Available via license: Public Domain Mark 1.0
Content may be subject to copyright.
Sleep, Sleepiness,
and Alcohol Use
Timothy Roehrs, Ph.D., and Thomas Roth, Ph.D.
The study of alcohol’s effects on sleep dates back to the late 1930s. Since then, an extensive
literature has described alcohol’s effects on the sleep of healthy, nonalcoholic people. For
example, studies found that in nonalcoholics who occasionally use alcohol, both high and
low doses of alcohol initially improve sleep, although high alcohol doses can result in sleep
disturbances during the second half of the nocturnal sleep period. Furthermore, people can
rapidly develop tolerance to the sedative effects of alcohol. Researchers have investigated the
interactive effects of alcohol with other determinants of daytime sleepiness. Such studies
indicate that alcohol interacts with sleep deprivation and sleep restriction to exacerbate
daytime sleepiness and alcohol-induced performance impairments. Alcohol’s effects on other
physiological functions during sleep have yet to be documented thoroughly and unequivocally.
K
EY WORDS: sleep disorder; physiological AODE (effects of alcohol or other drug use, abuse,
and dependence); REM (rapid eye movement) sleep; NREM (nonrapid eye movement) sleep;
circadian rhythm; melatonin; prolactin; body temperature; attention; time of day; insomnia;
dose-response relationship
A
lcohol affects sleep, daytime
alertness, and certain physiolog-
ical processes that occur during
sleep. Its impact on human sleep has
received much scientific study dating
back to early experiments by Kleitman
(1939), described in his book Sleep and
Wakefulness. In that monograph, the
author summarizes the effects that alcohol
consumed 60 minutes before bedtime
has on body temperature and motility
during sleep in healthy nonalcoholic
people. In the 1960s and 1970s, after
scientists had identified various sleep
states (e.g., rapid eye movement [REM]
sleep) and had standardized electrophys-
iological methods to document sleep,
research on alcohol’s effects on the
sleep of healthy nonalcoholic and non-
insomniac volunteers and on the sleep
of alcoholics increased substantially. More
recently, with the emergence of the field
of sleep-disorders medicine, researchers
and clinicians have focused their atten-
tion on alcohol’s effect on primary sleep
disorders, such as sleep apneas, which
are short (i.e., 10 to 30 seconds long)
episodes of breathing obstruction. This
attention to sleep disorders also has
sensitized investigators and clinicians
to the impact that disrupted and short-
ened sleep has on daytime alertness.
As a result, various studies have investi-
gated the potential interactive effects
of alcohol with daytime alertness and
daytime functioning in both healthy
people and patients with sleep disorders.
This article provides an overview of
alcohol’s effects on normal sleep, sleep
physiology, and daytime alertness in
nonalcoholic people. (The accompany-
ing article by Brower, pp. 110–125 in
this issue, discusses alcohol’s effects on
sleep in alcoholics.) The current article
reviews normal sleep physiology, describes
alcohol’s effects on the various sleep
states and sleep stages, and explores
some of the mechanisms through which
alcohol may exert those effects. It then
summarizes the relationship of noctur-
nal sleep to daytime alertness and how
TIMOTHY ROEHRS, PH.D., is director
of research at the Sleep Disorders and
Research Center of the Henry Ford
Hospital and adjunct professor of psychi-
atry in the Department of Psychiatry
and Behavioral Neuroscience, Wayne
State University, Detroit, Michigan.
T
HOMAS ROTH, PH.D., is division head
of the Sleep Disorders and Research
Center of the Henry Ford Hospital
and adjunct professor of psychiatry
in the Department of Psychiatry and
Behavioral Neuroscience, Wayne State
University, Detroit, Michigan.
This work has been supported by
National Institute on Alcohol Abuse
and Alcoholism grant R01–AA–11264.
Vol. 25, No. 2, 2001 101
alcohol affects this relationship. The
article ends with a discussion of alcohol’s
effects on sleep in people with primary
insomnia.
Normal Sleep Physiology
As most people know from their own
experience, sleep is not uniform through-
out the night. For example, at certain
times during the night, it is very diffi-
cult to wake a sleeping person, whereas
at other times, the slightest sound will
alert the sleeper. Extensive studies have
identified two different sleep states: REM
sleep and nonrapid eye movement
(NREM) sleep. Furthermore, NREM
sleep can be divided into four stages
based on how easy it is to arouse a sleeper
(i.e., how “deep” the sleep is).
These different sleep states and sleep
stages are defined based on scoring cri-
teria for three electrophysiological mea-
surements that were first published in
1968 and have been employed ever
since in sleep laboratories around the
world. The three electrophysiological
measurements are recorded simultane-
ously and comprise the following:
• The electroencephalogram (EEG),
which traces the electrical activity
of the brain through electrodes
placed on the scalp. These measure-
ments produce characteristic brain
waves called alpha, beta, delta, and
theta rhythms, which differ in their
frequencies.
• The electrooculogram (EOG),
which measures eye movements
through electrodes placed on the
skin around the eyes and records
tiny electric signals that occur
when the eyes move.
• The electromyogram (EMG), which
measures the electrical activity of
muscles through electrodes placed
on the skin in various body regions.
This technique can measure even
small muscle movement during
sleep, such as twitching.
The following paragraphs describe
how these measurements are used to
distinguish different sleep states and
sleep stages.
Stages of NREM and REM Sleep
When comparing the EEG readings
of various sleep stages, researchers and
clinicians assess the frequency of the
brain waves, measured in hertz (Hz),
and the size, or amplitude, of the brain
waves, measured in microvolts. Both
the frequency and amplitude of the
In addition to effects
on sleep initiation and
sleep maintenance,
researchers have found
that alcohol consistently
affects the proportions of
the various sleep stages.
brain waves, as well as the EOG and
EMG readings, differ for various stages
of wakefulness and sleep (see figure).
During active wakefulness (i.e., when
the person is awake and pursuing nor-
mal activities), the EEG is characterized
by high frequencies (i.e., 16 to 25 Hz)
and low voltage (i.e., 10 to 30 microvolts).
EOG readings during wakefulness
exhibit REMs, and EMG readings gen-
erally show a high amplitude indicative
of large muscle movements.
During relaxed wakefulness (i.e.,
when a person is awake but has his or
her eyes closed and is relaxed), the EEG
is characterized by a pattern of alpha
waves with a frequency of 8 to 12 Hz
and an amplitude of 20 to 40 microvolts.
EOG readings show slow, rolling move-
ments at the transition to NREM sleep.
EMG readings show reduced amplitudes.
During NREM sleep, the frequency
of the brain waves slows further, whereas
the amplitude continues to increase.
Thus, when the arousal threshold is
highest (i.e., sleep is “deepest”), the EEG
shows slow-wave sleep with a frequency
of 0.5 to 2.0 Hz and an amplitude of
75 microvolts or greater. EOG tracings
indicate cessation of eye movements,
and EMG readings are gradually reduced,
even though episodic repositioning of
the body and other motor events occur.
Based on the simultaneous analysis of
all three measurements, NREM sleep is
classified into four stages that are charac-
terized by increasing arousal thresholds.
Thus, stage 1 (i.e., drowsy sleep) has the
lowest arousal threshold; stage 2 (i.e.,
light sleep) is intermediate; and stages 3
and 4 (i.e., deep sleep), which collectively
are also called slow-wave sleep (SWS),
have the highest arousal threshold.
During REM sleep, cortical EEG
readings revert to the low-voltage-mixed-
frequency pattern seen during drowsy
sleep. The EOG displays the bursts of
rapid eye movements that give this stage
its name. The EMG is reduced to its
lowest level for the night. In fact, most
major voluntary muscle groups are par-
alyzed, because certain nerve cells in the
spinal cord (i.e., motor neurons) are not
responding to nerve signals. Arousal
thresholds in REM are relatively low,
similar to NREM stages 1 or 2.
Tonic and Phasic Periods
of REM Sleep
REM sleep can be further subdivided
into tonic and phasic periods. During
the tonic periods, which account for
the majority of REM sleep, muscle
tone is decreased and the EEG is simi-
lar to that seen during stage 1 NREM
sleep. These tonic periods are interrupted
by intermittent phasic REM events.
For example, the eye movements char-
acteristic of REM sleep occur in bursts
during these phasic periods, which are
followed by the tonic periods of EOG
quiescence. Coupled with the bursts
of eye movements are phasic muscle
twitches, typically involving peripheral
muscles, although the reduced muscle
tone (i.e., atonia) characteristic of the
tonic periods continues in most muscle
groups. In addition, bursts of activity
occur during the phasic periods in body
functions that are controlled by the
autonomic nervous system
1
; these bursts
1
The autonomic nervous system controls involuntary vital
functions, such as the activities of the heart, lungs,
gastrointestinal tract, and glands.
102 Alcohol Research & Health
Sleep, Sleepiness, and Alcohol Use
of activity are reflected by irregularities
in cardiopulmonary function (e.g.,
heart rate and breathing rate).
NREM–REM Cycles
An ultradian process—a biorhythm
with a cycle of less than 24 hours—
within sleep controls the alternation
between NREM and REM sleep
throughout the night. This ultradian
process creates cycles of NREM sleep
followed by REM sleep that last approx-
imately 90 to 120 minutes, yielding
four to five such cycles over a standard
8-hour sleep period. In the first two of
those cycles, slow-wave NREM sleep
predominates, whereas the REM periods
are generally quite short (i.e., 5 to 10
minutes). Conversely, in the last two or
three cycles, REM sleep predominates,
sometimes continuing uninterrupted for
30 to 40 minutes, and slow-wave NREM
sleep is almost nonexistent. (The signifi-
cance of this ultradian cycling of NREM
and REM sleep to alcohol’s effects on
sleep is described in the following section
of this article.)
Alcohol’s Effects
on Sleep Physiology
To assess alcohol’s effects on sleep,
investigators conducting a typical sleep
study administer alcohol to their subjects
approximately 30 to 60 minutes before
bedtime. As a result of this schedule,
alcohol concentrations in the breath or
blood usually peak at “lights-out.” Using
this approach, researchers have exten-
sively studied alcohol’s effects in healthy
people at doses ranging from 0.16 to
1.0 grams of alcohol per kilogram of
body weight (g/kg) (Williams and Salamy
1972). These doses, which correspond
to approximately one to six standard
drinks,
2
yield breath alcohol concentra-
tions (BrACs) as high as 0.105 percent.
3
Some studies using this range of alcohol
2
A standard drink is defined as one 12-ounce bottle of
beer or wine cooler, one 5-ounce glass of wine, and 1.5
ounces of 80 proof distilled spirits.
3
Breath alcohol concentrations are another way of
quantifying alcohol levels in the body and are approxi-
mately the same as blood alcohol concentrations after a
given alcohol dose.
doses reported that the study partici-
pants fell asleep faster (i.e., had reduced
sleep latency) than without alcohol con-
sumption. One study found an increased
sleep time at a low alcohol dose (i.e.,
0.16 g/kg) but detected no such effect
at higher alcohol doses (i.e., 0.32 and
0.64 g/kg) (Stone 1980).
Some investigators have separately
analyzed alcohol’s effects during the
first and second half of the nighttime
sleep period. These studies found that
particularly at higher alcohol doses,
increased wake periods or light stage 1
sleep periods occurred during the sec-
ond half of the sleep period (Williams
et al. 1983; Roehrs et al. 1991). This
second-half disruption of sleep continuity
is generally interpreted as a “rebound
effect” once alcohol has been completely
metabolized and eliminated from the
body. The term “rebound effect” means
that certain physiological variables (e.g.,
sleep variables, such as the amount of
REM sleep) change in the opposite
direction to the changes induced by alco-
hol and even exceed normal levels once
alcohol is eliminated from the body.
This effect results from the body’s adjust-
ment to the presence of alcohol during
the first half of the sleep period in an
effort to maintain a normal sleep pat-
tern. Once alcohol is eliminated from
the body, however, these adjustments
result in sleep disruption. This hypoth-
esis is supported by the known rate of
alcohol metabolism, which leads to a
decrease in BrAC of 0.01 to 0.02 per-
cent per hour. Given that in such
experiments, the typical peak BrACs
measured shortly before sleep are 0.06
to 0.08 percent, alcohol metabolism at
this rate would be completed within
4 to 5 hours of sleep onset; thus, the
sleep disruption during the second half
of the night would coincide with the
clearance of alcohol from the body.
In addition to these effects on sleep
initiation and sleep maintenance,
researchers have found that alcohol
consistently affects the proportions of
the various sleep stages. Thus, most
studies have reported a dose-dependent
suppression of REM sleep at least during
the first half of the sleep period (Williams
and Salamy 1972). As noted earlier,
the amount of REM sleep time is lower
during the first half of the night relative
to the second half of the night; conse-
quently, the full REM-suppressive effect
of alcohol is probably underestimated
in most studies. To determine alcohol’s
full effect on REM sleep, investigators
would need to administer an additional
alcohol dose in the middle of the night,
thereby causing alcohol’s peak concen-
trations to coincide with the majority
of REM sleep time. No such studies
have been conducted, however.
Those studies that have demonstrated
alcohol-induced REM suppression dur-
ing the first half of the sleep period also
have frequently found an REM rebound
(i.e., longer-than-normal REM periods)
during the second half of the night
(Williams and Salamy 1972). As a result,
the overall amount of REM sleep in
subjects receiving alcohol before sleep-
ing did not differ from that in subjects
receiving a nonalcoholic drink (i.e., a
placebo). As with the increased periods
of wakefulness or light sleep, the REM
rebound during the second half of the
night is associated with the completed
alcohol metabolism and elimination
from the body. The neurobiological
mechanisms responsible for the rebound
of either wakefulness or REM sleep are
still unknown.
Some studies also found an alcohol-
related increase in the amount of SWS
(i.e., stages 3 and 4 NREM sleep) in
the first half of the sleep period (Williams
and Salamy 1972). In addition to the
alcohol dose consumed, the basal (i.e.,
normal) level of SWS in the study pop-
ulation appeared to be the most likely
factor determining whether SWS was
increased. For example, in a study of
insomniacs who had lower amounts
of SWS than did healthy people when
taking a placebo—a typical finding
in insomniacs—SWS increased when
they consumed alcohol (Roehrs et al.
1999). Conversely, alcohol did not
affect SWS in a group of age-matched
healthy control subjects.
Another population that typically
shows lower levels of SWS compared
with healthy young adults are the elderly,
but no studies have assessed alcohol’s
effects on the sleep of healthy elderly
people. In sleep deprivation studies,
however, elderly participants show
Vol. 25, No. 2, 2001 103
increases in SWS on the recovery night
after the sleep-deprivation period; pos-
sibly alcohol could similarly promote
SWS in elderly people. This finding
does not imply, however, that alcohol
should be considered a potential sleep
therapy in elderly people, because toler-
ance to the SWS enhancement devel-
ops rapidly (Prinz et al. 1980).
Several studies have assessed the
effects of alcohol administration over
several nights. Such studies clearly
demonstrated that tolerance to alcohol’s
sedative and sleep-stage effects develops
within 3 nights (Williams and Salamy
1972) and that the percentages of SWS
and REM sleep return to basal levels
after that time. Furthermore, in some
studies, the discontinuation of nightly
alcohol administration resulted in a REM
sleep rebound—that is, an increase in
REM sleep beyond basal levels (Williams
and Salamy 1972). However, not all
studies found such a rebound effect.
This variability in results may be related
to several factors specific for each study,
including the basal level of REM sleep
in the participants, the degree of alcohol-
related REM suppression, the extent of
prior tolerance to REM suppression,
and the dose and duration of alcohol
administration.
Alcohol’s Effects
on Hormone Function
The sleep-wake cycle is organized in a
circadian rhythm. To track this rhythm
in humans, researchers tend to use
measurements of the core body temper-
ature and of the secretion of the hor-
mone melatonin from the pineal gland
in the brain, both of which fluctuate
in a typical pattern throughout the day.
Accordingly, one can also use these
measurements to assess alcohol’s effects
on the sleep-wake cycle. As noted earlier,
Kleitman (1939) first reported that
alcohol administration 60 minutes
before nocturnal bedtime altered body
temperature compared with placebo
administration. Thus, alcohol adminis-
tration initially resulted in a reduction
in core temperature, followed by a
rebound increase in temperature. Such
a temperature-reducing (i.e., hypother-
mic) effect of alcohol also has been
observed in numerous other studies.
Various hormones secreted by the
pituitary gland in the brain also show
circadian variations, with secretory
peaks occurring during the usual sleep
period. Some of these hormones are
linked to sleep—if sleep is delayed,
their secretory peaks also are delayed.
Conversely, the levels of other hormones
peak at the same time every night, even
if sleep is delayed. One of the pituitary
hormones linked to sleep is growth
hormone, whose secretion typically peaks
with the onset of SWS (Takahashi et al.
1969). In an early study, administration
of 0.8 g/kg alcohol before bedtime sup-
pressed growth-hormone secretion,
despite increasing the percentage of
SWS (Prinz et al. 1980). A later study
using two different alcohol doses—0.5
and 1.0 g/kg—similarly found that
alcohol suppressed growth-hormone
secretion at a dose-related rate (Ekman
et al. 1996). Thus, alcohol appears to
affect growth-hormone secretion and
SWS levels independently (i.e., to dis-
sociate growth hormone from SWS).
This hypothesis is further supported
by the results of repeated alcohol admin-
istration in the first study (Prinz et al.
1980). In that study, the alcohol-related
suppression of growth-hormone secretion
persisted over the 3 nights of alcohol
administration, whereas tolerance devel-
oped to the alcohol-related enhancement
of SWS. The clinical implications of
alcohol’s inhibitory effects on growth
hormone and the dissociation of growth
hormone and SWS are unclear, particu-
larly with chronic and excessive alcohol
use. Unfortunately, these provocative
findings have not been pursued further.
Another pituitary hormone linked
to sleep is prolactin
4
; the hormone’s
secretion peaks 4 to 5 hours after sleep
onset (Van Cauter and Turek 1994).
To date, researchers have not determined
conclusively whether alcohol affects
prolactin release. In the study by Ekman
and colleagues (1996), alcohol did not
affect prolactin levels. However, possi-
bly even at the 1.0 g/kg alcohol dose,
4
Together with other hormones, prolactin regulates growth
and development of the mammary glands and tinitiation
and maintenance of milk production in nursing women.
alcohol levels may no longer have been
high enough 4 to 5 hours after sleep
onset to affect prolactin secretion. Prinz
and colleagues (1980) did not measure
prolactin levels in their study.
Alcohol’s Effects
on Neurochemicals
Alcohol’s effects on central nervous sys-
tem (CNS) function are mediated by
its effects on various brain chemicals
(i.e., neurotransmitters and neuromod-
ulators) that are responsible for the
transmission of nerve signals from one
nerve cell (i.e., neuron) to the next.
These neurotransmitters are released
by the signal-emitting neuron and gen-
erally exert their actions by interacting
with certain molecules (i.e., receptors)
located on the surface of the signal-
receiving neuron. Particularly at low
doses, alcohol affects CNS function pri-
marily by interfering with the normal
actions of the neurotransmitters gamma-
aminobutyric acid (GABA) and gluta-
mate, both of which also play critical
roles in wake-sleep states (Koob 1996).
GABA is the major inhibitory neu-
rotransmitter system in the CNS—that
is, its interaction with the signal-receiving
neuron dampens the ability of that
neuron to generate a new nerve signal.
Evidence from studies using various
types of experimental approaches has
indicated that alcohol at low doses
enhances GABA’s actions on the signal-
receiving neuron, thereby reducing that
neuron’s ability to generate nerve signals
even further (Mihic and Harris 1996).
This observation is significant, because
many hypnotic drugs (i.e., barbiturates,
benzodiazepines, and the newer non-
benzodiazepine GABA agonists
5
) also
act by facilitating GABA function.
Scientists have long considered GABA
to play a major role in sleep (Jones
2000). For example, GABA-releasing
neurons are present in various brain
areas that are involved in the genera-
tion of SWS, such as the brainstem
reticular activation system, thalamus,
5
Agonists are substances that mimic the actions of
another molecule. For example, GABA agonists cause the
same reactions in other neurons as does GABA.
104 Alcohol Research & Health
Sleep, Sleepiness, and Alcohol Use
hypothalamus, and basal forebrain.
Thus, facilitation of GABA-mediated
inhibition is one possible explana-
tion for alcohol’s sedative and SWS-
promoting effects.
Glutamate is the major excitatory
neurotransmitter in the CNS—that is,
the interaction of glutamate with its
receptor activates the signal-receiving
neuron to generate a new nerve signal.
6
Antagonists are substances that inhibit or interfere with
the actions of another molecule. For example, glutamate
antagonists inhibit glutamate’s interactions with its receptors.
Four types of glutamate receptors have
been identified, including the NMDA
receptor (Tabakoff and Hoffman 1996).
Anatomically, glutamate-releasing neu-
rons also are present in some of the
brain areas that promote SWS, such as
the reticular activating system of the
brainstem and the forebrain (Jones 2000).
NMDA agonists produce seizures; con-
versely, some glutamate antagonists
6
are
used as sedatives and anesthetics (Jones
2000). Thus, glutamate is an important
element in wakefulness and activation.
Numerous biochemical and electro-
physiological studies have found that
alcohol inhibits NMDA-receptor func-
tion, thereby acting as a glutamate
antagonist (e.g., Tabakoff and Hoffman
1996). Consequently, alcohol inhibition
of NMDA function may be another
mechanism through which alcohol
derives its sedative effects.
In addition to GABA and the
glutamate-NMDA system, another agent
that only recently has been considered
a candidate for mediating alcohol’s sleep
effects is adenosine. This molecule is
not a neurotransmitter itself but modu-
Samples of electrophysiological measurements of various sleep stages. The four panels represent the measurements obtained
during (A) wakefulness; (B) stage 2 nonrapid eye movement (NREM) sleep (i.e., light sleep); (C) stages 3 to 4 NREM sleep
(deep or slow-wave sleep); and (D) rapid eye movement (REM) sleep, which is associated with dreaming. For each panel, the
graphs labeled LOC and ROC represent measurements of the left and right eye movements, respectively. The graph labeled
“chin” represents a measurement of small body movements, such as of the chin muscles. The graphs labeled C3-A2 and
O2–A1 represent two electroencephalogram (EEG) readings measuring brain activity in certain brain regions. Finally, the elec-
trocardiogram (EKG) measures the heart rate. Each sleep stage is characterized by a specific pattern of those readings. For
example, during REM sleep the eyes move rapidly compared with stage 2 NREM sleep. At the same time, the EEG readings
during REM sleep exhibit a higher frequency (i.e., number of waves per second) and a lower amplitude (i.e., height of the peaks
and valleys of the waves) compared with stage 2 NREM sleep.
A. Sample electrophysiological measurement of wakefulness B. Sample electrophysiological measurement of stage 2 NREM sleep
C. Sample electrophysiological measurement of stages 3 to 4 NREM sleep
D. Sample electrophysiological measurement of REM sleep
LOC
C3-A2
02-A1
EKG
EKG
LOC
Chin
ROC
ROC
Chin
C3-A2
02-A1
Vol. 25, No. 2, 2001 105
lates signal transmission by other neu-
rotransmitters, including GABA and
glutamate. In general, adenosine inhibits
the function of glutamate in the CNS
(Dunwiddie 1996). Alcohol appears to
facilitate these inhibitory modulatory
effects of adenosine through several
mechanisms, such as enhancing the
formation of adenosine; inhibiting the
return of released adenosine into the cells,
thereby prolonging its actions; and
enhancing adenosine-receptor function
(Dunwiddie 1996). Adenosine has been
hypothesized to function as the sleep
homeostat—the system that monitors
the accumulated amount of wakeful-
ness and sleep and signals the need for
sleep (Bennington and Heller 1995).
Its levels in the brain rise during waking
and decline during SWS. Thus, alcohol
REM suppressive effects
7
(Turner et al.
2000). In sum, alcohol’s REM-suppressive
effects may occur through glutamate-
related mechanisms, whereas its sedative
effects occur through GABA-related
mechanisms.
A
lcohol can
indirectly impair
daytime alertness and
performance through
to their performance after consuming a
placebo (Yesavage and Leirer 1986).
To investigate whether alcohol-
induced sleep disruption contributed
to subsequent performance impairment,
Roehrs and colleagues (1991) adminis-
tered alcohol to healthy people before
sleep, recorded their sleep, and assessed
the participants’ alertness and perfor-
mance throughout the following day.
The alcohol doses used resulted in a
BrAC of 0.06 percent before sleep. The
study found that this dose was associ-
ated with an increase in the amount of
stage 1 sleep in the second half of the
night. The next day, the investigators
assessed alertness using the Multiple
Sleep Latency Test (MSLT), a reliable
and well-validated electrophysiological
test. Performance was evaluated with
also may promote SWS and rapid sleep
onset by facilitating adenosine function.
The neurobiological mechanism
underlying alcohol’s suppression of REM
sleep is unclear. One neurotransmitter
considered to play an important role in
REM sleep is acetylcholine (Bennington
and Heller 1995). Like other neuro-
transmitters, this molecule acts through
several types of receptors, including
nicotinic receptors and muscarinic recep-
tors. To date, only minimal evidence
suggests a substantive alcohol effect on
acetylcholine. Furthermore, the evidence
that does exist indicates that alcohol’s
effects occur through the nicotinic acetyl-
choline receptor (Collins 1996); however,
acetylcholine-mediated induction of
REM sleep occurs through muscarinic
receptors (Bennington and Heller 1995).
Thus, it appears unlikely that the alcohol-
related suppression of REM sleep is
mediated by alcohol’s effects on the
acetylcholine system.
Glutamate also is involved in the
induction of some REM sleep phenom-
ena (Bennington and Heller 1995),
and alcohol’s inhibition of glutamate
was noted earlier in this article (Tabakoff
and Hoffman 1996). However, alcohol
does not appear to exert its sedative
and REM-suppressive effects through
the same mechanism (e.g., glutamate
inhibition), because both effects can
be experimentally dissociated. For exam-
ple, in a recent report, caffeine reversed
alcohol’s sedative effects but not its
effects on sleep.
its disruptive
Relation of Nocturnal
Sleep to Daytime Alertness
As mentioned earlier, the identification
and recognition of sleep disorders have
sensitized clinical researchers to the
importance of sleep quantity and conti-
nuity for optimal daytime alertness and
performance. In healthy people, even
relatively minimal (i.e., 1 to 3 hours)
reductions in nocturnal sleep time for
a single night can reduce alertness and
performance efficiency during the fol-
lowing day. Moreover, these effects can
accumulate across nights (Roehrs et al.
2000a). Similarly, a disruption of sleep
continuity by auditory stimuli, without
reductions in overall sleep time, results
in reduced alertness and performance
efficiency in healthy people (Roehrs et
al. 2000a). This fragmentation of sleep
continuity is characterized by increased
amounts of stage 1 sleep and brief
awakenings.
Several studies have evaluated next-day
performance and alertness in healthy
people who consumed alcohol before
bedtime. In one study, young pilots drank
alcohol between 6 p.m. and 9 p.m. in
quantities sufficient to result in blood
alcohol concentrations (BACs) of 0.10
and 0.12 percent right before bedtime.
The following morning, more than 14
hours after consuming alcohol and with
BACs at 0, the performance of pilots in
a flight simulator was impaired relative
tests of auditory vigilance, in which the
participants had to respond to a certain
sound, or divided attention tasks, in
which the participants had to perform
two tasks simultaneously (Roehrs et al.
2000a). The study found that in the
alcohol-consuming participants, next-
day alertness as measured by the MSLT
was reduced and divided-attention per-
formance was impaired (Roehrs et al.
1991), demonstrating that alcohol can
indirectly impair daytime alertness and
performance through its disruptive
effects on sleep. These reductions in
alertness and performance were relatively
minor in terms of percentage of the
baseline values; in the performance of
difficult tasks (e.g., driving a car or flying
an airplane), however, even such minor
impairments might have significant
consequences.
Direct Alcohol Effects
on Daytime Alertness
Although alcohol generally is classified
as a depressant drug, in fact it has both
sedative and stimulatory effects. These
differential (i.e., biphasic) effects are
dependent on the alcohol dose con-
sumed and on the phase of the BAC
(Pohorecky 1977). Thus, stimulatory
effects are evident primarily at low-to-
7
Although unlikely at the low dose used, caffeine’s own
REM-suppressive effects may have been responsible for
the REM suppression observed.
106 Alcohol Research & Health
Sleep, Sleepiness, and Alcohol Use
moderate alcohol doses and when BACs
ascend to a peak. Conversely, alcohol’s
sedative effects occur at higher alcohol
doses and when BACs decline. Nighttime
sleep studies that demonstrated alco-
hol’s sedative effects (i.e., reduced sleep
latencies) in healthy people typically
used alcohol doses that resulted in BrACs
above 0.05 percent (Williams and Salamy
1972). Furthermore, the alcohol gener-
ally was administered 30 to 60 minutes
before sleep, thus allowing for alcohol
concentrations to peak before bedtime.
In other studies that also were conducted
during the descending BAC phase,
alcohol reduced sleep latency, as measured
by a standard MSLT, and impaired
both attention and reaction-time per-
formance in a dose-dependent manner.
These impairing effects persisted for at
least 2 hours after the alcohol had been
completely metabolized as evidenced
by BrACs of 0 (Roehrs and Roth 1998).
Only one daytime study using a
modified MSLT assessed alcohol’s sleep
effects during both the ascending and
descending phase of the BrACs. That
study found increased sleep latencies at
peak BrACs relative to placebo, consis-
tent with alcohol’s stimulatory effects
under these conditions (Papineau et
al. 1988). During the subsequent
descending phase of the BrACs, how-
ever, sleep latencies were reduced rela-
tive to placebo, confirming alcohol’s
biphasic effects.
A series of studies explored the mod-
ulation of alcohol’s daytime sedative
and performance-disrupting effects by
a person’s basal level of sleepiness (Roehrs
and Roth 1998). In these studies, the
investigators first either shortened or
extended the participants’ scheduled
nocturnal sleep time and then adminis-
tered alcohol doses of 0.4 to 0.8 g/kg
the following day. Subsequently, the
researchers assessed the participants’
levels of sleepiness or alertness as well
as psychomotor performance for approx-
imately 8 hours. The results indicated
that the level of sleepiness or alertness
at the time of alcohol administration
altered alcohol’s subsequent sedating
and performance-disrupting effects.
Thus, increased sleepiness compounded
alcohol’s effects, whereas increased
alertness diminished alcohol’s effects.
Furthermore, the investigators observed
those effects whether they compared
sleepy versus alert healthy people,
whether they studied the same person
before and after both sleep restriction
and sleep extension, or whether they
studied the same person at various times
of the day when the levels of sleepiness
are known to differ according to the
typical circadian rhythm.
Relationships Between
Nocturnal Sleep, Daytime
Alertness, and Alcohol-
Consumption History
Until now, this article has explored
alcohol’s effects on nocturnal sleep and
daytime alertness. The relationship
between sleepiness-alertness and alcohol
consumption, however, may be bidirec-
tional. Thus, some survey and laboratory
data suggest that variations in the dura-
tion of nocturnal sleep and level of day-
time sleepiness may play an important
role in modulating alcohol consumption.
For example, a British survey found a
negative correlation between sleep times
and alcohol consumption in men—
that is, shorter periods of sleep were
associated with heavier drinking (Palmer
et al. 1980). Similarly, in a U.S. study
of young adults, participants who
reported needing only 6 hours of sleep
or less had an earlier age of drinking
onset and drank more per month than
did participants who needed more sleep
(Schuckit and Bernstein 1981), leading
the investigators to hypothesize that
short sleep is associated with heavier
alcohol intake.
Laboratory studies of alcohol and
mood have identified some interesting
relations between daytime sleepiness-
alertness and drinking. In such studies,
the participants’ preference for alcohol
is studied by offering them several bev-
erage choices presented in color-coded
cups in which the participants do not
know which of the cups contain an
alcoholic beverage. After the partici-
pants have tasted each beverage, they
can choose which beverage they prefer.
Using this procedure, de Wit and col-
leagues (1987, 1989) found that mod-
erate drinkers who preferred an alcohol
dose of 0.5 g/kg, which corresponds to
approximately three drinks, in the labo-
ratory tests felt less alert at that time
than did drinkers who did not prefer
alcohol. Furthermore, participants who
preferred alcohol in those studies gen-
erally experienced alcohol as increasing
their elation and vigor, whereas partici-
pants who did not prefer alcohol gener-
ally experienced alcohol as increasing
their sleepiness.
In an alcohol challenge study, in which
healthy young men received a certain
alcohol dose, the men’s drinking histo-
ries predicted their subjective responses
to alcohol (Schuckit and Klein 1991).
Those participants with histories of
greater alcohol consumption showed
less self-rated sleepiness after the alcohol
challenge than did participants with
histories of lower alcohol consumption.
Researchers do not know whether these
individual differences in response to
alcohol reflect different physiological
states (i.e., whether people are actually
more or less sleepy) or differences in
the perception of a common physiolog-
ical state (i.e., whether all people expe-
rience the same physiological state but
differ in whether they perceive that state
as “being sleepy”). In the latter case, the
different perceptions of alcohol’s effects
may result from differential expectations
regarding alcohol’s effects.
Alcohol’s Effects on
the Sleep of Insomniacs
Approximately 10 to 15 percent of the
U.S. general population experiences
difficulties falling asleep or maintaining
sleep, or suffer from nonrestorative sleep
(i.e., sleep that does not result in a feel-
ing of being rested) (Roehrs et al. 2000b).
Moreover, 30 percent of people with
persistent insomnia in the general pop-
ulation have reported using alcohol to
help them sleep in the past year, and 67
percent of those people have reported
that alcohol was effective in inducing
sleep (Ancoli-Israel and Roth 2000).
For several reasons, studies conducted
in healthy people sleeping at their usual
bedtimes, such as the studies reviewed
in this article, do not adequately repre-
sent the hypnotic potential of alcohol
Vol. 25, No. 2, 2001 107
in people with insomnia. First, in
healthy people, sleep latency and sleep
efficiency are already optimal, and fur-
ther improvement is difficult to demon-
strate. Consequently, as previously
noted, alcohol’s effects on measures of
sleep induction and maintenance in
healthy people are minimal and incon-
sistent. Second, the doses used in sleep
studies are generally much larger (i.e.,
resulting in BrACs greater than 0.05
percent, which corresponds to more
than three drinks) than the doses that
insomniacs typically report using (i.e.,
one to two drinks). Third, the same
alcohol dose may have different effects
in healthy people and insomniacs. A
recent study compared the effects of an
alcohol dose of 0.5 g/kg on the sleep of
insomniacs and age-matched healthy
people (Roehrs et al. 1999). In the
insomniacs, but not in the healthy con-
trol subjects, this alcohol dose improved
sleep compared with a placebo. Further-
more, the sleep disruption during the
second half of the night that occurs in
healthy people after higher alcohol doses
was not observed in the insomniacs.
Specifically, alcohol consumption in the
insomniacs increased their SWS to the
levels of the age-matched control subjects.
During a later phase of the same study
(Roehrs et al. 1999), the participants
also had an opportunity to choose
between beverages presented in color-
coded cups that contained various alcohol
concentrations or a placebo. The par-
ticipants had previously experienced all
of those beverages (i.e., they had taken
them one at a time before bedtime on
different nights) and were asked to choose
the beverage that would best help them
sleep. With this approach, the insomniacs
generally chose an alcohol-containing
beverage, whereas the healthy people
chose the placebo-containing beverage.
Furthermore, the average nightly alcohol
dose self-administered by the insomni-
acs was 0.45 g/kg (up to 0.6 g/kg was
possible), which is similar to the dose
previously shown to improve the sleep
of the insomniacs and similar to the dose
that insomniacs report using at home.
The epidemiological data and labo-
ratory study findings indicating the
preference for alcohol at bedtime by
insomniacs, compared with noninsom-
niacs, generate several questions. For
example, does this preference reflect the
use of alcohol as self-medication for a
sleep problem, as a way to improve mood,
or as a sleep medication that subsequently
becomes a “mood-altering” drug? And
In healthy people,
acute high alcohol
doses disturb sleep,
whereas in insomniacs,
beneficial.
lower doses may be
if alcohol use initially is, or ultimately
becomes, “mood-altering” behavior,
what are the “mood-altering” effects for
the insomniac that reinforce alcohol
consumption? Furthermore, do insom-
niacs develop tolerance to alcohol’s
sedative effects as do other people? Do
insomniacs increase their alcohol dose
in successive nights? Does hypnotic use
at night generalize to daytime use? And
ultimately, what are the risks associated
with the use of alcohol as a hypnotic?
All these issues have yet to be addressed.
But these data again suggest that the
alcohol-sleep relation is interactive—
that is, disturbed nocturnal sleep increases
the likelihood of alcohol use, and alco-
hol has the potential to influence sleep.
Summary
Alcohol has extensive effects on sleep
and daytime sleepiness. In healthy people,
acute high alcohol doses disturb sleep,
whereas in insomniacs, lower doses
may be beneficial. Data from healthy
people suggest, however, that tolerance
to alcohol’s sedative effects probably
develops rapidly. This tolerance devel-
opment may lead to excessive hypnotic
use and, possibly, excessive daytime use
for insomniacs.
The effects of alcohol appear to be
bidirectional in that nocturnal sleep
quantity and continuity and subsequent
levels of daytime sleepiness also influence
alcohol’s sedative and performance-
impairing effects. Sleep quality and day-
time sleepiness may also relate to rates
of alcohol drinking and become a gateway
to excessive alcohol use. To investigate
these issues and identify the mechanisms
underlying the relationship between
alcohol and sleep remain important
tasks, as does documenting alcohol’s
effects on other physiological functions
during sleep. ■
References
ANCOLI-ISRAEL, S., AND ROTH, T. Characteristics
of insomnia in the United States: Results of the
1991 National Sleep Foundation Survey. I. Sleep
22:S347–S353, 2000.
B
ENNINGTON, J.H., AND HELLER, H.C. Restoration
of brain energy metabolism as the function of sleep.
Progress in Neurobiology 45:347–365, 1995.
C
OLLINS, A.C. The nicotinic cholinergic receptor as
a potential site of ethanol action. In: Dietrich, R.A.,
and Erwin, V.G., eds. Pharmacological Effects of
Ethanol on the Nervous System. New York: CRC
Press, 1996. pp. 95–116.
DE WIT, H.; UHLENHUTH, E.H.; PIEZ, J.; AND
JOHANSON, C.E. Individual differences in behav-
ioral and subjective responses to alcohol. Alcoholism:
Clinical and Experimental Research 11:52–59, 1987.
DE
WIT, H.; PIERRI, J.; AND JOHANSON, C.E.
Assessing individual differences in ETOH prefer-
ence using a cumulative dosing procedure.
Psychopharmacology 98:113–119, 1989.
D
UNWIDDIE, T.V. Acute and chronic effects of
ethanol on the brain: Interactions of ethanol with
adenosine, adenosine transporters and adenosine
receptors. In: Dietrich, R.A., and Erwin, V.G., eds.
Pharmacological Effects of Ethanol on the Nervous
System. New York: CRC Press, 1996. pp. 147–162.
E
KMAN, A.C.; VAKKURI, O.; EKMAN, M.; ET AL.
Ethanol decreases nocturnal plasma levels of thy-
rotropin and growth hormone but not those of thy-
roid hormones or prolactin in man. Journal of
Clinical Endocrinology and Metabolism 81:2627–
2632, 1996.
J
ONES, B.E. Basic mechanisms of sleep-wake states.
In: Kryger, M.H.; Roth, T.; and Dement, W.C.,
eds. Principles and Practice of Sleep Medicine. 3d ed.
Philadelphia: W.B Saunders, 2000. pp. 134–154.
K
LEITMAN, N. Sleep and Wakefulness. Chicago:
University of Chicago Press, 1939.
K
OOB, G.F. The neuropharmacology of ethanol’s
behavioral action: New data, new paradigms, new
hope. In: Dietrich, R.A., and Erwin, V.G., eds.
Pharmacological Effects of Ethanol on the Nervous
System. New York: CRC Press, 1996. pp. 1–12.
108 Alcohol Research & Health
Sleep, Sleepiness, and Alcohol Use
MIHIC, S.J., AND HARRIS, R.A. Alcohol action at
the GABA receptor/chloride channel complex. In:
Dietrich, R.A., and Erwin, V.G., eds. Pharmacological
Effects of Ethanol on the Nervous System. New York:
CRC Press, 1996. pp. 51–72.
P
ALMER, C.D.; HARRISON, G.A.; AND HIORNS,
R.W. Association between smoking and drinking
and sleep duration. Annals of Human Biology
7:103–107, 1980.
P
APINEAU, K.; ROEHRS, T.; PETRUCELLI, N.;
R
OSENTHAL, L.; AND ROTH, T. Electrophysiological
assessment (Multiple Sleep Latency Test) of the
biphasic effects of ethanol in humans. Alcoholism:
Clinical and Experimental Research 22:231–235, 1988.
P
OHORECKY, L.A. Biphasic action of ethanol.
Biobehavior Reviews 1:231–240, 1977.
P
RINZ, P.; ROEHRS, T.; VITALIANO, P.; LINNOILA,
M.;
AND WEITZMAN, E. Effect of alcohol on sleep
and nighttime plasma growth hormone and cortisol
concentrations. Journal of Clinical Endocrinology
and Metabolism 51:759–764, 1980.
R
OEHRS, T., AND ROTH, T. State-altering actions
of ethanol, caffeine, and nicotine. In: Lydic, R., and
Baghdoyan, H.A., eds. Handbook of Behavioral
State Control—Cellular and Molecular Mechanisms.
New York: CRC Press, 1998. pp. 421–432.
R
OEHRS, T.; YOON, J.; AND ROTH T. Nocturnal and
next-day effects of ethanol and basal level of sleepi-
ness. Human Psychopharmacology 6:307–311, 1991.
R
OEHRS
, T.; P
APINEAU, K.; ROSENTHAL, L.; AND
ROTH, T. Ethanol as a hypnotic in insomniacs: Self
administration and effects of sleep and mood.
Neuropsychopharmacology 20:279–286, 1999.
R
OEHRS, T.; CARSKADON, M.A.; DEMENT
, W.C.;
AND ROTH, T. Daytime sleepiness and alertness. In:
Kryger, M.H.; Roth, T.; and Dement, W.C., eds.
Principles and Practice of Sleep Medicine. 3d ed.
Philadelphia: W.B. Saunders, 2000a. pp. 43–52.
R
OEHRS, T.; ZORICK, F.; AND ROTH
, T. Transient
and short-term insomnias. In: Kryger, M.H.; Roth,
T.; and Dement, W.C., eds. Principles and Practice
of Sleep Medicine. 3d ed. Philadelphia: W.B.
Saunders, 2000b. pp. 624–632.
S
CHUCKIT, M.A., AND BERNSTEIN, L.I. Sleep time
and drinking history: A hypothesis. American
Journal of Psychiatry 138:528–530, 1981.
S
CHUCKIT, M.A., AND KLEIN, J.L. Correlations
between drinking intensity and reaction of ethanol
and diazepam in healthy young men. Neuropsycho-
pharmacology 4(3):157–163, 1991.
S
TONE, B.M. Sleep and low doses of alcohol.
Electroencephalography and Clinical Neurophysiology
48:706–709, 1980.
T
ABAKOFF, B., AND HOFFMAN, P.L. Ethanol and
glutamate receptors. In: Dietrich, R.A., and Erwin,
V.G., eds. Pharmacological Effects of Ethanol on the
Nervous System. New York: CRC Press, 1996. pp.
73–93.
T
AKAHASHI, Y.; KIPNIS, D.M.; AND DAUGHADAY,
W.H. Growth hormone secretion during sleep.
Journal of Clinical Investigation 47:2079–2090, 1969.
T
URNER, L.; ROEHRS, T.; ROSENTHAL, L.; AND
ROTH, T. Ethanol and caffeine effects on the sleep
of insomniacs. Sleep 23:S48, 2000.
V
AN CAUTER, E., AND T
UREK, F.W. Endocrine and
other biological rhythms. In: Degroot, L.J., ed.
Endocrinology. Philadelphia: W.B. Saunders, 1994.
pp. 487–548.
W
ILLIAMS, D.L.; MACLEAN, A.W.; AND CAIRNS, J.
Dose-response effects of ethanol of the sleep of
young women. Journal of Studies on Alcohol 44(3):
515–523, 1983.
W
ILLIAMS, H., AND SALAMY, A. Alcohol and sleep.
In: Kissin, B., and Begleiter, H., eds. The Biology of
Alcoholism. New York: Plenum Press, 1972. pp.
435–483.
Y
ESAVAGE, J.A.,
AND LEIRER, V.O. Hangover effects
on aircraft pilots 14 hours after alcohol ingestion: A
preliminary report. American Journal of Psychiatry
143:1546–1550, 1986.
For a free subscription to
Alcohol Alert, write to: National Institute on Alcohol Abuse and Alcoholism,
Attn.: Alcohol Alert, Publications Distribution Center, P.O. Box 10686, Rockville, MD 20849–0686.
Fax: (202) 842–0418. Alcohol Alert is also available on NIAAA’s Web site (http://www.niaaa.nih.gov).
At What Blood Alcohol Level Is a Person
Too Impaired To Drive Safely?
N
ew findings relevant to this and other questions can be found in Alcohol Alert, the quarterly
bulletin published by the National Institute on Alcohol Abuse and Alcoholism. Alcohol Alert
provides timely information on alcohol research and treatment. Each issue addresses a specific
topic in alcohol research and summarizes critical findings in a brief, four-page, easy-to-read format.
Alcohol and Transportation Safety (No. 52)—discusses the effects that even low blood
alcohol concentration has on driving skills, and reviews strategies designed to reduce
alcohol-related crashes and repeat drinking-and-driving offenses.
Vol. 25, No. 2, 2001 109
