Neurol Sci
DOI 10.1007/s10072-017-2975-9
REVIEW ARTICLE
Sleep disturbance in mild cognitive impairment: a systematic
review of objective measures
Mingyue Hu 1 & Ping Zhang 1 & Chen Li 1 & Yongfei Tan 2 & Guichen Li 1 & Duo Xu 1 &
Li Chen 1,2
Received: 21 November 2016 / Accepted: 19 April 2017
# Springer-Verlag Italia 2017
Abstract Sleep disturbance frequently occurs in patients with
mild cognitive impairment (MCI) and appears to be involved
in the cellular and molecular mechanisms of cognitive decline.
The aim of this systematic review is to clarify whether patients
with MCI demonstrate alterations in certain sleep parameters:
total sleep time (TST), sleep efficiency (SE), sleep latency
(SL), rapid eye movement latency (REML), percent of rapid
eye movement (REM%), arousal index (AI), wake after sleep
onset (WASO), slow-wave sleep (SWS), periodic leg move-
ment in sleep (PLMS), and cyclic alternating pattern (CAP)
through polysomnography (PSG) and actigraphy. Databases
including PubMed, Web of Science, Embase, ScienceDirect,
Cochrane, CBM, CNKI, Wanfang Data, and VIP were
searched up to January 2016 to collect literature on the corre-
lation between sleep disturbance and MCI as assessed by ob-
jective measures. Meta-analysis was conducted using the
Review Manager 5.3 software. A total of ten case-control
studies involving 225 MCI patients and 235 healthy elders
(HE) were deemed eligible and included in our meta-analysis.
Every type of sleep disturbance was present in our studies
with significant differences in the MCI subtypes. Compared
with HE, overall MCI patients as a group expressed more SL
and less SE; MCI patients showed less TST and SE and more
SL and CAP; patients with amnestic mild cognitive impair-
ment (aMCI) had less AI; patients with non-amnestic mild
Mingyue Hu and Ping Zhang contributed equally to this work.
* Li Chen
chen_care@126.com
1
College of Nursing, Jilin University, 965 Xinjiang Road,
Changchun 130021, China
2
Department of Pharmacology, College of Basic Medical Sciences,
Jilin University, Changchun, China
cognitive impairment (naMCI) had more TST and less AI.
Patients with naMCI expressed more AI than those with
aMCI. The results indicate that MCI patients might experience
more serious sleep disturbance and that different MCI sub-
types have different patterns of sleep disturbance.
Keywords Mild cognitive impairment . Sleep disturbance .
PSG . Actigraphy . Meta-analysis
Introduction
In the World Alzheimer Report 2015, it is said that there are
over 9.9 million new cases of dementia each year worldwide,
implying one new case every 3.2 s [1]. Mild cognitive impair-
ment (MCI) is viewed as a transitional stage from normal
health to dementia [2, 3], with a conversion rate of
60.5~100% to dementia in 5~10 years [2]. Given the limita-
tions of therapeutic methods for dementia, an understanding
of the clinical features of MCI is crucial for the development
of preventive projects and early interventions.
Previous and recent studies all show that sleep disturbance
is one of the major causes of MCI [3, 4]. Researchers are also
investigating early predictors of MCI and have identified all
kinds of markers [58], and plenty of findings have suggested
that sleep disturbance is a risk factor of MCI or that MCI
patients experienced a more serious degree of sleep distur-
bance than healthy elders (HE) [2, 7, 8].
Sleep disturbance can be evaluated by subjective and ob-
jective measures [8]. For objective measures, medical detect-
ing and monitoring equipment has been widely used.
Actigraphy is highly correlated with polysomnography
(PSG) for detecting sleep states [8, 9]. The present review
includes a variety of results regarding the use of PSG and
actigraphy methods to evaluate sleep disturbance [1012].

Hence, our study summarized objective measures to evaluate
sleep disturbance and tried to explain the association between
sleep disturbance and MCI.
Materials and methods
Inclusion criteria
The PICOS approach was used for eligibility criteria [13]: (1)
Study participants should be diagnosed with MCI; (2) Control
groups should contain HE with normal cognition; (3) The
detection measures of the included articles should include
PSG and/or actigraphy; (4) The outcome should cover sleep
parameters; and (5) The continuous data and the 95% confi-
dence interval (CI) should be provided.
Exclusion criteria
(1) The sample did not meet all the PICOS standards; (2) The
study did not refer to a history of neurological conditions,
psychiatric disorders, and/or major medical illness; the use
of medication affecting the sleep-wake cycle (benzodiaze-
pines, etc.); the presence of depressive symptoms; complaints
of sleep-disordered breathing; movement disorders during
sleep; or unusual sleep schedules (i.e., shift work); (3) The
sample was a duplicate of another published paper without
novel information, although additional studies from the same
sample were included if they represented different outcome
variables; and (4) The study has missing data, and even
through a variety of methods, we were unable to contact the
authors.
Data source and search strategy
A systematic literature search was conducted to identify studies
that evaluated sleep disturbance in patients with MCI compared
to cognitively normal controls, and the measurement methods
were required to contain actigraphy or PSG as an objective mea-
sure. The databases include PubMed, Web of Science, Embase,
ScienceDirect, Cochrane, CBM, CNKI, Wanfang Data, and VIP.
All the databases were searched up to January 2016. The key
search terms (using the PubMed query as an example) were
Bmild cognitive impairment^[Mesh] OR Bcognitive impairment,
mild^[tiab] OR Bcognitive impairments, mild^[tiab] OR
Bimpairment, mild cognitive^[tiab] OR Bimpairments, mild
cognitive^[tiab] OR Bmild cognitive impairments^[tiab] OR
Bmental disorder^[Mesh] OR Bmental disorders^[tiab] OR
Bneurocognitive disorder^[tiab] OR Bneurocognitive
disorders^[tiab] OR Bcognitive disorder^[tiab] OR Bcognitive
disorders^[tiab] OR Bmci^[tiab], Bdyssomnias^[Mesh] OR
Bsleep wake disorders^[Mesh] OR Bsleep wake disorder^[tiab]
OR Bsleep deprivation^[Mesh] OR Bsleep disorder^[tiab] OR
Neurol Sci
Bsleep disorders^[tiab] OR Bsleep disturbance^[tiab] OR
Bcircadian rhythm^[tiab] OR Bcircadian rhythms^[tiab],
Bpolysomnography^[Mesh] OR Bpolysomnographies^[tiab]
OR Bmonitoring, sleep^[tiab] OR Bsleep monitoring^[tiab] OR
Bsomnography^[tiab] OR Bsomnographies^[Mesh],
Bactigraphy^[Mesh] OR Bactigraph^[tiab] OR Bactigraphic
recording^[tiab] OR Bwrist actigraph^[tiab] OR Bactigraph
recording^[tiab] OR Bwrist activity^[tiab] OR Brest
activity^[tiab] OR Bactivity^[tiab] OR Bsleep-wake
activity^[tiab] OR Baccelerometry^[Mesh] OR Bsensor^[tiab].
We also reviewed the references of the included studies to ex-
plore other potential candidates for inclusion. The languages
were restricted to English and Chinese.
Review selection and data extraction
In accordance with the formulated inclusion and exclusion
criteria, the titles were first reviewed for obvious exclusions.
After the abstract was read, if it was unclear whether the article
should be excluded, the full text was reviewed. Two reviewers
(HMY and LC) selected articles and extracted data indepen-
dently and finally compared their lists of articles to ensure that
the same studies had been included or excluded. The two
reviewers resolved the discrepancies by discussion. If the dis-
crepancies could not be solved, the final decision was made by
a third person (ZP).
Quality assessment
Two reviewers (HMY and LC) evaluated the included studies
independently. The tool used for evaluating the quality of the
studies was the Newcastle-Ottawa scale (NOS) [14]. The NOS
includes the three categories of selection (four items, four
stars), comparability (one item, two stars), and exposure (three
items, three stars) for a total of nine stars. A study can be
awarded a maximum of one star for each numbered item with-
in the selection and exposure categories. A maximum of two
stars can be given for comparability.
Statistical analysis
RevMan 5.3 was used in our meta-analysis. The standard
mean difference (SMD) with 95% CI was applied. To begin,
the fixed-effect model was used if the result of the heteroge-
neity test showed p > 0.05 and I2 < 50%. If not, the random-
effect model and subgroup analysis were used to analyze the
sources of heterogeneity. A p value <0.05 was considered
significant. Sensitivity analysis was used to analyze the stabil-
ity of the results, while a funnel plot was used to judge pub-
lication bias.
Neurol Sci

Fig. 1 PRISMA flowchart of

searching and selection guidelines

Publication bias Statistical analysis

The funnel plots were applied using RevMan 5.3. The SMD TST
as the x-axis and the SE (SMD) as the y-axis were used to
examine publication bias. We regarded heterogeneity as lower Eight [1012, 1618, 20] included studies showed that there
if the figure had bilateral symmetry. was no significant difference between the overall MCI and HE
groups (SMD = 0.19, 95% CI (0.75~0.47), p = 0.50)
(Fig. 2).
Results
Subgroup analysis of TST MCI subtypes: Seven [1012,
Search results 1618, 20] studies showed that the HE groups had a greater
mean than the MCI groups (SMD = 0.53, 95%CI (0.77~
The literature search method was conducted as shown in 0.28), p < 0.0001). Two [15, 20] studies showed no
Fig. 1: after the full text of the 61 potentially eligible articles
had been read, 10 studies fulfilled the eligibility criteria and Table 1 The NOS quality assessment of the ten included studies
were included [1012, 1521] (Fig. 1).
Studies Selection Comparability Exposure Total

Quality assessment of studies (NOS) Yu et al. [10] 8

Jiang et al. [11] 7
The Newcastle-Ottawa scale (NOS) was used in our study. All Hita-Yaez et al. [12] 7
the articles scored more than six stars, so all the articles were Westerberg et al. [15] 8
considered high-quality [14] (Table 1). Lou et al. [16] 8
Kim et al. [17] 8
Baseline characteristics of the included studies Maestri et al. [18] 7
Zhang et al. [19] 6
Information on all the ten articles, such as publication time, Naismith et al. [20] 9
countries, major criteria of MCI, and sample size, is listed in Hayes et al. [21] 8
Table 2.
Table 2 Main characteristics of the ten studies

Studies Year Participants country Major criteria for MCI Subtypes of Sex Age SD (years) Objective
of origin MCI measures
Cases (M / Controls (M / Cases Controls
F) F)

Yu et al. [10] 2009 Taiwan CDR/MMSE/blood test/thyroid test/GDS/CT MCI 9/4 7/5 75.1 10.7 76.4 8.3 PSG
Jiang et al. [11] 2013 Italy MRI/HIS/MoCA/Hamilton rating scale for MCI 38 / 18 33 / 15 63.0 2.5 62.0 3.6 PSG
depression/Hamilton anxiety rating scale
Hita-Yaez et al. 2013 Spain CDR/MMSE/DSM-IV/GDS/immediate and MCI 18 / 7 12 / 13 70.5 6.8 67.1 5.3 PSG
[12] delayed recall
Westerberg et al. 2010 US MMSE/trail making test parts A and B/Boston aMCI 2/8 3/7 71.1 0.0 72.5 0.0 Actigraphy
[15] naming test/word list learning and
recognition/constructions/category
fluency/CERAD/Wechsler memory scale
Lou et al. [16] 2008 China MMSE/ADL/HIS/CT/MRI MCI 6 / 11 9/9 70.0 6.0 69.0 6.0 PSG
Kim et al. [17] 2011 Korean Petersens criteria/CERAD-K/verbal MCI 9 / 21 9 / 21 68.0 4.09 67.4 3.8 PSG
fluency/Boston naming test/word list
memory/constructional praxis/word list
recall/word list recognition/constructional
recall/GDS/blood test/DSM-IV
Maestri et al. [18] 2015 Italy CDR/MMSE/MRI/blood test/immediate and MCI 6/7 4/5 68.5 7.0 69.2 12.6 PSG
delayed recall for Rey auditory verbal learning
test/attentive matrices/trail making test parts A
and B/digit span forward score
Zhang et al. [19] 2006 China MMSE/CDR/ADL/HIS/immediate and delayed MCI 21 / 28 74.0 4.8 71.6 5.3 PSG
memory test/digit symbol test
Naismith et al. [20] 2014 Australia Petersens criteria/attention, working memory, MCI 17 / 9 12 / 14 70.1 9.9 65.9 9.8 PSG Actigraphy
processing speed, verbal and visual learning and (sleep
memory, language, visuospatial, executive parameter
function/DSM-IV/MRI data from
PSG)
Hayes et al. [21] 2014 US Petersens criteria/GDS/MMSE/CDR/memory test/ aMCI naMCI 2 / 14 3 / 26 84.8 6.6 87.5 4.0 Actigraphy
language, attention, executive function test

M male, F female, MRI magnetic resonance imaging, CDR clinical dementia rating, MMSE minimum mental state examination, CERAD Consortium to Establish a Registry for Alzheimers Disease, HIS
Hachinski ischemic score, MoCA Montreal cognitive assessment, DSM diagnostic and statistical manual of mental disorders, GDS global deterioration scale, ADL activities of daily living, CT computed
tomography, MCI mild cognitive impairment, aMCI amnestic mild cognitive impairment, naMCI non-amnestic mild cognitive impairment, PSG polysomnograph.
Neurol Sci
Neurol Sci
Fig. 2 Meta-analysis of TST between the overall MCI and HE groups
significant difference between their amnestic mild cognitive
impairment (aMCI) and HE groups (SMD = 1.60, 95%CI
(1.25~4.44), p = 0.27). One [21] study showed that the
non-amnestic mild cognitive impairment (naMCI) group had
a greater mean total sleep time (TST) than the HE group
(SMD = 2.20, 95%CI (1.31~3.08), p < 0.00001) (Fig. 3).
Comparison between naMCI and aMCI: Only one [21]
study provided both naMCI and aMCI data. No significant
difference was observed between the naMCI and aMCI
groups [SMD = 0.61, 95%CI (1.65~0.43), p = 0.25]
(Fig. 4).
Different measurements: Seven [1012, 1618, 20] studies
provided detailed PSG data, while one [15, 21] provided
actigraphy data. In the PSG subgroup, longer TST was ob-
served in the HE groups (SMD = 0.53, 95%CI (0.77~
0.28), p < 0.0001). In the actigraphy subgroup, no significant
difference was observed between the overall MCI and HE
groups (SMD = 1.30, 95%CI (0.89~3.48), p = 0.25) (Fig. 5).
Fig. 3 Subgroup analysis of MCI subtypes on TST
SE
Six [10, 11, 1517, 19] studies showed that HE had a better
sleep efficiency (SE) (SMD = 0.68, 95%CI (0.92~0.45),
p < 0.00001) (figures not displayed).
Subgroup analysis of SE MCI subtypes: Five [10, 11, 16, 17,
19] studies provided MCI group data, while one [15] study
provided aMCI group data. In the MCI subgroup, greater SE
in HE was observed (SMD = 0.73, 95%CI (0.98~0.48),
p < 0.00001). In the aMCI subgroup, no significant difference
was observed between the aMCI and HE groups
(SMD = 0.11, 95%CI (0.99~0.77), p = 0.81).
Different measurements: Five [10, 11, 16, 17, 19] studies
provided detailed PSG data, while one [15] provided
actigraphy data. In the PSG subgroup, greater SE was ob-
served in HE (SMD = 0.73, 95%CI (0.98~0.48),
p < 0.0001). In the actigraphy subgroup, no significant

Fig. 4 Subgroup analysis of naMCI and aMCI on TST
difference was observed between the overall MCI and HE
groups (SMD = 0.11, 95%CI (0.99~0.77), p = 0.81).
SL
Six [1012, 15, 16, 19] studies showed that overall MCI pa-
tients have a longer sleep latency (SL) (SMD = 0.78, 95%CI
(0.53~1.02), p < 0.00001).
Subgroup analysis of SL MCI subtypes: Five [1012, 16, 19]
studies provided detailed MCI data, while one [15] provided
aMCI data. In the MCI subgroup, longer SL in MCI was
observed (SMD = 0.78, 95%CI (0.53~1.04), p < 0.0001). In
the aMCI subgroup, no significant difference was observed
between the aMCI and HE groups (SMD = 0.78, 95%CI
(0.53~1.02), p = 0.17).
REML
Five [11, 16, 1820] included studies showed that there was
no significant difference between the overall MCI and HE
groups (SMD = 0.27, 95%CI (0.24~0.77), p = 0.31).
Fig. 5 Subgroup analysis of different measurements in relation to TST
Neurol Sci
REM%
Six [10, 11, 1720] included studies showed that there was no
significant difference between the overall MCI and HE groups
(SMD = 0.70, 95%CI (1.59~0.19), p = 0.12).
AI
Six [10, 11, 1720] included studies showed that there was no
significant difference between the overall MCI and HE groups
(SMD = 0.02, 95%CI (1.46~1.51), p = 0.98).
Subgroup analysis of AI MCI subtypes: Six [1012, 16, 18,
19] studies showed that no significant difference was found
between the MCI groups and HE groups (SMD = 0.77,
95%CI (0.38~1.92), p = 1.09). One [21] study showed that
the aMCI group expressed a lower arousal index (AI) than the
HE group (SMD = 8.20, 95%CI (10.42~5.98),
p < 0.00001) and that the naMCI group expressed a lower
AI than the HE group (SMD = 3.54, 95%CI (4.63~
2.44), p < 0.00001).
Comparison between naMCI and aMCI: Only one [21]
study showed that the naMCI group had a greater AI than
Neurol Sci
the aMCI group (SMD = 2.80, 95%CI (4.31~1.30),
p = 0.0003).
Different measurements: Six [1012, 16, 18, 19] studies
provided detailed PSG data, while one [21] provided
actigraphy data. In both the PSG and actigraphy subgroups,
no significant difference was observed between the MCI and
HE groups (SMD = 0.77, 95%CI (0.38~1.92), p = 1.09;
SMD = 3.54, 95%CI (4.63~2.44), p < 0.00001).
WASO
Four [15, 17, 18, 20] included studies showed that there was
no significant difference between the overall MCI and HE
groups (SMD = 0.32, 95%CI (0.18~0.83), p = 0.21).
Subgroup analysis of WASO MCI subtypes: Three [17, 18,
20] studies provided MCI data, while one [15] provided aMCI
data. In both the MCI and aMCI subgroups, no significant
difference was observed between the MCI and HE groups
(SMD = 0.32, 95%CI (0.34~0.98), p = 0.35) or between
t h e aM C I an d H E g r ou p s ( S M D = 0. 28 , 95 % C I
(0.60~1.17), p = 0.53).
Different measurements: Three [17, 18, 20] studies provid-
ed PSG data, while one [15] provided actigraphy data. In both
the PSG and actigraphy subgroups, no significant difference
was observed between the MCI and HE groups (SMD = 0.32,
95%CI (0.34~0.98), p = 0.35; SMD = 0.28, 95%CI
(0.60~1.17), p = 0.53).
SWS
Three [11, 17, 20] included studies showed that there was no
significant difference between the overall MCI and HE groups
(SMD = 0.18, 95%CI (0.66~0.30), p = 0.45).
PLMS
Three [10, 11, 17] included studies showed that there was no
significant difference between the overall MCI and HE groups
(SMD = 0.72, 95%CI (0.38~1.82), p = 0.20).
CAP
One [18] included study showed that the MCI group
expressed less cyclic alternating pattern (CAP) than the HE
group (SMD = 1.23, 95%CI (2.15~0.30), p = 0.009).
Publication bias
We used RevMan 5.3 to make the funnel plots (figures are not
displayed). The figures appeared to have bilateral symmetry.
We can conclude that no publication bias was evident in our
Table 3 Publication bias of TST
Sleep parameters Eggers test Beggs test
TST t = 0.69 p > |t| = 0.510 z = 0.31 pr > z = 0.754
final result. We also used Stata 12.0 to retest the publication
bias of TST. The p value was >0.05 (Table 3).
Discussion
To the best of our knowledge, this is the first meta-analysis of
sleep disturbance in MCI as assessed by objective measures.
On the objective measures, we found that, compared with HE,
the overall MCI patients as a group expressed more SL and
less SE; MCI patients showed less TST and SE and more SL
and CAP; patients with aMCI had less AI; patients with
naMCI had more TST and less AI. Patients with naMCI
expressed more AI than those with aMCI.
Sleep disturbance and MCI
Several parameters of sleep architecture were identified in our
study. When compared with HE, MCI patients expressed
more SL, a greater decline in the length of TST. The naMCI
patients expressed more TST. Sleep loss could lead to short
sleep, tiredness, fatigue, and, ultimately, cognitive decline.
The identical findings from the Whitehall II study also showed
that some changes in sleep time (decrease from 8, 7, or 6 h to
less sleep time) are associated with cognitive dysfunction [22].
A previous study also examined the association of sleep du-
ration with cognitive function in a cohort of 1844 community-
dwelling women; women sleeping 5 h/night scored worse
than women sleeping 7 h/night in cognitive performance
[23]. In contrast to a prior study that, after controlling for
multiple potential confounders, reported a U-shaped associa-
tion between sleep time and cognitive decline, our study was
able to confirm only the left portion of the U. The reason
might be that our participants were all MCI patients and HE,
not patients with severe cognitive impairment or dementia,
meaning that almost all TST values were in the range of
5~8 h, not lower than 4 h or higher than 10 h. It also confirmed
that MCI is a transitional stage from normal to dementia [24].
Shorter sleep duration has been connected with greater -
amyloid (A) deposition, suggesting that sleep disturbance
leads to MCI and is not merely a consequence of brain lesions
but that A deposition in the areas involved in the control of
sleep could contribute to the worsening of sleep, causing a
further increase and accumulation of A [25]. We also found
that naMCI patients experienced more TST than HE; the rea-
son might be that naMCI patients have enough TST and

memory consolidation occurs during sleep [7]. Therefore, we
speculate that the memory deficits of MCI patients may be
caused by insufficient TST or increased SL.
SE of lower than 85% is usually used as one of the diag-
nostic criteria for insomnia [10]. Our study found that both the
overall MCI and unspecified MCI patients experienced lower
SE than HE. A related study investigated the association be-
tween sleep and cognition among 2932 community-dwelling
elderly persons, finding out that lower SE was consistently
related to MCI [26]. Another study examined 60 persons aged
55 years or older; the correlational analysis revealed that day-
time performance is related to SE and that a good night of
sleep is associated with better cognitive performance in good
sleepers. The reason for this phenomenon also might be A
deposition. A 2-year longitudinal study showed that cognitive
impairment is associated with SE and that, compared with
those without A deposition, the A deposition with cogni-
tive impairment group had worse SE (80.4 vs 83.7%) [27].
Compared with HE, MCI patients also had less CAP.
The result showed a CAP rate reduction that correlated
with the cognitive decline. Consistently, a previous
study showed a decreased amplitude and robustness of
the sleep-wake cycle, correlated with an increased risk
of cognitive decline [28]. The aMCI and naMCI patients
had a lower AI, and the effect was more severe in
naMCI than that in aMCI patients. A previous study
reported that the frequency of nocturnal AI was associ-
ated with accelerated cognitive decline [7]. These differ-
ences in sleep disruption between aMCI and naMCI
may be related to differences in the pathology underly-
ing these MCI subtypes [21].
In our study, not all the sleep parameters showed changes
in MCI or the MCI subtypes. Rapid eye movement latency
(REML), percent of rapid eye movement (REM%), wake after
sleep onset (WASO), slow-wave sleep (SWS), and periodic
leg movement in sleep (PLMS) showed no changes. This is
inconsistent with some previous studies. As confirmed by
prior study, rapid eye movement sleep behavior disorder
(RBD) was a preclinical marker for a kind of neurodegenera-
tive diseases, and it may impair cognition [29]. In NREM
stages N1, N2, and N3, research focuses on the effects of
SWS on cognition; results suggest that SWS is related to better
cognitive function [30]. WASO is the total amount of time
awake excluding the SL, determined by the PSG or
actigraphy. One study examined 39 patients by PSG, having
hypothesized that mental health would be affected by high
WASO [31]. A study reported that, after a motor learning task
performed during the day, the increase of SWS during the
night was related to improved performance [32]. In one epi-
demiological analysis on a cohort of 263 patients, MCI pa-
tients had a high frequency of sleep disturbances, and one of
the frequent symptoms was PLMS [24]. The reason might be
the limited sample size.
Neurol Sci
Objective measurement of sleep disturbance
Previous studies have shown that questionnaire tools (e.g., the
consensus sleep diary, Pittsburgh sleep quality index) and self-
reporting are applied at the point of care and rely on patients
and caregivers ability to accurately recall things [33]. In con-
trast, early research has suggested that actigraphy may be a
useful, cost-effective method for assessing specific sleep dis-
turbance [27]. PSG may be used for diagnosis and for evalu-
ation of various sleep parameters, and researchers have also
recommended routine overnight PSG to measure the sleep
state for elderly people [34]. Both measures were sensitive
in detecting sleep [9]. In our studies, subgroup analysis of
different measures showed that actigraphy is less accurate than
polysomnography but not does affect the results. Hence, the
best practice is to include both subjective and objective mea-
sures when examining sleep quality in elderly people [8].
Limitation
Our systematic review and meta-analysis is limited in that we
searched only the English- and Chinese-language literature.
Thus, some studies may have been missed, which resulted in
only ten studies being included. The small sample size includ-
ed is another main limitation, but we believe that with the
development of modern technology, there will be more and
more researchers using objective measures to explore the cor-
relation between sleep disturbance and MCI. A third limita-
tion is that the different criteria of MCI may affect the included
MCI groups. Fourth, there are only two studies referring to the
MCI subtypes of aMCI and naMCI; further study should em-
phasize the difference of sleep disturbance between aMCI and
naMCI. Finally, the existing sleep parameters were not com-
prehensive; further sleep parameters need to be explored.
Acknowledgments The authors thank the Jilin University, Dr.
Changgui Kou, Dr. Xiuying Zhang et al. for their guidance in the devel-
opment of this article.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflicts of
interest.
Funding The authors received no financial support for the article.
References
1. Prince M, Wimo A, Guerchet M, Ali GC, Wu YT, Prina M (2015)
The Global Impact of Dementia: an analysis of prevalence, inci-
dence, cost and trends. World Alzheimer Report
2. Baozhen D (2012) Knowledge, attitude, belief, and practice
(KABP/KAP) about Alzheimer disease among suspectable
Neurol Sci
individuals with mild cognitive impairment, Dissertation.
University of Wuhan, Wuhan.
3. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG,
Kokmen E (1999) Mild cognitive impairment: clinical characteri-
zation and outcome. Arch Neurol 56(3):303308
4. Lara E, Koyanagi A, Olaya B, Lobo A, Miret M, Tyrovolas S, Haro
JM (2016) Mild cognitive impairment in a Spanish representative
sample: prevalence and associated factors. International journal of
geriatric psychiatry
5. Yuan Y, Niu F, Liu Y, Lu N (2014) Zinc and its effects on oxidative
stress in Alzheimers disease. Neurol Sci 35(6):923928
6. ubuku HC, Yurtda M, Durak ZE, Ayta B, Gne HN, okal 23.
BG, Yolda TK, Durak (2016) Oxidative and nitrosative stress in
serum of patients with Parkinsons disease. Neurol Sci 37(11):
17931798
7. da Silva RAPC (2015) Sleep disturbances and mild cognitive im-
pairment: a review. Sleep science 8(1):3641
8. Landry GJ, Best JR, Liu-Ambrose T (2015) Measuring sleep qual-
ity in older adults: a comparison using subjective and objective
methods. Frontiers in aging neuroscience 7
9. Blood ML, Sack RL, Percy DC, Pen JC (1997) A comparison of
sleep detection by wrist actigraphy, behavioral response, and
polysomnography. Sleep 20(6):388395
10. Yu JM, Tseng IJ, Yuan RY, Sheu JJ, Liu HC, Hu CJ (2009) Low
sleep efficiency in patients with cognitive impairment. Acta Neurol
Taiwanica 18(2):9197
11. Jiang B, Ding C, Yao G, Yao C, Zhang Y, Ge J, Ferri R (2013)
Polysomnographic abnormalities in patients with vascular cognitive
impairment-no dementia. Sleep Med 14(11):10711075
12. Hita-Yaez E, Atienza M, Cantero JL (2013) Polysomnographic
and subjective sleep markers of mild cognitive impairment. Sleep
36(9):13271334
13. Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred
reporting items for systematic reviews and meta-analyses: the
PRISMA statement. Ann Intern Med 151(4):264269
14. Stang A (2010) Critical evaluation of the Newcastle-Ottawa scale
for the assessment of the quality of nonrandomized studies in meta-
analyses. Eur J Epidemiol 25(9):603605
15. Westerberg CE, Lundgren EM, Florczak SM, Mesulam MM,
Weintraub S, Zee PC, Paller KA (2010) Sleep influences the sever-
ity of memory disruption in amnestic mild cognitive impairment:
results from sleep self-assessment and continuous activity monitor-
ing. Alzheimer Dis Assoc Disord 24(4):325
16. Lou MZ, Lu YZ, Zhang ZF et al (2008) Evaluation on whole night
polysomnography in aged patients with mild cognitive impairment.
Chinese journal of new drugs and clinical remedies 27:109
17. Kim SJ, Lee JH, Lee DY, Jhoo JH, Woo JI (2011) Neurocognitive
dysfunction associated with sleep quality and sleep apnea in pa-
tients with mild cognitive impairment. Am J Geriatr Psychiatry
19(4):374381
18. Maestri M, Carnicelli L, Tognoni G, Di Coscio E, Giorgi
FS, Volpi L, Bonanni E (2015) Non-rapid eye movement
sleep instability in mild cognitive impairment: a pilot study.
Sleep Med 16(9):11391145
19. Zhang BH, Zhu ZB, He JG et al (2006) Sleep disorders in elderly
patients with mild cognitive impairment and the effect of melatonin
treatment. Chinese Journal of Neuromedicine 8:806808
20. Naismith SL, Hickie IB, Terpening Z, Rajaratnam SW, Hodges JR,
Bolitho S, Lewis SJ (2014) Circadian misalignment and sleep dis-
ruption in mild cognitive impairment. J Alzheimers Dis 38(4):857
866
21. Hayes TL, Riley T, Mattek N, Pavel M, Kaye JA (2014) Sleep
habits in mild cognitive impairment. Alzheimer Dis Assoc Disord
28(2):145
22. Ferrie JE, Shipley MJ, Akbaraly TN, Marmot MG, Kivimaki
M, Singh-Manoux A (2011) Change in sleep duration and
cognitive function: findings from the Whitehall II Study.
Sleep 34(5):565573
Tworoger SS, Lee S, Schernhammer ES, Grodstein F (2006) The
association of self-reported sleep duration, difficulty sleeping, and
snoring with cognitive function in older women. Alzheimer Dis
Assoc Disord 20(1):4148
24. Pistacchi M, Gioulis M, Contin F, Sanson F, Marsala SZ (2014)
Sleep disturbance and cognitive disorder: epidemiological analysis
in a cohort of 263 patients. Neurol Sci 35(12):19551962
25. Ju YES, McLeland JS, Toedebusch CD, Xiong C, Fagan AM,
Duntley SP, Holtzman DM (2013) Sleep quality and preclinical
Alzheimer disease. JAMA neurology 70(5):587593
26. Blackwell T, Yaffe K, Ancoli-Israel S, Schneider JL, Cauley JA,
Hillier TA, Stone KL (2006) Poor sleep is associated with impaired
cognitive function in older women: the study of osteoporotic frac-
tures. J Gerontol Ser A Biol Sci Med Sci 61(4):405410
27. Sadeh A, Hauri PJ, Kripke DF, Lavie P (1995) The role of
actigraphy in the evaluation of sleep disorders. Sleep 18(4):288
302
28. Tranah GJ, Blackwell T, Stone KL, Ancoli-Israel S, Paudel ML,
Ensrud KE, Cauley JA, Redline S, Hillier TA, Cummings SR, Yaffe
K (2011) Circadian activity rhythms and risk of incident dementia
and mild cognitive impairment in older women. Ann Neurol 70(5):
722732
29. Zhang X, Sun X, Wang J, Tang L, Xie A (2016) Prevalence
of rapid eye movement sleep behavior disorder (RBD) in
Parkinsons disease: a meta and meta-regression analysis.
Neurological Sciences 18
30. Aric D, Drago V, Foster PS, Heilman KM, Williamson J, Ferri R
(2010) Effects of NREM sleep instability on cognitive processing.
Sleep Med 11(8):791798
31. Moore P, Bardwell WA, Ancoli-Israel S, Dimsdale JE (2001)
Association between polysomnographic sleep measures and
health-related quality of life in obstructive sleep apnea. J Sleep
Res 10(4):303308
32. Huber R, Ghilardi MF, Massimini M, Tononi G (2004) Local sleep
and learning. Nature 430(6995):7881
33. Lyons BE, Austin D, Seelye A, Petersen J, Yeargers J, Riley T,
Kaye JA (2015) Corrigendum: Pervasive computing technologies
to continuously assess Alzheimers disease progression and inter-
vention efficacy. Frontiers in aging neuroscience 7
34. Kushida CA, Littner MR, Morgenthaler T, Alessi CA, Bailey D,
Coleman JR, Lee-Chiong T (2005) Practice parameters for the in-
dications for polysomnography and related procedures: an update
for 2005. Sleep 28(4):499521