Am J Respir Crit Care Med
Internet address: www.atsjournals.org
Vol 164. pp 698–703, 2001
The anatomical relationships between lymphoid, bony, and other
tissues affecting the shape of the upper airway in children with
obstructive sleep apnea syndrome (OSAS) have not been estab-
lished. We therefore compared the upper airway structure in 18
young children with OSAS (age 4.8
males) and an apnea index of 4.3
subjects (age, 4.9
2.0 yr; 12 males and 6 females). All subjects
underwent magnetic resonance imaging under sedation. Axial and
sagittal T1- and T2-weighted sequences were obtained. Images
were analyzed with image-processing software to obtain linear,
area, and volumetric measurements of the upper airway and the
tissues comprising the airway. The volume of the upper airway was
smaller in subjects with OSAS in comparison with control subjects
0.8 versus 2.5
tonsils were larger (9.9
3.9 and 9.1
2.2 cm; p
0.005 and p
umes of the mandible and tongue were similar in both groups;
however, the soft palate was larger in subjects with OSAS (3.5
1.2 cm; p
0.05). We conclude that in children with
moderate OSAS, the upper airway is restricted both by the ade-
noid and tonsils; however, the soft palate is also larger in this group,
adding further restriction.
3.9, with 18 matched control
2.1 yr; 12 males and 6 fe-
0.005) and the adenoid and
2.9 cm versus 6.4
0.0005, respectively). Vol-
nance imaging (MRI)
obstructive sleep apnea syndrome (OSAS); magnetic reso-
Obstructive sleep apnea syndrome (OSAS) is estimated to af-
fect 2% of children (1). The disorder most commonly occurs
between 2 and 6 yr of age and correlates with lymphoid hyper-
plasia during childhood (2, 3). Other important factors pro-
moting OSAS in children include craniofacial anomalies and
neuromuscular disorders that could affect the size, shape, and
collapsibility of the upper airway during sleep.
Traditional methods used in children to assess the upper
airway and tissues surrounding the airway include lateral neck
radiographs and cephalometry measurements (2–6). How-
ever, these methods are essentially two-dimensional and pro-
vide little information about lateral structures of the nasopha-
ryngeal and oropharyngeal regions. Newer techniques such as
acoustic reflection, ultrafast computed tomography, and mag-
netic resonance imaging (MRI) are readily available, and are
used in adults for more detailed analysis of the upper airway
structure (7–14). Of the above-cited methods, MRI provides
superior resolution of tissues composing the upper airway; it is
accurate, reproducible, and free of ionizing radiation. How-
ever, MRI has not been used in the past to systematically eval-
uate the upper airway in children with OSAS.
We hypothesized that the upper airway as assessed by MRI
in children with OSAS is smaller because of adenoid and/or
tonsillar overgrowth in comparison with control subjects. To
this end we quantified total upper airway volume and cross-
sectional area at expected sites of upper airway restriction in
the nasopharyngeal and oropharyngeal regions. In addition, we
measured the sizes of soft tissues and skeletal structures that
define the upper airway. Finally, we validated our methods by
examining the accuracy and reliability of these measurements.
Subjects with OSAS
Eighteen children were recruited from the pool of patients evaluated
for sleep-disordered breathing at the Children’s Hospital of Philadel-
phia (Philadelphia, PA). After OSAS was confirmed by polysomnog-
raphy, patients were allowed to undergo MRI of the upper airway un-
der sedation. The study was approved by the Institutional Review
Board of the Children’s Hospital of Philadelphia. Informed consent
was obtained from parents of the subjects.
Eighteen children with normal growth and development were
matched to subjects with OSAS by age, sex, ethnicity, weight, and height.
Control subjects were selected from patients who underwent head
MRI at the Children’s Hospital of Philadelphia for other medical indi-
cations. Exclusion criteria included (
assessed by a standard questionnaire,
tumor or a seizure disorder requiring therapy, (
sociated with any craniofacial anomaly, and (
ease such as asthma or bronchopulmonary dysplasia.
) subjects with scores
) genetic disorders as-
) chronic respiratory dis-
) evidence of a brain
For subjects with OSAS, polysomnography was performed 0–4 wk be-
fore MRI. Subjects were studied in the Sleep Disorders Center at the
Children’s Hospital of Philadelphia. Scoring of respiratory variables
was performed on the basis of standards set by the American Thoracic
Society and previous published data about children (15, 16; details in
online data supplement). Sleep stages were determined by the criteria
of Rechtschaffen and Kales (17).
A questionnaire regarding symptoms of sleep-disordered breathing,
based on the questionnaire developed by Brouillette and coworkers
(18), was used to assess the likelihood of OSAS in control subjects
and subjects with OSAS. On the basis of the questionnaire no subject
with a score
1 would be expected to have OSAS; a score between
1 and 3.5 is considered indeterminate, and a score
ered highly predictive of obstructive sleep apnea.
3.5 is consid-
Supported by grants HL-62408, HL-60287, and MO1-RR00240 from the Na-
tional Institutes of Health.
Correspondence and requests for reprints should be addressed to Raanan Arens,
M.D., Division of Pulmonary Medicine, Children’s Hospital of Philadelphia, 34th
Street and Civic Center Blvd., Philadelphia, PA 19104-4399. E-mail: arens@
This article has an online data supplement, which is accessible from this issue’s
table of contents online at www.atsjournals.org
Received in original form January 25, 2001 and in revised form April 18, 2001
Magnetic Resonance Imaging of the Upper Airway
Structure of Children with Obstructive Sleep
RAANAN ARENS, JOSEPH M. McDONOUGH, ANDREW T. COSTARINO, SOROOSH MAHBOUBI,
CATHERINE E. TAYAG-KIER, GREG MAISLIN, RICHARD J. SCHWAB, and ALLAN I. PACK
Divisions of Pulmonary Medicine, Anesthesia and Critical Care, and Radiology, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania;
and Center for Sleep and Respiratory Neurobiology, University of Pennsylvania Medical Center, University of Pennsylvania School of Medicine,
Arens, McDonough, Costarino,
: MRI of Upper Airway in Children with OSAS
Magnetic Resonance Imaging
MRI studies were performed in the Department of Radiology at the
Children’s Hospital of Philadelphia. All studies were performed un-
der sedation with intravenous pentobarbital (2–6 mg/kg), adminis-
tered in increments until sleep was achieved; a maximum of 200 mg
was administered. All subjects were monitored continuously by pulse
oximetry and observed by a critical care physician until recovery (
MRI was performed with a 1.5T Siemens (Iselin, NJ) Vision system.
Images were acquired with a commercially available head coil. The pa-
tient’s head was positioned supine in the soft tissue Frankfort plane. Ax-
ial and sagittal sequential T1-weighted (TR [repetition time], 650 ms; TE
[echo time], 14 ms) and T2-weighted (TR, 6,000 ms; TE, 90 ms) images
with 3-mm slice thickness and 1 NEX (number of excitations) were ob-
tained from the orbital cavity to the larynx and from the midline bilater-
ally, respectively (
online data supplement for MRI sequences).
Image Processing and Anatomic Measurements
Measurements from MR images were made with image processing soft-
ware (Volumetric Image Display and Analysis [VIDA]; Department of
Radiology, University of Iowa, Ames, IA) (9, 10, 13). Airway, soft tis-
sue, and bony structure segmentation was performed by manual tracing.
From an axial T1-weighted image at the level of the maximal tonsillar
cross-sectional area (Figure 1A), we determined the cross-sectional
area of the oropharyngeal airway, tonsils, pterygoids, and parapha-
ryngeal fat pads. The oropharyngeal airway was defined radiologically
as being bounded anteriorly by the soft palate or tongue, laterally by
the tonsils, and posteriorly by the pharyngeal constrictor muscle. At
this level we also performed a series of linear measurements along a
transverse line crossing the center of the tonsils (e.g., intertonsillar
width, bilateral tonsillar width, bilateral fat pad width, bilateral ptery-
goid width, and intermandibular distance). In addition, we measured
at other axial levels the largest adenoid cross-sectional area, maxilla
width, and the internal distance between mandibular heads.
From a midsagittal T1-weighted image (Figure 1B), we determined
the cross-sectional area of the nasopharyngeal airway, adenoid, soft
palate, tongue, mandible, and hard palate. The nasopharyngeal air-
way was defined radiologically as being bounded anteriorly by the
vomer, posteriorly by the adenoid, and inferiorly by a horizontal line
above the hard and soft palate. In addition, we obtained a series of lin-
ear measurements along an oblique line connecting the mental spine,
the center of the soft palate, and the clivus (i.e., tongue oblique, soft
palate oblique, airway oblique, adenoid oblique, and mental spine-cli-
vus oblique). We measured the length of the hard palate, defined
from anterior nasal spine to end of the palatine bone.
Adjacent axial slices were used to determine the volumes of the follow-
ing: combined upper airway (nasopharynx and oropharynx), adenoid,
tonsils, tongue, soft palate, and mandible. Volumetric measurements
(except tonsils and adenoid) were made from axial T1-weighted slices
spanning from the base of the orbital cavities to the epiglottis. For
measurements of adenoid and tonsils we used T2-weighted images be-
cause of their better resolution of lymphoid tissue. Tonsillar volume is
the combined volume of right and left tonsils.
online data supplement.) Accuracy of measurements was deter-
mined by using a set of commercial phantoms for lengths, areas, and
volumes spanning the measurements used in the study. Intraclass cor-
relation coefficient (ICC) was computed to assess the reliability of
MRI measurements (19, 20).
We hypothesized that children with OSA have a smaller upper airway
in comparison with matched control subjects. Therefore, the primary
outcome variable set was defined as (
tional airway area at the level of the maximal tonsillar cross-sectional
) midsagittal nasopharyngeal cross-sectional airway area, and (
upper airway volume. Secondary outcome measures included lengths,
areas, and volumes of the tissues surrounding the upper airway and
were considered hypothesis generating.
Means and standard deviations were used to summarize continu-
ous variables. For comparisons between the groups for MRI data, de-
mographics, and questionnaire data, we used a two-tailed unpaired
test, the Wilcoxon rank test, or
correlation was used to assess the linear correlation of volume percent
difference of each OSAS–control pair and the corresponding apnea–
hypopnea index. A p value
0.05 was considered significant.
) axial oropharyngeal cross-sec-
test as appropriate. The Pearson
We studied 18 children with OSAS, (mean age, 4.8
range, 1.9–9.3 yr) and 18 control subjects (mean age, 4.9
range 1.8–8.7 yr). Children with OSAS were not significantly dif-
ferent from control subjects with respect to age, sex, ethnicity,
height, or weight (Table 1). All controls had normal develop-
ment and cognitive function, intact tonsils and adenoid, and no
respiratory disorders or craniofacial anomalies. The primary in-
dications for head MRI in control subjects were as follows: sin-
gle seizure/febrile convulsion (10 subjects), migraine/headache
(4 subjects), head concussion (2 subjects), torticollis (1 subject),
and eye injury (1 subject). Thus, none of these clinical indica-
tions would be expected to affect the upper airway anatomy.
For subjects with OSAS the mean total sleep time during poly-
somnography was 7.3
1.0 h. The mean respiratory variable
values during this period were as follows: apnea index, 4.3
3.9; apnea–hypopnea index, 11.2
6.8; baseline arterial oxy-
Figure 1. (A) Anatomical outlines
of an axial T1 image at the level of
the maximal tonsillar area of a con-
trol subject. Transverse black arrow
represents the intermandibular dis-
tance. (B) Anatomical outlines of a
midsagittal T1 image of a control
subject. Oblique black arrow repre-
sents the mental spine–clivus oblique
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 1642001
gen saturation (Sa
these data suggest moderate OSAS in this group.
1%; and Sa
All controls had an apnea score
obstructive sleep apnea in this group (18), and as a group had
a mean apnea score of
0.8. This score was significantly
lower than the apnea score noted in the OSAS group, of 3.2
1, indicating absence of
Magnetic Resonance Imaging
The amount of sedation given before MRI was similar to both
groups. A mean of 3.8
1.1 and 4.0
per kilogram was given to subjects with OSAS and control sub-
jects, respectively. Representative midsagittal and axial images
of a control subject and a subject with OSAS are presented in
Figure 2. The MRI measurements of the various tissues form-
ing the upper airway of control subjects and children with OSAS
are presented graphically in Figures 3–5.
1.2 mg of pentobarbital
was significantly smaller in comparison with the control group,
0.8 versus 2.5
1.2 cm (p
tal nasopharyngeal airway cross-sectional area was smaller in
the children with OSAS, 0.6
0.005), as was the axial oropharyngeal cross-sectional airway
area at the level of maximal tonsillar cross-sectional area, 0.4
0.3 versus 0.7
0.5 cm (p
. The cross-sectional areas of the pterygoids, para-
pharyngeal fat pads, tongue, and soft palate obtained from the
axial and midsagittal images are presented along with the vol-
umetric measurements of the tongue and soft palate in Figure
3. The axial and sagittal cross-sectional areas of the soft tissues
were similar in both groups, as was the volume of the tongue.
However, we noted a significantly larger soft palate volume in
the OSAS group, 3.5
1.1 versus 2.7
Adenoid and tonsils
. As can be noted from all the measure-
ments obtained in Figure 4, both the adenoid and tonsils were
significantly larger in children with OSAS. The mean adenoid
volume in children with OSAS was 9.9
son with 6.4
2.3 cm in control subjects (p
mean tonsillar volume in children with OSAS was 9.1
2.2 cm in control subjects (p
Facial skeletal structure
. Figure 5 demonstrates the skeletal
findings of our study. The mandible width, length (as implied
by mental spine–clivus distance), and volume were similar in
both groups. Similarly, maxilla width, hard palate length, and
sagittal cross-sectional area did not differ between groups.
Thus, the similar skeletal findings in this set of measurements
do not suggest a primary skeletal difference between subjects
with OSAS and control subjects.
. The upper airway volume of children with OSAS
0.005). Similarly, midsagit-
0.3 versus 1.1
TABLE 1. DEMOGRAPHICS*
Subjects with OSAS
* Values represent means
Figure 2. T1-weighted midsagittal
and axial images of a control sub-
ject (C) and a subject with OSAS
(OSAS). Note in subjects with OSAS
the narrow nasopharyngeal (top right)
and oropharyngeal (lower right) air-
Arens, McDonough, Costarino,
: MRI of Upper Airway in Children with OSAS
Linear correlation of volume percent difference and apnea–
. Linear correlations between a volume per-
cent difference of each matched OSAS–control pair and ap-
nea–hypopnea index are presented in Figure 6A and 6B. We
noted no correlation between percent difference in airway
volume and apnea–hypopnea index (Figure 6A), and a posi-
tive correlation between the percent difference of the com-
bined tonsil and adenoid volume and apnea–hypopnea index
0.03; Figure 6B).
Validation of Methods
Analysis of data obtained from phantom MRI measurements
indicated, in general, that accuracy increases with the size of
the object measured. The ICC values for 13 representative pa-
rameters obtained include 1 measurement with a value of 0.78,
indicating substantial reliability, and 12 with values greater
than 0.88, indicating almost perfect reliability (20).
We used MRI to study the upper airway and surrounding tis-
sues as well as the lower face skeletal structure in young chil-
dren with OSAS. This is the first study of children with OSAS
to use MRI to delineate the airway in detail, the tissues sur-
rounding the airway, and the three-dimensional relationships
of these tissues. Our volumetric measurements indicate that
the adenoid and tonsils as well as the soft palate are signifi-
cantly larger in children with OSAS concomitant with signifi-
cantly smaller upper airway volume.
OSAS is common in children, with a prevalence of about 2%
(1). OSAS may occur in all age groups. However, the peak in-
cidence occurs between 2 and 6 yr of age, concomitant with rapid
growth of the lymphoid tissue during these years (3). Diagnosis
and management of OSAS are important because untreated
OSAS may lead to altered neurodevelopmental performance
(21–23), growth retardation (21, 24), pulmonary hypertension
(25), and corpulmonale (21, 26).
Our validation studies show that MRI provides both highly
accurate and reliable measurements of the various tissues com-
posing the upper airway. However, a disadvantage of MRI
compared with ultra-fast computed tomography (CT) scan is
its relatively long acquisition time (e.g., 2 min versus 10 s for
each sequence) and therefore its sensitivity to motion artifact.
To minimize movement during MR scanning, light sedation is
given routinely to all children in our institution younger than
8 yr of age. We are aware that sedation could have reduced
muscle tone impacting on airway measurements in our sub-
jects. We assume that because the same sedation protocol was
Figure 3. Comparison between soft tissue measurements of OSAS and
control groups. Measurements include combined pterygoid axial
cross-sectional areas (Pteryg. Axial), combined parapharyngeal fat pad
axial cross-sectional areas (Fat pad Axial), tongue midsagittal cross-sec-
tional area (Tongue Sag), soft palate midsagittal cross-sectional area
(SP Sag), tongue volume (Tongue Vol), and soft palate volume (SP Vol).
*p ? 0.05.
Figure 4. Comparison of adenoid and tonsil measurements of OSAS
and control groups. Measurements include maximal axial adenoid
cross-sectional area (AD Axial), adenoid midsagittal cross-sectional area
(AD Sag), tonsil axial cross-sectional area (Ton Axial), adenoid volume
(AD Vol), and tonsil volume (Ton Vol).
*p ? 0.005; **p ? 0.0005.
Figure 5. Comparison of facial skeletal measurements of OSAS and
control groups. Measurements include intermandibular head width
(Mand width), mental spine–clivus distance (Mental S-Clivus), mandible
volume (Mand Vol), maxilla width (Maxil width), midsagittal hard pal-
ate length (HP length), and midsagittal hard palate cross-sectional area
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 1642001
used for both groups, airway measurements would be affected
in a similar fashion. Sedation should not have affected the vol-
umetric measurements of muscle, lymphoid, and bone tissues.
To obtain the most reliable comparison, we performed a
case control study and matched subjects by age, height, weight,
sex, and ethnicity. All of these may influence the anatomy of
the structures studied, especially in respect to skeletal growth
(3). On the basis of volumetric measurements of the upper air-
way and cross-sectional area measurements of the midsagittal
and axial images, we conclude that the upper airway in chil-
dren with moderate OSAS is significantly restricted in both
the nasopharyneal and oropharyngeal regions. The main fac-
tors contributing to this restriction are oversized adenoid and
tonsils, which were noted to be 55 and 58% larger in volume,
respectively, in comparison with control subjects.
Moreover, the positive correlation between the percent dif-
ference of the combined tonsil and adenoid volume and the
apnea–hypopnea index found in the present study suggests
that volumetric measurements of these tissues may be a useful
way to predict severity of OSAS in these subjects. We found,
however, no correlation of percent difference in airway vol-
ume alone with severity of OSAS. It is possible that airway
shape, rather than simply volume, is a stronger predictor in
determining airway resistance and airway collapsibility.
We also noted the volume of the soft palate to be about
30% larger in subjects with OSAS, adding additional restric-
tion to the airway. This finding, not previously reported in
children, has been observed in adults with OSAS (27–29). His-
tological evaluation of the soft palate of adults who underwent
uvulopalatopharyngoplasty for OSAS indicated significant in-
flammatory changes and edema (27, 28). In another study,
adults treated for OSAS with continuous positive airway pres-
sure showed significant reduction in soft palate volume mea-
sured by MRI after several weeks of therapeutic use (29). In-
creased volume of the soft palate, hypothesized to be caused by
chronic vibratory injury due to snoring and leading to edema
and inflammation in adults, may be true in children as well.
Other soft tissues (excluding the tonsils) contributing to the
width of the lateral pharyngeal wall, such as the pterygoid
muscles, pharyngeal constrictor muscles, and parapharyngeal
fat pads, were similar in size in both groups studied. These tis-
sues did not contribute to airway restriction by thickening the
lateral pharyngeal wall as reported in adults with OSAS (13).
It is possible that these structures develop as risk factors later
in life and contribute to the adult form of OSAS.
Several studies applying cephalometry to children with
OSAS reported various facial skeleton differences, including
retroposed mandible and maxilla, low positioned hyoid bone,
and alterations in face length and width (30–33). On the basis
of these, it has been suggested that OSAS may be caused by a
combination of both adenotonsillar hypertrophy and abnor-
malities in the development of the lower face skeleton (32, 33).
However, another prospective study showed that most cepha-
lometric differences found in children reversed soon after ade-
notonsillectomy, suggesting that some of these differences de-
tected by cephalometry may be secondary or compensatory
and not a primary cause of OSAS (34).
Our set of skeletal measurements of children with no ap-
parent craniofacial anomaly does not suggest a primary skele-
tal difference between our OSAS and control groups. The rea-
son we did not find such a difference in our subjects can only
be speculated at this time. We used different techniques and
made measurements different from those performed in cepha-
lometry, including lateral measurements obtained by axial im-
ages. Our subjects were sedated and in the supine position dur-
ing imaging, compared with the upright awake subject studied
by cephalometry. It is conceivable that an altered mandibular–
skull or hyoid–tongue orientation detected in subjects in the
upright postion by cephalometry is positional in nature and
could not be confirmed in our study of children in the supine
position. Finally, differences in skeletal measurements may be
related to subject-control matching criteria. It would be help-
ful in future to analyze cephalometric and MR images simulta-
neously to draw more definite conclusions.
In summary, we used MRI to study the upper airway in nor-
mal children and in children with moderate OSAS. We noted
no significant lower face skeletal differences between these
groups. The upper airway was restricted in children with OSAS
by overgrowth of the adenoid and tonsils as well as a larger
soft palate. Our data suggest a positive correlation between
the percent difference in adenoid and tonsil volumes (OSAS
to control) and the apnea–hypopnea index. Because OSAS
represents a spectrum of disorders with a range of severity it
will be essential in future to determine more precisely the re-
lationships between these structures and the risk for, and se-
verity of, OSAS at a given age.
Acknowledgment: The authors thank the children and families who partici-
pated in the study, and Tamara D. Lee BSRT and Mimi Carroll for perform-
ing the MRI studies.
1. Redline S, Tishler PV, Schluchter M, Aylor J, Clark K, Graham G. Risk
factors for sleep-disordered breathing in children: associations with
obesity, race, and respiratory problems. Am J Respir Crit Care Med
2. Fujioka M, Young LW, Girdany BR. Radiographic evaluation of ade-
noidal size in children: adenoidal–nasopharyngeal ratio. Am J Roent-
3. Jeans WD, Fernando DC, Maw AR, Leighton BC. A longitudinal study
Figure 6. (A) Linear correlation between percent difference in airway
volume of each matched OSAS–control pair and the apnea–hypopnea
index (r ? 0.07). (B) Linear correlation between percent difference in
combined adenoid and tonsil volume of each matched OSAS–control
pair and the apnea–hypopnea index (r ? 0.51; p ? 0.03).
Arens, McDonough, Costarino, et al.: MRI of Upper Airway in Children with OSAS Download full-text
of the growth of the nasopharynx and its contents in normal children.
Br J Radiol 1981;54:117–121.
4. Fernbach SK, Brouillette RT, Riggs TW, Hunt CE. Radiologic evalua-
tion of adenoids and tonsils in children with obstructive sleep apnea:
plain films and fluoroscopy. Pediatr Radiol 1983;13:258–265.
5. Brooks LJ, Stephens BM, Bacevice AM. Adenoid size is related to se-
verity but not the number of episodes of obstructive sleep apnea in
children. J Pediatr 1998;132:682–686.
6. Guilleminault C, Partinen M, Praud JP, Quera-Salva MA, Powell N, Ri-
ley R. Morphometric facial changes and obstructive sleep apnea in
children. J Pediatr 1989;114:997–999.
7. Brown IG, Zamel N, Hoffstein V. Pharyngeal cross-sectional area in
normal men and women. J Appl Physiol 1986;61:890–895.
8. Fredberg JF, Wohl MEB, Glass GM, Dorkin HL. Airway area by acous-
tic reflections measured at the mouth. J Appl Physiol 1980;48:749–758.
9. Schwab RJ, Gefter WB, Pack AI, Hoffman EA. Dynamic imaging dur-
ing respiration in normal subjects. J Appl Physiol 1993;74:1504–1514.
10. Schwab RJ, Gefter WB, Hoffman EA, Gupta KB, Pack AI. Dynamic
upper airway imaging during awake respiration in normal subjects and
patients with sleep disordered breathing. Am Rev Respir Dis 1993;148:
11. Ryan CF, Lowe AA, Li D, Fluthom JA. Three dimensional upper airway
CT in obstructive sleep apnea. Am Rev Respir Dis 1991;144:428–432.
12. Burger CD, Stanson AW, Daniels BK, Sheedy PF, Shepard JW. Fast-
computed tomographic evaluation of the effect of route of breathing
on upper airway size and function in normal men. Chest 1993;103:
13. Schwab RJ, Gupta KB, Gefter WB, Hoffman EA, Pack AI. Upper air-
way soft tissue anatomy in normals and patients with sleep disordered
breathing: significance of the lateral pharyngeal walls. Am J Respir
Crit Care Med 1995;152:1673–1689.
14. Schwab RJ, Pack AI, Gupta KB, Metzger LJ, Oh E, Getsy JE, Hoffman
EA, Gefter WB. Upper airway and soft tissue structural changes in-
duced by CPAP in normal subjects. Am J Respir Crit Care Med 1996;
15. American Thoracic Society. Standards and indications for cardiopulmo-
nary sleep studies in children. Am J Respir Crit Care Med 1996;153:
16. Marcus CL, Omlin KJ, Basinski DJ, Bailey SL, Rachal AB, Von Pech-
mann WS, Keens TG, Ward SL. Normal polysomnogram values for
children and adolescents. Am Rev Respir Dis 1992;146:1235–1239.
17. Rechtschaffen A, Kales A. A manual of standardized terminology, tech-
niques and scoring systems for sleep stages on human subjects. Wash-
ington DC: National Institutes of Health; 1968, Publ. No. 204.
18. Brouillette R, Hanson D, David R, Klemka L, Szatkowski A, Fernbach
S, Hunt C. A diagnostic approach to suspected obstructive sleep ap-
nea in children. J Pediatr 1984;105:10–14.
19. Fleiss JK. The design and analysis of clinical experiments. New York:
John Wiley & Sons; 1986, Chapter 1.
20. Landis JR, Koch GG. The measurement of observer agreement for cate-
gorical data. Biometrics 1977;33:159–174.
21. Brouillette RT, Fernbach SK, Hunt CE. Obstructive sleep apnea in in-
fants and children. J Pediatr 1982;100:31–40.
22. Rhodes SK, Shimoda KC, Waid LR, O’Neil PM, Oexmann MJ, Collop
NA, Willi SM. Neurocognitive deficits in morbidly obese children
with obstructive sleep apnea. J Pediatr 1995;127:741–744.
23. Gozal D. Sleep disordered breathing and school performance in chil-
dren. Pediatrics 1998;102:616–620.
24. Marcus CL, Carroll JL, Koerner CB, Hamer A, Lutz J, Loughlin GM.
Determinants of growth in children with obstructive sleep apnea syn-
drome. J Pediatr 1994;125:556–562.
25. Perkin RM, Anas NG. Pulmonary hypertension in pediatric patients. J
26. Menashe VD, Ferrehi F, Miller M. Hypoventilation and cor pulmonale
due to chronic upper airway obstruction. J Pediatr 1965;67:198–203.
27. Sekosan M, Zakkar M, Wenig BL, Olopade CO, Rubinstein I. Inflam-
mation in the uvula mucosa of patients with obstructive sleep apnea.
28. Hamans EP, Van Marck EA, De Backer WA, Creten W, Van de Heyn-
ing PH. Morphometric analysis of the uvula in patients with sleep-re-
lated breathing disorders. Eur Arch Otorhinolaryngol 2000;257:232–236.
29. Ryan CF, Lowe AA, Li D, Fleetham JA. Magnetic resonance imaging of
the upper airway in obstructive sleep apnea before and after chronic
nasal continuous positive airway pressure therapy. Am Rev Respir Dis
30. Zucconi M, Caprioglio A, Calori G, Ferini-Strambi L, Oldani A, Cas-
tronovo C, Smirne S. Craniofacial modifications in children with ha-
bitual snoring and obstructive sleep apnoea: a case-control study. Eur
Respir J 1999;13:411–417.
31. Shintani T, Asakura K, Kataura A. Adenotonsillar hypertrophy and
skeletal morphology of children with obstructive sleep apnea syn-
drome. Acta Otolaryngol Suppl 1996;523:222–224.
32. Kawashima S, Niikuni N, Chia-hung L, Takahasi Y, Kohno M, Nakajima
I, Akasaka M, Sakata H, Akashi S. Cephalometric comparisons of
craniofacial and upper airway structures in young children with ob-
structive sleep apnea syndrome. Ear Nose Throat J 2000;79:499–502,
33. Shintani T, Asakura K, Kataura A. Evaluation of the role of adenoton-
sillar hypertrophy and facial morphology in children with obstructive
sleep apnea. ORL J Otorhinolaryngol Relat Spec 1997;59:286–291.
34. Agren K, Nordlander B, Linder-Aronsson S, Zettergren-Wijk L, Svan-
borg E. Children with nocturnal upper airway obstruction: postopera-
tive orthodontic and respiratory improvement. Acta Otolaryngol 1998;