Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
patients
Yan Zhang a, b, *, 1, Wanhua Wang a b, 1, Sijie Cai c, Qi Sheng c, Shenggui Pan d, Fang Shen a, b, Qing
Tang a, b, Yang Liu e
a
Department of Neurology, The Kunshan Affiliated Hospital of Jiangsu University, 215300 Kunshan, China b Department of
Neurology, The First People's Hospital of Kunshan, 215300 Kunshan, China c
Department of Respiratory Medicine, The Kunshan Affiliated Hospital of Jiangsu University, 215300 Kunshan, China d
Department of Rehabilitation Medicine, The Kunshan Affiliated Hospital of Jiangsu University, 215300 Kunshan, China e
Department of Neurology, Saarland University, 66421 Homburg, Germany
a r t i c l e i n f o
Article history: Received 19 October 2016 Received in revised form 20 November 2016 Accepted 28 November 2016 Available
online 4 February 2017
Keywords: Cognition Obstructive sleep apnea Prospective memory Sleep and stroke
a b s t r a c t
Background: Obstructive sleep apnea (OSA) is very common in stroke survivors. It potentially worsens the cognitive dysfunction
and inhibits their functional recovery. However, whether OSA independently damages the cognitive function in stroke patients is
unclear. A simple method for evaluating OSA- induced cognitive impairment is also missing. Methods: Forty-four stroke patients
six weeks after onset and 24 non-stroke patients with snoring were recruited for the polysomnographic study of OSA and sleep
architecture. Their cognitive status was evaluated with a validated Chinese version of Cambridge Prospective Memory Test. The
relationship between memory deficits and respiratory, sleeping, and dementia-related clinical variables were analyzed with
correlation and multiple linear regression tests. Results: OSA significantly and independently damaged time- and event-based
prospective memory in stroke patients, although it had less power than the stroke itself. The impairment of prospective memory
was correlated with increased apneaehypopnea index, decreased minimal and mean levels of peripheral oxygen saturation, and
disrupted sleeping continuity (reduced sleep efficiency and increased micro- arousal index). The further regression analysis
identified minimal levels of peripheral oxygen saturation and sleep efficiency to be the two most important predictors for the
decreased time-based prospective memory in stroke patients. Conclusions: OSA independently contributes to the cognitive
dysfunction in stroke patients, potentially through OSA-caused hypoxemia and sleeping discontinuity. The prospective memory
test is a simple but sensitive method to detect OSA-induced cognitive impairment in stroke patients. Proper therapies of OSA
might improve the cognitive function and increase the life quality of stroke patients.
© 2017 Elsevier B.V. All rights reserved.
1. Introduction
Obstructive sleep apnea (OSA) is characterized by repeated collapses of the upper airway during sleep, and is common in the
general population. Its mean prevalence is 22% and 17% in men and
women, respectively [1]. In stroke patients, the prevalence can rise to 72% [2]. OSA is an independent risk factor of stroke [3,4].
It also contributes to the occurrence of stroke through exaggerating hy- pertension, diabetes mellitus, and cardiovascular disease
[5e7]. Moreover, OSA has been shown to inhibit the recovery of stroke patients. Stroke patients with OSA showed lower
functional ca- pacity and spent longer time in the rehabilitation than non-OSA
DOI of original article: http://dx.doi.org/10.1016/j.sleep.2016.12.014. * Corresponding author. Department of Neurology, The
Kunshan Hospital Affiliated to Jiangsu University, Qianjinxi Road, No. 91, 215300 Kunshan, Jiangsu
stroke patients [8,9]. Fortunately, OSA is treatable with contin- uous positive airway pressure (CPAP) and maxillomandibular
advancement [10e12]. Thus, proper OSA therapy not only prevents
Province, China. Fax: þ86 512 57501112.
stroke, but also optimizes the stroke rehabilitation. Since
there are E-mail address: 2573560757@qq.com (Y. Zhang). 1 These two authors contributed equally to this study.
still more than 10 million major strokes every year in the world
http://dx.doi.org/10.1016/j.sleep.2016.11.028 1389-9457/© 2017 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Sleep Medicine
journal homepage: www.elsevier.com/locate/sleep
Sleep Medicine 33 (2017) 183e190
[13], a timely diagnosis of OSA in stroke patients and the following recovery training has become extremely important.
Cognitive dysfunction, shown as a variety of deficits in attention, executive function, memory, language and visuoperceptual
ability, is a common consequence of stroke [14,15]. Prospective studies have shown that impaired visuospatial construction and
memory, or inattention and perceptual disorders within one month after stroke, predicts a poor functional outcome after six
months [16,17]. Recently, OSA has been shown to worsen the attention, executive functioning, visuoperception, psychomotor
ability, and intelligence of stroke patients, which causes difficulty in the rehabilitation of stroke patients [18]. In the follow-up
study, CPAP therapy was able to significantly improve the cognitive function of stroke patients, although the effect of CPAP on
functional recovery is still under investigation [19]. Thus, OSA potentially damages the cognitive function, which subsequently
mediates the negative effects of OSA on stroke recovery. The improvement of cognitive function can be used to evaluate the
efficacy of anti-OSA therapy in stroke patients. However, studies investigating the relationship between OSA and cognitive
impairment in stroke patients are rare. In few pub- lished studies, there was not a control group of non-stroke patients with OSA,
which can be used to investigate the independent effects of OSA and stroke on cognitive dysfunction [18,19]. The underlying
pathophysiological mechanisms through which OSA impairs the cognitive function of stroke patients are unclear either. To
address these questions, our study recruited both stroke and non-stroke patients and carefully examined both respiratory
parameters and the structure of sleep. In order to find a method with which the clinicians (especially in local hospitals and
without extensive training for psychological analysis), can also successfully examine the cognitive status of stroke patients, we
used the prospective memory test to evaluate cognitive function; the patient was simply requested to perform a planned action at
a certain future time or after a certain event occurred [20].
2. Materials and methods
2.1. Patients
In this study, all 65 stroke patients (ischemic, hemorrhagic and subarachnoid hemorrhage) admitted to the Department of
Neurology, Kunshan affiliated Hospital of Jiangsu University, China, from June 2013 to June 2015 were invited, and 44 patients
were enrolled into this study according to the following criteria: (1) stroke confirmed by a neurologist, (2) age between 30 and 65
years old, (3) admission at six weeks after stroke, (4) able to participate in the sleep study and neuropsychological assessment,
and (5) sufficiently fluent in lingual communication. The exclusion criteria was as follows: (1) severe, unstable medical
conditions, respiratory failure, or history of severe congestive heart failure, (2) traumatic brain injury, (3) severe aphasia,
confusion, or severe psychiatric comorbidity, (4) central sleep apnea, or (5) any previ- ously diagnosed sleep diseases including
narcolepsy, periodic limb movement disorder and Parkinson's disease-related sleep disor- ders. In order to investigate the relative
contribution of OSA to cognitive dysfunction of stroke patients, we also recruited 24 non- stroke patients, who were over 18
years old and came to our outpatient service due to snoring (they had no previous diagnosis of neurological, psychiatric and/or
respiratory disease). All sub- jects received overnight polysomnography. Before poly- somnography, the subjects were evaluated
for sleeping and cognitive function by using Epworth sleepiness scale (ESS), Mini- mental state examination (MMSE), and
prospective memory test. The study was approved by the Ethics Committee of Jiangsu
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 184
University and written informed consent was obtained from each participant.
2.2. Polysomnographic study
All stroke patients underwent one night (!8 h) of polysomno- graphic study at the sleep laboratory of our department six
weeks after stroke onset. Non-stroke patients participated in the poly- somnographic (PSG) test without drinking alcoholic or
caffeinated drinks in the last two days. The overnight sleep polysomnograph (Alice 5 Diagnostic Sleep System; Philips
Healthcare, Andover, USA) included electroencephalography (C3/A2, C4/A1, O1/A2, and O2/ A1), electrooculography,
submental electromyography, bilateral anterior tibialis electromyography, electrocardiography, nasal airflow measurement
(sensed by both thermistor and pressure transducer), monitoring of thoracoabdominal movements, oxygen saturation, snoring and
body position.
Sleep stages and respiratory events were analyzed against the Sleep Medicine Criteria (American Academy, 2007) [21].
Apnea was defined as a reduction of airflow of !90% for at least 10 s and hypopnea was defined as a reduction of airflow of !50%
for at least 10 s followed by an oxygen desaturation of !3%. Apneas with thoracic motion, without thoracic motion, and with
initial lack of motion followed by respiratory effort, were classified as obstruc- tive, central, and mixed apneas, respectively.
OSA was diagnosed when at least 50% of the respiratory events were of the obstructive type. The apneaehypopnea index (AHI)
was defined as the mean number of apneas and hypopneas per hour in bed. Patients with AHI <5 were referred to as non-OSA
patients and enrolled in the control group of this study. Other OSA-associated respiratory pa- rameters, including minimum of
peripheral oxygen saturation (SpO2), mean of SpO2 and percentage of time with SpO2 at <90% were recorded. Sleep
architecture variables were also examined, such as sleep efficiency, sleep stages, and microarousal index. Sleep efficiency is
percentage of sleeping duration divided by the total time spent in the bed. Microarousal index was the number per hour of
microarousals which last 3e15 s.
2.3. Epworth sleepiness scale (ESS)
The ESS is a self-administered questionnaire that is widely used for subjective assessment of daytime sleepiness. It contains
eight items involving eight daily-life scenarios, with each item being assessed on a zero-to-three point scale. The total scores of
ESS range from 0 to 24. The cutoff point for excessive daytime sleepi- ness is set at >10 [22].
2.4. Mini-mental state examination
The Chinese version of the MMSE questionnaire consists of several subscales: orientation, immediate and short-term memory
recall, attention/calculation, language, and visuospatial skills. Cognitive deficit on the MMSE is defined as a score less than 27
[23].
2.5. Prospective memory test
The prospective memory test was assessed with a validated Chinese version of Cambridge Prospective Memory Test (CAM-
PROMPT) [24]. The CAMPROMPT comprises six prospective mem- ory tasks which are either cued by events (event-based
prospective memory tasks; n 1⁄4 3) or by time (time-based prospective memory tasks; n 1⁄4 3). In this test, participants were
asked to remember carrying out three time- and three event-based prospective mem- ory tasks while working on a number of
other activities (pencil and
with paper tasks, such as a general knowledge quiz or word-finding
zero and one, respectively), and the
dementia-related factors puzzle) for a 20-minute period. The CAMPROMPT generates
[age, gender, hypertension, diabetes, smoking,
drinking, education, scores on all six tasks, each with a maximum score of six; higher
body mass index (BMI), cholesterol and
homocysteine]. To deter- scores indicate better prospective memory performance.
mine the relative contributing power of respiratory and
sleeping All participants were asked to repeat the instructions before
disorders in the cognitive dysfunction of stroke
patients, all respi- beginning the CAMPROMPT to ensure that they understood what
ratory and sleeping parameters (together with the
above described they had to do during the test. The explanation of assay instructions
dementia-related factors) were analyzed as
independent variables was repeated to participants who appeared uncertain about the
to predict the dependent variable: time- or event-based
prospec- test. Upon completion of the CAMPROMPT, participants were again
tive memory, with a multiple linear regression analysis
with the asked to repeat the requirements of the test to make sure they
method of “stepwise.” In this analysis, non-stroke
patients were not remembered and understood the instructions. Two well-trained
included. psychiatrists had assessed the prospective
memory and other cognitive functions.
3. Results
2.6. Statistical analysis
3.1. Characteristics of the participants
Data analyses were performed using SPSS (version 19.0, IBM,
To investigate the effects of OSA on cognitive
function of stroke New York, USA). We used descriptive statistics to characterize the
patients, we screened 65 stroke patients and found 44
patients sample in terms of demographic, clinical, sleep, and stroke vari-
eligible. Twenty-one patients were discharged due to
their severe ables. To compare values between two independent groups, we
or exaggerated symptoms, which made the
polysomnographic and used parametric (Student t test) and nonparametric tests (c2
test)
cognitive studies impossible. The polysomnographic
and cognitive where appropriate. To test the differences among three groups,
examinations for all recruited patients were performed
six weeks one-way analysis of variance (ANOVA) and c2
test were used. The
after stroke onset. Out of 44 stroke patients, nine
(20.45%) received Pearson correlation analysis was performed to assess the rela-
a diagnosis of mild OSA (15 > AHI ! 5) and 18
patients (40.91%) tionship between time- or event-based prospective memory and
were diagnosed with moderate-to-severe OSA (AHI !
15). Seven- respiratory parameters (eg, AHI, minimum of SpO2, mean of SpO2
teen patients (38.64%) without OSA (AHI < 5) were
considered as and the percentage of time with SpO2 at <90%) and sleep param-
non-OSA control stroke patients. All patients except
one patient eters (eg, sleep efficiency, sleep time at stages 1, 2, 3 and rapid eye
with moderate to severe OSA were diagnosed with
stroke for the movement, and microarousal index). To investigate whether OSA
first time (Table 1). The functional status of patients
after stroke was an independent risk factor of cognitive dysfunction in stroke
was evaluated with NIHSS, which showed that the
functional re- patients, stroke and non-stroke patients were pooled into one
covery was not different between patients with
different severity of single study group. A multiple linear regression analysis with the
OSA (Table 1; one-way ANOVA, p > 0.05). Other
clinical data of method of “enter” was used to predict time- or event-based pro-
participating stroke patients was shown in Table 1.
The age (range: spective memory with potential predictors: OSA (without OSA,
35e73 years; 57.3 ± 9.6 years), sex distribution (male:
70.5%), time with mild and moderate-to-severe OSA were coded with zero, one
of education (range: 6e15 years; 8.3 ± 2.8 years),
obesity shown and two, respectively), stroke (with and without stroke were coded
with BMI (25.3 ± 3.3), rates of diabetes (18.2%), smoking (34.1%) and
Table 1 Characteristics of stroke patients.
Characteristics Moderate to severe OSA (n 1⁄4 18) Mild OSA (n 1⁄4 9) Non-OSA (n 1⁄4 17) p Value
Age (year) 59.4 ± 10.2 58.2 ± 9.1 54.6 ± 9.1 0.336a Sex (males) 13 (72.2%) 7 (77.8%) 11 (64.7%) 0.768
b
Educations (year) 8.5 ± 3.3 8.0 ± 3.4 8.1 ± 2.1 0.888
a
BMI 26.3 ± 3.0 24.8 ± 2.4 24.6 ± 3.9 0.295a Hypertension 17 (94.4%) 6 (66.7%) 10 (58.8%) 0.042
b
Diabetes 6 (33.3%) 0 2 (11.8%) 0.072
b
Smoking 3 (16.7%) 5 (55.6%) 7 (41.2%) 0.097
b
Drinking 3 (16.7%) 2 (22.2%) 1 (5.9%) 0.456b Cholesterol 4.6 ± 0.8 4.6 ± 1.2 4.7 ± 1.1 0.966
a
Homocysteine 17.0 ± 11.7 18.3 ± 13.4 14.8 ± 3.9 0.672
a
Stroke type 0.527b
Ischemia 11 (61.1%) 7 (77.8%) 13 (76.5%) Hemorrhage 7 (38.9%) 2 (22.2%) 4 (23.5%) Subarachnoid hemorrhage 0 0 0 Stroke
location 0.779
b
Cortex 2 (11.1%) 2 (22.2%) 2 (11.8%) Cerebellum 2 (11.1%) 0 1 (5.9%) Brain stem 14 (77.8%) 7 (77.8%) 14 (82.4%) Recurrent
stroke 1 (5.6%) 0 (0%) 0 (0%) 0.478
b
Minimum of SpO2 79.1 ± 5.0 89.4 ± 3.64 92.0 ± 2.2 <0.001
a
Mean of SpO2 93.8 ± 3.2 97.1 ± 1.17 96.4 ± 1.3 0.001
a
Percentage of time with SpO2 at <90% 6.7 ± 13.2 0.2 ± 0.3 0.0 ± 0.0 0.049
a
Apneaehypopnea index 38.2 ± 17.4 8.1 ± 2.3 2.7 ± 1.3 <0.001
a
NIHSS 3.5 ± 3.5 2.7 ± 2.6 3.4 ± 2.7 0.780
a
Values are presented as mean (standard deviation) or n (%). BMI, body mass index; OSA, obstructive sleep apnea; SpO2,
peripheral oxygen saturation; NIHSS, National Institutes of Health Stroke Scale.
a
One-way ANOVA. b c2 test.
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 185
with drinking (13.6%), serum concentrations of cholesterol
various severity of OSA did not differ in the
scores (Fig. 1A; (4.6 ± 1.0 mmol/L) and homocysteine (16.4 ± 9.8 mmol/L), as well as
one-way ANOVA, p > 0.05). stroke types (ischemia,
hemorrhage, or subarachnoid hemorrhage)
In non-stroke patients, we have observed that both
time- and and locations (cortex, cerebellum, or brain stem) were not signifi-
event-based prospective memory significantly
declined in OSA cantly different between different OSA groups (one-way ANOVA or
patients compared with non-OSA controls (Fig. 1B; t
test, p < 0.001). c2
test, p > 0.05). However, the rate of hypertension in the group of
The MMSE did not show difference between these
two groups of patients (94.4%) suffering from moderate-to-severe OSA was
non-stroke patients (Fig. 1B). significantly higher than
that in the groups of patients with mild
We continued to ask whether OSA independently
contributed to (66.7%) or no OSA (58.8%) (c2
test, p < 0.05). AHI and other respi-
the cognitive dysfunction of stroke patients and how
strong the ratory parameters, such as minimum of SpO2, mean of SpO2 and
contribution was. We pooled stroke and non-stroke
patients, and percentage of time with SpO2 at <90% were significantly changed
performed a multiple linear regression (with the
method of in patients with moderate-severe OSA compared to other patients
“enter”) to examine whether OSA and stroke
independently pre- with mild or no OSA (one-way ANOVA, p < 0.05).
dicted time- or event-based prospective memory. Age,
gender, In 24 non-stroke snoring patients, 17 (70.83%) received a diag-
hypertension, diabetes, smoking, drinking, education,
BMI, nosis of OSA (AHI ! 5), among which 16 patients (66.67%) were
cholesterol, and homocysteine were also included as
independent diagnosed of moderate-to-severe OSA (AHI ! 15). We have
variables in the regression analysis as they potentially
affect the observed that BMI and cholesterol, which were associated with
cognitive function. These 12 independent variables
were able to obesity, were significantly higher in OSA group than in the non-OSA
reliably predict the time and event-based prospective
memory (F- controls (p < 0.05; see Table 2). Interestingly, the serum concen-
tests for time-based prospective memory: F (12, 55)
1⁄4 22.139, tration of homocysteine was also strongly increased by OSA (from
p < 0.0001, R
2
1⁄4 0.828; and for event-based
prospective memory: F 10.5 ± 2.8 to 16.0 ± 4.9, t test, p 1⁄4 0.011). It was similar to stroke
(12, 55) 1⁄4 7.045, p < 0.0001, R patients that snoring
patients with OSA showed significantly higher AHI and reduced saturation of oxygen in the peripheral circulation (p < 0.05; see
Table 2).
3.2. OSA is independently involved in the cognitive dysfunction of stroke patients
To evaluate the effects of OSA on cognitive functions of stroke patients, we performed MMSE and time- or event-based
prospec- tive memory assays. Three groups of stroke patients with different severity of OSA had shown significant difference in
the prospective memory (Fig. 1; one-way ANOVA, p < 0.05). In the further post-hoc Bonferroni tests, we observed that
time-based prospective memory was already impaired in patients with mild OSA as compared to non-OSA patients (Fig. 1A;
one-way ANOVA, p < 0.001); whereas, the event-based prospective memory in patients with mild OSA was not different from
that in non-OSA stroke patients. When the two groups with mild OSA and non-OSA stroke patients were combined and
compared with the group of stroke patients with moderate-to-severe OSA, we clearly observed that event-based prospective
memory was also impaired in moderate-to-severe OSA patients (Fig. 1A; t test, p 1⁄4 0.003). In MMSE, all patients
2
1⁄4 0.606). Indeed, OSA and stroke were two independent predictors for the loss of both time-
and event-based prospective memory, although stroke had relatively more power than OSA to impair the prospective memory
(Table 3; The coefficients of stroke and OSA were À8.16 and À2.68 for the prediction of time-based prospective memory, and
À5.51 and À1.38 for event-based prospective memory, respectively; p < 0.01). For the event-based prospective memory, the
serum concentration of cholesterol was another significant predictor (p < 0.05), although it had very weak power (with the
coefficient of À0.68) compared to stroke and OSA.
3.3. OSA-associated hypoxemia is correlated with the impaired cognitive functions in stroke patients
In order to understand how OSA impaired the prospective memory, we made correlation tests between the time- or event-
based prospective memory and AHI and other OSA-associated respiratory parameters. Indeed, time-based prospective memory
was significantly correlated with AHI, minimal or mean level of SpO2, and percentage of time, during which SpO2 was less than
90% (Table 4; Pearson correlation tests, p < 0.05). The event-based prospective memory was significantly correlated with AHI
and minimal level of SpO2, but not with mean level of SpO2, and
Table 2 Characteristics of patients with snoring.
Characteristics OSA (n 1⁄4 17) Non-OSA (n 1⁄4 7) p Value
Age (year) 43.9 ± 10.8 39.7 ± 12.9 0.423a Sex (males) 15 (88.2%) 5 (71.4%) 0.315
b
Educations (year) 12.2 ± 3.9 12.0 ± 4.1 0.896
a
BMI 28.5 ± 4.0 24.4 ± 2.3 0.019
a
Hypertension 11 (64.7%) 3 (42.9%) 0.324
b
Diabetes 5 (29.4) 3 (42.9%) 0.525
b
Smoking 8 (47.1%) 1 (14.3%) 0.132
b
Drinking 2 (11.8) 0 0.343
b
Cholesterol 5.0 ± 1.1 4.3 ± 0.3 0.019
a
Homocysteine 16.0 ± 4.9 10.5 ± 2.8 0.011
a
Minimum of SpO2 71.4 ± 15.6 92.7 ± 1.8 0.000
a
Mean of SpO2 93.0 ± 6.6 96.4 ± 1.3 0.008
a
Percentage of time with SpO2 at <90% 17.4 ± 22.8 0 0.006
a
Apneaehypopnea index 50.0 ± 23.8 1.5 ± 1.1 <0.001
a
Values are presented as mean (standard deviation) or n (%). BMI, body mass index; OSA, obstructive sleep apnea; SpO2,
peripheral oxygen saturation.
a
Student t test. b c2 test.
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 186
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 187
Fig. 1. Obstructive sleep apnea impairs prospective memory. Forty-four stroke and 24 non-stroke patients with snoring were
recruited and examined with poly- somnography for the diagnosis of OSA and with Cambridge Prospective Memory Test and
MMSE for their cognitive status. A, Cognitive scores of MMSE and time-based prospective memory (PM) between stroke
patients with different severity of OSA were compared with one-way ANOVA followed by post-hoc Bonferroni tests; The score
of event-based PM in moderate-to-severe OSA group was compared to the score of combined groups of stroke patients with
non-OSA and mild OSA with t test (n 1⁄4 17, 9 and 18 for non-OSA, mild and moderate-to-severe OSA, respectively). B,
Cognitive scores between snoring patients with and without OSA were compared with t tests (n 1⁄4 7 and 17 for non-OSA and
OSA, respectively). **: p < 0.01 and ***: p < 0.001.
Table 3 Linear regression analysis to predict time or event-based prospective memory.
Independent variables TPM EPM
Coefficients p Value Coefficients p Value
(Constant) 17.87 <0.001 15.77 0.001 Stroke À8.16 <0.001 À5.51 <0.001 OSA À2.68 <0.001 À1.38 0.002 age À0.03 0.515 0.02
0.641 Gender À0.06 0.946 0.15 0.874 Hypertension 1.18 0.082 0.89 0.244 Diabetes À1.12 0.117 À0.23 0.773 Cholesterol À0.03
0.925 À0.68 0.047 Homocysteine 0.02 0.510 0.06 0.170 Body mass index 0.08 0.389 0.06 0.588 Smoking À0.28 0.714 À0.14
0.871 Drinking À1.04 0.302 À0.32 0.782 Education 0.06 0.656 0.18 0.237
TPM, time-based prospective memory; EPM, event-based prospective memory; OSA, obstructive sleep apnea.
percentage of time with SpO2 at <90% (Table 4; Pearson correlation tests, p < 0.05). When coefficients in different correlation
tests were compared, we observed that the minimal level of SpO2 (which referred to the hypoxemia in patients), was the most
prominently correlated with the impaired prospective memory (Table 4). Of note, in non-stroke patients, both time- and
even-based prospec- tive memory was not correlated with AHI and reduction of pe- ripheral oxygen saturation (data not shown).
3.4. OSA-associated impairment of sleep efficiency is correlated with damaged cognitive functions in stroke patients
OSA causes sleep fragmentation, which potentially damages cognitive function [25]. We continued to examine effects of
OSA on the architecture of sleep. The sleep efficiency of stroke patients without OSA was 84.6 ± 4.8%, which significantly
decreased in stroke patients with mild (75.4 ± 12.8%) and moderate-to-severe OSA (78.9 ± 8.8%) (Table 5; one-way ANOVA, p
< 0.05). The microarousal index differed between these three groups of stroke patients with different severity of OSA (Table 5;
non-OSA: 10.1 ± 4.3; mild OSA: 11.9 ± 6.8; and moderate-to-severe OSA: 34.7 ± 17.8; one-way ANOVA, p < 0.05). The
percentages of time during stage 1, 2, or 3, or during the rapid eye movement (REM) stage among total sleeping time, were not
different between these three groups of stroke patients (Table 5; one-way ANOVA, p > 0.05); however, when all stroke patients
with mild to severe OSA were combined and compared to patients without OSA, we have observed that the time at stage 3 [also
named slow wave sleep (SWS)] was significantly shortened by OSA (14.6 ± 3.5% in non-OSA patients and 11.0 ± 5.7% in
patients with OSA; t test, p 1⁄4 0.025). Similarly, OSA significantly shortened the sleeping time at stage 3 and increased the
microarousal index in patients suffering from snoring (Table 6; t test, p < 0.05). The examination with a self- administered
questionnaire, ESS, had not shown the effects of OSA on the daytime sleepiness in both stroke and snoring patients (Tables 5 and
6; p > 0.05).
When we made correlation tests between the prospective memory and variable sleep parameters, we observed time- but not
event-based prospective memory was significantly correlated with sleep efficiency and microarousal index in stroke patients with
similar coefficient values (Table 7; p < 0.05). In non-stroke patients, neither time- nor even-based prospective memory was
correlated with all parameters related to the sleep structure (data not shown).
3.5. Hypoxemia and sleep disruption independently contribute to the cognitive dysfunction of stroke patients
After we observed that both hypoxemia and sleep efficiency were correlated with the prospective memory, we further
determined their relative contribution to the OSA-induced cognitive dysfunction in stroke patients with a multiple linear
regression with stepwise method. Among the 21 different respiratory parameters (eg, AHI, minimal or mean level of SpO2, and
percentage of time with SpO2 at <90%), sleep parameters (eg, sleep efficiency, sleeping time at stage 1, 2, 3 and REM, and
microarousal index), and dementia-related clinical factors (age, gender, hypertension, diabetes, smoking, drinking, education,
BMI, cholesterol, homocysteine, and NIHSS), three independent variables: minimal level of SpO2 (Coefficients: 0.286, p <
0.001), sleep efficiency (Coefficients: 0.127, p 1⁄4 0.004) and age (Coefficients: À0.134, p 1⁄4 0.002) were significantly entered
the statistical model to predict the time-based memory [F (3, 40) 1⁄4 20.049, p < 0.001, R
2
1⁄4 0.601]. Similarly, in the prediction of event-based prospective memory, AHI (Coefficients:
À0.079, p < 0.004), cholesterol (Coefficients: À1.259, p 1⁄4 0.002) and
Table 4 Correlation tests between prospective memory and variable OSA parameters in stroke patients.
AHI Minimum of SpO2 Mean of SpO2 Percentage of time with SpO2 at <90%
TPM Pearson correlation À0.507 0.608 0.364 À0.302
p Value <0.001 <0.001 0.015 0.046 EPM Pearson correlation À0.343 0.356 0.162 À0.206
p Value 0.022 0.018 0.294 0.179
AHI, apneaehypopnea index; SpO2, peripheral oxygen saturation; TPM, time-based prospective memory; EPM, event-based
prospective memory.
Table 5 Sleep architecture of stroke patients.
Moderate-to-severe OSA (n 1⁄4 18) mild OSA (n 1⁄4 9) non-OSA (n 1⁄4 17) p Value
a
Sleep efficiency 78.9 ± 8.8 75.4 ± 12.8 84.6 ± 4.8 0.028 Stage 1 (%) 12.6 ± 7.4 11.0 ± 4.2 10.1 ± 4.4 0.433 Stage 2 (%) 58.7 ±
5.6 59.4 ± 6.9 55.6 ± 4.8 0.168 Stage 3 (%) 11.2 ± 6.4 10.4 ± 4.4 14.6 ± 3.5 0.076 REM (%) 17.5 ± 7.2 19.3 ± 5.8 19.8 ± 4.7
0.509 Microarousal index 34.7 ± 17.8 11.9 ± 6.8 10.1 ± 4.3 <0.001 ESS 11.6 ± 2.3 11.3 ± 2.5 11.5 ± 1.7 0.947
Values are presented as mean (standard deviation). OSA, obstructive sleep apnea; REM, rapid eye movement; ESS, Epworth
sleepiness scale.
a
One-way ANOVA.
Table 6 Sleep architecture of snoring patients.
Sleep parameter OSA (n 1⁄4 17) Non-OSA (n 1⁄4 7) p Value
declarative and retrospective memory, and followed supervisory executive functions [26]. Impairment of prospective memory oc-
a
curs frequently after stroke with a similar rate of retrospective
Sleep efficiency Stage 1 (%) Stage 2 (%) 90.8 ± 5.2 88.3 ± 12.5 11.1 ± 5.5 6.6 ± 3.0 62.3 ± 6.6 57.2 ± 8.3 0.493 0.055 0.121
memory, but more often than impairments in attention and exec- utive functioning [27]. Prospective memory tests have been used
to identify memory deficits in Hong Kong Chinese stroke patients
Stage 3 (%) 4.3 ± 4.7 13.2 ± 5.2 <0.001
with 95.5% of sensitivity and 55.9% of specificity [28].
We examined REM (%) 22.2 ± 4.5 23.0 ± 4.2 0.714
prospective memory to evaluate the effect of OSA on
cognitive Microarousal index 43.8 ± 22.0 7.4 ± 3.7 ESS 12.9 ± 3.1 12.7 ± 1.4 <0.001
0.893
function. We did observe that prospective memory, especially time- based prospective memory, is impaired in stroke patients in
an OSA Values are presented as mean (standard deviation). OSA, obstructive sleep apnea; REM, rapid eye movement; ESS,
Epworth sleepiness scale. a
Student t test.
severity-dependent manner; event-based prospective memory appears to be less affected by OSA. After MMSE failed to detect
cognitive dysfunction in stroke patients with OSA in our and other's studies [29], the prospective memory test proved to be a
preferable testing method, as it is simpler than the domain-specific neuro- education (Coefficients: 0.483, p 1⁄4 0.001) were
significantly entered
psychological tests and will be more acceptable to
clinicians. the statistical model [F (3, 40) 1⁄4 9.953, p < 0.001, and R
2
1⁄4 0.427].
Stroke as a major brain disorder often damages cognitive func- tion [14,15]. In published studies, OSA has been associated
with
4. Discussion
severer cognitive dysfunction in stroke patients and the pattern of OSA-related cognitive impairment different from the typical
OSA is very common in stroke survivors. In our study, 61.36% stroke patients suffered from OSA. OSA was reported to
exaggerate cognitive dysfunction of stroke patients, which potentially impaired their rehabilitation [18,19]. Our study has
clarified that OSA is an independent risk factor in additional to stroke impairing the cognitive function (eg, prospective memory)
of stroke patients. As a potential underlying pathogenic mechanism, OSA decreases the oxygen supply to the brain and reduces
sleep efficiency.
Prospective memory is a form of memory that involves remembering to perform a planned action at a certain future time point
or after a certain event occurs [20]. It requires episodic,
stroke-caused cognitive deficits [18,19]; however, it remains diffi- cult to conclude whether OSA independently contributes to
cognitive impairment in stroke patients. In our study, we recruited non-stroke patients in additional to stroke patients. After
stroke was used as a potential predictor of prospective memory deficits in a multiple linear regression analysis, we observed that
both OSA and stroke are two independent risk factors with stroke having more power than OSA. In non-stroke patients, we also
observed that OSA impairs prospective memory, which is in accordance with the previous observations that OSA negatively
affects cognitive functioning in the general population [30]. It should be also noted
Table 7 Correlation tests between prospective memory and variable sleep parameters in stroke patients.
Sleep efficiency Stage 1 Stage 2 Stage 3 REM Microarousal index
TPM Pearson correlation 0.382 À0.178 À0.194 0.292 0.101 À0.394
p Value 0.011 0.246 0.208 0.055 0.513 0.008 EPM Pearson correlation 0.140 À0.041 0.089 À0.070 0.015 À0.260
p Value 0.364 0.790 0.564 0.653 0.925 0.088
REM, rapid eye movement; TPM, time-based prospective memory; EPM, event-based prospective memory.
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 188
that the toxic effect of OSA on prospective memory in stroke patients is more pronounced than in non-stroke patients, which
might be due to the less capacity of the brain of stroke patients than that of snoring patients to compensate OSA-induced neuronal
degeneration.
We also investigated the potential pathophysiological mechanisms mediating OSA-induced cognitive deficits in stroke
patients. Both blood oxygen levels (eg, minimal and mean level of SpO2) and air flow through the respiratory tracts (eg, AHI)
were correlated to the impaired prospective memory in our experiments. However, the values of coefficients in correlation tests
and the regression analysis both demonstrated that the minimal level of SpO2 was the most important risk factor in the loss of
time-based prospective memory. AHI was detected as an independent predic- tor of the decline of event-based prospective
memory. However, the coefficient of AHI was tiny compared to the coefficients for education and the cholesterol concentration
in blood. A number of previous studies have associated OSA-related cognitive dysfunction with hypoxia [30e32]. Our study may
further suggest that the extremely low level of oxygen in the blood (even lasting for short time) is more important to induce the
neurodegeneration in the brain than the hypoxemia, which might last for longer time but with a relative higher level of oxygen in
the blood.
OSA causes fragmentation of sleep due to repeated arousals [25]. In a study of patients with insomnia, sleep discontinuity was
shown to damage the attention and episodic memory [33]. In OSA patients, the number of microarousals predicts deficits of
episodic memory, especially the recollection of spatial and temporal contexts [34]. In our study, we have observed that the
decrease of prospective memory is correlated with sleep efficiency and microarousal index, although only sleep efficiency
entered the regression analysis to predict deficit of time-based prospective memory.
It should be noted that in both stroke and snoring patients, OSA tended to shorten sleeping time at the slow wave stage. SWS
is essential in the consolidation of declarative memory in young and healthy people [35e39]. As a potential mechanism, the active
behavior first generates a labile information store in the hippo- campus; during SWS, this temporally encoded hippocampal infor-
mation is reactivated and transmitted to the cerebral cortex and integrated into a more permanent memory [40]. However, the
episodic memory-promoting effect of SWS is lost during aging [39,41]. In epileptic patients with accelerated long-term
forgetting, SWS adversely affects the declarative memory [42]. In stroke pa- tients, SWS activity increases over the infarct area
and decreases at the peri-infarct area [43]. Moreover, arteriolosclerosis and subcor- tical infarcts in older adults are associated
with sleep fragmentation [44]. Thus, the effects of SWS on the formation of declarative memory in stroke patients remain to be
determined.
Most importantly, OSA is a treatable disorder. CPAP therapy has been shown to improve episodic memory, attention, and
executive- functioning in non-stroke OSA patients [45e47]. Recently, the similar therapeutic effect was also observed in stroke
patients suffering of OSA [19]. CPAP and other anti-OSA therapies also improve sleep efficiency and reduce sleep fragmentation
[46,48].
There are several limitations of our study. First, the sample size of our study is small. Second, there might be a selection bias
in the sample of non-stroke snoring patients. It is expected that only these patients with severe OSA-related symptoms, like
daytime sleepi- ness and significant sleep disturbance, visited our hospital. It might explain why 70.83% snoring patients suffered
from OSA and most of them were diagnosed with severe OSA; the percentage of OSA patients is higher than that of stroke
patients. Third, stroke patients were significantly older than non-stroke control patients in our study. During aging, the amount of
SWS decreases and the frag- mentation of SWS increases, which results in decreased slow-wave
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 189
activity [49]. Indeed, the sleep-dependent consolidation of declar- ative memory has been observed to decline with aging,
especially in untreated OSA patients [39,41,50]. Thus, the contributing power of stroke was potentially overestimated in the
regression analysis to predict prospective memory loss. Fourth, we did not use a comprehensive neuropsychological assessment
battery to evaluate cognitive function of stroke patients. We are not able to compare the prospective memory deficits with other
OSA-induced domain- specific cognitive dysfunction. Thus, we are unclear whether the prospective memory assay is the best
method to detect cognitive impairment in stroke patients with OSA.
5. Conclusion
Our study demonstrates that OSA significantly and indepen- dently contributes to cognitive dysfunction in Chinese stroke pa-
tients, which is associated with hypoxemia and disrupted continuity of sleep. In stroke patients with OSA, proper anti-OSA
therapies might improve the cognitive function, facilitate the functional recovery and increase the quality of life of stroke
patients.
Funding sources
This work was supported by the program “Clinical and basic research of brain” (KYC004) from the Kunshan affiliated
hospital of Jiangsu University, and the Suzhou development plan for science and technology in 2013 (SYSD2013022, to YZ).
Conflict of interest
The authors declare that they have no conflict of interest in regard to this work.
The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking
on the following link: http://dx.doi.org/10.1016/j.sleep.2016.11.028.
References
[1] Franklin KA, Lindberg E. Obstructive sleep apnea is a common disorder in the population-a review on the epidemiology of
sleep apnea. J Thorac Dis 2015;7: 1311e22. [2] Johnson KG, Johnson DC. Frequency of sleep apnea in stroke and TIA patients:
a meta-analysis. J Clin Sleep Med 2010;6:131e7. [3] Valham F, Mooe T, Rabben T, et al. Increased risk of stroke in patients
with coronary artery disease and sleep apnea: a 10-year follow-up. Circulation 2008;118:955e60. [4] Yaggi HK, Concato J,
Kernan WN, et al. Obstructive sleep apnea as a risk factor
for stroke and death. N Engl J Med 2005;353:2034e41. [5] Lee CH, Sethi R, Li R, et al. Obstructive sleep apnea and
cardiovascular events after percutaneous coronary intervention. Circulation 2016;133:2008e17. [6] Lindberg E, Theorell-Haglow
J, Svensson M, et al. Sleep apnea and glucose metabolism: a long-term follow-up in a community-based sample. Chest
2012;142:935e42. [7] Peppard PE, Young T, Palta M, et al. Prospective study of the association be- tween sleep-disordered
breathing and hypertension. N Engl J Med 2000;342: 1378e84. [8] Kaneko Y, Hajek VE, Zivanovic V, et al. Relationship of
sleep apnea to func- tional capacity and length of hospitalization following stroke. Sleep 2003;26: 293e7. [9] Yan-fang S,
Yu-ping W. Sleep-disordered breathing: impact on functional
outcome of ischemic stroke patients. Sleep Med 2009;10:717e9. [10] Bratton DJ, Gaisl T, Wons AM, et al. CPAP vs
mandibular advancement devices and blood pressure in patients with obstructive sleep apnea: a systematic review and
meta-analysis. JAMA 2015;314:2280e93. [11] Salord N, Fortuna AM, Monasterio C, et al. A randomized controlled trial of
continuous positive airway pressure on glucose tolerance in obese patients with obstructive sleep apnea. Sleep 2016;39:35e41.
[12] Zaghi S, Holty JC, Certal V, et al. Maxillomandibular advancement for treat- ment of obstructive sleep apnea: a
meta-analysis. JAMA Otolaryngol Head Neck Surg 2015:1e9. [13] Rothwell PM. AVERT: a major milestone in stroke research.
Lancet 2015;386:
7e9.
[14] Leys D, Henon H, Mackowiak-Cordoliani MA, et al. Poststroke dementia.
Lancet Neurol 2005;4:752e9. [15] Levine DA, Galecki AT, Langa KM, et al. Trajectory of cognitive decline after
incident stroke. JAMA 2015;314:41e51. [16] Park J, Lee G, Lee SU, et al. The impact of acute phase domain-specific
cognitive function on post-stroke functional recovery. Ann Rehabil Med 2016;40:214e22. [17] Nys GM, van Zandvoort MJ, de
Kort PL, et al. The prognostic value of domain- specific cognitive abilities in acute first-ever stroke. Neurology 2005;64: 821e7.
[18] Aaronson JA, van Bennekom CA, Hofman WF, et al. Obstructive sleep apnea is related to impaired cognitive and functional
status after stroke. Sleep 2015;38:1431e7. [19] Aaronson JA, Hofman WF, van Bennekom CA, et al. Effects of continuous
positive airway pressure on cognitive and functional outcome of stroke pa- tients with obstructive sleep apnea: a randomized
controlled trial. J Clin Sleep Med 2016;12:533e41. [20] McDaniel MA, Einstein GO. The neuropsychology of prospective
memory in normal aging: a componential approach. Neuropsychologia 2011;49: 2147e55. [21] Berry RB, Budhiraja R, Gottlieb
DJ, et al. Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated
Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine. J Clin Sleep Med
2012;8: 597e619. [22] Johns MW. Sensitivity and specificity of the multiple sleep latency test (MSLT), the maintenance of
wakefulness test and the Epworth sleepiness scale: failure of the MSLT as a gold standard. J Sleep Res 2000;9:5e11. [23] Yu ES,
Liu WT, Levy P, et al. Cognitive impairment among elderly adults in
Shanghai, China. J Gerontol 1989;44:S97e106. [24] Wilson BA, Emslie H, Foley J, et al. The Cambridge prospective
memory test.
London: Harcourt Assessment; 2005. [25] Kimoff RJ. Sleep fragmentation in obstructive sleep apnea. Sleep 1996;19:
S61e6. [26] Martin T, McDaniel MA, Guynn MJ, et al. Brain regions and their dynamics in prospective memory retrieval: a
MEG study. Int J Psychophysiol 2007;64: 247e58. [27] Kant N, van den Berg E, van Zandvoort MJ, et al. Functional correlates
of
prospective memory in stroke. Neuropsychologia 2014;60:77e83. [28] Man DW, Chan MK, Yip CC. Validation of the
Cambridge prospective memory test (Hong Kong Chinese version) for people with stroke. Neuropsychol Rehabil
2015;25:895e912. [29] Sandberg O, Franklin KA, Bucht G, et al. Sleep apnea, delirium, depressed mood, cognition, and ADL
ability after stroke. J Am Geriatr Soc 2001;49: 391e7. [30] Yaffe K, Laffan AM, Harrison SL, et al. Sleep-disordered breathing,
hypoxia, and risk of mild cognitive impairment and dementia in older women. JAMA 2011;306:613e9. [31] Gelber RP, Redline
S, Ross GW, et al. Associations of brain lesions at autopsy
with polysomnography features before death. Neurology 2015;84:296e303.
Y. Zhang et al. / Sleep Medicine 33 (2017) 183e190 190
[32] Launer LJ, Hughes TM, White LR. Microinfarcts, brain atrophy, and cognitive function: the Honolulu Asia aging study
autopsy study. Ann Neurol 2011;70: 774e80. [33] Fortier-Brochu E, Morin CM. Cognitive impairment in individuals with
insomnia: clinical significance and correlates. Sleep 2014;37:1787e98. [34] Daurat A, Foret J, Bret-Dibat JL, et al. Spatial
and temporal memories are affected by sleep fragmentation in obstructive sleep apnea syndrome. J Clin Exp Neuropsychol
2008;30:91e101. [35] Deliens G, Leproult R, Neu D, et al. Rapid eye movement and non-rapid eye movement sleep contributions
in memory consolidation and resistance to retroactive interference for verbal material. Sleep 2013;36:1875e83. [36] Axmacher N,
Haupt S, Fernandez G, et al. The role of sleep in declarative memory consolidationedirect evidence by intracranial EEG. Cereb
Cortex 2008;18:500e7. [37] Cedernaes J, Rangtell FH, Axelsson EK, et al. Short sleep makes declarative
memories vulnerable to stress in humans. Sleep 2015;38:1861e8. [38] Griessenberger H, Hoedlmoser K, Heib DP, et al.
Consolidation of temporal
order in episodic memories. Biol Psychol 2012;91:150e5. [39] Scullin MK. Sleep, memory, and aging: the link between
slow-wave sleep and episodic memory changes from younger to older adults. Psychol Aging 2013;28:105e14. [40] Mitra A,
Snyder AZ, Hacker CD, et al. Human cortical-hippocampal dialogue in
wake and slow-wave sleep. Proc Natl Acad Sci U S A 2016;113:E6868e76. [41] Baran B, Mantua J, Spencer RM. Age-related
changes in the sleep-dependent
reorganization of declarative memories. J Cogn Neurosci 2016;28:792e802. [42] Atherton KE, Nobre AC, Lazar AS, et al.
Slow wave sleep and accelerated
forgetting. Cortex 2016;84:80e9. [43] Poryazova R, Huber R, Khatami R, et al. Topographic sleep EEG changes in the
acute and chronic stage of hemispheric stroke. J Sleep Res 2015;24:54e65. [44] Lim AS, Yu L, Schneider JA, et al. Sleep
fragmentation, cerebral arterio- losclerosis, and brain infarct pathology in community-dwelling older people. Stroke
2016;47:516e8. [45] Dalmases M, Sole-Padulles C, Torres M, et al. Effect of CPAP on cognition, brain function, and structure
among elderly patients with OSA: a randomized Pilot study. Chest 2015;148:1214e23. [46] Kushida CA, Nichols DA, Holmes
TH, et al. Effects of continuous positive airway pressure on neurocognitive function in obstructive sleep apnea pa- tients: the
Apnea Positive Pressure Long-term Efficacy Study (APPLES). Sleep 2012;35:1593e602. [47] Castronovo V, Scifo P, Castellano
A, et al. White matter integrity in obstructive
sleep apnea before and after treatment. Sleep 2014;37:1465e75. [48] Hetzenecker A, Escourrou P, Kuna ST, et al. Treatment
of sleep apnea in chronic heart failure patients with auto-servo ventilation improves sleep fragmentation: a randomized controlled
trial. Sleep Med 2016;17:25e31. [49] Varga AW, Ducca EL, Kishi A, et al. Effects of aging on slow-wave sleep dy- namics and
human spatial navigational memory consolidation. Neurobiol Aging 2016;42:142e9. [50] Djonlagic I, Guo M, Matteis P, et al.
Untreated sleep-disordered breathing: links to aging-related decline in sleep-dependent memory consolidation. PLoS One
2014;9:e85918.