Three-dimensional Upper Airway Computed Tomography in Obstructive Sleep Apnea A Prospective Study in Patients Treated by Uvulopalatopharyngoplasty1-3
C. FRANCIS RYAN,4 ALAN A. LOWE, DAVID LI, and JOHN A. FLEETHAM
Obstructive sleep apnea (GSA) occurs because of recurrent occlusion of the upper airway during sleep. Upper airway patency depends on a balance of forces. The negative pressure produced when the thoracic inspiratory muscles contract tends to occlude the upper airway unless adequately opposed by the upper airway dilating muscles (1). Any increase in inspiratory resistance due to a narrow airway predisposes to upper airway occlusion during sleep. Computed tomographic (CT) studies have shown that the upper airway is narrower in patients with GSA (2-5).
Uvulopalatopharyngoplasty (UPPP) was proposed as a surgical treatment for GSA by Fujita and associates (6). UPPP enlarges the upper airway by removing excessive distal palatal tissue while preserving the function of the proximal palatal musculature. The success of this operation varies considerably (7-13), and some of this variability has been accounted for by differences in patient selection (9, 11) or surgical technique (10). Computed tomographic studies of the upper airway in GSA have shown that patients who have airway narrowing in the retropalatal airway (14), or narrow tongues (15), are more likely to obtain a good response following UPPP. We performed awake upper airway three-dimensional CT scans on 60 consecutive patients with GSA before UPPP to determine whether upper airway, tongue, or soft palate size predicts the response to UPPP. Methods Patients Patients with OSA (~ 5 apneas/h) and associated daytime symptoms were considered for uvulopalatopharyngoplasty. Before surgery each patient was informed of the differ428
SUMMARY The success of uvulopalatopharyngoplasty in treating obstructive sleep apnea varies considerably. Some of this variability may be accounted for by differences in the site of upper airway narrowing. Todetermine whether preoperative awake upper airway and soft tissue volumes predict the response to uvulopalatopharyngoplasty, preoperative awake computed tomograms (CT) of the upper airway were performed on 60 consecutive patients with symptomatic obstructive sleep apnea. Tracings were made from the CT scans of upper airway, tongue, and soft palate. Computer software was used to determine the cross-sectional area and volume of the upper airway, tongue, and soft palate. Patients underwent overnight polysomnograms before and 3 months after uvulopalatopharyngoplasty. Tongue volume was larger (p < 0.02) and both upper airway to tongue volume (p < 0.0005) and oropharynx to soft palate volume ratios (p < 0.01) were smaller in obese patients. A good response to uvulopalatopharyngoplasty as defined by a postoperative apnea index of < 5 apneas/h or a reduction in apnea index ~ 50% was seen In 50 patients (83%). Patients who had a good response had a smaller oropharyngeal cross-sectional area (p < 0.01),a smaller upper airway volume (p < 0.05), a smaller upper airway to tongue volume ratio (p < 0.01), and a smaller oropharynx to soft palate volume ratio (p < 0.05). Obese patients with obstructive sleep apnea have larger tongues and smaller upper airways relative to tongue and soft palate size. Patients with smaller upper airways, particularly relative to tongue and soft palate size, have a good response to uvulopalatopharyngoplasty. AM REV RESPIR DIS 1991; 144:428-432
ent therapeutic options currently available (weight loss, dental appliances, protriptyline, and nasal continuous positive airway pressure). Patients who underwent UPPP either refused to consider any other treatment or had previously had an unsuccessful trial of another form of treatment. UPPP was not performed in three patients because of severe cardiorespiratory disease and the risk of general anesthesia. There were no other inclusion or exclusion criteria. Sixty consecutive patients who underwent UPPP had a preoperative upper airway CT scan and follow-up polysomnography. The patients were 56 men and 4 women, age 47 ± 1.5 yr (mean ± SEM), body mass index 30.1 ± 0.6 kg/rn". UPPP was performed using the same technique (16)by one of two surgeons within 2 months of the initial polysomnogram. Polysomnography Each patient had an overnight polysomnogram performed before and 3 months after UPPP. Sleep and its various stages was documented by standard electroencephalographic (EEG), electrooculographic (EOG), and electromyographic (EMG) criteria (17). EEG was recorded with electrodes applied at C 3-A2 and CCA 1 (according to the International 10-20 system), and EMG activity was
recorded from the submental muscles. Apneas were defined as cessation of airflow at the nose and mouth for longer than 10sand weredocumented by an infrared CO 2analyzer (Model LB-2; Beckman Instruments, Inc., Schiller Park, IL), which recorded from both the nose and mouth. A single ECG lead (modified V2) was monitored to detect cardiac arrhythmias. Sa02was monitored continuously with a pulse oximeter (Model N-100; Nellcor, Inc., Hayward, CA) attached to the index finger. Chest wall movement was moni(Received in original form March 5, 1990 and in revised form November 30, 1990) 1 FrOl':1 the Departments of Medicine, Clinical Dental Sciences, and Radiology, University of British Columbia, Vancouver, British Columbia, Canada. 2 Supported by the British Columbia Health Care Research Foundation and by Grant No. MA-3849 from the Medical Research Council of Canada. J Correspondence and requests for reprints should be addressed to Dr. John A. Fleetham, Department of Medicine, University Hospital (UBC), 2211 Wesbrook Mall, Vancouver, BC V6T 2B5, Canada. 4 Research Fellow of the British Columbia Lung Association.
3D UPPER AIRWAY SIZE IN SLEEP APNEA TREATED BY UPPP
tored by a respiratory inductive plethysmograph (Respitrace'"; Ambulatory Monitoring Equipment, Ardsley, NY). The data were recorded on a I5-channel polygraph (Model 78; Grass Instruments Co., Quincy, MA) at a paper speed of 10 mm/s. Sleep stage and respiratory variables were analyzed independently. A microcomputer (PC-AT®; IBM, Boca Raton, FL) continuously monitored and stored arterial oxygen saturation, chest wall movement, and heart rate data on a mass storage medium. The entire record was manually scored for sleep stage and apnea type and duration and these data merged with the remainder of the data already stored. Severity of sleep apnea was assessed in terms of total apnea time/total sleep time (TST), number of apnea/TST (apnea index), and mean and minimum overnight Sao..
Computed Tomography Each patient had an awake supine CT scan (Somatom" DR 2; Siemens Electric, Ltd., Erlangen, Germany) of the upper airway performed before UPPP. The patient's head was positioned with the soft tissue Frankfort plane (tragus of the ear to soft tissue orbitale) perpendicular to the floor. Scans wereperformed during tidal breathing. The patient was instructed not to move or talk during the scan, to relax with the back teeth lightly contacted, and not to swallow during anyone scan. Patients were instructed to hold their index fin-
ger upright to confirm wakefulness throughout the scan. Contiguous scans were obtained at 8-mm intervals from the Frankfort plane to the sixth cervical vertebra. The scanning time for each image was 3 s. The estimated radiation dose associated with a complete CT scan of the upper airway was 0.5 rad. Linear attenuation coefficients for each pixel within each scanned section were stored digitally on a magnetic disk and were available for video display as an image analog to the digital distribution. A fixed window level of 30 and a window width of 210 were used to view muscle. A center point was fixed before scanning, and a magnification factor of 3.5 was used. Tracings were made on acetate paper of each of the slices for the upper airway (nasopharynx, oropharynx, and hypopharynx), tongue, and soft palate. The tongue was delineated to include all intrinsic and extrinsic muscles (genioglossus, hypoglossus, and styloglossus), but the muscles of the floor of the mouth (digastric, mylohyoid, and geniohyoid) were not included. Tracings of the soft palate were made from its tip to the junction with the hard palate. Tracings were also made of the outlines of the nasopharynx, oropharynx, and hypopharynx. The nasopharynx extended from the rostallimit of the airway to the levelof the hard palate, the oropharynx extended from the hard palate to the rostral tip of the epiglottis, and the hypopharynx extended to the level of the laryngeal introitus.
Each slice was digitized (Model 9874A; Hewlett-Packard Co., Andover, MA) and a crosshair cursor was used to enter the contour of each structure into a computer (lOOOE Series; Hewlett-Packard Co., Andover, MA), as has been previously described in detail (4). All contours were corrected to real size values by adjustments of a 5-cm scale digitized on the images. The minimum and mean upper airway and oropharyngeal cross-sectional areas were calculated. The site of minimum upper airway cross-sectional area was determined for each patient. The slices were oriented and stacked at 8-mm intervals in accordance with the fixed center point registered before scanning. Volume measurements, based on the cross-sectional areas of a series of contiguous slices and on the thickness of the slices, were calculated for the upper airway (nasopharynx, oropharynx, and hypopharynx), tongue, and soft palate as previouslydescribed (4). A three-dimensional reconstruction of the tongue and upper airway of a typical patient with GSA is shown in figure 1.
Statistical Analysis A two-tailed paired Student's t test was used to compare anthropometric and polysomnographic data before and after UPPP. Subgroup analysis was performed comparing patients with a "good" versusthose with a "poor" response to UPPP. A good response to UPPP was defined as a reduction in apnea index of ~ 50070 or to < 5 apneas/h 3 months after UPPP. A two-tailed unpaired Student's t test was used for comparison between these subgroups of patients. Linear regression analysis was used to determine the relation between preoperative anthropometric, polysomnographic, and CT measurements. Results
Fig. 1. Three-dimensional reconstruction of upper airway (blue) and tongue (red) in a patient with obstructive sleep apnea; right anterior oblique (upper left), left posterior oblique (upper right), and superior (lower) views.
The anthropometric and polysomnographic data for the 60 patients before and 3 months after UPPP are presented in table 1. In general, the patients were middle-aged and obese and had moderate to severeobstructive sleep apnea with associated arterial oxygen desaturation. There was a small decrease in body mass index following UPPP (p < 0.05). Apnea frequency and duration decreased (p < 0.0005), and mean and minimum asleep arterial oxygen saturation increased (p < 0.0005). Fifty patients (83070) had a good response to UPPP as defined by our preoperative criteria. Thirty-six patients (60070) had a post-UPPP apnea index of < 5/h TST, 14 (23%) had a ~ 50070 reduction in apnea index, 5 (8.5%) had a < 50% reduction in apnea index, and 5 (8.5%) had an increase in apnea index following UPPP. The decrease in body mass index following Uppp was not significantly different between patients who had a good and a
RYAN, LOWE, L1, AND FLEETHAM TABLE 1
47.5 ± 1.5 30.1 ± 0.6
47.8 ± 1.5 29.6 ± 0.6 6 ± 1.0 4 ± 1.0 94 ± 1.0 74 ± 2.0
Definition of abbreviations: UPPP • Mean ± SEM.
27 ± 2.2
21 ± 3.3 92 ± 0.5 64 ± 2.4
< 0.0005 < 0.0005 < 0.0005
40 ' - - - -.........._ _----'-_ _----J'--_------' 10 20 30 40 50 Body Mass Index (kg/m 2 )
< 0.0005 sleep time.
PREOPERATIVE ANTHROPOMETRIC AND POLYSOMNOGRAPHIC DATA COMPARING PATIENTS WHO HAD A GOOD AND A POOR RESPONSE FOLLOWING UVULOPALATOP HARYNGOPLASTY·
Age, yr Body mass index, kg/m 2 Apnea index, apnealh TST Total apnea time, %TST Mean Sa0 2 , % Minimum Sa0 2, %
~ ." ... • .
. ,......•: -.• . •
Significance (p value)
Age, yr Body mass index, kg/m 2 Apnea index, apneas/h TST Total apnea time, %TST Mean Sa0 2 , % Minimum Sa0 2 , %
~ ~ CL
Significance (p value)
21.1 ± 5.2 86.9 ± 2.4 47.0 ± 9.4
y =0.47 - 7.8 x 10 -3 x (b) P < 00005
0.0 '----'-_-'----'-_.....I...-~_....1...._"__---l 10 20 30 40 50 Body Mass Index (kg/m 2 )
NSt NS NS
< 0.001 < 0.01
y =2.7 - 5.1 x 10 -2 x P < 0.01
• Mean ± SEM. Not significant. For definition of other abbreviations, see table 1.
.. . ... .",
poor response to UPPP (- 0.5 ± 0.8 versus -004 ± 004 kg/m"). The preoperative anthropometric and polysomnographic data for the good and poor responders are presented in table 2. There wereno differences in age, body mass index, or apnea frequency or duration between good and poor responders. Good responders had less severe arterial oxygen desaturation during sleep preoperatively than poor responders. The minimum upper airway crosssectional area was located in the nasopharynx in 5 patients (8070), in the oropharynx in 49 patients (82%), and in the hypopharynx in 6 patients (10%). Its location was similar in the good and poor responders. Upper airway cross-sectional area measurements for good and poor responders are shown in figure 2. Upper airway cross-sectional area was smaller in the good responders, but this difference was only significant in the retroglossal oropharynx (p < 0.01). Tongue volume was directly related to body mass index (r = 0.32, p < 0.02)(figure 3a), and both upper airway to tongue volume ratio (r = - 0048, p < 0.0005)(figure 3b) and oropharynx to soft palate volume ratio (r = - 0.38, p < 0.01) (figure 3c) were inversely related to body mass index in the 60 patients. There was no relationship between upper airway volumes and either age or polysomnographic data.
The preoperative upper airway CT data for the good and poor responders are presented in table 3. Good responders had a smaller upper airway volume (p < 0.05), a smaller upper airway to tongue volume ratio (p < 0.01), and a smaller oropharynx to soft palate volume ratio (p < 0.05) compared with poor responders. Discussion
These results confirm that UPPP is an effectivetreatment for some patients with symptomatic OSA. This study also demonstrates that preoperative awake
·24 -16 -8 0 8 16 24 32 40 48 56 64 72 80 88 Nasopharynx Oropharynx Hypopharynx
Airway length (mm) Fig. 2. Preoperative upper airway cross-sectional areas in good (n = 50) and poor (n = 10) responders to UPPP. Zero represents the level of the hard palate. Data points represent mean values ± SEM. Closed circles = good response; open circles = poor response. Asterisk indicates p < 0.01.
.'·7. t·..· •
0'----.....1...----1...------'--------' 10 20 30 40 50 Body Mass Index (kg/m 2 ) Fig. 3. Relation between body mass index and (a) tongue volume, (b) upper airway to tongue volume ratio, and (c) oropharynx to soft palate VOlume ratio in 60 patients with obstructive sleep apnea.
three-dimensional upper airway measurements are related to obesity and differ between patients who have a good response and those who have a poor response to UPPP. Obese patients with OSA have large tongues and small upper airways relative to tongue and soft palate size. Patients with small upper airways, particularly relative to tongue and soft palate size, have a good response to UPPP. Three-dimensional computed tomography is being used increasingly to provide a better understanding of the interaction between different structures in a variety of diseases (18, 19) and to help plan reconstructive surgery (20). Threedimensional CT measurements provide a detailed analysis ofthe relationship between the upper airway and its surrounding soft tissues. It is important to determine the size of both the upper airway and the adjacent soft tissues to completely assess upper airway morphology in patients with OSA. Furthermore, volumetric measurements of soft tissue structures
30 UPPER AIRWAY SIZE IN SLEEP APNEA TREATED BY UPPP
TABLE 3 PREOPERATIVE UPPER AIRWAY COMPUTED TOMOGRAPHIC DATA COMPARING PATIENTS WHO HAD A GOOD AND A POOR RESPONSE FOLLOWING UVULOPALATOPHARYNGOPLASTY· Good Response Cross-sectional area, ernUpper airway minimum Upper airway mean Oropharynx minimum Oropharynx mean Volume, ml Nasopharynx Oropharynx Hypopharynx Upper airway Tongue Soft palate Upper airway, tongue Oropharynx, soft palate
Significance (p value)
0.86 2.70 0.91 2.10
± ± ± ±
0.05 0.09 0.06 0.09
0.91 3.19 0.99 2.61
± ± ± ±
0.24 0.32 0.28 0.30
NS NS NS NS
4.6 10.0 5.4 20.0 91.4 10.0 0.22 1.0
± ± ± ± ± ± ± ±
0.3 0.5 0.5 0.8 2.3 0.4 0.01 0.8
4.6 12.2 8.1 24.9 84.6 8.6 0.30 1.6
± ± ± ± ± ± ± ±
1.2 1.6 1.7 3.0 5.6 1.2 0.04 0.32
NS NS NS < 0.05 NS NS < 0.01 < 0.05
• Mean ± SEM.
may provide more useful information than measurements of cross-sectional area, as they are unaffected by neuromuscular factors. Upper airway muscle activity can change the cross-sectional areas of both the upper airway lumen and the surrounding soft tissues but cannot affect the volume of the soft tissue structures. The validity of both CT and magnetic resonance imaging volumetric measurements of the upper airway has been confirmed by fluid displacement of tongues dissected from cadavers (21) and in control objects (22). Our study reports upper airway, tongue, and soft palate volumes in a large sample of patients with OSA. Our measurements of upper airway and tongue volume are similar to those in our initial report of 25 patients with OSA (4). There is a large interpatient variability in upper airway, tongue, and soft palate size. Obese patients have larger tongues and smaller upper airways relative to tongue and soft palate size.This supports a causal link between obesity and the abnormal upper airway morphology observed in patients with OSA. A variety of studies have demonstrated abnormalities of upper airway configuration in awake OSA patients (2-5), but it is unclear whether these abnormalities are due to anatomic or neuromuscular factors. A recent study of the upper airway using magnetic resonance imaging has shown excess fat deposition in the soft palate and tongue and surrounding the collapsible segment of the pharynx in patients with OSA compared with weight-matched control subjects (23).A quantitative morphometric study of palatal tissue removed by Uppp has confirmed that OSA patients have excess fat in the soft palate (24). Fat deposition may contribute to an enlarge-
ment of the soft palate and tongue and a reduction in upper airway crosssectional area and volume in obese patients with OSA. The success of UPPP varies considerably, and some of this variability may be accounted for by differences in preoperative age, weight, and severity of OSA. Guilleminault and associates (7) reported objective data on 35 of 150 patients and suggested that the 8 patients who had a poor response to UPPP tended to be more obese. Fujita and coworkers (8) in a consecutive case series of 66 patients reported the opposite in that the 33 patients with a greater than 50070 reduction in apnea index were more obese. In a recent consecutive case seriesof 34 patients, 65070 had a greater than 50070 reduction in apnea index 6 months after UPPP (12). These patients had a lower body mass index and less severeOSA compared with the patients who had a poor response following UPPP. In our study we found no differences in age, body mass index, or apnea frequency or duration between patients who had a good response and those who had a poor response to UPPP. We observed less severe sleep arterial oxygen desaturation preoperatively in those patients who had a good response to UPPP. A variety of reports have attempted to identify the preoperative upper airv ay abnormalities associated with a successful outcome following UPPP. Fujita and coworkers (8) assessed the site of upper airway obstruction by clinical examination and suggested that patients with primarily oropharyngeal obstruction had a better response rate to UPPP than patients with both oropharyngeal and hypopharyngeal obstruction. Riley and associates (25) retrospectively compared cephalometrics between five patients who had
a good response following UPPP and nine patients who had a poor response following UPPP. Patients with a narrow posterior airway space and inferior hyoid bone position had a poor result following UPPP. No differences in cephalometries were found between responders and nonresponders in a recent consecutive case series of 34 patients (12). Other reports have examined the site of initial airway occlusion during apnea in an attempt to predict response to UPPP. Sher and associates (9) utilized nasopharyngoscopy during a Muller maneuver to determine the site of airway occlusion. Of 171 patients with OSA, 101 (59070) occluded their superior oropharynx rather than hypopharynx during a Muller maneuver. Thirty of these patients underwent UPPP and 87% had a > 500/0 reduction in apnea index. Katsantonis and Walsh (11) performed somnofluoroscopy on 26 consecutive patients before UPPP. Ten (67070) of 15patients had a successful response to UPPP when initial airway occlusion occurred in the retropalatal airway. One (9070) of 11 patients had a successful response when the site of initial airway occlusion was in more caudal regions of the upper airway. CT upper airway evaluation of awake patients with OSA before UPPP has been reported in two recent studies (14, 15); however, neither performed volumetric measurements of the upper airway and surrounding soft tissue structures. Larsson and colleagues (15) reported that 21 (66070) of 32 consecutive patients were cured 6 months following UPPP. The tongue width was smaller in the patients who had a good response to surgery. Shepard and Thawley (14) reported that 8 (35070) of 23 patients had a good response with a 50070 reduction in apnea plus hypopnea index following UPPP. Although there were no significant differences in mean or minimum upper airway cross-sectional area between the good and poor responders, airway crosssectional area 20 mm below the hard palate was borderline smaller in the good compared with the poor responders. Furthermore, UPPP failed to improve the four patients in whom minimum airway cross-sectional area was greater than 1 ern" before surgery. They concluded that patients with minimum airway crosssectional area less than 1.0 ern- located 20 mm below the hard palate were most likely to obtain a favorable response following UPPP. Our study was also carried out prospectively on consecutive patients and had a larger sample size. The patients were unselected apart from cardiorespiratory status, and there was com-
plete follow-up of all patients. Our response rate to UPPP is higher than generally reported for unselected patients (8, 12). This is most likely due to a surgical technique that removes as much redundant oropharyngeal tissue as possible while preserving the function of the proximal palatal musculature (16). Our results clearly show that patients with a smaller airway,particularly relative to tongue and soft palate size, have a good response to UPPP. The tongue, which forms a large part of the anterior boundary of the upper airway, may contribute to the upper airway narrowing in OSA either by an absolute increase in its size or by an abnormal position. The tongue and soft palate lie in close apposition when the mouth is closed and during obstructive apnea (3). An enlarged or abnormally positioned tongue may push the soft palate rostrally, causing narrowing of the retropalatal oropharynx, which is frequently observed in patients with OSA. Our findings are consistent with the current understanding of the pathogenesis of OSA (1). Patency of the upper airway depends on a balance of pressures. The transmural pressure at which the passive pharynx closes (P close) reflects the collapsibility of the upper airway and is determined by the structural characteristics of the upper airway soft tissues. Pharyngeal muscle pressure (P mus) is the positive pressure generated by contraction of upper airway dilating muscles. Intraluminal pressure (P lumen) is subatmospheric during inspiration as a result of the negative intrathoracic pressure that occurs during inspiratory muscle contraction. Upper airway occlusion occurs when the sum of intraluminal and pharyngeal muscle pressures is less than pharyngeal closing pressure (P lumen + P mus < P close) (26). UPPP may reverse the balance of this equation in favor of upper airway patency by several mechanisms. In patients with a good response to UPPP the surgery enlarges the oropharynx (14), which reduces upstream resistance and results in less negative intraluminal pressure (tP lumen). Acoustic reflection studies (27) have also demonstrated that upper airway collapsibility is reduced following UPPP, which decreases airway closing pressure (tP close). Patients who have a good response to UPPP have a smaller upper airway before surgery, which causes an increase in
RYAN, LOWE, L1, AND FLEETHAM
both upper airway resistance (tP lumen) and collapsibility (tP close). This airway narrowing makes these patients more likely to benefit from the changes in upper airway size and collapsibility that occur after UPPP. Other factors, such as neuromuscular activity and coordination, may be more important in the pathogenesis of OSA in those patients with large airways who respond poorly to UPPP. We have used three-dimensional upper airway CT to advance our understanding of the pathogenesis of OSA and of the mechanisms of action of UPPP. In the future, three-dimensional upper airway CT computer simulations may prove useful in predicting upper airway changes following UPPP or other treatments for OSA, such as weight loss, craniofacial surgery, or intraoral appliances. Acknowledgment The writers thank Dr. R. 1. Dickson and Dr. A. Blokmanis for allowing us to study their patients, Mrs. M. Wong and Mr. B. Sinclair for their technical assistance, and Mrs. B. Robillard for her secretarial assistance.
References 1. Remmers JE, DeGroot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 1978; 44:931-8. 2. Haponik EF, Smith PL, Bohlman ME, Allen RP, Goldman SM, Bleeker ER. Computerized tomography in obstructive sleep apnea. Correlation of airway size with physiology during sleep and wakefulness. Am Rev Respir Dis 1983; 127: 221-6. 3. Suratt PM, Dee P, Atkinson RL, Armstrong P, Wilhoit SC. Fluoroscopic and computed tomographic features of the pharyngeal airway in obstructive sleep apnea. Am Rev Respir Dis 1983; 127:487-92. 4. Lowe AA, Gionhaku N, Takeuchi K, Fleetham JA. Three dimensional CT reconstructions of tongue and airway in adult subjects with obstructive sleep apnea. Am J Orthod Dentofacial Orthop 1986; 90:364-74. 5. Horner RL, Shea SA, McIvor J, Guz A. Pharyngeal size and shape during wakefulness and sleep in patients with obstructive sleep apnea. Q J Med 1989; 72:719-35. 6. Fujita S, Conway W, Zorick F, Roth T. Surgical correction of anatomic abnormalities in obstructive sleep apnea syndrome. Uvulopalatopharyngoplasty, Otolaryngol Head Neck Surg 1981; 89:923-34. 7. Guilleminault C, Hayes B, Smith L, Simmons FB. Palatopharyngoplasty and obstructive sleep apnea syndrome. Bull Eur Physiopathol Respir 1983; 19:595-9. 8. Fujita S, Conway WA, Sicklesteel JM, et al. Evaluation of the effectiveness of uvulopalatopharyngoplasty. Laryngoscope 1985; 95:70-4. 9. Sher AE, Thorpy MJ, Shprintzen RJ, Spiel-
man AJ, Burack B, McGregor PA. Predictive value of Muller maneuver in selection of patients for uvulopalatopharyngoplasty. Laryngoscope 1985; 95:1483-7. 10. Kimmelman CP, Levine SB, Shore ET, Millman RP. Uvulopalatopharyngoplasty: a comparison of two techniques. Laryngoscope 1985; 95: 1488-90. 11. Katsantonis GP, Walsh JK. Somnofluoroscopy: its role in the selection of candidates for uvulopalatopharyngoplasty. Otolaryngol Head Neck Surg 1986; 94:56-60. 12. Gislason T, Lindholm CE, Almqvist M, et al. Uvulopalatopharyngoplasty in the sleep apnea syndrome. Arch Otolaryngol Head Neck Surg 1988; 114:45-51. 13. Walker EB, Frith RW, Harding DA, Cant BR. Uvulopalatopharyngoplasty in severe idiopathic obstructive sleep apnoea syndrome. Thorax 1989; 44:205-8. 14. Shepard JW, Thawley SE. Evaluation of the upper airway by computerized tomography in patients undergoing uvulopalatopharyngoplasty for obstructive sleep apnea. Am Rev Respir Dis 1989; 140:711-16. 15. Larsson S-G, Gislason T, Lindholm CEo Computed tomography of the oropharynx in obstructive sleep apnea. Acta Radiol 1988; 29:401-5. 16. Dickson RI, Blokmanis A. Treatment of obstructive sleep apnea by uvulopalatopharyngoplasty. Laryngoscope 1987; 97:1054-9. 17. Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring systems for sleep stages of human subjects. Los Angeles: UCLA, Brain Information Service/Brain Research Institute, 1968. 18. Burk DL, Mears DC, Kennedy WH, Cooperstein LA, Hebert DL. Three-dimensional computed tomography of acetabular fractures. Radiology 1985; 155:183-6. 19. Stern RL, Cline HE, Johnson GA, Ravin CEo Three-dimensional imaging of the thoracic cavity. Invest Radiol 1989; 24:282-8. 20. Vannier MW, Marsh JL, Warren JO. Three dimensional CT reconstruction images for craniofacial surgical planning and evaluation. Radiology 1984; 150:179-84. 21. Roehm EG. Computer tomographic measurement of tongue volume relative to its surrounding space. M.Sc. Thesis, University of Manitoba, 1981. 22. Abbey NC, Block AJ, Green D, Mancuso A, Hellard DW. Measurement of pharyngeal volume by digitized magnetic resonance imaging: effect of nasal continuous positive airway pressure. Am Rev Respir Dis 1989; 140:717-23. 23. Horner RL, Mohiaddin RH, Lowell DO, et al. Sites and sizes of fat deposits around the pharynx in obese patients with obstructive sleep apnea and weight matched controls. Eur Respir J 1989; 2:613-22. 24. Stauffer JL, Buick MK, Bixler EO, et al. Morphology of the uvula in obstructive sleep apnea. Am Rev Respir Dis 1989; 140:724-8. 25. Riley R, Guilleminault C, Powell N, Blair Simmons F. Palatopharyngoplasty failure, cephalometric roentgenograms and obstructive sleep apnea. Otolaryngol Head Neck Surg 1985; 93:240-4. 26. Remmers JE. Anatomy and physiology of upper airway obstruction. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine. Philadelphia: W.B. Saunders, 1989; 525-36. 27. Wright S, Haight J, zamel N, Hoffstein V. Change in pharyngeal properties after uvulopalatopharyngoplasty. Laryngoscope 1989; 99:62-5.