EPIDEMIOLOGY
Effect of Ethnicity on Sleep: Complexities for Epidemiologic Research
Carl J. Stepnowsky, Jr., PhD; Polly J. Moore, PhD; Joel E. Dimsdale, MD
Department of Psychiatry, University of California, San Diego
Study Objectives: The goal of this study was to examine whether there
were ethnic differences in polysomnographically recorded sleep, either in
the controlled laboratory environment or in the home setting.
Design: Prospective study of ethnic differences in stress physiology and
sleep.
Setting: Two sleep recordings were performed on consecutive nights in a
hospital-based sleep laboratory, followed 1 to 4 weeks later by a third
sleep recording in the subject’s home.
Participants: 51 employed healthy adult subjects, aged 15 to 50 years.
24 self-identified as black, and 27 as white.
Interventions. None.
Measurements and Results: Blacks had less slow wave sleep than did
whites in both the sleep laboratory and in the home. Blacks had signifi-
cantly more slow wave sleep at home compared to the hospital setting,
while the reverse was true for whites. This location-by-ethnicity interaction
could not be accounted for by depression ratings or social class.
Conclusions: The home setting is generally considered to be more ecologically valid than the controlled hospital-based laboratory setting for the
monitoring of sleep. These data suggest that ethnicities may respond differentially to the sleeping environment. This observation may need to be
taken into account in future epidemiologic studies of sleep.
Key Words: Ethnicity, home sleep recording, polysomnography, sleep
architecture, sleep location, sleep quality
Citation: Stepnowsky, Jr. CJ, Moore PJ, Dimsdale JE. Effect of ethnicity
on sleep: complexities for epidemiologic research. SLEEP 2003;26
(3):329-32.
INTRODUCTION
part of the aging process.
When measuring sleep by standard laboratory-based polysomnography, Profant and colleagues5 reported in 2 separate samples that blacks
had longer total sleep time, more minutes of rapid eye movement (REM)
sleep, and lower percentage of deep sleep, suggesting possible ethnic
differences in sleep architecture.
In terms of the sleep of other ethnicities, there is a surprising lack of
studies examining the possible differences in sleep architecture in Asians
and American Indians. In 1 small study examining the relationship
between ethnicity and sleep patterns in normal controls, Hispanics tended to have higher REM density than did other ethnic groups, and blacks
had more stage 1 and 2 sleep and less stage 4 sleep than did other ethnic
groups.6 The authors concluded that sleep patterns are quite consistent
across cultures. In a study of REM sleep and depression, depressed Hispanics and whites had more REM sleep than did depressed African
Americans and Asians.7 Most sleep and ethnicity studies focus on the
prevalence of sleep-related breathing disorder.
Oddly missing from this area of research is the recognition that there
may be site effects on sleep, and these effects may differ by ethnicity.
That is, people of various ethnicities may sleep differently in different
locations. For instance, there are ethnic variations in response to hospitalization itself. In response to admission to the hospital, whites drop
their blood pressure more than do blacks.8
Attention to these “site-of-testing” effects is important for sleep
research because of the dramatic growth of home sleep monitoring.
Home sleep monitoring correlates well with laboratory monitoring of
sleep,9 but it is generally accepted that the lab environment is different
from the home. For example, the first-night effect is known to occur in
the laboratory10,11 but is not typically found when sleep is studied in the
home.12,13 Home monitoring has the benefit of ecologic validity and
increased generalizability relative to studies performed in the laboratory.
Given the increased emphasis on acquiring meaningful ethnic data in
sleep, we wanted to explicitly examine the site effect as it relates to
objective measurement of sleep quality.
GIVEN THE GROWING INTEREST IN THE EFFECTS OF ETHNICITY ON HEALTH, it is surprising that there have not been more
studies on ethnic variation in sleep. Ethnicity may significantly impact
perceived sleep quality as well as objectively recorded polysomnographic sleep measures.
In terms of subjective self-report, the findings have been mixed. Karacan and colleagues1 surveyed a random sample of 1645 individuals living in Florida and found that blacks self-reported more sleep complaints
than did whites; however, this finding did not remain after controlling
for gender, age, socioeconomic status, and marital status. More recently,
Ancoli-Israel and colleagues2 found that elderly blacks reported less satisfaction with sleep, more difficulty falling asleep, more frequent morning headaches, and more inadvertent naps during the day (ie, more daytime sleepiness) than did whites. On the other hand, 2 studies found that
blacks had fewer sleep complaints than whites. Blazer and colleagues3
found that blacks reported a lower frequency of sleep complaints concerning trouble falling asleep, wakeful sleep, restless sleep, early morning awakening, and awakening not feeling refreshed. These findings
held even after controlling for age, education, presence of chronic health
conditions, self-rated health, and depression. In one of the largest epidemiologic studies to date (n=1118), Jean-Louis and colleagues4 found
that elderly whites self-reported significantly more sleep problems than
did elderly blacks. Five sleep questions were asked: “Do you depend on
medicine to sleep?” “Do you have difficulty falling asleep?” “Do you
wake up often during the night?” “Do you wake up early or wake feeling tired?” “Do you sleep during the day for more than two hours?”
These authors speculated that: 1) blacks may underreport health problems and 2) blacks may be more likely to accept sleep disturbance as
Disclosure Statement
This work was supported by NIH grants HL36005, HL44915, RR0827, and T32
MH18399.
Submitted for publication August 2002
Accepted for publication December 2002
Address correspondence to: Carl J. Stepnowsky, PhD, University of California,
San Diego, Department of Psychiatry (0804), 9500 Gilman Drive, La Jolla, CA
92093-0804; Tel: 619-543-5207; Fax: 619-543-7519;
E-mail: [email protected]
SLEEP, Vol. 26, No. 3, 2003
METHODS
Participants
Volunteers were located through word-of-mouth referral or through public advertisements. Sixty participants gave written informed consent.
329
Effect of Ethnicity on Sleep—Stepnowsky et al
Measurement
The University of California, San Diego, Institutional Review Board
approved the protocol.
Participants who completed the full protocol included 24 individuals
with self-reported ethnicity as black and 27 individuals as white. Therefore, the designations black and white will be used throughout this paper.
Inclusion criteria were ages between 25 and 50, body weight between
90% and 130% of ideal, blood pressure of less than 180/110 while off of
medications, current full-time employment (defined as 1 or more nonvolunteer jobs totaling 30 or more hours per week), and caffeine intake
less than 600 mg (about 6 cups) per day.
Exclusion criteria consisted of being a shift worker, taking medications other than antihypertensive agents, or the presence of any of the
following medical conditions: congestive heart failure, pulmonary disease requiring ongoing treatment, symptomatic coronary or cerebral vascular disease (eg, history of myocardial infarction, angina, stroke, or
transient ischemic attack), history of life-threatening arrhythmias, cardiomyopathy, history of psychosis, current drug or alcohol abuse, known
secondary hypertension, creatinine levels greater than 1.4mg/dL, kidney
disease, renal bruit on physical examination, prior diagnosis or treatment
of diabetes, fasting blood glucose greater than 120mg/dL, or known
sleep disorder. Women were additionally excluded if they had a diagnosis of premenstrual syndrome, were taking oral contraceptives, or were
pregnant at the time of the study.
Sleep in the Clinical Research Center and at home was recorded using
a polysomnograph (Embla, Flaga Medical, Reykjavik, Iceland) that
recorded central and occipital electroencephalogram derivations (C3,
C4, O1, O2), bilateral electrooculogram (left and right outer canthus),
submental and anterior tibialis electromyogram, electrocardiogram,
nasal/oral airflow using a thermistor and nasal cannula, respiratory effort
using chest and abdominal inductance belts, and finger pulse oximetry.
Participants underwent an adaptation sleep study on night 1 in the hospital. Sleep set-up typically began at 2000 and took about 1 hour.
Patients were instructed to go to sleep at their normal bedtime. “Lights
on” was scheduled for 0600. The second-night recording was obtained
in identical fashion and constituted our measure of “hospital sleep.”
Patients were set-up for polysomnography in their homes between 1900
and 2000. Patients were instructed to go to sleep and awaken on their
normal schedule.
Full polysomnography was performed identically in both the laboratory and the home. Sensors and electrodes were secured with combinations of tape, gauze with an adhesive backing (Hypafix), electrical conducting gels, and collodion. Signals were visualized on computer monitor (laptop in the home), and impedance values were checked. The electroencephalogram, electrooculogram, and electromyogram electrodes
were replaced if individual paired impedance values were greater than 5
kOhms. The sleep equipment was removed the following morning by a
research technician or the participant.
Apneas were defined as decrements in airflow of at least 90% from
baseline for a period of at least 10 seconds. Hypopneas were defined as
decrements in airflow from baseline between 50% and 90% for a period
of at least 10 seconds. The apnea-hypopnea index (AHI) was defined as
the number of apneas plus hypopneas per hour of sleep. The definition
of an arousal from sleep was based on the criteria published in the 1992
American Sleep Disorders Association report on electroencephalographic arousals.14 The total arousal index was defined as the number of
arousals per hour of sleep. Sleep staging was manually scored according
to the Rechtschaffen and Kales criteria15 on Embla’s sleep software,
Somnologica version 3.1. A minimum of 4 hours of scorable sleep was
necessary for the sleep data to be included in the analyses. All sleep scorers had interrater reliability indexes (κ) greater than 0.85 for staging,
arousal, and respiratory variables.
We considered some consolidated variables that might confound the
relationship between ethnicity and sleep. The possible confounding variables were age, AHI, depressive symptoms, social class, and blood pressure level.
Procedure
Patients receiving antihypertensive medication had their medication
tapered for 2 weeks prior to the sleep study. Participants were admitted
to University of California, San Diego Clinical Research Center for 2
nights of overnight polysomnography. They were then studied in their
homes 7 to 31 days later with an additional night of polysomnography.
Table 1—Participant Baseline Characteristics (mean ± SD)
Characteristic
Whites (n=27)
Blacks (n=24)
Sex (male/female)
19/8
12/12
Age (years)
37.7 ± 7.3 (26-47)
40.4 ± 7.4 (27–50)
Social Class
3.1 ± 1.4
3.3 ± 1.3
Blood pressure (at intake)
Systolic (mmHg)
128.2 ± 14.9
130.2 ± 18.7
Diastolic (mmHg)
76.6 ± 9.6
75.9 ± 11.1
2
26.3 ± 4.6
29.3 ± 6.3
Body mass index (kg/m )
Center for Epidemiological Studies
—Depression Scale
8.5 ± 7.2
13.0 ± 10.8
Apnea-hypopnea index
Hospital
12.1 ± 12.2 (0.30-58.4)
9.7 ± 7.9 (1.0-26.6)
Home
10.6 ± 6.1 (2.9-26.7)
11.8 ± 10.4 (0.20-40)
Arousal Index (events per hour)
Hospital
6.5 ± 4.0
7.7 ± 4.4
Home
6.9 ± 3.6
8.2 ± 5.0
Sleep efficiency (percentage of sleep time during sleep period)
Hospital
91.6% ± 7.1%
90.5% ± 7.6%
Home
91.0% ± 6.9%
87.5% ± 8.7%
Center for Epidemiological Studies—Depression Scale
The Center for Epidemiological Studies—Depression (CES-D) Scale
is a frequently used 20-item self-report scale that has been shown to be
reliable and valid for assessing depressive symptoms.16 In a variety of
populations, a cutoff score of 16 or greater has been shown to differentiate nondepressed subjects from those meeting diagnostic criteria for
dysthymia or major depression.17,18
Table 2—Sleep Characteristics (mean ± SD)
Characteristic
Total Sleep Time (minutes)
Stage 1
Percentage
Minutes
Stage 2
Percentage
Minutes
Slow Wave Sleep*
Percentage
Minutes
Rapid eye movement sleep
Percentage
Minutes
Whites
Hospital
Home
Blacks
Hospital
Home
389.7 ± 44.4
412.1 ± 57.0 371.8 ± 51.9
381.3 ± 84.0
5.7 ± 3.7
21.8 ± 13.6
7.2 ± 4.0
7.2 ± 2.8
28.5 ± 13.7 26.5 ± 9.5
8.8 ± 4.7
34.2 ± 17.7
62.1 ± 7.1
242.1 ± 38.4
62.9 ± 7.4 66.5 ± 6.9
258.4 ± 42.3 246.8 ± 40.3
64.6 ± 8.1
251.7 ± 67.6
12.4 ± 8.5
49.3 ± 35.1
9.0 ± 6.6
4.4 ± 4.2
37.8 ± 28.4 15.3 ± 16.6
7.2 ± 8.9
23.1 ± 28.1
19.8 ± 5.6
76.5 ± 87.5
20.9 ± 5.2 21.9 ± 6.4
87.5 ± 27.9 79.9 ± 26.5
19.4 ± 6.0
76.9 ± 31.5
Social Class
Social class was determined according to the method developed by
Hollingshead19 that takes into account education and occupation levels.
Scores range from 1 to 5, with lower scores indicating higher social
class.
Data were analyzed with SPSS v10.1 (Chicago, IL) using a repeated
measures ANCOVA with site as the repeated measure and ethnicity as a
factor and adjusting for covariates of age and AHI.
RESULTS
Participant characteristics are summarized in Table 1. Nine participants had incomplete data, resulting in a final sample size of 51. The eth-
* Repeated Measures ANCOVA significant at the p <. 05 level
SLEEP, Vol. 26, No. 3, 2003
330
Effect of Ethnicity on Sleep—Stepnowsky et al
nicities did not differ in age, body mass index, blood pressure level,
social class, or depression level.
A repeated measures ANCOVA was performed entering the sleepstage variables measured in minutes as the repeated measures across
location, ethnicity as the between-subjects factor, and age and AHI as the
covariates. The omnibus multivariate test revealed a significant location
by ethnicity interaction [Wilks’ Lambda = .708; F (4, 44) = 4.54; p =
.004]. Univariate tests revealed the location by ethnicity interaction was
significant for slow wave sleep (SWS) [F(1, 47) = 12.24; p = .001].
Between-subjects analysis showed that the ethnicities differed on
amount of SWS (F (1, 47) = 7.82; p = .007). Specifically, compared to
whites, blacks had fewer minutes of SWS, both in the hospital (blacks:
15.3 ± 16.6; whites = 49.3 ± 35.1) and home setting (blacks: 23.1 ± 28.1;
whites: 37.8 ± 28.4). Table 2 shows the means and standard deviations
for the groups on each of the sleep-stage variables by ethnicity, and Figure 1 illustrates this in graphic form. Figure 2 highlights the interaction
of sleep location by ethnicity. No other sleep-stage variables (stage 1,
stage 2, or REM) or measures of sleep continuity (arousal index, sleep
efficiency) were found to be significantly different across location or
ethnicity.
Follow-up analyses showed that SWS differed by site for each ethnicity such that whites had more SWS in the hospital than they did at
home (F = 10.10; p = .004). The reverse was true for blacks: blacks had
more SWS at home compared to the night in the hospital (F = 4.40; p =
.047). None of the variables that might have accounted for this interaction (social class or depressive symptoms) were significantly different
between the ethnicities.
DISCUSSION
M inutes
Blacks had significantly less deep sleep than did whites, whether they
were sleeping in the hospital laboratory or at home. These findings held
after including AHI as a covariate in the
repeated measure ANCOVA and constitute a partial replication of our earlier
observation.5 Additional studies with
larger sample sizes would be necessary to
determine what factor or factors are driving this ethnic difference. Unexpectedly,
this study also found that blacks
increased the amount of SWS from the
hospital laboratory to the home, while
whites had decreased SWS at home compared to the hospital recording. Sleep
apnea severity, as measured by the AHI,
did not change across locations.
Using SWS as an indicator, both
blacks and whites appear to be sensitive
to changes in their sleeping environment,
albeit in different directions. It is unclear
what is inducing the location effect. Our
analyses did not indicate that the groups
differed on depressive symptoms or
social class. Sleep quality is highly sensitive to several factors that cannot be controlled in the home, such as ambient
noise, temperature, and light, as well as
Figure 1—Graphic illustration of the distribution of polysomnographic sleep stages by ethnicity and sleeping location. SWS, the only
sleep-onset times and final awakening
sleep stage that was significantly different, is depicted at the top of the bars. There were no significant differences for any other
times. The subject’s choice of nighttime
sleep/wake stage. Abbreviations: SWS, slow wave sleep; REM, rapid eye movement sleep; WASO, wake after sleep onset.
activities is more restricted in hospital
settings as well. These factors, or any combination thereof, may be con60
tributing to the findings of this study.
The present study partially replicates the recent report of ethnic differences in laboratory-based polysomnography by Profant et al,5 name50
ly that blacks had less SWS than did whites, even when controlling for
AHI. Here, however, blacks did not have more total sleep time nor more
REM sleep than did whites. One possibility is that our laboratory stud40
ies enforced a consistent final awakening time of 0600, which may have
artificially curtailed customary total sleep time for many subjects. Doing
Whites
so would not only shorten total sleep time, but likely decrease REM time
30
as well. Thus, the logistics of this study might have made it less likely to
Blacks
detect differences in total sleep time and REM sleep between blacks and
whites. If this were true, however, we would then have expected to see
20
total sleep time and REM differences between blacks and whites in the
ad lib home sleep setting.
It is intriguing that blacks had diminished SWS in the hospital relative
10
to the home environment. To a certain extent, this observation is compatible with that of Mills et al,8 showing that blacks’ blood pressure falls
less in the hospital, an observation the authors attributed to blacks’ lower
level of comfort with the hospital setting. The current study is clearly
0
Hospital
Home
preliminary but represents a unique set of data that begins to address the
Figure 2—Interaction of ethnicity by sleeping location for slow wave sleep (in minutes).
question of polysomnographic differences between blacks and whites
Error bars represent mean ± SE.
and between different locations of sleep. Other larger-scale studies are
SLEEP, Vol. 26, No. 3, 2003
331
Effect of Ethnicity on Sleep—Stepnowsky et al
needed to replicate this finding, and multiple nights in each environment
will be required to tease apart those effects. It may also be worthwhile to
explore polysomnographic differences in other ethnic groups as well.
Although we believe these findings are of interest, we acknowledge
that the findings stem from a convenience sample. Therefore, we must
be cautious in concluding that our sample is representative of blacks and
whites in general. This limits the generalizations that can be made about
differences between these two ethnic groups. Nonetheless, these findings may have implications for the measurement of physiologic processes in the laboratory versus home environments for other ethnic
groups and may have clinical implications for the measurement of sleep
in different ethnic groups as well.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Karacan I, Thornby JI, Anch M et al. Prevalence of sleep disturbance in a primarily
urban Florida county. Soc Sci Med 1976; 10:239-44.
Ancoli-Israel S, Klauber MR, Stepnowsky C, Estline E, Chinn A, Fell R. Sleep-disordered breathing in African-American elderly. Am J Respir Crit Care Med
1995;152:1946-9.
Blazer DG, Hays JC, Foley DJ. Sleep complaints in older adults: a racial comparison. J
Gerontol Med Sci 1995;50A:M280-4.
Jean-Louis G, Magai CM, Cohen CI et al. Ethnic differences in self-reported sleep problems in older adults. Sleep 2001;24:926-33.
Profant J, Ancoli-Israel S, Dimsdale JE. Are there ethnic differences in sleep architecture? Am J Hum Biol 2002;14:321-6.
Rao U, Poland RE, Lutchmansingh P, Ott GE, McCracken JT, Lin KM. Relationship
between ethnicity and sleep patterns in normal controls: implications for psychopathology and treatment. J Psychiatr Res 1999; 33:419-26.
Poland RE, Rao U, Lutchmansingh P, et al. REM sleep in depression is influenced by
ethnicity. Psychiatry Res 1999;88:95-105.
Mills PJ, Berry CC, Dimsdale JE. Race affects the decline in blood pressure with hospitalization. Am J Hypertens 1997;10:1091-6.
Ferber R, Millman R, Coppola M, et al. Portable recording in the assessment of obstructive sleep apnea. Sleep 1994;17:378-92.
Agnew HW, Webb WB, Williams RL. The first night effect: an EEG study of sleep. Psychophysiology 1966;2:263-6.
Browman CP, Cartwright RD. The first-night effect on sleep and dreams. Biol Psychiatry 1980;15:809-12.
Coates TJ, George JM, Killen JD, Marchini E, Hamilton S, Thorensen CE. First night
effects in good sleepers and sleep-maintenance insomniacs when recorded at home.
Sleep 1981;4:293-8.
Sharpley AL, Solomon RA, Cowen PJ. Evaluation of first night effect using ambulatory monitoring and automatic sleep stage analysis. Sleep 1988;11:273-6.
EEG arousals: scoring rules and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 1992;15:17384
Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and
scoring system for sleep stages of human subjects. Los Angeles: Brain Information Service/ Brain Research Institute, UCLA, 1968.
Patterson TL, Semple SJ, Temoshok LR, et al. Depressive symptoms among HIV positive men: life stress, coping and social support. J Appl Biobehavioral Res 1993;1:64-87.
Murphy J. Psychiatric Instrument Development for Primary Care Research: Self-report
questionnaire. Bethesda, MD: NIMH; 1982. Report on Contract #80M0142810ID.
Schulberg HC, Saul M, McClelland M, Gangul M, Christy W, Frank RL. Assessing
depression in primary medical and psychiatric practices. Arch Gen Psychiatry
1985;42:1164-70.
Hollingshead AB. Two-factor index of social position. New Haven, CT: Department of
Sociology, Yale University; 1957.
SLEEP, Vol. 26, No. 3, 2003
332
Effect of Ethnicity on Sleep—Stepnowsky et al