Sleep. 16(7):632--637
© 1993 American Sleep Disorders Association and Sleep Research Society
Patterns of Sleep Episodes in Young and Elderly
Adults During a 36-]Hour Constant Routine
Daniell. Buysse, Timothy H. Monk, Charles F. Reynolds III, Deborah Mesiano,
Patricia R. Houck and David 1. Kupfer
Sleep and Chronobiology Center, Department of Psychiatry, University of Pittsburgh School of Medicine,
Pittsburgh, Pennsylvania, U.S.A.
The deterioration of nocturnal sleep and increase in daytime napping that accompany aging may
represent changes in homeostatic or circadian aspects of sleep regulation. We addressed these possibilities by
examining patterns of unintended sleep episodes, subjective sleepiness and mood during 36 hours of constant
wakeful bed rest in 17 healthy 20-30-year-olds (n = 9 men, n = 8 women) and 18 healthy 80-year-olds (n = 11
men, n = 7 women). Subjects had continuous electroencephalographic (EEG) monitoring and core body temperature
measurements and completed visual analog scale ratings of sleepiness and mood each hour. Young and elderly
groups both showed significant linear and circadian trends in sleep episodes across the interval. Group x time
interaction effects revealed a stronger linear trend in the young vs. elderly group, but no difference in circadian
trends. Analysis of age groups by sex demonstrated more complex interactions, with elderly men showing weaker
rhythmic (circadian) trends than elderly women. Subjective ratings of sleepiness and mood showed no difference
in temporal trends between young and elderly groups. These data support the hypothesis that age-related sleep
changes result primarily from decreases in homeostatic, rather than circadian, sleep propensity. They also confirm
the presence of sex differences in sleep patterns among the elderly and support the distinction between objective
and subjective measures of sleepiness. Key Words: Aging-Sleep-Circadian rhythms-Sleep deprivation-Core
body temperature- Homeostasis.
Summary:
Aging and gender have clear and consistent effects
on subjective and objective measures of nocturnal sleep
(1-3). Aging also affects daytime sleepiness. Several
studies have documented increased napping in the elderly (4-6), and observational studies in nursing homes
demonstrate polyphasic sleep/wake patterns in many
residents (7,8). However, objective sleep latency studies have shown both increased (9,10) and decreas{~d
(11,12) daytime sleepiness in the elderly, compared to
young adult subjects.
A number offactors-including subjects' prior sleep
histories, specific sleep pathologies, daytime activity
levels and physical and neuropsychiatric health - may
affect age-related changes in sleep-wake patterns.
However, changes in these patterns could also result
from changes in homeostatic and circadian sleep propensities in the elderly.
We addressed these issues by examining temporal
patterns of unintended sleep episodes during 36 hours
Accepted for publication June 1993.
Address correspondence and reprint requests to Dr. Buysse, 3811
O'Hara St., Room E-1128, Pittsburgh, Pennsylvania 15213, U.S.A.
of constant wakeful bed rest in healthy elderly and
young subjects. The major goal of the protocol was to
define age-related changes in circadian rhythms of core
body temperature, neuroendocrine measures, mood and
performance, while minimizing effects due to medical
neuropsychiatric illness. However, investigators have
long recognized the occurrence of unintended "microsleeps" during sleep deprivation studies (e.g. 13).
We postulated that the frequency and pattern of these
sleep episodes would reflect underlying circadian and
homeostatic sleep propensities. For instance, a linear
(monotonic) increase in sleep episodes during the 36
hours would indicate a homeostatic increase in sleep
tendency across the interval; a peak in the occurrence
of sleep episodes followed by a subsequent decrease
would indicate a rhythmic (circadian or ultradian) process. This methodology complements previous studies
that have either assessed subjects' habitual behavior,
or measured their sleep tendency on repeated discrete
trials, e.g. multiple sleep latency test (MSLT) (9,11,12).
In preliminary analyses, we have found that healthy
elderly subjects maintain robust circadian rhythms of
some measures (including core body temperature and
cortisol secretion), but that elderly men show decreased
632
SLEEP EPISODES DURING A CONSTANT ROUTINE
circadian modulation of psychomotor performance
measures, relative to young men (14,15). These findings suggest differential effects of aging on endogenous
circadian rhythms compared to behavioral rhythms
requiring cortical or brainstem input. Therefore, we
also sought to determine whether patterns of unintended sleep episodes would follow a time course resembling that of endogenous circadian rhythms such
as body temperature (with robust circadian patterns in
both young and old) or that of psychomotor performance rhythms (with less circadian modulation in the
elderly). Based on our results using the MSL T (II), we
hypothesized that elderly subjects would have fewer
sleep episodes than young adults during the constant
routine, but that both groups would show similar circadian patterns of sleep episodes.
METHODS
Subjects
Subjects included nine healthy young men and eight
healthy young women between 21 and 28 years of age
and 11 elderly men and seven elderly women between
80 and 86 years old. These subjects overlap partially
with those in other reports of sleep, circadian characteristics and performance in the elderly (3,15,17).
We have not previously reported data pertaining to
sleep episodes during the constant routine, nor any data
from the young women.
Subjects had no current or past psychiatric disorder
according to the Schedule for Affective Disorders (18)/
Research Diagnostic Criteria (19), no acute or unstable
medical/neurological disorder and no specific sleep
complaint or disorder. Physical examinations, electrocardiograms (ECG), laboratory blood work and two
nights of polysomnography verified that these conditions were met. Medications taken by subjects did not
have substantial effects on sleep; specific medications
included erythromycin (for acne; one young man), nonsteroidal anti-inflammatory drugs (three elderly men),
oral hypoglycemic (one elderly man), calcium channel
blocking agent (one elderly man), I-thyroxine (one elderly man), diuretic (one elderly man), p.r.n. antihistamines (four doses; one young man), ophthalmic
iJ-blocker (two elderly men), systemic iJ-blocker (one
elderly woman) and oral contraceptives (three young
women). Young women were studied during the first
2 weeks of their menstrual cycles.
The level of any depressive symptoms was rated with
the Hamilton Rating Scale for Depression (20), cognitive status was assessed with the Folstein Mini-Mental State Examination (21) and sleep quality was assessed with the Pittsburgh Sleep Quality Index (PSQI)
(22).
633
Study procedures
Previous reports (15,17) detail the study methods.
The protocol consisted of 60 consecutive hours spent
in temporal (but not social) isolation in the Sleep and
Chronobiology Center. For the first 24 hours, subjects
followed their habitual sleep-wake cycle. For the next
36 hours, subjects followed a constant routine of wakeful bed rest. Meals were replaced with hourly allotments ofliquid nutritional supplement equal to 1!24th
of the usual daily caloric intake.
On-line rectal temperatures were recorded every 1
minute and plasma was sampled through an intravenous catheter every 20 minutes for determination of
cortisol and melatonin over the entire 60 hours. Subjects completed a series of computer-based mood performance tests every hour; results for a subgroup of
the current subjects have been reported elsewhere (15).
In the current report, we present only the results of 10
cm visual analog scale ratings of sleepiness and global
affect [a subjective measure of mood state (23)] as a
contrast to electroencephalographic (EEG) sleep measures.
Nocturnal sleep was recorded using a routine EEG
sleep montage, consisting of C 3 or C 4 EEG referenced
to Al + A2 (high frequency filter setting 30 Hz, low
frequency filter setting 0.3 Hz), bilateral electrooculogram (EOG) referenced to Al + A2 and submental
electromyogram (EMG). Sleep was scored in 60-second epochs using standard criteria (24). During the
constant routine, subjects were monitored continuously with the same EEG sleep montage. Polysomnographic technologists monitored the record and
awakened the subject if sleep was detected. In addition,
research assistants constantly monitored computer
testing and on-line temperature, interacted socially with
subjects and enforced wakefulness when necessary.
Unintended sleep episodes during the constant routine were scored by polysomnographic technologists
whose reliability for traditional sleep stage measurements is ::::85% in monthly reliability studies. Sleep
episodes were identified by the appearance of theta and
delta EEG frequencies, often accompanied by slow
rolling eye movements or a decrease in muscle tone.
Less frequently, sleep could be detected by the presence
of K-complexes of sleep spindles. A minimum duration of 3 seconds was required for scoring sleep episodes. The 36-hour study was divided into three 12hour segments and scored without knowledge of the
specific time segment.
As a measure of inter-rater reliability, 10 technicians
each scored sleep episodes in a 24-hour segment from
one constant routine study in a young adult. The number of sleep episodes scored each hour by each technician was entered into an analysis of variation (ANOSleep, Vol. 16. No.7, 1993
634
D. J. BUYSSE ET AL.
vA), from which we calculated an intraclass correlation
(assuming fixed raters) of 0.74 (F9207 = 6.20, P <
0.0001), indicating good overall agreement.
RESULTS
Baseline sleep and circadian measures
Polysomnographic measures from the night before
the constant routine differed between elderly and young
Statistical analyses
subjects in expected directions. Specifically, young
adults had significantly better sleep efficiency, longer
Clinical and polysomnographic differences between
sleep duration and more slow wave sleep. Young womgroups were tested with ANOVAs and x2 contingency
en had later habitual and laboratory sleep times than
tables. To test our main hypothesis regarding sleep
elderly men. Estimates of core body temperature
episodes, we first compared the entire group of young
rhythm amplitude and acrophase (derived from bestadults to the entire group of elderly adults. As a more
fit sinusoids for each subject) showed significantly
exploratory analysis, we then compared young and elgreater amplitude in elderly than young adult women,
derly subjects subdivided by sex into four subgroups
but no other group differences.
(young/elderly and men/women). We analyzed sleep
episodes in the following way: a) The number of sleep
episodes was tabulated for each of the 36 hours of the Sleep episodes during the constant routine:
study in each subject. b) Because of the low overall young (n = 17) vs. elderly (n = 18) subjects
rate of sleep episodes, we collapsed these data into the
The young adult group had a significantly greater
number of sleep episodes per three hours across the
mean
number of sleep episodes than the elderly group
study, yielding 12 3-hour blocks per subject. c) Because
(68.1
±
36.4 vs. 25.8 ± 19.2; F= 11.91, p = 0.002).
the number of sleep episodes was not normally disThe
temporal
pattern of sleep episodes also differed
tributed, we performed square-root transformations on
between
the
groups,
as indicated by a significant group
each subject's data. d) Mean values were calculated for
x
time
interaction
effect
(F = 3.47, p = 0.0006). The
the number of sleep episodes (square-root transinteraction
resulted
from
differences
in the linear trend,
formed) in each 3-hour block for each group of subwhich
also
showed
a
significant
group
x time interjects. e) Statistical analyses were conducted on these
(F
=
12.86,
p
=
0.001).
Quadratic
and
action
effect
group data.
cubic
trends
did
not
show
significant
group
x
time
We analyzed sleep episode data using repeated-measures ANOVA with orthogonal contrasts, with age interactions, indicating similar rhythmic trends in the
group (young vs. elderly) as one factor and 3-hour time young and elderly groups. Orthogonal contrasts for
blocks as the repeated measure. Huyn-Feldt probabil- young and elderly groups analyzed separately conities are reported for overall effects due to time and firmed a stronger linear (monotonic) trend in the young
group x time interactions. Orthogonal contrasts ex- adult group than the elderly group, but no age-related
amine linear, quadratic and cubic trends, as well as differences in the strength of quadratic or cubic (rhythgroup x time interaction effects. Linear trends indi- mic) trends for sleep episodes during the constant roucated monotonic changes as a function of time into tine (data available upon request).
study and were interpreted as indicators of a homeostatic sleep propensity. Quadratic and cubic trends indicated one or two "turning points" in the data and Sleep episodes during the constant routine
were interpreted as evidence of rhythmic (circadian or young vs. elderly subjects subdivided by sex
ultradian) trends in sleep propensity. We repeated the
Data for the four subgroups (young/old, men/womorthogonal contrast analyses for elderly and young
groups separately to confirm the strength oflinear, qua- en) again showed significant differences in the number
dratic and cubic components in each group. We also of sleep episodes (F= 4.28, p = 0.01), and in the overall
report exploratory analyses subdividing each age group temporal pattern (significant group x time interaction
by sex in order to determine whether young-elderly effect F = 1.75, p = 0.02) (Fig. 1). This four-way group
x time interaction resulted from group differences in
differences were different among men and women.
We used identical ANOVA procedures for visual the linear trend (F = 4.67, p = 0.008), with no interanalog scale ratings of subjective sleepiness and global actions observed for quadratic or cubic trends. Results
affect. These ratings were selected to contrast subjec- of orthogonal contrast analyses for each of the four
tive experiences of sleepiness and mood from physi- groups demonstrate strong linear trends for both young
ologically-defined sleep episodes. Ratings were square- groups and statistically significant rhythmic (quadratic
root transformed and averaged over 3-hour blocks for and/or cubic) trends in each group except elderly men
analysis.
(data available upon request).
Sleep, Vol. 16, No.7, 1993
SLEEP EPISODES DURING A CONSTANT ROUTINE
SLEEP EPISODES DURING A 36-HOUR CONSTANT ROUTINE
SUBJECTlVE SLEEPINESS DURING A 36-HOUR CONSTANT ROUTlNE
100
12
SUBJEClS
I.
635
o
o
OLD FnlALLS CN..1)
••
OLD WALLS CN=l1)
8.
.!2!!...
•
SUBJECTS
o
•
o
yOu·· ........ !l!=!L
~
OLD MAW ,"=111
YOu .......... <l!:!L
OLD ""AlES CN=1)
.YOu •• ~
.. .
...a
~
20
I.
~oo \1.00 \<§l0 \!.oo 1.\00 ~oo 0,>00 (/>00 ~oo <J.oo -e,oo .,,00 1.\00
TIME
FIG. 1. Mean number of sleep episodes per 3-hour interval for
each of the four subgroups. Each point represents the average, within
subgroup, ofthe total number of episodes during the 3-hour intervaL
Data in the figure are back-transformed into original units from the
mean data using square-root transformed data for each subject (see
Methods).
Patterns of sUbjective sleepiness and
global affect during the constant routine
We first examined subjective sleepiness and global
affect for the entire group of young adults vs. the group
of elderly adults. Despite a trend toward greater sleepiness in the young group (F = 3.72, p = 0.06), neither
the overall group x time interaction effect nor linear,
quadratic or cubic group x time interaction effects
were statistically significant. This indicates overall
similar temporal patterns of subjective sleepiness in
young and elderly groups, including significant linear
(F = 46.53, p < 0.0001) and quadratic (F = 23.44, p
< 0.0001) trends. Significant group x time interaction
effects were not observed between young and elderly
groups for global affect, despite significant temporal
trends for linear (F = 15.03, p = 0.0005) and quadratic
(F = 10.57, p = 0.003) components.
We also compared subjective sleepiness for the four
subgroups (young/elderly, men/women) (Fig. 2). The
group x time interaction effect was not significant,
indicating similar temporal trends in all groups. Within
individual subgroups, significant linear trends were observed in each group, and significant rhythmic trends
(quadratic or cubic) were found in each group except
elderly women (data available upon request). For global affect in the four subgroups (Fig. 3), no significant
overall group or group x time interaction effects were
noted. In general, global affect showed weaker temporal trends than sleep episodes or sleepiness ratings
in all subgroups.
FIG. 2. Mean visual analog scale ratings of sleepiness per 3-hour
interval for each of the four subgroups. Each point represents the
average, within subgroup, of the mean sleepiness rating during the
3-hour interval. Data in the figure are back-transformed into original
units from the mean data using square-root transformed data for
each subject (see Methods). Higher numbers indicate greater sleepiness.
DISCUSSION
Compared to young adults, eldterly subjects had fewer episodes of inadvertent sleep, a smaller linear
(monotonic) increase in sleep episodes and no difference in rhythmic (circadian, ultradian) trends in unintended sleep episodes across the 36-hour constant
GLOBAL AFFECT DURING A 36-HOUR CONSTANT ROUTINE
10.
.0
8.
SUBJECTS
o ""L.£S
• you.L1W.£S <l!:!L
o
OLD
04=11)
OLD ....Al.ESfH=1)
•
2.
YOu.£!!l'~
I.
00,00 \1.00 \<§l0 \&00 1.\00 1.~O 0,>00 O~OO 00,00 <J.OO ~O \!.OO 1.'f:l0
TIME
FIG. 3. Mean visual analog scale ratings of global affect (mood)
per 3-hour interval for each of the four subgroups. Each point represents the average, within subgroup, of the mean global affect rating
during the 3-hour interval. Data in the figure are back-transformed
into original units from the mean data using square-root transformed
data for each subject (see Methods). Higher numbers indicate a greater sense of well-being.
Sleep. Vol. 16. No.7. 1993
636
D. J. BUYSSE ET AL.
routine. However, sex differences were also noted. Elderly men did not have significant rhythmic trends in
sleep episodes, but elderly women had strong rhythmic
trends. Subjective ratings of sleepiness and mood revealed no differences in temporal trends for elderly and
young subjects. Our results suggest that aging diminishes homeostatic sleep propensity (indicated by linear
trends) in both men and women, but that circadian
sleep propensity (indicated by quadratic and cubic
trends) is also diminished in elderly men. Objectively
defined sleep episodes do not follow the same temporal
pattern as subjective sleepiness and mood.
Although we know of no previous investigations that
used similar methodology, the current results can be
interpreted in the context of previous self-report, observational and MSLT studies. Studies of daytime napping generally report an age-related increase in nap
frequency, particularly among retired adults (5,6,25)
(see refs. I and 26 for review). Monitoring studies,
using behavioral observation, wrist actigraphy or EEG
monitoring, demonstrate increased daytime sleep and
reduced nocturnal sleep in the elderly, particularly those
with neurocognitive disorders (7,8,27,28). MSLT studies prove more controversial. Available evidence supports either an increase (10) or a decrease (11,12) in
daytime sleep tendency in older adults compared to
young adults.
In a circadian modification of the MSLT procedure,
Richardson et al. (29) found no difference in the circadian pattern of sleep latency between young and elderly subjects. Our results generally support this finding, despite differences in study procedures and demand
characteristics. For instance, our study was conducted
under the instruction to resist sleep, and Richardson's
with the expectation of sleep onset. Our subjects were
also older than Richardson's (mean of82 vs. 70 years).
Taken together, previous studies suggest that agerelated changes in daytime sleepiness may result from:
1) reduced circadian variation in sleep tendency; 2) the
emergence of a "circasemidian" (12-hour) rhythm of
sleep propensity (30) or 3) a change in mean level or
homeostatic regulation of sleepiness. Our current data
support the third possibility in elderly subjects of both
sexes, and the first possibility only in elderly men. The
current analyses do not adequately address the circasemidian possibility.
Subgroup analyses suggested sex ·differences in the
temporal pattern of sleep episodes, with elderly men
showing weaker rhythmic trends than elderly women.
This observation supports previous findings of sexspecific changes in aging circadian rhythms, such as
core body temperature (16), that generally show larger
amplitude rhythms in elderly women. It also supports
our finding that elderly men show weaker circadian
influences in performance and activation than young
Sleep. Vol. 16. No.7. 1993
men in this protocol (15). Finally, it is interesting to
note that our group of young women actually had a
smaller core body temperature amplitude than elderly
women as well as weaker rhythmic trends in sleep episodes. Thus, age and sex appear to interact in complex
ways regarding the circadian regulation of sleep tendency.
Acknowledgment: This study was supported by NIA grant
AG06836 (T.H.M., C.F.R., D.J.B.) and NIMH grant
MH48891 (D.J.B.).
REFERENCES
1. Bliwise DL. Sleep in normal aging and dementia. Sleep 1993;
16:40-81.
2. Dement WC, Laughton EM, Carskadon MA. "White paper" on
sleep and aging. JAm Geriatr Soc 1982;30:25-50.
3. Buysse DJ, Reynolds CF, Monk TH, Hoch CC, Yeager AL,
Kupfer DJ. Quantification of subjective sleep quality in healthy
elderly men and women using the Pittsburgh Sleep Quality Index
(PSQI). Sleep 1991;14:331-8.
4. Buysse DJ, Browman KE, Monk TH, Reynolds CF, Fasiczka
AL, Kupfer DJ. Napping and 24-hour sleep/wake patterns in
healthy elderly and young adults. JAm Geriatr Soc (in press).
5. Tune GS. Sleep and wakefulness in 509 normal human adults.
Br J Med PsychoI1969;42:75-80.
6. Webb WB. Patterns of sleep in healthy 50-60 year old males
and females. Res Comm Psychol. Psychiatry, Behav 1981;6: 13340.
7. Jacobs D, Ancoli-Israel S, Parker L, Kripke DF. Twenty-fourhour sleep-wake patterns in a nursing home population. Psychol
Aging 1989;4:352-6.
8. Regestein QR, Morris J. Daily sleep patterns observed among
institutionalized elderly residents. JAm Geriatr Soc 1987;35:
767-72.
9. Carskadon MA, Van Den Hoed J, Dement We. Insomnia and
sleep disturbances in the aged; sleep and daytime sleepiness in
the elderly. J Geriatr Psychiatry 1980; 13: 135-51.
10. Carskadon MA. Ontogeny of human sleepiness as measured by
sleep latency. In: Dinges DF, Broughton RJ, eds. Sleep and
alertness. New York: Raven Press, 1989:53-69.
11. Reynolds CF, Jennings JR, Hoch CC, et al. Daytime sleepiness
in the healthy "old old": a comparison with young adults. JAm
Geriatr Soc 1991;39:957-62.
12. Levine B, Roehrs T, Zorick F, Roth T. Daytime sleepiness in
young adults. Sleep 1988; 11:39-46.
13. Southmayd SE, Cairns J, Delva NJ, Letemendia FJJ, Brunet
DG. Awake, perchance asleep? Br J Psychiatry 1986; 148:748-9.
14. Buysse DJ, Monk TH, Reynolds CF, et al. Circadian rhythms
in the healthy elderly. In: Kuna ST, Suratt PM, Remmers JE,
eds. Sleep and respiration in aging adults. New York: Elsevier,
1991:31-8.
15. Monk TH, Buysse DJ, Reynolds CF, Jarrett DB, Kupfer DJ.
Rhythmic versus homeostatic influences on mood, activation
and performance in the elderly. J Gerontal: Psychol 1992;47:
221-7.
16. Moe KE, Prinz PN, Vitiello MV, Marks AL, Larsen LH. Healthy
elderly women and men have different entrained circadian temperature rhythms. JAm Geriatr Soc 1991;39:383-7.
17. Monk TH, Reynolds CF, Buysse DJ, et al. Circadian characteristics of healthy 80-year-olds and their relationship to objectively recorded sleep. J Gerontal 1991;46:MI71-5.
18. Endicott J, Spitzer RL. A diagnostic interview. The schedule
for affective disorders and schizophrenia. Arch Gen Psychiatry
1978;35:837-44.
19. Spitzer RL, Endicott J, Robins E. Research diagnostic criteria:
rationale and reliability. Arch Gen Psychiatry 1978;35:773-82.
20. Hamilton M. A rating scale for depression. J Neural Neurasurg
Psychiatr 1960;23:56-62.
SLEEP EPISODES DURING A CONSTANT ROUTINE
21. Folstein MF, Folstein SW, McHugh PRo Mini-mental state: a
practical method for grading the cognitive state of patients for
the clinician. J Psych Res 1975;12:189-98.
22. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ.
The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989;28: 193-213.
23. Monk TH. A visual analogue scale technique to measure global
vigor and affect. Psychiatry Res 1989;27:89-99.
24. Rechtschaffen A, Kales A, eds. A manual oj standardized terminology, techniques and scoring system Jar sleep stages oj human subjects. Washington DC: U.S. Government Printing Office, 1968.
25. Spiegel R. Questionnaire data from the present subject sample.
In: Sleep and sleeplessness in advanced age. New York: Spectrum
Publications, Inc., 1982:23-37.
26. Dinges DF. Napping patterns and effects in human adults. In:
Dinges DF, Broughton RJ, eds. Sleep and alertness. New York:
Raven Press, 1989: 171-204.
637
27. Prinz PN, Peskind ER, Vitaliano PP, et al. Changes in the sleep
and waking EEGs of nondemented and demented elderly subjects. JAm Geriatr Soc 1982;30:86-93.
28. Witting W, Kwa IH, Eikelenboom P, Mirmiran M, Swaab DF.
Alterations in the circadian rest-activity rhythm in aging and
Alzheimer's Disease. BioI Psychiatry 1990;27:563-72.
29. Richardson GS, Carskadon MA, Orav EJ, Dement WC. Circadian variation of sleep tendency in elderly and young adult
subjects. Sleep 1982;5:S82-594.
30. Broughton RJ. Chronobiological aspects and models of sleep
and napping. In: Dinges DF, Broughton, RJ, eds. Sleep and
alertness. New York: Raven Press, 1989:71-98.
31. Hoch CC, Reynolds CF, Monk TH, et al. Comparison of sleepdisordered breathing among healthy elderly in the seventh, eighth,
and ninth decades oflife. Sleep 1990; 13:502-11.
32. Ancoli-Israel S, Kripke DF, Klauber MR, Mason WJ, Fell R,
Kaplan O. Sleep-disordered breathing in community-dwelling
elderly. Sleep 1991;14:486-95.
Sleep. Vol. 16. No.7. 1993