.
the line joining porion (Po) and orbital (Or) points Co-A Maxillary length Distance in mm between condylion (Co) and subspinal (A) points ANS-PNS Maxillary length Distance in mm between the anterior nasal spine (ANS) and the posterior nasal spine (PNS) points Co-Gn Mandibular length Distance in mm between condylion (Co) and gnation (Gn) points.SNA Antero-posterior position of the maxilla Angle formed by nasion (N), sella (between a line perpendicular to FH plane at nasion (N) point and subspinal (A) point.N-Pog ┴ FH Antero-posterior relation of the mandible Distance in mm between a line perpendicular to FH plane at nasion (N) point and pogonion (Pog) point FMA Mandibular plane angle Angle formed between the mandibular plane and FH plane Mandibular plane: Line joining gnation (Gn) and menton (Me) points APPENDIx 1 Lateral cephalometric landmarks Continued on next page

TAble 1 Subject characteristics and study results Subject Sex Age, years body mass index, kg/m 2 RDI RDI supine Peff without MAD, cmH 2 O Peff with MAD, cmH 2 O Mask used in first half of each night*
*The maximal available continuous positive airway pressure was 20 cmH 2 O; † Clinically significant improvement in oronasal effective pressure (Peff) for correction of obstructive respiratory events in supine stage 2 sleep with versus without the mandibular advancement device (MAD).RDI Respiratory disturbance index A B suggesting no direct effect of the MAD on airway obstruction.In the other subject (subject 5), nasal Peff increased by 2 cmH 2 O (Table 1).Oronasal Peff decreased in three subjects, remained >20 cmH 2 O in two and increased in one (subject 5).Two individuals, who could not be treated by oronasal mask alone (Peff > 20 cmH 2 O), achieved oronasal Peff 19 cmH 2 O and 18 cmH 2 O, respectively, with the MAD (Figure