Association between severity of obstructive sleep apnea and its common symptoms varies by race, ethnicity, and sex

The US Preventive Services Task Force recently released guidelines suggesting little evidence of benefit to screening asymptomatic adults for obstructive sleep apnea (OSA). Our goal was to provide important context to this statement. Specifically, we examined associations between common OSA symptoms, excessive daytime sleepiness and snoring, and OSA severity for different racial/ethnic and sex groups.

Analyses were performed on 2 samples. One combined 2 observational studies that included full polysomnography, the Epworth Sleepiness Scale (ESS), and questions about snoring (mean [standard deviation] age 39 [15.2] years). The second sample was the Multi-Ethnic Study of Atherosclerosis study of older adults (mean [standard deviation] age 69 [9.1] years), which also included polysomnography, ESS and a question about snoring. Apnea-hypopnea index represented OSA severity. For each racial/ethnic-sex group we estimated correlations between apnea-hypopnea index and ESS and the sensitivity and specificity of excessive daytime sleepiness (ESS >10) or frequent snoring to predict moderate-to-severe OSA (apnea-hypopnea index >15 events/h).

A weak significant correlation between OSA severity and ESS was found only in White men in the first sample and Black men in the second sample. Screening tool characteristics for ESS and snoring were poor except for moderate specificity in some racial/ethnic-sex groups.

Excessive daytime sleepiness and snoring are commonly used to identify symptomatic patients. Our results suggest that the accuracy of these symptoms to identify OSA varies by race/ethnicity and sex. Therefore, focus on common symptoms as an OSA screen could systematically leave out certain patient populations who would benefit from treatment.

Stepney D, Attarian HP, Knutson KL. Association between severity of obstructive sleep apnea and its common symptoms varies by race, ethnicity, and sex. J Clin Sleep Med. 2023;19(10):1727-1733.