Sleep-disordered breathing (SDB) disturbances are very prevalent in developed countries. Since it was estimated over 10 years ago, the prevalence of SDB is probably higher now because of the dramatic increase in body weight in the populations of these countries.¹ Given the large increase in mortality and morbidity outcomes associated with the diagnosis of SDB, the diagnosis of a nocturnal breathing disorder should no longer be confirmed solely by conventional in-laboratory polysomnographic recordings. This justifies the need for abbreviated monitoring during sleep to be part of the assessment of SDB and the tremendous effort developed by the sleep research community to evaluate the diagnostic value of abbreviated recordings.

The study by Jobin et al² reported in this issue of Thorax (see p 422) is the first comparative study that does not use in-laboratory polysomnographic recordings as the gold standard, and is thus an important step towards evaluating the merits of abbreviated recording techniques. This is a major upheaval in the field of sleep medicine, and opens the way to realistic assessments of abbreviated recording techniques in real-life conditions that avoid costly, time-consuming in-laboratory polysomnographic recordings. It is, however, reasonable to wonder whether the authors should have proceeded more cautiously by starting with level 2 monitoring techniques (ie, an unattended complete polysomnographic study) as a reference, which would allow the influence of home monitoring on cardiorespiratory variables to be evaluated while, at the same time, taking potential differences in sleep characteristics into consideration. The authors did not explain why electrophysiological variables, which can be recorded using the Suzanne apparatus, were not collected. At a minimum, the reference portable monitoring technique should be designed to interfere minimally with sleep quality. The level 3 device used by Jobin et al may not fully meet these requirements due to the cumbersome equipment, but the latest generation of recording systems should correct these potential pitfalls.

Despite the tremendous interest in the use of abbreviated monitoring by the medical community, American medical societies (APSS, ACCP, ATS) have, until recently, maintained that portable monitoring devices are not accurate enough to be used in an ambulatory setting for the management of SDB.³ A number of reasons may account for the discrepancy between the official recommendations of medical societies and the widespread use of abbreviated monitoring by the medical sleep community (apart from the potential impact of differences in reimbursement rules in certain countries). One is the very large disparity in the recorded signals and in the recording and signal processing techniques of the devices that have been tested (such as oximetry, breathing sounds, sophisticated cardiac rhythm analysis (heart rate variability), respiratory impedance signals, pulse transit time, arterial tonometry). In this regard, night-time oximetry recordings remain the most extensively investigated technique, and it is somewhat paradoxical that a typical desaturation/resaturation profile per se may not be considered as a diagnostic finding given that a repetitive fall in arterial oxygen saturation (Sao₂) is recognised as a cornerstone of the capacity of sleep recordings to identify SDB⁴ and that the accuracy of Sao₂ recording techniques (probes, software analysis including artefact deletion, sampling frequency, averaging time, signal processing) has dramatically improved in recent years.

The discrepancies in the diagnostic performance of oximetry recording techniques reflect the specificity of the data obtained with a given recording system, but also indicate the need to have access to, and to examine, raw data to satisfactorily interpret abbreviated recordings. Considering that “oximetry” refers to a wide variety of different techniques with different diagnostic performances,⁵ the term “oximetry” is meaningless when used to designate an investigation category. The work of Jobin et al illustrates this point since the desaturation profiles of the two oximeters they tested were different. Expertise with portable monitoring thus has to be developed in each sleep centre and should take into account the usefulness and limits of portable monitoring devices in the investigation strategy for individual patients.

Another concern is the moving target of the validation of technological development procedures, particularly with respect to the choice of the signal(s) to be recorded and the way tracings should be scored. One example of this is the lack of change in the normal threshold of the apnoea-hypopnoea index when highly sensitive signals such as nasal pressure recordings are used instead of the less sensitive signals used to select the 5 events/h threshold.⁴

A further reason for the discrepancy between official recommendations and the widespread use of portable monitoring is the unrealistic diagnostic value that has been attributed to abbreviated recording techniques in the investigation of SDB; that is, that the results of these recordings have usually been considered in the binary fashion where a negative test is useless in the diagnosis of sleep apnoea. However, it must be pointed out that portable monitors have been extensively tested in patients who are referred to tertiary centres independent of underlying symptoms and morbid conditions. This is not the ideal situation for using these devices. The circumstances of individual patients (logistical considerations, ability to install captors and to check whether the signals are adequate) should be taken into consideration, as should the factors known to influence the variables being measured. For example, in the presence of SDB, the potential of abnormal nocturnal oximetry is probably negatively influenced by the absence of excess weight and/or comorbid conditions. The same reasoning applies to the investigation of patients whose comorbid conditions may be worsened by SDB, independent of any vigilance disturbances. This may be a growing problem since cardiovascular and metabolic morbidity factors have been linked to the number of nocturnal desaturation/resaturation events^4,6 through their effect on systemic inflammation and sympathetic activity.^7,8 In situations where diurnal somnolence is not an issue, oximetry may be the test of choice because, unlike sleepiness investigation strategies, negative oximetry recordings eliminate the contribution of nocturnal breathing disorders as a deleterious factor of a patient’s comorbidity. Diagnostic studies conducted in these populations, rather than in patients referred to sleep clinics, should help to refine these diagnostic criteria by targeting patients who best fit the diagnostic value of a given abbreviated recording technique. Combined with an increase in the availability of portable monitors, this would appear to be the only effective strategy to reduce investigation delays and costs. This further implies that training and educational strategies adapted to the condition of individual patients are required for polysomnography to no longer be the “one size fits all” approach to investigating SDB.

Abbreviated recording techniques should become an integral part of a stepwise investigation strategy where the results of sleep recordings must respond to the questions raised by the clinician about a patient’s specific clinical condition and pre-test probability. Future approaches should thus take into account the condition of individual patients and the clinical justification for conducting ambulatory recordings during sleep to select the most appropriate investigation strategy and to refine the diagnostic procedure by using level 4 to level 1 recording techniques in a stepwise fashion.