Review
Connections between sleep and cognition in older adults Kristine Yaffe, Cherie M Falvey, Tina Hoang
Sleep disturbances and cognitive impairment are common in older adults. Mounting evidence points to a potential connection between sleep and cognitive function. Findings from observational studies support a role for sleep disturbances (particularly for sleep duration, sleep fragmentation, and sleep-disordered breathing) in the development of cognitive impairment. Less consistent evidence exists for associations of insomnia and circadian rhythm dysfunction with cognition. These findings suggest that the sleep–wake cycle plays a crucial part in brain ageing, pointing to a potential avenue for improvement of cognitive outcomes in people at risk of cognitive decline and dementia. Several biological mechanisms might underlie the association between sleep and cognition, but these pathways are not completely understood. Future studies that aim to clarify the association between sleep and cognition might help to identify people at risk of cognitive disorders and to facilitate the development of novel therapies to treat and potentially prevent both sleep disturbances and cognitive impairment.
Introduction As people age, changes in sleep patterns often occur, including a decrease in total sleep duration and efficiency, an increase in sleep fragmentation, greater difficulty falling asleep, and less time in rapid eye movement (REM) sleep and slow wave sleep.1–3 Although some of these changes might be a function of normal ageing, others might result from underlying medical conditions4–7 or a neurodegenerative process.2 The sleep–wake cycle is regulated by complex interactions among brain regions and neurotransmitter systems8 (figure 1), many of which are implicated in memory and cognitive function.9–11 Perhaps as a result of this shared circuitry, sleep problems are common in people with Alzheimer’s disease and other dementias9,12–16 (panel). In adults with Alzheimer’s disease, circadian rhythm dysfunction and so-called sundowning (increased confusion and restlessness at the end of the day and into the night) are frequently reported and are thought to result from degeneration of the suprachiasmatic nucleus (known as the master clock regulating circadian rhythms) and the cholinergic neurons of the nucleus basalis of Meynert.9,15 Similarly, REM behavioural sleep disorders often accompany dementia with Lewy bodies and might indicate degeneration of brainstem structures, although the specific anatomical regions have not been determined.15–20 Although an increased prevalence of sleep disturbances in people with dementia has been well documented and is often thought to result from neurodegeneration, focus has turned towards the possibility that sleep disturbances can also increase the risk of cognitive decline and dementia. The aim of this Review is to discuss research investigating how sleep disturbances in older adults (≥55 years) might predict or contribute to cognitive decline and dementia. We present evidence from observational studies of the risk of cognitive decline and dementia associated with sleep disturbances in older adults, including studies of insomnia, sleep quality, sleep duration, excessive daytime sleepiness, sleep-disordered breathing, and circadian rhythm disturbances. We then www.thelancet.com/neurology Vol 13 October 2014
discuss potential biological mechanisms underlying the association followed by results from treatment trials assessing potential cognitive benefits of improved sleep.
Lancet Neurol 2014; 13: 1017–28 Departments of Psychiatry, Neurology, and Epidemiology and Biostatistics, University of California, San Francisco, CA, USA (Prof K Yaffe MD, C M Falvey MPH); and Northern California Institute for Research and Education, San Francisco, CA, USA (T Hoang MPH) Correspondence to: Prof Kristine Yaffe, University of California, San Francisco, 4150 Clement Street, San Francisco, CA 94121, USA kristine.yaff[email protected]
Epidemiological studies of sleep and cognition Insomnia Insomnia is characterised by chronic problems with falling asleep, staying asleep, or poor sleep quality that occurs for 1 month or longer, causes substantial functional impairment, does not occur solely in the context of another sleep or mental disorder, and is not associated with substance use or a medical comorbidity.21,22 An estimated 25% of older adults have either insomnia or insomnia symptoms.1,23 In cross-sectional studies of mainly middle-aged (mean age 47 years [SD 15]) adults,24 insomnia has been consistently associated with worse cognitive function; however, only a few studies have assessed whether insomnia in older adults increases the risk of cognitive decline and dementia and results are mixed (table 1). In a cross-sectional study25 of Italian adults aged 65 years or older, insomnia was not associated with cognitive impairment or dementia. By contrast, findings from a retrospective cohort study26 using medical records showed that people with clinically diagnosed insomnia and long-term use of hypnotics were twice as likely to develop dementia over 3 years of follow-up than those without insomnia.26 Conflicting results have also been reported in the two prospective studies that assessed self-reported insomnia and cognitive decline, with findings from one study showing that insomnia is associated with an increased risk of cognitive decline in men and in women who had symptoms of depression,27 and findings from another showing no association for cognitive decline or dementia in older men.28 Discrepant results across studies might be due partly to the heterogeneity of study methods and design, particularly for insomnia measures and cognitive outcomes. Conclusions regarding the direction of the association between insomnia and cognitive decline are difficult to reach because these studies were either crosssectional,25 retrospective,26 or had a short duration of follow-up (≤3 years).27,28 1017
Review
Cortex
Thalamus GABA (–) Basal forebrain Acetylcholine (+), glutamate (+), adenosine (–), nitric oxide (–), GABA (–) Hypothalamus Orexin (+), histamine (+), modulating peptides (–), GABA(–) Brainstem Acetylcholine (+), serotonin (+), noradrenaline (+)
Figure 1: Brain regions and neurotransmitter systems involved in sleep regulation Sleep and wakefulness are regulated by nuclei in the brainstem, hypothalamus, basal forebrain, and thalamus. These regions send projections to the cortex and release excitatory and inhibitory neurotransmitters that help to regulate the sleep–wake cycle.8 (+)=excitatory and (–)=inhibitory.
Panel: Prevalent sleep disturbances in people with dementia Alzheimer’s disease • Circadian rhythm dysfunction: • Fragmented nocturnal sleep • Excessive daytime sleepiness • Phase delay in activity rhythms • Amplitude of melatonin rhythms • Sundowning • Changes in sleep architecture: • Wake after sleep onset • Latency to onset of first rapid eye movement episode • Total sleep time • Sleep efficiency • Rapid eye movement and slow wave sleep • Indeterminate non-rapid eye movement sleep patterns • Sleep-disordered breathing Dementia with Lewy bodies • Rapid eye movement behavioural sleep disorder • Hypersomnia • Periodic limb movements in sleep Vascular dementia • Sleep-disordered breathing Frontotemporal dementia Few studies have reported on sleep disorders in those with frontotemporal dementia; however, there is some evidence to suggest that patients might have a phase advance or an increased prevalence of excessive daytime sleepiness
Structural brain imaging studies have also produced conflicting results, with some,29–31 but not all,32,33 showing an association between insomnia and reduced brain 1018
volumes, including volumes of the hippocampus and orbitofrontal and parietal grey matter. However, these studies were done mainly in middle-aged adults and results might not be generalisable to older age groups. Therefore, whether insomnia is a risk factor for cognitive decline and dementia or whether it is an early marker of disease remains undetermined. Additional prospective studies with a long duration of follow-up and standardised measures of insomnia and cognitive outcomes are needed to help to answer these questions.
Sleep quality Sleep complaints are common in older adults, with around half of older people (≥55 years) reporting problems with initiation or maintenance of sleep.34 Studies of self-reported sleep complaints in older adults have produced mixed results,35–44 but suggest a link between measures of reduced sleep quality and worse cognitive outcomes. In a group of older adults living in the community (ie, not living in nursing homes or assisted living facilities), people with poor sleep quality (measured with the Pittsburgh Sleep Quality Index [PSQI]) had worse performance in tests of working memory, attentional set shifting, and abstract problem solving but not in tests of processing speed, inhibitory function, or episodic memory than those with good sleep quality.35 Findings from another cross-sectional study36 showed reduced performance in measures of verbal knowledge, long-term memory, and visuospatial reasoning associated with self-reported longer sleep onset latency. In women participating in the Nurse’s Health Study,37 those who reported regularly having difficulty falling or staying asleep had worse global cognitive scores in cross-sectional analysis, but changes in cognitive scores over follow-up were not different in longitudinal analysis, than did those who reported that they rarely had difficulty sleeping. Findings from two other cross-sectional studies did not show an association between selfreported sleep quality and cognitive measures.38,39 In four prospective studies using self-reported sleep measures, fi ndings from one study showed that reduced sleep quality was associated with an increased risk of cognitive decline 40 and fi ndings from the other three studies reported an increased risk of cognitive impairment41,42 or dementia.42,43 These studies showed that those who reported sleep disturbances were roughly two to four times more likely to develop a cognitive impairment than were those who did not. However, fi ndings from two other studies showed no association with cognitive decline over 2 years37 and 8 years.44 Thus, most prospective studies suggest that sleep quality is linked to an increased likelihood of cognitive impairment. Although subjective measures of sleep provide a general estimation of sleep quality, the specialty is moving towards the use of objective measures to better www.thelancet.com/neurology Vol 13 October 2014
Review
Sample
Sleep measure
Cognitive measure
Results
Insomnia was determined by interview and defined as reporting ≥1 insomnia symptoms ≥3 times per week
Cognition was measured with the No association MMSE; those with score 9 point drop in CASI score; dementia was assessed by specialist physicians
No association
Cross-sectional studies Merlino et al, 201025
750 adults aged ≥65 years
Retrospective case–control studies Chen et al, 201226
Prospective studies
MMSE=Mini-Mental State Examination. DSM-IV=Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. ICD-9=International Classification of Diseases, Ninth Revision. HR=hazard ratio. SPMSQ=Pfeiffer’s Short Portable Mental Status Questionnaire. OR=odds ratio. CASI=Cognitive Abilities Screening Instrument.
Table 1: Observational studies of insomnia and cognitive function in older adults
detail the association between sleep and cognitive outcomes in older adults. Findings from most studies using actigraphy, which provides an estimate of sleep time and fragmentation by detailing periods of inactivity (indicative of sleep) and activity (indicative of wakefulness), have supported an association between reduced sleep quality and poor cognitive outcomes.39,45–48 In a cross-sectional study of older women (≥65 years), actigraphy measures of disturbed sleep, including lower sleep efficiency (percentage of time in bed spent sleeping after lights were turned off ), higher sleep latency, wake after sleep onset (WASO) time and reported napping for 2 h per day or longer, had a higher risk of cognitive impairment than did those who had normal sleep indicators.45 A similar cross-sectional study in older men living in the community showed reduced cognitive function associated with actigraphymeasured WASO.39 In another study of older women, actigraphy measures of lower sleep efficiency were associated with worse executive function in those with low, but not high, levels of physical activity, suggesting that exercise might hold promise for improvement of cognitive outcomes in those with poor sleep quality.46 In patients with mild cognitive impairment, actigraphic measures of sleep fragmentation have been associated with an increased degree of cognitive impairment.47 Increased sleep fragmentation has also been associated with an increased risk of incident Alzheimer’s disease www.thelancet.com/neurology Vol 13 October 2014
and increased rate of cognitive decline in a prospective study of older adults living in the community.48 Furthermore, fi ndings from studies that have assessed biomarkers associated with risk of Alzheimer’s disease have shown that reduced sleep quality could potentially play a part in dementia pathology. Increased sleep fragmentation has been associated with increased amyloid-β deposition (measured by low CSF concentrations of amyloid-β1–42) in cognitively healthy older adults,49 and better sleep consolidation seems to reduce the incidence of Alzheimer’s disease, cognitive decline, and neurofibrillary tangle density in those with APOE ε4 allele.50 Results from these studies suggest that even mild forms of sleep disturbances might affect the development of cognitive impairment and dementia in older adults.
Sleep duration More than a third of adults report regularly having a shorter or longer sleep duration than the typical duration of 7–9 h of sleep per day,51 and mounting evidence suggests that sleep duration might be predictive of cognitive outcomes in older adults (table 2). In cross-sectional studies36,39,52,53 using self-reported assessments of sleep duration, fi ndings showed associations with reduced cognitive function in participants reporting long sleep duration, whereas another54 showed associations for long and short sleep 1019
Review
Sample
Sleep measure
Cognitive measure
Results
Long sleep duration (>9 h) was associated with reduced verbal short-term memory adjusted for depression, age, education, hypnotic drug use, and comorbidities
Cross-sectional studies Schmutte et al, 200736
375 adults aged 75–85 years
Self-reported nocturnal TST
A battery of ten neuropsychological tests
Blackwell et al, 201139
3132 older men aged ≥65 years
Actigraphy-measured TST and self-reported nocturnal TST sleep
Tests of global and executive function, and Self-reported long (>8 h) sleep duration was associated with worse performance on visual attention all cognitive measures; adjusted for demographics, comorbidities, medication, and health behaviours
Blackwell et al, 200645
2932 women aged ≥65 years
Actigraphy-measured TST
Tests of global and executive function
No association
Faubel et al, 200952
3212 adults aged ≥60 years
Self-reported typical daily TST
Test of global cognitive function
Long (≥11 h) sleep duration was associated with worse global cognition (mean difference –1·48, 95% CI –2·12 to –0·85); adjusted for demographics, comorbidities, and treatments
Ramos et al, 201353
927 adults, mean age of 75 years
Self-reported nocturnal TST
A test of global cognitive function
Long sleep (≥ 9 h) was inversely associated with cognitive performance (adjusted β –0·06, SE 0·03, p=0·012), whereas short sleep was not associated with cognition; adjusted for demographics, comorbidities, and treatment
Xu et al, 201154
28 670 adults aged 50–85 years
Self-reported daily TST
Tests of global cognitive function and memory
Short (3–4 h) and long (≥10 h) sleep durations were associated with memory impairment (OR 1·29, 95% CI 1·07–1·56; and OR 1·52, 95% CI 1·25–1·86, respectively); adjusted for demographics, comorbidities, and health behaviours
Tworoger et al, 200637
1884 women aged 70–81 years
Self-reported daily TST
Global cognitive score comprised of the Z-scores from six cognitive tests
Short (≤5 h) sleep duration was associated with an increased risk of impairment in cross-sectional analysis (OR 2·19, 95% CI 1·10–4·39), but not with change in cognition over 2 years; adjusted for demographics, health behaviours, blood pressure, living alone, mental health index, and tranquiliser use
Potvin et al, 201241
1664 adults aged 65–96 years
Sleep duration was determined by the PSQI
Cognitive decline was defined as a score below 1 SD on the MMSE in a normal distribution; incident cognitive impairment was defined as a loss of at least 2 MMSE points over follow-up
In women, long (≥9 h) sleep duration was associated with impairment (OR 2·10, 95% CI 1·10–4·00); in men, short (≤5 h) sleep duration was associated with impairment (OR 2·91, 95% CI 1·24–6·82); adjusted for age, education, baseline MMSE score, comorbidities, and psychotropic drug use
Benito-Leon et al, 200955
3286 adults aged ≥65 years
Self-reported daily TST
Dementia was assessed by neurological and medical examinations
Long (≥9 h) sleep duration was associated with a greater risk for dementia over 3 years (RR 2·18, 95% CI 1·09–4·37); adjusted for age, education, smoking, and alcohol
Loerbroks et al, 201056
695 adults aged ≥70 years
Self-reported nocturnal TST
A battery of standard cognitive tests
Increases in sleep duration from 7–8 h at baseline to ≥9 h over almost 9 years was associated with greater cognitive impairment (OR 2·1, 95% CI 1·0–4·5) than in those who slept 7–8 h at both timepoints; adjusted for demographics, health behaviours, BMI, sleep treatment, and comorbidities
Ferrie et al, 201157
5413 adults aged 45–69 years
Self-reported nocturnal TST during week nights
A battery of standard cognitive tests
Long (≥9 h), short (≤5 h), and changes in sleep duration in either direction were associated with lower scores on most cognitive tests than for those who slept 7 h; adjusted for demographics and corrected for multiple testing
Virta et al, 201358
2336 adults aged ≥65 years
Self-reported daily TST and hours of nocturnal sleep needed to be alert the following day
Two telephone screening instruments
Short (8 h) sleep durations were associated with lower cognitive scores (β =–0·79, p=0·019; and β=–1·61, p
Connections between sleep and cognition in older adults Kristine Yaffe, Cherie M Falvey, Tina Hoang
Sleep disturbances and cognitive impairment are common in older adults. Mounting evidence points to a potential connection between sleep and cognitive function. Findings from observational studies support a role for sleep disturbances (particularly for sleep duration, sleep fragmentation, and sleep-disordered breathing) in the development of cognitive impairment. Less consistent evidence exists for associations of insomnia and circadian rhythm dysfunction with cognition. These findings suggest that the sleep–wake cycle plays a crucial part in brain ageing, pointing to a potential avenue for improvement of cognitive outcomes in people at risk of cognitive decline and dementia. Several biological mechanisms might underlie the association between sleep and cognition, but these pathways are not completely understood. Future studies that aim to clarify the association between sleep and cognition might help to identify people at risk of cognitive disorders and to facilitate the development of novel therapies to treat and potentially prevent both sleep disturbances and cognitive impairment.
Introduction As people age, changes in sleep patterns often occur, including a decrease in total sleep duration and efficiency, an increase in sleep fragmentation, greater difficulty falling asleep, and less time in rapid eye movement (REM) sleep and slow wave sleep.1–3 Although some of these changes might be a function of normal ageing, others might result from underlying medical conditions4–7 or a neurodegenerative process.2 The sleep–wake cycle is regulated by complex interactions among brain regions and neurotransmitter systems8 (figure 1), many of which are implicated in memory and cognitive function.9–11 Perhaps as a result of this shared circuitry, sleep problems are common in people with Alzheimer’s disease and other dementias9,12–16 (panel). In adults with Alzheimer’s disease, circadian rhythm dysfunction and so-called sundowning (increased confusion and restlessness at the end of the day and into the night) are frequently reported and are thought to result from degeneration of the suprachiasmatic nucleus (known as the master clock regulating circadian rhythms) and the cholinergic neurons of the nucleus basalis of Meynert.9,15 Similarly, REM behavioural sleep disorders often accompany dementia with Lewy bodies and might indicate degeneration of brainstem structures, although the specific anatomical regions have not been determined.15–20 Although an increased prevalence of sleep disturbances in people with dementia has been well documented and is often thought to result from neurodegeneration, focus has turned towards the possibility that sleep disturbances can also increase the risk of cognitive decline and dementia. The aim of this Review is to discuss research investigating how sleep disturbances in older adults (≥55 years) might predict or contribute to cognitive decline and dementia. We present evidence from observational studies of the risk of cognitive decline and dementia associated with sleep disturbances in older adults, including studies of insomnia, sleep quality, sleep duration, excessive daytime sleepiness, sleep-disordered breathing, and circadian rhythm disturbances. We then www.thelancet.com/neurology Vol 13 October 2014
discuss potential biological mechanisms underlying the association followed by results from treatment trials assessing potential cognitive benefits of improved sleep.
Lancet Neurol 2014; 13: 1017–28 Departments of Psychiatry, Neurology, and Epidemiology and Biostatistics, University of California, San Francisco, CA, USA (Prof K Yaffe MD, C M Falvey MPH); and Northern California Institute for Research and Education, San Francisco, CA, USA (T Hoang MPH) Correspondence to: Prof Kristine Yaffe, University of California, San Francisco, 4150 Clement Street, San Francisco, CA 94121, USA kristine.yaff[email protected]
Epidemiological studies of sleep and cognition Insomnia Insomnia is characterised by chronic problems with falling asleep, staying asleep, or poor sleep quality that occurs for 1 month or longer, causes substantial functional impairment, does not occur solely in the context of another sleep or mental disorder, and is not associated with substance use or a medical comorbidity.21,22 An estimated 25% of older adults have either insomnia or insomnia symptoms.1,23 In cross-sectional studies of mainly middle-aged (mean age 47 years [SD 15]) adults,24 insomnia has been consistently associated with worse cognitive function; however, only a few studies have assessed whether insomnia in older adults increases the risk of cognitive decline and dementia and results are mixed (table 1). In a cross-sectional study25 of Italian adults aged 65 years or older, insomnia was not associated with cognitive impairment or dementia. By contrast, findings from a retrospective cohort study26 using medical records showed that people with clinically diagnosed insomnia and long-term use of hypnotics were twice as likely to develop dementia over 3 years of follow-up than those without insomnia.26 Conflicting results have also been reported in the two prospective studies that assessed self-reported insomnia and cognitive decline, with findings from one study showing that insomnia is associated with an increased risk of cognitive decline in men and in women who had symptoms of depression,27 and findings from another showing no association for cognitive decline or dementia in older men.28 Discrepant results across studies might be due partly to the heterogeneity of study methods and design, particularly for insomnia measures and cognitive outcomes. Conclusions regarding the direction of the association between insomnia and cognitive decline are difficult to reach because these studies were either crosssectional,25 retrospective,26 or had a short duration of follow-up (≤3 years).27,28 1017
Review
Cortex
Thalamus GABA (–) Basal forebrain Acetylcholine (+), glutamate (+), adenosine (–), nitric oxide (–), GABA (–) Hypothalamus Orexin (+), histamine (+), modulating peptides (–), GABA(–) Brainstem Acetylcholine (+), serotonin (+), noradrenaline (+)
Figure 1: Brain regions and neurotransmitter systems involved in sleep regulation Sleep and wakefulness are regulated by nuclei in the brainstem, hypothalamus, basal forebrain, and thalamus. These regions send projections to the cortex and release excitatory and inhibitory neurotransmitters that help to regulate the sleep–wake cycle.8 (+)=excitatory and (–)=inhibitory.
Panel: Prevalent sleep disturbances in people with dementia Alzheimer’s disease • Circadian rhythm dysfunction: • Fragmented nocturnal sleep • Excessive daytime sleepiness • Phase delay in activity rhythms • Amplitude of melatonin rhythms • Sundowning • Changes in sleep architecture: • Wake after sleep onset • Latency to onset of first rapid eye movement episode • Total sleep time • Sleep efficiency • Rapid eye movement and slow wave sleep • Indeterminate non-rapid eye movement sleep patterns • Sleep-disordered breathing Dementia with Lewy bodies • Rapid eye movement behavioural sleep disorder • Hypersomnia • Periodic limb movements in sleep Vascular dementia • Sleep-disordered breathing Frontotemporal dementia Few studies have reported on sleep disorders in those with frontotemporal dementia; however, there is some evidence to suggest that patients might have a phase advance or an increased prevalence of excessive daytime sleepiness
Structural brain imaging studies have also produced conflicting results, with some,29–31 but not all,32,33 showing an association between insomnia and reduced brain 1018
volumes, including volumes of the hippocampus and orbitofrontal and parietal grey matter. However, these studies were done mainly in middle-aged adults and results might not be generalisable to older age groups. Therefore, whether insomnia is a risk factor for cognitive decline and dementia or whether it is an early marker of disease remains undetermined. Additional prospective studies with a long duration of follow-up and standardised measures of insomnia and cognitive outcomes are needed to help to answer these questions.
Sleep quality Sleep complaints are common in older adults, with around half of older people (≥55 years) reporting problems with initiation or maintenance of sleep.34 Studies of self-reported sleep complaints in older adults have produced mixed results,35–44 but suggest a link between measures of reduced sleep quality and worse cognitive outcomes. In a group of older adults living in the community (ie, not living in nursing homes or assisted living facilities), people with poor sleep quality (measured with the Pittsburgh Sleep Quality Index [PSQI]) had worse performance in tests of working memory, attentional set shifting, and abstract problem solving but not in tests of processing speed, inhibitory function, or episodic memory than those with good sleep quality.35 Findings from another cross-sectional study36 showed reduced performance in measures of verbal knowledge, long-term memory, and visuospatial reasoning associated with self-reported longer sleep onset latency. In women participating in the Nurse’s Health Study,37 those who reported regularly having difficulty falling or staying asleep had worse global cognitive scores in cross-sectional analysis, but changes in cognitive scores over follow-up were not different in longitudinal analysis, than did those who reported that they rarely had difficulty sleeping. Findings from two other cross-sectional studies did not show an association between selfreported sleep quality and cognitive measures.38,39 In four prospective studies using self-reported sleep measures, fi ndings from one study showed that reduced sleep quality was associated with an increased risk of cognitive decline 40 and fi ndings from the other three studies reported an increased risk of cognitive impairment41,42 or dementia.42,43 These studies showed that those who reported sleep disturbances were roughly two to four times more likely to develop a cognitive impairment than were those who did not. However, fi ndings from two other studies showed no association with cognitive decline over 2 years37 and 8 years.44 Thus, most prospective studies suggest that sleep quality is linked to an increased likelihood of cognitive impairment. Although subjective measures of sleep provide a general estimation of sleep quality, the specialty is moving towards the use of objective measures to better www.thelancet.com/neurology Vol 13 October 2014
Review
Sample
Sleep measure
Cognitive measure
Results
Insomnia was determined by interview and defined as reporting ≥1 insomnia symptoms ≥3 times per week
Cognition was measured with the No association MMSE; those with score 9 point drop in CASI score; dementia was assessed by specialist physicians
No association
Cross-sectional studies Merlino et al, 201025
750 adults aged ≥65 years
Retrospective case–control studies Chen et al, 201226
Prospective studies
MMSE=Mini-Mental State Examination. DSM-IV=Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition. ICD-9=International Classification of Diseases, Ninth Revision. HR=hazard ratio. SPMSQ=Pfeiffer’s Short Portable Mental Status Questionnaire. OR=odds ratio. CASI=Cognitive Abilities Screening Instrument.
Table 1: Observational studies of insomnia and cognitive function in older adults
detail the association between sleep and cognitive outcomes in older adults. Findings from most studies using actigraphy, which provides an estimate of sleep time and fragmentation by detailing periods of inactivity (indicative of sleep) and activity (indicative of wakefulness), have supported an association between reduced sleep quality and poor cognitive outcomes.39,45–48 In a cross-sectional study of older women (≥65 years), actigraphy measures of disturbed sleep, including lower sleep efficiency (percentage of time in bed spent sleeping after lights were turned off ), higher sleep latency, wake after sleep onset (WASO) time and reported napping for 2 h per day or longer, had a higher risk of cognitive impairment than did those who had normal sleep indicators.45 A similar cross-sectional study in older men living in the community showed reduced cognitive function associated with actigraphymeasured WASO.39 In another study of older women, actigraphy measures of lower sleep efficiency were associated with worse executive function in those with low, but not high, levels of physical activity, suggesting that exercise might hold promise for improvement of cognitive outcomes in those with poor sleep quality.46 In patients with mild cognitive impairment, actigraphic measures of sleep fragmentation have been associated with an increased degree of cognitive impairment.47 Increased sleep fragmentation has also been associated with an increased risk of incident Alzheimer’s disease www.thelancet.com/neurology Vol 13 October 2014
and increased rate of cognitive decline in a prospective study of older adults living in the community.48 Furthermore, fi ndings from studies that have assessed biomarkers associated with risk of Alzheimer’s disease have shown that reduced sleep quality could potentially play a part in dementia pathology. Increased sleep fragmentation has been associated with increased amyloid-β deposition (measured by low CSF concentrations of amyloid-β1–42) in cognitively healthy older adults,49 and better sleep consolidation seems to reduce the incidence of Alzheimer’s disease, cognitive decline, and neurofibrillary tangle density in those with APOE ε4 allele.50 Results from these studies suggest that even mild forms of sleep disturbances might affect the development of cognitive impairment and dementia in older adults.
Sleep duration More than a third of adults report regularly having a shorter or longer sleep duration than the typical duration of 7–9 h of sleep per day,51 and mounting evidence suggests that sleep duration might be predictive of cognitive outcomes in older adults (table 2). In cross-sectional studies36,39,52,53 using self-reported assessments of sleep duration, fi ndings showed associations with reduced cognitive function in participants reporting long sleep duration, whereas another54 showed associations for long and short sleep 1019
Review
Sample
Sleep measure
Cognitive measure
Results
Long sleep duration (>9 h) was associated with reduced verbal short-term memory adjusted for depression, age, education, hypnotic drug use, and comorbidities
Cross-sectional studies Schmutte et al, 200736
375 adults aged 75–85 years
Self-reported nocturnal TST
A battery of ten neuropsychological tests
Blackwell et al, 201139
3132 older men aged ≥65 years
Actigraphy-measured TST and self-reported nocturnal TST sleep
Tests of global and executive function, and Self-reported long (>8 h) sleep duration was associated with worse performance on visual attention all cognitive measures; adjusted for demographics, comorbidities, medication, and health behaviours
Blackwell et al, 200645
2932 women aged ≥65 years
Actigraphy-measured TST
Tests of global and executive function
No association
Faubel et al, 200952
3212 adults aged ≥60 years
Self-reported typical daily TST
Test of global cognitive function
Long (≥11 h) sleep duration was associated with worse global cognition (mean difference –1·48, 95% CI –2·12 to –0·85); adjusted for demographics, comorbidities, and treatments
Ramos et al, 201353
927 adults, mean age of 75 years
Self-reported nocturnal TST
A test of global cognitive function
Long sleep (≥ 9 h) was inversely associated with cognitive performance (adjusted β –0·06, SE 0·03, p=0·012), whereas short sleep was not associated with cognition; adjusted for demographics, comorbidities, and treatment
Xu et al, 201154
28 670 adults aged 50–85 years
Self-reported daily TST
Tests of global cognitive function and memory
Short (3–4 h) and long (≥10 h) sleep durations were associated with memory impairment (OR 1·29, 95% CI 1·07–1·56; and OR 1·52, 95% CI 1·25–1·86, respectively); adjusted for demographics, comorbidities, and health behaviours
Tworoger et al, 200637
1884 women aged 70–81 years
Self-reported daily TST
Global cognitive score comprised of the Z-scores from six cognitive tests
Short (≤5 h) sleep duration was associated with an increased risk of impairment in cross-sectional analysis (OR 2·19, 95% CI 1·10–4·39), but not with change in cognition over 2 years; adjusted for demographics, health behaviours, blood pressure, living alone, mental health index, and tranquiliser use
Potvin et al, 201241
1664 adults aged 65–96 years
Sleep duration was determined by the PSQI
Cognitive decline was defined as a score below 1 SD on the MMSE in a normal distribution; incident cognitive impairment was defined as a loss of at least 2 MMSE points over follow-up
In women, long (≥9 h) sleep duration was associated with impairment (OR 2·10, 95% CI 1·10–4·00); in men, short (≤5 h) sleep duration was associated with impairment (OR 2·91, 95% CI 1·24–6·82); adjusted for age, education, baseline MMSE score, comorbidities, and psychotropic drug use
Benito-Leon et al, 200955
3286 adults aged ≥65 years
Self-reported daily TST
Dementia was assessed by neurological and medical examinations
Long (≥9 h) sleep duration was associated with a greater risk for dementia over 3 years (RR 2·18, 95% CI 1·09–4·37); adjusted for age, education, smoking, and alcohol
Loerbroks et al, 201056
695 adults aged ≥70 years
Self-reported nocturnal TST
A battery of standard cognitive tests
Increases in sleep duration from 7–8 h at baseline to ≥9 h over almost 9 years was associated with greater cognitive impairment (OR 2·1, 95% CI 1·0–4·5) than in those who slept 7–8 h at both timepoints; adjusted for demographics, health behaviours, BMI, sleep treatment, and comorbidities
Ferrie et al, 201157
5413 adults aged 45–69 years
Self-reported nocturnal TST during week nights
A battery of standard cognitive tests
Long (≥9 h), short (≤5 h), and changes in sleep duration in either direction were associated with lower scores on most cognitive tests than for those who slept 7 h; adjusted for demographics and corrected for multiple testing
Virta et al, 201358
2336 adults aged ≥65 years
Self-reported daily TST and hours of nocturnal sleep needed to be alert the following day
Two telephone screening instruments
Short (8 h) sleep durations were associated with lower cognitive scores (β =–0·79, p=0·019; and β=–1·61, p
