OBJECTIVES: There is a paucity of longitudinal studies with sleep efficiency (SE) as an outcome measure. Our objective was to examine potential risk factors for incident reduced SE among community-dwelling women in late life.

METHODS: We studied 700 women (mean age 82.5 [SD=3.0] years) with a SE ≥70% at the Year 16 (2002-04) visit of the Study of Osteoporotic Fractures with a follow-up measure of SE at the Year 20 (2006-08) visit. SE (percentage of time sleeping while in bed) at both visits was measured using a wrist actigraph with data collected for an average of four 24-hour periods. Women were classified as having incident reduced SE if they had SE <70% at Year 20. Logistic regression was used to estimate the associations between potential risk factors (demographics, lifestyle, use of medications, self-reported medical conditions, functional impairment, frailty, mental and physical health) at Year 16 and reduced SE at Year 20. The association of each candidate risk factor with reduced SE at Year 20 was examined in models adjusted for age, clinical site and continuous SE at Year 16. Candidate risk factors with Benjamin Hochberg false-discovery rate q-values <0.10 were included in a final multivariate model.

RESULTS: Among the 700 eligible women, 62 (8.9%) developed incident reduced SE between the Year 16 and Year 20 visits. After adjusting for age, site and baseline SE, antidepressant use [OR=3.06; 95% CI: 1.50-6.25], benzodiazepine use [OR=2.97; 95% CI: 1.30-6.80] and the presence of hypertension [OR=2.83; 95% CI: 1.47-5.45] at Year 16 were independently associated with a higher odds of having reduced SE at follow-up.

CONCLUSIONS:These findings suggest that antidepressant use, benzodiazepine use and hypertension are risk factors or markers for the development of reduced sleep efficiency in older women. Future studies are warranted to examine the underlying mechanisms for these associations.