16 August 2012. At the Alzheimer’s Association International Conference held 14-19 July 2012 in Vancouver, Canada, both epidemiological and lab findings converged to point a finger at sleep in connection to cognitive decline. In particular, objective sleep measures suggesting that sleep disorders predict decline added a new layer of experimental data to older self-reported hints that how long and how well one sleeps might affect brain health later in life.

“It is unusual to have [multiple] studies coming together at the same time to say the same thing,” said Constantine Lyketsos, Johns Hopkins Bayview Medical Center, Baltimore, Maryland. “Sleep disruptions throughout the lifespan are probably accelerators of cognitive aging and risk factors for dementia,” he said.

Kristine Yaffe, University of California, San Francisco, presented her recent epidemiological findings on sleep—some published and some not. Yaffe and colleagues followed a prospective, multicenter cohort of community-dwelling women, all of whom were 65 or older and free of dementia at enrollment. Fifteen years into the study, about 3,000 of those women underwent actigraphy, whereby a watch-like device worn on the wrist measures sleep by recording movement over several days. They also underwent polysomnography, a lab-based sleep study that takes more in-depth measurements of overnight sleep patterns. Five years after those tests, the scientists checked to see who had since developed mild cognitive impairment (MCI) or dementia.

Yaffe found that several sleep-related problems raised the risk of subsequent cognitive decline. For one thing, a late shift in the women’s circadian rhythm, where their mean peak daily activity moved by two hours from early to late afternoon, was linked to an almost two-fold increased risk for developing MCI or dementia (see Tranah et al., 2011). Separately, sleep-disordered breathing (or sleep apnea) at baseline also heightened the risk of developing MCI or dementia almost twofold. This latter finding may be explained by hypoxia rather than sleep fragmentation, said Yaffe (see ARF related news story on Yaffe et al., 2011). In addition, a delay in falling asleep and less efficient sleep both increased the risk of being placed in a nursing home five years later.

“The effect is probably bidirectional, meaning disordered sleep may cause dementia at the same time that dementia causes sleep trouble. But at least we are showing that sleep problems are predicting five-year outcomes related to nursing home placement and MCI and dementia,” Yaffe said at a press briefing. The mechanisms are not yet clear, but besides hypoxia they could have to do with Aβ accumulation, she added (see Part 2). The findings matter, she said, “because sleep problems could be treatable, and intervention might delay or prevent some of the cognitive sequelae.” A next step would be to do a randomized controlled trial that tests whether improving the quality of sleep delays cognitive problems, Yaffe said.

One poster presentation did look at sleep apnea treatment, though the study was retrospective rather than prospective. Anne-Cécile Troussière, Université Lille Nord de France, presented her exploratory finding that mild to moderate AD patients with sleep apnea fare better if they use a continuous positive airway pressure (CPAP) device, which treats sleep apnea by applying a mild pressure to keep airways open. The nine patients who refused CPAP treatment declined by three points on the Mini-Mental State Examination (MMSE) over five years, whereas the 28 patients who used CPAP regularly slipped by less than one point. CPAP treatment has previously been reported to slow or improve cognitive decline in AD patients (see Ancoli-Israel et al., 2008, and Cooke et al., 2009).

Another poster supported Yaffe’s suggestion that hypoxia may be to blame for the negative effects of sleep apnea. Emily Clionsky, a private provider from Clionsky Neuro Systems, Inc., Springfield, Massachusetts, found that her memory clinic patients with cognitive complaints, MCI, or dementia have a higher prevalence of hypoxia than previously realized. In her database of 353 patients, 64 percent were hypoxic at some point during the day, either while sitting, walking, or during the night. Previous estimates had put that number at 43 percent (see Reynolds et al., 1985). Clionsky said no good data existed for the prevalence in the healthy elderly population, but that estimates for normal middle-aged people ranged from 2-40 percent.

Taken together, these findings suggest that better sleep benefits the aging brain. In keeping with the theme that one should enjoy everything in moderation, however, some data claim that too much sleep, as well as too little, could be a bad thing. Elizabeth Devore, Brigham and Women’s Hospital, Boston, Massachusetts, analyzed data from the Nurses' Health Study of more than 15,000 women who had been self-reporting their sleep times at mid-life and later in life through periodic questionnaires. From about 2000 until 2006, four separate telephone sessions assessed the women’s cognitive ability. The women were 70 or older at their initial cognitive assessment.

Women who reported the shortest (five hours) and longest (nine hours) sleep durations at both ages had the worst average cognitive scores at the end of the study. Participants in the middle, averaging seven hours of sleep each night, fared the best. In addition, women whose sleep duration changed the most, i.e., by more than two hours per day, between middle and later life, did worse on the cognitive tests than women who slept the same number of hours throughout the study. “In general, these findings indicate that extreme sleep duration or greater changes over time may contribute to cognitive decrements in older adults,” Devore said during a press conference announcing her results. “This could potentially lead to sleep and circadian-based strategies for mitigating cognitive impairment in Alzheimer’s disease.” The accuracy of self-reported data is generally difficult to verify, though Devore did report a correlation of 0.8 between self-report and a six-day sleep diary kept by a small substudy of 260 women. Other scientists cautioned that how long an aging person sleeps may be influenced by many factors.

Claudine Berr, INSERM, Montpellier, France, presented data from the French Three City (3C) Study, which has followed almost 9,300 people in Bordeaux, Dijon, and Montpellier for more than a decade. Surveys at baseline assessed participants’ sleep complaints, asking whether they had excessive daytime sleepiness, were awake during the night or in the early morning, if they had a hard time falling asleep, or if they slept poorly. Follow-up assessments every two or three years looked for cognitive decline, defined as a drop in MMSE score by four points. Reported by about 18 percent of patients, excessive daytime sleepiness alone boosted the risk of cognitive decline by 30 percent. “It is important to recognize and, if possible, to treat sleep problems, particularly excessive daytime sleepiness,” Berr said.—Gwyneth Dickey Zakaib.

This is Part 1 of a two-part series. See also Part 2.