UCSF

Study objectives

To examine associations of objectively and subjectively measured sleep with subsequent cognitive decline.

Design

A population-based longitudinal study.

Setting

Six centers in the United States.

Participants

Participants were 2,822 cognitively intact community-dwelling older men (mean age 76.0 ± 5.3 y) followed over 3.4 ± 0.5 y.

Interventions

None.

Measurements and results

OBJECTIVELY MEASURED SLEEP PREDICTORS FROM WRIST ACTIGRAPHY: total sleep time (TST), sleep efficiency (SE), wake after sleep onset (WASO), number of long wake episodes (LWEP). Self-reported sleep predictors: sleep quality (Pittsburgh Sleep Quality Index [PSQI]), daytime sleepiness (Epworth Sleepiness Scale [ESS]), TST. Clinically significant cognitive decline: five-point decline on the Modified Mini-Mental State examination (3MS), change score for the Trails B test time in the worse decile. Associations of sleep predictors and cognitive decline were examined with logistic regression and linear mixed models. After multivariable adjustment, higher levels of WASO and LWEP and lower SE were associated with an 1.4 to 1.5-fold increase in odds of clinically significant decline (odds ratio 95% confidence interval) Trails B test: SE < 70% versus SE ≥ 70%: 1.53 (1.07, 2.18); WASO ≥ 90 min versus WASO < 90 min: 1.47 (1.09, 1.98); eight or more LWEP versus fewer than eight: 1.38 (1.02, 1.86). 3MS: eight or more LWEP versus fewer than eight: 1.36 (1.09, 1.71), with modest relationships to linear change in cognition over time. PSQI was related to decline in Trails B performance (3 sec/y per standard deviation increase).

Conclusions

Among older community-dwelling men, reduced sleep efficiency, greater nighttime wakefulness, greater number of long wake episodes, and poor self-reported sleep quality were associated with subsequent cognitive decline.