Magnetic Resonance Imaging of the Upper Airway in Obstructive Sleep Apnea before and after Chronic Nasal Continuous Positive Airway Pressure Therapy1-3
C. FRANCIS RYAN,4 ALAN A. LOWE, DAVID LI, and JOHN A. FLEETHAM Introduction
Obstructive sleep apnea (OSA) 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 load due to a narrow airway predisposes to upper airway occlusion. Snoring is the hallmark of patients with OSA and is caused by partial upper airway obstruction and high-frequency oscillation of the soft palate (2). Many patients with OSA complain of pharyngeal swelling on awakening and appear to have an inflamed and swollen oropharynx, which improves as the day progresses (3). Upper airway edema may occur secondary to OSA and subsequently exacerbate the OSA by causing further upper airway narrowing. Nasal continuous positive airway pressure (CPAP) is a very effective form of treatment for OSA. It was initially proposed that nasal CPAP provides a physical "pressure" splint preventing upper airway occlusion by pushing the soft palate and tongue forward and away from the posterior oropharyngeal wall (4). Preliminary reports indicate that patients initially demonstrate an improvement in the severity of their OSA if nasal CPAP is discontinued after a period of chronic treatment (5, 6). However, after a few days the sleep apnea returns to the previous degree of severity. One explanation for this phenomenon is that nasal CPAP relieves the snoring, thereby allowing a resolution of upper airway narrowing due to edema, which gradually reaccumulates when the nasal CPAP is discontinued. Magnetic resonance imaging (MRI) produces high-resolution images without the use of ionizing radiation. MRI depends on the tissue relaxation times and regional concentrations of hydrogen nu-
SUMMARY Magnetic resonance Imaging (MRI) provides high-resolution Images of the upper airway and Is useful for assessing conditions associated with Increased tissue water content. Todetermine whether nasal continuous positive airway pressure (CPAP)changes awake upper airway morphology in obstructive sleep apnea (OSA), we performed awake upper airway MRI scans on five male patients with moderate to severe OSA before and after 4 to 6 wk of nasal CPAP therapy. MRI scans were performed using spin echo pulse sequences to examine detailed anatomy and Inversion recovery sequences to assess mucosal water content. Patients did not have nasal CPAP applied during the MRI scans. Axial and sagittal Images were obtained, and tracings were made of the upper airway, tongue, and soft palate. Utilizing computer graphics, cross-sectional areas and volumes were calculated for each anatomic structure. A subjective grading system was used to assess upper airway mucosal water content. Pharyngeal volume and minimum pharyngeal cross-sectional area increased (p < 0.05) and tongue volume decreased (p < 0.01)following chronic nasal CPAPtherapy. The Increase In pharyngeal volume occurred mainly in the oropharynx (p < 0.01). Upper airway mucosal water content decreased In the oropharynx (p < 0.05). We conclude that chronic nasal CPAPtherapy during sleep In patients with OSA produces changes in awake upper airway morphology. These changes may be due to resolution of upper airway edema. The upper airway of patients with OSA AM REV RESPIR DIS 1991; 144:939-944 can be accurately and repeatedly assessed using MRI.
cleito produce an image (7). Because each water molecule has two hydrogen nuclei, MRI is ideally suited to assessing conditions associated with increased tissue water content. We performed awake MRI scans on patients with OSA before and after chronic nasal CPAP therapy to determine whether upper airway morphology changes with chronic nasal CPAP therapy. Methods Patients Five male patients with newly diagnosed moderate to severe symptomatic OSA, who elected to be treated with nasal CPAP, were recruited to the study. Patients whose weight wasgreater than 115 kg wereexcluded because of the size limitations of the MRI scanner. Patients who were edentulous were also excluded because of difficulty in defining the boundaries of the tongue on the upper airway scan. Each patient gaveinformed written consent, and the study protocol was approved by the University of British Columbia Clinical Screening Committee for Research Involving Human Subjects. Polysomnography Each patient had a detailed overnight polysomnogram performed before and after a
period of chronic nasal CPAP therapy. The initial study was performed without nasal CPAP, and the second study was performed with nasal CPAP. Sleep and its various stages was documented by standard electroencephalographic (EEG), electrooculographic (EOG), and e1ectromyographic (EMG) criteria. EEG was recorded with electrodes applied at C3-A2 and CcA, (according to the International 1020 system), and EMG activity was recorded from the submental muscles. Apneas were documented by a Beckman LB-2 infrared CO 2 analyzer (Beckman Instruments Inc., Schillar Park, IL), which recorded from both the nose and mouth. A singleECG (modified V2)
(Receivedin originalform November 15, 1989 and in revised form May 3, 1991) 1 From the Departments of Medicine, Clinical Dental Sciences, and Radiology, University of British Columbia, Vancouver, British Columbia, Canada. 2 Supported by the British Columbia Lung Association and the British Columbia Health Care Research Foundation. 3 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 British Columbia Lung Association Research Fellow.
939
RYAN, LOWE, LI, AND FLEETHAM
940
was monitored to detect card iac arrhythmias. Sao 2was monitored continuously with a Nellcor N-l00 pulse oximeter (Nellcor Inc., Hayward, CA). Chest wall movement was monitored by a respiratory inductive plethysmograph (Respitrace'"; Ambulatory Monitoring Equ ipment Inc., Ardsley, NY). Snoring was monitored by a microphone 3 feet from the patient and maximum decibels recorded (Realistics 33-2050; Tandy Corp., Fort Worth, TX) . The data were recorded on a 15-channel polygraph (Model 78; Grass Instrument Co., Quincy, MA) at a paper speed of 10 mm/s. Sleep stage and respiratory variables were analyzed independently. A microcomputer (PCAT; IBM, Boca Raton, FL) cont inuously 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 were merged with the remainder 0 f 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), number of apnea and hypopnea/TST (respiratory disturbance index), maximal snoring, and mean and minimal overnight Sa02.
Magnetic Resonance Imaging Each patient had an awake MRI scan of the upper airway performed before and after 4 to 6 wk of chronic nasal CPAP therapy. At the time of each scan, the patient was weighed and had a clinical inspection of the oropharynx. All scans were performed without nasal CPAP therapy.Scans wereperformed at 8 A.M. after an overnight sleep to optimize the likelihood of detecting upper airway edema. The patient's head was placed in a holding frame with the soft tissue Frankfort plane (tragus of the ear to soft tissue orbitale) perpendicular to the floor and the chin supported by a chin strap. Scans were performed during tidal breathing. The patient was instructed not to move or talk during the scan and to swallow in a manner consistent with comfort. Verbal communication was maintained with the patient, who responded by act ivating a call button to confirm wakefulness throughout the scan. MRI scans were performed using a Picker Vista'" MR 1100 superconducting scanner (Picker International, Cleveland , OH) operating at a field strength of 0.15T. Sagittal and axial images were ob tained with a standard head coil. Using a balanced spin echo (SE) pulse sequence with a repetition time (Th) of 2,066 ms and echo time (Ts) of 40 ms, contiguous 5-mm thick slices were obtained from the Frankfort plane to the sixth cervical vertebra to viewdetailed anatomy (figure I). The acquis ition time was 17.6 min using an acquisition matrix of 512 x 256 views with two excitations. A short tau, or inversion time (11), inversion recovery (STIR) pulse sequence was obtained at IO-mmintervals from the Frankfort plane to the sixth cervical vertebra to assessupper airwayedema. The parameters used in the STIR sequences were a 'll of 125 ms,
Fig. 1. Midline sagittal magnetic resonance image from the spin echo sequence with superimposed 5-mm grid indicating the levels of the axial slices .
a Th of 2,066 ms, a Th of 40 ms using an acquisition matrix of 512 x 128viewswith four excitations, and an acquisition time of 17.6 min. A short 11was selected to suppress signal from fatty tissue with its shorter T I relaxation time and produce an additive effect for tissues with longer T I and T 2 relaxation times. In the clinical context of these scans, this improves the sensitivity for detection of highintensity signal from water. The axial SE and STIR sequence scans wereperformed at identical levels in each patient before and after nasal CPAP by using the midline sagittal image and grid (figure 1) from the prenasal CPAP scan to ensure that images were obtained at the same leveland angle on the postnasal CPAP scan. Tracings of each of the slices were made from the SE sequences on acetate paper for the airway, tongue, and soft palate (figure 2) by one investigator, who was blinded to the sequence of the scans. The tongue was delineated to include all intrinsic and extrinsic muscles (genioglossus, hyoglossus, 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 right and left nasal passages, the nasopharynx, oropharynx, and hypopharynx. The nasopharynx extended from the roof of the airway to the level of the hard palate. The oropharynx extended down to the uppermost 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 cross-hair cursor was used to enter the contour of each structure into a computer (lOooE Series; HewlettPackard Co., Andover, MA), as has been previously described in detail (8). All contours were corrected to real size values by adjustments of a 5-cm scale digitized on the images. The slices were oriented and stacked at 5-mm intervals in accordance with a 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 (5 mm), for the upper airway, tongue, and soft palate were calculated as has been previously described in deta il (8). The STIR images were examined using a visual grading system to assess upper airway mucosal water content. Mucosal water was quantitated by observing the amount of highintensit y signal in the upper airway mucosa (figure 3). An arbitrary scale of 1 to 4 was developed to indicate a range of mucosal water from minimal to severe. The upper airway was divided into four segments (nasal cavities, nasopharynx, oropharynx, and hypopharynx) corresponding to the anatomic divisions outlined for the analysis of the SE sequences. All scans were examined in random
Fig. 2. Axial slice at the level of the oropharynx from a spin echo sequence image of a patient with obstructive sleep apnea (T = tongue; P = soft palate; 0 = oropharynx).
MRI OF THE UPPER AIRWAY IN OBSTRUCTIVE SLEEP APNEA
941
Fig. 3. Inversion recovery sequence magnetic resonance imaging scans of Patient 3 at the same level of the oropharynx before. with a mucosal water score of 4 (figure 3a). and afte r. with a mucosal water score of 1 (figure 3b), chronic nasal CPAP therapy. The arrow in figure 3a indicates increased mucosal water in the posterolateral oropharyngeal wall.
sequence by three independent obser vers in a blinded fashion for the intensity, thickness, and craniocaudal extent of signal within each segment. Each segment was scored separately, and the scores were then added to obtain a total upper airway score. Scores for each segment and the total upper airway scoreswere averaged and corrected to the nearest whole number. The total upper airway mucosal water score wasdivided by 4. The prenasal CPAP STIR images were repeated in two patients following the administration of an oral anti cholinergic agent (propantheline, 30 mg) to inhibit upper airway secretions. Scans were repeated when the patients confirmed that they had a dry mouth.
Nasal CPAP Following the initial overnight polysomnogram and upper airway MRI scan, patients were commenced on nasal CPAP therapy (Healthdyne 7001, Marietta, GA)at a pressure (range 8 to 10 em H 2 0 ) demonstrated to prevent apnea and snoring in all sleep stages and body positions. They were instructed to use the nasal CPAP during sleep on a regular and consistent basis until the repeat MRI scan. Statistical Analysis A two-tailed Student's t test for analysis of paired data was used to compare weight, polysomnographic data, and MRI measurements before and after chronic nasal CPAP therapy. A two-tailed Wilcoxon matched-pairs signed-rank test was used to compare upper airway mucosal water scores before and after chronic nasal CPAP therapy. No correction was made for multiple comparisons because the variables were not completely indepen-
dent. Kendall's coefficient of concordance was used to assess agreement between the three observers in visually grading the MRI scans. Linear regression analysis was used to determine the relationship between prenasal CPAP apnea severity, improvement in apnea severity, or change in mucosal water score and changes in upper airway morphology.
Results Five male patients (age range 40 to 71 yr) were studied. They were obese and had moderate to severe symptomatic OSA. The 4 to 6 wk of nasal CPAP therapy effectively relieved both OSA and snoring and was not associated with any significant weight change (table 1). Three patients were smokers and three drank
alcohol regularly, and these habits did not change during the study. The minimum pharyngeal crosssectional area was in the oropharynx in three patients and the hypopharynx in two patients, and its location remained unchanged in each patient following chronic nasal CPAP therapy. The minimal pharyngeal cross-sectional area (CSAmin) increased from 14 ± 6 to 40 ± 14 mm" (mean ± SD) (p < 0.05) and pharyngeal volume increased from 9.3 ± 3.3 to 1l.8 ± 4.0 ml (p < 0.05) (figure 4) following chronic nasal CPAP therapy. The majority of the increase in pharyngeal volume occurred in the oropharynx, which increased from 2.6 ± 1.7 to 5.3 ± 2.7 ml (p < 0.01), whereas there was no significant change in either nasopharyngeal or hypo pharyngeal volume (figure 5). Tongue volume decreased from 63.5 ± 9.1 to 60.1 ± lOA ml (p < 0.01). The soft palate volume decreased from 12.9 ± 4.8 to 11.4 ± 2.4 ml, but this change was not statistically significant (figure 6). The individual data for pharyngeal cross-sectional area and upper airway volumes before and after chronic nasal CPAP therapy are shown in table 2. There was no relationship between prenasal CPAP apnea severity or improvement in apnea severity during nasal CPAP therapy and changes in upper airway morphology following chronic nasal CPAP therapy. The change in minimal pharyngeal CSA was positively correlated with the change in mucosal water score in the corresponding anatomic region (r = 0.84, p < 0.05). Clinical examination of the oropharynx revealed improvement in the erythema and swelling in all five patients after chronic nasal CPAP therapy. The upper airway mucosal water scores based on the visual grading system of the STIR sequence MRI scans are shown in tables
TABLE 1 EFFECT OF CHRONIC NASAL CPAP THERAPY ON POLYSOMNOGRAPHIC VARIABLES AND WEIGHT IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
Total apnea time . %TST Apnea index. Apnealh TST Respiratory disturbance index. apnea + hypopnealTST Sao, mean, % Sao, minimum. % Snor ing, decibels Weight, kg
Before Nasal CPAP
With Nasal CPAP
p Value
22 ± 13 25 ± 15
1 ± 1 1 ± 1
< 0.001 < 0.001
3:!:6 94 ± 3 84 :!: 4 0 92 :!: 11
< 0.001 < 0.001 < 0.001 < 0.001
52 84 47 65 91
:!: 20 ± 8 ± 17 :I: 8 :!: 12
NS
Definition of abbreviations: CPAP = continuous positive airway pressure; TST = total sleep time; NS = nonsignificant. • Mean ± SO.
RYAN, LOWE, LI, AND FLEETHAM
942
ynx, where the score decreased from 3.4 ± 0.5 to 1.2 ± 0.4 (p < 0.05). The STIR sequence MRI images of the oropharynx of Patient 3 before and after chronic nasal CPAP therapy are shown in figure 3. There was minimal interobserver difference in the visual grading of the MRI scans (p < 0.005).
50
P < 0.05 I=SEM 40
30
Discussion
20
10
PRECPAP
POSTCPAP
PHARYNGEAL CSA
MIN
Fig. 4. Mean minimal pharyngeal cross-sectional area (CSAmin), and pharyngeal volume before and after chronic nasal CPAP therapy in five patients with obstructive sleep apnea.
P = NS 1= SEM
~
p < 0.01 I=SEM
3
~
2
"
W
"3
§?
3
W
:::J
PRE CPAP
POSTCPAP
PRE CPAP
POSTCPAP
OROPHARYNX
NASOPHARYNX
PRECPAP
POSTCPAP
HYPOPHARYNX
Fig. 5. Mean volumes of nasopharynx, oropharynx, and hypopharynx before and after chronic nasal CPAP therapy in five patients with obstructive sleep apnea.
80
P < 0.Q1 1= SEM 70
30
p ; NS
1= SEM
E
Fig. 6. Mean volumes of tongue and soft palate before and after chronic nasal CPAP therapy in five patients with obstructive sleep apn ea.
w
3" §?
80
50
PRE CPAP
POST CPAP
TONGUE
PRE CPAP
POST CPAP
SOFT PALATE
3 and 4. There was no difference in the upper airway mucosal water scores before and after the administration of propantheline (table 3). Total upper air-
way mucosal water score decreased from 2.8 ± 0.2 to 1.8 ± 0.5 (p < 0.05) following chronic nasal CPAP therapy (table 4). This was most marked in the orophar-
These results show that in patients with OSA, chronic nasal CPAP therapy during sleep produces changes in awake upper airway morphology that are independent of the direct pressure effect of nasal CPAP. Pharyngeal volume and minimal pharyngeal CSA increase and tongue volume decreases. There is an alteration in the visual appearance of the MRI scans, which suggests that these changes in upper airway morphology are due to a decrease in edema. There is a marked interpatient variability in the change in minimum pharyngeal CSA following chronic nasal CPAP therapy that may be related to differences in upper airway edema before nasal CPAP therapy. This study also demonstrates the potential value of MRI in the serial evaluation of the upper airway in patients with OSA. Our measurements of pharyngeal and tongue volume before nasal CPAP therapy are similar to our previous report, which utilized computer tomography (CT) to assess upper airway morphology in patients with OSA (8). The validity of both CT and MRI volumetric measurements of upper airway structures has been confirmed by comparison with volumes obtained by fluid displacement of tongues dissected from cadavers (9) and control objects (10). The pharyngeal volumes in our study are smaller than those recently reported by Abbey and associates with MRI (10). These two studies are not comparable, however, because Abbey and associates studied patients with chronic snoring rather than OSA and there were differences in the boundaries of the pharynx between the two studies. A recent study utilizing MRI reported fat deposits in the posterior oropharynx and soft palate of patients with OSA (11). We selected STIR sequences to suppress signal from fat, thereby improving the sensitivity of highintensity signals for tissue water (12). Our visual grading system of the MRI scans was both arbitrary and subjective. This, however, is no different from a variety of previously reported visual lung radiologic and pathologic scores (13), and there was good agreement between our three
MRI OF THE UPPER AIRWAY IN OBSTRUCTIVE SLEEP APNEA
943
TABLE 2
The efficacy of nasal CPAP is dependent on the degree of positive pressure applied to the upper airway. Sullivan and coworkers originally proposed (4) that nasal CPAP acts as a "pneumatic splint" to the upper airway. This explanation is supported by studies using endoscopy (14), acoustic reflection (15), CT (16), and MRI (10),which have shown increases in pharyngeal size during nasal CPAP. Nasal CPAP could also increase airway patency by reflexly activating upper airway muscles, but recent studies have shown that upper airway muscle activity is either reduced or unchanged (16, 17) during nasal CPAP therapy. There is also some evidence that nasal CPAP has a separate effect not directly related to the acute dilatation of the upper airway. Patients with OSA may initially show a reduction in their apnea index if nasal CPAP is discontinued after a period of chronic treatment (3, 5, 6). The OSA then returns to the previous degree of severity within several days. The increase in pharyngeal size which we have demonstrated following chronic nasal CPAP therapy supports the concept that there is a separate beneficial effect of nasal CPAP independent of the acute dilatation of the upper airway. Furthermore, the change in the visual appearance of the STIR gap MRI images of the oropharynx supports the hypothesis (18)that nasal CPAP therapy may help relieve upper airway obstruction in patients with OSA by reducing upper airway edema. A reduction in upper airway edema could increase upper airway size by sev-
PHARYNGEAL CROSS-SECTIONAL AREA AND UPPER AIRWAY VOLUMES BEFORE AND AFTER CHRONIC NASAL CPAP THERAPY IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
Pharyngeal CSAmin Location Pre Post mm' Pre Post Nasal cavities volume, ml Pre Post Nasopharyngeal volume, ml Pre Post Oropharyngeal volume, ml Pre Post Hypopharyngeai volume, ml Pre Post Pharyngeal volume, rnl Pre Post Tongue volume, ml Pre Post Soft palate volume, ml Pre Post
Mean ± SD
2
3
4
5
0 0
H H
0 0
0 0
H H
5 46
19 17
9 42
18 55
18 38
13.3 10.8
13.7 13.6
11.5 16.7
14.3 18.4
7.7 13.8
12.1 ± 2.7 14.7 ± 3.0
2.3 2.6
5.9 4.0
4.6 4.7
1.9 1.2
0.4 0.7
3.0 ± 2.2 2.6 ± 1.7
1.7 4.2
1.6 5.2
3.0 5.9
5.5 9.3
1.4 1.9
2.6 ± 1.7 5.3 ± 2.7t
5.5 3.7
3.6 2.7
2.2 3.0
4.9 6.2
1.9 3.3
3.6 ± 1.6 3.8 ± 1.4
9.5 10.5
11.1 11.9
9.8 13.9
12.2 16.7
3.7 5.9
9.3 ± 3.3 11.8 ± 4.0"
55.4 50.5
73.6 70.8
73.2 71.9
58.2 55.0
56.9 52.5
63.5 ± 9.1 60.1 ± 10At
10.1 10.1
10.2 12.6
21.1 14.9
13.2 10.5
10.1 8.7
12.9 ± 4.8 11.4 ± 2.4
Patient
14 ± 6 40 ± 14"
Definition of abbreviations: Pharyngeal CSAmin ee minimal pharyngeal cross-sectional area;Pre ~ beforechronictherapywith nasal continuous positive airway pressure; Post ~ after chronictherapywith nasal continuous positive airway pressure; 0 ~ oropharynx; H ~ hypopharynx; pharyngeal volume ~ nasopharyngeal + oropharyngeal + hypopharyngeal volume. • Differentfrom prsnasal CPAP measurement, p < 0.05. t Differentfrom prenasal CPAP measurement, p < 0.01.
independent blinded observers. We believethe change in the visual appearance of the STIR gap MRI scans is most likely due to a decrease in upper airway ede-
rna rather than mucosal surface fluid, because the appearance of the upper airway was not altered by administration of an anticholinergic agent.
TABLE 3
TABLE 4
UPPER AIRWAY MUCOSAL WATER SCORES BEFORE AND AFTER PROPANTHELINE
UPPER AIRWAY MUCOSAL WATER SCORES BEFORE AND AFTER CHRONIC NASAL CPAP THERAPY IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
2
Patient
Mean
2
Patient Mucosal water score" Nasal cavities Pre Post Nasopharynx Pre Post Oropharynx Pre Post Hypopharynx Pre Post Total upper airway Pre Post
4 4
3 2
3.5 3.0
3 3
2 2
2.5 2.5
3 3
4 4
3.5 3.5
2 2
1 2
1.5 2.0
3.0 3.0
2.5 2.5
2.8 2.8
Definition of abbreviations: Pre ~ beforeoral administration of propantheline, 30 mg; Post ~ after oral administration of propantheline, 30 mg. • Mucosal waterscore:1 ~ minimal;2 ~ mild; 3 ~ moder-
ate; 4 = severe.
Mucosal water score" Nasal cavities Pre Post Nasopharynx Pre Post Oropharynx Pre Post Hypopharynx Pre Post Total upper airway Pre Post
3
4
5
Mean ± SD
4 4
3 4
3 1
3 1
4 1
3.4 ± 0.5 2.2 ± 1.6
3 2
2 2
2 1
4 4
3 4
2.8 ± 0.8 2.6 ± 1.3
3 1
4 1
4 1
3 1
3 2
3.4 ± 0.5 1.2 ± OAt
2 2
1.6 ± 0.5 1.4 ± 0.5
3.0 2.2
2.8 ± 0.2 1.8 ± 0.5t
2 2 3.0 2.2
2 1 2.5 2.0
2.7 1.0
2.7 1.7
Definition of abbreviations: Pre ~ beforechronictherapywith nasalcontinuous positiveairwaypressure; Post = after chronic therapy with nasal continuous positive airway pressure. • Mucosal water score: 1 = minimal; 2 = mild, 3 = moderate; 4 = severe . t Differentfrom prenasal CPAPscore, p < 0.05.
RYAN, LOWE, L1, AND FLEETHAM
944
eral mechanisms. A reduction in upper airway wall thickness due to a decrease in edema would increase upper airway luminal dimensions. Alternatively, a decrease in upper airway edema may reduce the collapsibility of the upper airway and produce an increase in upper airway size for a given level of upper airway dilator muscle activity (1). Although both the tongue and soft palate decreased in volume following chronic nasal CPAP therapy, this change was only statistically significant for tongue volume. This apparent discrepancy may be due to problems differentiating between the tongue and soft palate on MRI scans as these two structures lie in close apposition when the mouth is closed. There are other potential explanations for the change in upper airway size following chronic nasal CPAP therapy. Changes in body weight or sleep state between MRI scans do not account for our results because body weight did not change during the 4- to 6-wk period and both MRI scans occurred during wakefulness. The change in airway size could be related to the improvement in sleep architecture and gas exchange associated with relief of the OSA. Sleep fragmentation has been demonstrated to reduce respiratory responses to chemical stimuli (19),and long-term nasal CPAP therapy improves hypercapnic ventilatory responses in patients with OSA (20). Increased respiratory drive to the upper airway dilator muscles could increase awake airway size following chronic nasal CPAP therapy, but there is no evidence to support this hypothesis in our study. We used MRI of the upper airway to advance our understanding of the pathogenesis of OSA and of the mechanisms of action of nasal CPAP therapy. We
showed that upper airway morphology can be accurately and repeatedly assessed in a noninvasive way using MRI. The avoidance of ionizing radiation and the improved definition of areas of increased tissue water content, such as edema, and fat make MRI particularly suitable for sequential studies of the upper airway. Magnetic resonance imaging may prove to be a useful method for assessing upper airway changes produced by other treatments for OSA, such as weight reduction, upper airway surgery, and intraoral appliances. Acknowledgment The writers thank Mary Wong for her technical assistance in the analysis of the data, Cathy Jardine, Leslie Castley, Sharon Hall, and Monique Genton for performing the MRI scans, and BerniceRobillard for her secretarial assistance. References 1. Remmers JE, de Groot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 1978; 44:931-8. 2. Rodenstein DO, Stanescu DC. The soft palate and breathing. Am RevRespirDis 1986;134:311-25. 3. Sullivan CE, Grunstein RR. Continuous positive airways pressure in sleep-disordered breathing. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine. Philadelphia: W. B. Saunders, 1989: 559-70. 4. Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversalof obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet 1981; 1:862-5. 5. Schneider H, Becker H, Boke M, et al. Reexposition of NCPAP therapy on sleepapnea patients. Eur Respir J 1989; 2(Suppl. 5:402S). 6. Olsen LG, Rolfe IE, Saunders NA. Tolerance of nasal CPAP treatment of obstructive sleep apnea, and its effects on disease severity. Am Rev Respir Dis 1990; 141:866A. 7. Pykett IL, Newhouse JH, Buonanno FS, et al. Principles of nuclear magnetic resonance imaging. Radiology 1982; 143:157-68. 8. LoweAA, 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. 9. Roehm EO. Computer tomographic measurement of tongue volume relative to its surrounding space. M.Sc. Thesis, University of Manitoba, 1981. 10. 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. 11. Horner RL, Mohiaddin RH, Lowell DG, et al. Sites and sizes of fat deposits around the pharynx in obese patients with obstructive sleep apnoea and weight matched controls. Eur Respir J 1989; 2:613-22. 12. Bydder GM, Young IR. MR imaging: clinical use of the inversion recovery sequence. J Comput Assist Tomogr 1985; 9:659-75. 13. Bergin C, Muller N, Nichols DM, et al. The diagnosisof emphysema. A computer tomographicpathologic correlation. Am Rev Respir Dis 1986; 133:541-6. 14. Popper RA, Leidlinger MJ, Williams AJ. Endoscopic observations of the pharyngeal airwayduring treatment of obstructive sleep apnea with nasal continuous positive airway pressure-a pneumatic splint. West J Med 1986; 144:83-5. 15. Brown I, McClean P, Boucher R, Zamel N, Hoffstein V.Changes in pharyngeal cross-sectional area with posture and application of continuous positive airway pressure in patients with obstructive sleep apnea. Am Rev Respir Dis 1987; 136: 628-32. 16. Kuna ST, Bedi DG, Ryckman C. Effect of nasal airway positive pressure on upper airway size and configuration. Am Rev Respir Dis 1988; 138: 969-75. 17. Strohl KP, Redline S. Nasal CPAP therapy, upper airway muscle activation, and obstructive sleep apnea. Am Rev Respir Dis 1986; 134:555-8. 18. Weil rv, Cherniack NS, Dempsey lA, et al. NHLBI workshop summary. Respiratory disorders of sleep. Pathophysiology, clinicalimplications and therapeutic approaches. Am Rev Respir Dis 1987; 136:755-61. 19. Coopers K, Phillips B. Effect of short-term sleep loss on breathing. J Appl Physiol 1982; 53: 855-8. 20. Berthon-Jones M, Sullivan CEoTime course of change in ventilatory response to CO 2 with long term CPAP therapy for obstructive sleep apnea. Am Rev Respir Dis 1987; 135:144-7.
C. FRANCIS RYAN,4 ALAN A. LOWE, DAVID LI, and JOHN A. FLEETHAM Introduction
Obstructive sleep apnea (OSA) 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 load due to a narrow airway predisposes to upper airway occlusion. Snoring is the hallmark of patients with OSA and is caused by partial upper airway obstruction and high-frequency oscillation of the soft palate (2). Many patients with OSA complain of pharyngeal swelling on awakening and appear to have an inflamed and swollen oropharynx, which improves as the day progresses (3). Upper airway edema may occur secondary to OSA and subsequently exacerbate the OSA by causing further upper airway narrowing. Nasal continuous positive airway pressure (CPAP) is a very effective form of treatment for OSA. It was initially proposed that nasal CPAP provides a physical "pressure" splint preventing upper airway occlusion by pushing the soft palate and tongue forward and away from the posterior oropharyngeal wall (4). Preliminary reports indicate that patients initially demonstrate an improvement in the severity of their OSA if nasal CPAP is discontinued after a period of chronic treatment (5, 6). However, after a few days the sleep apnea returns to the previous degree of severity. One explanation for this phenomenon is that nasal CPAP relieves the snoring, thereby allowing a resolution of upper airway narrowing due to edema, which gradually reaccumulates when the nasal CPAP is discontinued. Magnetic resonance imaging (MRI) produces high-resolution images without the use of ionizing radiation. MRI depends on the tissue relaxation times and regional concentrations of hydrogen nu-
SUMMARY Magnetic resonance Imaging (MRI) provides high-resolution Images of the upper airway and Is useful for assessing conditions associated with Increased tissue water content. Todetermine whether nasal continuous positive airway pressure (CPAP)changes awake upper airway morphology in obstructive sleep apnea (OSA), we performed awake upper airway MRI scans on five male patients with moderate to severe OSA before and after 4 to 6 wk of nasal CPAP therapy. MRI scans were performed using spin echo pulse sequences to examine detailed anatomy and Inversion recovery sequences to assess mucosal water content. Patients did not have nasal CPAP applied during the MRI scans. Axial and sagittal Images were obtained, and tracings were made of the upper airway, tongue, and soft palate. Utilizing computer graphics, cross-sectional areas and volumes were calculated for each anatomic structure. A subjective grading system was used to assess upper airway mucosal water content. Pharyngeal volume and minimum pharyngeal cross-sectional area increased (p < 0.05) and tongue volume decreased (p < 0.01)following chronic nasal CPAPtherapy. The Increase In pharyngeal volume occurred mainly in the oropharynx (p < 0.01). Upper airway mucosal water content decreased In the oropharynx (p < 0.05). We conclude that chronic nasal CPAPtherapy during sleep In patients with OSA produces changes in awake upper airway morphology. These changes may be due to resolution of upper airway edema. The upper airway of patients with OSA AM REV RESPIR DIS 1991; 144:939-944 can be accurately and repeatedly assessed using MRI.
cleito produce an image (7). Because each water molecule has two hydrogen nuclei, MRI is ideally suited to assessing conditions associated with increased tissue water content. We performed awake MRI scans on patients with OSA before and after chronic nasal CPAP therapy to determine whether upper airway morphology changes with chronic nasal CPAP therapy. Methods Patients Five male patients with newly diagnosed moderate to severe symptomatic OSA, who elected to be treated with nasal CPAP, were recruited to the study. Patients whose weight wasgreater than 115 kg wereexcluded because of the size limitations of the MRI scanner. Patients who were edentulous were also excluded because of difficulty in defining the boundaries of the tongue on the upper airway scan. Each patient gaveinformed written consent, and the study protocol was approved by the University of British Columbia Clinical Screening Committee for Research Involving Human Subjects. Polysomnography Each patient had a detailed overnight polysomnogram performed before and after a
period of chronic nasal CPAP therapy. The initial study was performed without nasal CPAP, and the second study was performed with nasal CPAP. Sleep and its various stages was documented by standard electroencephalographic (EEG), electrooculographic (EOG), and e1ectromyographic (EMG) criteria. EEG was recorded with electrodes applied at C3-A2 and CcA, (according to the International 1020 system), and EMG activity was recorded from the submental muscles. Apneas were documented by a Beckman LB-2 infrared CO 2 analyzer (Beckman Instruments Inc., Schillar Park, IL), which recorded from both the nose and mouth. A singleECG (modified V2)
(Receivedin originalform November 15, 1989 and in revised form May 3, 1991) 1 From the Departments of Medicine, Clinical Dental Sciences, and Radiology, University of British Columbia, Vancouver, British Columbia, Canada. 2 Supported by the British Columbia Lung Association and the British Columbia Health Care Research Foundation. 3 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 British Columbia Lung Association Research Fellow.
939
RYAN, LOWE, LI, AND FLEETHAM
940
was monitored to detect card iac arrhythmias. Sao 2was monitored continuously with a Nellcor N-l00 pulse oximeter (Nellcor Inc., Hayward, CA). Chest wall movement was monitored by a respiratory inductive plethysmograph (Respitrace'"; Ambulatory Monitoring Equ ipment Inc., Ardsley, NY). Snoring was monitored by a microphone 3 feet from the patient and maximum decibels recorded (Realistics 33-2050; Tandy Corp., Fort Worth, TX) . The data were recorded on a 15-channel polygraph (Model 78; Grass Instrument Co., Quincy, MA) at a paper speed of 10 mm/s. Sleep stage and respiratory variables were analyzed independently. A microcomputer (PCAT; IBM, Boca Raton, FL) cont inuously 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 were merged with the remainder 0 f 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), number of apnea and hypopnea/TST (respiratory disturbance index), maximal snoring, and mean and minimal overnight Sa02.
Magnetic Resonance Imaging Each patient had an awake MRI scan of the upper airway performed before and after 4 to 6 wk of chronic nasal CPAP therapy. At the time of each scan, the patient was weighed and had a clinical inspection of the oropharynx. All scans were performed without nasal CPAP therapy.Scans wereperformed at 8 A.M. after an overnight sleep to optimize the likelihood of detecting upper airway edema. The patient's head was placed in a holding frame with the soft tissue Frankfort plane (tragus of the ear to soft tissue orbitale) perpendicular to the floor and the chin supported by a chin strap. Scans were performed during tidal breathing. The patient was instructed not to move or talk during the scan and to swallow in a manner consistent with comfort. Verbal communication was maintained with the patient, who responded by act ivating a call button to confirm wakefulness throughout the scan. MRI scans were performed using a Picker Vista'" MR 1100 superconducting scanner (Picker International, Cleveland , OH) operating at a field strength of 0.15T. Sagittal and axial images were ob tained with a standard head coil. Using a balanced spin echo (SE) pulse sequence with a repetition time (Th) of 2,066 ms and echo time (Ts) of 40 ms, contiguous 5-mm thick slices were obtained from the Frankfort plane to the sixth cervical vertebra to viewdetailed anatomy (figure I). The acquis ition time was 17.6 min using an acquisition matrix of 512 x 256 views with two excitations. A short tau, or inversion time (11), inversion recovery (STIR) pulse sequence was obtained at IO-mmintervals from the Frankfort plane to the sixth cervical vertebra to assessupper airwayedema. The parameters used in the STIR sequences were a 'll of 125 ms,
Fig. 1. Midline sagittal magnetic resonance image from the spin echo sequence with superimposed 5-mm grid indicating the levels of the axial slices .
a Th of 2,066 ms, a Th of 40 ms using an acquisition matrix of 512 x 128viewswith four excitations, and an acquisition time of 17.6 min. A short 11was selected to suppress signal from fatty tissue with its shorter T I relaxation time and produce an additive effect for tissues with longer T I and T 2 relaxation times. In the clinical context of these scans, this improves the sensitivity for detection of highintensity signal from water. The axial SE and STIR sequence scans wereperformed at identical levels in each patient before and after nasal CPAP by using the midline sagittal image and grid (figure 1) from the prenasal CPAP scan to ensure that images were obtained at the same leveland angle on the postnasal CPAP scan. Tracings of each of the slices were made from the SE sequences on acetate paper for the airway, tongue, and soft palate (figure 2) by one investigator, who was blinded to the sequence of the scans. The tongue was delineated to include all intrinsic and extrinsic muscles (genioglossus, hyoglossus, 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 right and left nasal passages, the nasopharynx, oropharynx, and hypopharynx. The nasopharynx extended from the roof of the airway to the level of the hard palate. The oropharynx extended down to the uppermost 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 cross-hair cursor was used to enter the contour of each structure into a computer (lOooE Series; HewlettPackard Co., Andover, MA), as has been previously described in detail (8). All contours were corrected to real size values by adjustments of a 5-cm scale digitized on the images. The slices were oriented and stacked at 5-mm intervals in accordance with a 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 (5 mm), for the upper airway, tongue, and soft palate were calculated as has been previously described in deta il (8). The STIR images were examined using a visual grading system to assess upper airway mucosal water content. Mucosal water was quantitated by observing the amount of highintensit y signal in the upper airway mucosa (figure 3). An arbitrary scale of 1 to 4 was developed to indicate a range of mucosal water from minimal to severe. The upper airway was divided into four segments (nasal cavities, nasopharynx, oropharynx, and hypopharynx) corresponding to the anatomic divisions outlined for the analysis of the SE sequences. All scans were examined in random
Fig. 2. Axial slice at the level of the oropharynx from a spin echo sequence image of a patient with obstructive sleep apnea (T = tongue; P = soft palate; 0 = oropharynx).
MRI OF THE UPPER AIRWAY IN OBSTRUCTIVE SLEEP APNEA
941
Fig. 3. Inversion recovery sequence magnetic resonance imaging scans of Patient 3 at the same level of the oropharynx before. with a mucosal water score of 4 (figure 3a). and afte r. with a mucosal water score of 1 (figure 3b), chronic nasal CPAP therapy. The arrow in figure 3a indicates increased mucosal water in the posterolateral oropharyngeal wall.
sequence by three independent obser vers in a blinded fashion for the intensity, thickness, and craniocaudal extent of signal within each segment. Each segment was scored separately, and the scores were then added to obtain a total upper airway score. Scores for each segment and the total upper airway scoreswere averaged and corrected to the nearest whole number. The total upper airway mucosal water score wasdivided by 4. The prenasal CPAP STIR images were repeated in two patients following the administration of an oral anti cholinergic agent (propantheline, 30 mg) to inhibit upper airway secretions. Scans were repeated when the patients confirmed that they had a dry mouth.
Nasal CPAP Following the initial overnight polysomnogram and upper airway MRI scan, patients were commenced on nasal CPAP therapy (Healthdyne 7001, Marietta, GA)at a pressure (range 8 to 10 em H 2 0 ) demonstrated to prevent apnea and snoring in all sleep stages and body positions. They were instructed to use the nasal CPAP during sleep on a regular and consistent basis until the repeat MRI scan. Statistical Analysis A two-tailed Student's t test for analysis of paired data was used to compare weight, polysomnographic data, and MRI measurements before and after chronic nasal CPAP therapy. A two-tailed Wilcoxon matched-pairs signed-rank test was used to compare upper airway mucosal water scores before and after chronic nasal CPAP therapy. No correction was made for multiple comparisons because the variables were not completely indepen-
dent. Kendall's coefficient of concordance was used to assess agreement between the three observers in visually grading the MRI scans. Linear regression analysis was used to determine the relationship between prenasal CPAP apnea severity, improvement in apnea severity, or change in mucosal water score and changes in upper airway morphology.
Results Five male patients (age range 40 to 71 yr) were studied. They were obese and had moderate to severe symptomatic OSA. The 4 to 6 wk of nasal CPAP therapy effectively relieved both OSA and snoring and was not associated with any significant weight change (table 1). Three patients were smokers and three drank
alcohol regularly, and these habits did not change during the study. The minimum pharyngeal crosssectional area was in the oropharynx in three patients and the hypopharynx in two patients, and its location remained unchanged in each patient following chronic nasal CPAP therapy. The minimal pharyngeal cross-sectional area (CSAmin) increased from 14 ± 6 to 40 ± 14 mm" (mean ± SD) (p < 0.05) and pharyngeal volume increased from 9.3 ± 3.3 to 1l.8 ± 4.0 ml (p < 0.05) (figure 4) following chronic nasal CPAP therapy. The majority of the increase in pharyngeal volume occurred in the oropharynx, which increased from 2.6 ± 1.7 to 5.3 ± 2.7 ml (p < 0.01), whereas there was no significant change in either nasopharyngeal or hypo pharyngeal volume (figure 5). Tongue volume decreased from 63.5 ± 9.1 to 60.1 ± lOA ml (p < 0.01). The soft palate volume decreased from 12.9 ± 4.8 to 11.4 ± 2.4 ml, but this change was not statistically significant (figure 6). The individual data for pharyngeal cross-sectional area and upper airway volumes before and after chronic nasal CPAP therapy are shown in table 2. There was no relationship between prenasal CPAP apnea severity or improvement in apnea severity during nasal CPAP therapy and changes in upper airway morphology following chronic nasal CPAP therapy. The change in minimal pharyngeal CSA was positively correlated with the change in mucosal water score in the corresponding anatomic region (r = 0.84, p < 0.05). Clinical examination of the oropharynx revealed improvement in the erythema and swelling in all five patients after chronic nasal CPAP therapy. The upper airway mucosal water scores based on the visual grading system of the STIR sequence MRI scans are shown in tables
TABLE 1 EFFECT OF CHRONIC NASAL CPAP THERAPY ON POLYSOMNOGRAPHIC VARIABLES AND WEIGHT IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
Total apnea time . %TST Apnea index. Apnealh TST Respiratory disturbance index. apnea + hypopnealTST Sao, mean, % Sao, minimum. % Snor ing, decibels Weight, kg
Before Nasal CPAP
With Nasal CPAP
p Value
22 ± 13 25 ± 15
1 ± 1 1 ± 1
< 0.001 < 0.001
3:!:6 94 ± 3 84 :!: 4 0 92 :!: 11
< 0.001 < 0.001 < 0.001 < 0.001
52 84 47 65 91
:!: 20 ± 8 ± 17 :I: 8 :!: 12
NS
Definition of abbreviations: CPAP = continuous positive airway pressure; TST = total sleep time; NS = nonsignificant. • Mean ± SO.
RYAN, LOWE, LI, AND FLEETHAM
942
ynx, where the score decreased from 3.4 ± 0.5 to 1.2 ± 0.4 (p < 0.05). The STIR sequence MRI images of the oropharynx of Patient 3 before and after chronic nasal CPAP therapy are shown in figure 3. There was minimal interobserver difference in the visual grading of the MRI scans (p < 0.005).
50
P < 0.05 I=SEM 40
30
Discussion
20
10
PRECPAP
POSTCPAP
PHARYNGEAL CSA
MIN
Fig. 4. Mean minimal pharyngeal cross-sectional area (CSAmin), and pharyngeal volume before and after chronic nasal CPAP therapy in five patients with obstructive sleep apnea.
P = NS 1= SEM
~
p < 0.01 I=SEM
3
~
2
"
W
"3
§?
3
W
:::J
PRE CPAP
POSTCPAP
PRE CPAP
POSTCPAP
OROPHARYNX
NASOPHARYNX
PRECPAP
POSTCPAP
HYPOPHARYNX
Fig. 5. Mean volumes of nasopharynx, oropharynx, and hypopharynx before and after chronic nasal CPAP therapy in five patients with obstructive sleep apnea.
80
P < 0.Q1 1= SEM 70
30
p ; NS
1= SEM
E
Fig. 6. Mean volumes of tongue and soft palate before and after chronic nasal CPAP therapy in five patients with obstructive sleep apn ea.
w
3" §?
80
50
PRE CPAP
POST CPAP
TONGUE
PRE CPAP
POST CPAP
SOFT PALATE
3 and 4. There was no difference in the upper airway mucosal water scores before and after the administration of propantheline (table 3). Total upper air-
way mucosal water score decreased from 2.8 ± 0.2 to 1.8 ± 0.5 (p < 0.05) following chronic nasal CPAP therapy (table 4). This was most marked in the orophar-
These results show that in patients with OSA, chronic nasal CPAP therapy during sleep produces changes in awake upper airway morphology that are independent of the direct pressure effect of nasal CPAP. Pharyngeal volume and minimal pharyngeal CSA increase and tongue volume decreases. There is an alteration in the visual appearance of the MRI scans, which suggests that these changes in upper airway morphology are due to a decrease in edema. There is a marked interpatient variability in the change in minimum pharyngeal CSA following chronic nasal CPAP therapy that may be related to differences in upper airway edema before nasal CPAP therapy. This study also demonstrates the potential value of MRI in the serial evaluation of the upper airway in patients with OSA. Our measurements of pharyngeal and tongue volume before nasal CPAP therapy are similar to our previous report, which utilized computer tomography (CT) to assess upper airway morphology in patients with OSA (8). The validity of both CT and MRI volumetric measurements of upper airway structures has been confirmed by comparison with volumes obtained by fluid displacement of tongues dissected from cadavers (9) and control objects (10). The pharyngeal volumes in our study are smaller than those recently reported by Abbey and associates with MRI (10). These two studies are not comparable, however, because Abbey and associates studied patients with chronic snoring rather than OSA and there were differences in the boundaries of the pharynx between the two studies. A recent study utilizing MRI reported fat deposits in the posterior oropharynx and soft palate of patients with OSA (11). We selected STIR sequences to suppress signal from fat, thereby improving the sensitivity of highintensity signals for tissue water (12). Our visual grading system of the MRI scans was both arbitrary and subjective. This, however, is no different from a variety of previously reported visual lung radiologic and pathologic scores (13), and there was good agreement between our three
MRI OF THE UPPER AIRWAY IN OBSTRUCTIVE SLEEP APNEA
943
TABLE 2
The efficacy of nasal CPAP is dependent on the degree of positive pressure applied to the upper airway. Sullivan and coworkers originally proposed (4) that nasal CPAP acts as a "pneumatic splint" to the upper airway. This explanation is supported by studies using endoscopy (14), acoustic reflection (15), CT (16), and MRI (10),which have shown increases in pharyngeal size during nasal CPAP. Nasal CPAP could also increase airway patency by reflexly activating upper airway muscles, but recent studies have shown that upper airway muscle activity is either reduced or unchanged (16, 17) during nasal CPAP therapy. There is also some evidence that nasal CPAP has a separate effect not directly related to the acute dilatation of the upper airway. Patients with OSA may initially show a reduction in their apnea index if nasal CPAP is discontinued after a period of chronic treatment (3, 5, 6). The OSA then returns to the previous degree of severity within several days. The increase in pharyngeal size which we have demonstrated following chronic nasal CPAP therapy supports the concept that there is a separate beneficial effect of nasal CPAP independent of the acute dilatation of the upper airway. Furthermore, the change in the visual appearance of the STIR gap MRI images of the oropharynx supports the hypothesis (18)that nasal CPAP therapy may help relieve upper airway obstruction in patients with OSA by reducing upper airway edema. A reduction in upper airway edema could increase upper airway size by sev-
PHARYNGEAL CROSS-SECTIONAL AREA AND UPPER AIRWAY VOLUMES BEFORE AND AFTER CHRONIC NASAL CPAP THERAPY IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
Pharyngeal CSAmin Location Pre Post mm' Pre Post Nasal cavities volume, ml Pre Post Nasopharyngeal volume, ml Pre Post Oropharyngeal volume, ml Pre Post Hypopharyngeai volume, ml Pre Post Pharyngeal volume, rnl Pre Post Tongue volume, ml Pre Post Soft palate volume, ml Pre Post
Mean ± SD
2
3
4
5
0 0
H H
0 0
0 0
H H
5 46
19 17
9 42
18 55
18 38
13.3 10.8
13.7 13.6
11.5 16.7
14.3 18.4
7.7 13.8
12.1 ± 2.7 14.7 ± 3.0
2.3 2.6
5.9 4.0
4.6 4.7
1.9 1.2
0.4 0.7
3.0 ± 2.2 2.6 ± 1.7
1.7 4.2
1.6 5.2
3.0 5.9
5.5 9.3
1.4 1.9
2.6 ± 1.7 5.3 ± 2.7t
5.5 3.7
3.6 2.7
2.2 3.0
4.9 6.2
1.9 3.3
3.6 ± 1.6 3.8 ± 1.4
9.5 10.5
11.1 11.9
9.8 13.9
12.2 16.7
3.7 5.9
9.3 ± 3.3 11.8 ± 4.0"
55.4 50.5
73.6 70.8
73.2 71.9
58.2 55.0
56.9 52.5
63.5 ± 9.1 60.1 ± 10At
10.1 10.1
10.2 12.6
21.1 14.9
13.2 10.5
10.1 8.7
12.9 ± 4.8 11.4 ± 2.4
Patient
14 ± 6 40 ± 14"
Definition of abbreviations: Pharyngeal CSAmin ee minimal pharyngeal cross-sectional area;Pre ~ beforechronictherapywith nasal continuous positive airway pressure; Post ~ after chronictherapywith nasal continuous positive airway pressure; 0 ~ oropharynx; H ~ hypopharynx; pharyngeal volume ~ nasopharyngeal + oropharyngeal + hypopharyngeal volume. • Differentfrom prsnasal CPAP measurement, p < 0.05. t Differentfrom prenasal CPAP measurement, p < 0.01.
independent blinded observers. We believethe change in the visual appearance of the STIR gap MRI scans is most likely due to a decrease in upper airway ede-
rna rather than mucosal surface fluid, because the appearance of the upper airway was not altered by administration of an anticholinergic agent.
TABLE 3
TABLE 4
UPPER AIRWAY MUCOSAL WATER SCORES BEFORE AND AFTER PROPANTHELINE
UPPER AIRWAY MUCOSAL WATER SCORES BEFORE AND AFTER CHRONIC NASAL CPAP THERAPY IN FIVE MALE PATIENTS WITH OBSTRUCTIVE SLEEP APNEA
2
Patient
Mean
2
Patient Mucosal water score" Nasal cavities Pre Post Nasopharynx Pre Post Oropharynx Pre Post Hypopharynx Pre Post Total upper airway Pre Post
4 4
3 2
3.5 3.0
3 3
2 2
2.5 2.5
3 3
4 4
3.5 3.5
2 2
1 2
1.5 2.0
3.0 3.0
2.5 2.5
2.8 2.8
Definition of abbreviations: Pre ~ beforeoral administration of propantheline, 30 mg; Post ~ after oral administration of propantheline, 30 mg. • Mucosal waterscore:1 ~ minimal;2 ~ mild; 3 ~ moder-
ate; 4 = severe.
Mucosal water score" Nasal cavities Pre Post Nasopharynx Pre Post Oropharynx Pre Post Hypopharynx Pre Post Total upper airway Pre Post
3
4
5
Mean ± SD
4 4
3 4
3 1
3 1
4 1
3.4 ± 0.5 2.2 ± 1.6
3 2
2 2
2 1
4 4
3 4
2.8 ± 0.8 2.6 ± 1.3
3 1
4 1
4 1
3 1
3 2
3.4 ± 0.5 1.2 ± OAt
2 2
1.6 ± 0.5 1.4 ± 0.5
3.0 2.2
2.8 ± 0.2 1.8 ± 0.5t
2 2 3.0 2.2
2 1 2.5 2.0
2.7 1.0
2.7 1.7
Definition of abbreviations: Pre ~ beforechronictherapywith nasalcontinuous positiveairwaypressure; Post = after chronic therapy with nasal continuous positive airway pressure. • Mucosal water score: 1 = minimal; 2 = mild, 3 = moderate; 4 = severe . t Differentfrom prenasal CPAPscore, p < 0.05.
RYAN, LOWE, L1, AND FLEETHAM
944
eral mechanisms. A reduction in upper airway wall thickness due to a decrease in edema would increase upper airway luminal dimensions. Alternatively, a decrease in upper airway edema may reduce the collapsibility of the upper airway and produce an increase in upper airway size for a given level of upper airway dilator muscle activity (1). Although both the tongue and soft palate decreased in volume following chronic nasal CPAP therapy, this change was only statistically significant for tongue volume. This apparent discrepancy may be due to problems differentiating between the tongue and soft palate on MRI scans as these two structures lie in close apposition when the mouth is closed. There are other potential explanations for the change in upper airway size following chronic nasal CPAP therapy. Changes in body weight or sleep state between MRI scans do not account for our results because body weight did not change during the 4- to 6-wk period and both MRI scans occurred during wakefulness. The change in airway size could be related to the improvement in sleep architecture and gas exchange associated with relief of the OSA. Sleep fragmentation has been demonstrated to reduce respiratory responses to chemical stimuli (19),and long-term nasal CPAP therapy improves hypercapnic ventilatory responses in patients with OSA (20). Increased respiratory drive to the upper airway dilator muscles could increase awake airway size following chronic nasal CPAP therapy, but there is no evidence to support this hypothesis in our study. We used MRI of the upper airway to advance our understanding of the pathogenesis of OSA and of the mechanisms of action of nasal CPAP therapy. We
showed that upper airway morphology can be accurately and repeatedly assessed in a noninvasive way using MRI. The avoidance of ionizing radiation and the improved definition of areas of increased tissue water content, such as edema, and fat make MRI particularly suitable for sequential studies of the upper airway. Magnetic resonance imaging may prove to be a useful method for assessing upper airway changes produced by other treatments for OSA, such as weight reduction, upper airway surgery, and intraoral appliances. Acknowledgment The writers thank Mary Wong for her technical assistance in the analysis of the data, Cathy Jardine, Leslie Castley, Sharon Hall, and Monique Genton for performing the MRI scans, and BerniceRobillard for her secretarial assistance. References 1. Remmers JE, de Groot WJ, Sauerland EK, Anch AM. Pathogenesis of upper airway occlusion during sleep. J Appl Physiol 1978; 44:931-8. 2. Rodenstein DO, Stanescu DC. The soft palate and breathing. Am RevRespirDis 1986;134:311-25. 3. Sullivan CE, Grunstein RR. Continuous positive airways pressure in sleep-disordered breathing. In: Kryger MH, Roth T, Dement WC, eds. Principles and practice of sleep medicine. Philadelphia: W. B. Saunders, 1989: 559-70. 4. Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Reversalof obstructive sleep apnoea by continuous positive airway pressure applied through the nares. Lancet 1981; 1:862-5. 5. Schneider H, Becker H, Boke M, et al. Reexposition of NCPAP therapy on sleepapnea patients. Eur Respir J 1989; 2(Suppl. 5:402S). 6. Olsen LG, Rolfe IE, Saunders NA. Tolerance of nasal CPAP treatment of obstructive sleep apnea, and its effects on disease severity. Am Rev Respir Dis 1990; 141:866A. 7. Pykett IL, Newhouse JH, Buonanno FS, et al. Principles of nuclear magnetic resonance imaging. Radiology 1982; 143:157-68. 8. LoweAA, 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. 9. Roehm EO. Computer tomographic measurement of tongue volume relative to its surrounding space. M.Sc. Thesis, University of Manitoba, 1981. 10. 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. 11. Horner RL, Mohiaddin RH, Lowell DG, et al. Sites and sizes of fat deposits around the pharynx in obese patients with obstructive sleep apnoea and weight matched controls. Eur Respir J 1989; 2:613-22. 12. Bydder GM, Young IR. MR imaging: clinical use of the inversion recovery sequence. J Comput Assist Tomogr 1985; 9:659-75. 13. Bergin C, Muller N, Nichols DM, et al. The diagnosisof emphysema. A computer tomographicpathologic correlation. Am Rev Respir Dis 1986; 133:541-6. 14. Popper RA, Leidlinger MJ, Williams AJ. Endoscopic observations of the pharyngeal airwayduring treatment of obstructive sleep apnea with nasal continuous positive airway pressure-a pneumatic splint. West J Med 1986; 144:83-5. 15. Brown I, McClean P, Boucher R, Zamel N, Hoffstein V.Changes in pharyngeal cross-sectional area with posture and application of continuous positive airway pressure in patients with obstructive sleep apnea. Am Rev Respir Dis 1987; 136: 628-32. 16. Kuna ST, Bedi DG, Ryckman C. Effect of nasal airway positive pressure on upper airway size and configuration. Am Rev Respir Dis 1988; 138: 969-75. 17. Strohl KP, Redline S. Nasal CPAP therapy, upper airway muscle activation, and obstructive sleep apnea. Am Rev Respir Dis 1986; 134:555-8. 18. Weil rv, Cherniack NS, Dempsey lA, et al. NHLBI workshop summary. Respiratory disorders of sleep. Pathophysiology, clinicalimplications and therapeutic approaches. Am Rev Respir Dis 1987; 136:755-61. 19. Coopers K, Phillips B. Effect of short-term sleep loss on breathing. J Appl Physiol 1982; 53: 855-8. 20. Berthon-Jones M, Sullivan CEoTime course of change in ventilatory response to CO 2 with long term CPAP therapy for obstructive sleep apnea. Am Rev Respir Dis 1987; 135:144-7.
