The Effect of Nightly Nasal CPAP Treatment on Underlying Obstructive Sleep Apnea and Pharyngeal Size* Nancy A. Collop, M.D.;t A. jay Block, M.D., F.C.C.P.;:t: and Don Hellard, A.S.

Nasal continuous positive airway pressure (CPAP) is an effective treatment for obstructive sleep apnea (OSA). It is usually prescribed for nightly use; however, recent studies show that patients often do not wear the appliance consistently. Previous studies have also suggested that nasal CPAP may improve a patient's underlying OSA even when the maslc is not in place. We investigated 12 men with OSA to see if nasal CPAP used nightly for six weelcs would improve their underlying sleep-disordered breathing. We also studied pharyngeal volumes measured using magnetic resonance imaging and a computer-controlled digitizing pad. Patients with more severe OSA had improvement after six

weelcs; however, they still demonstrated signi6cant OSA. Patients with less severe OSA did not have signi6cant change. We were unable to show a significant clifTerence in any patient's awake pharyngeal volumes. We cooclude that patients with OSA should be encouraged to wear their nasal CPAP machines regularly. (Chull991; 99:855-60)

Nasal continuous positive airway pressure (CPAP) introduced in 1981 by Sullivan et al' has revolutionized the treatment of obstructive sleep apnea (OSA). It has been shown conclusively to reduce or abolish the number of apneic and hypopneic events and improve the concomitant oxygen desaturation, 2 decrease snoring,3 and improve daytime hypersomnolence.• For most patients, nasal CPAP is well tolerated; however, recent studies have shown that longterm compliance is not 100 percent.s.7 Adverse side effects such as mask discomfort, drying of nasal passages, and noise contributed to noncompliance and often patients wore their masks intermittently. It has also been suggested that patients have improvement in their underlying OSA after regularly using nasal CPAP for a period of time even when the mask is not in place.2 •11•9 Possible reasons for this are reduction of pharyngeal mucosal edema, improvement in sleep fragmentation, concomitant weight loss, or increased upper airway muscle tone. If it could be conclusively shown that patients with OSA improve their underlying sleep-disordered breathing after using nasal CPAP for a specified period of time, it is possible that patients would not have to use the appliance nightly and may only need to use it intermittently. We set out to look at newly diagnosed

OSA patients who began use of nasal CPAP on enrollment in the study and were followed up over six weeks. Sleep studies were done without nasal CPAP at the initiation of therapy and at the end of six weeks. Magnetic resonance imaging of the upper airway, awake and without nasal CPAP, was also performed at the onset of the study, at two weeks, and at the end of six weeks to see if changes in pharyngeal volume occurred.

•From the Gainesville VA Medical Center. University of Florida, Gainesville, Florida. Supported by a grant from the American Lung Association of Florida. tFellow, Division of Pulmonary Medicine. *Professor of Medic.ine. Manuscript received July 3; revision accepted September 24. Reprint request8: Dr: Bkx:lc , j. llillu Miller Health Center. Univemty of Florida, Cainemlle 32610

SPT=slee~riod time; TST::a.total sleep time; AID•~ hypopoea . • MDI=mean deutunllioo iDcla; SSI•sleep stUdy I; SS2=sleep study 2; HCJ"Z=hydroddorochiazj; HTN =hypertension; MBI s: mean body indes; BDI'"" respiratory disturbance index

M ETHODS AND MATERIALS Subject~

A total of 12 patients were enrolled in this study, all male. All of the patients had symptoms of daytime hypersomnolence and loud snoring. Patients were enrolled after they had undergone a full night of polysomnography at our VA sleep laboratory. The presence of OSA was defined by an apnea-hypopnea index (apneas plus hypopneaslbours of total sleep time) of more than 6ve events per hour with accompanying oxygen desaturation to less than 85 percent. In addition, all patients demonstrated marked improvement with nasal CPAP and were able to tolerate wearing it. Patients enrolled in this study could not weigh more than 146.25 kg because that amount is the weight limit of our magnetic resonance scanner. The patients had a variety of other medical problems and were receiving a variety of medications (Thble 1). Three patients currently smoked cigarettes and six additional patients had smoked cigarettes in the past. All patients volunteered for this study and infonned consent was obtained. This experimental protocol was approved by the Institutional Review Boards of the ]. Hillis Miller Health Center and the Gainesville VA Medical Center. Sleep Thchnlques

Patients undetwent two sleep studies for this study. All studies were done at the Gainesville VA Medical Center (GVAMC) sleep laboratory. The first sleep study used for analysis was actually the screening study done to diagnose OSA. The second polysomnognaphy was done at the end of the study, which was approximately six CHEST I 99 I 4 I APRIL. 1991

855

Table 1-lblienta' Medic411Norden and Medication. Subjects

Medleal Disorders

Current Medications

1 2

Craves' disease Hyperlipidemis, HTN,• gastric re8ux

3

Diabetes. coronary artery disease None HTN, • hyperlipidemia

Synthroid Terazosin, Ranitidine, Triamterene/HCTZ, t Diclofe.n ac Insulin, Furosemide, Digoxin None Hcrz. t Hydralazine, Cuanfacine, Pindolol None Albuterol MOlt. Doxepin Beclamethasone Nasal Spray, Clipizide Prednisone, methotrexate, Naproxen, Cimetidine Prednisone, Digoxin, Clipizide, Colchicine, Chlorthalidone lsordil

4

5

6 7

8 9

10 11 12

Cirrbosis Asthma, diabetes, osteoarthritis. allergic rhinitis Hyperlipidemia, rheumatoid arthritis Psoriatic arthritis, atrial 6brillation, diabetes SIP acoustic neuroma removal , angina None Chronic lymphocytic lymphoma, emphysema, allergic rhinitis, osteoarthritis

None Albuterol and lpratropium Bromide MDJ'st, Ibuprofen, Beclomethasone Nasal Spray, Furosemide, Theophylline

•HTN = hypertension. tHCTL= hydrochlorthiazide. *MDI= metered dose inhaler.

weelcs later (this time varied slightly for each patient).

Polysomnogn~phy was done for the entire night using standard Recbtscha.tren/Kales montage.•• The polygraph (model 790; Crass Instruments, Braintree, MA) simultaneously recorded frontal electroencephalograms (EEC), eJectro..oculogrs (EOC), submental electromyograms (EMC), electrocardiogram (lead 2), and microphone tracings. The microphone used to detect snores was a TR-21 miniature microphone (Crass) that generates a large signal (1 to 5 mV) In response to noisy airflow. This was taped directly above the sternal notch and detected all snores and recorded them on the polygraph. Chest and abdominal movements were impedance sensed with surface e lectrodes. airflow was sensed by a combined nasal-i>ral thermistor (Rocheste~Eiectromed, Tampa, FL), and oxygen saturation was measured with an ear oximeter (3700 pulse oximeter, Obmeda, Louisville, CO). These were simultaneously recorded on a physiograph (Model DMI4B, Naroo, Houston, TX). All patients were videotaped using a camera (RCA TC2511 US) with infrared lighting. Split screen imaging allowed us to view patient and sleep recordings simultaneously. For all sleep studies, patients were put to bed at approximately 10 Pl\4 . In the initial sleep study, patients were monitored for at least 3 hours (180 mioutes) without tbe nasal CPAP mask on with 2 exceptions. One patient bad such severe oxygen desaturation (oximeter reading 14 percent at nadir) that he was only monitored without nasal CPAP for 149 minutes. The second patient also bad severe oxygen desaturation (oximeter reading 63 percent at nadir) along with arrhythmias and was monitored for 130 minutes. The remainder of the night was spent with nasal CPAP in place. For the seoond sleep study, all patients were monitored for more than 220 minutes without the nasal CPAP mask in place and then spent the remainder of the night with the mask on.

NtuOJCPAP For the first sleep study, after initial monitoring, when it was clear that the patient bad OSA, a nasal CPAP mask was applied. Our machine was a commercial brand (SieepEasy Ill, provided b y Respironics, Inc, Monroeville, PA). The pressure was adjusted in each patient individually and was ultimately set at the pressure that abolished most apneas and hypopneas, kept oxygen saturation at more than 90 percent, and eliminated snoring. The day following their initial sleep study, each patient was given detailed instructions on how to use a nasal CPAP machine and watched an informational video about the use of nasal CPAP (distributed by Respironics). In addition, eight of the 12 subjects slept an additional night (supervised but without polysomnography) at the CVAMC to assure that they could tolerate nasal CPAP. Patient.s were given a nasal CPAP machine to take home with them. (The machine provided was a SleepEasy Ill supplied by Respironics, Inc.) Each machine had an internal valve that was adjusted to their prescribed pressure and an hour meter that recorded hours of use. Patients were instructed to use their machine nightly and were given a diary in which to record when the machine was turned on each night and when the machine was turned off each morning. The patients did not know that the machine had an internal clock. At the end of approximately six weeks, the patient returned with his nasal CPAP machine and his diary for the second sleep study. As stated previously, this sleep study was done in the same manner as the initial sleep study with the exception that in some patients the monitoring time prior to the placement of nasal CPAP was longer.

Magnetic Resonance Imaging The morning immediately following their initial sleep study, each subject underwent magnetic resonance imaging (MRJ) of the upper airway. All studies we re done on a 1.5-Testa superconductive magnet (Siemenc, Sweden) with specialized external receive-only coils. Scans were obtained in the awake, supine position using the threedimensional Fourier transform spin echo technique. Slices were obtained in a "slab," making the m truly contiguous, with a 4-mm thickness. Pulse repetition time was 15 ms, echo delay time was 0.50 s. Images were T, weighted and each acquisition took 3.8 minutes. Sagittal views were obtained. The patients were instructed to relax but not to sleep or swallow during the scan. Two weeks after the initial sleep study, the patients returned for a second MRI of the upper airway and a third scan was done the morning after their second sleep study, six weeks later. All MRI scans were performed awake , without nasal CPAP in place. DefinitioN

Standard procedures were used to quantitate sleep events. Desaturation was defined as a reduction in oxygen saturation by 4 percent or more. Apnea was a cessation of airftow at the nose and mouth for greater than 10 s. Hypopneas were scored as a decrease in inspiratory Bow coupled with desaturation. Sleep period time (SPT) was de6ned as the time from the onset of sleep to the last awakening in the morning. Total sleep time (1'S1) was sleep period time less any time the subject was awake after falling asleep. An apnea-hypopnea index (AHI) was defined as all the apneas plus all the hypopneas divided by TST.

Dota Analyril For each sleep study, respiratory events {apneas, hypopneas) were analyzed for type and duration of the event as well as the duration and amount of desaturation. An AHI and a mean desaturation index (M OJ) were calculated for each sleep study. The M 01 was calculated by summing maximal desaturations for all events that occurred without nasal CPAP in place and dividing by that number of events. Nightly Nasal CPAP T1'881rnent o1 OSA (Co/lop, BJock, Hellatd)

Sleep stages were scored by the method of Rechtschalfen and Kales.'" The time and percentage of sleep stages 1 +2, 3+4 and rapid eye movement (REM) were recorded for time with and without nasal CPAP for all sleep studies. Computation of pharyngeal volumes was done as described previously." A computer program using a digitizing pad {Numonics 2210, Montgomeryville, PA) was used. In summary, the pharyngeal airspace was outlined with a romputer mouse using the beginning of the soft palate as the superior landmark and the back of the epiglottis as the inferior landmark. An area (in sqiUI(e centimeters) is calculated for each slice and then multiplied by 0.4 em (the thickness of each slice) to approltimate a volume. The sequence is repeated on each slice that involves the airway (n = 7 to 10) and then added together to give a composite volume. In addition, we divided the airway into three separate sections and calculated volumes for each section. The uppermost section was designated as the transpalatal volume and extended from the beginning of the soft palate (the end of the nasal choanae) to the end of the soft palate (the tip of the uvula). The lowermost section was designated as the hypopharyngeal volume and it extended from the tip of the epiglottis to the base of the epiglottis. The mid-section was designated as the oropharyngeal volume and included the section between the transpalatal and hypopharyngeal volumes. We have validated this technique using rontrol objects." Since a different scanner was used than in the previous article, we again scanned rontrol objects with the same technique as the patients. Known volumes of mineral oil and cheese were scanned and digitized. The mineral oil was scanned while sitting in "boats" of clay molded in various shapes to emulate the contour of the upper airway. The cheese (mozzarella) was also cut into various shapes and water volume displacement was used to determine its actual volume. The digitized volume was compared with the actual volume. The