Positional sleep apnea revisited (2025)

INTRODUCTION

The heterogeneity of sleep apnea phenotypes is far greater than first conceived. As we have learned more about them, we are redesigning the practice of clinical sleep medicine. As we identify ever-increasing numbers of obstructive sleep apnea (OSA) phenotypes, we have to determine the utility of the grouping and how it affects health and wellness.

In this issue of the Journal of Clinical Sleep Medicine, Duce etal1 present a thorough examination of data from the Sleep Heart Health Study identifying the prevalence of the most common sleep apnea phenotype, positional obstructive sleep apnea, in varied community settings. Its presence in the population of patients studied in sleep laboratories has been found to vary from 50–60%,2 but few studies have looked at those who have not presented with a report of sleep problems. A well-done Swiss study by Heinzer etal3 reported on a community population that was fairly homogeneous and less obese than in the United States. The US data reported here were obtained from the Sleep Heart Health Study,4 a multiethnic cohort.

The authors identified positional OSA (POSA) and position-independent OSA (non-POSA) and also looked at 2 subgroups of POSA: supine-isolated (siOSA), in which the nonsupine apnea-hypopnea index (AHI) is < 5 events/h, and supine-predominant (spOSA), when the AHI is at least twice the nonsupine but nonsupine events are > 5 events/h. Only polysomnograms with at least 30 minutes each of supine and nonsupine sleep were included. They found that positional OSA was predominant in the community population in the same range as the clinical studies but that apnea was less severe. Interestingly, there was no difference in AHI between the POSA and the non-POSA groups. However, the subgroup analysis showed that those with spOSA were more likely to be on a variety of medications for comorbid conditions, such as antihypertensives, lipid-lowering agents, and oral hypoglycemics, compared with the siOSA group.

This study identifies a group at high risk for subsequent cardiovascular events and, because they are younger than those who present to sleep clinics, they appear to be a group to target for diagnosis and possible intervention. There are some concerns as to how to go about this. It is assumed that 1 night of testing is definitive. However, that is not necessarily so.

The night-to-night variability of POSA was addressed by Joosten etal.5 They performed 2 polysomnograms within 4 nights on 44 patients and found that the AHI ratio of supine:nonsupine AHI of 2:1, as is currently used, did not reliably identify POSA from one study to another, but a 4:1 ratio did in men. None of the ratios tested between 2 and 5 was reliable for women. The reasons for this are not clear but were felt to reflect sex differences in airway anatomy and physiology.

Instrumentation may also confound reproducibility, as well as applicability to different patient groups. The type of flow sensor may affect scoring of hypopneas (the differences between thermistor and nasal pressure recordings are well-known),6 as do brands of oximeters7; in a recent study8 of various oximeters in stable patients in intensive care units with indwelling arterial lines SpO₂ was ± 4%. If home sleep apnea tests are used, carbon monoxide levels from smoking and/or the use of indoor wood stoves for heating alter the accuracy of the desaturation index and Tc90 (cumulative time spent with SpO₂ < 90%). Also, the accuracy of the various position sensors is unknown.

After those questions are answered, the question of therapy for the 2 groups arises. It is tempting to posit positional therapies for the siOSA group. The authors suggest that this subgroup should be considered for such an intervention. However, there is no device that is currently recommended for primary therapy of POSA. Multiple devices of various types9^–12 have been tested with varying degrees of success. Most of them are effective in the short term, but by 6 months most have less than 10% adherence. It seems that getting someone with mild disease and few symptoms to adhere to therapy is difficult no matter what the intervention is.

This paper has painted a picture of the prevalence of POSA and its subtypes in a representative community cohort. In the 30-year interval since these data were collected, the population has aged and grown heavier. It is likely that the prevalence is even higher now. Nevertheless, the need for further studies to identify the cohorts most in need of aggressive therapy and those whose benefits from interventions are minimal is clear. The identification of the spOSA subgroup needs to be made earlier and interventions considered to mitigate later health effects, because their comorbidities indicate that they are at high risk for untoward effects from their sleep-disordered breathing. To this end, perhaps clinical polysomnograms and home sleep apnea test reports should include clear statements regarding how much time was spent in the supine vs the nonsupine position, as well as how many minutes of supine rapid eye movement sleep were captured, rather than simply including “supine rapid eye movement sleep was present.” For men, comments on the reliability of single-night assessments of positionality appear to be warranted from the work of Joosten etal, and for women, a disclaimer that there is significant variability in POSA from night to night might also be included.

The siOSA group is tantalizing in that positional therapy seems so obvious and inexpensive. However, the lack of long-term adherence thus far reflects the unwillingness of someone with no symptoms or minimal symptoms to tolerate any inconvenience.

Identifying OSA phenotypes is not enough. Learning which groups require early intervention will occur when they can be found on routine sleep studies. The groups in this study clearly can be found relatively easily if we are willing to look for them. Further research is needed to know what the next steps should be.

DISCLOSURE STATEMENT

The author reports no conflicts of interest.

REFERENCES