A Sleep Laboratory Evaluation of an Automatic Positive Airway Pressure System for Treatment of Obstructive Sleep Apnea

Transcription

1 A Sleep Laboratory Evaluation of an Automatic Positive Airway Pressure System for Treatment of Obstructive Sleep Apnea Khosrow Behbehani, 1 Fu-Chung Yen, 1 Edgar A. Lucas, 2 and John R. Burk 2 (1) Joint Biomedical Engineering, The University of Texas at Arlington and The University of Texas Southwestern Medical School at Dallas; (2) Sleep Consultants, Fort Worth, Tex Study Objectives: This paper compares the performance of an experimental nasal positive airway pressure device that automatically adjusts the level of applied pressure (APAP) with the performance of a conventional continuous positive airway pressure (CPAP) in a sleep laboratory study. Design: In a randomized sequence, conventional CPAP therapy was applied for 1 night (CPAP night) and APAP therapy the following night (APAP night). Setting: The study was conducted in an accredited sleep disorders center. Patients or Participants: Twenty-six men and 5 women between the ages of 35 to 73 (51±9.6) years with body mass index 35.82±8.35 (kg/m 2 ) who were diagnosed (using standard nocturnal polysomnography [NPSG] methods) as having OSA syndrome were studied. The subjects were treated with conventional CPAP for approximately 8 (7.79±3.16) weeks at home prior to their participation in this study. Measurements and Results: All standard polysomnography data and nasal mask pressures were recorded using a computer-based data acquisition system. Sleep and respiratory data were scored by a registered polysomnographer. The mean apnea-hypopnea index (AHI) for subjects for the NPSG night was 55.2±33.7. It dropped to 4.2±3.8 for the CPAP night and to 5.4±5.4 for the APAP night. There was no significant (p=0.05) difference between mean AHI indices, sleep stages, sleep stage shifts, and snore arousals for CPAP night and APAP night. However, all the measures showed significant (p=0.05) improvement over NPSG night. The mean of APAP applied pressure (8.4±3.3 cm H 2 O) was significantly (p=0.05) lower than the prescribed pressure (11.5±3.1 cm H 2 O), but there was no significant (p=0.05) difference between the maximum APAP applied pressure (12.8±4.3 cm H 2 O) and the prescribed pressure (11.5±3.1 cm H 2 O). All mean comparison tests were carried out using two-tailed statistics. Conclusions: APAP appears to be as effective as CPAP in treating OSA patients. APAP delivers the same level of therapy as CPAP, but it reduces the average airway pressure while providing needed peak pressures. Key words: CPAP; APAP; auto CPAP; obstructive sleep apnea; smart CPAP Accepted for publication April, 1998 Address correspondence and requests for reprints to Khosrow Behbehani PhD, PE, Biomedical Engineering Program, P.O. Box 19138, Arlington, TX MORE THAN 90% of sleep disorders patients tested at sleep disorders centers in the United States suffer from obstructive sleep apnea (OSA) syndrome. 1 Sleep apnea syndrome refers to a set of symptoms resulting from respiratory impairment during sleep. A study estimated that as many as 4% of men and 2% of middle-aged women (30-60 years) of the United States work force suffer from sleep apnea syndrome. 2 In OSA, the respiratory effort (ie, diaphragm movement) continues despite a partial or complete collapse of the upper airway that results in reduction or complete loss of airflow to the lungs. The most commonly used treatment for OSA is nasal continuous positive airway pressure (CPAP). A CPAP device includes a blower and a flexible hose connected to a 485

2 nasal mask secured over the patient s nose during sleep, 3-6 which applies a constant pressure to the patient s airway. This pressurized air column acts as pneumatic splint to prevent airway collapse. The prescribed CPAP pressure level is usually determined by testing the patient in a sleep laboratory. Specifically, in our sleep center, pressure is manually titrated to eliminate all airway obstructions and snoring for all sleep stages and body positions, including supine REM sleep, where pressure needs are often the greatest. Although CPAP therapy is widely used, it has known limitations. Since a CPAP unit applies the prescribed pressure uniformly throughout the night, patients often complain about the increased expiratory effort or sensation of forced air through their nostrils, especially prior to sleep onset. 7 One attempt to reduce pressure-related complaints has been use of a pressure-ramp feature that slowly increases the CPAP-applied pressure from a low level to the prescribed pressure over a fixed interval (eg, 30 minutes). This gradual increase in pressure allows sleep onset to take place at a pressure lower than the prescribed pressure. Another attempt to alleviate the pressure-related discomfort has been the advent of a bilevel pressure CPAP unit called BiPAP. A BiPAP unit allows independent adjustment of inspiratory and expiratory pressures. Sanders et al 7 reported that BiPAP can effectively treat OSA with lower levels of expiratory pressure than traditional CPAP. In testing 13 OSA patients, they showed that they could treat the patients with the same mean inspiratory pressure as their mean prescribed CPAP pressure, but with significantly (p<0.0001) lower mean expiratory pressure. They indicated that lower BiPAP pressures might reduce the adverse effects associated with nasal CPAP therapy and improve long-term therapeutic compliance. However, CPAP units with both the pressure-ramp feature and BiPAP maintain constant pressures through the entire night. It is known that, in some patients, the amount of pressure required to prevent airway collapse varies significantly throughout the night and from night to night. 7,9 Sleep position, sleep stage, and airway congestion are among the most significant parameters that affect the level of pressure required to keep the airway patent. CPAP devices have been developed which automatically adjust the applied pressure in real time to establish a minimum level that is adequate for preventing airway collapse. Such devices are referred to as Auto-CPAP (or APAP for brevity). 7,10 An APAP device appears to address both the pressurerelated complaints and the inadequacy of CPAP in consistently applying the necessary pressure. However, the therapeutic efficacy of APAP has not been objectively established. This paper compares pressure and sleep-architecture data for an experimental APAP and a conventional CPAP device using a random crossover, 2-consecutivenight study design in 31 OSA syndrome patients who had been treated with conventional CPAP therapy for a limited time (»8 weeks). METHODS Subjects Thirty-one volunteer subjects, 26 men and 5 women, between the ages of 35 and 73, were recruited from patients diagnosed with OSA by standard nocturnal polysomnography (NPSG). 14 Institutional Review Board (IRB) approvals were obtained prior to the conduct of this study. Because subject selection was on a voluntary basis, the recruitment was not consecutive. Some of the invited patients chose not to participate for a variety of reasons, such as lack of time or disdain for further involvement with the sleep-study process. Aside from having OSA, no other OSA-related criteria such as weight, prescribed CPAP pressure level or severity of OSA were used in recruiting subjects. However, if a patient also had a severe medical condition, such as class 2 through class 4 heart failure, that patient was not recruited. Table 1 shows a summary of the biometrics of the subjects who participated in the study. Each subject had been titrated for a prescription pressure therapy in an accredited sleep center (All Saints Sleep Disorders Center, Fort Worth, Tex), as will be described below. Following NPSG study, patients were maintained on conventional CPAP therapy at their prescribed CPAP pressure for an average of approximately 8 weeks (range: 3 to 18 weeks) prior to the present study. Most of the subjects received routine follow-up visits to ensure adequate compliance with the therapy. All subjects reported regular nightly use of CPAP at home. Protocol The prescribed pressure for each subject was determined by starting the patient with 4-6 cm H 2 O of CPAP pressure while polysomnographic (PSG) recordings of electroencephalogram (EEG) (C3-A2, C4-A1), electrooculogram (EOG) (LOG, ROG), electrocardiogram (ECG), chin electromyogram (EMG), and EMG of both anterior tibialis muscles were obtained. Respiratory activity was monitored by thermistors at the nose and mouth and by chest and abdominal piezoelectric elastic bands (EPM systems). Blood oxygen saturation was measured using a pulse oximeter (Ohmeda Biox III). A surface microphone (model , Tandy Corporation, Fort Worth, Tex) 11 was placed over the lateral thyroid cartilage in the cervical triangle to record tracheal sounds. All leads were connected to a 21-channel polygraph 12 and recorded at a paper speed of 10 mm/second. By monitoring the patient and the polysomnography data, the CPAP pressure was raised by 2 cm H 2 O whenever the patient experienced full obstruction. 486

3 In the case of partial obstructions, the pressure was raised by 1 cm H 2 O. After each pressure increase, the pressure was left unchanged for at least 2 minutes before another pressure increase was made. Pressure increases ceased when all partial and complete obstructions were eliminated for all sleep stages and body positions. The final pressure was selected as the prescribed pressure for treating patient with a CPAP machine. Following the use of conventional CPAP at home, each subject participated in a 2-consecutive-night study in our laboratory. With the sequence of test nights chosen randomly, conventional CPAP therapy was applied during one night (CPAP night) and APAP therapy the following night (APAP night). The patient test protocol for this study aimed at minimizing the changes in patient condition by conducting the tests on 2 consecutive nights. The subjects were not aware which therapy was to be applied on any given night. The experimental APAP unit used in this study senses pharyngeal wall vibrations (PWVs) associated with snoring and assumed to be precursors to airway obstruction. The algorithm increases airway pressure to eliminate PWV and preclude airway collapse 12 ; in most cases it eliminated snoring. It was built by interfacing a computer with a CPAP machine (BiPAP, Respironics Corp., Murrysville, Pa). The computer adjusted the level of applied pressure using nasal mask pressure as feedback (ICSensor Model 120A, ICSensor, Milpitas, Calif) and a PWV-detection algorithm. The detection algorithm examined the frequency and amplitude of the pressure oscillations in the nasal mask. 13 The computer started the APAP therapy at 4 cm H 2 O and initiated the detection of PWVs. If three PWV episodes were detected and no two consecutive episodes were separated by more than 20 seconds, then the applied APAP pressure was automatically raised by 1 cm H 2 O. After the pressure increase, if additional PWV episodes occurred, the pressure was raised again by 1 cm H 2 O. This continued until either no PWV was detected or a maximum pressure of 20 cm H 2 O was reached. Hence, the maximum applied pressure could not exceed 20 cm H 2 O. Conversely, if no PWV episodes were detected for a 5-minute interval, the applied APAP pressure was reduced at a rate of 1 cm H 2 O per each 5-minute interval until the preselected minimum pressure (4 cm H 2 O) was reached. On each study night, polysomnography (PSG) data, as described above, was recorded. PSG data was recorded both on paper with conventional ink pen and in a dedicated computer. However, to independently monitor the APAP automatic pressure adjustments, a small pressure transducer (Microswitch, 163PC01D48, Freeport, Ill) was attached proximal to the nasal mask to measure the delivered gas pressure. For off-line analysis, the output of the transducer was digitized and stored in a second computer, independent of the computer that stored PSG data. The gas delivery Figure 1. Mean AHI for the 3 study nights. The error bars represent one standard deviation from the mean. Figure 2. Comparison of the mean prescribed pressure (CPAP) with the mean of the time-weighted average (APAP_Mean) and maximum (APAP_MAX) of APAP-applied pressure. The error bars represent one standard deviation from the mean. tubes and the transducer wires were put in a harness to minimize patient disturbance during the night. The subjects generally stayed in bed for approximately 8 hours, and PSG was terminated after their final awakening in the morning. Analysis PSG data were scored either by one of the authors (EAL) or another registered polysomnographer, blind to the mode of therapy and in accordance with the criteria of Rechtschaffen and Kales, 14 except that 30-second epochs were used. Sleep stages were assigned to these epochs prior to assessing the tracing for respiratory events. Apnea was scored when airflow decreased by 75% for at least 10 seconds. Hypopnea was scored either when airflow dimin- 487

4 Figure 3. Mean amount of sleep stages for NPSG, CPAP, and APAP nights. The error bars represent one standard deviation from the mean. Table 1. Subject biometric data Male (26 Subjects) ished by 25% and oximetry readings fell more than 2%, or when snore arousal occurred. Snore arousals were scored if the amplitude of the microphone output increased leading to indications of behavioral or EEG arousal concomitant with a cessation of snoring in the absence of apnea or hypopnea criteria. The total of stages 1, 2, combined stages 3 and 4, rapid eye movement (REM), and awake time were computed. The apnea-hypopnea index (AHI) was obtained by dividing the total number of apnea and hypopnea episodes by total sleep time. Sleep stage shifts reflecting all changes occurring in sleep-awake staging were also computed. Since the APAP unit varied the applied pressure throughout the night, the independently recorded mask pressure was analyzed to obtain the minimum, maximum, and time-weighted mean pressures. PSG data, as well as the pressure data, were compared for the CPAP and APAP nights using paired two-tailed t tests. In all comparisons, the level of significance was set to p=0.05. Female (5 Subjects) RESULTS Total (31 Subjects) Mean Standard Range Mean Standard Range Mean Standard Range Deviation Deviation Deviation Weight (lb.) Height (in) Body Mass (Kg/m 2 ) Age (year) As predicted, the APAP system automatically adjusted airway pressure to prevent airway occlusion. The APAPapplied pressure was usually less than the prescribed pressure for most of the night and occasionally greater than the prescribed pressure level. Similar patterns were obtained for other subjects participating in the study. For the 31 subjects tested, there were on the average 44±22 incidences of pressure increase and 37±18 of pressure decrease. The mean of the total number of pressure changes was 84±39. Based on these averages we analyzed the EEG traces across each pressure change for four of the subjects (randomly selected), using criteria for scoring EEG arousals 15 and excluding the leg movement induced arousal. 16 Only one arousal was noted out of a total of 476 APAP pressure adjustments that were analyzed. Figure 1 shows the mean AHI for NPSG, CPAP, and APAP nights. Table 2 summarizes the results of statistical comparison of the mean values of respiratory and sleep- 488

5 Table 2. Statistical comparison of the means of respiratory and sleep-quality metrics Parameters APAP vs CPAP APAP vs NPSG CPAP vs NPSG AHI NS S S Snore arousal NS S S Stage 1 duration NS S S Stage 2 duration NS S S Stage duration NS S S REM NS S S Number of sleep stage shifts NS S S Time in bed NS NS NS Mean APAP press. vs CPAP prescribed pressure S - - Max. APAP press. vscpap prescribed pressure NS - - NS: There is no statistically significant difference between the means tested (using two-tailed statistical test; p=0.05) S: There is a statistically significant difference between the means tested (using two-tailed statistical test; p=0.05) architecture measures. Comparison of means of AHI for the three therapy modes showed that AHI was significantly lower for CPAP and APAP nights compared with NPSG night (Fig.1), but the AHI for APAP and CPAP nights did not differ significantly. APAP mean pressures were significantly lower than the mean prescribed (CPAP) pressure (Fig. 2 and Table 2). However, the mean of the maximum pressures applied by APAP (APAP_MAX) did not differ significantly from the prescribed CPAP pressure (Fig. 2). The mean amount of sleep stage 1 for NPSG night was significantly greater than for CPAP or APAP nights. Stage 2, REM, and delta stage each had a greater mean for CPAP as well as APAP nights compared to NPSG night (Fig. 3 and Table 2). There was no significant difference between the mean of these measures for APAP vs CPAP nights (Table 2). Comparison of the means for the total time in bed, total sleep time, and total awake time showed that mean values of these measures were not significantly different for the NPSG, CPAP and APAP nights (Fig. 3 and Table 2). For the NPSG nights, the number of snore arousals was significantly greater than the number of snore arousals for APAP or CPAP night, but the mean number of snore arousals for the CPAP and APAP nights did not differ significantly (Fig. 4 and Table 2). The mean number of sleep stage shifts for CPAP and APAP nights was significantly lower than the number of shifts for NPSG night, but the number of stage shifts for CPAP and APAP nights did not differ significantly (Fig. 5 and Table 2). Figure 4. Mean number of snore arousals for NPSG, CPAP, and APAP nights. The error bars represent one standard deviation from the mean. Figure 5. Mean number of sleep stage shifts for NPSG, CPAP, and APAP nights. The error bars represent one standard deviation from the mean. DISCUSSION Comparison of the AHI for APAP and CPAP therapy shows that the APAP system reduced the number of apnea and hypopnea events for OSA patients as effectively as the CPAP unit; that is, APAP lowered the AHI to the same level as CPAP. Hence, it appears that APAP treats OSA as well as CPAP. Since the APAP unit automatically varied the pressure that it applies to the airway, it was of interest to explore the relationship between the prescribed pressure (applied by 489

6 CPAP) and the APAP-applied pressure values. As indicated earlier, the prescribed pressures for the subjects were determined, on the average, 8 weeks prior to the start of the present study. Therefore, there is a possibility that if a titration study was conducted immediately before the present study, a different prescribed pressure for some of the subjects might have resulted. Barring this possibility, the timeweighted average of APAP-applied pressures had a lower mean than prescribed pressure (26%). This shows that APAP, on the average, applies lower pressure across the night than prescribed pressure. The clinical ramifications of this finding on upper-airway problems associated with long term use of CPAP therapy (such as nasal congestion, insufflation, etc.) are not known at this time. However, APAP provides a means to investigate potential long-term benefits of lower pressure without compromising the patient s therapy. The pressure needed to eliminate PWVs throughout the night was not constant; that is, for most of the night the applied pressure was below, but it also was occasionally above the prescribed pressure. Since the results of this study indicate that APAP efficaciously treats OSA, it follows that APAP applied pressures are appropriate. In other words, the results clearly demonstrate that during a given night it may be necessary to apply pressures that are different from the prescribed pressure to eliminate PWV. This observation suggests that CPAP-applied pressure at times is either greater than or less than the minimum required pressure to prevent airway obstruction. Further, APAP therapy provides a lower mean airway pressure while still providing the needed peak pressures. Additionally, statistical equivalence of the mean prescribed pressure and the mean of the maximum of APAP-applied pressure suggests that the maximum pressures automatically generated by APAP on the average are the same as the mean of prescribed pressures; the maximum APAP applied pressure is, on the average, the same as the manually titrated prescribed pressure (see Protocol section). Based on this observation, the optimum pressure varies across the night and from night to night in a dynamic manner. The finding that APAP pressure adjustments rarely (1 in 476) resulted in EEG arousals indicates that changes in APAP pressure did not disturb sleep. The lack of sleep disturbance by APAP can be further inferred by examining the sleep architecture measures. Specifically, the significant increase in the mean total of stage 2, stages 3 and 4 combined, and REM sleep, as well as the reduction in the number of sleep stage shifts and snore arousals, were similar for both CPAP and APAP nights (in comparison to NPSG nights). Subjects in this study had been treated with CPAP for an average of approximately 8 weeks before the first night of the study. It is possible that some of the physiologic effects of CPAP therapy have an impact on the polysomnography measurements that were made. However, by randomly assigning the order of APAP and CPAP tests for this study, the impact of prior CPAP treatment on the results for both therapy modes (APAP and CPAP) should be equal, as both were preceded by several weeks of CPAP therapy at home. It is important to note that the automatic pressure-adjustment algorithm of APAP is based on detecting PWV. This approach is based on the assumption that in patients with OSA syndrome, PWV precedes apnea, hypopnea, and snore arousal events. Consequently, for patients in whom this assumption is true, APAP aims to prevent such events. Therefore, APAP can have an advantage over a reactive system that would increase the pressure only after an occlusion has occurred. Alternatively, if a patient exhibits OSA without having PWV, APAP would not respond to adjust the pressure. While PWV is a hallmark of OSA and is generally present in OSA patients, patients with CSA or patients who have had soft-palate surgery may not produce sufficient PWV to activate the APAP unit. Specifically, while APAP automatically seeks to find the pressure that eliminates PWV, patients should be examined for evidence of OSA prior to treating them with APAP, and evaluated afterwards for the therapy efficacy. Conversely, it is possible that not all PWV in an OSA patient would lead to apnea or hypopnea. For this reason, the algorithm for APAP pressure control requires that multiple PWV events take place before the pressure is raised. This approach diminishes the likelihood of that APAP might respond to occasional PWV events that do not disturb sleep. It is also noted that APAP does not respond to central sleep apnea. In comparing APAP and CPAP therapy using the measures employed here, APAP therapy exhibits the same level of efficacy as CPAP therapy. However, it appears that APAP offers a significant advantage over CPAP therapy, as it adapts to patient respiratory needs for variable pressure throughout the night. In this study, we focused only on objective measurements of polysomnography to determine the performance of APAP relative to CPAP. It may be of interest for future studies to include some subjective measures, such patients impression of the comfort of APAP therapy vs CPAP. We are currently investigating the pressure variations from night to night in APAP therapy for OSA patients. CONCLUSIONS The results of this study demonstrate that APAP therapy is as effective as conventional CPAP therapy in lowering the AHI and improving sleep architecture. Pressure changes by APAP did not cause sleep disruption. Peak pressures in APAP were not significantly different from prescribed pressure, but the mean applied pressure was 490

7 reduced significantly across the night by APAP. APAP applied pressures were generally lower, but were occasionally greater than the prescribed pressure. ACKNOWLEDGMENTS This research was supported in part by a grant from the National Institutes of Health (1R15 HL A1) and a grant from Texas Higher Education Coordinating Board. REFERENCES 1. Parkes, JD. Sleep and Its Disorders. pages , Publisher: Saunder, London, Philadelphia, Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The Occurrence of Sleep-Disordered Breathing Among Middle-Aged Adults. N Engl J Med 1993;328: Acebo C, Watson RK, Bakos L, Thoman EB. Sleep and Apnea in the Elderly: Reliability and Validity of 24-Hour Recordings in the Home. Sleep 1991;14(1): Martin RJ. Cardiorespiratory Disorders During Sleep. Futura Publishing company, Mount Kisko, NY, Sanders MH, Moore SE, Eveslage J. CPAP Via Nasal Mask: A Treatment For Occlusive Sleep Apnea. Chest 83(1): , Perez-Guerra F, The treatment of obstructive sleep apnea. Tex Med 1987;83: Burk JR, Lucas EA, Axe JR, Behbehani K, Yen F. Auto-CPAP in the Treatment of Obstructive Sleep Apnea: A new Approach Annual Meeting Abstracts, Association of Professional Sleep Societies 6th Annual Meeting, 61, Sanders MH, Kern N. Obstructive Sleep Apnea Treated by Independently Adjusted Inspiratory and Expiratory Positive Airway Pressures via Nasal Mask. Chest, 98(2): , Berthon-Jones M. Feasibility of a Self-Setting CPAP Machine. Sleep 1993;16:S120-S Sadrnoori B. Demand Positive Airway Pressure System and Obstructive Sleep Apnea. Innovative Medical Systems Inc., Hampton, NH, Daniels B, Harris C. Technical Tidings: A Simple Device for Detection of Snoring. APSS Newsletter, 4 (1): 37, March Behbehani K, Yen FC, Burk JR, Lucas EA, Axe JR. Automatic Control of Airway Pressure for Treatment of Obstructive Sleep Apnea. IEEE Trans Biomed Eng 1995;42 (10): Axe JR, Behbehani K, Burk JR, Lucas EA, Yen FC, "Methods and Apparatus for Controlling Sleep Disorder Breathing", U.S. Patent No. 5,548,137, Oct. 17, Rechtschaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Bethesda, MD: US Government Printing Office, 1968 (National Institutes of Health Publication No. 204). 15. A Preliminary Report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association, EEG Arousals: Scoring Rules and Examples. Sleep 1992;15(2). 16. Bonnet M, Carely D, Carskadon M, Easton P, Guilleminault C, Harper R, Hayes B, Hirshkowitz M, Perkilis K, Keenan S, Pressman M, Roehrs T, Smith J, Walsh J, Weber S, Westbrook P, Jordan B. Recording and Scoring Leg Movements. Sleep 1993;16(8). 491