Quantitative Aspiration During Sleep in

Normal Subjects*
Kevin Gleeson , MD; Douglas F. Eggli, MD; and Steven L. Maxwell, DO, FCCP

Study objective: To determine the within-subject variability and to estimate the quantity of occult
aspiration of nasopharyngeal secretions during sleep in normal humans.
Design: Prospective duplicate full-night sleep studies.
Setting: Pulmonary sleep laboratory, university hospital.
Participants: Ten normal male volunteers aged 22 to 55 years.
Interventions: Two full-night polysomnographic recordings with infusion of 2 mUh radioactive
99
mTc tracer into the nasopharynx through a small catheter during EEG-documented sleep.
Standard lung scans were conducted immediately following final awakening. Aspiration was
defined as the presence of radioactivity in the pulmonary parenchyma on two separate views.
Results: A mean sleep efficiency of 85.72.6% was found with no difference between the two
study nights. A total of 5 of the 10 subjects studied had tracer evident in the pulmonary
parenchyma following final awakening. Three had the tracer apparent following the first-night
study and four had tracer apparent following the second-night study. Thus, two subjects aspirated
on both nights. Comparing the subjects who aspirated with those who did not, no significant
difference could be found for age, time spent in bed, sleep efficiency, apnea-hypopnea index,
arousal plus awakening index, or percent of sleep time spent in a supine position. The quantities
of tracer aspirated were on the order of magnitude of 0.01 to 0.2 mL.
Conclusions: Aspiration measured by this technique occurs commonly in healthy young men
during sleep, is unrelated to sleep quality, and is variable within subjects studied on more than
one occasion. The quantity aspirated is of an order of magnitude likely to contain bacterial
organisms in physiologically significant quantities. (CHEST 1997; 111:1266-72)

Key words: aspiration; pneumonia; radionuclide; sleep

Aspiration of upper airway secretions and coloniz-
ing microorganisms is thought to be a major
pathogenic event preceding most cases of bacterial
pneumonia acquired outside (community-acquired
pneumonia), and especially within, the hospital (nos-
ocomial pneumonia). It is now widely believed that
silent aspiration into the intrathoracic airways is a
common occurrence in normal subjects during sleep.
Pneumonia is usually prevented by mechanical and
immunologic defense mechanisms such as cough,
mucocilia1y escalator clearance, airway and alveolar
phagocytic cells, and immunoglobulins. According to
this theory, circumstances that either increase the
volume of aspirated secretions or impair normal host
*From the Departments of Medicine (Drs. Gleeson and Max-
well) and Radiology (Dr. Eggli), The Milton S. He rshey Medical
Center, The Pennsylvania State University, Hershey.
Manuscript received October 7, 1996; revision accepted Decem-
ber ll.
Reprint requests: Dr. Gleeson, Pulmonary/Critical Care Division,
The Milton S. Hershey M edical Center, The Penns ylvania State
University, 500 University Ave, Hershey, PA 17033
defenses are risk factors for developing bacterial
pneumonia with microorganisms colonizing the su-
praglottic airways .1
The most direct evidence supporting the hy-
pothesis that silent aspiration occurs during sleep
in normal subjects was published in 1978 by
Huxley and coworkers 2 who injected small boluses
of a solution of 111 In chloride into the noses of 20
normal subjects, and 10 patients with stupor or
coma, during apparent nocturnal sleep. In 9 of
these 20 subjects, some 111 In was found in lung
parenchyma the following day. These authors
noted that those who slept most soundly all aspi-
rated, with restless sleepers not aspirating, and
concluded that sound sleep was a risk factor for
aspiration of upper airway secretions . In furth er
support of this conclusion, 7 of the 10 patients
they studied with depressed consciousness aspi-
rated. In contrast to these results, a second group
of investigators , using a slightly different tech-
nique, found apparent nocturnal aspiration in only
1 of 10 normal subjects, while 10 of 14 patients
with previous community-acquired pneumonia

1266 Clinical Investigations

had 111 In evident in lung parenchyma when
scanned the following morning. 3
These two studies share the weakness of not
documenting sleep using standard EEG criteria4 or
any other quantifiable criteria, thereby limiting the
validity of their observations concerning sleep quality
and quantity. In addition, neither study addressed
the issue of whether nocturnal aspiration is subject-
specific. If silent aspiration occurs during sleep in
50% of subjects studied, does this mean that this half
of the population aspirates nightly, with the remain-
ing half never aspirating, or does it imply that the
entire population can be expected to aspirate fre-
quently, but not every night? This question could be
directly addressed only by studying both aspirators
and nonaspirators on multiple occasions. To improve
our understanding of silent aspiration during sleep,
we studied a group of 10 subjects during two full-
night polysomnograms each. We monitored sleep
stages while staining the upper airway secretions
with 99 mTc sulfur colloid, scanning the thorax the
morning after each study to ascertain if aspiration
had occurred occultly the previous night. We hy-
pothesized that aspiration during sleep, determined
by this technique, would occur randomly across the
group on the two nights, rather than recurrently
within several individual subjects, suggesting that
aspiration during sleep is a universal rather than
subject-specific event in normal subjects.
MATERIALS AND METHODS
Subjects and Equipment
We studied 10 healthy men aged 22 to 55 years. None had any
health or sleep complaint or took any medication. All were asked
to refrain from caffeine-containing beverages for 12 h prior to
being studied. Electro-oculogram, EEG, and chin electromyo-
gram were monitored using standard silver paste-on electrodes.
Respiration was recorded qualitatively using thermistors and a
C0 2 sampling catheter attached to a loosely fitting facial mask. In
addition, a noncalibrated chest and abdominal vest was used to
record movement of the respiratory apparatus. Artelial hemoglo-
bin oxygen saturation was measured using an ear oximeter (Biox
II; Ohmeda Monitoling Systems; Louisville, Colo). The patient's
body position during sleep was monitored using a sleep posture
monitor (Sleep Technologies; Chicago). Signals from all of the
above devices were recorded on a polygraph recorder (Grass
78D; Grass Instrument Co; West Warwick, RI). In addition, a
complete video and audiotape recording of the sleeping period
was made, which employed an electronic signal for precise
correlation with the polygraph hard copy record.
Access to the nasopharynx was attained by placing a pliable
polyethylene tube approximately 2 em into either naris and
securing it to the nose with adhesive tape. This catheter was
attached by a low deadspace, 100-cm extension tubing to a 30-mL
sylinge containing approximately 37 MBq of 99 mTc sulfur colloid
in a 10-mL volume. Radionuclide infusion into the nose could
thus be controlled using a variable pressure infusion pump (IVOX
560; Farley, Inc; Simi Valley, Calif).
Protocol
All studies were conducted during regular nocturnal hours (10
PM to 7 AM) in a sound-proofed experimental bedroom. Follow-
ing w1itten informed consent in accord with a protocol approved
by the Institutional Review Board, each subject was provided
vvith a hospital surgical ("scrub") shirt to wear and was then
attached to the above recording equipment. Next, the polyethyl-
ene catheter was inserted into one naris and securely taped to the
nose, and the subject was allowed to fall asleep.
Once a stable sleep stage was achieved, the infusion of
technetium solution into the nose was begun at a rate of 2 mUh.
Provided that the subject remained in bed for the 5-h infusion
period, the infusion was uninterrupted during this interval. If the
subject awoke and needed to urinate, the infusion was discontin-
ued until a stable sleep stage was reestablished.
Upon the patient's final awakening, the catheter was removed
from the nose and isolated with the syringe. Then the mask with
the attached thermistor/capnograph probe, pillow case, hospital
surgical shirt, and bedsheets were collected to assure isolation of
radioactive matelial. The subject was given a 6-oz glass of water
to drink and was transported immediately, with the items men-
tioned above, to the nuclear medicine suite where subjects were
imaged using a rectangular field-of-view gamma camera with a
low-energy general purpose collimator.
Data Analysis
Sleep stages were defin ed in accord with standard critelia.
Sleep efficiency was calculated as the time asleep divided by the
total time spent in bed. Apnea was defined as the absence of
airflow signals from the thermistor or the capnograph for > 10 s;
hypopnea was defined as a reduction in these qualitative airflow
signals when associated with a decrease in arterial oxygen
saturation of 4%. The apnea-hypopnea index could thus be
calculated as the total number of these events occurring pe r hour
of sleep. Similarly, the arousal and awakening index was calcu-
lated as the total number of arousals and awakenings occurring
per hour of sleep.
Body position duling sleep was determined from the sleep
posture monitor signal that was scored by 30-s epochs as either
prone, supine, or on the right or left side. Thus, body position
could be quantitated for desired intervals of the night for
correlation with sites of radionuclide aspiration into the pulmo-
nary parenchyma.
Radionuclide scans were interpreted by a single radiologist
with specific expertise in interpretation of nuclear studies, but no
knowledge of the specific hypotheses being tested. To be inter-
preted as a positive (abnormal) scan, radioactivity had to be
evident in the thorax, away from midline GI structures, in at least
two separate views. The volume of radioisotope suspension
aspirated was calculated from the number of counts measured in
lung parenchyma, the half-life of 99 mTc, and the known concen-
tration of the radionuclide in the original suspension. The counts
in lung parenchyma were quantitated by employing the geomet-
ric mean method. This method corrects for depth and compen-
sates for attenuation by projecting all radioactivity to the mid-
plane of the chest on both anteroposterior and lateral views.
Statistical analysis was performed using a standard program
(Microsoft-Excel; Roselle, Ill). The valious measures of sleep
quality and quantity were compared for study night 1 vs study
night 2 v.ithin individuals using two-tailed paired t testing. These
same variables were compared for subjects who aspirated com-
pared to those who did not aspirate using two-tailed nonpaired t
testing. To determine if there was any relationship between the
quantity aspirated and the charactelistics of sleep on individual
CHEST/111/5/MAY, 1997 1267
 Table 1-Age, Sleep Variables, and Total Quantity of Colloid Aspirated in All Subjects on Both Study Nights*
Av.aken in~ and Quantity of
Sleep Period, Sleep E ffi ciencv, A)~d!~~-Ik\,~R~sea Arousal In c ex, pe r %Sleep Time Colloid
mm % , per Hou r Hou r % 314 Sleep Supine Aspirated, m L
Subject! I I I I I I I I I I 1 I I
Age, yr "'igh t I "'ight 2 Night 1 :"Jight 2 Night I Night 2 N ig ht 1 ght 2
Ni ~ight I 1\ight 2 Night I Night 2 1\ight 1 Night 2

1/22 416 371 92.7 96.2 1.2 0 14.2 9.8 16.0 26.3 85 74.8 0.019
2125 295 460 54.2 97 0 0 11.7 8.4 21.0 15.7 60.2 72 ..5 0 0
3139 .384 402.5 92.6 93.3 0.2 1.9 13.4 17.1 5.1 5.2 38.3 66.2 0.148 0.044
4/45 341 377 92.3 90.5 19.3 7.9 31.5 24. 9 13.1 .5.3 100 100 O.Oll 0.024
S/34 356.5 362.5 91.6 74.2 0.9 1.3 9.7 26.8 2.8 5.9 37.7 12.3 0 0
6/20 369.5 384 89.9 9 1.5 0 ..5 0.3 15.2 10.4 17.2 14.4 49.3 50.7 0 0
7/57 434 423.5 62.1 75.6 0.9 0.9 27.6 19.7 0 0.6 4.8 30.0 0.031
8/24 343.5 486 85.6 82 2.7 4. 1 16. 1 27.2 8.8 15.3 90.8 72.7 Post
9/32 :390.5 437.5 89 78.2 1.7 3.3 23.5 32.8 10.2 .59.8 36.7
10/42 419 437 99.7 85.7 3.7 0.6 24.9 25.7 8.9 17.1 66.0 20.0 0
Mean SE M 387.9:!:9.2 4 14.2:!: 13.0 85.0:!:4.6 86.4:!:2.7 3. 1:!:1.8 2.0:!:0.8 18.8:!:2.4 19.5:!: 1.8 10.3:!:2.1 10.8:!:2.6 59.2:!:9.0 53.6:!:8.9
34:!:.3.7

*Sleep efficiency time asleep/time in bed (%); % 3/4 sleep-percentage of sleep time spent in stages 3 or 4 sleep.
1
Quantitation lost.

study nights , linear correlation coefficients were calculated. When the five subjects who aspirated on either
Statistical significance was defined as a p value <0.05. night were compared with the five subjects who did
not aspirate, no significant difference was found for
age (37.46.6 years vs 31.63.8 years; p=0.4), sleep
RESULTS period (398.211.8 min vs 401.214.0 min), sleep
Full-night sleep studies were obtained in duplicate efficiency (83.84.7% vs 85.13.1 %), apnea-hypop-
in 10 subjects with a mean interval of 34.03.7 days nea index (3.92.5 events per hour vs 1.20.5
between studies. Individual subject data for sleep events per hour; p=0.32), arousal plus awakenings
efficiency, apnea-hypopnea index, arousal-awaken- index (20.33.0 events per hour vs 19.13.6 events
ing index, percentage of sleep time spent in stages 3 per hour), percent of sleep time spent in stages 3 and
plus 4, and percent of sleep time spent in supine 4 (9.63.5% vs 11.52.7%), or percentage of sleep
position can be found in Table 1 for both nights, As time spent in supine position (66.314.4% vs
can be asce1tained from Table 1, there is no differ- 46.56.6%; p=0.25).
ence apparent for any of these variables comparing To determine if the quantity of radionuclide aspi-
the hvo study nights, corresponding to statistical rated might be associated with measures of sleep
comparison revealing no significant differences. quality, linear correlation coefficients were calcu-
Therefore, individual subject data not pertaining to a lated between the quantity aspirated and the various
specific study night will be expressed as the mean of parameters of sleep quality. No significant correla-
these two nights. tion was found between the quantity aspirated and
Despite the somewhat cumbersome paraphernalia sleep efficiency, apnea-hypopnea index, arousal-
required for monitoring, the subjects slept reason- awakening index, percent time spent in stages 3 and
ably well with a mean sleep efficiency of 85.72.6%. 4 sleep, percentage of time spent in rapid eye
The intranasal infusion of the 99 mTc was tolerated movement sleep, or percentage of sleep time spent
without apparent difficulty, with all subjects receiv- in supine position.
ing the full 10 mL of suspension during the noctur-
Table 2-Subjects, Location, and Quantitation of
nal study period, Tracer in Lung
Of the 10 subjects studied, 3 had the technetium
tracer evident in lung parenchyma following the first Quantity of
sleep study, and 4 had tracer in the lung following Subject Night Location of Aspirated A spirated
No. Aspirated Tracer Tracer, mL
the second study. Because two subjects had radio-
isotope apparent in lung parenchyma following both 1 2 L midlung 0.019
3 1 R upper lung 0.129
study nights, a total of 5 of the lO subjects aspirated
89 L midlung 0.019
mTc colloid on one study night or the other. The 2 R upper lung 0.024
quantity of radioisotope suspension aspirated, and 2 L midlung 0.020
the location in the lung where it was found, can be 4 1 L midlung 0.011
seen in Table 2. The lung scans obtained in two of 2 R upper lobe 0.023
the patients who aspirated are depicted in Figures 1 7 1 L upper lobe 0.031
8 1 Data lost Data lost
and 2.

1268 Clinical Investigations

 First Study

t ' '
, .. .
" L,t 1 'I
~ " '
1 .,.\~~''" J
'' ' I
, .. ) (

...' ',;,_:
'
.
- II ~L
:.

Anterior Posterior

Second Study
F IGU RE 1 . Shown are th e ante ri or and posterior lung images obtained on study night l (top ) and study
night 2 (bottom ) for subject 4 who aspirated into th e le ft midlung fi eld on ni ght l and the right midlung
fi eld on ni ght 2. M=a cobalt marker at th e sternal notch; S=trace r accumulation in the stom ach;
E=tracer activity in th e esophagus. The regions of tracer aspiration are ci rcled for quantitation.

Two of the five subjects aspirated tracer on both
study nights. One of these two aspirated unilaterally
on both nights: on the le ft during study night 1, and
on the right during study night 2. The second subject
who aspirated on both nights aspirated bilaterally on
both occasions. Analysis of demographic and sleep
study data again failed to identify any features
unique to these two subjects that might suggest why
aspiration occurred on both study nights (Table 1).
DISCUSSION
In this study, we have shown that a suspension of
99
mTc-labeled sulfur colloid, which is slowly admin-
istered through the nose during nocturnal sle ep, can
be found in the pulmonary parenchyma by standard
scanning techniques the following morning in 50% of
normal subjects studied on two separate nights . The
presence of aspiration in individual subjects on one
study night does not indicate that aspiration deter-
mined by this technique will be observed on addi-
tional study nights, as three of the five aspirating
subjects aspirated on one night only. Considering
age, sleep quality, degrees of sleep-disordered
breathing, and body position, no factors can be
identified that predict which subjects will aspirate on
one or both study nights.
The proportion of normal subjects aspirating dur-
ing sleep is comparable to that observed b yHuxley et
aF who found 9 of 20 normal subjects to aspirate
when studied on a single occasion. Huxley and
coworkers also suggested that all of their subjects
who slept soundly appeared to aspirate, with the
restless sleepers not aspirating; these observations
were made qualitatively without the benefit of sleep
staging. We were unable to confirm these observa-
tions with simultaneous sleep stage monitoring as no
measure of sleep efficiency or tim e spent in any
particular sleep stage appeared to either favor or
protect from nocturnal aspiration.

CHEST / 111 / 5 / MAY, 1997 1269

 First Study

Anterior Posterior

Second Study
F IGURE 2. Shown are an te lio r and posterior lung mages i obtained on study nights l an d 2 fo r subj ect
3 w ho aspirated bilaterally on both study nights. Image labeli ng is the same as Figme l. A large tracer
signal i sevident in th e righ t micllung field on stu dy ni ght l. Regions of isotope aspi ration are not icrcled
in thi s particular exampfe .

In contrast, Kikuchi et aP found only 1 of 10
control subjects to aspirate radioactive tracer during
nocturnal behavioral sleep. These authors postulated
that the technique of injecting a 1-mL solution of
radionuclide into the nose over 1 to 2 min, at 30-min
intervals, as employed b y Huxley and coworkers,2
may have produced an overestimate of the preva-
lence of nocturnal aspiration and excessive sleep
disruption. Kikuchi et aP endeavored to overcome
these limitations by fixing a gauze packet containing
a mixture of 1 11 In chloride, saline solution, and
cellulose powder to the subject's te eth with dental
adhesive shortly before the subject retired to sleep in
a hospital room. Because sleep is not documented b y
polygraphic variables, and the influence of either
technique on nocturnal aspiration is not known, it is
impossible to know if these different techniques
account for the different results in these two studies.
Other differences possibly accounting for the rela-
tively low (10%) incidence of nocturnal aspiration
found by Kikuchi and coworkers are the relatively
older and sicker p atients in their study group. As
both increasing age5 and chronic medical illness 6
reduce sleep quality and duration, it is possible that
these patients slept much more poorly than either
the normal, younger subjects of Huxley et al or our
own, with not enough sustained sleep occurring to
allow aspiration to occur.
It is unlikel~ that our choice of 99 mTc sulfur colloid
rather than 1 1In chloride as a radioactive tracer
accounts for any differences between our results and
those of previous investigations. We chose this com-
pound for several properties that made it pa1ticularly
suitable for this study. First, it is insoluble in water
and resists absorption through mucous membranes
or the GI tract, properties that in part account for its
approval b y the US Food and Drug Administration
for oral administration. Second, these same proper-
ties also assure that any radionuclide found in the
lung parenchyma had to have been aspirated into

1270 Clinical Investigations

that location. There is no real possibility that the
technetium suspension was absorbed into the sys-
temic circulation and then transported and somehow
lod@ed in pulmonary tissue. Finally, the 6-h half-life
of 9 mTc allowed us to perform a second study 8 days
following the first with no possibility of persistent
pulmonary parenchymal radioactivity confounding
our results.
Several limitations of the present work need to be
acknowledged. First, our data do not show that
aspiration definitely occurs while the subjects are
asleep . This would require "real-time" scanning dur-
ing sustained, EEG-proven sleep. Alternatively, the
study design could control for sleep, a measure that
would require an additional night or two of study
wherein the subjects were kept awake throughout
the night but otherwise studied in a similar fashion.
Uniformly normal morning scan results follovving a
full night of supine wakefuln ess, however, would still
not establish that an abnormal scan follovving a night
of sleep is a result of aspiration during sleep. Real-
time scanning during sleep would still be necessary,
and this is beyond our current technical capacity.
Second, the technique we employed to stain the
upper ainvay secretions with radionuclide may have
contributed to the aspiration we observed. To mini-
mize this possibility, we modified the technique of
Huxley and coworkers 2 by introducing the radioac-
tive tracer by continuous infusion at the rate of 2
mUh rather than the bolus injections of 1 mL every
30 min employed in their study. Whether this change
was effective or if either technique of introducing
this tracer into the nasopharynx influenced nocturnal
aspiration, we cannot determine. 'Ve know of no
method for introducing a tracer to identify nocturnal
aspiration in normal subjects that does not introduce
this possibility.
These data leave several questions unanswered .
What is the mechanism of aspiration during sleep? It
is knovm that the sleep state impairs the cough
reflex. Arousal from sleep is required before cough
or swallowing will occur in response to upper ainvay
stimuli.7 We speculate that aspiration occurs pas-
sively during sleep as a result of this sleep-induced
impairment of mechanical ainvay defenses. An at-
tractive corollary to this hypothesis is that factors that
tend to promote sleep, or increase the threshold for
arousal from sleep, would predispose to aspiration.
In support of this mechanism , Huxley and cowork-
ers2 observed that aspiration appeared to correlate
with sound sleep: the subjects who slept most
soundly appeared most apt to aspirate and vice versa.
In addition, 70% of their patients with reduced
consciousness, and presumably s le ep which was
more sustained than normal, were found to aspirate
during the nocturnal sleep period. However, our
own data, the first (to our knowledge ) to document
sleep objectively and then determine if aspiration has
occurred, fail to support the hypothesis that normal
uninterrupted sleep is a 1isk factor for nocturnal
aspiration . It is notabl e in this regard that the
arousal/awakening index (number of arousals and
awakenings per hour of sleep ) was higher than would
be expected in normal subjects, 8 presumably due to
the nasal catheter and intranasal infusion. These
frequent arousals from sleep most likely account for
the relatively low percentage of sleep time these
subjects spent in the deeper sle ep stages (stages 3/4).
This underrepresentation of uninterrupted "deep
sleep" may have reduced the quantity of aspiration
we observed, in accord with the above hypothesis.
However, the lack of a positive correlation between
the percentage or duration of stages 3/4 sleep and
the magnitude of aspirated tracer in our subjects
suggests that these two factors are unrelated, regard-
less of sleep quality. Whether our findings were
altered by relatively poor sleep quality we cannot
know. In addition, we can offer no insight pertaining
to pathologic or drug-induced states of impaired
consciousness on nocturnal aspiration as we included
no such patients in our study.
What is the significance of aspiration of a radioactive
tracer introduced into the upper ai1way during sleep?
This event is presumed to reflect the physiologic
aspiration of pharyngeal ainvay contents during sleep in
normal subjects, an event which is in tum believed to
lead to pulmonary infection with organisms known to
be normal colonizers of the upper ainvay (mouth, nose,
and oronasopharynx). Although we can provide no
direct evidence to support or refute this general hy-
pothesis, our quantitative scanning data do provide an
estimate of the volume of material aspirated from
which some deductions can be made.
Extrapolating from the number of counts found in
each hemithorax into which aspiration occurred, the
volume of aspirated 99 mTc sulfur-colloid suspension
was calculated. As can be seen in Table 1, the order
of magnitude of aspirate ranged from 1.1 to
12.9X 10- 2 mL. This calculation assumes that only
the tracer material, vvith no uppe r ainvay secretions,
is aspirated. Although we believe that the trace r
surely stains upper ainvay secretions and is therefore
a marker for a larger actual volume of aspirated
material, we know of no available information that
provides the basis for the a n estimation of the
volume of nocturnal nasal or salivary gland secre-
tions. As a result, we have no way of determining a
factor of tracer dilution in these subjects, and the
magnitude of aspiration calculated is considered a
minimum value. Upper ainvay secretions in normal
subjects contain on the order of 106 aerobic organ-
isms (mostly streptococci) and 108 anaerobic organ-
CHEST I 111 I 5 I MAY, 1997 1271
isms 9 per milliliter. Therefore, the minimum quantity
of potentially pathogenic organisms aspirated into the
lung in the current study is on the order of 104 to 105 .
The clinical significance of this bolus of microorgan-
isms aspirated into the lung in man is unknown but
undoubtedly relates to the pathogenicity of the organ-
isms aspirated. There are some data from experimental
animal work that suggest that this order of magnitude
of aspiration can be significant. For example, if 105
Staphylococcus aureus organisms are introduced into
the distal airway of a mouse, they are promptly cleared
by host macrophages; 106 microorganisms introduced
in a similar fashion will provoke a neutrophil response,
suggesting that the host macrophage system is over-
whelmed, and a granulocytic response is required to
contain the organisms; 10 108 organisms introduced
identically will cause pneumonia, sepsis, and death
within 12 h. In a different model, the intratracheal
instillation of > 3 X 103 Klebsiella pneunwniae uni-
formly produce K pneumoniae pneumonia and death in
squirrel monkeys. 11 There are obviously a variety of
factors that determine the clinical significance of the
occult aspiration we have observed in this study. Con-
sidering all available infonnation, we find no reason to
dissent from the conclusion that occult aspiration oc-
curring during sleep is likely to be a mechanism that
introduces organisms into the distal airway, which
under some circumstances produces pneumonia in
human subjects.
1272
REFERENCES
1 Gleeson K, Reynolds HY. Life-threatening pneu monia. Clin
Chest Med 1994; 15:581-60'2
2 Huxley EJ, Viroslav J, Gray WR, eta!. Pharyngeal aspiration
in normal subjects and patients with dep ressed consciousness .
Am J M ed 1978; 64:564-86
3 Kikuchi R, Watabe N, Konno T, eta!. High incidence of silent
aspiration in elderly patients with community-acquired pneu-
monia. Am J Respir Crit Care Med 1994; 150:'251-53
4 Hechtschaffen A, Kales A. A manual of standardized termi-
nology, technique and scoring system for sleep stages of
human sleep. Los Angeles: Brain Information Service, Brain
Information Institute, University of California, 1968
5 Carskadon MA, VanDen Hoed J, Dement WC. Insomnia and
sleep disturbances in the aged: sleep and daytime sleepiness
in th e elderly. J Geriatr Psychiatty Neurol 1980; 13:135-51
6 Krachman SL, D'Alonzo GE, Criner GJ. Sleep in the inten-
sive care unit. Chest 1995; 107:1713-'20
7 Sullivan CE, Murphy E, Kozar LF, et a!. Waking and
ventilatory responses to laryngeal stimulation in sleeping
dogs. J Appl Physiol 1978; 45:681-89
8 Mathur R, Douglas NJ. Frequency of EEG arousals from
nocturnal sleep in normal subjects. Sleep 1995; 18:330-33
9 Toews GB. Nosocomial pneumonia. Clin Chest Med 1987;
8:467-79
10 Onofrio JM , Towes GB, Kipscomb MF, e t a!. Granulocyte-
alveolar-macrophage interaction in the pulmonary clearance
of Staphylococcus aureus. Am Hev Respir Dis 1983; 127:
335-41
11 Berendt HF. Helationship of method of administration to
respiratory virulence of Klebsiella pneumoniae for mice and
squirrel monkeys. Infect Immun 1978; 20:581-87
Clinical Investigations