0541 Racial/Ethnic Disparities in Links of Obstructive Sleep Apnea Diagnosis and Treatment with Cognitive Impairment in Middle-Aged and Older Adults

Abstract Introduction Obstructive sleep apnea (OSA) is linked to cognitive impairment, and racial/ethnic disparities exist in both OSA and dementia. We previously identified disparities in OSA diagnosis and treatment among middle-aged and older adults in the Health and Retirement Study (HRS). Here, we tested whether the association between OSA diagnosis/treatment status and incident cognitive impairment differed by race/ethnicity. Methods Data came from the 2016-2020 waves of the HRS, a nationally representative study of middle-aged and older adults in the United States. We included participants who had probable or diagnosed OSA in 2016 determined by a modified STOP-Bang measure or self-reported OSA diagnosis. Self-reported OSA treatments included CPAP, oral appliance, and surgery. We classified participants as: probable-undiagnosed OSA (i.e., no self-reported diagnosis and met criteria for probable OSA on the modified STOP-Bang); diagnosed-untreated OSA; or diagnosed-treated OSA. Our outcome was a diagnosis of either cognitively impaired, no dementia (CIND) or dementia per the Langa-Weir Classification of Cognitive Function in waves 2018 and 2020. We assessed the association of OSA diagnosis/treatment status with incident cognitive impairment using generalized estimating equations. Effect modification by race/ethnicity (Hispanic/Latino, Non-Hispanic Black (NHB), Non-Hispanic White (NHW), Non-Hispanic Other Race) was assessed via an interaction term. Results We studied 3,881 participants who had probable or diagnosed OSA and were cognitively unimpaired in 2016 (mean 63.0±9.7 years, 52.6% female, 40% minoritized race/ethnicity). Among probable-undiagnosed OSA participants: compared to NHW participants, NHB and Hispanic/Latino participants had 105% (OR=2.05, 95% CI 1.48, 2.84) and 48% (OR=1.48, 95% CI 1.02, 2.11) higher odds of cognitive impairment, respectively. In main effects, diagnosed-treated participants had 22% lower odds of incident cognitive impairment compared to probable-undiagnosed OSA participants (OR=0.78, 95% CI 0.61, 0.99). We found evidence for potential effect modification by race/ethnicity (p for interaction = 0.051). Among NHB participants, diagnosed-treated participants had 54% lower odds of incident cognitive impairment than probable-undiagnosed participants (OR=0.46, 95% CI 0.27, 0.79); these associations were non-significant in NHW and Other racial/ethnic groups. Conclusion Undiagnosed NHB and Hispanic/Latino individuals with OSA may be at increased risk for cognitive impairment. NHB individuals may have particularly strong cognitive benefits from OSA treatment. Support (if any) National Institute on Aging (R01AG079391).