INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Since it was first described by Sullivan and colleagues (1), nasal continuous positive airway pressure (nCPAP) has developed into the standard form of treatment of obstructive sleep apnea syndrome (OSAS). CPAP treatment involves the application throughout the entire treatment period of a constant pressure titrated manually on the basis of the data provided by polysomnography (PSG). The aim of this titration is to determine the lowest pressure capable of reliably preventing all obstructive events during the course of the night. As further developments of this form of treatment, automatic devices (APAP) were designed that make it possible to continuously adjust the pressure to the level actually needed to keep the upper airways patent. With such devices, two objectives are possible: either to determine the pressure level for fixed CPAP home treatment (autotitrating systems), or to replace fixed CPAP by automatically varying pressure during the entire treatment period (autoadjusting systems). In the latter way, it is hoped to reduce pressure-associated side effects of CPAP, thus avoiding treatment failures because of noncompliance. Earlier automated nCPAP devices measured airway obstruction on the basis of a reduction in airflow, flattening of the inspiration flow contour, vibrations of the pharyngeal wall, or the flow regime provided by the CPAP compressor (2).

In 1956, DuBois and colleagues (12) described the forced oscillation technique for the measurement of complex resistances in the airways (impedance). In contrast to flow, flattening, and esophageal pressure, measurement of impedance is independent of patient activity and ventilatory effort, and directly reflects the resistance offered by the airways. Using this method, upper airway obstructions during sleep, that is, obstructive apneas and hypopneas, and the upper airway resistance syndrome could be detected noninvasively (13). In such situations, impedance correlated with the esophageal pressure. During polysomnography, it was possible to distinguish central apneas with open airways from obstructive events (13- 15). In addition, Navajas and colleagues (16) showed that manual nCPAP titration based on impedance measurement permitted an adequate determination of the pressure for constant-pressure treatment. The constant CPAP determined in this way did not differ significantly from that determined by conventional titration. The automatic device based on the forced oscillation technique (APAP_FOT) enables the treatment pressure to be adjusted to the actual requirement (autoadjusting device) (19). In short-term observations in the sleep laboratory, we were able to show that with APAP_FOT, effective treatment of OSAS was possible with significantly reduced mean treatment pressures as compared with constant CPAP (19).

To date, few data are available on the continuous use of automated treatment systems (8, 22), in particular in comparison with the standard form of treatment (9, 10, 23, 24). As far as we are aware, no results have yet been published on the application of APAP_FOT under these conditions. In a prospective, randomized study comparing APAP_FOT with standard treatment with constant CPAP, therefore, we investigated the following points:

1. Effectiveness of APAP_FOT treatment in terms of polysomnography data over 6 wk.

2. Profile of treatment pressures in continuous use.

3. Patient acceptance and side effects of APAP_FOT.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Fifty-two patients (45 men, 7 women, 54.7 ± 10.1 yr of age; BMI, 32.4 ± 5.8 kg/m²) (all data given as means ± SD) referred to a university sleep laboratory by pneumologists and general practioners (7/99 to 10/99) were recruited to the study. They underwent polysomnography for the first time in order to investigate hypersomnia, and OSAS was diagnosed (AHI  10/h). All the patients underwent a basic examination of lung function (FEV₁, 104.9 ± 26% predicted; Rtot, 0.25 ± 0.1 kPa/ L/s, Body Master Lab, Jaeger, Höchberg, Germany), and all gave their written consent to participate. The study was approved by the Ethics Commission of the University of Witten/Herdecke. Forty-seven patients completed the study. Of five patients who terminated their participation prematurely, one patient failed to turn up for either of the follow-up examinations. Two patients receiving automated treatment and two receiving constant-pressure CPAP left the study. They were all recommended to continue treatment with standard CPAP. One patient was excluded after a bronchial carcinoma had been diagnosed.

Impedance and the APAP_FOT Device

The forced oscillation technique (FOT) is a noninvasive method for determining complex resistances in the upper and lower airways (12). With the aid of a pump, an oscillating flow is applied to the airways. During the measurement with oscillating flows, not only "real" resistances but also "imaginary" resistances are involved (inertance: inertia of the air column moved, capacitance: extensibility of airways and chest). The term impedance thus includes these complex resistances and describes the mechanical properties of the upper airways, that is, the degree of obstruction in the sleep apnea syndrome (Figure 1).

View larger version (23K): [in this window] [in a new window]   Figure 1.   Schema of the autoadjusting system APAPFOT and description of the complex airway resistance (impedance). The patients breathe via a mask and tube. An oscillating flow of small volume (< 1ml/stroke, 20 Hz) generated by a pump is superimposed on the respiratory flow (Q0). A pressure signal comprising the treatment pressure produced by the CPAP generator and the oscillation pressure is picked up at the mask. Because the tube and device represent a constant resistance (Z0 reference resistance), variations in the measured pressure signal result from changes in airway resistance, and thus reflect the degree of obstruction. When an alternating flow (the oscillating flow) is applied to the airways, not only real resistances, as in the case of uniform constant flows, but also imaginary resistances are also involved. These include inertance (I: inertia of the moved column of air) and capacitance (C: compressibility of air and elasticity of the thoracopulmonary system). This complex of real and imaginary resistances is known as impedance.

As described elsewhere, (19) the system investigated here (Somnosmart; Weinmann, Hamburg, Germany) is an automatic CPAP device that regulates the treatment pressure in accordance with the impedance of the upper airways. The patient breathes through a tube to which an oscillating flow is applied. This oscillating flow is split between the airways and the device unit. A pressure sensor at the mask picks up a signal that comprises the current treatment pressure (CPAP) and the pressure generated by the oscillating measuring flow (oscillation pressure). Together the tube and device form an unchanging reference resistance, and the relative changes in oscillation pressure thus reflect the impedance of the upper airways, and serve to regulate the CPAP generator via a central processing unit. The frequency of the applied oscillations is 20 Hz. Measurement of flow via a pneumotachograph, and thus computation of absolute impedance, is not provided (19).

During the first 5 min of each application, the device determines a reference value during quiet breathing, which serves as baseline and forms the basis for pressure regulation. An increase in pressure is triggered when the reference value is exceeded by a factor of 1.6 (obstructive event) (19). The pressure increase increment is 0.2 cm H₂O/s. At the end of an obstructive event, the treatment pressure is decreased again at a maximum rate of 0.1 cm H₂O/s (19). Under APAP_FOT the treatment pressure can be varied between 4 and 18 cm H₂O. The algorithm defines a central respiratory disturbance when the oscillatory pressure signal is equal to, or not more than 10% lower than, the reference. During these events the pressure is kept constant as long as 2 min.

The patients breathed via a mass-produced standard nose mask (Bubble Mask; ResMed, North Ryde, Australia, SomnoMask; Weinmann, Hamburg, Germany). Humidifiers were not used because we intended to compare the amount of side effects with the different modes. However, the APAP_FOT device functions properly with a humidifier in-line or other masks. Nasal prongs could not be used because the pressure sensor can only be applied in a mask. Constant CPAP treatment was performed without a pressure ramp.

Study Design

After establishing daytime sleepiness with the Epworth Sleepiness Scale (ESS) (25), and performing diagnostic polysomnography, manual titration of the treatment pressure was carried out. During this procedure, the treatment pressure was increased in incremental steps of 1 cm H₂O/h until respiratory disturbances were minimized and respiration-related arousals reduced to  5/h. The pressure thus titrated (P man) was used as the treatment pressure under constant-pressure CPAP. On the following two nights, the effect of treatment with constant CPAP (CPAP-first night) and APAP_FOT (APAP_FOT-first night) was studied in the sleep laboratory with PSG. Thereafter, the patients were treated in randomized order for 6 wk each with both constant CPAP and APAP_FOT. Twenty-four patients began the study with APAP_FOT, 28 with constant CPAP. At the end of each 6-wk treatment period, PSG was carried out in the sleep laboratory (CPAP-6-wk and APAP_FOT-6 wk), and the patients were questioned about possible side effects and their personal assessment of the treatment. At the end of the study, the patients were asked which form of treatment they would prefer for home use. The compliance data (days/time used) and the pressure data were read out of the integrated processor. The detailed testing in the laboratory was performed to confirm adequate treatment before starting the home therapy (CPAP-first night and APAP_FOT-first night). After the 6-wk periods of regular use we wanted to establish the benefits based on both polysomnographic and subjective criteria (CPAP-6 wk and APAP_FOT-6 wk).

The physicians and technicians doing the manual CPAP titration were not involved in either the evaluation of the results of PSG or the questioning of the patients. Those who did the questioning and cared for the patients in the hospital and sleep laboratory and those who evaluated the results of PSG were blinded to the mode of treatment employed. Only the person who managed the randomization knew about the order of the treatment modes in the individual patients. This person adjusted the devices to the different treatment modes.

The patients were not informed about the sequence in which the two treatment modes were applied. In order to avoid product-specific differences such as design, they used only one device for both treatment periods, which permits ventilation in both modes. The patients were informed that they would be treated in two different modes, one with constant pressure, one with variable pressure, but they were not told which mode was the actual standard of treatment and which was the new mode. The patients were asked to contact one of us (W.R.) in case of problems with the treatment during the 6-wk periods. For this purpose a telephone emergency hotline (24-h) was at their disposal.

Polysomnography

The following parameters were recorded: electroencephalogram C4A1 or C3A2, submental or pretibial electromyogram, electrooculogram, effort (thoracic and abdominal induction plethysmography, Consumable Sensor Respiratory Band; Compumedics, Abbotsford, Australia), respiratory flow (thermo-elements, Thermistor Typ 5K3A1; Delta Regeltechnik, München, Germany), snoring signals (laryngeal microphone, ProTec, Woodinville, WA), oxygen saturation (finger pulse oximeter, Oxisensor II D25; Nellcor Puriton Bennett Inc., Pleasanton, CA). Data were recorded electronically (Compumedics). The polysomnography was scored manually. The analysis of the sleep stages was carried out in accordance with the guidelines of Rechtschaffen and Kales (26) and the ASDA criteria (27). Arousals were defined as respiration-induced when they occurred at the earliest with the onset, at the latest 2 s after the end, of an apnea or hypopnea. An apnea was defined as the cessation of respiratory flow for at least 10 s. A hypopnea was defined as a reduction in effort of 50% in comparison with baseline for at least 10 s together with a decrease in oxygen saturation of at least 4%. To quantify snoring, the number of epochs (30 s/page) with evidence of microphone signals for at least 2 s outside of movement artefacts were counted.

Statistics

For the computation of significant differences in dependent parameters (between baseline measurement and each treatment mode, and between the two modes), the ANOVA test with Bonferroni correction (p < 0.05) was used. The McNemar test was used to detect significant differences in device preference. To establish a relationship between treatment pressure and anthropometric and polysomnographic data, Pearson's correlation and simple regression were determined. To determine distribution as normal the Kolmogorov-Smirnov test was used (28).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Both constant CPAP and APAP_FOT improved respiratory disturbances highly significantly at both measuring time points. AHI was reduced from 35.1 ± 26/h (baseline) to 5.3 ± 5.6 (APAP_FOT-first night), 4.6 ± 4.8 (CPAP-first night), 5.0 ± 5.2 (APAP_FOT-6 wk) and 4.3 ± 6.3 (CPAP-6 wk) (p < 0.001 between baseline and each treatment mode). Snoring was also decreased highly significantly in comparison with baseline: 49 ± 36/h (baseline), 10 ± 17 (APAP_FOT-first night), 6 ± 11 (CPAP-first night), 13 ± 20 (APAP_FOT-6 wk), and 6 ± 13 (CPAP-6 wk) (p < 0.001 between baseline and each treatment mode) (see Table E1 in online data supplement at www.atsjournals.org).

Also the total number of arousals and the number of respiration-related arousals were highly significantly reduced. Arousals decreased from 34.0 ± 21.7/h (baseline) to 10.3 ± 6.4 (APAP_FOT-first night), 9.9 ± 5.5 (CPAP-first night), 10.9 ± 5.7 (APAP_FOT-6 wk), and 12.6 ± 8.3 (CPAP-6 wk) (p < 0.001 between baseline and each treatment mode). With both modes, a significant improvement in REM sleep was observed. Both modes showed an improvement in daytime sleepiness, which reached the level of significance after 6 wk of treatment with APAP_FOT (Table E1). No significant differences were found between the two treatment modes in terms of polysomnographic data.

Under APAP_FOT, significantly lower mean pressures were seen at both time points (first night: 6.6 ± 2.4 cm H₂O, p < 0.05; and 6 wk: 5.7 ± 1.8 cm H₂O, p < 0.001) in comparison with manually titrated treatment pressure (7.8 ± 1.5) (see Table E2 in online data supplement). The 95th percentile of the treatment pressure under APAP_FOT over 6 wk (9.0 ± 3.9) did not differ significantly from the constant CPAP. Although the device could apply treatment pressures between 4 and 18 cm H₂O, the maximum pressure during the 6-wk period with APAP_FOT was 12.1 ± 4.1 cm H₂O. For 81.5 ± 21% of the 6-wk treatment period, the treatment pressure under APAP_FOT was lower than the manually titrated pressure (Figure 2, Table E2). A significant correlation was found between BMI and Pman in CPAP, and Pmean in APAP_FOT (Figure 3). In all subgroups of patients (grouped by BMI, age, sex, baseline AHI, ESS, and manually titrated treatment pressure), a reduction in Pmean vis-a-vis Pman was observed.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Pressure profile in relation to treatment time. The figure shows the portions of the treatment time of the first night (left columns) and the 6-wk period (right columns) in which the treatment pressure with APAPFOT was lower than (closed column), equal to (hatched column) or higher than (open column) the manually titrated pressure. Under APAPFOT, the pressure is lower than constant CPAP in more than 80% of the 6-wk period.

View larger version (18K): [in this window] [in a new window]   Figure 3.   Correlation of body mass index and treatment pressures. A significant correlation was found between BMI and mean treatment pressure both under CPAP and APAPFOT. For all values of BMI, a significant difference was seen between CPAP pressure and Pmean under APAPFOT.

The patients used the devices 98.4% of the days, under both APAP_FOT and constant CPAP. The duration of use was 315.4 ± 94.7 min/d (APAP_FOT) and 315.4 ± 97.4 min/d (constant CPAP) (NS). Of the 52 patients initially recruited, 47 (90.4%) completed the study. Of these 47 patients, 35 (75%) gave preference to APAP_FOT for long-term treatment at home, and 12 opted for constant CPAP (p < 0.01). Those showing a preference for APAP_FOT had used this form of treatment for 334.9 ± 88.4 min/d, compared with 320.0 ± 108.8 min/d for CPAP (NS). Those preferring constant CPAP had used this mode for 323.5 ± 76.7 min/d and APAP_FOT for 283.8 ± 92.5 min/d (p < 0.01). Side effects were mild under both modes, and no significant differences were observable. As expected, under APAP_FOT, the patients were more aware of pressure variations. These variations in pressure under APAP_FOT were also significantly more strongly felt to be unpleasant (scale 0 to 6: APAP_FOT, 2.7 ± 2.3; CPAP, 1.3 ± 1.9; p < 0.01). This was particularly the case in those patients who on conclusion of the study gave preference to CPAP (APAP_FOT, 4.4 ± 2.2; CPAP, 2.4 ± 2.2; p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As far as we are aware, our study presents for the first time data on patient acceptance and self-assessment of continuous home treatment using an automatic CPAP device in comparison with constant CPAP. Also for the first time it investigates the question wether the autoadjusting CPAP treatment on the basis of impedance measurement (APAP_FOT) permits adequate treatment of OSAS under this condition. APAP_FOT proved to be equally as effective as the standard treatment with constant CPAP in terms of the respiration parameters and sleep profile. In agreement with the polysomnographic findings, including the number of arousals, we were also able to demonstrate the effectiveness of the therapy on the basis of subjective patient criteria. Thus, daytime sleepiness as measured on the Epworth Sleepiness Scale significantly improved after 6 wk of treatment.

Normalization of respiratory disturbances and sleep architecture under APAP_FOT was achieved with a significant reduction in mean treatment pressure measured over the whole observation period of more than 2 cm H₂O, in comparison with constant CPAP (Figure 4). The pressure profile and the size of the pressure reduction under APAP_FOT were investigated for possible dependence on anthropometric data and the data of baseline polysomnography. For each of the treatment modes, a significant correlation was found between body mass index (BMI) and treatment pressure (Figure 3). The extent of the reduction in treatment pressure under APAP_FOT was equal in all subgroups of patients with differences in age, BMI, sex, AHI, sleep profile, or compliance. The reduction in pressure under APAP_FOT confirmed our earlier results obtained in short-term treatment (19). The large majority of the patients investigated gave preference to autoadjusting CPAP treatment for long-term home use. It may be assumed that this highly significant difference (p < 0.01) in acceptance is based on the reduction in the treatment pressure under APAP_FOT.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Distribution of the differences in treatment pressure. The differences between constant CPAP and the mean treatment pressure with APAPFOT were normally distributed in the population. Positive values indicate a reduction of the pressure under APAPFOT as compared with constant CPAP.

Studies by Condos and colleagues (29) and Monserrat and coworkers (30) showed that a reduction in constant CPAP pressure of 2 mbar was associated with partial collapse of the upper airways (29, 30). Our present data are not incompatible with this. Although a highly significant reduction in the mean pressure over a period of 6 wk was to be seen under APAP_FOT, the 95th percentile of the pressure range applied did not differ significantly from manually titrated constant CPAP. This implies that, in phases of high upper airways resistance, the pressure applied was identical to, or higher than, that titrated manually. This means that there was no constant reduction in treatment pressure; rather, the obvious reduction in mean treatment pressure was due to pressure-lowering during phases of low pressure requirement. In our view, the data on sleep microstructure and macrostructure show that APAP_FOT reduced Pmean while still ensuring appropriate treatment of OSAS. The method of manual titration included oxygen desaturation in contrast to the impedance-controlled treatment. However, we do not believe that one of the algorithms is more rigorous in titration than the other because the amounts of oxygen desaturation and respiratory disturbances did not differ significantly between the two treatment modes.

To date, only few data are available on continuous treatment with autoadjusting CPAP systems, in particular in direct comparison with constant CPAP. In an open study involving patients pretreated with constant CPAP, Boudewyns and colleagues (22) discovered a significant improvement in the polysomnographic data with an autoadjusting CPAP system, as had been seen earlier with constant CPAP. Although that study did not collect data on compliance, four of the 15 patients investigated complained of an increase in excessive daytime sleepiness under autoadjusting treatment. Teschler and colleagues (23) reported a reduction in median pressure over a period of 8 wk under the automatic device based on flattening, but no significant differences in compliance data. Meurice and colleagues (10) found a significantly higher daily use over a 3-wk period as compared with constant CPAP. In a cohort study lasting 3 to 6 mo, Konermann and colleagues (24) demonstrated a significant improvement in the numbers of days of use of the automatic device with identical duration of application/d. In our randomized crossover study, no significant differences were to be found in the mean compliance data of the overall group. We believe that this is due to the fact that in our study, both systems were utilized on more than 98% of the days of the observation period, and for more than 5.3 h per day. We observed a greater compliance (significant in constant CPAP) on the part of patients vis-a-vis the respective system to which they finally gave preference. Therefore, if patients have the opportunity to test both modes their preference may be an indicator of the compliance. Subjective data on sleep quality, self-assessment, and patients' preference were not reported by Teschler (23) or Konermann (24). However, as we did not directly compare APAP_FOT with other autoadjusting devices our data do not prove the superiority of one method.

According to Badia and colleagues (31) accurate measurement of impedance and detection of leakage mandate the simultaneous determination of airflow and pressure variations. Furthermore, in the event of leakages at the mouth, the elevation of the soft palate might close the posterior nasal airway. This closure results in an increase in impedance, which would prompt a pressure increase of the automatic device. The APAP_FOT device does not prevent such an occurrence since it has no integrated facility for flow measurement. Although it is conceivable that these disturbances might trigger a pressure increase, a significant decrease in treatment pressure while providing effective treatment of OSAS has nevertheless been shown for APAP_FOT over 6 wk of therapy too.

On the basis of the results of this study, therefore, we conclude that (1) autoadjusting treatment based on measurement of impedance enables effective treatment of OSAS over 6 wk at home. As is the case with conventional treatment with constant CPAP, a correlation between treatment pressure and BMI is found for APAP_FOT too. (2) Under APAP_FOT, a significant reduction in treatment pressure is achieved. (3) The majority of patients give preference to the automated system for long-term home treatment.