Abstract
Abstract Introduction We aimed to determine if age or sex modifies associations between sleep-disordered breathing (SDB), sleep duration and severe phenotypes of combined SDB/sleep duration with 7-year neurocognitive change (NC) in a diverse sample of U.S. Hispanic/Latinos. Methods We analyzed data of 5,235 adults 50-80 years of age from SOL-INCA, an ancillary to the Hispanic Community Health Study/Study of Latinos that determines the risk factors for NC. The main outcome was NC after a mean follow-up of 7-years on measures of memory (SEVLT sum and SEVLT recall), language (word fluency), processing speed (DSS) and a cognitive impairment screener. We evaluated the effect of baseline SDB (AHI ≥ 15), sleepiness (Epworth Sleepiness Scale, ESS ≥ 10), self-reported sleep duration (i.e. <6 hours, 6-9 hours, ≥ 9 hours), age and sex on NC. Survey linear regression models with interaction terms were used to examine the relationship between SDB, sleep duration, combinations of SDB and sleep duration phenotypes and NC. Depression, vascular risk, sleep medication, and study site were entered into all models as covariates. Results Overall, the mean age was 56.0 years, 54.8% females, 62.2% completed high school, 17.3% had SDB, 6.6% had short sleep,and 14.8% had long sleep. Sleep duration and SDB were not associated with NC. There was a significant interaction between agexSDB+sleep duration on delayed recall (F10,599= 2.40, p=0.01) and processing speed (F10,597= 2.55, p=0.01). Combined SDB + short sleep was associated with decline in processing speed (β=-0.6, 95% CI= [-1.2, -0.1], and combined SDB+long sleep was associated with decline in verbal memory (β=-0.9, 95% CI=[-1.7, -0.2] in adults aged ≥ 65 years. There was no association in participants aged <65 years and no sex differences. Conclusion Age, but not sex, modified the association between SDB and sleep duration with decline on processing speed and verbal memory. Sleep interventions tailored for older adults may be useful in slowing or preventing neurocognitive decline. Support This work is supported by National Institute on Aging (R01AG048642, RF1AG054548, R01AG061022, and R21AG056952).