Journal of

PHYSIOLOGICAL
ANTHROPOLOGY
and Applied Human Science

Effects of Bathing and Hot Footbath on Sleep in

Winter

Eun-Jung Sung and Yutaka Tochihara

Department of Ergonomics, Kyushu Institute of Design

 in Control were greater subsequently enables
than under the other 60% of them to sleep
conditions. Rapid eye well in winter. Upon
Abstract. The effects of movement (REM) sleep
humidity of the bedroom,
investigating the effects
daily bathing and hot decreased under
bedding, background
of increases in body
footbath (immersion of Condition B compared
noise, human voices,
temperature on sleep, it
feet in hot water) in with Condition F, and
lighting, and exposure to
has been found that
winter on the sleep stage 3 was greater
morning sunrays. In fact,
rectal temperature
behavior of nine healthy under the latter
those factors have been
increased of 1.8 (Horne
female volunteers were condition compared with
found to be primary
and Reid, 1985) and 1.4°
studied. Subjects were Control. As such, the
causes of sleep
C (Jordan et al., 1990)
assigned to three sleep subjective sleep
disturbance (Miyazawa et
markedly enhanced slow
conditions: sleep after sensations were better
al., 1988). A commonly
wave sleep (SWS).
bathing (Condition B), under the two treatment
reported sleep
These increases in rectal
sleep after hot footbath conditions. These results
disturbance in winter is
temperature are
(Condition F), and sleep suggest that both daily
due to ‘cold feet’. Yanase
considerably higher with
without either treatment bathing and hot
(1981) noted that 33% of
daily bathing, as
(Control). footbath before sleeping
female students
observed in Japan.
Polysomnograms facilitates earlier sleep
participating in her study
Ohnaka (1997) and
(consisting of electro- onset. A hot footbath is
complained of sleep
Dorsey et al. (1996)
encephalograph, especially
disturbance due to cold
have reported that
electrooculograph, and recommendable for the
sensations in the feet.
respective thermal
electromyograph) were handicapped, elderly,
This problem may be
increases of 0.8 and 0.9°
obtained, and body and disabled, who are
typical of Japanese
C can be achieved by
movements during unable to enjoy regular
people, as most houses in
daily bathing. However,
sleep were measured baths easily and safely.
Japan are not equipped
studies on the body
while monitoring both with central heating
(J Physiol temperature/sleep
the rectal and skin systems, and
Anthropol, 19 relationship have been
temperatures of (1): 21-27, temperature of bedroom
limited.
subjects. In addition, 2000) can be quite low and
Immersing feet in
subjective sleep improper in winter
hot water may impose
sensations were (Miyazawa et al., 1985;
Keywords: bathing, less of a physical burden
obtained with a hot footbath, sleep, Kanda et al., 1996). The
on the human body than
questionnaire answered rectal temperature, results of several studies
subjecting the whole
immediately by the skin temperature conducted on
body to bathing. A hot
subjects on awakening. temperature changes in a
footbath improves blood
The rectal temperature typical Japanese bedroom
circulation to eventually
increased by Introduction in winter have shown that
generate a comfortably
approximately 1.0° C room temperature can
warm body temperature
under Condition B, but Sleep is often range from 2 to 14° C
(Inami et al., 1988).
this elevation was not disturbed by a number (Ikeda et al., 1994;
According to Genta
observed under of factors, including Hasekawa et al., 1985;
(1981), immersion of
Condition F compared surrounding factors such Kanda et al., 1999). This
feet in hot water before
with Control. In contrast, as temperature and extremely low and
sleeping refreshes a
the respective increases uncomfortable thermal
whole body and may be
in the mean skin environment may
useful for facilitating
temperature of contribute to sleep
sleep. Hot footbath,
participants subjected disturbances in winter.
therefore, has frequently
to bathing and hot
been employed as a
footbath were greater As a counter
therapeutic technique in
than those of Control, measure to eliminate
the nursing field (Genta,
although these this reason for sleep
1981; Falkenbach et al.,
temperature differences disturbances in winter,
1997; Inami et al., 1988;
became negligible 2 h bathing before bedtime
after subjects went to is traditionally practiced
bed. The sleep onset in Japan. According to
latency was shortened Kanda et al. (1995),
under both conditions more than 80% of
compared with Control. subjects take baths to
Body movements during attain a higher body
the first 30 min of sleep temperature, which
22 Effects of Bathing and Hot Footbath on Sleep in
Winter
Takamata et al., 1997). level in hot water (42° C) for 30 min from 22:10. Each
Although bathing and hot footbath before subject was asked to go to sleep at 22:45 hr, and the
sleeping may be expected to reduce sleep lights were turned off at 23:00 hr to complement a
disturbances and facilitate sleeping in winter, few consistently dark environment. During sleep,
studies on the effects of bathing on sleep have physiological data were recorded. When subjects woke
been carried out. Therefore, the present study has up at 06:00 hr the next morning, they were asked to
attempted to systematically examine evaluate their sleeping conditions of the previous night
physiological effects of daily bathing (whole body) using a standardized questionnaire. The Control
and hot foot-bathing on human sleep in winter. followed the same regimen and was deprived of any
external thermal stimuli before sleeping.
Methods Condition B, subjects immersed their bodies in hot
water (40° C) up to the shoulder level for 20 min from
Subjects 22:10, while subjects under Condition F were required
Nine healthy female volunteers (8 college to soak their legs up to the knees
students, 1 researcher) with a mean age of 26.9
years (range: 21–40 years) participated in the
study. Physical characteristics including the
average (SD) height (158.4 ± 4.7 cm), average
body weight (51.4 ± 4.1 kg), and Rohrer index
(129.7 ± 13.1) were derived from the subjects.
Subjects avoided participation in the survey
during menstruation. All participants were
requested not to take any chemical stimulants
(e.g. caffeine, medicine, alcohol, etc.) and to
avoid excessive exercise or napping during the
experimental period.

Environmental conditions
Room temperature and relative humidity were
controlled at 10° C and RH50% respectively, to
simulate the typical thermal conditions of a Japanese
bedroom in winter. Participants were subjected to
each of following three conditions: immersion of a
whole body into a hot bath (Condition B), immersion
of feet up to the knees in hot water (Condition F),
and non-exposure to either (Control) before going to
bed. Subjects, clothed in standard long-sleeved
pajamas (100% cotton), slept on the mattress
(wool/polyester: 50/50) layered with a sheet (100%
cotton), and covered their bodies with a blanket
(100% acryl) and a comforter (wool/polyester:
50/50).

Procedures
The experiment extended over a 5-day period.
Each subject spent two nights adapting: one night
without electrode attachment followed by full
electrodes attachment on the next night. From the
third night, physiological indices were measured.
Each subject was allocated randomly to one of the
three conditions.
The experiment started at 19:30 hr each day.
Subjects were assigned to the respective groups on
arrival at the laboratory. An hour after dining at the
laboratory (all participants were given the same meal),
electrodes and sensors were attached to the respective
recording sites on the body, and each subject was led
into a previously prepared bedroom at 21:50. For
Measurements
Electroencephalograms (EEG; C3, C4, O1),
electrooculograms (EOG), and electromyograms
(EMG) were monitored with an EEG recorder (EEG-
5214, Nihon Kouden). Sleep polygraphs were
analyzed at 30-sec epochs according to the standard
criteria of Rechtschaffen and Kales (1968). The sleep
onset latency was defined as the time (min) from
lights-out to the starting point of continuous (more
than 3 min) stage 1 or deeper sleep, and sleep
efficiency as the ratio of total sleeping time to the
time after lights-out. Body movements during sleep
were recorded using a Static Charge Sensitive Bed
(SCSB: Biorecoy Co., Finland) placed under the
mattress (Alihanka, 1982). A probe for measuring
rectal temperature (Tre) was inserted 12 cm into the
rectum of subject. Skin temperature was measured
at the forehead, abdomen, forearm, hand, thigh,
calf, and instep. The mean skin temperature (Tsk)
was calculated by Hardy and DuBois’s (1938) seven-
point method. Tre and Tsk were recorded at 1-min
intervals using a thermistor datalogger (LT8A, Gram
Corp.).
Subjective sleep sensations were derived from the
Oguri-Shirakawa-Azumi (OSA) sleep questionnaire
(Oguri et al., 1985), which psychologically estimates the
quality of a night’s sleep according to five different
factors (sleepiness, sleep maintenance, anxiety,
synthetic sleep, and falling asleep) related to subjective
sleep feelings. A higher score on the questionnaire
indicates a better sleep sensation.

Statistical analysis
All data are expressed in the text and tables as
the mean (± SD). To analyze the data on body
temperature and body movement, a three-way
ANOVA was performed to assess the influence of
subject, condition, and time of day. For analyzing the
data on sleep stages, a two-way ANOVA was
performed to assess the influence of subject and
condition. In addition, multiple comparisons between
conditions were performed using the paired t-test.
Differences of p<0.05 were considered significant for
all statistical analyses.

Results

Sleep characteristics
Based on the results of the sleep parameters observed
 Sung, E-J et al. 23

Table 1 Mean sleep parameters under three conditions: bathing (Condition B), hot footbath
(Condition F), and non-treatment (Control) before going to bed
Condition B Condition F Control
Total time of
Wake 21.94 (19.26) 26.17 (21.36) 40.56 (26.07)
MT 0.06 ( 0.17) 0.11 ( 0.33) 0.22 ( 0.44)
Wake + MT 22.00 (19.34) 26.28 (21.62) 40.78 (26.18)
REM** 50.61 (20.68) a 72.28 (14.76) 64.72 (13.06)
Stage 1 44.72 (18.30) 43.44 (14.03) 42.94 (15.12)
Stage 2* 249.17 (34.06) a 219.50 (36.47) 219.33 (48.83)
Stage 3* 35.11 (23.97) 42.17 c 27.17 (19.70)
(23.13)
Stage 4 18.44 (27.08) 16.33 (20.82) 25.06 (32.48)
SWS (S3+S4) 53.56 (33.53) 58.50 (32.06) 52.22 (39.25)
First 210 min
Wake 14.56 (17.76) 17.33 (15.57) 33.39 (23.32)
MT 0.06 ( 0.17) 0.00 ( 0.00) 0.11 ( 0.33)
Wake + MT** 14.61 (17.81) b 17.33 (15.57) c 33.50 (23.41)
REM* 15.78 (13.18) a 29.50 ( 8.10) c 16.72 (13.00)
Stage 1 18.78 (16.00) 18.72 ( 9.61) 20.50 (12.98)
Stage 2 111.44 (22.50) 100.06 (29.24) 100.28 (32.96)
Stage 3 29.44 (23.37) 29.78 (20.69) 21.78 (19.89)
Stage 4 15.50 (23.72) 14.61 (19.98) 17.22 (19.34)
SWS (S3+S4) 44.94 (29.35) 44.39 (28.14) 39.00 (23.31)
Second 210 min
Wake 7.39 ( 8.83) 8.83 (11.28) 7.17 ( 7.88)
MT 0.00 ( 0.00) 0.11 ( 0.33) 0.11 ( 0.33)
Wake + MT 7.39 ( 8.83) 8.94 (11.47) 7.28 ( 7.98)
REM* 34.83 (10.89) b 42.78 (13.93) 48.00 (14.28)
Stage 1 25.94 (10.16) 24.72 ( 7.21) 22.44 ( 7.24)
Stage 2* 137.67 (19.34) a 119.44 (20.77) 119.06 (21.45)
Stage 3* 5.67 ( 7.89) 12.39 ( 9.43) c 5.39 ( 8.14)
Stage 4 2.94 ( 6.06) 1.72 ( 5.17) 7.83 (15.54)
SWS (S3+S4) 8.61 (11.00) 14.11 (12.75) 13.22 (22.20)
Total sleep time (min) 398.06 (19.26) 393.83 (21.36) 379.39 (26.16)
Sleep efficiency (%) 94.76 ( 4.57) 93.77 ( 5.09) 90.34 ( 6.20)
Sleep onset latency (min)* 11.22 (17.06) b 15.50 (14.37) c 29.39 (23.38)
REM latency (min) 137.44 (43.72) 99.39 (39.37) 109.61 (49.59)
First REM duration (min) 11.50 ( 7.26) 13.83 (12.96) 13.61 (13.83)
Wake after sleep onset (min) 9.61 ( 9.88) 10.67 (12.54) 11.17 (11.28)

Average (SD) min. MT, movement time; SWS, slow wave sleep. *p<0.05, **p<0.01 in
ANOVA. a: Significant difference between Conditions B and F. b: Significant difference
between Condition B and Control. c: Significant difference between Condition F and Control.
for each condition (Table To compare the
1), some sleep stages differences among the
showed significant conditions over the
(p<0.05) differences whole time- course of
corresponding to the sleep, the total
condition. Although the sleeping time was
total time of wake and divided into two blocks
movement time (MT) of 210 min each. There
during sleep increased were significant
slightly in Control differences in REM in
compared with the other both blocks under
conditions, significant three conditions.
differences were not However, Wake and MT
established between showed a significant
any two of the three (p<0.01) increase only
conditions. There was a under Control during
significant difference the first 210 min, and
(p<0.01) in the rapid significant differences
eye movement (REM) (p<0.05) were
sleep period between obtained for stages 2
Conditions B and F, with and 3 among the
the REM sleep period conditions during the
under Condition B being second 210 min.
shorter than under While the sleep onset
Condition F. However, latency (min) was
stage 2 showed a significantly (p<0.05)
significant (p<0.05) delayed under the Control
increase under compared to the other
Condition B compared conditions; sleep
to Condition F. SWS efficiency tended to be
increased more under lower for Control,
Condition F than under although this difference
Control, although a was insignificant. In
significant difference Condition B, the REM
was only established for latency was delayed,
stage 3 (p<0.05). while the first REM
duration was shortened
with no significant
differences. Wake after
24 Effects of Bathing and Hot Footbath on Sleep in
Winter
 Condition B increased
by 1.0 (± 0.44) ° C, while
Tre did not alter under
Condition F. At the time
of lights-out (23:00 hr),
Tre under Condition B
was significantly
(p<0.05) higher under
the other conditions:
Condition B (37.6 ±
0.45° C) vs Condition F
(36.9 ± 0.23° C) vs
Control (36.9 ± 0.47° C).
Tsk under the Control
was lower than under the
other conditions,
indicating significant
(p<0.05) differences until
00:30 hr. However,
negligible differences in
Fig. 1 Changes in the Tsk were found among
rectal temperature and three conditions after
mean skin temperature 01:00 hr. The rising
under three conditions:
values of Tsk under
bathing (Condition B),
Conditions B and F (*;
hot footbath (Condition
F), and no treatment registered 7.9 (± 1.21)
(Control) before sleep. and 3.5 (± 0.76) ° C,
Values are the means (± respectively. At the time
standard deviations) for of lights-out, Tsk in the
nine subjects. The Control was significantly
statistical significance
(p<0.05) lower than in
of Conditions B vs F (*;
the subjects under the
p<0.05), Condition B vs
Control (o; p<0.05), and other conditions:
Condition F vs Control Condition B (34.7 ± 1.02°
(+; p<0.05) was C) vs Condition F (33.9 ±
evaluated with t-tests. 0.55° C)
vs Control (30.5 ±
3.26° C).
sleep onset did not
show any specific
pattern. Body movements
Body movements
during sleep under each
Body temperatures
condition (Fig. 2)
Tre and Tsk under
showed differences with
three conditions (Fig. 1)
respect to subjects,
were significantly
(p<0.05) affected by the
subjects, conditions,
and time of day. Tre
under Condition B was
higher than under the
other conditions, and
the reading persisted
above 36.5° C day (p<0.05). Numerous
throughout the night.
There were significant
(p<0.05) differences in
Tre among three
conditions from 22:30 to
04:30 hr. Tre under
the first 30 min after
going to bed under
Control were
significantly (p<0.05)
more frequent than
under the other
conditions: Condition B
(10.8 ± 20.6) vs
Condition F (11.3 ± 15.0)
vs Control (33.0 ± 40.7)
sec/30 min. In addition,
there were significantly
fewer body movements
under Condition F
(p<0.05) than under
Fig. 2 Changes in body Control from 02:30 to
movement during sleep
03:00 and from 03:30 to
under three conditions
(see Fig. 1). Data are 04:00 hr.
expressed as the
average movement Subjective sleep
time (sec) per 30 min sensations
during sleep. The
statistical significance
The scores for
was verified by t-test of subjective sleep sensations
Condition B vs Control
under Control were
p<0.05) and
significantly lower than
Condition F vs and
Control (+; p<0.05). under the other conditions.
The scores for all five
factors of OSA under the
Control were lower than
under Conditions B and F
(Fig. 3). In particular, the
scores for factors related to
falling asleep under the
Control were significantly
lower
Fig. 3 Scales of the five
factors (F1-F5) of OSA
involving sleepiness
in the morning (F1),
sleep maintenance
(F2), anxiety (F3),
synthetic sleep (F4),
and falling asleep
(F5).
conditions, and time of
body movements during
the first 90 min after
going to bed were
observed under Control
compared to the other
conditions. In particular,
body movements during
 Sung, E-J et al. 25
Table 2 Correlation coefficients between body temperatures upon going

to bed (22:45 hr) and sleep parameters

Tsk (° C) Tre (° C)
Sleep onset latency (min) － 0.6220** － 0.0410
Sleep efficiency (%) 0.4769* 0.0383
Wake+MT (%) － 0.4751* － 0.0419
Wake after sleep onset (min) 0.0781 － 0.0339
SWS latency (min) 0.0129 0.0868
REM latency (min) 0.2288 0.4642*
Slow wave sleep (%) － 0.2501 － 0.1287
Stage (%) 0.1846 － 0.2385
3
Stage in the first half of sleep (%) 0.2205 － 0.1999
3
REM (%) 0.0395 － 0.4729*

Tsk, Mean skin temperature; Tre, Rectal temperature. *p<0.05; **p<0.01.

 heating on the body bathed subjects during
during sleep has been the first 3 h of sleep
extensively researched. (Kanda et al., 1999), and
than those under the Horne and Reid (1985) Sleep onset latency the results of the
other conditions. have reported that an was significantly present show that
Subjects under Control increase of up to 1.8° C shortened under bathing and hot
complained of in the average Tre of Conditions B and F footbath are effective in
difficulties in falling subjects can be compared to the promoting rapid sleep
asleep due to cold achieved when they are Control, and the sleep onset. Furthermore,
feet. Based on the immersed in hot water efficiency tended to results similarly
subjective sleep index maintained at 41° C for increase under the favorable in terms of
derived from OSA, three consecutive cycles experimental conditions subjective sleep
sleep in the two of 30 min each. In (Table 1). A negative sensation were also
experimental addition, Jordan et al. correlation between obtained. Based on all
conditions was (1990) have found that sleep onset latency and factors of sleep
significantly (p<0.05) Tre of subjects after Tsk at the time of going sensation, our findings
better than under bathing in water to bed coupled with a show that sleep after
Control: Condition B maintained between 42° positive correlation bathing or hot footbath
(871.7 ± 120.3) vs C and 43° C increased between sleep efficiency is generally evaluated
Condition F (864.0 ± by up to 1.6° C when and Tsk suggested that more positively than
176.2) vs Control they are subjected to a higher Tsk at the time that occurring under
(600.0 ± 200.1). two 40-min bathing of going to bed Control (Fig. 3).
sessions. In our study, facilitates falling asleep Some studies have
Correlation between Tre increased by in winter. However, the reported a SWS increase
body temperatures approximately 1.0° C correlations between Tre due to body heating
and sleep parameters after bathing. This at the time of going to (Horne and Moore, 1985;
Correlation increase correlated well bed and sleep Horne and Reid, 1985;
coefficients between the with values (0.8–0.9° C) parameters were not Jordan et al., 1990). In
body temperature upon observed in previous significant. Furthermore, this study, however, there
going to bed (at 22:45 studies (Ohnaka, 1997; Tsk under Control was 2° were no significant
hr) and the sleep Dorsey et al., 1996). C lower than under the differences in SWS among
parameters are shown However, Tre after other conditions at three conditions, despite
in Table 2. These results immersing feet in hot lights-out (23:00 hr, Fig. a significant difference
illustrate negative water was hardly 1). Therefore, it was being observed between
correlations between affected at all. deduced that an Condition F and Control
the sleep onset latency elevation in Tre is not only during stage 3 (Table
and Tsk as well as always necessary to 1). Extended sleeping
wake+MT and Tsk (r= － shorten the sleep onset time was primarily
latency. observed in stage 2 under
0.62, p<0.01; r= －
0.48, p<0.05), thus Several studies Condition B. In
have reported that body comparison with other
establishing a positive
movement is a good studies on body heating,
correlation between
index for evaluating the SWS duration (52–59
sleep efficiency and Tsk
sleep (Kanda et al., min) in
(r=0.48, p<0.05).
1999; Ohnaka et al.,
Furthermore, a
1995; Ohnaka et al.,
significant correlation
1988). During the first
between REM sleep and
30 min of sleep under
Tre was also observed
Control, a higher
(r=0.46, p<0.05 in REM
frequency of body
latency; r= － 0.47, movement was
p<0.05 in REM period). measured (Fig. 2), which
also suggests that
Discussion bathing or hot
footbathing facilitates
The primary sleep onset in winter.
purpose of this study Body movements have
was to investigate the been observed to be
effects of daily bathing more frequent in non-
in hot water on sleep. bathed subjects than in
The effect of passive