How to assess, diagnose, refer and treat adult obstructive sleep apnoea: a commentary on the choices

Obstructive sleep apnoea (OSA) is a condition characterised by repetitive occlusions of the upper airway during sleep, resulting in arousals and sleep fragmentation. It impacts on daytime vigilance1and contributes to cognitive dysfunction2and mood disorders.3 It is a source of lost productivity in the workplace4 and increases motor vehicle accident risk.5 OSA has also been implicated as a cause of hypertension,6 with studies showing small but consistent falls in blood pressure following continuous positive airway pressure (CPAP) treatment.7 Epidemiological studies have also shown OSA to be independently associated with an increased risk of diabetes8 and cardiovascular disease,9,10 although definitive evidence for a causal link with these diseases awaits the results of large-scale randomised controlled trials of OSA treatment. In the early 1990s, the prevalence of OSA in the community in the United States, determined by polysomnography (PSG), was shown to be 24% of adult men and 9% of women,11 with recent evidence suggesting a further increase due to the obesity epidemic and an ageing population.12 OSA is now recognised as a major public health and economic burden, with an estimated cost to the Australian community of more than $5.1 billion a year in health care and indirect costs.4

The purpose of this article is to describe the key issues in evaluation and management of OSA, to assist health care professionals to better engage in OSA management. We outline several evidence-based models of care that could be scaled up to allow the primary care physician to have a greater role in addressing the high burden of OSA in the community. To do this, primary care health professionals must be skilled in identifying those at high risk of OSA who are likely to benefit from treatment and must know which investigation to order, what treatments to recommend, and when specialist referral is needed.

Despite the high prevalence of OSA, most patients are minimally symptomatic. About 15% of patients have moderate to severe sleep apnoea.13 The vital issue in clinical practice is to identify those with OSA who have clinically important disease. Lack of clarity around goals of treatment can lead to excessive investigation, inappropriate treatment and patient disengagement. We believe the major goals of management of OSA should be:

• to identify and offer treatment to symptomatic patients, regardless of disease severity, whose safety and quality of life is affected;

• to identify and offer treatment to patients with severe OSA determined by PSG, regardless of symptoms, who may be at risk of adverse health outcomes; and

• to modify adverse lifestyle factors that contribute to OSA pathogenesis and other poor health outcomes. This may include advice on diet and exercise to lose weight, and encouragement to reduce alcohol intake and stop smoking.

Optimal outcomes are usually achieved through an initial identification of the presenting clinical triggers, an evaluation of the symptom profile, and an exploration of the patient’s treatment preferences and capacity to afford or comply with the range of treatment options. The workup for OSA must start with a careful clinical assessment to identify patients who are likely to benefit from treatment. Clinicians must then select an investigation: either in-laboratory PSG, home-based PSG or simplified limited channel sleep testing. The test result must then be coupled closely with the clinical assessment to inform a personalised treatment plan. This plan should identify adverse lifestyle factors, overlapping sleep disorders and medical comorbidities (eg, hypertension, diabetes, depression, dyslipidaemia), and consider these when advising on OSA-specific treatments.14 Box 1 depicts an algorithm that may assist the primary care practitioner with this process. There is no one preferred treatment for OSA but rather a range of options of proven effectiveness that can be applied individually or in combination, depending on patient preference, symptoms, OSA severity, comorbidities and other health risk factors (Box 2). Development of a personalised treatment plan requires the active involvement of the patient, partner and family in goal-setting.14 For the health professional, it requires that they be sufficiently familiar with the field and the practical application of each of the available OSA investigations and treatment options. The complexity of this process has meant that OSA has been traditionally managed by a relatively small specialised workforce using the gold standard, in-laboratory PSG. However, patient access to specialist sleep services has been limited and alone will not be able to cope with the evidently large burden of disease.12

Given these service barriers, various simplified, lower-cost clinical models have been developed for OSA. These have incorporated screening questionnaires to identify patients at high risk of OSA,15-19 simplified testing with home-based PSG or limited channel sleep studies (typically without sleep electroencephalography) and selected use of automatically titrating CPAP devices that lessen the need for supervised in-laboratory CPAP titrations. If patients are identified as having a high pretest probability of OSA and if major comorbidities and overlapping sleep disorders are excluded (Box 3), the use of home-based PSG or limited channel sleep testing and automatically titrating CPAP has been shown to produce similar or non-inferior patient outcomes to more traditional specialist referral and in-laboratory PSG approaches.20-23 Further, it has been shown that with suitable training and support from a specialist sleep centre, these management approaches can be applied effectively to uncomplicated OSA patients by nurses and primary care physicians.24

A major challenge is how to translate and upscale these research findings from controlled settings to the “real world” to meet the demonstrably high community burden of disease while ensuring high-quality, holistic patient care. The current availability of open-access PSG has enabled primary care practitioners to become more involved in the care of OSA patients. However, few of these services currently select for high pretest probability of disease, nor do they train or adequately support the referring health care professional to ensure that they are sufficiently knowledgeable in assessment and personalised treatment of OSA. Some patients accessing this service model who are found to have uncomplicated moderate to severe symptomatic OSA may adhere to CPAP treatment and be successfully managed in primary care. However, there are no data available on the overall adequacy of CPAP treatment for patients with milder or complex OSA, or overlapping sleep disorders and adverse lifestyle issues. There is also a lack of such data for patients who refuse CPAP treatment or for whom the treatment is unsuccessful.

If sleep service delivery at the primary care level is to be upscaled and the recently validated simplified models of care for OSA translated into routine care, there will need to be greater awareness around clinical assessment, goals of OSA treatment and the various available treatment options. Objectives of the clinical assessment are to determine the motivating factor(s) for presentation and the patient’s symptom profile and pretest probability of disease, and to identify modifiable adverse lifestyle factors and co-occurring sleep problems. Collectively, these factors will have an important influence on the investigation and management pathway.

In Australia, Medicare reimbursement is provided for full PSG, conducted in either supervised (in-laboratory) or unsupervised (home) settings. For home-based PSG, patients are connected to the electrodes and sensors at the facility on the afternoon of the test and return home, or self-connect in their own home before bed, according to written, verbal or audiovisual instructions. Patients may perceive an increased level of convenience and comfort with testing in their own home, sensing a more sleep-conducive environment. When directly compared, one study showed 50% of patients preferred home-based testing, 25% preferred laboratory-based testing and 25% had no preference.28 Comparison of home-based versus in-laboratory PSG showed reduced total cost for home testing28 and high overall satisfaction rates for both forms of testing.28 However, home-based PSG is associated with higher test-failure rates, partial signal loss producing equivocal results,28,29 and a tendency to underestimate sleep apnoea severity.29 More severe sleep apnoea (high pretest probability) may overcome the shortfalls of partial signal loss and any tendency to underestimate severity, and this will improve diagnostic accuracy.

A comprehensive overview of all treatment options is beyond the scope of this manuscript, and international guidelines are available.31 Options are summarised in Box 2. Identification of sleep apnoea is a teachable moment for the health care professional to guide the patient and suggest interventions to modify lifestyle and reduce weight. Thereafter, treatment considerations ought to extend beyond CPAP. Studies have shown large variation (17%–71%) in adherence to optimal CPAP (defined as average of ≥ 4 hours per night).32 More recent randomised controlled trial evidence suggests that mandibular advancement splints may be as effective as CPAP across a range of OSA severity,33 although there is limited information on longer-term compliance. Newer surgical techniques are emerging for OSA and the combination of uvulopalatopharyngoplasty, tonsillectomy where appropriate, and a low-morbidity technique to reduce tissue volume at the tongue base shows promise in highly selected patients for whom conventional therapies such as CPAP or a mandibular advancement splint have been unsuccessful.34 Adults with gross tonsillar hypertrophy and sleep apnoea are uncommon but often do very well following tonsillectomy. Some preliminary success is reported with nasal positive expiratory pressure devices,35 although long-term adherence to treatment is unknown and patient selection needs more evaluation.

Overall, treatment for OSA includes a range of options, all of which have their unique challenges. Cost is a consideration and commitment is required to achieve long-term adherence. There is a risk that patients may not persevere with the treatment plan if the clinical assessment was not patient-focused and did not address key presenting symptoms or motivators. A negative experience has consequences in terms of lost opportunity if the patient withdraws from the therapeutic process. All patients should be clinically reassessed after a treatment option is tried, to ensure the treatment has been effective in controlling both OSA and its symptoms.

The specialty of sleep medicine now has a robust curriculum, encompassing both respiratory and non-respiratory sleep disorders, and requires a full year of dedicated training. This will see larger numbers of specialists with sufficient skills to assist with managing the public health burden of OSA. Telehealth will also enable sleep specialists to assist health care practitioners and patients in rural and regional communities.

However, the large burden of disease is likely to be best served in the long term by an expanded trained pool of primary care and other health care providers working alongside sleep and respiratory specialists. In this model of care, sleep specialists working in a multidisciplinary environment would have as their major clinical focus complex or atypical OSA cases (eg, those with comorbidities or overlapping sleep disorders) or treatment failures. All modalities of sleep testing will be used in accordance with existing validated algorithms. These models of care will take time to evolve and will require changes to clinical guidelines and accreditation standards, the upskilling of the health care workforce, and government and private sector policy changes with respect to reimbursement.

1 Algorithm for management of obtructive sleep apnoea (OSA) in the community

PSG = polysomnography.