Content uploaded by Ahmed S. Bahammam
Author content
All content in this area was uploaded by Ahmed S. Bahammam on Oct 12, 2017
Content may be subject to copyright.
Sleep and Biological Rhythms 2004; 2: 135– 143
© 2004 Japanese Society of Sleep Research 135
Blackwell Science, LtdOxford, UKSBRSleep and Biological Rhythms 2004; 2: 000–0001446-92352004 Japanese Society of Sleep Research22135143Original ArticleRamadan Fasting and SleepA BaHammam
Correspondence: Dr Ahmed BaHammam, FRCP, FCCP,
Sleep Disorders Centre, Respiratory Unit, Department of
Medicine 38, College of Medicine, King Saud University,
Box 2925, Riyadh, Saudi Arabia. Email:
ashammam2@yahoo.com
Accepted for publication 20 April 2004
ORIGINAL ARTICLE
Effect of fasting during Ramadan on sleep architecture,
daytime sleepiness and sleep pattern
Ahmed BAHAMMAM
Sleep Disorders Center, Respiratory Unit, Department of Medicine, College of Medicine, King Saud University, Riyadh,
Saudi Arabia
Abstract
Fasting during Ramadan is distinct from regular voluntary or experimental fasting. This project was
conducted to objectively assess the effect of Ramadan fasting on sleep architecture, daytime sleep-
iness and the circadian cycle of melatonin level. Eight healthy volunteers reported to the Sleep Dis-
orders Center on four occasions for polysomnography and multiple sleep latency tests: 1) an initial
visit for adaptation; 2) 2 weeks before Ramadan (BL); and 3,4) during the first and third weeks of
Ramadan (R1, R3). Salivary melatonin level was measured using radioimmunoassay. Sleep latency at
night was significantly shorter and the amount of rapid eye movement sleep was significantly less,
at R3 compared to BL. There was no difference in multiple sleep latency test data between BL and
Ramadan. Although melatonin level kept the same circadian pattern at BL, R1 and R3, it had a flatter
slope and a significantly lower peak at midnight (00:00) at R1 and R3. This study showed a signif-
icant reduction in sleep latency and rapid eye movement sleep during the third week of Ramadan
fasting. Otherwise, there was no significant effect of Ramadan on sleep architecture and assessment
revealed no increase in daytime sleepiness. Although melatonin level had the same circadian pattern
during Ramadan, the level of the hormone dropped significantly from baseline.
Keywords: Ramadan, fasting, sleep, melatonin.
INTRODUCTION
Fasting during the month of Ramadan is prescribed for
Muslims as one of the five major pillars of Islam. For
Ramadan, Muslims abstain from food, drink and smok-
ing during the daytime; i.e. from dawn to sunset. The
pattern of meals changes as well, with two to three main
meals usually taken: breakfast at sunset, dinner after
night prayer (Esha), and another meal before sunrise
(Suhur). Over one billion Muslims fast during this
month every year.
During Ramadan, physiological changes are expected
to result from both long-term dietary restriction and
partial sleep loss. Islamic fasting is distinct from regular
voluntary or experimental fasting by the limited dura-
tion of the fasting period within the 24-h cycle; by absti-
nence of the performer of the fast from drinking and
smoking; by the change in the usual circadian pattern of
eating which causes caloric intake increases at night;
and by the month-long duration of the practice which
may allow adaptation to the new regimen. Therefore,
one may assume that physiological changes occurring
during Islamic fasting may be different from those dur-
ing an experimental fast.1
Moreover, Ramadan is a lunar month, meaning that
its position within the calendar year changes over time.
The Islamic (Hijra) year, containing 354 days, is 11 days
shorter than the Gregorian year.
In a previous work using the Epworth Sleepiness
Scale (ESS) with a group of medical students, we found
a significant increase in daytime sleepiness during
A BaHammam
136 Sleep and Biological Rhythms 2004; 2: 135 –143
Ramadan even though the participants’ total sleep time
during Ramadan did not differ from baseline.2
The fact that meals for Ramadan are taken exclusively
at night has been proposed as a mechanism possibly
affecting circadian rhythm during the fast.3 Recently,
Rocky et al. reported increased objective daytime sleep-
iness during Ramadan fasting using the Multiple Sleep
Latency Test (MSLT) associated with decreased total
sleep time (TST), increased sleep latency and changes in
diurnal body temperature.4 However, in the protocol of
the above paper, dinner (the main meal) was given
shortly (1 h) before bedtime, which might have affected
nocturnal sleep. In turn, reduction in TST might have
induced the increased daytime sleepiness. Moreover,
during MSLT, the use of a portable, at-home polysom-
nography-recording device in the above project forced
the operator to program the computer to end the test
20 min after the beginning of recording regardless of
sleep onset. Therefore the participants did not sleep
long enough to progress to rapid eye movement (REM)
sleep, which made comparison of sleep stages between
naps before and during Ramadan less accurate.
Given this proposed alteration in the circadian pat-
tern of body temperature, we might also expect to find
a change in the circadian secretion of melatonin. Mela-
tonin is considered to be one of the best markers for cir-
cadian rhythm disruption.5 The circadian rhythm of
melatonin secretion has not yet been investigated sys-
tematically under controlled conditions at the beginning
and end of Ramadan to evaluate possible changes over
time and whether adaptation occurs by the end of the
month. Therefore, the present project was designed to
examine the effect of Islamic fasting on sleep architec-
ture, daytime sleepiness and the circadian pattern of
melatonin level by using polysomnography (PSG) and
the MSLT, along with other laboratory and subjective
measures, during the first and third weeks of Ramadan
(R1 and R3). In contrast to Rocky’s papers, the dinner
meal in the present study was given 3 h before bedtime
to control the possible effect of late meals on nocturnal
sleep and circadian rhythm.
MATERIALS AND METHODS
This observational study with repeated measures in a
non-random sample of volunteers was approved by the
Ethics Committee of the College of Medicine at King
Saud University. The study was conducted during the
month preceding Ramadan (Shaban) and the month of
Ramadan in the year 1423 Hijra (between 24 October
and 29 November 2002). During that period, fasting
for Ramadan commenced at around 04:45–05:00 and
breakfast was eaten at around 17:00–17:15.
Study group
Eight healthy, Muslim volunteers who were not taking
any regular medications and who do not drink alcohol
participated in the study after giving their informed con-
sent. Mean age was 31.8 ± 2 years and body mass index
(BMI) was 25 ± 2.2. Additional exclusion criteria were
sleep complaints, smoking, and/or addiction to caffein-
ated beverages. Before enrolment in the study, each vol-
unteer received a medical check-up and a data form
including demographic and sleep habits information
was completed. All selected participants had fixed day-
time working hours and a regular sleep-wake schedule
during weekdays; they also had the same type of work,
tasks and working hours before and during Ramadan.
Study protocol
Participants were asked to complete a daily sleep diary
when not in the Sleep Disorders Center (SDC). TST, nap
duration, wake-up time and bedtime for study days
(baseline and Ramadan) not in the SDC were calculated
on a daily basis and the means were used in our analysis.
Subjective sleepiness was assessed using the Epworth
Sleepiness Scale (ESS).6
Participants reported to the SDC at King Khalid Uni-
versity Hospital (KKUH) at 16:00 on four occasions,
spending approximately 24 h in the SDC on each
occasion:
1Initial visit for medical check-up and adaptation to
the environment.
2Baseline, or BL for PSG and an MSLT the next day.
3During R1 for PSG and MSLT.
4During R3 for PSG and MSLT.
When in the SDC, participants took three meals per
day with fixed caloric values and fixed proportions of
carbohydrate, fat and protein based on their ideal body
weights at the beginning of the study. Food items con-
taining large amounts of 5-hyroxytryptamine were
avoided on the study nights and days. At BL, dinner was
served at 20:00, breakfast at 07:45 and lunch at 12:00.
During Ramadan, breakfast was served at sunset around
17:00–17:15, dinner at 21:00 and a predawn meal
(Suhur) at 04:30.
Body weight and body fat were measured for each
participant at every SDC visit. Serum glucose level was
estimated at 15:00 on study days.
Ramadan Fasting and Sleep
Sleep and Biological Rhythms 2004; 2: 135 –143 137
Body fat measurement
FUTREX-5000 Analyzers (FUTREX, Gaithersburg, MD,
USA) were used to determine body fat composition. The
FUTREX-5000 emits near-infrared light into the body at
very precise frequencies (938 nm and 948 nm) at which
body fat absorbs the light and lean body mass reflects
the light. The validity of this method has been previ-
ously documented.7 The measurement was taken at the
midpoint of each participant’s dominant bicep.
Polysomnography
Overnight sleep studies with PSG were performed in
the SDC. Alice 4 (Respironics Inc., Murrysville, PA,
USA) diagnostic equipment was used for data acquisi-
tion and the data were then downloaded to an IBM PC.
A trained technician connected the participants to the
monitor (as per the standard hook-up process) and
stayed in the SDC throughout the study, monitoring the
recording in the control room. Before Ramadan, PSG
recording started at midnight (00:00) and ended at
07:30. During Ramadan, PSG recording started at
00:00 and continued until the participants were woken
up at 04:30 for Suhur; PSG was then resumed at 05:00–
05:30 and participants were allowed to sleep until
08:00. Participants were asked to avoid any napping
during the SDC days. During PSG recording, light
intensity was maintained <20 lux.
During PSG, the following parameters were moni-
tored: brain activity by four electroencephalographic
(EEG) placements (C1-A4, C2-A3, O1-A4 and O2-A3);
muscle tone and leg movements by chin and leg elec-
tromyography (EMG); eye movements by electro-
oculography (EOG); heart rate by electrocardiography
(EKG); oxygen saturation by finger pulse oximeter;
chest and abdominal wall movements by thoracic and
abdominal belts; air flow by thermistor; sleep position
by position sensor; and snoring by microphone.
Analysis and scoring of
polysomnography data
Page-by-page analysis and scoring of the electronic raw
data was done manually to determine lights out, lights
on, total time in bed (TIB), TST and sleep latency. Scor-
ing included the sleep stages and their proportion to
TST, slow wave sleep latency and duration, REM latency
and duration, number of sleep cycles, stage shifts (total
number of changes in sleep state from lights out to lights
on) and arousals according to established criteria.8,9 All
sleep study data were analyzed by the author. Reports
were generated using Alice 4 software.
Multiple Sleep Latency Test
A standard MSLT was performed in accordance with the
American Academy of Sleep Medicine guidelines.10
Beginning 2 h after participants awoke in the SDC, four
tests (naps) were performed 2 h apart, using a routine
PSG hook-up but omitting the leg EMG and cardiores-
piratory parameters. After awakening, participants were
asked to get out of bed, get dressed in street clothes,
avoid sleeping between naps and avoid any vigorous
activity 15 min before each nap. At each nap, partici-
pants were asked to remove their shoes, loosen their
clothing and lie in bed for 5 min before lights out. If
sleep occurred, the nap was continued for 15 min after
sleep onset. Sleep onset was determined by the first
epoch of any stage of sleep, including stage 1.10 If no
sleep occurred, the nap was terminated after 20 min The
technician performed calibrations prior to each nap.
Analysis and scoring of multiple sleep
latency test data
Manual analysis of the MSLT data included sleep latency
from lights out to first epoch of sleep, amount of each
sleep stage, mean latencies to sleep for all naps, wake
efficiency (WE; WE = 100 – sleep efficiency), sleep
onset (SO) frequency and number of sleep-onset REM
(SOREM) periods. The absence of sleep at any nap
opportunity was recorded as a sleep latency of 20 min
and included in the overall analysis of mean latency.
Delta or slow wave activity (SWA) during non-REM
(NREM) sleep is considered to be a classic marker of the
homeostatic process; on the other hand, high levels of
EEG alpha activity during wakefulness and NREM sleep
may be indicative of high levels of alertness.11 Therefore,
SWA (0.5–5 Hz) during NREM sleep and alpha activity
(8.5–12.5 Hz) during wakefulness and NREM sleep
were analyzed for each nap, using a Fast Fourier
Transformation (FFT) algorithm. Activity in the spindle
range (12–15 Hz) was also analyzed during NREM
sleep.
Melatonin level
Melatonin level was estimated three times during each
24-h SDC session, using saliva samples collected on the
following schedule: morning at 08:00, afternoon at
04:00 and midnight at 00:00.
A BaHammam
138 Sleep and Biological Rhythms 2004; 2: 135 –143
Each sample consisted of 5 mL of saliva collected
according to a special technique. Participants were
asked to rinse their mouths with water before each col-
lection and to avoid coughing or throat clearing into the
collection tube. Saliva samples were deposited directly
into Petri dishes by each participant and subsequently
stored on ice until transport and storage at -20∞C. (We
also ensured that on the days of the tests, the partici-
pants avoided caffeine and did not consume substances
containing melatonin or melatonin precursors, e.g.
banana, kiwifruit, tomato, artificial colorants and nuts).
Light exposure while in the SDC was the same before
and during Ramadan. Samples collection at night was
done under dim light (<20 lux) conditions.
Melatonin levels were determined by means of a
highly sensitive radioimmunoassay (RIA) kit (IBL-
Hamburg, Germany) using 125I-melatonin (4 mCi) anti-
serum. Assay sensitivity was 2.5 pg/mL. The intra- and
interassay coefficients of variation were 8% to 8.1% and
14.8% to 15%, respectively. Results were expressed as
the concentration of melatonin in pg/mL. The scientific
literature has demonstrated excellent circadian plotting
using salivary melatonin analysis.12,13
Body temperature
Oral temperature was measured using a high-precision
medical thermometer on study days at 08:00, 16:00 and
00:00.
Statistical analysis
Comparisons between BL, R1 and R3 were performed
using one-way repeated-measures ANOVA for continu-
ous variables. The chi-square test was used for discrete
variables. Results were considered statistically signifi-
cant at the P = 0.05 level. Data is expressed as mean ±
standard deviation (SD).
RESULTS
Sleep schedule at home
Based on the clinical data collected from the completed
sleep diaries, we found that, compared to BL, bedtime
was delayed at R1 by about 1 h and 18 min and by
about 1 h and 36 min at R3 (Table 1). On the other
hand, wake-up time did not change much, which
resulted in a significant reduction in TST at night during
Ramadan: TST at BL was 6.8 ± 1.1 h, versus 4.7 ± 1.7 h
at R1 and 4.8 ± 1.8 h at R3 (P < 0.05). However, nap-
ping during the daytime compensated for the deficit in
TST at night. TST + nap was not significantly different at
BL, R1 and R3. ESS scores were not different at BL, R1
and R3.
Body fat and serum glucose
The percentage of body fat increased over time during
Ramadan, from 20.6 ± 4.0% at BL, to 25.2 ± 4.0% at R1
and 26.2 ± 6.0% at R3 (P = 0.038 for BL versus R3).
Serum glucose measurements at 15:00 showed no dif-
ference between BL, R1 and R3: 6.2 ± 0.4, 5.5 ± 0.5 and
5.7 ± 0.5 mmol/L, respectively.
Polysomnography findings
Table 2 summarizes the PSG findings. There was no sig-
nificant difference between TIB and TST at BL, R1 and
R3. Sleep latency (SL) was shorter at R3 compared to BL
(18.5 ± 10.0 min versus 35.0 ± 12.0; p = 0.05). There
was no significant difference in REM latency at R1
(106 ± 42 min) and R3 (84.1 ± 27.0) compared to BL
(71.6 ± 19.0). Slow wave sleep latency did not differ sig-
nificantly between BL, R1 and R3; nor did arousal index
or stage shifts.
The proportion of different NREM sleep stages in rela-
tion to TST did not change significantly in Ramadan
Table 1 Sleep schedules when sleeping at home, Epworth Sleepiness Scale (ESS) scores and the percentage of total body fat
before Ramadan (BL) and during the first week (R1) and third week (R3) of Ramadan
BL R1 R3 P-value
Bedtime 23.7 ± 1.3 1 ± 2.3 1.3 ± 2.8 NS
Wake-up time 6.6 ± 1.0 5.7 ± 2.0 6.7 ± 2.0 NS
ESS 3.0 ± 3.0 4.2 ± 2.9 3.6 ± 1.2 NS
TST 6.8 ± 1.1 4.7 ± 1.7* 4.8 ± 1.8* <0.05
TST + nap 8.4 ± 1.4 7.1 ± 1.8 8.2 ± 1.7 NS
Body fat percentage 20.6 ± 4.0% 25.2 ± 4.0 26.2 ± 6.0* 0.038
*The difference is statistically significant compared to baseline (BL).
Ramadan Fasting and Sleep
Sleep and Biological Rhythms 2004; 2: 135 –143 139
compared to BL. However, total NREM sleep in propor-
tion to TST was significantly greater at R3 compared to
BL (89.8 ± 1.3% versus 75.9 ± 3.1%; P < 0.01). On the
other hand, REM sleep was greater at BL compared
to R3 (24.1 ± 3.9% versus 10.2 ± 5.0%; P < 0.01).
Although the average number of REM cycles did not
change, the duration of REM in every cycle was different
between BL, R1 and R3 (Table 2). REM duration was
significantly less for the first cycle at R1 compared to BL
and for the second and third cycles at R3 compared to
BL.
No difference in cardiorespiratory parameters was
found between BL and Ramadan.
MSLT findings
Table 3 summarizes the MSLT findings. There was no
significant difference in SL of individual naps or their
mean between BL, R1 and R3. The time of day of the
nap had no effect on SL before Ramadan (BL) or at R3.
At R1, participants were most sleepy at the 12:00 nap
(SL, 5.5 ± 5.0 min, P < 0.05).
Further examination of wake efficiency (WE) and SO
frequency revealed no significant difference between BL,
R1 and R3 in either the mean or the individual naps.
Analysis of sleep stages during the MSLT showed no
significant difference in the mean of all naps between BL
and Ramadan. However, analysis of the individual naps
revealed a significant reduction in stage 1 at the 16:00
nap at BL compared to R1 and R3. Stage 2 was also sig-
nificantly less at the 10:00 nap at R1 compared to BL.
None of the participants developed deep sleep during
the naps.
The frequency of SOREM showed no difference
between BL, R1 and R3. Analysis of the EEG absolute
powers in the alpha (8.5–12 Hz) frequency ranges was
performed for waking epochs and NREM sleep and the
delta (0.5–5 Hz) and spindle activity for NREM sleep, of
the daytime naps. Spectral analysis of EEG activity
revealed no difference between BL and Ramadan.
Table 2 Polysomnographic characteristics before Ramadan (BL) and during the first week (R1) and third week (R3) of
Ramadan. NS: non-significant, TIB: time in bed, TST: total sleep time, SPT: sleep period time, REM 1, 2 and 3: first, second
and third REM cycle, PLM: periodic leg movements index, RR: respiratory rate, HR: heart rate
BL R1 R3 P-value
TIB 440.2 ± 9.3 472.6 ± 10.9 476.3 ± 4.8 NS
TST 361.7 ± 39 379.0 ± 36.0 378.0 ± 35.0 NS
Sleep latency 35.0 ± 12.0 30.4 ± 17 18.5 ± 10* 0.05
TST/TIB 82.0 ± 9.0 80.0 ± 6.8 79.4 ± 8.0 NS
TST/SPT 92.0 ± 5.0 85.6 ± 4.0 82.6 ± 5.0* 0.05
REM onset (minutes) 71.6 ± 19.0 106.0 ± 42.0 84.1 ± 27.0 NS
SWS latency (minutes) 36.3 ± 14.0 32.6 ± 16.0 52.1 ± 32.0 NS
Stage 1% (TST) 10.9 ± 6.6 11.0 ± 4.6 20.6 ± 11.2 NS
Stage 2% (TST) 55.6 ± 13.0 57.0 ± 9.7 54.3 ± 5.6 NS
Deep sleep percentage (TST) 8.4 ± 5.0 8.6 ± 5.0 7.4 ± 4.7 NS
REM percentage (TST) 24.1 ± 3.9 21.5 ± 4.2 10.2 ± 5.0* 0.003
NREM percentage (TST) 75.9 ± 3.1 76.7 ± 3.9 89.8 ± 1.3*†0.008
REM cycles 3.6 ± 0.9 3.4 ± 0.9 3.2 ± 1.4 NS
REM1 (minutes) 15.0 ± 9.0 8.0 ± 8.0* 19.9 ± 11.0 0.05
REM2 (minutes) 32.6 ± 16.4 20.4 ± 7.5 10.7 ± 8.2* 0.01
REM3 (minutes) 40.9 ± 13.3 52.4 ± 9.7 32.9 ± 16.6†0.01
Arousal index 19.9 ± 6.8 18.1 ± 10.9 19.0 ± 11.0 NS
Stage shifts 94.0 ± 23.5 95.0 ± 23.8 101.8 ± 22.0 NS
PLM 17.6 ± 10.0 14.0 ± 9.7 18.6 ± 17.7 NS
RR REM 13.9 ± 1.4 15.3 ± 0.9 15.2 ± 1.2 NS
RR NREM 12.9 ± 1.1 13.9 ± 1.1 13.9 ± 1.4 NS
HR REM 66.6 ± 6.9 70.2 ± 7.2 74.8 ± 10.8 NS
HR NREM 65.4 ± 6.8 71.1 ± 8.8 70.6 ± 6.3 NS
*The difference is statistically significant compared to baseline (BL).
†The difference between R1 and R3 is statistically significant.
A BaHammam
140 Sleep and Biological Rhythms 2004; 2: 135 –143
Melatonin level and body temperature
Although melatonin level kept the same circadian pat-
tern for BL, R1 and R3, it had a flatter slope and a sig-
nificantly lower peak at 00:00 at R1 and R3 (BL,
18.1 ± 5.5 pg/mL; R1, 5.9 ± 8.0 pg/mL and R3, 4.1 ±
7.0 pg/mL; P = 0.02) (Fig. 1). Melatonin level at 16:00
was also significantly lower at R1 (0.14 ± 0.1 pg/mL)
and R3 (0.21 ± 0.1) compared to BL (0.62 ± 0.37;
P < 0.05). At 08:00 melatonin level was not significantly
different between BL (2.01 ± 1 pg/mL), R1 (1.2 ±
1.1 pg/mL) and R3 (3.9 ± 2.7 pg/mL).
Oral temperature measurements revealed no circa-
dian difference in body temperature between BL, R1 and
R3.
DISCUSSION
This study was conducted in volunteers who kept sim-
ilar schedules of working hours and wake-up times
before and during Ramadan in order to assess the effects
of intermittent Islamic fasting on sleep. The fact that this
project had to take place within a limited time (the first
and third weeks of Ramadan only) put some constraints
on the number of recruited volunteers. In the future, a
multicenter study (within one country, to control for
cultural factors) could involve a bigger sample size.
In the present study, body fat increased significantly
by the end of Ramadan. This finding concurs with those
of Frost et al.,14 who reported a significant increase in
caloric, fat, carbohydrate and protein intake in a sample
of 15 young Saudis, as well as a significant increase in
body weight, despite a significant reduction in meal fre-
quency.
Our results also showed a significant reduction in SL
at R3. This contrasts with the findings of Rocky et al.4,15
who reported an increase in SL during Ramadan. How-
ever, Michalsen et al.16 reported no change in SL during
short-term modified fasting. A possible explanation for
the difference between our findings and Rocky et al.4,15
is the time difference between dinnertime and bedtime.
In Rocky’s work, the difference between dinnertime and
bedtime was 1 h (dinner was served at 22:30 and PSG
recording started at 23:30). On the other hand, the dif-
ference in the present study was 3 h (dinner was served
at 21:00 and PSG recording started at 00:00). It is quite
possible that late dinners affect nocturnal sleep. There-
fore, the relationship between dinnertime during
Ramadan and nocturnal sleep should be considered in
future research.
In the present study, REM sleep was reduced at R3, in
accordance with the findings of Rocky et al.15 during
Ramadan. A shift in cortisol and insulin rhythms has
Table 3 MSLT characteristics before Ramadan (BL) and during the first week (R1) and third week (R3) of Ramadan. NS: non-
significant, SOREM: sleep onset REM
BL R1 R3 P-value
Mean sleep latency (minutes) 10.0 ± 4.5 8.6 ± 6 8.7 ± 5.5 NS
Sleep latency Nap 1 (minutes) 10.0 ± 4.9 11.4 ± 7.8 8.3 ± 7.4 NS
Sleep latency Nap 2 (minutes) 7.5 ± 2.7 5.5 ± 5.1 10.1 ± 7.6 NS
Sleep latency Nap 3 (minutes) 11.0 ± 5.1 8.7 ± 8.5 9.0 ± 8.0 NS
Sleep latency Nap 4 (minutes) 11.7 ± 8.2 8.8 ± 7.0 7.2 ± 4.9 NS
SOREM frequency Nap 1 0.4 ± 0.5 0.4 ± 0.5 0 NS
SOREM frequency Nap 2 0.4 ± 0.5 0.4 ± 0.5 0.4 ± 0.5 NS
SOREM frequency Nap 3 0.2 ± 0.4 0 0.2 ± 0.4 NS
SOREM frequency Nap 4 0.4 ± 0.5 0.2 ± 0.4 0.2 ± 0.4 NS
SOREM (mean) 0.35 ± 0.5 0.25 ± 0.4 0.25 ± 0.4 NS
Figure 1 Circadian pattern of salivary melatonin level dur-
ing baseline (BL), first week of Ramadan (R1) and third week
of Ramadan (R3). a and b indicate significant difference
(P < 0.05; a = BL versus R1, b = BL versus R3).
Melatonin level
20
18
16
14
12
10
8
6
4
2
0
pg per ml
12 midnight 8AM 4PM
BL
R1
R3
a, b
a, b
Ramadan Fasting and Sleep
Sleep and Biological Rhythms 2004; 2: 135 –143 141
been reported during Ramadan, with an increased level
at night. 17–20 It has been suggested that the nocturnal
rise of these hormones results in a reduction in REM
sleep,21–23 which may partially explain the reported
reduction in REM sleep during Ramadan. In a recent
study measuring rectal temperature continuously via
a thermistor probe, the reduction in REM during
Ramadan was attributed to an observed increase in noc-
turnal body temperature.15 Another possible explana-
tion for the reduction in REM sleep during Ramadan is
the interruption of sleep in the early morning (predawn)
for Suhur, which usually includes a larger amount of
REM sleep. We tried to compensate for that time by
delaying wake up time by 30 min (TST was not different
between BL, R1 and R3).
Michalsen et al.16 reported a significant reduction of
periodic leg movement (PLM) with fasting of six of their
subjects. We could not find any difference in PLM before
and during Ramadan; however, the mean PLM in our
participants was not high compared to the mean PLM in
those of Michalsen et al.16
Apart from the above results, we found that sleep
architecture was within normal parameters and revealed
no difference between baseline and Ramadan. This is
consistent with other reported findings of improved
sleep quality during experimental short-term fasting.16
We could not find any difference in daytime sleepi-
ness, either subjectively using the ESS or objectively
using the MSLT. This contrasts with our previous
report,2 which showed increased ESS scores during
Ramadan in medical students. This difference may be
attributed to the different groups studied. The previous
report was in a group of medical students who may have
irregular sleep habits or a shortening of mean sleep
length due to lifestyle constraints. MSLT data were ana-
lyzed for SL, SO frequency and WE. WE has been
reported by some to be a more useful MSLT measure,24
whereas others have suggested that daytime sleepiness
seems to be better appreciated using SO frequency.25
Using all of these methods in the present study, we
could not find any difference in daytime sleepiness
between baseline and Ramadan.
On the other hand, researchers have long recognized
that fasting alters the sleep-wakefulness pattern. Inves-
tigators have found that food deprivation increases wak-
ing and markedly reduces REM sleep.26–28 Interestingly,
these changes occur only during the normal circadian
feeding period of the respective species.29 However, the
magnitudes of fasting-induced changes are highly
dependent on prior nutritional state and energy
reserves. Highly significant changes in waking and
reduction in NREM and REM sleep time have been
reported in fasted lean rats with minimal changes in
fasted obese rats.27 Fasting in such experiments is usu-
ally more prolonged than fasting during Ramadan, so
we do not know if the above can be applied to Islamic
intermittent fasting.
Melatonin is considered to be the best marker for cir-
cadian rhythm.5,30 Individual melatonin profiles are
highly reproducible and are less subject to masking fac-
tors than are other rhythm markers such as core tem-
perature and cortisol.31 Although there were significant
drops in melatonin levels at 00:00 and 16:00 during
Ramadan, melatonin profiles kept the same trend dur-
ing Ramadan, but with a flatter slope. As we only mea-
sured melatonin three times per 24 h, with the last
measurement at midnight, a peak melatonin value
delayed until after 00:00 would not have been mea-
sured. Our findings concur with those of Bogdan
et al.,31 who reported a flatter slope and a smaller night
peak of melatonin during Ramadan. Future studies
should check melatonin level more frequently after
midnight.
Melatonin level has also been reported to decrease
during short-term fasting.16,32 As in the present paper,
delay in bedtime during Ramadan has been reported
previously. 2,33 This raises the question: does this sudden
delay in sleep-wake schedule affect melatonin secretion?
In a study by Morgan et al.,34 melatonin and body tem-
perature data showed no shift in the endogenous clock
during a 27-h schedule imposed on a group of volun-
teers. It is important to underline here that, in our study,
light-exposure characteristics were almost the same
during Ramadan and the baseline period except for the
duration (which was about 1.5 h longer during
Ramadan).
The exact mechanisms behind the reduced levels of
melatonin are unknown. However, postulated theories
to explain this finding are: (i) increased nocturnal cor-
tisol level during Ramadan; (ii) decreased melatonin
synthesis secondary to decreased glucose provision; and
(iii) decreased tryptophan provision (tryptophan is an
essential amino acid that cannot be synthesized by
human cells).
A nocturnal rise in cortisol level during Ramadan
has been reported previously.18,19 Brismar et al.35
showed that reduced cortisol production induced by
metyrapone stimulates human pinealocytes and
increases melatonin secretion. On the other hand, Beck-
Friis et al.36 have shown that administration of dexam-
ethasone caused significant reduction in melatonin level
in healthy participants.
A BaHammam
142 Sleep and Biological Rhythms 2004; 2: 135 –143
Malnourished rats have been found to have less mela-
tonin in the pineal gland than normally fed animals
have.37 This effect has been attributed to mild hypogly-
cemia causing reduction in N-acetyltransferase (NAT, an
enzyme involved in the synthesis of melatonin) because
levels of this enzyme decrease in starving rats.38 Röjd-
mark and Wetterberg32 demonstrated a reduced melato-
nin level in fasting participants with mild hypoglycemia;
glucose supplementation during fasting returned the
decreased melatonin level to normal. Although this the-
ory is plausible, it may not explain a decreased melato-
nin level during Ramadan fasting, as hypoglycemia is
not a known manifestation of Islamic fasting in healthy
participants.1 The tryptophan theory also seems
unlikely, as glucose supplementation restored normal
melatonin secretion.32
In summary, the present study showed that Ramadan
fasting resulted in significant reduction in SL and REM
sleep during the third week of Ramadan; otherwise,
there was no significant effect on sleep architecture.
Objective assessment of daytime sleepiness revealed no
increase in daytime sleepiness during Ramadan.
Although melatonin kept the same circadian pattern
during Ramadan, the level of the hormone dropped sig-
nificantly from baseline.
ACKNOWLEDGEMENTS
This project was supported by a grant from the College
of Medicine Research Center (CMRC), King Saud
University
REFERENCES
1Fereidoun A. Research in Islamic fasting and health. Ann.
Saudi. Med. 2002; 22: 186–91.
2BaHammam A. Sleep pattern, daytime sleepiness and
eating habits during the month of Ramadan. Sleep and
Hypnosis. 2003; 5: 163–70.
3Rocky R, Iraki L, HajKhlifa R, Ghazal NL, Hakkou F.
Daytime alertness, mood, psychomotor performance and
oral temperature during Ramadan intermittent fasting.
Ann. Nutr. Metab. 2000; 44: 101–7.
4Rocky R, Chapotot F, Taoudi M, Benaji B, Hakkou F, El
Khalifi H, Buguet A. Daytime sleepiness during Ramadan
intermittent fasting: polysomnographic and quantitative
waking EEG study. J. Sleep Res. 2003; 12: 95–101.
5Miles A, Thomas R. Melatonin – a diagnostic marker
in laboratory medicine? In: Miles, A, Philbrick, DRS,
Thompson, C, eds. Melatonin: Clinical Perspectives.
Oxford, Oxford University Press 1988; 253–79.
6Johns MW. A new method for measuring daytime sleep-
iness: the Epworth Sleepiness Scale. Sleep 1991; 14:
540–5.
7Wilson WL, Heyward VH. Validation of the near-infrared
interactance method for Native Indian, black, Hispanic
and white men 19–57 years. Basic Life Sci. 1993; 60:
389–92.
8Rechtschaffen A, Kales A, eds. A Manual of Standardized
Terminology, Techniques and Scoring System for Sleep Stages
of Human Subjects. NIH Publication #204 Washington:
DC. US. Government Printing Office, 1968.
9American Sleep Disorders Association Atlas Task Force.
EEG arousals: scoring rules and examples. Sleep 1992;
15: 174–84.
10 Thorpy MJ, Westbrook P, Ferber R, Fredrickson P,
Mahowald M, Perez-Guerra F, Reite M, Smith P. The
clinical use of the multiple sleep latency test. An Amer-
ican Sleep Disorders Association report. Sleep 1992; 15:
268–76.
11 Cajochen C, Dijk DJ. Electroencephalographic activity
during wakefulness, rapid eye movement and non-rapid
eye movement sleep in humans: comparison of their cir-
cadian and homeostatic modulation. Sleep Biol. Rhythm
2003; 1: 85–95.
12 Nowak R, McMillen C, Redman J, Short RV. The corre-
lation between serum and salivary melatonin concentra-
tions and urinary 6-hydroxymelatonin sulphate
excretion rates: two non-invasive techniques for moni-
toring human circadian rhythmicity. J. Clin. Endocrinol.
1987; 27: 445–52.
13 Laakso ML, Porkka-Heiskanen T, Alila A, Stenberg D,
Johansson G. Correlation between salivary and serum
melatonin: dependence on serum melatonin levels. J.
Pineal Res. 1990; 9: 39–50.
14 Frost G, Pirani S. Meal frequency and nutritional intake
during Ramadan: a pilot study. Hum. Nutr. Appl. Nutr.
1987; 41: 47–50.
15 Rocky R, Chapotot F, Hakkou F, Taoudi M, Buguet A.
Sleep during Ramadan intermittent fasting. J. Sleep Res.
2001; 10: 319–27.
16 Michalsen A, Schlegel F, Rodenbeck A, Ludtke R,
Huether G, Teschelr H, Dobos GJ. Effects of short-term
modified fasting on sleep patterns and daytime vigilance
in non-obese subjects: results of a pilot study. Ann. Nutr.
Metab. 2003; 47: 194–200.
17 Al-Hadramy MS, Hawaii TH, Abdelwahab SM. Altered
cortisol levels in relation to Ramadan. Eur. J. Clin. Nutr.
1988; 42: 359–62.
18 Sliman NA, Ajlouni KS, Khatib FA. Effect of fasting
Ramadan on some blood hormones in healthy Muslim
males. Mu’tah J. Res. Studies. 1993; 8: 91–109.
19 Elati J, Beji C, Danguir J. Increased fat oxidation during
Ramadan fasting in healthy women: an adaptive mech-
anism for body-weight maintenance. Am. J. Clin. Nutr.
1995; 62: 302–7.
Ramadan Fasting and Sleep
Sleep and Biological Rhythms 2004; 2: 135 –143 143
20 Iraki L, Bogdan A, Hakkou F, Amrani N, Abkari A,
Touitou Y. Ramadan diet restrictions modify the circa-
dian time structure in humans. Study on plasma gastrin,
insulin, glucose and calcium and on gastric pH. J Clin.
Endocrinol Metab. 1997; 82: 1261–73.
21 Krieger DT, Glick SM. Sleep EEG states and plasma
growth hormone concentration in states of endogenous
and exogenous hypercortisolemia or ACTH elevation. J.
Clin. Endocrinol. Metab. 1974; 39: 986–1000.
22 Sangiah S, Caldwell DF. Reduction of rapid eye move-
ment (REM) sleep by glucose alone or glucose and insu-
lin in rats. Life Sci. 1988; 42: 1425–9.
23 Born J, Spath-Schwalbe E, Schwakenhofer H, Kern W,
Fehm HL. Influences of corticotropin-releasing hor-
mone, adreno-corticotropin and cortisol on sleep on
normal man. J. Clin. Endocrinol Metab. 1989; 68: 904–11.
24 Pollak C. How should the multiple sleep latency test be
analyzed? Sleep 1997; 20: 34–9.
25 Clodoré M, Benoit O, Foret J, Bouard G. The multiple
sleep latency test: individual variability and time of the
day effect in normal young adults. Sleep 1990; 13: 385–
94.
26 Borbély AA. Sleep in the rat during food deprivation and
subsequent reinstitution of food. Brain Res. 1977; 124:
457–71.
27 Danguir J, Nicolaidis S. Dependence of sleep on nutri-
ents’ availability. Physiol. Behav. 1979; 22: 735–40.
28 Rashotte ME, Pastukhov IF, Poliakov EL, Henderson RP.
Vigilance states and body temperature during the circa-
dian cycle in fed and fasted pigeons (Columbia Livia).
Am. J. Physiol. 1998; 275: 690–702.
29 Willie JT, Chemelli RM, Sinton CM, Yanagisawa M. To
eat or to sleep? Orexin in the regulation of feeding and
wakefulness. Annu. Rev. Neurosci. 2001; 24: 429–58.
30 Arendt J. Melatonin and the Mammalian Pineal Gland.
London: UK. Chapman & Hall, 1995.
31 Bogdan A, Bouchareb B, Touitou Y. Ramadan fasting
alters endocrine and neuroendocrine circadian patterns.
Meal-time as a synchronizer in humans? Life Sci. 2001;
68: 1607–15.
32 Röjdmark S, Wetterberg L. Short-term fasting inhibits
the nocturnal melatonin secretion in healthy man. J. Clin.
Endocrinol. 1989; 30: 452–7.
33 Taoudi M, Rocky R, Toufiq J, Benaji B, Hakkou F. Epi-
demiological study: chronotype and daytime sleepiness
before and during Ramadan. Therapie 1999; 54: 567–72.
34 Morgan L, Arendt J, Owens D, Folkard S, Hampton S,
Deacon S, English J, Ribeiro D, Taylor K. Effects of the
endogenous clock and sleep time on melatonin, insulin,
glucose and lipid metabolism. J. Endocrinol. 1998; 157:
443–51.
35 Brismar K, Werner S, Wetterberg L. Melatonin and cor-
ticosteroid response to metyrapone in patients with
pituitary disease. In: Reiter, RJ, ed. The Pineal and its
Hormones. New York, NY: Liss 1982, 283–92.
36 Beck-Friis J, Ljunggren JG, Thoren M, von Rosen D,
Kjellman BF, Wetterberg L. Melatonin, cortisol and
ACTH in patients with major depressive disorder and
healthy humans with special reference to the outcome of
the dexamethasone suppression test. Psychoneuroendo-
crinol 1985; 10: 173–86.
37 Chick L, Ho AK, Brown GM. Effects of 24-h serum and
pineal melatonin content in male rats. Acta Endocrinol.
1987; 115: 507–13.
38 Welker HA, Vollrath L. The effects of a number of
short-term exogenous stimuli on pineal serotonin-N-
acetyltransferase activity in rats. J. Neural Trans. 1984;
59: 68–80.
