Does Treating Mild OSA Lessen Daytime Sleepiness?

Aaron B. Holley, MD


November 28, 2016

Debating the Treatment of Mild Obstructive Sleep Apnea

A few months ago, I reviewed a position statement on mild obstructive sleep apnea (OSA) published by the American Thoracic Society (ATS).[1] I concluded that the evidence to support treatment efficacy for this subgroup of patients was weak, and that the third edition of the International Classification of Sleep Disorders erred in expanding the OSA definition to include more patients.

As part of a lecture given at the 2016 CHEST conference, I was tasked with identifying patient subgroups with mild OSA who stood to benefit from treatment. Specifically, the moderator of the session wanted to know whether a particular subgroup of patients with mild OSA could, with treatment, experience a reduction in daytime sleepiness. This required a look at the evidence used to support the recommendations in the recent ATS statement.

The first issue, acknowledged by the statement authors, is that the hypopnea criteria used to define OSA differed among the articles they cited. For example, the second question in the statement is: "Does treatment of mild OSA, in comparison to no treatment, prevent or reduce adverse neurocognitive consequences and MVAs and improve quality of life?" The authors found 11 treatment trials in mild OSA that assessed sleepiness as an outcome. Four required a decline in oxygen desaturation of 4% to score an event as hypopnea,[2,3,4,5] two required a 3%-4% decline in desaturation or an arousal,[6,7] and five required neither.[8,9,10,11,12]

It has been well documented that the definition of hypopnea dramatically affects the total apnea/hypopnea index (AHI).[13,14,15] In fact, the AHI can triple when going from an arousal to a 4% desaturation requirement. The variability is greatest when the AHI is low, as it is with mild disease. When summarizing the studies cited by the ATS statement, it must be acknowledged that although they all purported to be examining mild OSA, the differences in hypopnea definitions mean that the patients were not the same across trials.

Assessing the Evidence: CPAP and Sleepiness

Five of the trials included by the ATS assessed continuous positive airway pressure (CPAP) as the treatment intervention for mild disease. Four were randomized controlled trials (RCTs), three used an oral placebo as the comparison,[8,9,10]and the fourth used a sham CPAP device.[5] The sixth trial was a prospective cohort study without a comparison arm.[6]

The oral placebo studies were small, and they generally showed an improvement in subjective (Epworth Sleepiness Scale [ESS]) but not objective (maintenance of wakefulness testing or mean sleep latency testing) sleepiness. The RCT using sham CPAP had the largest sample size but did not show a difference in subjective or objective sleepiness. The prospective cohort study showed a large improvement, but it was the weakest methodologically and included only eight patients with mild OSA.

Other studies tested such interventions as oral appliance therapy, weight reduction, or surgery. The oral appliance studies showed a reduction in daytime sleepiness (ESS), but both were observational.[7,16] The weight reduction studies were randomized and well done.[2,3,4] They showed consistent weight reduction during follow-up and normalization of the AHI, but no change in symptoms.

Finally, the surgery studies had conflicting findings. An RCT showed no effect on subjective outcomes,[11] whereas a very small observational study did show an effect.[12]

Two RCTs were not included because they recruited patients with mild and moderate OSA and the data for those with mild disease could not be isolated. The first was a smaller study[10] (42 participants, of whom only 28 finished the study) published in 2002. A placebo pill was used for the comparison group, and no difference in objective or subjective outcomes was found.

The second RCT that was not included is arguably the most important. The CATNAP study[17] randomly assigned 239 patients with mild to moderate OSA (mean AHI, 12.8 ± 6.4; approximately two thirds of the group fell within the mild AHI range) to receive CPAP vs sham CPAP. The study was designed to include symptomatic patients (mean ESS score, 14-15), and a 3% oxygen desaturation criterion was used to score hypopneas. The study found a significant difference in Functional Outcomes of Sleep Questionnaire (FOSQ) score that favored CPAP over sham therapy.

The Answer Is Still "No"

Given these data, what conclusions were drawn at the CHEST 2016 conference panel? For my part, I was unable to identify a specific mild OSA subgroup that could benefit from treatment. Largely on the basis of the CATNAP trial, I believe that the persons most likely to benefit will be profoundly sleepy at baseline and will be selected using oxygen desaturation criteria to identify hypopneas. Even then, if there is a benefit to be achieved, it is likely to be very small. The change in FOSQ score in the CATNAP trial was tiny, and it is debatable whether it surpassed the minimal clinically important difference for that score.[18] The 2016 ATS statement reached a similar conclusion.[1]

I'll reiterate what I have said before: The stakes here are really quite high. If you take the Wisconsin Sleep Cohort Data,[19] factor in the 4% desaturation they required for hypopneas, and adjust it for 2013 demographics,[20] you find that approximately 1 in 4 middle-aged adults (aged 30-60 years) will have mild OSA. That's a huge number. Let's say for the sake of argument that the subjective, symptomatic benefit from treatment is real. We know it's tiny. Should our health system invest in trying to achieve this small benefit by spending on CPAP for one fourth of the adult population? Never mind the issues with poor compliance or the inflated prevalence seen with 2012 hypopnea criteria.

My preference is clear: We shouldn't be paying for treatment of mild OSA. The cost-efficacy just isn't there.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.