You are here

Thermoregulation in Patients With Obstructive Sleep Apnea

Dr. Ibrahim

Authors:  Sherin M. Ibrahim, DO; Stephen Lund, MD; Jon Freeman, PhD

Previous research has  demonstrated a  strong  relationship between a decline in
core body temperature and the initiation of sleep.   This relationship is  supported  by
observations that the circadian rhythm  of core body temperature (CBT) is inversely
correlated with changes in sleep duration and propensity.
   Early studies of
thermoregulation and sleep indicate that sleep onset is associated with both increased heat
loss and reduced heat production.
  Recently, distal skin temperature has shown promise
as a predictor of sleep onset latency.
   Distal skin regions have shown increases in
temperature of approximately 0.5-1.0 degree Celsius in the evening prior to sleep onset.
These increases in distal skin temperature have been linked with objective and subjective
The circadian pattern of heat loss and heat production is not a precise sine wave;
heat loss is more dominant in the evening than reduction of heat production and vice
versa in the morning.
    The onset of sleep is seen on the downward portion of the CBT curve when
the distal skin temperatures are increasing most rapidly.
  The increase in distal skin temperature is best explained by the
presence of dense arteriovenous anastomoses (AVA) in such areas as the hands and feet.
Vasodilation of these distal AVAs promotes rapid heat loss.  The vasodilation is
prompted by the  withdrawal of sympathetic constrictor tone and potentiated by
   Preceding sleep, reduced activation of sympathetic constrictor innervation
relaxes circular muscles of AVA shunts allowing greater inflow of heated blood from the
core, thus promoting heat loss to the environment through the skin surface.
Although the usefulness of measuring distal skin temperature has been illustrated
in patients with sleep onset insomnia, there has been limited application of it to patients
with obstructive sleep apnea (OSA).  It is not known whether the presence of OSA has an
effect on the circadian rhythm of body temperature.  Furthermore, would the severity of
OSA show a significant change as compared to healthy subjects?  The aim of this study
was to investigate if OSA and severity alters the known  relationship between
thermoregulation and sleep.
Patients that presented to the Sleep Disorders Institute for investigation of OSA
were initially screened.   No compensation was given for participation in the research. 
There were inclusion criteria for each participant.  Participants must (a) be between the
ages of 18 to 55; (b) have no previous diagnosis or treatment of OSA; (c) be able to read
and understand the consent form; (d) not have a history of peripheral vascular disease or
some other medical disorder known to effect body temperature such as an uncontrolled
thyroid; (e) not have sleep onset insomnia (>30 minutes); and (f) not be a shift worker. 
Twenty-two patients met inclusion criteria.  However, during the nocturnal
polysomnography  (NPSG), two of the patients met criteria to be split (have CPAP
initiated due to severity of OSA)  and were excluded.  The study was approved by the
BioMed IRB.

The  study  utilized between-groups, mixed model design of patients  who
presented to the Sleep Disorders Institute for investigation of the presence of OSA. 
Anonymity and confidentiality was maintained throughout the course of the research. As
each selected patient’s information  was included in the data base, the patient  was
identified by their initials only.  Once the data base  was completed, the  initials were
converted into a new identifier number beginning with 01 and continuing upwards. 

Patients presenting for the NPSG had a standard montage of electrodes applied. 
These included the electroencephalograph (EEG)  F3,  F4, C3, C4, O1, O2, A1 and A2,
electrooculograph (EOG) and chin electromyography (EMG) electrodes.  Monitoring of
thoracic and abdominal movements, nasal flow assessed by pressure transducer and oronasal airflow assessed by thermocouple device, and pulse oximetry were also completed
per standard protocol. NSPG data were collected using Grass Gamma (v.4.3, 4.8)
systems.  A skin temperature thermistor (YSI  400  Series Probe, Yellow Springs
Instrument Co Inc, OH, USA) was attached on the dominant hand using supplied
adhesive pads and reinforced with thin, porous surgical tape.
Sleep  staging was performed using criteria of Rechtschaffen and Kales and
respiratory and myoclonic events were scored according to AASM criteria.  The scoring
of sleep staging and respiratory events were completed by standard SDI scorers that were
blind to the purposes of this study.  Based on AHI, subjects were stratified into three
groups: 1) Subjects with AHI < 19 were the contrasts for this study (include mild OSA),
2) moderate OSA (AHI = 20 – 39) and 3) severe OSA (AHI > 40). 
  All temperature data were
collected in 5-minute intervals on a Respironics mini-logger 2000 data recorder (accuracy
±0.1  °C, Mini-Mitter Co., Inc. Bend, OR).  Ambient room temperature was maintained
between 20 to 23 °C throughout the study.   Subjects were instructed to go to sleep
according to their habitual lights out time. 

Statistical Analysis
A repeated measures  ANOVA approach  was utilized to  measure temperature
across time and  between  groups.   Temperature  was the dependent variable and was
plotted  over time. Temperature plots were adjusted for lights out based patient self
reports.   Temperature samples were taken at 5 minute intervals and condensed into 30
minute windows for analysis.  Table 1 indicates the windows of temperature analyses that
were adjusted to the NPSG lights out time.  Although time points were collected as far
out as point Q, points O-Q were eliminated from analysis due to too many missing time
Group membership defined by the presence and severity of OSA was the
independent variable.   Comparisons by group were made for sleep variables including
sleep latency, wake after sleep onset (WASO), sleep efficiency, percentage of stages 1, 2,
3, 4 and REM sleep, AHI and lowest SaO2.   Demographics and sleep variables were
examined with a one-way ANOVA with Tukey’s HSD utilized for post-hoc analysis.

Tables 2 shows results for demographic and sleep staging  variables.  Although
groups differed with respect to anticipated areas  (e.g. BMI, lowest SaO2, % of Stage 1
and 2 sleep); groups also differed by age.  The severe OSA group was significantly older
than the contrast group (p=0.004).  For this reason, subsequent analysis of temperature by
time by group statistically controlled for differences in age.
Data was missing for 3  time points for 2 subjects.  This was a result of the
temperature probe coming off of the finger while in the bathroom.  Missing data was
handled by taking a mean value of distal temperatures between adjacent  time  points. 

Table 1 : Temperature Time Measurement
A Pre-lights Out- Lights Out
B 0 – 30 minutes
C 31 – 60 minutes
D 1 – 1.5 hours
E 1.5 – 2 hours
F 2 – 2.5 hours
G 2.5 – 3 hours
H 3 – 3.5 hours
I 3.5 – 4 hours
J 4 – 4.5 hours
K 4.5 – 5 hours
L 5 – 5.5 hours
M 5.5 – 6 hours
N 6 – 6.5 hours


Table 2 : Demographic and Sleep Staging Variables
Variable Contrast (Mild)

  Mean  ± sd Moderate Mean ± sd Severe Mean ±sd P value
Age  30.56 ±9.46  35.8 ±3.11  49.17 ±11.29  .004*
BMI  26.89 ±5.2  28.8 ±4.92  35.17 ±6.3  .033*
AHI 5.1 ±5.04  29.36 ±5.45  65.9 ±19.54  .000
Sleep Latency 12.33 ±8.24  5.6 ±6.23  14.17 ±12.27  .299
Latency to Persistent Sleep 24.44 ±20.05  9.4 ±11.43  47 ±62.71  .261
WASO  53.89 ±24.69 62 ±38.43  110.17 ±62.17  .056
TST 378.5 ±44.72  350.5 ±40.41  320.08 ±77.32  .165
Sleep Efficiency 84.54 ±7.02  83.72 ±9.86  71.38 ±14.45  .064
Stage 1%  10.43 ±4.71  12.44 ±3.38  26.42 ±17.53  .022*
Stage 2%  54.63 ±7.79  49.16 ±7.73 34.33 ±16.31  .009*
Stage 3%  4.91 ±3.6  5.54 ±2.86   2.08 ±4.3 .251
Stage 4%   3.42 ±4.07 2.82 ±2.71 0.35 ±0.86 .193
REM %  13.63 ±6.94   15.4 ±3.55 11.7 ±9.45 .701
REM Latency 207.56 ±99.69 106.6 ±48.87 184.83 ±112.53  .185
Low SaO2 92.11 ±1.9  87.6 ±4.39  81.67 ±10.29  .017*

*Severe > Contrast (Tuley’s HSD)

Van Den Heuvel, et al. Changes in sleepiness & body temperature precede nocturnal sleep onset:
Evidence from a polysomnographic study in young men. J. Sleep Res. 1998, 7: 159-166.
Kreider, et al. Oxygen consumption and body temperatures during the night. J. Appl. Physiol., 1958, 12:
Lack, et al. Acute finger temperature changes preceding onsets over a 45-h period. J. Sleep Res. 2002,
Kubo, et al. Sleep stage and skin temperature regulation during night-sleep in winter.  Psychiatry Clin.
Neurosci., 1999, 53: 121-123.
Gilbert, et al. Attenuations of sleep propensity, core hypothermia, and peripheral heat loss after
temazepam tolerance. Am. J. Physiol., 200, 279:R1980-R1987.
Krauchi, Wirz-Justice (1994). Circadian rhythm of heat production, heart rate, and skin and core body
temperature under unmasking conditions in men. Am. J. Physiol. 267:R819-R826.
Krauchi, et al. A functional link between distal vasodilation and sleep-onset latency? Am. J. Physiol.,
2000, 278:R741-R748.
Krauchi, Wirz-Justice, et al. Thermoregulatory effects of melatonin in relation to sleepiness. 
Chronobiology Int.,2006,  23: 475-484.
Lack, et al. Acute finger temperature changes preceding onsets over a 45-h period. J. Sleep Res. 2002,
Van Den Heuvel, et al. Changes in sleepiness & body temperature precede nocturnal sleep onset:
Evidence from a polysomnographic study in young men. J. Sleep Res. 1998, 7: 159-166.