Assessment of sleep and circadian rhythm disorders

27. Charlson ME, Pompei P, Ales KL et al.. A new method of clas- 30. Manderbacka K. Examining what self-rated health question is
sifying prognostic comorbidity in longitudinal studies: devel- understood to mean by respondents. Scand J Soc Med 1998;
opment and validation. J Chronic Dis 1987; 40: 373–83. 26: 145–53.
28. Wolinsky FD, Callahan CM, Fitzgerald JF et al.. Changes in 31. Walker JD, Maxwell CJ, Hogan DB et al.. Does self-rated
functional status and the risks of subsequent nursing home health predict survival in older persons with cognitive impair-
placement and death. J Gerontol 1993; 48: S94–101. ment? J Am Geriatr Soc 2004; 52: 1895–900.
29. Idler EL, Kasl S. Health perceptions and survival: do global
evaluations of health status really predict mortality? Journal of Received 29 November 2012; accepted in revised form
Gerontology 1991; 46: S55–65. 22 May 2013

Age and Ageing 2014; 43: 57–63 © The Author 2013. Published by Oxford University Press on behalf of the British Geriatrics Society.
doi: 10.1093/ageing/aft153 All rights reserved. For Permissions, please email: journals.permissions@oup.com

Assessment of sleep and circadian rhythm

disorders in the very old: the Newcastle 85+
Cohort Study
KIRSTIE N. ANDERSON, MICHAEL CATT, JOANNA COLLERTON, KAREN DAVIES, THOMAS VON ZGLINICKI,
THOMAS B. L. KIRKWOOD, CAROL JAGGER
Institute for Ageing and Health, Newcastle University, Campus for Ageing and Vitality, Newcastle upon Tyne NE4 5PL, UK
Address correspondence to: K. N. Anderson. Tel: +44(0)1912823833; Fax: +44(0)1912825027. Email: kirstie.anderson@nuth.nhs.uk

Abstract

Objectives: to examine the association between subjective and objective measures of sleep and wake and other health para-
meters in a cohort of the very old.
Design: a population-based cohort study.
Setting: primary care, North East England.
Participants: four hundred and twenty-one men and women, aged 87–89, recruited to the Newcastle 85+ Study cohort.
Methods: sleep questionnaires were administered and sleep–wake patterns were assessed over 5–7 days with a novel wrist tri-
axial accelerometer. Associations between sleep measures and various health parameters, including mortality at 24 months,
were examined.
Results: only 16% of participants perceived their sleep as severely disturbed as assessed with questionnaire responses. Wrist accel-
erometry showed marked variation between normal and abnormal sleep–wake cycles that did not correlate with the participants’
perception of sleep. Impaired sleep–wake cycles were significantly associated with cognitive impairment, disability, depression,
increased falls, body mass index and arthritis but not with any other specific disease markers and with decreased survival.
Conclusions: commonly used sleep questionnaires do not differentiate well between those with objectively determined disturbance
of sleep–wake cycles and those with normal cycles. Abnormal sleep–wake patterns are associated with institutionalisation, cognitive
impairment, disability, depression and arthritis but not with other diseases; there is also an association with reduced survival.

Keywords: accelerometry, sleep, 85+, circadian rhythm, novel accelerometer, impaired sleep and mortality, lder people

Introduction have at least one chronic sleep-related problem with high rates
of sleep apnoea, insomnia and restless legs [1, 2]. Changes in
Sleep complaints and disruptions in the sleep–wake cycle are circadian rhythms can be demonstrated with advancing age,
common in older people. Over 50% of adults aged over 65 with a decline in the cortisol and melatonin rhythms that

57
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 K. N. Anderson et al.

entrain day–night activity patterns [3, 4]. With ageing there is within their normal home environment. The GENEA records
reduced amplitude of the circadian rhythm, a phase shift (in continuous acceleration rather than epoch-based summary
particular phase advance) [5] and a loss of the robustness of measures more typical of studies employing commercial acti-
the rhythm. However, within older populations there is signifi- watches (actigraphy) and has been validated for monitoring ac-
cant variability between individuals with some hypothesising tivity levels [13]. Only participants who had worn the watch
an age-related disruption to the suprachiasmatic nucleus [6]. continuously for 5 days or more within the 7-day period were
Poor sleep in older adults is associated with an increase in included in the analysis. In those with seven full days of moni-
mortality, accidents and falls and a decline in health status toring, 5 and 7 day measures were compared and no significant
[2]. There is a particular association between disrupted circa- differences were found.
dian rhythms and dementia [7]. For the very old, variously The following accelerometry measures were obtained: the
defined as those aged 80+ or 85+, sleep and circadian least active 5 h period (L5); the most active 5 h period (M5);
rhythm data are limited yet these are now the fastest growing the mean rest-active transition count in the least active 5 h
age sector of the population [8]. (L5_TN) and the difference between M5 and L5 (ΔM5L5).
We have assessed the sleep and wake patterns of a large, L5 and M5 activity counts were averaged for the entire 5-day
well characterised UK cohort of people aged 87–89 years, in- period.
cluding in institutional care, and report the extent of sleep Participants were divided into quartiles for each accelero-
disorders, the contribution of different diseases and condi- metry measure and then the middle two quartiles were
tions to disrupted sleep and whether very old individuals grouped together to represent low, medium and high activity
with sleep disorders have a higher risk of mortality. counts for all the subsequent analyses. We investigated three
objective measures of sleep disturbance, defining abnormal
sleep patterns as those in the lowest quartile of ΔM5L5and
Methods the highest quartiles of L5 and L5_TN.
Participants
The study was nested in the Newcastle 85+ Study, a Morbidity and mortality measures
population-based longitudinal study of health and ageing in the Measures assessed for association with sleep disorders were
very old [9]. People living in Newcastle or North Tyneside (NE those indicated in younger populations: disability (difficulties
England) were recruited at around age 85 through general in 17 daily activities coded as 0 difficulties/1-6/7-12/13-17),
practitioner patient lists; those living in institutions and the cog- depressive symptomatology using the Geriatric Depression
nitively impaired were included. Following baseline assessment Scale (ref) and coded as none/mild/severe, cognition assessed
(Phase 1: 2006–7, n = 849) Newcastle 85+ Study participants by the Mini-Mental State Examination (MMSE), number of
were re-assessed at 18 months (Phase 2: 2007–9, n = 630) and falls in last year (none/1/2/3+) and body mass index (BMI),
again at 36 months (Phase 3: 2009–10, n = 484). Loss between (assessed at 36 months). Data on pre-existing diseases were
phases 1 and 3 was mainly due to deaths (63.6%, 232/365) extracted from the general practice medical records and
with the remainder due to drop out. Participants were invited included presence at baseline, of ischaemic heart disease, arth-
to take part in this sleep study as part of the 36-month follow- ritis, cerebrovascular disease, cancer (excluding non-melanoma
up assessment when participants were aged 87–89 years. All skin cancer) in last 5 years, chronic obstructive airways disease,
assessments were conducted in the participant’s usual resi- as well as a total chronic disease count (18 diseases and condi-
dence (home or institution). Ethical approval was obtained tions coded as none/1–2/3–6/7+) [9]. The National Health
from Newcastle and North Tyneside Local Research Ethics Service Medical Research Information Service provided the
Committee. Written informed consent was obtained from par- date of death. Survival time was measured from date of
ticipants and where people lacked the capacity to consent, a actigraphy to death or censored at 1/8/2011 (mean follow-up
formal written opinion was sought from a relative/carer in ac- 24 months).
cordance with the UK Mental Capacity Act [10].
Statistical analysis
Sleep questionnaires Associations between measures of disturbed sleep and socio-
A research nurse administered the Pittsburgh Sleep demographic and health characteristics of participants were
Questionnaire Inventory (PSQI) [11] and the Epworth assessed by Mann–Whitney U tests (ESS, PSQI) or Kruskal–
Sleepiness Score (ESS) [12] as subjective measures of sleep Wallis tests (accelerometry measures in three categories) for
and wake. These well-validated questionnaires assess sleep– ordinal data (disability score, disease count, depression, BMI,
wake disturbance and daytime sleepiness, respectively. self-rated health, MMSE, number of falls) and Chi-square
tests for categorical variables (sex, where living, presence of
individual diseases). The associations between accelerometry
Accelerometry measures measures of disturbed sleep and subjective measures (ESS
Participants wore a triaxial accelerometer (GENEA, Unilever) and PSQI scores) were assessed by Kruskal–Wallis tests. The
on their right wrist continuously for a period of 5–7 days impact of objective and subjective sleep measures on

58
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 Assessment of sleep and circadian rhythm disorders

subsequent survival was investigated by Cox proportional Association between accelerometry measures of
hazards regression models with adjustment initially for sex, disturbed sleep and subjective sleep measures
then added disease count, disability, cognitive function and Of the objective measures of abnormal sleep patterns only
BMI. P-values were two-sided; the level of significance was ΔM5L5 was associated with significantly higher ESS scores
0.05. Missing values were excluded from the data set and per- (Mann–Whitney U test, P < 0.007), most participants report-
centages were calculated from the number of valid ing daytime sleepiness having little difference in levels of activ-
responses. ity between the most and least active 5 h (low ΔM5L5)
(Supplementary data are available in Age and Ageing online,
Table S4). In addition participants reporting daytime sleepiness
Results were significantly more likely to have low activity levels in the
Sleep questionnaires most active 5 h (low M5) (Mann–Whitney U test, P = 0.012).
The global PSQI score was not significantly associated
Four hundred and twenty-seven/484 participants completed with any of the accelerometry measures of disturbed sleep
both PSQI and ESS questionnaires without any missing data; (Supplementary data are available in Age and Ageing online,
63.5% were women (156 men: 271 women), the majority Table S4), although those with poorer sleep quality (PSQI
(80.3%) lived independently at home, 41% rated their health component 1) were significantly more likely to have higher
as excellent or very good, 7.3% had no difficulty with activities activity levels in the least active 5 h (high L5 activity) (Kruskal–
of daily living and 69.9% had no cognitive impairment Wallis test, P = 0.014) (data not shown).
(MMSE score 26–30). Some disturbance of sleep (PSQI >5)
was reported by 50.1% although only 15.7% had severe sleep
disturbance (PSQI >10). Daytime sleepiness (ESS >10) was Association between accelerometry measures of
reported by only 10.8% (Supplementary data are available in disturbed sleep and other health parameters
Age and Ageing online, Table S1). Participants with abnormal sleep–wake cycles defined by little
Those reporting disturbed sleep (PSQI >5) were more difference in activity between the most and least active periods
likely to be female, to report poorer self-rated health, to have (ΔM5L5) were significantly more likely to have poorer cogni-
depressive symptomatology, were less likely to have cerebro- tive function, depressive symptomatology, more falls, be obese,
vascular disease but had a similar prevalence of other diseases, arthritis, higher disease count and more disability (Table 1).
disability, cognitive impairment and obesity as participants not Neither of the two other objective measures of sleep abnor-
experiencing disturbed sleep (Supplementary data are available mality (L5 and L5_TN) appeared to be significantly associated
in Age and Ageing online, Table S2). Participants who had sig- with sociodemographic characteristics, disease burden or
nificant daytime sleepiness (ESS >10) were more likely to have functioning.
higher levels of disability and depressive symptomatology than
those without daytime sleepiness (Supplementary data are
available in Age and Ageing online, Table S3). Subjective and objective measures of sleep and
mortality
Accelerometry Only abnormal sleep–wake cycles, defined by ΔM5L5, was
significantly associated with reduced survival after adjust-
Of the 427 participants with complete sleep scores, 337 (79%) ment for sex (P = 0.002). After further adjustment for dis-
wore the accelerometer for at least five continuous days. When ability, cognitive function, disease count and BMI ΔM5L5
compared with those with accelerometry data (n = 337, 130 remained significantly associated with subsequent 2-year
M: 207 F), those without accelerometry (n = 90) were more mortality (P = 0.018) (Figure 2), participants with abnormal
likely to live in an institution (χ2, P < 0.001), more disabled sleep-wake cycles (low ΔM5L5) having over 3 times the risk
(Mann–Whitney U, P = 0.003), had worse cognitive function of death (hazard ratio: 3.18, 95% CI: 1.25–8.09) compared
(Mann–Whitney U, P < 0.001) and reported more falls with participants having high differentiation between active
(Mann–Whitney U, P = 0.04) although PSQI and ESS scores and inactive periods (high ΔM5L5). Neither of the two other
were similar (data not shown). objective measures or the subjective measures were signifi-
Accelerometry provided a clear representation of daily cantly associated with survival (L5: P = 0.91; L5_TN
activity and a robust assessment of sleep–wake cycles with P = 0.23; PSQI: P = 0.77; ESS P = 0.18).
decreased activity during the night and increased activity
during the day. There was marked variation between parti-
cipants, with some individuals being much more active Discussion
during the day and demonstrating much greater differences
between day and night time activity patterns. Data from This population-based study is the first to subjectively and
two participants are shown in Figure 1. Although the total objectively assess the sleep and circadian rhythms of a UK
amount of physical activity varies between the two indivi- cohort of very old people (87–89 years) and report the rela-
duals, it is possible to distinguish a sleep-wake pattern tionship with morbidity, functioning and mortality.
in both. Significantly disturbed sleep (PSQI >10) was reported by

59
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 K. N. Anderson et al.

Figure 1. Accelerometry Data. Accelerometry data over a 7-day period of continuous recording with two different participants.
Both participants have cycles of decreased activity corresponding to a sleep period (marked with arrow) and increased activity during
wake. However, the total amount of activity varies between the two participants.

only 16% of the very old and significant daytime sleepiness and with both polysomnography and accelerometry [14–16].
by only 10.8%. Depressive symptomatology was higher in In these studies, the PSQI was better than the ESS at distin-
those with disturbed sleep and those with daytime sleepiness guishing those with normal sleep from those with sleep dis-
but no other disease or cognitive impairment was significant- orders. Raised PSQI score was associated with female sex
ly associated with the presence of reported sleep disorders. and depressive symptoms, and this was confirmed by our
In contrast, very old people with disturbed sleep–wake study. Other studies have demonstrated that the PSQI and
cycles, objectively assessed using accelerometry, were more ESS correlate poorly with objective measures of sleep assess-
likely were had significantly worse cognitive function, higher ment in older populations [17]. There was only a weak associ-
BMI, more falls, more depressive symptoms, and were more ation between proven sleep apnoea and elevated ESS in the
likely to have arthritis and a higher number of diseases in total. large MrOS cohort study (mean age of 75.6) and no associ-
We found no evidence of association between the objective ation with PSQI [18, 19]. In the MrOS study 13% of men
and subjective measures of sleep disturbance with the excep- had ESS >10 which matches the low reporting of significant
tion of the difference in activity levels between the most and sleep disturbance in our cohort (16% PSQI >10 and 11%
least active periods (ΔM5L5) and the ESS. However, partici- ESS >10). There are a number of possible explanations for
pants with little difference in activity between the most and low reporting of subjective sleep disturbance: this population
least active periods (ΔM5L5) had significantly worse survival may have different expectations of sleep and wake; cognitive
even after adjustment for potential confounding factors. impairment may have affected the accurate completion of
Previous studies have suggested that the ESS and PSQI the questionnaires and the large number of participants who
reliably detect sleep disturbance in adults over 60 both alone lived and slept alone may have affected accurate reporting for

60
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 Assessment of sleep and circadian rhythm disorders

Table 1. Health measures by objective measure of disturbed sleep

Difference between M5 L5 activity (ΔM5L5)
Low Medium Higha P-value
Percent (n) Percent (n) Percent (n)
....................................................................................
Sex
Female 53.0 (44) 61.0 (105) 70.7 (58) 0.06b
Where living
Standard housing 78.3 (65) 83.6 (143) 85.4 (70) 0.06b
Sheltered with warden 13.3 (11) 13.5 (23) 14.6 (12)
Institution 8.4 (7) 2.9 (5) 0.0 (0)
Self-rated health
Excellent/very good 32.9 (27) 43.5 (74) 48.8 (40) 0.13c
Good 45.1 (37) 35.9 (61) 36.6 (30)
Fair/poor 22.0 (18) 20.6 (35) 14.6 (12)
Disability score
Fully independent (0) 3.7 (3) 5.8 (10) 13.4 (11) <0.001c
1–6 29.3 (24) 60.8 (104) 63.4 (52)
7–12 42.7 (35) 26.3 (45) 22.0 (18)
13–17 24.4 (20) 7.0 (12) 1.2 (1)
Mini-Mental State Examination score
No cog imp (26–30) 60.2 (50) 74.4 (128) 85.4 (70) <0.001c
Mild cog imp (22–25) 24.1 (20) 22.1 (38) 9.8 (8)
Mod cog imp (18–21) 10.8 (9) 3.5 (6) 4.9 (4)
Severe cog imp (0–17) 4.8 (4) 0.0 (0) 0.0 (0)
Geriatric depression score
None 64.2 (52) 83.7 (144) 88.9 (72) <0.001c
Mild 21.0 (17) 12.2 (21) 4.9 (4)
Severe 14.8 (12) 4.1 (7) 6.2 (5)
Number of falls
None 47.6 (39) 52.9 (91) 62.2 (51) 0.04c
1 19.5 (16) 22.1 (38) 24.4 (20)
2 17.1 (14) 12.8 (22) 9.8 (8)
3+ 15.9 (13) 12.2 (21) 3.7 (3)
Body mass index
<18.5 12.9 (9) 9.0 (15) 14.8 (12) 0.002c
18.5–<25 32.9 (23) 54.2 (90) 58.0 (47)
25–<30 35.7 (25) 29.5 (49) 27.2 (22)
30+ 18.6 (13) 7.2 (12) 0.0 (0)
Presence of
Arthritis 69.9 (58) 57.0 (98) 47.6 (39) 0.01b
Ischaemic heart disease 36.1 (30) 30.8 (53) 29.3 (24) 0.59b
Cerebrovascular disease 18.1 (15) 15.1 (26) 19.5 (16) 0.65b
Cancer <5 years 6.0 (5) 7.0 (12) 6.1 (5) 0.94b
Chronic obstructive airways disease 16.9 (14) 18.6 (32) 12.2 (10) 0.44b
Disease count
None 0.0 (0) 0.0 (0) 0.0 (0) 0.015c
1–2 9.0 (7) 6.1 (10) 19.0 (15)
3–6 76.9 (60) 84.0 (137) 74.7 (59)
7+ 14.1 (11) 9.8 (16) 6.3 (5)
a
Corresponds to abnormal sleep–wake cycle.
b
Chi-square test.
c
Kruskal–Wallis test.

some responses. Our results suggest that these widely used studies across community-based populations described the in-
screening tools do not accurately identify disrupted sleep and creasingly fragmented sleep that occurred during ageing, with
wake measured more objectively in the very old, although a tendency for older subjects to have an advanced sleep phase
our findings of associations between sleep disturbance and compared with younger subjects in several large population
both female gender and depressive symptoms confirm these studies [3, 5]. Limitations of previous studies include the vari-
across a broader age range. ability and lack of standardisation of analysis methods, al-
Accelerometry or actigraphy has been used to assess sleep though many use non-parametric circadian rhythm analysis.
and wake patterns for many years and is well established in This is the first study to use continuous wave accelerometry in
both clinical practice and research (for review 20). Early a very old population; this technique allows for analysis of raw

61
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 K. N. Anderson et al.

further informed causes of fragmented sleep, though this

was likely to have overburdened very old adults and reduced
study participation. Finally, both sleep questions and accel-
erometry were only included in the 3-year follow-up when
participants were aged 87–89. However, the baseline cohort
was broadly representative of 85 year olds in England and
Wales [9] and as the main reason for withdrawal from the
study was death, the surviving cohort should also be repre-
sentative of 87–89 year olds.
In conclusion, reported sleep disturbance affects a minor-
ity of the very old. However, these commonly used sleep
questionnaires do not differentiate well between those
normal and abnormal sleep–wake cycles determined
objectively.
This study confirms that objectively measured, abnormal
sleep–wake cycles are strongly correlated with cognitive im-
pairment, depressive symptoms, falls, higher BMI and arth-
ritis (and higher number of diseases in total) and also with
increased mortality in the very old.

Figure 2. Survival following actigraphy monitoring (mean Key points

follow-up 24 months), adjusted for sex, disability, cognitive func-
tion, disease count and BMI. • Primary sleep disorders are common and increase with age.
• The 85+ report little subjective sleep disturbance.
• Objective sleep disturbance measured with actigraphy is
data enabling systematic exploration of alternative measures common.
with potentially greater sensitivity appropriate for a group with • Objective sleep disturbance is associated with increased
reduced total physical activity. mortality.
By far the largest studies using accelerometry in the
elderly to date has been the US, multicentre MrOS and SOF
studies that recruited over 6,000 community-dwelling men
and women over 67 years of age and used at least five or
more days of accelerometry to assess rest activity rhythms Acknowledgements
and sleep–wake cycles in relation to a wide variety of health
parameters. In women (mean age 83.5 years) actigraphically Thanks are especially due to the older people of Newcastle
measured disturbed sleep was consistently related to poorer and North Tyneside for the generous donation of their time
cognition although total sleep time was not [21]. A subset of and personal information to make the study possible. We
younger participants (mean age 67.5) were followed up for appreciate the support of NHS North of Tyne (Newcastle
3.5 years and a small but definite association between dis- Primary Care Trust) and local general practices. We thank
turbed rest/activity rhythms and increased all-cause and car- Sally Barker, June Edwards, Joan Hughes and Judith Hunt
diovascular mortality was found [22]. In men, there has (research nurses); Pauline Potts (data manager) and Lucy
recently been an association described between frailty, poor Farfort ( project secretary).
objective sleep measures and increased mortality [23]. In our
cohort, we found reduced survival in those with abnormal
sleep–wake cycles and decreased activity during the most
active period of the day (M5). Funding
Strengths of the study included its population-based
sample allowing assessment of those living independently The core Newcastle 85+ Study was supported by the UK
and within institutions, and those with cognitive impairment. Medical Research Council and the Biotechnology and
Obtaining information on diseases from medical records Biological Sciences Research Council (grant reference
increased data reliability [9]. The novel triaxial accelerometer G0500997), the Dunhill Medical Trust (grant reference
allowed greater sensitivity in data analysis compared with R124/0509) and NHS North of Tyne (Newcastle Primary
existing commercial devices. Our study has some limitations: Care Trust). The sleep study was supported by the UK
few participants had abnormal PSQI and ESS which reduces NIHR Biomedical Research Centre for Age and Age related
our ability to detect weaker associations; sleep diaries were disease award to the Newcastle upon Tyne Hospitals NHS
not completed alongside accelerometry which may have Foundation Trust (grant reference BH091228).

62
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017
 Assessment of sleep and circadian rhythm disorders

Supplementary data 13. Esliger DW, Rowlands AV, Hurst TL, Catt M, Murray P, Eston
RG. Validation of the GENEA Accelerometer. Med Sci Sports
Supplementary data mentioned in the text is available to sub- Exerc 2011; 43: 1085–93.
scribers in Age and Ageing online. 14. Buysse DJ, Cheng Y, Germain A et al. Night-to-night sleep
variability in older adults with and without chronic insomnia.
Sleep Med 2010; 11: 56–64.
15. Spira AP, Beaudreau SA, Stone KL et al. Reliability and validity
References of the Pittsburgh Sleep Quality Index and the Epworth
Sleepiness Scale in older men. Osteoporotic fractures in Men
1. Foley DJ, Monjan AA, Brown SL, Simonsick EM, Wallace RB, study. J Gerontol A Biol Sci Med Sci 2012; 67: 433–9.
Blazer DG. Sleep complaints among elderly persons: an epide- 16. Beaudreau SA, Spira AP, Stewart A et al. Validation of the
miologic study of three communities. Sleep 1995; 18: 425–32. Pittsburgh Sleep Quality Index and the Epworth Sleepiness
2. Ancoli-Israel S. Sleep and its disorders in ageing populations. Scale in older black and white women. Study of Osteoporotic
Sleep Med 2009; 10(Suppl. 1): 7–11. Fractures. Sleep Med 2012; 13: 36–42.
3. Lieberman HR, Wurtman JJ, Teicher MH. Circadian rhythms 17. Buysse DJ, Hall ML, Strollo PJ et al. Relationships between the
of activity in healthy young and elderly humans. Neurobiol Pittsburgh Sleep Quality Index (PSQI), Epworth Sleepiness
Aging 1989; 10: 259–65. Scale (ESS), and clinical/polysomnographic measures in a
4. Mazzoccoli G, Vendemiale G, La Viola M et al. Circadian varia- community sample. J Clin Sleep Med 2008; 4: 563–71.
tions of cortisol, melatonin and lymphocyte subpopulations in 18. Kezirian EJ, Harrison SL, Ancoli-Israel S et al. Behavioral cor-
geriatric age. Int J Immunopathol Pharmacol 2010; 23: 289–96. relates of sleep-disordered breathing in older men. Study of
5. Monk TH, Buysse DJ, Billy BD et al. Circadian type and bed- Osteoporotic in Men Research Group. Sleep 2009; 32: 253–61.
timing regularity in 654 retired seniors: correlations with sub- 19. Kezirian EJ, Harrison SL, Ancoli-Israel S et al. Behavioral cor-
jective sleep measures. Sleep 2011; 34: 235–9. relates of sleep-disordered breathing in older women. Study of
6. Hu K, Van Someren EJ, Shea SA, Scheer FA. Reduction of Osteoporotic Fractures Research Group. Sleep 2007; 30:
scale invariance of activity fluctuations with aging and alzhei- 1181–8.
mer’s disease: involvement of the circadian pacemaker. Proc 20. Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W,
Natl Acad Sci USA 2009; 106: 2490–4. Pollak CP. The role of actigraphy in the study of sleep and cir-
7. Hatfield CF, Herbert J, van Someren EJ, Hodges JR, Hastings cadian rhythms. Sleep 2003; 26: 342–92.
MH. Disrupted daily activity/rest cycles in relation to daily 21. Blackwell T, Yaffe K, Ancoli-Israel S, Redline S et al.
cortisol rhythms of home-dwelling patients with early Association of sleep characteristics and cognition in older
Alzheimer’s dementia. Brain 2004; 127: 1061–74. community-dwelling men: the MrOS sleep study. Osteoporotic
8. United Nations Department of Economic and Social Affairs. Fractures in Men (MrOS) Study Group. Sleep 2011; 34:
Population Division. World Population Ageing: 2009: United 1347–56.
Nations, 2009. 22. Paudel ML, Taylor BC, Ancoli-Israel S et al. Rest/activity
9. Collerton J, Barrass K, Bond J et al. The Newcastle 85+ study: rhythms and mortality rates in older men: MrOS Sleep Study.
biological, clinical and psychosocial factors associated with Osteoporotic Fractures in Men (MrOS) Study. Chronobiol Int
healthy ageing: study protocol. BMC Geriatr 2007; 7: 14. 2010; 27: 363–77.
10. Parliament: Mental Capacity Act. Stationery office. 2005. 23. Ensrud KE, Blackwell TL, Ancoli-Israel S et al. Sleep distur-
11. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer bances and risk of frailty and mortality in older men. Sleep
DJ. The Pittsburgh Sleep Quality Index: a new instrument for Med 2012; 13: 1217–25.
psychiatric practice and research. Psychiatry Res 1989; 28:
193–213.
12. Johns MW. Sleepiness in different situations measured by the Received 29 November 2012; accepted in revised form
Epworth sleepiness scale. Sleep 1994; 17: 703–10. 22 May 2013

63
Downloaded from https://academic.oup.com/ageing/article-abstract/43/1/57/24018
by guest
on 19 November 2017