Acta Otolaryngol (Stockh) 1999; 119: 880 – 885

Middle Ear Pressure: Effect of Body Position and Sleep

BO TIDEHOLM, MARIANNE BRATTMO and BJO8 RN CARLBORG

From the Department of Oto-rhino-laryngology, Di6ision of Audiology, Uni6ersity Hospital, Malmoe, Uni6ersity of Lund, Sweden

Tideholm B, Brattmo M, Carlborg B. Middle ear pressure: effect of body position and sleep. Acta Otolaryngol (Stockh)
1999; 119: 880 – 885.
The aim of this study was to investigate whether a middle ear (ME) pressure rise noted during sleep is an effect of the
recumbent position, the state of sleep or Eustachian tube (ET) function, respectively. Eleven subjects with no history of
ear disease were subjected to continuous, 24-h, direct ME pressure measurements and tubal function tests, respectively.
At the start of the measurements the subjects were in the erect position. This position was maintained for at least 2 h
(during which they were permitted to conduct normal everyday activities). The subjects then rested in the recumbent
position for at least 2 h, after which they again resumed the erect position with normal activities for a period of at least
2 h. Then the subjects went to sleep in the recumbent position. A mean pressure rise in the ME of 36.4 daPa was seen
during sleep compared with being awake in the same position. The rise was not caused by tubal opening and active
insufflation of air via the ET. The number of tubal openings was significantly fewer during sleep compared with resting
awake and the erect position, respectively. The ME pressure was not significantly different in the erect position compared
with the recumbent position while resting awake. Tubal function tests demonstrated results in accordance with normal
tubal function for all subjects. In conclusion, the state of sleep induced a ME pressure rise, not the recumbent position
per se. These findings support the significance of a two-directional gas diffusion for regulation of the ME pressure. Key
words: pressure regulation, gas diffusion.

INTRODUCTION demonstrated a ME pressure rise in the recumbent

Several investigations have demonstrated that a posi-
tive middle ear (ME) pressure may arise sponta-
neously in the recumbent position during the night or
early in the morning (1 – 3). The physiological mecha-
nism for such a phenomena is not fully understood
(4). The finding may have important implications,
since it seems incompatible with the traditional con-
cept of ME pressure regulation and the role of the
Eustachian tube (ET). The classic theory ‘‘hydrops ex
vacuo’’ would need to be modified or replaced by a
more complex model of a two-directional gas diffu-
sion over the ME mucosa (3, 5 – 7). However, there
are still many questions. A spontaneously arising ME
pressure has been reported from studies with inter-
mittent and indirect recordings of the pressure (1, 2,
8). The effect occurred seemingly without contribu-
tion from the ET. To exclude such a possibility,
direct, continuous measurements would be required.
Although meticulous, intermittent and indirect stud-
ies cannot fully exclude the possibility of an ET effect
contributing to the pressure rise. Tympanometry per-
formed intermittently during the night is also likely to
have been associated with a variable state of sleep
and awareness among the patients (1, 2, 8). The ME
gas composition, which is claimed to be a key factor
for the ME pressure rise, has been extensively studied
in anaesthetized animals (9). Moreover, changes in
gas composition related to a ME pressure increase
during the night is yet to be verified in the normal
human ME. Also, it is unclear whether a pressure rise
is related to the recumbent position, sleep or a combi-
nation of sleep and body position. Knight et al.
position in awake patients with presumed ET ob-
struction due to common cold (10). In their study the
change in posture had no significant effect on ME
pressure in healthy subjects. In previous investiga-
tions we confirmed a ME pressure rise of low rate
during the night in the recumbent position (3). How-
ever, it was not possible to evaluate whether the effect
was related to the body position or the sleeping
condition. In order to study the association between a
pressure change and a specific activity, body position,
sleep, tubal openings or a combination of factors it
seems essential to monitor the ME pressure continu-
ously during physiological conditions that are as
normal as possible. Such a method is available and
has previously been described in detail (11).
The aim of this study was to establish whether a
ME pressure rise during the night is an effect of the
recumbent position per se, an effect of sleep or a of
change in tubal function. The ME pressure was
recorded in the erect and recumbent position in the
awake and sleeping state during the night.
MATERIALS AND METHODS
Eleven healthy subjects were selected with the follow-
ing criteria: no history of ear disease; no current or
previous ear disease during adult life, but childhood
secretory or acute otitis media was accepted; normal
history of ME pressure equilibration during flying
and diving; hearing threshold level (PTA) better than
20 dB HL; normal tympanometry and otomicro-
scopy.

© 1999 Scandinavian University Press. ISSN 0001-6489

Acta Otolaryngol (Stockh) 119
The mean age of the subjects was 28 years (range
24–44 years). A small myringotomy was made in the
tympanic membrane. A small drop of 90% phenol
was used as topical anaesthesia.
Continuous 24 h ME pressure measurements and
tubal function tests were performed in all 11 subjects.
The method for ME pressure measurement was de-
veloped by our group and has been described previ-
ously in detail (11). A commercial ear protector
(Comfit®) with tubing was tightly fitted deep into the
external ear canal. The tubing was coupled to a
piezo-electric pressure transducer ( SenSym®). The
transducer was connected to an instrument amplifier
and via a low pass filter to an 8 bit AD-converter.
The pressure resolution was 3.9 daPa. When evaluat-
ing the tubal function with requirements of very high
temporal resolution, equipment was used with a sam-
pling rate of 18.2 Hz.
For continuous long-term pressure measurements
the equipment included a crystal-driven clock and a
digital memory. The sampling rate was 1.25 Hz and
the sampling interval was 0.8 s. This was previously
found adequate for detecting pressure changes ex-
ceeding 3.9 daPa during 0.8 s (11). Thus, significant
rapid pressure changes could be detected and low rate
pressure trends analysed in detail. The requirements
of temperature and baseline stability were well met
(11). The apparatus measuring 120×60 × 35 mm was
conveniently worn in a specially designed chest har-
ness. The capacity of the memory was sufficient for
27 h of continuous pressure recording. The connec-
tions of the tubing were checked regularly and signs
of leakage were objectively tested before and after
each experiment by inducing a standard pressure
change into the ME (11).
Each subject was initially subjected to tubal func-
tion tests and thereafter to continuous ambulatory
pressure measurement for 24 h. Microscopy after the
measurement in each subject were performed in order
to a establish that the myringotomy still was open.
Tubal function tests were performed according to a
test protocol previously described in detail (11, 12).
The tests included Valsalva’s and Toynbee’s manoeu-
vres, powerful sniff provocations, determination of
the opening pressure of the ET and the pressure
equilibration capacity respectively. The opening pres-
sure of the ET was determined by introducing into
the ME a positive pressure at an even rate. The
opening pressure was defined as the level at which a
verified rapid pressure change \ 10 daPa towards
ambient pressure was recorded in the ME.
The continuous long term ME pressure measure-
ments were performed in all 11 subjects according to
the following specific instructions (Fig. 1):
Middle ear pressure 881
1. erect position during at least 2 h of normal activ-
ities, followed by;
2. recumbent position for at least 2 h while staying
fully awake, followed by;
3. erect position during at least 2 h of normal activ-
ities, followed by;
4. sleep in the recumbent position for the night.
A notation protocol was used for entering the
exact time of each change in body position and for
entering the estimated time of falling asleep as exactly
as possible. The measurements occurred while ambu-
latory and off clinic.
Otomicroscopy was performed 3 weeks after the
measurement for evaluation of the iatrogenic perfora-
tion of the tympanic membrane.
The investigation was approved by the Research
Committee on Ethics, University of Lund, Sweden
(LU 185–93).
Analysis
The entire pressure recordings were examined in de-
tail and correlated to the exact time of changes in
body position, resting and sleep as given by the
notation protocols of the subjects. Pressure equilibra-
tion or tubal opening during the 24 h measurements
was defined as a rapid pressure change \ 10 daPa
towards ambient pressure, resulting in a new tran-
siently stable pressure level. The mean pressure for
each hour of the measurement was calculated using
the area under the graph. An hour was split into
1,200 equal portions and a mean value for each hour
was calculated from this. The difference in pressure
between erect, resting and sleeping conditions respec-
tively was calculated. Wilcoxon signed rank test was
used for the statistical calculations. A value of p5
0.05 was considered significant.
Fig. 1. Experimental set-up. The mean ME pressure and
results of the statistical comparison are given.
882 B. Tideholm et al.
RESULTS
The continuous measurements demonstrated a simi-
lar pressure pattern for all subjects but one. This
subject had signs of leakage from the ear canal during
the major part of the recording and was excluded
from the study. The remaining 10 subjects demon-
strated predominantly slow trends of pressure
changes. Few rapid pressure changes, meeting the
criteria of tubal opening in this study, were noted.
The mean number of tubal openings in the erect
position were 9.4/h. In the recumbent position while
resting awake a mean of 8.4/h openings were noted.
During sleep the mean value was 3.2/h. The number
of tubal openings were significantly fewer during
sleep compared to erect and resting awake conditions
respectively (p = 0.014 and p= 0.049, respectively).
The number of tubal openings was not significantly
different in the erect and resting awake conditions
(p=0.667).
The mean pressure for the 2 h of erect position,
preceding the resting condition, was − 7.1 daPa com-
pared with −17.0 daPa for the second episode of
erect position (Fig. 1). The difference was not statisti-
cally significant (p = 0.059). The mean pressure for
the total of 4 h in erect position was − 12.1 daPa.
The mean pressure for the 2 h of resting awake in
recumbent position was −20.1 daPa. A paired com-
parison between the erect and the recumbent awake
conditions revealed no statistically significant pres-
sure difference (p =0.139).
According to the notation protocols all subjects
slept in the recumbent position for at least 4 consecu-
tive hours. The ME pressure was seen to rise at a
slow rate. Few pressure changes indicative of tubal
openings occurred. But shortly after an opening the
ME pressure again started to rise at a slow rate. The
mean pressure for the first 2 h after going to sleep in
the recumbent position was +15.9 daPa compared
with +16.8 daPa for the subsequent 2 h of sleep
(Fig. 1). There was no significant difference and the
mean pressure for the first 4 h of sleep was +16.3
daPa.
A paired comparison between the 4 h in erect
position and 4 h of sleep revealed a significantly
higher pressure during sleep (p =0.007). The differ-
ence in mean pressure was 28.4 daPa. A significantly
higher pressure during sleep was demonstrated also in
a paired comparison with the recumbent position
while resting awake (p =0.009) (Fig. 2). The differ-
ence in mean pressure was 36.4 daPa.
Results from the tubal function tests are summa-
rized in Table I. All 10 subjects were able to perform
Valsalva’s manoeuvre. A direct pressure rise to ap-
proximately 500 daPa was detected in the ME of all
Acta Otolaryngol (Stockh) 119
Fig. 2. ME pressure during 10 h of continuous measure-
ments in 10 subjects. Comparison of the mean ME pressure
pattern during erect and resting awake conditions (thin line)
vs the subsequent erect and sleeping conditions (thick line).
The graph was calculated using a sliding mean during 5
min for 10 subjects. The time of change in body position is
illustrated by 0.
subjects. Equilibration tests demonstrated that all
subjects could equilibrate + 200 daPa in 5 4 degluti-
tions. In 2 subjects the ET opened before reaching
+ 200 daPa. The opening pressure of the ET ranged
from 50 to 340 daPa.
Equilibration of negative pressure was successful in
7 subjects. In one of these subjects the ET was
spontaneously opened before reaching − 200 daPa.
Three subjects could not equilibrate − 200 daPa in
54 deglutitions and had to use Valsalva’s manoeu-
vre, i.e. negative test results. Three subjects could
change the ME pressure with Toynbee’s manoeuvre.
Only one of the subjects could change the ME pres-
sure with a powerful sniff.
At the end of the registrations most subjects re-
ported a mild discomfort from the ear. The investiga-
tion did not result in any side-effects such as infection
in the ME or in the ear canal. All tympanic mem-
brane perforations were healed within 3 weeks.
DISCUSSION
A positive pressure spontaneously arising in the ME
during the night has been reported in several investi-
gations using different methods (1, 3, 8, 13). The
underlying mechanisms have not been fully clarified,
and it is unclear whether the pressure rise is related to
the recumbent position during the night, the state of
sleep, changes in ET function or ME gas composition
respectively. A spontaneous pressure rise not related
to tubal openings or active insufflation of air via the
ET may have important implications for the theory
of ME pressure regulation. Thus, it is of importance
to elucidate this phenomena by continuous, direct
ME pressure measurements under circumstances as
close to normal as possible, i.e. ambulatory with no
Acta Otolaryngol (Stockh) 119 Middle ear pressure 883

restrictions on normal everyday and night-time activ-
ities. A limited number of subjects were used in this
study. But, the results seem representative for a group
of subjects with normal ME function. The subjects
were selected according to the same criteria as used in
the previous study of long term ME pressure in
normal subjects (3). A comparison between the two
studies demonstrated similar results for tubal func-
tion tests as well as for the long-term pressure pattern
during the erect and sleeping conditions, respectively
(3).
The present investigation confirmed previous re-
sults of a slight negative ME pressure in erect posi-
tion and added the information of a slight mean
negative ME pressure also in the recumbent position
while subjects were resting awake. There was no
statistically significant pressure difference. For vari-
ous reasons a pressure rise might have been expected
when resting awake in the recumbent position, e.g.
presumed reduction in tubal openings, venous con-
gestion in the ME mucosa and pressure equipment
sensitive to positional changes (14, 15). However, the
number of tubal openings, as defined in the study was
not significantly different in the 2 separate conditions.
The pressure sensor used reads pressure relative to
ambient pressure and consequently is not influenced
by changes in body position. Further, venous conges-
tion in the recumbent position, reported by others,
did not significantly influence the mean ME pressure
(16). One previous investigation has reported on a
pressure increase in patients awake in the recumbent
position (10). The results were obtained from patients
with common cold and a presumed ET obstruction.
The patients were restrained to the location of the
investigation and the ME pressure measured indi-
rectly and intermittently with tympanometry. In their
study five healthy subjects were also investigated, and
in all of them postural change did not significantly
change the ME pressure. The pressure results from
these healthy subjects are in accordance with our
results.
This investigation demonstrated a mean pressure
rise in ME during the night while the subjects were
sleeping in the recumbent position. The pressure rise
was significant compared with the erect as well as the
recumbent position when the subjects were resting
awake. The pressure increase was not related to
differences in body position, since this was similar
when resting awake and sleeping. Neither can the rise
be attributed to differences in pre-experimental ME
pressure levels, since there was no significant pressure
difference before going to rest and before going to
sleep. The pressure rise during the night was of slow
rate, not compatible with an effect of tubal opening
and active insufflation of room air. The increase
seemed to start after each tubal opening, a pattern
not normally seen when the subjects were awake (3).
The number of tubal openings was significantly fewer
during sleep compared with the other two conditions.
It is our belief that diminished tubal ventilation dur-
ing sleep is not the cause of the pressure rise, but that
it helped to reveal the process. A plausible explana-
tion for a pressure rise during sleep might be given by
differences in blood gas composition in the resting
and sleeping conditions (17–19).
A ME pressure increase associated with few tubal
openings seems to indicate a positive turnover of gas
in the ME cavity during sleep. A two-directional gas
diffusion over the ME mucosa is likely to occur. It
has been demonstrated that pulmonal ventilation falls
by approximately 20% during sleep leading to a slight
decrease in PaO2 and elevation of PaCO2 (17, 18).
There is a minor ventilation/perfusion mismatch in
the lung. This mismatch increases in the recumbent
position. However, functional residual capacity de-
creases and diffusing capacity increases in normal

Table I. Results from tubal function tests in 10 subjects. ME pressure in daPa. Positi6e (Pos): the pro6ocation
altered the ME pressure; negati6e (Neg): the pro6ocation did not alter the ME pressure.
Equilibration test

Sub Valsalva Toynbee Sniff +200 −200 Opening pressure (daPa)

1 Pos Neg Neg Pos Pos 270

2 Pos Neg Neg Pos Pos 120
3 Pos Pos Neg Pos Pos 310
4 Pos Pos Neg Pos Pos 270
5 Pos Neg Neg Pos Neg 280
6 Pos Neg Neg Pos Neg 320
7 Pos Neg Neg Pos Neg 340
8 Pos Neg Neg Pos Pos 320
9 Pos Neg Neg Pos Pos 310
10 Pos Pos Pos Pos Pos 50
884 B. Tideholm et al.
subjects when they change position from erect to
recumbent position. The effects of these factors are
normally opposing not resulting in any net change in
the blood gas composition.
During sleep this may induce a net transport of
CO2 to the ME and a net transport of O2 from the
ME. Since CO2 is transported much more rapidly
over the ME mucosa compared with O2, a greater
transport of gas is likely to occur to the ME than
from the ME (20, 21). Such a situation would lead to
a net increase in ME pressure during sleep as found
in this investigation. This reasoning is supported by
several experimental studies performed under condi-
tions other than rest and sleep (7, 9, 22, 23). Similar
effects on blood gas composition have not been docu-
mented in subjects resting awake in the recumbent
position (24).
This investigation had neither objective means to
verify the exact time of actual sleep nor objective
parameters to assure the state of awareness during
rest. Thus, it cannot be excluded that some subject
fell asleep for a brief time while resting or was awake
for a short period noted as sleep. However, the
subjects were thoroughly notified of the significance
of the information given and were accordingly in-
structed to note times for the specific events as ex-
actly as possible. Furthermore, we compared the ME
pressures for the first 2 h of sleep with the subsequent
2 h of sleep. There was no significant ME pressure
difference. Since we did find a significant difference
between resting awake and sleeping, the lack of ob-
jective control does not seem to have had a negative
influence on the power of the statistical calculations.
CONCLUSION
Subjects with normal tubal function had a mean
pressure rise in the ME of 36.4 daPa during sleep in
the recumbent position compared with being awake
in the same position. There were no indications that
tubal opening or active insufflation of air contributed
to the pressure rise. The ME pressure was not signifi-
cantly different in the erect position during normal
every day activities compared with the recumbent
position while resting awake. The study indicates that
the state of sleep induced a ME pressure rise, not the
recumbent position per se. These findings support the
significance of a two-directional gas diffusion for the
regulation of the ME pressure and give little support
to the traditional concept of hydrops ex 6acuo.
ACKNOWLEDGMENTS
This study was supported by grants from Helga Hjerpsteds
fund, the Swedish Society for Otolaryngology, Head and
Acta Otolaryngol (Stockh) 119
Neck Surgery and the Research funds of the University
Hospital, Malmoe.
REFERENCES
1. Hergils L, Magnuson B. Morning pressure in the mid-
dle ear. Arch Otolaryngol Head Neck Surg 1985; 111:
86–9.
2. Shinkawa H, Okiysu T, Yusa T, Yamamuro M,
Kaneko Y. Positive intratympanic pressure in the
morning and its aetiology. Acta Otolaryngol (Stockh) ;
Suppl 1987; 435: 107 – 11.
3. Tideholm B, Carlborg B, Jönsson S, Bylander-Groth
A. Continuous long-term measurement of the middle
ear pressure in subjects without a history of ear disease.
Acta Otolaryngol (Stockh) 1998; 118: 369 – 74.
4. Hergils L, Magnuson B. Middle ear pressure under
basal conditions. Arch Otolaryngol Head Neck Surg
1987; 113: 829 – 32.
5. Sadé J, Luntz M. Gas diffusion in the middle ear. Acta
Otolaryngol (Stockh) 1991; 111: 354 – 7.
6. Doyle WJ. Increases in middle ear pressure resulting
from counter diffusion of oxygen and carbon dioxide
into the middle ear of monkeys. Acta Otolaryngol
(Stockh) 1997; 117: 708 – 13.
7. Ostfeld EJ, Silberberg A. Theoretic analysis of middle
ear gas composition under conditions of nonphysio-
logic ventilation. Ann Otol Rhinol Laryngol 1992; 101:
445 – 51.
8. Bylander AKH, Ivarsson A, Tjernström O8 , Andréasson
L. Middle ear pressure variations during 24 hours in
children. Ann Otol Rhinol Laryngol ; Suppl 1985; 120:
33–5.
9. Yee AL, Cantekin EI. Middle ear pressure after
changes in steady state. Acta Otolaryngol (Stockh)
1987; 104: 261 – 9.
10. Knight L, Eccles R. The effect of postural change and
upper respiratory tract infection on middle ear pres-
sure. Acta Otolaryngol (Stockh) 1991; 111: 1075 –82.
11. Tideholm B, Jönsson S, Carlborg B, Welinder R, Gren-
ner J. Continuous 24-hour measurement of middle ear
pressure. Acta Otolaryngol (Stockh) 1996; 116: 581–8.
12. Elner A, , Ingelstedt S, Ivarsson A. The normal function
of the Eustachian tube. Acta Otolaryngol (Stockh)
1971; 72: 320 – 8.
13. Luntz M, Sadé J. Daily fluctuations of middle ear
pressure in atelectatic ears. Ann Otol Rhin Laryngol
1990; 99: 201 – 4.
14. Rundcrantz H. Posture and Eustachian tube function.
Acta Otolaryngol (Stockh) 1970; suppl 263: 15 – 7.
15. Rundcrantz H. Posture and Eustachian tube function.
Acta Otolaryngol (Stockh) 1969; 68: 279 – 92.
16. Andréasson L, Ingelstedt S, Ivarsson A. Pressure de-
pendent variation in the volume of the mucosal lining
of the middle ear. Acta Otolaryngol (Stockh) 1976; 81:
442–9.
17. Douglas NJ, White DP, Pickett CK, Weil JV, Zwillich
CW. Respiration during sleep in normal man. Thorax
1982; 37: 840 – 4.
18. Stradling JR, Chadwick GA, Frew AJ. Changes in
ventilation and its components in normal subjects dur-
ing sleep. Thorax 1985; 40: 364 – 70.
19. Schafer T. Variability of vigilance and ventilation:
studies on the control of respiration during sleep.
Respir Physiol 1998; 114: 37 – 48.
Acta Otolaryngol (Stockh) 119 Middle ear pressure 885

20. VanLiew HD. Oxygen and carbon dioxide permeability ness and sleep in normal human subjects. J Appl
of subcutaneous pockets. Am J Physeol 1962; 202: Physiol 1995; 78: 1339 – 49.
53–8.
21. Sadé J, Saba K, Luntz M. Middle ear as a gas pocket. Submitted August 10, 1999; accepted September 9, 1999
Ann Otol Rhinol Larynol 1990; 99: 529–34.
22. Ikarachi F. The effect of respiratory mode on human Address for correspondence:
middle ear pressure. Aris Nasus Larynx 1998; 25: 349 – Bo Tideholm, MD
54. Department of Oto-rhino-laryngology
23. Mover-Lev H, Levy D, Luntz M, Marell M, Ar A, University Hospital
Sadé J. Dependence of middle ear gas composition on SE-205 02 Malmö
pulmonary ventilation. Ann Otol Rhinol Laryngol Sweden
1997; 106: 314–9.
Fax: + 46 40-337066
24. Morrell MJ, Harty HR, Adams L, Guz A. Changes in
E-mail: bo.tideholm@oron.mas.lu.se
total pulmonary resistance and pCO2 between wakeful-

.
Copyright of Acta Oto-Laryngologica is the property of Taylor & Francis Ltd and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.