Obstructive Sleep Apnea (OSA) Treatment & Management

Updated: Sep 15, 2020
  • Author: Himanshu Wickramasinghe, MD, MBBS; Chief Editor: Zab Mosenifar, MD, FACP, FCCP  more...
  • Print

Approach Considerations

Obstructive sleep apnea (OSA) should be diagnosed and treated promptly. Board-certified sleep specialists evaluate polysomnography (PSG) results and make treatment recommendations for OSA patients. Treatment depends in part on the severity of the sleep-disordered breathing (SDB). People with mild apnea have a wider variety of options, while people with moderate-to-severe apnea should be treated with nasal continuous positive airway pressure (CPAP).

General and behavioral measures, such as weight loss, avoidance of alcohol for 4-6 hours prior to bedtime, and sleeping on one’s side rather than on the stomach or back, are elements of conservative nonsurgical treatment. In a 2006 practice parameter, both weight loss and positional therapy were rated as “guidelines,” indicating a patient care strategy with a moderate degree of evidence. [153, 154]

Because obesity is a major predictive factor for OSA, weight reduction reduces the risk of OSA. The best data suggest that a 10% reduction in weight leads to a 26% reduction in the respiratory disturbance index (RDI). Benefits of weight reduction in patients with SDB include the following:

  • Decreased RDI

  • Lowered blood pressure

  • Improved pulmonary function and arterial blood gas values

  • Improved sleep structure and snoring

  • Possible reduction of optimum CPAP pressure required

Weight gain is one of the most important determinants of relapse of OSA after surgical treatment. Although accomplishing and maintaining weight reduction are difficult, the results are extremely beneficial when patients can do so. The treatment approach to SDB is not complete if weight reduction is not addressed in patients who are obese.

Mechanical measures include positive airway pressure with a CPAP or bilevel positive airway pressure (BiPAP) device and oral appliance (OA) therapy. CPAP is the standard treatment option for OSA and generally can reverse this condition quickly with the appropriate titration of devices.

OAs are indicated for (1) patients with mild-to-moderate OSA who prefer oral appliances to CPAP devices, (2) patients with mild-to-moderate OSA who do not respond to CPAP therapy, and (3) patients with mild-to-moderate OSA in whom treatment attempts with CPAP devices fail. They should not be considered effective therapy for patients with severe OSA.

Pharmacologic therapy is not part of primary treatment. No clinically useful drug therapy is currently available, except in certain cases of excessive sleepiness remaining after apparently successful treatment.

From least invasive and effective to most invasive and effective, treatments can be summarized as follows:

  • All patients should be offered nasal CPAP therapy first.

  • In patients with mild-to-severe obstructive sleep apnea who refuse or reject nasal CPAP therapy, BiPAP therapy should be tried next. If this therapy fails or is rejected, OA therapy should be considered.

  • OAs may be considered first-line therapy for patients with mild OSA, particularly if they are unwilling to try nasal CPAP therapy.

  • All interventions to improve tolerance of CPAP therapy should be attempted prior to deciding that treatment has failed in a particular patient.

  • Patients in whom noninvasive medical therapy (eg, CPAP, BiPAP, OAs) fails should be offered surgical options. Patients should be made aware of the success rates for each surgical procedure. They should be informed that they might require more than 1 surgical procedure, some fairly extensive, to cure OSA. Refer patients only to centers that have personnel experienced in these special surgical techniques.

Improving treatment adherence is important to the care of OSA patients. Whereas adherence in OSA patients is comparable to that in patients taking medications, such as statins, a body of research on adherence seems to have been largely ignored and needs to be integrated into sleep medicine clinical practice.

Studies showing how to improve CPAP adherence exist as well and should be integrated into a standard CPAP follow-up program to improve adherence; the same could be said for OA therapy to the degree that some of the methods and assessment are common to both treatments.

Unlike CPAP/BiPAP treatment adherence, OA treatment adherence is not objectively measured. Therefore, studies comparing adherence between OA and CPAP/BiPAP therapy cannot be considered with confidence in the outcome. As when CPAP/BiPAP did not have objective adherence data, OA treatment adherence is probably lower than published values that have relied on patient or practitioner self-report.

Sleep-related breathing disorder continuum

The concept of the sleep-related breathing disorder (SRBD) continuum (see Pathophysiology) suggests that optimal OSA treatment must correct OSA, upper airway resistance syndrome (UARS), and snoring. If it does not eliminate all 3 problems, the symptoms and the pathophysiological process that was evident at the start of disease recur. Therefore, in the treatment of SRBD, CPAP corrects OSA first, UARS next, and snoring last.

An unlikely occurrence is snoring being corrected before OSA and/or UARS; if this is thought to have occurred, then consideration should be given to the integrity of the snoring microphone.

Consider whether snoring has been correctly interpreted on PSG during a CPAP titration. When a mask leak occurs, the noise may be transferred by the microphone to the PSG snore channel and may sound like snoring. One can determine the difference between snoring and a CPAP mask leak because snoring occurs at the point of peak inspiration and the beginning of expiration; mask leak occurs during expiration.

Consider whether the patient has had upper airway (UA) corrective surgery. If pharyngeal tissue has been eliminated, snoring may not occur, but OSA can develop (so-called silent apnea).


Nasal CPAP Therapy

Initially described in 1981, nasal CPAP therapy is the most effective treatment for OSA, and it has become the standard of care for this condition. (It is also effective for treating mixed apneas and some central apneas.)

The CPAP device consists of a blower unit that produces continuous positive-pressure airflow. This airflow is usually applied at the nose and is then directed through the UA. CPAP increases the caliber of the airway in the retropalatal and retroglossal regions (see the image below). It increases the lateral dimensions of the UA and thins the lateral pharyngeal walls, which are thicker in patients with obstructive sleep apnea than in people without obstructive sleep apnea.

Top image is 3-dimensional surface renderings of t Top image is 3-dimensional surface renderings of the upper airway demonstrating the effect of progressive increases in continuous positive airway pressure (CPAP) from 0-15 cm of water on upper-airway volume in a patient with upper airway narrowing. CPAP significantly increases airway volume in the retropalatal (RP) and retroglossal (RG) regions. Bottom image is soft tissue images in the same patient in the RP region at analogous levels of CPAP. With increasing CPAP, the upper airway progressively enlarges, particularly in the lateral dimension. Note the progressive thinning of the lateral pharyngeal walls as the level of CPAP increases. Little movement occurs in the parapharyngeal fat pads, the white structures lateral to the airway. The first image in each series depicts the baseline upper airway narrowing present in this patient.

Effectively, CPAP acts as a pneumatic splint to maintain UA patency during sleep, preventing the soft tissues from collapsing. By this mechanism, it effectively eliminates the apneas and/or hypopneas, decreases the arousals, and normalizes the oxygen saturation (see the image below).

Effect of nasal continuous positive airway pressur Effect of nasal continuous positive airway pressure (CPAP) on oxygen saturation in sleep apnea. The upper portion of this figure shows the raw oxygen saturation trace from 1 night of a sleep study. Below the raw trace are vertical lines that indicate the presence of either an apnea or hypopnea. Before CPAP, frequent respiratory events with significant desaturations occurred. During the night, CPAP was applied, resulting in the elimination of the apnea and hypopneas and normalization of the oxygen trace.

Guidelines for use

Patients with severe SDB (respiratory disturbance index [RDI] >20-30) should be treated irrespective of their symptoms because of the increased risk of cardiovascular morbidity. Patients with an RDI of 5-20 should be treated if they have symptoms or coexistent cardiovascular disease. Patients with UARS may need CPAP therapy.

Medicare guidelines specify criteria for ordering CPAP for patients with OSA. All patients with an apnea-hypopnea index (AHI) greater than 15 are considered eligible for CPAP, regardless of symptomatology. For patients with an AHI of 5-14.9, CPAP is indicated only if the patient has one of the following: excessive daytime sleepiness (EDS), hypertension, or cardiovascular disease.

Most sleep center physicians still titrate CPAP during a sleep study, either as a second night of study or during the second half of a diagnostic study (ie, split-night PSG [see Workup]). Proper titration includes identifying the minimum CPAP level that abolishes obstructive apneas and/or hypopneas, oxyhemoglobin desaturation, respiratory effort–related arousals (RERAs), and snoring in all sleep stages and all sleep positions. The pressure needed is typically 5-20 cm H2 O. Guidelines for positive-pressure titration have been published. [155, 156, 157]

Currently, CPAP devices are available that automatically change pressures based on the presence and/or absence of OSA (auto–positive airway pressure, or auto-PAP). The rationale for auto-titrating devices is that the pressure required to treat OSA may vary over the course of the night and between different nights, sleep stages, and body positions, with the variations not captured by a one-night titration study.

In theory, the mean pressure delivered by auto-PAP devices is lower than that delivered with fixed CPAP; however, no studies have shown increased patient compliance with auto-PAP devices. [158] In fact, a randomized crossover study comparing fixed and variable pressure CPAP, the largest to date (N = 200), found a marginal increase in hours used per night (0.2 h) but no difference in patient preference, which actually showed an order effect (patients preferred the type of device first used in the study). [159]

Guidelines from 2008 indicate that auto-PAP devices may be used during an attended sleep study to determine a single pressure for use at home (guideline recommendation). In addition, some evidence supports use in the unattended setting to determine a single pressure for home use (option recommendation). [160]


Application of adequate levels of nasal CPAP during sleep almost always resolves obstructive apnea and/or hypopnea, oxyhemoglobin desaturation, RERAs, and snoring from sleep. It also results in adequate sleep continuity.

CPAP has been shown to improve daytime sleepiness, mood, and cognitive function in people with both mild and moderate apnea. [161, 162]

CPAP has also been shown to decrease blood pressure, primarily in patients with severe OSA. [92, 93, 91] Evidence also indicates that it may improve the left ventricular ejection fraction in patients with congestive heart failure and OSA. [99] CPAP plus an antihypertensive medication may synergistically improve systemic hypertension. [95] In addition, it improves right-side heart function and pulmonary hypertension.

A study of 86 patients with sleep apnea, including 75 who had metabolic syndrome, suggests that CPAP is associated with a lower risk for heart disease, stroke, and diabetes. Study participants were treated for 3 months with either CPAP or sham CPAP, followed by a month of no treatment and 3 additional months of the opposite treatment. Of patients treated with CPAP, 13% no longer met diagnostic criteria for metabolic syndrome, compared with 1% of patients in the sham-CPAP control group. CPAP use was also associated with significant weight loss. [163]

CPAP has also been shown to increase quality of life [164] and decrease health care costs. [165] Prospective cohort studies suggest that CPAP reduces mortality in OSA. [46, 54] The benefits parallel those observed after tracheostomy.

Although many OSA patients note an immediate improvement in alertness, concentration, and memory, achieving maximum improvement in neurocognitive symptoms may take as long as 2 months. Follow-up visits should be scheduled at least once after CPAP treatment is first started and at least yearly thereafter. Follow-up evaluation is required to ensure symptomatic improvement, CPAP adherence, and equipment maintenance.

In an attempt to determine to what extent CPAP benefits its users, the authors of a randomized controlled trial evaluated the effects of stoppping CPAP treatment. Results show a rapid recurrence of OSA and sleepiness within a few days of CPAP withdrawal. Also, study participants experienced deteriorated endothelial function and a marked increase in heart rate after 2 weeks. [166]

In one meta-analysis of three randomized placebo-controlled trials, nearly 30% of the treatment benefit among high users of CPAP was due to patients’ expected benefit of treatment due to their knowledge of hours of device use. [167]


CPAP adherence is key to patients obtaining benefits from its use. Unfortunately, adherence may be poor. Evidence indicates that many patients do not accept (or even initiate) CPAP therapy and that up to 25% do not regularly follow up with a sleep physician, [168] with most of these patients being nonadherent to therapy.

Compliance with CPAP therapy can be objectively measured. [46] Most modern CPAP devices measure both "machine-on" and "mask-on" times, with the mask-on time used to measure compliance. Compliance data are downloaded onto electronic chips from which compliance reports can be downloaded during follow-up appointments. These reports allow the physician to have real-time data on compliance, allowing him or her to immediately address problems. Examples of a compliance report are shown in the image below.

Examples of good (upper panel) and poor (lower pan Examples of good (upper panel) and poor (lower panel) compliance. In the upper panel, the patient is using continuous positive airway pressure (CPAP) most nights and generally for more than 4 hours (solid black line). In the lower panel, the patient is using CPAP infrequently and, when used, is wearing the CPAP device for less than 4 hours.

Many studies have examined how many nightly hours use is necessary for CPAP to render salutary effects. A study published in 2009 demonstrated that persons who wore their CPAP device for 5.6 hours per night experienced a slight decrease in blood pressure. [94] CPAP adherence was related to the AHI and decreased scores on the Epworth Sleepiness Scale (ESS); this effect was robust because patients with abnormal ESS scores were excluded from the study

Generally, patients have been considered compliant if they use their CPAP device more than 4 hours per night, 5-7 nights per week. However, a 2007 study questioned what is optimal CPAP usage. [169] In this study, nightly duration of use required to normalize daytime functioning was examined for 3 common functional measures. The thresholds above which further improvements were less likely were 4 hours for the ESS score, 6 hours for the multiple sleep latency test (MSLT), and 7.5 hours for the Functional Outcomes of Sleep Questionnaire.

This study illustrates that patients who are considered to have good compliance by accepted definitions may, in fact, be undertreated and that patients should be encouraged to use positive airway pressure for at least 7 hours per night.

Evidence now indicates that CPAP compliance may be determined as early as 2 weeks after CPAP is initiated. [170, 171, 172] Patients who consistently use their positive airway pressure device at 6 months were, on average, using the device for more than 2 hours more per night during the first 2 weeks and were more ready and confident to continue use. On the other hand, intermittent users were more likely to report adverse effects from the device, general discomfort, and that the device was too inconvenient.

In a study of long-term compliance, 68% of patients were using a CPAP machine at 5 years. Predictors of long-term compliance were baseline AHI and degree of sleepiness. However, the best predictor of compliance was regular use at 3 months of therapy, indicating that physicians must work to increase patient compliance early in the treatment period. [173]

CPAP adherence has often been assessed outside the context of adherence in other areas of medicine. In that context, CPAP adherence rates appear dismal.

However, in a prospective study of severe OSA patients investigating whether adherence or nonadherence to CPAP treatment predicted adherence to 3 well-known protective cardiovascular medications, CPAP adherence did not predict adherence with these medications. [123] The adherence with the medications was not surprisingly low, given what is known about adherence (81-95% adherence to the medication). Again, with severe OSA (AHI > 30) and comorbid heart disease, medication adherence was low, despite CPAP adherence.

In another study, patients who consistently refilled lipid-lowering medications were more adherent to CPAP. [174] However, being married was the most powerful predictor of adherence. Once marital status entered the regression model, CPAP adherence and medication adherence were not significant predictors of adherence. This study demonstrates the complexity of the assessment of adherence.

A systemic review and meta-analysis by Bakker and Marshall examined the use of flexible pressure delivery of PAP. [175] Although flexible pressure delivery intends to improve comfort and compliance through reduction of pressure during early exhalation, the study found that this flexible pressure modification does not significantly improve compliance with CPAP in patients with OSA.

Split-night PSG does not adversely affect short-term CPAP adherence in patients with OSA. [176] Additionally, full-night PSG performed so patients can have a full night in the sleep disorders center to adapt to CPAP has not been shown to improve adherence. [176]

An individualized approach in which reasons for poor compliance are systematically investigated is essential to improving compliance. Poorly adherent patients should be asked about mask fit, mask leak, sinus/nasal congestion, mouth breathing, and general sleep habits to determine reasons for nonadherence.

Interventions to improve compliance include (1) attendance in a group clinic with education provided, (2) group cognitive behavioral therapy provided as two 1-hour sessions that include an educational talk and video of real CPAP users, (3) written literature and weekly phone calls during the first month of use, (4) use of nasal humidification, (5) alternative mask interfaces, and (6) prompt attention to adverse effects and nasal/sinus congestion (see Complications and adverse effects). [177, 178, 179, 180, 181, 182]

Patient education with sleep specialists has been shown to improve adherence. Patients who attended the authors’ short information program showed higher daily usage and lower subjective daytime sleepiness. These results suggest that patients on CPAP therapy may benefit from education, even after a longer treatment period. [183]

Pressure-relief CPAP, a system that lowers the pressure at the onset of expiration, is hypothesized to improve adherence by reducing the uncomfortable sensation of breathing against high pressure while maintaining a patent upper airway. However, a review of the literature has not found that adherence to therapy is greater in patients using pressure-relief CPAP. [158] Despite this, the author’s center frequently prescribes pressure-relief CPAP for all patients who require a setting of greater than 10 cm water.

Although an average of 20-40% of patients do not use the prescribed therapy, some sleep disorder centers have achieved greater than 90-95% adherence rates with CPAP therapy. In the authors’ experience, regular, close, and personalized follow-up greatly enhances adherence.

Complications and adverse effects

Pressure- and airflow-related complications include a sensation of suffocation or claustrophobia, difficulty exhaling, inability to sleep, musculoskeletal chest discomfort, aerophagia, and sinus discomfort. Pneumothorax and/or pneumomediastinum (extremely rare), pneumoencephalos (isolated case report), and tympanic membrane rupture (rare) also can occur.

Patients with claustrophobia may try using nasal pillows or behavioral management. If patients feel a sensation of increased resistance to expiration, use of a CPAP unit with a ramp feature is indicated. This unit permits the patient to fall asleep with little or no pressure applied, and the pressure gradually increases to the set optimal level over a predetermined interval (usually 15-30 min). BiPAP may be used as an alternative.

Mask-related problems include skin abrasions, rash, and conjunctivitis (due to air leaks). If excessive air leaks through the mouth, patients should use a chin strap to keep their mouths closed or they should try an oronasal mask. Consider consultation with an otolaryngologist to rule out sinus dysfunction. If a poorly fitting mask causes skin breakdown and/or air leaks, patients should try masks of different sizes and/or models; a variety of interfaces are now available.

Nasal problems can include rhinorrhea, nasal congestion, epistaxis, and nasal and/or oral dryness. Nasal congestion can be treated with antihistamines and/or topical corticosteroids. Nasal dryness can be treated with topical saline sprays or humidification. If the air generated by the unit is too cold, the patient should use a heated humidifier.

Other problems include noise and spousal intolerance.

Medicare and Medicaid coverage

For Medicare and Medicaid patients, regulations state that the coverage of CPAP is initially limited to a 12-week period for beneficiaries diagnosed with OSA as determined Centers for Medicare and Medicaid Services (CMS) criteria. CPAP is subsequently covered for those beneficiaries diagnosed with OSA whose OSA improved as a result of CPAP treatment during this 12-week period.

CPAP for adults is covered by Medicare and Medicaid when OSA is diagnosed using a clinical evaluation and findings from one of the following assessments are positive:

  • PSG performed in a sleep laboratory

  • Unattended home sleep monitoring device of Type II

  • Unattended home sleep monitoring device of Type III

  • Unattended home sleep monitoring device of Type IV, measuring at least 3 channels

For more information, see Decision Memo for Continuous Positive Airway Pressure (CPAP) Therapy for Obstructive Sleep Apnea (OSA) (CAG-00093R2).

Go to Obstructive Sleep Apnea, Home Sleep Monitoring for complete information on this topic.


BiPAP Therapy

Some patients require the use of BiPAP. In contrast to CPAP, which delivers a constant pressure during both inspiration and expiration, BiPAP permits independent adjustment of the pressures delivered during inspiration and expiration. The levels are set so that the expiratory positive airway pressure eliminates apneas and the inspiratory positive airway pressure eliminates hypopneas.

The ability to adjust inspiratory and expiratory pressures independently results in lower mean airway pressures compared with those of CPAP. In a given patient, the expiratory positive airway pressure level that must be applied is lower than the corresponding CPAP level required to maintain airway patency.

BiPAP is generally used in patients who cannot tolerate high CPAP pressures (ie, patients who experience difficult exhalations) or who have barotrauma complications (eg, ear infections, bloating). Many laboratories automatically place a patient on BPAP if the CPAP level needs to be increased above 15 cm water. However, BiPAP is too expensive to be used as first-line therapy, and it has no distinct advantages over CPAP therapy.

Compliance with BPAP has not been demonstrated to be better than compliance with CPAP. [184, 185]


Oral Appliance Therapy

OAs act by moving (pulling) the tongue forward or by moving the mandible and soft palate anteriorly, enlarging the posterior airspace. They open or dilate the airway.

Newer designs have separate upper and lower parts that are attached to each other and that allow for adjustability and jaw mobility. A minimum percentage of protrusion to effectively treat OSA is 6-10 mm or up to 75% of the maximum protrusion the patient is capable of performing upon request at the initial examination. This protrusion advance distance is necessary for the OA to be effective. The more protrusion gained, the lower the AHI at the treatment assessment time point.

Multiple different devices are available on the market. At present, the 3 basic designs of OAs used to treat sleep-related breathing disorders (SRBDs) are mandibular repositioners, tongue-retaining devices (TRDs), and palatal-lifting devices. More than 40 OAs are available to manage SRBD and obstructive sleep apnea.

Go to Oral Appliances in Snoring and Obstructive Sleep Apnea for complete information on this topic.

Guidelines for use

The American Academy of Sleep Medicine (AASM) has published practice parameters and a review of the use of OAs in persons with OSA. [186, 187] These parameters include the following recommendations:

  • The presence or absence of OSA must be determined before treatment with OAs is started in order to identify patients at risk because of complications of sleep apnea and to provide a baseline to establish the effectiveness of subsequent treatment.

  • For patients with OSA, the desired outcome of treatment includes resolution of the clinical signs and symptoms of OSA and normalization of the patient’s AHI and oxyhemoglobin saturation.

  • Although OA therapy is not as effective as CPAP, OAs are indicated for use in patients with mild-to-moderate OSA who prefer OAs to CPAP, those whose condition does not respond to CPAP, those who are not appropriate candidates for CPAP, and those in whom attempted CPAP or behavioral measures (eg, weight loss, changing sleeping positions) fail.

  • Patients with severe OSA should receive an initial trial of nasal CPAP because CPAP is more effective than OA therapy. UA surgery may also supersede the use of OAs in patients for whom these operations are predicted to be highly effective in treating sleep apnea.

  • To ensure satisfactory therapeutic benefit from OAs, patients with OSA should undergo PSG or an attended cardiorespiratory (type 3) sleep study with the OA in place after final adjustment of fit is performed.

According to the standards and guidelines listed above, the major role for OA therapy appears to be the treatment of patients with mild-to-moderate OSA who cannot tolerate CPAP (and BiPAP) therapy. These devices are relatively unlikely to benefit patients with severe OSA. Clinicians and patients prefer a titratable device, such as a mandibular repositioner, because it can be adjusted to improve both effectiveness and comfort.

Patients should receive a complete evaluation by a sleep disorders specialist and a dental professional, both of whom should be experienced in OA therapy; their close collaboration is required. Follow-up PSG after final fitting helps confirm that OA therapy is treating OSA adequately; no other method is available for this purpose (the device should not be titrated during the course of an assessment study). [188] . OA devices may resolve snoring without adequately treating OSA.

Remember that the AHI may increase with OA treatment. A change in AHI may be due to weight change between the first study and the final OA titration, not always due to the OA appliance itself. Also recognize that the AHI may vary from night to night because of the degree of severity of the obstructive sleep apnea itself; persons with mild-to-moderate obstructive sleep apnea have higher AHI variability compared with those who have more severe obstructive sleep apnea, and OA therapy is not typically used in patients with more severe obstructive sleep apnea.

Night-to-night variability in supine sleep position also occurs, and supine is the position in which OA has been thought to render its greatest benefit, including rapid eye movement (REM) sleep percentage. REM sleep percentage can increase with adaptation to the testing procedure, and it can increase if a patient is taking a REM-suppressant medication that was discontinued (eg, selective serotonin reuptake inhibitors). All these factors, among others, can account for changes in the AHI.

Contraindications for OA treatment include the following:

  • Less than 6-10 teeth in each arch

  • Patient unable to protrude the mandible forward and open the jaw widely

  • Preexisting temporomandibular joint problems

  • Severe bruxism

  • Patient with full dentures (cannot use a mandibular repositioner but could use a TRD)


Multiple small cohort studies have shown that OAs effectively lower the AHI and improve overnight sleep architecture.

The biomechanical factors responsible for the effectiveness of OAs are not completely understood. Increased tone of UA musculature is thought to be the predominant influence on the caliber and volume of the airway. Key among these muscles is the genioglossus muscle. When the jaw is opened, the genioglossus (with other muscles in the pharyngeal airway) influences the airway to improve its caliber and stability.

OAs thin the lateral pharyngeal walls by exerting traction. According to imaging studies, the size of the lateral pharyngeal fat pads and the thickness of the lateral pharyngeal muscular walls are greater in patients with apnea than in healthy subjects. For OA therapy to be successful, the lateral dimension of the airway may be critical; however, data suggest that total airway space, measured as a 3-dimensional image using cone-beam computed tomography (CT) scanning, may be the most powerful predictor for OSA severity.

The ideal treatment goal for OA therapy, as with CPAP treatment, is an AHI of less than 5 and snoring resolution. The OA literature has generally regarded an AHI less than 10 as a treatment success. However, an AHI of greater than 5 is associated with adverse health consequences; therefore, an AHI of less than 5 without snoring should be the ideal treatment goal.

The 4 key variables that contribute to OA treatment efficacy are (1) mild-to-moderate OSA (AHI < 30), (2) mandibular repositioner protrusion distance greater than 70% of baseline, (3) higher supine AHI relative to lateral sleep position AHI, and (4) a lower body mass index (BMI).

Treatment success with mandibular repositioners (and with OAs in general) appears to be inversely related to the initial RDI. A growing body of evidence now suggests that the severity of OSA is predictive of the response to OAs.

Most studies that have examined the efficacy of OAs have used mandibular repositioners. However, a 2009 study investigated the effect of TRDs on OSA severity. [189] If ideal treatment is an AHI lower than 5 and no snoring, few patients met that goal. If ideal treatment is an AHI lower than 10, then 31% met that goal. Response variation was wide, with a standard deviation for AHI of 19.5. Two predictors of response were found using multiple regression: age and protrusion distance (mean protrusion, approximately 7±1.5 mm).

A review of the literature by the American Sleep Disorders Association (ASDA) indicated the following findings [188] :

  • Overall, 51% of patients studied achieved an RDI of less than 10 with OA therapy.

  • Of patients who had a pretreatment RDI of greater than 20, 39% continued to have an RDI above this level despite OA therapy. At least 1 randomized controlled trial demonstrated that OAs have better success rates in patients with mild OSA (81%) than in those with moderate (60%) or severe (25%) OSA.

  • Continuous adjustment or replacement, as needed, improves success rates with OAs in the long term.

  • No patient characteristics predicted success with OA therapy.

  • No particular OA had any advantages over the others studied.

  • Some patients have an increase in AHI with OA treatment.

  • Endpoints assessed in the studies of OAs varied and included an RDI of less than 10, an RDI of less than 20, or a greater than 50% reduction in the AHI. This variation made the comparison of results difficult. Furthermore, many studies did not stratify patients by severity of OSA.

  • OAs were more likely to be successful in patients with low BMIs, at a young age, with a small neck circumference, with a short soft palate, or with a small oropharynx and in treating positional OSA, as based on retrospective data analysis.

Subsequent prospective controlled clinical trials to compare OA therapy with nasal CPAP therapy to treat OSA and snoring demonstrated the following results:

  • Treatment was successful in 55% of patients using an adjustable mandibular repositioner and in 48% of patients using a nonadjustable mandibular repositioner; treatment success was defined as a posttreatment RDI of less than or equal to 10, with symptomatic relief

  • CPAP therapy was superior to mandibular repositioning in normalizing the RDI, reducing snoring, and improving oxygenation

  • CPAP and adjustable mandibular repositioning equally improved daytime sleepiness

  • Most patients preferred mandibular repositioning therapy to CPAP therapy

In general, comparative studies show that CPAP is more effective than oral appliances in lowering the AHI to less than 5-10 events per hour. [190] However, many other outcomes, such as sleepiness and cognitive functioning, are not different between the 2 devices. When asked which device the participant would use at home, the responses varied; in some studies, the participants favored CPAP and in others, they favored the oral appliance.

In a trial comparing health effects after 1 month each of CPAP and mandibular advancement device (MAD) therapy in 126 patients with moderate-to-severe OSA, the 2 treatments yielded comparable improvements in neurobehavioral outcomes and disease-specific quality-of-life outcomes; neither improved blood pressure. CPAP was more effective in reducing the apnea-hypopnea index, but the difference appeared to be offset by a higher treatment compliance in the MAD group. MAD was also more effective in improving 4 general quality-of-life domains. [191]

In one study, OAs were more effective than uvulopalatopharyngoplasty (UPPP) in the treatment of OSA. OAs may also be useful in managing OSAS if surgery fails.

The use of OAs in clinical practice is limited because of the difficulty in predicting the therapeutic response of individual patients. Tsai et al used a remote-controlled device to titrate OA treatment during a single-night sleep study to predict the therapeutic response. [192] In concept, this approach is similar to titrating nasal CPAP during a single-night sleep study.

Raphaelson et al [193] and Petelle et al [194] first demonstrated the titration of mandibular advancement during a sleep study. Petelle et al demonstrated that determining the optimum level of mandibular advancement required for an individual patient during a single-night study is possible. Of note, Tsai et al did not report the same.

Apart from raising the possibility of predicting therapeutic responses in individual patients, this titration approach potentially provides an opportunity to determine the optimum therapeutic dose of mandibular advancement required during a single-night sleep study.

Further work is required in this area because it could greatly affect the use of OAs in SRBDs. This research may yield the method required to identify patients who may respond to OAs and to determine the optimum level of advancement required for an individual patient.


Adherence rates are not as well defined with OAs as they are with CPAP, because CPAP units have internal computerized recorders that capture data that can give feedback regarding hours of use each night. Most studies to date have used subjective measurement tools to determine OA adherence percentage. Unfortunately, this is a flawed approach, because self-report is unreliable; moreover, when adherence is compared with CPAP (when objective adherence has been measured), the outcome is meaningless.

Given that caveat, available evidence suggests that OA adherence is not high. Most studies have shown that fewer than 50% of patients use their appliance over time, and adherence is not universally defined. If self-report of use is inflated, as it typically is, actual adherence with OA therapy may be much lower.

In one study, only 52% of patients stated that they were using the TRD after 5 years. Of those, 79% wore the TRD more than 4 nights per week. Among nonusers, 47% went back to CPAP, 12% tried another type of OA and were satisfied, and 41% remained untreated.

This subanalysis of what patients who were noncompliant did is informative and instructive for other studies. First, it shows that CPAP may be preferred after another therapy has been tried over time, and, as mentioned, preference is not synonymous with adherence. Second, patients who do not adhere to one treatment do not automatically seek another treatment. Hence, the use of a TRD is not benign, because 41% of those who did not continue TRD treatment did not receive OSA treatment at all for 1-36 months.

This report by the authors was enlightening, because many studies do not go on to ask what alternatives were sought by the patient. Not seeking other treatment is a very important follow-up question to ask in nonresponders to any OSA treatment.

Adherence declines over time and is thought to be largely due to temporomandibular joint (TMJ) problems. Median adherence over the first year, for those who accepted OA treatment, has been shown to be 77%.

Each oral appliance device (>40 devices available) can have a unique adverse effect that leads to nonadherence. Given the literature, TMJ, occlusal changes, excessive salivation, and discomfort are likely to be chief among the reasons for nonadherence to effective OA treatment. As with all treatment, ineffectiveness is among the most prominent reasons patients do not adhere to treatment.

To assess the above mentioned concerns of OA treatment, OSA patients treated with an OA need to return to determine optimal fit and then at 6 months, 1 year, and annually thereafter. This standard is not different from good medical care for any sleep medical treatment (eg, prescription medicine, weight loss, CPAP, surgery).

Multiple PSG studies may be required to document a therapeutic response. Home studies seem indicated in these cases, with final verification using a full PSG in the sleep center.

Long-term follow up is very important in OA therapy. Insufficient data exist to determine if damage to teeth and joints occur in patients using OA therapy.

As noted, adherence data for OA treatment derive from subjective reports, whereas CPAP adherence can be monitored objectively. The development of objective measures of OA treatment would be of benefit in the future for patients and so that assessment of the adherence can be meaningfully compared between OA treatment, CPAP, treatment, and surgical correction over time.

Complications and adverse effects

Complications or adverse effects include excessive salivation, dental misalignment with bite change, and tooth movement. Additionally, OAs may aggravate or cause adverse occlusal changes (as early as 6 mo into treatment), TMJ disease, myofascial pain, tooth pain, gagging, gum irritation, salivation, TMJ sounds and morning-after occlusal changes, and/or tongue pain (most common with TRDs—from 67-68% of patients reported one or more of these adverse effects, and the adverse effects were noted to occur at any time during treatment).

Patients may also object to having an appliance in their mouth throughout the night.

Insurance coverage and cost

Cost may be a barrier to OA therapy. Although not all dental practitioners charge the same fee, common amounts range from $300 to more than $2,500.

The lack of long-term studies with OA therapy may limit the clinician from choosing it as an option. Unlike CPAP or surgery, OA therapy is often not covered by medical insurance plans. Check with individual insurance carriers to verify coverage.

The ASDA states, “Economic assessment, focusing on both short- and long-range costs (inclusive of needed follow-up and indirect costs of OA therapy) is needed so that OA therapy can be compared with alternate therapies through cost and effectiveness analyses. Research is needed to clarify design characteristics most beneficial in given patient groups, so that device selection is driven by data that are more precise.” [188]


Surgical Correction of the Upper Airway

Surgical correction of the upper airway (UA) is still performed but is not considered primary therapy for OSA. The theoretical advantage of surgery is that if the patient is cured, compliance with CPAP or OA therapy is no longer an issue. However, a primary reason why surgery has not become a standard therapy is the lack of any long-term outcome studies showing that surgical correction continues to be effective 5 or more years after it is performed.

Surgical care for OSA patients should not be seen as a "last ditch" effort in treatment of OSA patients. Surgery may be initial therapy for patients with mild OSA (RDI < 20, lowest oxyhemoglobin saturation >90%) if medical therapy is refused or rejected and if the patients are medically stable enough to undergo the procedure. Surgery may be indicated if noninvasive medical therapy, nasal CPAP, or OA fails to effectively treat OSA or is rejected by the patient.

Surgery is also indicated in patients who have a specific underlying abnormality that is causing the OSA. Three of 200 adults with OSA have a specific space-occupying lesion that causes an UA obstruction. Although surgical correction of such an abnormality (ie, tonsillectomy) can potentially cure OSA, most adult patients do not have such correctible lesions.

If the patient opts for surgery, ensure the following:

  • Surgery should be performed by a qualified ear, nose, and throat surgeon.

  • Surgery should be based on the location of collapse.

  • The patient should be willing to undergo combination or multiple surgeries.

In some patients, tracheostomy or CPAP therapy is required in the perioperative period to ensure a safe airway.

For additional reading on surgery in OSA, the reader is referred to a 2005 pro-con debate in the Journal of Clinical Sleep Medicine by Powell ("Upper Airway Surgery Does Have a Major Role in the Treatment of Obstructive Sleep Apnea: ‘The Tail End of the Dog’") and Phillips ("Upper Airway Surgery Does Not Have a Major Role in the Treatment of Sleep Apnea"). [195, 196]

Go to Surgical Approach to Snoring and Obstructive Sleep Apnea for complete information on this topic.

Anatomic considerations

Functional division of the pharynx into the retropalatal-oropharyngeal region (posterior to the soft palate) and the retrolingual-hypopharyngeal region (posterior to the vertical portion of the tongue) has been proposed. Obstruction in patients with SDB is classified into 3 types according to region. Type I is obstruction in only the retropalatal region. Type II is obstruction in both the retropalatal and retrolingual regions. Type III is obstruction in only the retrolingual region.

Choice of surgical procedure

A full listing of possible surgical procedures for OSA is available elsewhere. [197] Commonly performed procedures include the following:

  • UPPP is resection of the uvula and soft palate. It is effective in approximately 40% of patients, but predicting which patients will benefit from the procedure is problematic. Long term, patients with treatment success often present with a recurrence of symptoms, especially if they continue to gain weight.

  • Craniofacial reconstruction involves advancement of the tongue (ie, genioglossus advancement with hyoid myotomy [GAHM]) or the maxillomandibular bones (ie, maxillomandibular osteotomy [MMO]). These operations should be performed only at centers with expert personnel. Short-term success rates are approximately 70% for GAHM and 95% for MMO. No good long-term studies have been performed to evaluate the success for either GAHM or MMO.

  • Tracheostomy provides definitive correction because it bypasses the obstruction. It is recommended for patients with very severe OSA, especially if the patient does not tolerate CPAP or has cor pulmonale.

Different surgical procedures have been proposed for patients with different levels of obstruction. UPPP may correct type I obstruction. GAHM may correct type III obstruction. MMO may correct obstruction at all levels.

AASM recommendations for surgical treatment of OSA

The ASDA published guidelines for the surgical treatment of OSA in 1995. [198]

The current position of the AASM on surgical therapy for OSA states the following:

  • Surgery is indicated only if noninvasive types of therapy have not worked or are rejected by the patient.

  • Tracheostomy is the only operation shown to be nearly 100% effective as a sole procedure for OSA.

  • UPPP is indicated for retropalatal obstruction.

  • For retrolingual (types II and III), mandibular maxillary advancement is the most promising (>90%) surgical approach.

Riley-Powell-Stanford surgical protocol

Because several sites of obstruction may be responsible, a systematic approach for selecting surgery has been developed. [199, 200] This is the Riley-Powell-Stanford surgical protocol designed in 1988. The protocol has 2 phases. Phase I consists of the UPPP and GAHM procedures, and phase II consists of the more complicated MMO procedure. Patients who are not adequately treated with phase I surgery are offered phase II surgery.

For phase I surgery, perform UPPP for patients with type I obstruction, GAHM for patients with type III obstruction, and simultaneous UPPP and GAHM for patients with type II obstruction. The overall success rate for phase I surgery is approximately 61%, although patients with severe OSA (RDI >60, lowest oxyhemoglobin saturation < 70%) have a success rate of only 42%.

Phase II surgery consists of MMO, in which the jaw is advanced anteriorly. With the phased protocol, the success rate has been in excess of 90% for phase II surgery.


UPPP is the most common surgical procedure performed for adults with OSA. It involves removal of the tonsils (if present), the uvula, the distal margin of the soft palate, and the redundant pharyngeal tissue, as well as reshaping of the soft tissues in the lateral pharyngeal walls.

AASM recommendations for UPPP state that "UPPP, with or without tonsillectomy, may be appropriate for patients with narrowing or collapse in the retropalatal region. Good preoperative evaluation does not guarantee surgical success; the effectiveness of the UPPP is variable, and the procedure should be considered when nonsurgical treatment options, such as CPAP, have been considered." It is important to note that enlarged adenoids and tonsils, in the adult OSA patient, are rarely a singular cause of OSA. As such, it is not a recommended surgery alone in the adult OSA patient.

The surgical success rate is not high with UPPP alone, approximately 50% when surgical success is defined as both 50% reduction in RDI and/or apnea index, and a postoperative RDI of less than 20 (or apnea index < 10). This rate is despite preselection of patients with type I obstruction by using imaging and endoscopic studies. This finding highlights the inadequacy of the methods available to identify sites of UA obstruction. This success rate, when compared with the potential complications of the surgery, has served to make the UPPP alone a rarely recommended procedure in the author’s center.

Complications from UPPP should be explained to patients in detail, as with all procedures; however, this is particularly true for those patients who use their voice to make a professional living (eg, singers, muscicians who play instruments that require maneuvers involving the areas where tissue will be removed).

Complications can include the following:

  • Change in voice, singing, or the playing of musical instruments that require changes in the soft palate to play

  • Pain with swallowing and pain with speech, usually for 1-2 weeks postoperatively

  • Hemorrhage (2-4%)

  • Swallowing difficulties, particularly regurgitation of food

  • Long-term pharyngeal discomfort

  • Disturbance in taste

  • Numbness of tongue

  • Nasopharyngeal stenosis

  • Creation of silent apnea

Silent apnea may result post-UPPP. Silent apnea refers to a condition in which the vibration of the tissues that caused snoring during airway collapse remains; thus, OSA persists but snoring does not. The decision to reevaluate OSA postsurgically should not depend on a postive report of snoring. If the patient still snores, however, this indicates that the surgery was not curative of obstructive sleep-related breathing disturbance (see the image below).

Sleep-related disordered breathing continuum rangi Sleep-related disordered breathing continuum ranging from simple snoring to obstructive sleep apnea (OSA). Upper airway resistance syndrome (UARS) occupies an intermediate position between these extremes. Note areas of overlap among the conditions.

Two studies showed that UPPP may make OSA worse, as it did in 31% of the patient population studied. Previous UPPP reduces the maximal level of pressure that patients who require CPAP therapy can tolerate. It may also compromise subsequent CPAP therapy by promoting mouth leaking. Uvulopalatopharyngoglossoplasty (UPPPG) is a modified UPPP with limited resection of the base of the tongue in which both the retropalatal and retrolingual regions of the UA are enlarged.

Genioglossus advancement with hyoid myotomy

In GAHM, the genioglossus muscle is repositioned anteriorly through an inferior mandibular osteotomy (genioglossus advancement). This maneuver places the pharyngeal muscles and the base of the tongue on tension and expands the airway. The hyoid is suspended to the superior edge of the larynx and fixed in this position, adding to the effect of genioglossus advancement.

Maxillomandibular osteotomy

In MMO, the midface, palate, and mandible are moved forward, increasing the space behind the tongue and increasing tension on the genioglossus muscle. This operation is more extensive than any of the others described. It is usually reserved for patients in whom other treatment modalities fail.


Tracheostomy bypasses the UA and is the most effective surgical procedure for treatment of OSA; it is virtually 100% effective. Unfortunately, tracheostomy is a disfiguring procedure and decreases the patient’s quality of life. Tracheostomy is now reserved for patients with severe OSA in whom other medical and surgical treatment modalities fail. Tracheostomy is also used for airway protection during UA reconstructive surgery.

Other surgical options

Laser-assisted uvulopalatoplasty is successful for reducing snoring in 90% of patients, but the success rate in patients with SDB is not clear. It may cause more scarring than UPPP, and it could potentially worsen apnea. Worsened OSA has been observed in the early postoperative period after laser-assisted uvulopalatoplasty. Laser-assisted uvulopalatoplasty is not recommended for the treatment of OSA until further data are available.

Laser midline glossectomy and lingualplasty are performed to enlarge the retrolingual region by using a laser to remove a portion of the posterior aspect of the tongue. The role of these procedures in the management of SDB has yet to be defined.

Nasal surgery includes septoplasty, turbinectomy, and polypectomy and may be useful as an adjunct to other procedures or to improve CPAP adherence. Nasal surgery by itself is rarely effective for the treatment of OSA.

Radiofrequency volumetric tissue reduction of soft palate

Recent interest has been generated in a new technique, pioneered by Powell and associates, [201] in which radiofrequency (RF) energy is used to ablate the soft palate. The US Food and Drug Administration (FDA) has approved this procedure for the treatment of snoring and OSA.

A midline soft palate submucosal scar is created by using a needle electrode inserted near the border of the hard palate and directing it toward the uvula. Pulses of RF energy are delivered, resulting in tissue necrosis and needle-tract fibrosis over subsequent weeks to months.

A study of 22 patients with mild SDB demonstrated reduced palatal tissue volume and improved symptoms in all subjects. [202] However, no data are available regarding improvement of RDI and oxyhemoglobin saturation. Follow-up over 12-18 months revealed that approximately 41% of patients who underwent RF volumetric reduction of the soft palate developed recurrent snoring. Evidence showed postsurgical improvement in the severity of esophageal pressure swings, indicating that this treatment may be useful in patients with UARS.

One study of RF volumetric reduction of the soft palate in 12 patients demonstrated success in treating snoring, but data regarding adequate treatment of SDB are lacking. [203] Data from large controlled studies are required before this technique can be recommended for the treatment of SDB. RF volumetric reduction appears to decrease morbidity compared with UPPP, laser-assisted uvulopalatoplasty, and lingualplasty.

Finally, animal studies of RF volumetric reduction of the tongue have shown volume reduction in tongue tissue after treatment. Results of human studies are pending.

Bariatric surgery

Bariatric surgery as therapy for OSA has been investigated in several nonrandomized, uncontrolled studies, with most showing a decrease in the AHI with weight loss. The 2006 practice parameter on medical therapies for OSAHS lists bariatric surgery as an option for OSA, although with limited evidence. [153]

Unfortunately, studies have not examined presurgery and postsurgery RDI or AHI using a design in which subjects were randomly assigned to a weight-loss surgery group and followed over time. At best, improvements in clinical symptoms have occurred after weight-loss surgery; however, again because no PSG data were available before or after weight-loss surgery and because subjects could not be randomly assigned to a weight-loss surgery group or control group, these data suggest that weight loss reduces OSA symptoms.

It is unknown whether this is a placebo effect, as objective reduction in the RDI or AHI as determined by the criterion standard of PSG, was not permitted by the human subjects committee. [204]

In a position paper on the treatment of OSA, the AASM stated that "[b]ariatric surgery should be considered as an adjunct to less invasive and rapidly active first-line therapies such as [positive airway pressure] for patients who have OSA and meet the currently published guidelines for bariatric surgery (Consensus). The remission rate for OSA two years after bariatric surgery, related to the amount of weight lost, is 40%, emphasizing the need for ongoing clinical follow-up of these patients." [205]

Guidelines for bariatric surgery have been published by the Society of American Gastrointestinal and Endoscope Surgeons (SAGES). [206]

Perioperative management of OSA is of special concern to anesthesiologists, and best practice papers have been published to address these concerns. [207]


Factors that increase the likelihood of successful surgery include (1) lower AHI, (2) lower BMI, (3) the location of collapse (surgeries targeted specifically to collapse at either the nasopharynx or oropharynx improve outcome), (4) the degree of mandibular protrusion (better outcomes are achieved in patients with clear deficiencies), and (5) the presence of fewer comorbidities.

The success of surgical procedures for OSA depends on accurate identification of the site of obstruction in the UA. Modalities available for identifying the site of obstruction include lateral cephalometry, endoscopy, fluoroscopy, computed tomography (CT) scanning, and magnetic resonance imaging (MRI). The accuracy of these methods in identifying the sites of obstruction is not clear. Success rates for UPPP are only approximately 50% despite preselection of patients with type I obstruction.

Data regarding surgical therapy for OSA are mainly from case series. The phased protocol of Riley-Powell-Stanford holds promise for achieving cure in patients with OSA, but further data from controlled clinical trials are needed to decide its role in the overall management of OSA.

The success rates quoted are from select centers with surgeons highly skilled in these special procedures. These results cannot be extrapolated to the general population of patients with OSA. All patients undergoing surgery for treatment of OSA should undergo follow-up PSG.

Several studies have compared surgery with both CPAP and OAs. In one randomized study comparing temperature-controlled RF tissue ablation of the tongue with CPAP, CPAP was more effective at lowering the AHI but no differences were noted in functional outcome measures such as sleepiness at approximately 1 month of treatment. [208]

In a retrospective cohort study, UPPP was found to be associated with decreased mortality compared with CPAP [209] over a 3- to 5-year period of follow-up. Additionally, one long-term outcome study compared UPPP with OA therapy. In this study, 63% of the oral appliance subjects had a normal AHI at 4-year follow-up, compared with only 33% for the UPPP subjects. [210]


Management of Residual Sleepiness Despite Apparently Effective Treatment

Residual excessive daytime sleepiness (EDS) after apparently effective treatment with CPAP is a commonly encountered problem. An exact percentage of patients has not been determined but has been estimated at 5% in one study. [211] Residual EDS is generally considered present if the ESS score remains higher than 10 after treatment.

Some of these patients may not have been "optimally" treated; rather, they were put on a positive pressure setting which, during the sleep disorders center positive airway pressure study, is sufficient to correct OSA. They may not be optimally treated for a number of reasons that need to be considered first, before placing patients in this subgroup of patients.

Reasons for suboptimal OSA treatment may include the following:

  • Positive airway pressure titration may be insufficient to fully treat OSA and snoring. Research suggests that pressures used to correct OSA and snoring are typically lower than those needed to abolish OSA and snoring. [1] Consider empirically increasing the expiratory pressure by 2 cm H2 O.

  • Once OSA and snoring are treated, patients may be noted to have complex sleep apnea (ie, residual central sleep apneas that do not resolve spontaneously). Complex sleep apnea is beyond the scope of this article.

  • Patients may not be using the positive pressure machine at home sufficiently to obtain adequate benefit to abolish sleepiness during waking hours.

  • A change in medication may have occurred that may decrease arousal from sleep (eg, benzodiazepines), use of alcohol may have changed, or a medication that suppressed rapid eye movement (REM) sleep percentage was discontinued on the titration night, but once stopped, an increase in REM sleep percentage may be correlated with OSA exacerbation (OSA is typically worse in REM sleep).

  • Sildenafil (Viagra) may be increasing the severity of OSA. Sildenafil prolongs the action of cyclic guanosine monophosphate (cGMP) and nitric oxide by inhibiting cGMP-specific phosphodiesterase 5. Nitric oxide promotes upper airway congestion, muscle relaxation, and pulmonary vasodilation and may be the mechanism by which sildenafil exacerbates OSA. [212]

  • A medical disorder that causes excessive sleepiness (eg, hypothyroidism, even if subclinical) may be present.

  • Weight gain may have occurred.

  • The patient may not be complying or may only be partially complying with their CPAP (or BiPAP) device usage. Only rarely do new CPAP devices not have cards that stores data regarding daily positive airway pressure use. All patients in the author’s sleep clinic are encouraged to bring their device with them so the clinicians can download data from their CPAP machine’s data card. Using the device properly and routinely (eg, >5 h per night, >90% of the time) seems necessary to reduce or correct sleepiness.

  • Another sleep disorder known to have hypersomnia as a major presenting symptom (eg, insufficient sleep syndrome, narcolepsy) may be present. Insufficient sleep syndrome is the most common cause of hypersomnia, and OSA is more common among patients who have narcolepsy (a 30% incidence rate vs 1-4% in the population). Consider a diagnosis of sleepiness in addition to OSA (eg, narcolepsy).

If residual sleepiness after treatment is present, a series of questions must be asked to determine its etiology. The residual sleepiness should never be considered idiopathic until all questions have been answered satisfactorily. The algorithm used by the author can be found in the image below and the questions that follow it.

Approach to a patient with excessive daytime sleep Approach to a patient with excessive daytime sleepiness after treatment with nasal continuous positive airway pressure.

See the list below:

  • Question 1: Is the patient compliant with CPAP? Compliance has been generally defined in the modafinil studies as greater than 4 hours of use, at least 5 nights per week. Patient self-reporting may not be adequate to determine compliance. Objective monitoring with one of the newer CPAP units that measures compliance may be necessary. If identified, work with the patient to improve compliance.

  • Question 2: Is the titration pressure adequate? This is particularly important in 2 situations. The first is if the patient had a difficult first titration, resulting in a prescription that may not be adequate to relieve symptoms. The second is if the patient develops symptoms months to years after initial relief, generally due to weight gain. In both cases, repeat titration is recommended. Some authorities advocate autotitration CPAP units in these situations.

  • Question 3: Is the patient sleeping sufficiently or performing shift work? Chronic insufficient sleep and shift work are 2 of the more common causes of sleepiness. In both cases, the patient must be counseled to sleep more.

  • Question 4: Is the patient on any medications that cause sleepiness. The most common medications causing this are the over-the-counter antihistamines. Other common agents are tricyclic antidepressants, benzodiazepines, narcotics, and many neurologic and psychiatric medications.

  • Question 5: Does the patient have evidence of another sleep disorder that could cause sleepiness? A full evaluation by a sleep physician can often determine a second sleep disorder, such as narcolepsy or periodic limb movement disorder.

If the above-mentioned potential causes of excessive sleepiness have been excluded, among others that the author may not have considered herein, use of stimulants to treat excessive sleepiness is indicated.

Several medications may be considered for treatment of residual sleepiness: the stimulants modafinil and armodafinil, or solriamfetol, a dopamine/norepinephrine reuptake inhibitor (DNRI). These agents are approved by the FDA for treating residual sleepiness despite optimal treatment of OSA using positive airway pressure therapy.

Modafinil seems most effective when used at the higher dose of 400 mg/d, whereas fatigue seems to be better treated with lower doses of the medication (100-200 mg/d).

If modafinil does not help at higher doses, the authors then consider armodafinil. Armodafinil reaches a peak plasma level nearly as quickly as modafinil and has a larger area under the curve for a given dose; additionally, the duration of action of armodafinil continues at the higher dosage throughout the day. Because of this action, armodafinil has more potency and most often requires only once-a-day dosing (taken in the morning for daytime workers and in the evening for nighttime workers).

FDA approval of modafinil was based on several studies carried out in this patient population. [213, 214, 215] The largest of these studies was a double-blind, randomized, placebo-controlled study in which subjects received either placebo or modafinil (200 mg/d in week 1, 400 mg/d in weeks 2-4) for 4 weeks. [214] . Subjects had an AHI of 15 or more, ESS of 10 or more, and CPAP use of 4 hours per night or more or 5-7 nights or more during 3 weeks of home monitoring.

One hundred fifty-seven patients were randomized (77 modafinil, 80 placebo), with 143 completing the study (66 modafinil, 77 placebo). The primary efficacy measures were the Epworth Sleepiness Score (ESS), daytime sleep latency based on multiple sleep latency test (MSLT) results, and CPAP use. Both the ESS and daytime sleep latency improved in the modafinil group. No difference was noted in CPAP use between groups.

The effectiveness of modafinil in this clinical situation has been confirmed by a randomized, double-blind study of modafinil versus placebo for 12 weeks, which showed that the efficacy of modafinil, as measured subjectively by the ESS and objectively by the maintenance of wakefulness test (MWT), does not change over an extended period (12 wk). [79]

One concern clinicians share with the use of modafinil (or other stimulants) in the management of OSA is whether improvement in alertness with the use of these agents may lead to noncompliance with CPAP therapy. This is an important concern, because stimulants do not control the SDB, resulting in worsening symptom control and potentially increasing the risk of cardiovascular morbidity.

Fortunately, this does not appear to be a major problem at present. The 2 largest trials did not show a decrease in CPAP usage with modafinil. [214, 79] However, decreased CPAP usage was noted in the smaller randomized study of modafinil [213] and in an open-label extension trial. [215] Therefore, if a patient is prescribed a stimulant, compliance with CPAP must be closely monitored.

More data are needed to increase assurance that use of stimulants does not lead to noncompliance. In the authors’ experience, it is unlikely that stimulant use is sufficient to correct the severe sleepiness OSA patients experience; hence, both positive airway pressure treatment and stimulant use are preferred by patients.

Approval of solriamfetol for excessive daytime sleepiness in adults with OSA was based on the TONES 3 randomized controlled trial. A total of 476 patients with OSA were randomized to receive solriamfetol 37.5 mg, 75 mg, or 150 mg, or placebo, once daily. Coprimary endpoints (Maintenance of Wakefulness Test sleep latency and Epworth Sleepiness Scale score) were met at all solriamfetol doses (P< .05), with dose-dependent effects observed at week 1 maintained over the 12-week study duration. All doses except 37.5 mg resulted in higher percentages of participants reporting improvement on Patient Global Impression of Change (key secondary endpoint; I < .05). [216]

The use of stimulant medications (ie, modafinil, armodafinil) or a DNRI (ie, solriamfetol) should be reserved for patients optimally treated with CPAP and who remain sleepy. However, there is not yet a consensus on the role of stimulant medications to enhance alertness, and research in nonoptimally treated OSA patients has not been performed to determine to what degree, if any, subjective and objective changes occur when the medications are used for OSA patients not treated ideally with CPAP.


Prevention of Obstructive Sleep Apnea

Patients should restrict their body positions during sleep. SDB is worse in the supine position, and some patients have apnea only in this position. Preventing the patient from assuming the supine position by using devices such as a snore ball (eg, a tennis ball sewn onto the back of the patient’s pajamas) or a gravity-activated position monitor may be useful. However, these devices are cumbersome and appear to benefit only those patients with mild OSA.

Patients with marked obesity may benefit from sleeping in an upright position. Additionally, the FDA has approved a specially designed pillow (PillowPositive) for the treatment of snoring and mild OSA, which maintains the patient’s head and neck position during sleep to optimize UA patency.

Patients should avoid smoking. Smoking increases the risk of snoring and apnea. Smoking cessation appears to decrease the risk. Individuals who smoke are also more likely than those who do not smoke to report problems with going to sleep, maintaining sleep, and daytime somnolence.

Patients should avoid drinking alcohol and using other sedatives known to make apnea worse. Finally, patients should avoid sleep deprivation.



All patients with signs or symptoms of OSA should be referred to a sleep disorders center for an evaluation by a sleep physician and a PSG study. A comprehensive sleep evaluation is recommended because as many as 25% of sleep patients have more than one sleep disorder, many of which are only identified as a result of a consultation with a sleep specialist.

Any patient with loud habitual snoring and any other feature of OSA who is being considered for surgery should be referred for a sleep study prior to surgery. This is important to rule out OSA because the surgery is likely to correct the snoring but may not correct the apneas or hypopneas, which are associated with other morbidities.

Patients should also undergo complete evaluation by a dental professional, who should be experienced in OA therapy and should work in close collaboration with the sleep disorders specialist.


Long-Term Monitoring


Once diagnosed with OSA and started on nasal CPAP, patients require regular follow-up with a sleep specialist. Most patients are seen within 2 months of initiating CPAP to determine if it has been effective in alleviating symptoms (eg, daytime sleepiness is substantially reduced or eliminated), to troubleshoot problems preventing regular use of the CPAP, and to reinforce the importance of daily use.

Further follow-up depends on whether the CPAP has been effective.

  • If CPAP has been effective, the patient is generally seen at 6- to 12-month intervals to troubleshoot new problems, to reinforce daily use, and to be certain the CPAP remains effective.

  • If CPAP has not been effective, problems preventing use are identified, steps are taken to eliminate problems, and the patient is seen at 2- to 3-month intervals until use is regular and the CPAP is alleviating symptoms. Repeat titration may be necessary.

  • Routine repeat PSG is generally not indicated for patients who report improved symptoms. Repeat CPAP titrations are usually performed in patients without effective relief of their symptoms despite intervention or in patients who had relief of symptoms but present months to years later reporting the return of symptoms, generally in association with weight gain.

A study by Kuna et al assessed home monitoring and determined that functional outcome and treatment adherence in patients evaluated using a home testing algorithm was not clinically inferior to the results found among patients receiving standard, in-laboratory PSG. [217]

Assessment of driving and flying risk

Assess the risk of driving in any patient with OSA. Criteria that increase this risk are as follows:

  • The patient has had previous motor vehicle accidents.

  • The patient has had near-miss incidents while driving.

  • The patient has evidence of severe daytime sleepiness or impaired driving performance.

Immediately warn patients at highest risk of the potential dangers of driving while sleepy—specifically, of the potential personal and social risk. Provide additional counseling depending on other risk factors (eg, occupation). Advise patients to not drive until their OSA is treated adequately. Provide additional counseling to family members as appropriate, and help patients explore alternatives to driving if they are unaware of their sleepiness or if they are unwilling to acknowledge their increased risk.

Document in writing any warnings, concerns, and/or recommendations given to the patient. This documentation reinforces the importance of the message to the patient and helps reduce the risk of legal liability for medical personnel.

Whether and under what circumstances patients with sleep apnea should be reported to the licensing authority depend on the laws of the state. Those who take care of patients with OSA must be aware of state statutes or regulations regarding reporting of high-risk drivers.

Laws regarding impaired drivers, including those with OSA, vary from state to state. In some states, the clinician is obligated to report patients under specific conditions (ie, mandatory reporting statute), whereas other states permit reporting but do not require it (ie, permissive reporting statute). Mandatory statutes take 1 of the 2 following approaches:

  • With the categorical approach, the clinician is obligated to report patients who have specified medical conditions, such as epilepsy. In these states, the reporting obligation is based on diagnosis alone.

  • With the functional approach, the clinician is required to report patients with certain medical conditions only if the clinician believes that a condition impairs the patient’s driving ability.

Each clinician is obligated to adhere to the requirements of the law in the specific state of practice, even if those laws do not reflect sound public policy or medical evidence. Irrespective of whether statutory reporting is required, clinicians may be liable for damages if a patient with obstructive sleep apnea injures himself or herself or someone else while driving.

American Thoracic Society guidelines on reporting of patients to the appropriate state authorities are as follows [74] :

  • In states with permissive reporting mechanisms, the clinician should at a minimum notify the state’s department of motor vehicles when a highest-risk patient (ie, a patient with OSA with severe daytime sleepiness and a previous motor vehicle accident or near-miss incident) (1) is unwilling to restrict driving until effective treatment is started, (2) is noncompliant or unwilling to accept treatment, or (3) has an untreatable condition or one that is not amenable to expeditious treatment (≤2 mo of diagnosis).

  • Increased occupational exposure to driving or increased occupational risk for an accident of substantial importance to the public may be other indications for reporting. Examples of patients in this situation are truck drivers who are transporting hazardous waste and school bus drivers.

  • A diagnosis of OSA without additional risk factors for impaired driving should not be the basis for reporting a patient unless required by state law.

Categorical reporting may be most appropriate in the context of occupational licenses, but this is arguable. At a minimum, the threshold for suspecting an increased driving risk due to sleepiness should be low given the increased hazard.

The US Department of Transportation convened a group of respiratory experts at its Conference on Pulmonary/Respiratory Disorders in Commercial Drivers in September 1990. The group recommends that operators with suspected sleep apnea should not be medically qualified for commercial vehicle operation “until the diagnosis has been eliminated or accurately treated.”

A US Federal Aviation Administration specification letter entitled “Sleep Apnea Evaluation Specifications” states that the complications of OSA present a risk to flying safety and recommends an initial workup, acceptable treatments, and follow-up for pilots being evaluated for OSA.