Localization of Upper Airway Collapse during Sleep in Patients with Obstructive Sleep Apnea 1 - 3

JOHN W. SHEPARD, JR. and STANLEY E. THAWLEY

Introduction

The surgical success rate for uvulopalatopharyngoplasty (UPPP) in the treatment of obstructive sleep apnea (OSA) has varied considerably in the published series with the general consensus being that one-half of the patients will obtain a greater than 50070 reduction in the frequency of disordered breathing events (1-6). Because UPPP has been documented to increase the cross-sectional area of the upper airway (UA) at the level of the resected soft palate and uvula (7), failure to improve after surgery is usually attributed to the inclusion of patients who preoperatively collapse their UA caudal to the soft palate and uvula. In contrast, patients with preoperative collapse confined to the retropalatal or velopharyngeal segment of the UA are considered to be the optimal candidates for UPPP (5, 8). Several methods have been used to evaluate the regions over which the UA collapses during sleep including fiberoptic pharyngoscopy (9, 10), fluoroscopy (8, 11, 12), fast-CT scanning (13), and physiologic measurements of pressures within the UA (14, 15). Although fiberoptic pharyngoscopy successfully visualizes the proximal site of collapse, the caudal extent of collapse is not precisely defined. Using fluoroscopy, Suratt and coworkers (11) observed that UA collapse began proximally at the levelof the soft palate and extended caudally down the UA to a variable extent in a small group of patients with OSA. Stein and colleagues (13), who used fastCT scanning to directly image the region over which UA collapse occurs during sleep, also observed that collapse was initiated proximally and extended caudally to a variable extent in individual patients. Using fluoroscopy, Katsantonis and Walsh (8) reported a greater success rate with UPPP in subjects with collapse confined to the segment of the UA cephalad to the mid portion of the second cervical vertebral compared with subjects demonstrating collapse in more caudal regions. 1350

SUMMARY The present study was conducted to determine the effects of body position and sleep state, as well as the effect of uvulopalatopharyngoplasty (UPPP)on the regions over which the upper airway (UA) collapses during sleep. To accomplish this goal, 18 male patients with obstructive sleep apnea (OSA) underwent overnight polysomnography with simultaneous monitoring of pressures in the posterior nasopharynx, oropharynx, hypopharynx, and esophagus. From the profile of pressures recorded in the UA and esophagus, the regions over which the UA collapses during apneas could be determined. The patients were 54 ±14 y of age and were grossly obese with a body mass index of 37 ± 2 kg/m 2 • They had moderately severe OSA with a mean apnea plus hypopnea index of 62 ± 8 per hour. During NREM sleep, 10 of the 18 (56%) patients had collapse confined to the velopharyngeal or retropalatal segment of the upper airway. The remaining 44% of the patients demonstrated collapse of the retroglossal segment of the oropharynx located caudal to the inferior margin of the soft palate. Upper airway collapse at the level of the hyoid bone was not observed during NREM sleep. Observations made during REM sleep in nine patients demonstrated that collapse occurred in a more caudal segment of the UA in seven patients during REM than during NREM sleep. The effect of sleep position was evaluated in 10 patients and found to have little affect on the extent over which the UA collapsed during sleep independent of sleep state. The effects of UPPP on regional UA collapse were evaluated in a small group of six patients. The results indicated that patients with UA collapse confined to the velopharynx preoperatively may have persistent collapse of this level after surgery and that failure to respond is not always due AM REV RESPIR DIS 1990; 141:1350-1355 to collapse of more caudal segments of the UA.

Several studies have used recordings of pressure at multiple sites within the UA to define regions of collapse during sleep (14, 15). Although these studies did not report the effects of body position or REM sleep on the region over which the UA collapses, approximately 50070 of patients with OSA had collapse confined to the velopharynx during NREM sleep. The fact that the remaining one-half of the patients demonstrated collapse of a more caudal segment suggested that this may be an important pathophysiologic variable contributing to the approximately 50070 failure rate for UPPP. The importance of body position in determining whether or not the UA collapses during sleep has been documented in several studies (16-19). More than 50070 of patients with OSA will have repetitive disordered breathing events in the supine sleep position, with greatly diminished apneic activity when sleeping in the lateral decubitus positions. It is not known whether body position affects the region over which the UA collapses. In addition to the importance of body position, sleep state is known to affect

whether or not the UA collapses. Stages 3 and 4 of NREM sleep have been associated with both a decrease in apneic activity (20) and a greater resistance to UA collapse compared with those during Stages 1 and 2 (21). In contrast, REM sleep is associated with a greater propensity for UA collapse, presumbably because of UA muscle atonia. The present study was conducted to determine the effects of body position and sleep state, as well as the effects of UPPP on the regions over which the UA collapses during sleep.

(Receivedin originalform July 5, 1989 and in revised form October 13, 1989) 1 From the Sleep Disorders Center and the Division of Thoracic Diseases, Mayo Clinic and Mayo Foundation, Rochester, Minnesota, and the Veterans Administration Medical Center, St. Louis, Missouri. 2 Supported by the Mayo Clinic Foundation and by the Veterans Administration. 3 Correspondence and requests for reprints should be addressed to John W. Shepard, Jr., M.D., Sleep Disorders Center, Mayo Clinic, 200 SW First Street, Rochester, MN 55905.

1351

SITE OF UPPER AIRWAY COLLAPSE IN SLEEP APNEA

Methods

Patients Eighteen male subjects with a polysomnographically confirmed diagnosis of obstructive sleep apnea and apnea plus hypopnea index (AHI) > 10 per hour were recruited for participation in this study. Informed consent was obtained in compliance with the guidelines of the St. Louis Veterans Administration Medical Center. Two of the 18 subjects had previously undergone UPPP. One of these patients was restudied six months after revision of his UPPP, and five others were restudied 3.5 ± 0.8 months after de novo UPPP performed according to the method described by Fujita and coworkers (22). Arterial blood was obtained in all patients during wakefulness in the seated position and analyzed for Pa02, Pac0 2, and pH using a Model IL-813 gas analyzer (Instrumentation Laboratories, Lexington, MA). Spirometry was performed in 15 subjects with a dry rolling-seal spirometer (Horizon System; Sensor Medics Corp., Anaheim, CA). Spirometric data were expressed as a percentage of previously published normal values (23).

Sleep Study Each patient underwent overnight polysomnography using standard techniques to record electroencephalographic (EEG) activity from central leads (C 3-A2/C 4-A,), submental electromyographic (EMG) activity, and electrooculographic (EO G) activity. Nasal and oral thermistors were used to monitor air flow, abdominal and thoracic strain gauges were used to document respiratory effort, and ear oximeters (Biox 2A; Biox Technology, Boulder, CO) to monitor arterial oxyhemoglobin saturation. A multichannel polygraph (Model 78; Grass Instruments, Quincy, MA) was used to record the data, which were scored in 30-s epics according to standard criteria (24). Apneas and hypopneas were scored according to the criteria of Block and colleagues (20). The AHI was computed as a frequency of these disordered breathing events per hour of sleep. Baseline values for Sa02 were averaged to derive mean high Sa02, whereas the values for Sa02 at the nadir of the desaturations were averaged to compute mean low Sao.; The lowest value for Sa02 for each subject was also recorded. A multiport catheter was constructed to monitor pressures in the UA and esophagus during this study. The catheter consisted of an esophageal balloon 12 em in length (A and E Medical Co., Farmingdale, NJ) placed at the end of polyethylene tubing 120 em in length with an inner diameter of 1.27 mm to record esophageal pressures. Three separate polyethylene catheters, each with an internal diameter of 0.86 mm and 95 em in length, were affixed to the esophageal catheter with silas tic silicon adhesive (Dow Corning, Midland, MI) using silastic 382 medical grade elastomer as the catalyst to form this multiport catheter. The distal port used to record pressure in the hypopharynx (Ph) was locat-

ed 250 mm proximal to the distal tip of the esophageal balloon. The catheter used to record pressure in the oropharynx (Po) just distal to the inferior margin of the soft palate and uvula was located 35 mm proximal to the hypopharyngeal pressure port, and the catheter tip used to record pressures in the nasopharynx (Pn) was 40 mm proximal to the oropharyngeal pressure port (Figure 1). The cross-sectional area ofthe catheter at its thickest segment in the nose was 30 mm-. A Model MP45 pressure transducer (Validyne Corp., Northridge, CA) was used to record pressures in the esophageal balloon, which was filled with 0.5 ml of air. Statham P23ID pressure transducers (Statham Instruments, Oxnard, CA) were used to record pressures from the catheters placed in the upper airway. These catheters were filled with normal saline and periodically flushed as required throughout the evening. The response characteristics of this multiport pressure catheter were carefully evaluated in a Plexiglas'" box using a Harvard animal ventilator (Harvard Apparatus Co., South Natick, MA) to generate cyclical fluctuations in pressure. After calibration using a water manometer, essentially identical values for pressure were recorded from the esophageal balloon and three fluidfilled catheters at frequencies as great as 0.5 cps. At higher frequencies there was a phase lag in the pressures recorded from the esophageal balloon compared with recordings made with the saline-filled catheters. Careful attention was paid to the design and construction of the pressure ports used to record pressures in the upper airway. Silastic adhesive was circumferentially removed

Pn Po Ph

1"fJ I I IVelophlllyf1X I Oropharynx I I I

Hypopharynx

I

I I I I I

iii

o

10 20 30 40 50 60 70

CT scan

level, mm

Fig. 1. This diagram schematically depicts the positioning of the catheter used to record pressures in the posterior nasopharynx (Pn), the oropharynx below the inferior margin of the soft palate (Po), the hypopharynx (Ph), and the esophagus (Pes). The velopharyngeal segment of the upper airway was defined as the retropalatal air space from the level of the hard palate to the inferior margin of the soft palate. The oropharynx is defined as the retroglossal region of the upper airway extending from the inferior margin of the soft palate to the tip of the epiglottis. The hypopharynx was defined as the region of the upper airway extending from the tip of the epiglottis to the laryngeal vestibule.

around the distal tip of each of the UA pressure ports such that occlusion of almost the entire circumference around the catheter tip would be required before UA pressure fluctuations would cease to be recorded. This design permitted prolonged recordings of pressure in the UA in contrast to catheters constructed with side holes, which would occlude whenever the side with the hole would contact the wall of the UA. Despite the design of the pressure ports, occasional dampening of the pressure signals would occur, presumably from inspissated secretions. At these times 1 ml of normal saline would be instilled to clear the line. Copper wire was imbedded in the silicone adhesive between the hypopharyngeal and oropharyngeal pressure ports to assist with accurate placement of the catheter. After topical anaesthesia with 4% Xylocaine, the catheter was placed with the oropharyngeal pressure port located just below the inferior margin of the soft palate and uvula under direct visual examination of the posterior oropharynx. The patient was then instructed to close his mouth and, while breathing through the nose in the supine position, a radiograph of the lateral neck was obtained. When required, the position of the catheter was adjusted to place the oropharyngeal pressure port just distal to the inferior margin of the soft palate at the base of the tongue. This resulted in hypopharyngeal pressures being recorded in the hypopharynx at the level of the body of the hyoid bone, whereas the proximal pressure port was generally located in the posterior nasopharynx at the level of the hard palate. Once positioned, the catheter was marked at the distal nares and secured with adhesive tape, and its position was verified throughout the night as well as at the end of the study. Regions of UA collapse were determined by changes in the pressures recorded along the UA during apneas. Collapse of the velopharyngeal segment was indicated by the absence of negative pressure deflection in P n with the negative pressures recorded in the orpharynx and hypopharynx closely approximating the swings in esophageal pressure (figure 2). Collapse was considered to extend into the oropharyngeal segment when Po ceased to decrease in parallel with the pressure recorded in the hypopharynx (figure 3). Collapse of the hypopharyngeal segment was documented when Ph failed to decrease in parallel with esophageal pressure (figure 4). The study was terminated when the patient woke up or whenever the patient requested that the UA catheter be removed.

Statistical Analysis The data are presented as mean ± SEM. The paired and nonpaired t statistics (two-tailed) were used to test for differences in mean values. Results

The mean age for the patient group was 54 ± 3 yr (range, 26 to 74 yr). Mean

1352

SHEPARD AND THAWLEY

sso

x 2

100-_ "

-soOo! -----

TABLE 1

90-

Fig. 2. This figure demonstrates the pattern of pressures recorded from the upper airway in a patient with collapse limited to the velopharyngeal segment of the airway. Note the absence of negative pressure deflections in the nasopharynx . with parallel reductions in pressures in the oropharynx . hypopharynx. and esophagus .

Fig. 3. This figure depicts the most common pattern of obstruction observed extending into the oropharynx. Although the reduction in Ph parallels the decreases in esophageal pressure. there is no negative deflection in Pn. indicating the velopharynx is obstructed at the onsel of inspiration. The pressure in the oropharynx initially becomes negative and is then truncated at the pressure at which collap se of this segment occurs.

sso " 10 0 2

S0Oo!

90

Fig. 4. This pattern was observed during REM sleep in the same patient illustrated in figure 3 in whom collapse extends inferiorly into the hypopharyngeal segment of the airway indicated by the truncation of negativepressuredeflection in Ph relative to esophageal pressure.

height was 176 ± 2 ern (range, 161 to 185 ern). The patients were grossly obese with a body mass index of 37 ± 2 kg/m! (range, 25 to 48 kg/m-), The results of the arterial blood gas analysis and spirometry indicated that the group was mildly hypercapnic and hypoxemic awake at rest with borderline reductions in forced vital capacity, but a normal FEV. and FEV./FVC ratio (table I), Two pa tients had mild chronic obstructive pul monary disease with FEV.lFVC ratios of 62 and 67ITJo, The results of overnight polysomnography indicated that the group had relatively severe obstructive sleep apnea in terms of the frequency of disordered breathing events, AHI = 62

± 8 per hour, but only moderate oxyhemoglobin desaturation with a mean low Sa0 2 of 86 ± IlTJo. The multiport catheter was positioned with the hypopharyngeal pressure port an average of 17.2 ± 0.4 em below the nares corresponding to the midportion of the fourth cervical vertebral body (figure 5), The oropharyngeal port was generally located at the level of the caudal segment of the second cervical vertebral body, but it ranged as high as its midportion and as low as the intervertebral disc space between C2 and C3. The regions of upper airway collapse during sleep are illustrated in figure 6 for both NREM and REM sleep. The region

ARTERIAL BLOOD GAS . SPIROMETRIC. AND POLYSOMNOGRAPHIC TEST RESULTS IN 16 PATIENTS WITH OBSTRUCTIVE SLEEP APNEA Test Results

Arterial blood gas Pao,. mm Hg Paco,. mm Hg pH Spirometry FVC. 0lll predicted FEV, . % predicted FEV,/FVC. % Polysomnography TST. min REM. % AHI. nih Duration. s Mean high Sao,. % Mean low Sao,. % Lowest Sao,. °lll

Mean ± SEM

Range

70 ± 2 45 ± 1 7.39 ± 0.Q1

59-69 34-53 7.34-7.46

63 ± 5

66 ± 5 76 ± 2 267 ± 24 4 ± 1 62 ± 6

21 ± 2 94 ± 1 66 ± 1

73 ± 2

60-131' 63-112' 62-67" 44-402 0-12 14-119 14-36 69-96 79-90 44-66

Definition of abbreviations: TST = total sleep time; AHI apnea plus hypopnea index. • n = 15.

of collapse for each patient remained constant throughout the night for a given sleep state and body position. No differences were observed between mixed and obstructive apneas in the region of collapse. During NREM sleep, collapse was confined to the velopharyngeal segment of the VA in 10 of the 18 patients (56ITJo). In eight patients (44lTJo), collapse extended into the oropharyngeal region below the level of the inferior margin of the soft palate at the base of the tongue. There were no significant differences in age, weight, body mass index, spirometry, arterial blood determinations, or polysomnographic variables between the patients with collapse confined to the velopharyngeal segment and those with collapse extending into the oropharynx at the base of tongue. In six of the eight patients in whom collapse extended caudally into the oropharyngeal segment, the oropharynx was initially patent at apnea onset. This pattern of sequential collapse, illustrated in figure 3, is characterized by initial negative deflections in Po with the negative pressure being truncated when intraluminal suction pressure became sufficiently negative to collapse the airway. The magnitude of the negative pressure required to produce collapse of the oropharyngeal segment frequently increased with increasing apnea duration. Only two patients demonstrated initial collapse of the oropharyngeal segment, indicated by the absence of negative pressure generation in the oropharynx. Observations made during REM sleep

1353

SITE OF UPPER AIRWAY COLLAPSE IN SLEEP APNEA

Fig. 5. Lateral radiograph of the neck indicating the typical catheter position for the recording of oropharyngeal (Po) and hypopharyngeal (Ph) pressures.

Fig. 6. Caudal extent of upper airway collapse in patients with OSA during NREM (n = 18)and REM (n = 9) sleep. Each dot represents the region of collapse in an individual patient.

Veloph8JYllX

• ••• •

Oroph8JY!lX

•••

•

Hypopharynx

NREM sleep

REM sleep

NREMVelopharynx

Fig. 7. Caudal extent of upper airway collapse in the supine and lateral decubitus (side) positions in NREM (n = 10) and REM (n = 2) sleep.

Oropharynx

REM 0--0

••

--.•

• o

Hypopharynx

in nine patients documented that collapse extended into a more caudal segment of the VA in seven (78%) of the nine patients during REM than during NREM sleep. The two patients who did not show collapse of a more caudal segment of the VA during REM sleep had collapse confined to the velopharynx. The effects of body position on regional VA collapse during REM and NREM sleep are illustrated in figure 7. NREM sleep was observed in both the supine and lateral decubitus positions in 10patients. In eight of these patients, collapse occurred over identical segments of the VA in both sleep positions. The other two subjects had velopharyngeal collapse in the supine position, which extended into the oropharyngeal segment when the patient slept in the lateral decubitus position. The two patients in whom REM sleep was observed in both the supine and lateral decubitus positions showed collapse over identical segments in both positions. The six patients studied before and after VPPP showed no significant change in AHI, with a mean of 71 ± 16per hour before surgery compared with 69 ± 17 per hour after surgery. The effects of this surgical procedure on the region of VA collapse during NREM sleep are presented in figure 8. Prior to surgery, three patients had collapse confined to the velopharynx. Each of these subjects demonstrated collapse of the same velopharyngeal segment after surgery. The three patients in whom collapse extended into the oropharynx before surgery showed a variable response. After VPPP, one subject collapsed his VA over the same segment, whereas another collapsed only the velopharyngeal segment. The third patient demonstrated collapse down to the level of the hypopharynx after surgery. Discussion

Supine

Side

Velopharynx

Fig. 8. Regions of UA collapse before and after uvulopalatopharyngoplasty (UPPP) in six patients with OSA.

Oropharynx

Hypopharynx

Before

After

UPPP

UPPP

The results of the present study, demonstrating velopharyngeal collapse during NREM sleep in 56% of the patients, with collapse of a more caudal segment in 46070, are in excellent agreement with the results obtained in previous studies. In the study conducted by Hudgel (14),46% of nine patients showed collapse of the velopharyngeal segment, with 56% collapsing more caudally in the retroglossal region, which he defined as hypopharyngeal collapse. Similar results were reported by Chaban and coworkers (15) who documented that collapse was confined to the velopharynx in 50% of the 10 patients they studied. Because the position

1354

of the oropharyngeal catheter is critical to the velopharyngeal segment, whereas to the results obtained, it should be not- the other three had collapse extending ined that the oropharyngeal pressure port to the oropharynx. The only patient to was placed an average of 13.7 ± 0.3 cm have a "successful" result from this pro(range, 10.4 to 15.9 em) from the exter- cedure based on a reduction in AHI from nal nares in the present study compared 52 to 23 per hour had collapse extending with a distance of 10to 12em in the study into the oropharynx before surgery, with conducted by Hudgel (14).The more dis- collapse confined to the velopharynx aftal catheter placement in the present ter surgery. The oropharyngeal pressure study may have accounted for the slight- port was positioned the same distance ly greater frequency of obstruction be- from the external nares in the postopering confined to the velopharyngeal seg- ative and preoperative studies. Although ment. Because neck radiographs were not the number of patients in the present obtained in Hudgel's study, the position- study was small, the results suggest that ing of the oropharyngeal catheter port the performance of VPPP in patients relative to the cervical vertebra was not with collapse confined to the velopharynreported. However, the positioning of this geal segment does not guarantee a sucpressure port in the present study was cessful outcome. Furthermore, the results highly compatible with that of Chaban document persistence of collapse in the and coworkers, whose results indicated proximal segment of the VA. Careful asthat the inferior margin of the soft pal- sessment of regional VA collapse will ate corresponded with the midportion of have to be undertaken both before and the second cervical vertebral body. Vari- after VPPP in larger numbers of subations in the length of the palate and uvu- jects before definitive conclusions can be la likely account for the small variations reached. However, these preliminary rein catheter placement relative to the cer- sults suggest that modifications of this vical vertebrae in our patients. In addi- surgical procedure to further enhance the tion, substantial differences in patient patency of the more cephalad regions of height in the present study likely con- the VA may need to be undertaken if its tributed to the variability in catheter success rate is to be improved. It should be mentioned that the region length between the external nares and the oropharyngeal pressure port. Because the of VA collapse remained remarkably distances between the hypopharyngeal, constant throughout the night, with the oropharyngeal, and nasopharyngeal pres- exception of more extensive collapse dursure ports were fixed, differences in pa- ing REM sleep. This observation suggests tient height and placement of the oro- that valid conclusions on regional colpharyngeal port resulted in pressure being lapse can be reached with monitoring exrecorded from slightly different regions of tending only through the first episode of the nasopharynx and hypopharynx, In all REM sleep. Topical anaesthesia did not cases, the hypopharyngeal pressure port appear to alter the segment over which was visualized as being located in the cau- the VA collapsed as there was no differdal segment of the hypopharynx on the ence observed early compared with later lateral neck roentgenograms. in the study. Most of the subjects demonBody position was found to have little strating collapse in the oropharyngeal reeffect on the region over which the VA gion showed the pattern illustrated in figcollapses during both REM and NREM ure 3. As apnea duration increased, the sleep. This finding has not previously magnitude of the suction pressure rebeen reported and suggests that although quired to collapse the VA would often position may substantially affect wheth- increase to more negative values as repetier or not collapse occurs, once it does tive inspiratory efforts were made. The occur, sleep position does not have a ma- greater resistance to VA collapse, indijor influence. An additional new find- cated by the recording of more negative ing in the present study was that the pressure, would be consistent with an inlength of VA collapse was more exten- crease in chemical drive (hypercapniasive during REM than during NREM hypoxia) activating VA dilator muscles sleep. Collapse extended into the hypo- prior to arousal and apnea termination. pharyngeal segment of the VA during Failure to observe negative pressure deREM sleep in 78070 of the patients. This flections with the onset of inspiratory efresult is consistent with atonia of the VA fort in either the nasopharyngeal or muscles during REM sleep, allowing col- oropharyngeal pressure ports indicates that the respective segments of the VA lapse over a longer segment. Preoperatively, three of six patients un- were collapsed prior to inspiration. This dergoing VPPP had collapse confined result is consistent with observations re-

SHEPARD AND THAWLEY

cently made with fast-CT scanning during wakefulness showing collapse of the VA at end-expiration in selected patients with OSA (25). In summary, 56% of 18 patients with OSA had VA collapse confined to the velopharyngeal segment of the VA during NREM sleep. The remaining subjects had collapse extending into the oropharyngeal segment. During REM sleep, collapse extended caudally into the hypopharynx in the majority of patients. Body position had little effect on the extent of VA collapse independent of sleep state. The results obtained in a small number of patients studied before and after VPPP suggest that having collapse confined to the velopharynx preoperatively does not guarantee a successful outcome. In addition, failure to improve after VPPP was most frequently associated with persistence of collapse in the proximal segment of the airway. These preliminary observations suggest that the generally held concept that failure to respond to VPPP is due to collapse of more caudal segments of the VA may not be correct. Acknowledgment The writers thank Debra Grither, Mike Garrison, and Bruce Daniels for technical assistance, and Cathy Nelson for preparation of the manuscript. References I. Fujita S, Conway W, Zorick F, et 01. Evaluation of the effectiveness of uvulopalatopharyngoplasty. Laryngoscope 1985; 95:70-4. 2. Simmons FB, Guilleminault C, Silvestri R. Snoring, and some obstructive sleep apnea, can be cured by oropharyngeal surgery. Palatopharyngoplasty. Arch Otolaryngol Head Neck Surg 1983; 109:503-7. 3. Simmons FB, Guilleminault C, Miles LE. The palatopharyngoplasty operation for snoring and sleep apnea: an interim report. Otolaryngol Head Neck Surg 1984; 92:375-80. 4. Hernandez SF. Palatopharyngoplasty for the obstructive sleep apnea syndrome: technique and preliminary report of results in ten patients. Am J Otolaryngol 1982; 3:229-34. 5. Sher AE, Thorpy MJ, Shprintzen RJ, Spielman AJ, Burack B, McGregor PA. Predictive value of Muller maneuver in selection of patients for uvulopalatopharyngoplasty. Laryngoscope 1985; 95:1483-7. 6. Kimmelman CP, LevineSB, Shore ET, Millman RP. Uvulopalatopharyngoplasty: a comparison of two techniques. Laryngoscope 1985; 95:1488-90. 7. Shepard JW Jr, Thawley SE. Evaluation of the upper airway by computerized tomography in patients undergoing uvulopalatopharyngoplasty for obstructive sleep apnea. Am Rev Respir Dis 1989; 140:711-6. 8. Katsantonis GP, Walsh JK. Somnofluoroscopy: its role in the selection of candidates for uvulopalatopharyngoplasty. Otolaryngol Head Neck Surg 1986; 94:56-60. 9. Weitzman ED, Pollak CP, Borowiecki B, Bu-

1355

SITE OF UPPER AIRWAY COLLAPSE IN SLEEP APNEA

rack B, Shprintzen R, Rakoff S. The hypersomniasleep apnea syndrome: site and mechanism of upper airway obstruction. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R. Liss, 1978; 235-48. 10. Hill MW, Guilleminault C, Simmons FB. Fiber-optic and EMG studies in hypersomnia-sleep apnea syndrome. In: Guilleminault C, Dement WC, eds. Sleep apnea syndromes. New York: Alan R. Liss, 1978; 249-58. 11. 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. 12. Walsh JK, Katsantonis GP, Schweitzer PK, VerdeIN, Muehlback M. Somnofluoroscopy: cineradiographic observation of obstructive sleep apnea. Sleep 1985; 8:294-7. 13. Stein MG, Gamsu G, de Geer G, Golden JA, Crumley RL, Webb WR. Cine CT in obstructive sleep apnea. AJR 1987; 148:1069-74. 14. Hudgel DW. Variable site of airway narrow-

ing among obstructive sleep apnea patients. J Appl Physiol 1986; 61:1403-9. 15. Chaban R, Cole P, Hoffstein V. Site of upper airway obstruction in patients with idiopathic obstructive sleep apnea. Laryngoscope 1988; 98:641-7. 16. Cartwright RD, Lloyd S, Lilie J, Kravitz H. Sleep position training as treatment for sleep apnea syndrome: a preliminary study. Sleep 1985; 8: 87-94. 17. Phillips BA, Okeson J, Paesani D, Gilmore R. Effect of sleep position on sleep apnea and parafunctional activity. Chest 1986; 90:424-9. 18. McEvoy RD, Sharp DJ, Thornton AT.The effects of posture on obstructive sleep apnea. Am Rev Respir Dis 1986; 133:662-6. 19. Lloyd SR, Cartwright RD. Physiologic basis oftherapy for sleep apnea. Am Rev Respir Dis 1987; 136:525-6. 20. Block AJ, Boysen PG, Wynne JW, Hunt LA. Sleep apnea, hypopnea, and oxygen desaturation in normal subjects. A strong male predominance. N Engl J Med 1979; 300:513-7. 21. Issa FG, Sullivan CEo Upper airway closing

pressures in obstructive apnea. J Appl Physiol1984; 57:520-7. 22. 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. 23. Morris JF, Koski A, Johnson LC. Spirometric standards for healthy nonsmoking adults. Am Rev Respir Dis 1971; 103:57-69. 24. Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Washington, D.C.: U.S. Government Printing Office, 1968. 25. Shepard JW Jr, Stanson AW, Sheedy PF, Westbrook PRoFast-CT evaluation of the upper airway during wakefulnessin patients with obstructive sleep apnea. In: Suratt PM, Remmers J, eds. Proceedings of the first international symposium on sleep and respiration. New York: Alan R. Liss, 1990(In Press).