0
Clinical Investigations: SLEEP |

Validation of the POLY-MESAM Seven-Channel Ambulatory Recording Unit* FREE TO VIEW

Thomas Verse, MD; Wolfgang Pirsig, MD; Bernhard Junge-Hülsing, MD; Béatrice Kroker, MD
Author and Funding Information

*From the University of Ulm, Otorhinolaryngological Clinic Section for Rhinology und Rhonchopathies, Ulm, Germany.

Correspondence to: Thomas Verse, MD, Universitäts-Hals-Nasen-Ohren-Klinik Ulm, Prittwitzstrasse 43, 89075 Ulm, Germany;



Chest. 2000;117(6):1613-1618. doi:10.1378/chest.117.6.1613
Text Size: A A A
Published online

Study objectives: To determine the recording capabilities of the POLY-MESAM (PM) unit (MAP; Martinsried, Germany), an American Sleep Disorders Association level III system, and to compare it with simultaneous 12-channel polysomnography in the sleep laboratory.

Measurements and results: Fifty-three patients (49 men and 4 women) with obstructive sleep-related breathing disorders of varying severity were included. The apnea-hypopnea indexes (AHIs) obtained using the two methods differed significantly from each other, although the correlation was close. The PM unit produced false-negative results in patients with mild to moderate obstructive sleep apnea (OSA). The sensitivity of the PM unit in detecting patients with an AHI > 10 was 92%, while the specificity was 96.3%.

Conclusions: The correlation of AHIs obtained with polysomnography and with the PM unit is close. However, in some cases, the PM may underestimate OSA parameters. The PM unit produces false-negative results in patients with mild to moderate OSA. While inpatient polysomnography remains the “gold standard,” the PM unit may provide an inexpensive alternative in some special cases.

Figures in this Article

The“ gold standard” for the diagnosis of respiratory abnormalities during sleep is a full night of sleep for the patient while being observed and recorded in the laboratory with polysomnography. This is classified by the American Sleep Disorders Association (ASDA) as a level I sleep study.1 This procedure is both labor and resource intensive, and, so, alternatives are desirable. A variety of devices that can be used at home are available, and these have been categorized by the ASDA into the following three levels.

  1. Level II studies. These are comprehensive studies using portable polysomnography devices that measure both respiratory and sleep variables.

  2. Level III studies. These are studies using unattended recording devices (modified portable sleep apnea testing) that allow for the assessment of cardiorespiratory variables only.

  3. Level IV studies. These studies comprise continuous single or double bioparameter recordings.

One level III device that measures thoracic and abdominal respiratory movements and combined oronasal flow is the POLY-MESAM (PM) unit (MAP; Martinsried, Germany).2 This device is an upgrade from its predecessor the MESAM4. The purpose of this study is to determine the recording capabilities of the PM unit compared to those of simultaneously performed 12-channel attended polysomnography in the sleep laboratory.

Patients who attended our clinic for sleep-related respiratory disorders and snoring were randomly selected. Those with chronic obstructive lung disease and cardiac insufficiency (New York Heart Association score > 2) were excluded. Fifty-three subjects were recruited and admitted to the hospital, were examined further, and underwent simultaneous polygraphy and polysomnography. For polygraphy, the PM unit was used, and for polysomnography the SIDAS GS system (Stimotron; Wendelstein, Germany) was used. There were 49 men and 4 women in the study with the following characteristics: mean (± SD) age, 48.1 ± 10.8 years (men, 48.4 ± 10.7 years; women, 44.3 ± 10.9 years); and average body mass index, 27.4 ± 4.9 kg/m2 (men, 27.0 ± 4.5 kg/m2; women, 32.0 ± 6.3 kg/m2).

The PM unit has been described earlier.2 It is a small monitor that can be attached to the patient and, in the standard version, has the following seven channels: (1) a flow sensor for joint registration of nasal and oral breath flow; (2) a laryngeal microphone; (3) a three-channel ECG; one stress-sensitive belt each for the (4) thorax and (5) abdomen; (6) a positional sensor for determination of body position; and (7) a pulse oximeter. A computer (IBM; New York, NY) is needed for the evaluation, and the software used in our department (POLY-MESAM, version 1.42; MAP; Martinsried, Germany) automatically calculates the apnea-hypopnea index (AHI), the apnea index (AI), the hypopnea index (HI), the oxygen desaturation index (ODI), the heart rate variation index, and the movement index. Besides presentation of the data in the form of diverse graphs, evaluation of the raw data is also possible.

Polysomnography was performed on an inpatient basis (level I sleep study) using the SIDAS GS system, which in this study recorded 12 channels. There were similar cardiorespiratory channels to the PM unit but also two EEG channels, two electro-oculography channels, and two electromyogram channels (sensors to the anterior tibialis and genioglossus muscles).

Patients were admitted to the clinic at 9:00 pm, and the sensors of both units were attached by technicians in the sleep laboratory. Measurements were recorded between 10:30 pm and 5:30 am. Periods during which patients were awake during the recording time were purposely not documented. The next morning, patients were disconnected from the machines and were discharged.

The evaluation of the recordings was performed in a double-blinded manner by examiners who were not familiar with the patients or with the results of the other recording method. For the SIDAS GS system, only manual findings returned by an experienced examiner have been used because the software for automatic evaluation of sleep stages delivers partially incorrect data. Therefore, computer-generated findings were rejected. For the PM unit both automated and hand-scored data analysis were provided and discussed. For the PM unit, the indexes for AHI, AI, HI, and ODI are based on the total measurement period from 10:30 pm to 5:30 am, while for complete polysomnography recordings, these parameters were recorded for the entire period of sleep. Thus, periods of possible wakefulness were deducted from the total measurement time of polysomnography.

The analysis of sleep stages was performed according to the guidelines of Rechtschaffen and Kales3 in 30-s intervals. Arousals were counted in the polysomnograms. In both systems, a reduction in oxygen was recorded for oxygen desaturations of at least 4%, independent of the simultaneous occurrence of an apnea or hypopnea episode. The minimum duration of an apnea or hypopnea episode was set at 10 s. Apneas were diagnosed in cases in which respiratory flow was reduced to < 20% of the normal level, while hypopneas were designated as reductions in flow of 50 to 80%, independent of the presence of oxygen desaturations.

Authorities are divided on the diagnosis of sleep apnea. Some require an AHI of at least 10,46 others require 20.4,7 Other authorities use an AI of at least 10.7 The findings from polygraphy and polysomnography were compared, and sensitivity and specificity were calculated for AHI values of 10, 15, and 20 and AI values of 5 and 10.

In order to test the hypothesis that measurement of the AHI is comparable using both methods, the paired t test was used. Representation of the AHI findings was performed using the method of Bland and Altman.8

In general, values recorded by the PM unit were lower than those returned by standard polysomnography. This can be seen in Table 1 , which presents data regarding the ODI, AHI, AI, HI, and arousal index (ie, the number of arousals per hour of sleep) in the complete patient group. Raw data for individual patients are presented in Table 2 .

The evaluation of data returned by polysomnography yielded AHI values between 0.2 and 75.9 (mean [± SD], 17.9 ± 18.1). In comparison, findings from polygraphy resulted in AHI values between 0 and 73 (mean, 17.1 ± 15.5). Figure 1 compares the findings of level I studies (the SIDAS GS system) and level III studies (the PM unit). The coefficient of correlation is 0.965. Figure 2 displays the respective AIs (coefficient of correlation, 0.971). The coefficients of correlation for the HI and the ODI were 0.761 and 0.709, respectively.

For the AHI parameter, the paired t test at 52 degrees of freedom yielded a value of 3.52. The value t0.95 for the 95% confidence interval stands at 1.68. The original base hypothesis that identical results would be obtained using both level I (polysomnography) and level III devices (the PM unit) is, thus, disproved. Figure 3 shows the difference between the AHI values for both types of measurements in comparison with their averages. It is clear that the PM unit underestimates the polysomnography findings.

Findings for the AI parameter were comparable. The paired t test yields a t value of 4.4 (t0.95 = 1.68). Therefore, the AI values also did not tally in a statistically significant manner.

During polysomnography, 21 patients had an AHI of ≥ 20, 23 patients had an AHI of > 15, and 25 patients had an AHI of at least 10. Twenty patients had an AI of at least 10, while the remaining 33 patients had an AI < 10. Twenty-three patients had an AI of at least 5.

In contrast, findings returned by the PM system correctly identified 15 of 21 patients with an AHI of at least 20, 20 of 23 patients with an AHI of > 15, and 23 of 25 patients with an AHI of at least 10 as suffering from sleep apnea. The values for sensitivity are shown in Table 3 . Similarly, 17 of 20 patients with an AI of at least 10 and 21 of 23 patients with an AI of at least 5 were correctly identified (Table 3).

In the present study, 26 of 27 patients with AHIs < 10 were correctly identified as not suffering from sleep apnea. This corresponds to a specificity of 96.3%. Similarly, 29 of 30 patients with AHIs < 15 and 30 of 31 patients with AHIs < 20 were identified correctly. This corresponds to specificities of 96.7% and 96.8%, respectively. Concentrating on the AI, all 30 patients with AIs < 5 and all 33 patients with AIs < 10 were correctly classified using the PM unit (Table 3).

This study and previous studies comparing level III devices with the “gold standard” of polysomnography suffer from the inherent problem that for study purposes both types of measurement systems are used while being attended in the laboratory. Of course, by definition, level III devices are not designed to be used in this way. Laboratory use may give a false degree of reliability that is not found when used in the home.

In the present study, there was a high degree of correlation between the findings of the two measurement systems for both AHI (r = 0.965) and AI (r = 0.971) (Fig 1, 2). Due to larger variation, the correlation between findings returned for HI and ODI is less pronounced despite comparable median values (Table 1). Nevertheless, it was not possible to demonstrate statistically that the polysomnography test method and the level III study using the PM unit yielded identical test results.

Considering the total patient group (Table 1), one sees that values for AHI and AI measured by polygraphy (PM unit) are lower than those returned by polysomnography. The higher the value for AHI or AI using polysomnography, the greater the degree to which the level III study underestimates the result (Fig 3). This is probably due to differences in measurement scope. Whereas the PM unit bases its index scores on the total measurement time (or total time in bed [TIB]), the corresponding indexes scores are calculated by polysomnography on the basis of total sleeping time (TST).5 It is conceivable, for example, that a patient may actually sleep for only 50% of the ambulatory recording time. In this instance, the indexes measured with polygraphy would be only half as high as they would be in reality. In the present study, the average TIB (PM unit) was 400 min, and the average TST (polysomnography) was 321 min. Therefore, one may assume that the PM unit may significantly underestimate the severity of disease in individual cases.

If the TIB is used as the basis for the polysomnographic findings instead of the TST, the average deviation of AHI is 0.45 ± 3.79. The t test yields a value for t of 0.864. It can, therefore, be shown that AHI values do not statistically differ from each other (p > 0.95). In other words the differences between AHI measured with polysomnography vs the PM unit disappear if the AHI measured with polysomnography is calculated on the basis of total recording time rather than TST.

The sensitivity of a method is based on the number of patients with a given disease that are correctly identified as suffering from that disease. In the present study, the PM system failed to identify 2 of 25 patients with an AHI of at least 10 and 5 of 21 patients with an AHI of at least 20 as suffering from sleep apnea (false-negative). This corresponds to sensitivities of 92% and 71.4%, respectively. Similarly, of 23 patients with an AI of at least 5, 2 patients were not correctly identified, resulting in a sensitivity of 91%. Three of 20 patients with an AI of at least 10 were not identified, corresponding to a sensitivity of 85%. It seems clear that the higher that the AHI and AI levels have been set, the more difficulty the PM unit has in correctly classifying patients. In other words, the PM unit underestimates obstructive sleep apnea (OSA) with increasing severity.

Conversely, the specificity of a method corresponds to the proportion of study subjects not affected by a given disease that is correctly classified as not having that disease. In the present study, 26 of 27 patients with an AHI < 10 and all 30 patients with an AI < 5 were correctly identified by the PM unit as not suffering from sleep apnea. Hence, the specificity of the method is 96.3 to 100% (Table 2).

As can be seen from the issues discussed above, depending on the definition used, a certain number of patients affected by sleep apnea may escape detection in a level III study using the PM unit and, therefore, will not be referred for further diagnostic procedures in the sleep laboratory.

Our results show that up to 20% of OSA patients were not recognized by the PM unit. Peer-reviewed evidence in the medical literature confirms that cardiorespiratory sleep studies (level III studies) can produce false-negative results in patients with mild to moderate apnea and should be reserved for patients for whom the probability of having moderate to severe OSA is high.9Severe cases of OSA are usually apparent at clinical examination.10 A patient who continues to have excessive daytime sleepiness but a negative cardiorespiratory sleep study usually requires further testing with a full polysomnography.

The indications for a level III study with the PM unit can be derived from the above discussion and from the following recommendations of the ASDA for the use of portable polysomnography units and related examination methods.1,9

Standard attended polysomnography (level I study) is the accepted test for the diagnosis and determination of the severity and treatment of OSA.1 The PM unit might be an acceptable alternative in the following instances: (1) for patients with severe clinical symptoms that are indicative of a diagnosis of OSA, and when the initiation of treatment is urgent and standard polysomnography is not readily available; and (2) for patients unable to be studied in the sleep laboratory.

A comparison of the PM and earlier MESAM4 units regarding their respective sensitivity and specificity is of interest. In the literature, sensitivity is indicated to be between 78% and 100%, and specificity between 63% and 100%.1,57 The number of patients included ranged from 37 to 68. In each instance, only the best results (manual interpretation57) are given. It is of note that both the PM and MESAM4 units yield comparable results. The superiority of manual interpretation is often cited.57 It is possible that, in other studies, the examiners performing the manual scoring devoted more time and effort, paying particular attention to pulse rate.2 The advantage of using the more costly and complex PM unit is the possibility of discriminating different kinds of breathing events due to the additional use of the thermistor for combined oronasal flow and thorax and abdomen belts. This distinction, however, is uncertain unless esophageal balloons are used. In any case, however, the precision of the PM unit can be significantly increased by the addition of an objective measurement device for oronasal respiratory flow, such as the pneumotachograph.5,11 On the basis of its above-described application, the PM unit appears to be an improvement over its predecessor, the MESAM4 unit.

An earlier study of 104 patients showed that the PM unit was well-suited for ambulatory use.2 Faulty recordings were returned in 5.8% of patients. Two sensors (one thermistor for oronasal respiration and one pulse oximeter probe) broke and had to be replaced. The average total time spent by physicians was 33 min per patient (patients were attached to the recording devices in the sleep laboratory), and 22 min when the patients attached themselves to recording units. Scoring times averaged 9.3 min for the whole group. Hence, a level III study using the PM unit is significantly less expensive than a fully attended level I study in the sleep laboratory. Though, according to the reported sensitivities and specificities, repeat testing with fully attended polysomnography after testing with the PM unit might negate the initial decrease expense.

The correlation of AHIs using polysomnography and the PM unit is close. However, in some cases, the use of the PM unit results in an underestimation of the AHI. The PM unit produces false-negative results in patients with mild to moderate OSA. Although inpatient polysomnography remains the “gold standard” because of that reason, the PM unit may provide an inexpensive alternative for the indications recommended by the ASDA.1

Abbreviations: AHI = apnea-hypopnea index; AI = apnea index; ASDA = American Sleep Disorders Association; HI = hypopnea index; ODI = oxygen desaturation index; OSA = obstructive sleep apnea; PM = POLY-MESAM; TIB = time in bed; TST = total sleeping time

The POLY-MESAM unit was loaned to the authors by the MAP Company (16 Fraunhoferstrasse, 82152 Martinsried, Germany). No person at MAP has read the manuscript or has been involved in any of the research, which is the sole responsibility of the authors.

Table Graphic Jump Location
Table 1. Group Results of Polysomnography and Ambulatory Recordings With the PM Unit*
* 

Values given as mean ± SD.

Table Graphic Jump Location
Table 2. Data for Each Patient Comparing Polysomnography vs Ambulatory Recording With the PM Unit*
* 

PG = polygraphy; PSG = polysomnography; M = male; F = female.

Figure Jump LinkFigure 1. Comparison of AHI value pairs determined by polysomnography and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location
Figure Jump LinkFigure 2. Comparison of AI value pairs determined by polysomnography and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location
Figure Jump LinkFigure 3. Difference in AHI vs average AHI determined by polysomnography (PSG) and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location
Table Graphic Jump Location
Table 3. Specificity and Sensitivity of the PM Unit
Ferber, R, Millman, R, Coppola, M, et al (1994) ASDA standards of practice: practice parameters for the use of portable recording in the assessment of obstructive sleep apnea.Sleep17,378-392. [PubMed]
 
Verse, Th, Junge-Hülsing, B, Kroker, BA, et al First results of a prospective study validating the method of ambulatory polysomnography using the POLY-MESAM unit.Sleep Breathing1997;2,56-64. [PubMed]
 
Rechtschaffen A, Kales A. A manual of standardized terminology techniques and scoring system for sleep stages in human subjects. Bethesda, MD: National Institutes of Health, 1968; Publication No. 204.
 
Cohrs, S, Herrendorf, G, Kinkelbur, J, et al Validierung eines ambulanten Screeninggerätes für schlafbezogene Atmungsstörungen: APNOESCREEN I.Somnologie1998;2,8-13. [CrossRef]
 
Koziej, M, Cieslicki, JK, Gorzelak, K, et al Hand-scoring of MESAM4 recordings is more accurate than automatic analysis in screening for obstructive sleep apnoea.Eur Respir J1994;7,1771-1775. [CrossRef] [PubMed]
 
Stoohs, R, Guilleminault, C MESAM4: an ambulatory device for the detection of patients at risk for obstructive sleep apnea syndrome (OSAS).Chest1992;101,1221-1227. [CrossRef] [PubMed]
 
Roos, M, Althaus, W, Riehl, C, et al Vergleichender Einsatz von MESAM4 und Polysomnographie bei schlafbezogenen Atmungsstörungen (SBAS).Pneumologie1993;47,112-118. [PubMed]
 
Bland, JM, Altman, DG Statistical methods for assessing agreement between two methods of clinical measurement.Lancet1986;1,307-310. [PubMed]
 
American Sleep Disorders Association.. Practice parameters for the indications for polysomnography and related procedures.Sleep1997;20,406-422. [PubMed]
 
Guilleminault, C Dement, WC eds. Sleep apnea syndrome. 1978; Alan R Liss. New York, NY:.
 
Teschler, H, Hoheisel, G, Schumann, H, et al Validierung des Sleep-Dic-Porti-Systems für die ambulante schlafapnoediagnostik.Pneumologie1995;49,496-501. [PubMed]
 

Figures

Figure Jump LinkFigure 1. Comparison of AHI value pairs determined by polysomnography and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location
Figure Jump LinkFigure 2. Comparison of AI value pairs determined by polysomnography and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location
Figure Jump LinkFigure 3. Difference in AHI vs average AHI determined by polysomnography (PSG) and a cardiorespiratory sleep study with the PM unit.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1. Group Results of Polysomnography and Ambulatory Recordings With the PM Unit*
* 

Values given as mean ± SD.

Table Graphic Jump Location
Table 2. Data for Each Patient Comparing Polysomnography vs Ambulatory Recording With the PM Unit*
* 

PG = polygraphy; PSG = polysomnography; M = male; F = female.

Table Graphic Jump Location
Table 3. Specificity and Sensitivity of the PM Unit

References

Ferber, R, Millman, R, Coppola, M, et al (1994) ASDA standards of practice: practice parameters for the use of portable recording in the assessment of obstructive sleep apnea.Sleep17,378-392. [PubMed]
 
Verse, Th, Junge-Hülsing, B, Kroker, BA, et al First results of a prospective study validating the method of ambulatory polysomnography using the POLY-MESAM unit.Sleep Breathing1997;2,56-64. [PubMed]
 
Rechtschaffen A, Kales A. A manual of standardized terminology techniques and scoring system for sleep stages in human subjects. Bethesda, MD: National Institutes of Health, 1968; Publication No. 204.
 
Cohrs, S, Herrendorf, G, Kinkelbur, J, et al Validierung eines ambulanten Screeninggerätes für schlafbezogene Atmungsstörungen: APNOESCREEN I.Somnologie1998;2,8-13. [CrossRef]
 
Koziej, M, Cieslicki, JK, Gorzelak, K, et al Hand-scoring of MESAM4 recordings is more accurate than automatic analysis in screening for obstructive sleep apnoea.Eur Respir J1994;7,1771-1775. [CrossRef] [PubMed]
 
Stoohs, R, Guilleminault, C MESAM4: an ambulatory device for the detection of patients at risk for obstructive sleep apnea syndrome (OSAS).Chest1992;101,1221-1227. [CrossRef] [PubMed]
 
Roos, M, Althaus, W, Riehl, C, et al Vergleichender Einsatz von MESAM4 und Polysomnographie bei schlafbezogenen Atmungsstörungen (SBAS).Pneumologie1993;47,112-118. [PubMed]
 
Bland, JM, Altman, DG Statistical methods for assessing agreement between two methods of clinical measurement.Lancet1986;1,307-310. [PubMed]
 
American Sleep Disorders Association.. Practice parameters for the indications for polysomnography and related procedures.Sleep1997;20,406-422. [PubMed]
 
Guilleminault, C Dement, WC eds. Sleep apnea syndrome. 1978; Alan R Liss. New York, NY:.
 
Teschler, H, Hoheisel, G, Schumann, H, et al Validierung des Sleep-Dic-Porti-Systems für die ambulante schlafapnoediagnostik.Pneumologie1995;49,496-501. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
Description and Validation of the Apnea Risk Evaluation System*: A Novel Method To Diagnose Sleep Apnea-Hypopnea in the Home
PubMed Articles
Guidelines
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543