Screening for sleep-related breathing disorders by transthoracic impedance recording integrated into a Holter ECG system

Transcription

1 J. Sleep Res. (2006) 15, Screening for sleep-related breathing disorders by transthoracic impedance recording integrated into a Holter ECG system ANDREAS MUELLER, INGO FIETZE, RICHARD VOELKER, STEPHAN EDDICKS, MARTIN GLOS, GERT BAUMANN and HEINZ THERES Charite University Medical Centre, Medical Division, Berlin, Germany Accepted in revised form 18 August 2006; received 7 June 2006 SUMMARY In patients with arrhythmias, coincidence with sleep-related breathing disorders (SRBD) is high and of clinical relevance. Electrocardiogram-derived (ECG) parameters have been developed for SRBD screening, but it has proved necessary to exclude patients with frequent arrhythmias. Holter-based screening tools, easy to use, are therefore warranted. The goal of our study was to evaluate the diagnostic accuracy, with respect to SRBD detection, of transthoracic impedance recording (TTIR) integrated into a Holter System. Our investigation consisted of 2 phases. In phase 1 we compared the performance of TTIR to that of in-hospital polysomnography (PSG) in 56 patients (46 male, mean age 57). In phase 2 we compared TTIR to results from an ambulatory polygraphy (PG) system in 180 patients (143 male, mean age 56). We scored apnea and hypopnea from P(S)G, and derived a respiratory-disturbance index (P(S)G-RDI). TTIR was analyzed semi-automatically. Reduction of the impedance amplitude by more than 50% over 10 s was scored as apnea/hypopnea, with consequent calculation of TTIR- RDI. In phase 1, 20 out of 56 patients revealed a PSG-RDI > 10 h )1. TTIR-RDI in 19 patients from this group was >10 h )1 (sensitivity 95%, specificity 97.2%, positive predictive value 95%, negative predictive value 97.2%, interclass correlation coefficient 0.98). In phase 2, 46 of 180 patients revealed a PSG-RDI > 10 h )1. TTIR-RDI in 37 out of this group was >10 h )1 (sensitivity 80.4%, specificity 92.5%, positive predictive value 78.7%, negative predictive value 93.2%, interclass correlation coefficient 0.92). TTIR integrated into a Holter ECG system and tested in a large patient cohort demonstrates acceptable high accuracy in detection of SRBD. Arrhythmia analysis and screening for SRBD can be performed in a single-step approch. keywords recording ambulatory screening, holter ecg, sleep apnea, transthoracic impedance INTRODUCTION The prevalence of obstructive sleep apnea syndrome (OSAS) (i.e. an apnea/hypopnea score of 5 or higher and daytime hypersomnolence) among the middle-aged is estimated at 2% in women and 4% in men (Young et al., 1993). Recent studies have revealed the high coincidence of OSAS and cardiovascular diseases. OSAS leads to a rise in blood pressure Correspondence: Dr. med. Andreas Mueller, Charité University Medical Centre, Medical Division, Department for Cardiology and Angiology, Charitéplatz 1, Berlin, Germany. Tel.: +49 (0) ; fax: +49 (0) ; andreas.mueller@ charite.de (Davies et al., 2000), which is reversed by OSAS treatment (Becker et al., 2003; Pepperell et al., 2002). Sleep-related breathing disorders (SRBD) are a frequent concomitant disease in patients with pacemakers and implantable cardioverter defibrillators (ICD) (Fietze et al., 2000; Fries et al., 1999). Cheyne Stokes respiration, furthermore, is an indicator for reduced survival in heart failure. Small-scale studies, with the objective of relieving SRBDs in this patient population, have shown promising results (Bradley and Floras, 2003; Sin et al., 2000). In contrast, a recent study did not reveal a reduction of survival by the use of continuous positive airway pressure (CPAP) in this patient group (Bradley et al., 2005). Ó 2006 European Sleep Research Society 455

2 456 A. Mueller et al. The current ambulatory diagnostic approach includes patient history (e.g. daytime sleepiness), physical examination (e.g. body mass index and blood pressure), and screening with ambulatory monitoring and with the use of multichannel polygraphic recorders. These devices record information on thoracic excursion, nasal flow, arterial oxygen saturation (SaO 2 ), pulse rate, as well as other parameters including body position, snoring, and electrocardiogram (ECG) results. In light of the high coincidence between SRBD and cardiac arrhythmias (Grimm et al., 2000; Kanagala et al., 2003) it appears promising to develop strategies that allow screening for SRBD with application of Holter ECG parameters. Until now, application of these algorithms has been limited, as they are not effective in the presence of arrhythmias such as atrial fibrillation, or of paced ventricular rhythms and autonomic dysfunction (Roche et al., 2002; Stein et al., 2003a). Transthoracic impedance recording (TTIR) is known as a reliable tool for the detection of intrathoracic volume changes (Frerichs et al., 2002; Hahn et al., 1995; Kunst et al., 2000). It is used to detect respiration in various contexts, e.g. in ICU monitoring systems and with pacemakers. Recently published data have disclosed high sensitivity and specificity in transthoracic impedance signals recorded by pacemakers in screening for SRBD (Scharf et al., 2004). The goal of our study was to evaluate the sensitivity and specificity of TTIR, integrated into a Holter ECG system, for screening of SRBD, when compared with standard in-laboratory and ambulatory poly(somno)graphy. METHODS General aspects The study was organized in two consecutive phases. In phase 1, TTIR was compared with results of in-laboratory sleep studies using full polysomnography as gold standard for the diagnosis of SRBD. In phase 2, TTIR was compared with results from an ambulatory screening device for the detection of SRBD. Whereas phase 1 was conducted exclusively in the Charite University Medical Centre, in Berlin, Germany, patients for phase 2 were additionally enrolled at Achenbach Hospital in Ko nigs Wusterhausen, Germany. The study was approved by the Ethics Committee of Charite University Medical Centre, in conformity with the Declaration of Helsinki. Written informed consent was obtained from all patients prior to the study. applied. A total of 243 patients were asked for informed consent, and 199 agreed to participate in the study. In 180 patients, recordings demonstrated sufficient quality for analysis. Patients were excluded from the analysis for the following reasons: incomplete ambulatory polygraphy because of patient discomfort and termination of recording (n ¼ 3), dislodgement of nasal flow sensor (n ¼ 9), or technical issues, e.g. electrode dislodgement (n ¼ 5) and battery depletion (n ¼ 2) related to ECG recording. Study procedure and equipment Holter ECG and transthoracic impedance recording We applied a Holter ECG system, the CardioMemÒ, manufactured by the company getemed, of Teltow, Germany. In addition to monitoring three ECG channels with this system, we applied it to record transthoracic impedance for 24 h via standard ECG chest electrodes. No additional electrodes are necessary, as impedance recording is performed via the same electrodes used for ECG recording. We analyzed ECGs with a standard software package (CardioDay Ò, getemed, Teltow, Germany). Rhythm disturbances were scored semiautomatically and by beat-by-beat inspection. Transthoracic impedance is obtained by continuously passing an alternating current through the chest via two Holter ECG channel 1 electrodes. The waveform used is a square wave (approx. 27 khz) having a peak-to-peak amplitude of 20 la. The resulting voltage alteration is amplified and applied to a bandpass filter from 0.2 to 4 Hz, before being sampled and digitized by a 12-bit A to D converter. The dynamic range of ±3 X covers the full-scale range of the A to D converter. The sampling rate for ECG and impedance is 256 Hz. The resulting impedance values are stored on a compact flash memory card, together with ECG data. Transthoracic impedance was filtered (lowpass with 0.5 Hz) to minimize cardiac and movement artifacts. Typical recordings during normal breathing and during apnea are depicted in Figs 1 and 2. Further investigation of the TTIR signal was conducted by a single scorer, blinded to results of sleep parameter analysis, with Study population In phase 1, patients were recruited during a 6-month enrolment phase. All patients undergoing in-laboratory sleep study were eligible who presented with suspected OSAS, as evidenced by ambulatory sleep apnea monitoring and/or symptoms. A total of 143 subjects were screened, and 56 patients gave informed consent. For phase 2, all patients undergoing routine Holter ECG monitoring were accepted as candidates for SRBD screening. No exclusion criteria were Figure 1. Typical transthoracic impedance recording during normal breathing.

3 Screening for sleep-related breathing disorders 457 inspection of the underlying signals and redefined apneas/ hypopneas as necessary. Finally, we calculated the number of apneas and hypopneas, as well as the RDI. A board-certified sleep-medicine specialist, blinded to the results of TTIR, performed analysis of PG data. Statistical analysis Figure 2. Typical transthoracic impedance recording (TTIR) during central apnea; this figure shows that detection apneas in TTIR is possible. assistance of digital signal-processing software. As reference, we calculated a moving average of the peak-to-peak amplitude over 60 s. This value was continuously updated, as changes in body position can cause considerable alteration in this baseline value. A drop in the impedance amplitude of more than 50% of the reference value over 10 s was scored as apnea/hypopnea, with subsequent calculation of a TTIR respiratory disturbance index (TTIR-RDI). To enable calculation of total sleep time (TST), patients were asked to keep diaries and record information on sleep onset/termination and awakenings. In-laboratory polysomnography In phase 1, we employed EmblaÒ/Somnologica TM (Embla, Reykjavik, Iceland) for polysomnographic recording and analysis. The following parameters were acquired: oral and nasal air flow, thoracic and abdominal excursions, body position, EEG, ECG, EMG, and peripheral oxygen saturation. Analysis was conducted in accordance with criteria set by the American Academy of Sleep Medicine Task Force (1999). The analysis of polysomnographic (PSG) data was performed by a board-certified sleep-medicine specialist, blinded to the results of TTIR. As a final step, we calculated the number of apneas and hypopneas, as well as the PSG-RDI. Ambulatory polygraphy In phase 2, we investigated patients with an ambulatory monitoring system (ApnoeScreen ProÒ, Jaeger, Wu rzburg, Germany), in addition to the use of the Holter ECG. This PG system records the following parameters: nasal and oral air flow (via thermistor), peripheral oxygen saturation and pulse rate, body position, body activity, and snoring (recorded by microphone). Using this system, we scored a reduction in air flow of more than 88% over 10 s as apnea, and a reduction between 88% and 50%, as hypopnea. A significant reduction in oxygen saturation was defined as 4% reduction over 10 s. After automatic analysis, we performed visual For baseline characteristics, we calculated the total number (n), the mean value with standard deviations (SD), and/or percentages (%). After conducting independent analysis of inlaboratory and ambulatory monitoring, in addition to TTIR, we performed further statistical analysis. The correlation between the TTIR-RDI and PSG-RDI was assessed using the Pearson correlation coefficient and Bland Altman plots. A threshold of PSG-RDI or PG-RDI > 10 h )1 was used as the cut-off point for SRBD diagnosis. According to common thresholds in sleep studies, we used a cut-off threshold of >10 h )1 in TTIR to define suspected SRBD. Receiver operating characteristic (ROC) analysis was carried out to evaluate the TTIR diagnostic capability by varying the threshold of impedance events per hour. Based on these threshold definitions, ROC curves were derived, and areas under the curves (AUC) were calculated. Whereas an AUC of 0.5 characterizes an insufficient test, an AUC of close to 1 represents the optimal result. RESULTS Baseline characteristics Baseline characteristics are shown in Table 1. A total of 236 patients (189 men and 47 women) were included in the study; 56 patients were enrolled in phase 1 and 180 patients, in phase 2. The mean age was 55.9 years. Average body mass index was 29. Current smokers constituted 19% of the patient cohort. Of the total cohort, 29.7% had a history of coronary artery disease, and 15.7% had already experienced at least one myocardial infarction. Medication was as follows: 45.3% were being treated with beta-blockers, 44.5% received ACE inhibitors, and 39.4% took aspirin. Comparison of TTIR versus polysomnography All 56 patients were included in the definitive analysis. Polysomnography revealed SRBD with an RDI > 10 h )1 in 20 patients (Table 2a). Mean PSG-RDI was 14.5 h )1 (min 0, max h )1, STD 33 h )1 ). In the patient cohort with PSG- RDI>10h )1, mean PSG-RDI was 37.2 h )1 (min 10.1 h )1, max h )1, STD 33 h )1 ). Of 20 patients with an PSG-RDI > 10 h )1, 19 demonstrated a TTIR-RDI > 10 h )1. We obtained the following results: sensitivity 95%, specificity 97%, positive predictive value 95%, and negative predictive value 97%. Analysis (Fig. 3) revealed excellent correlation (r ¼ 0.98) between PSG-RDI and TTIR- RDI. ROC (AUC 0.989; 95% CI ) and Bland Altman results are shown in Figs 4 and 5.

4 458 A. Mueller et al. Phase 1: TTIR versus polysomomnography (in laboratory) Phase 2: TTIR versus polygraphy (home study) Total Table 1 Baseline characteristics Patients, n Age, mean (SD), years 55.5 (1.4) 56.8 (11.51) 55.9 (12.5) Men, no./total (%) 44/56 (79) 143/180 (79) 189/236 (80.0) Body mass index, kg m )2, 30.2 (5.3) 28.4 (4.8) 28.8 (5.0) mean (SD) Systolic blood pressure, mmhg, (16.5) (10.4) (16.33) mean (SD) Diastolic blood pressure, mmhg, 84.5 (10.2) 78 (12.1) (10.71) mean (SD) Clinical history Smoker, n (%) 12 (21) 32 (18) 44 (18.6) Non-smoker, n (%) 31 (56) 75 (41) 106 (45) Ex-smoker, n (%) 13 (23) 73 (41) 86 (36.4) Coronary artery disease, n (%) 6 (11) 65 (36) 71 (30.1) Myocardial infarction, n(%) 5 (9) 32 (18) 37 (15.7) Hyperlipidemia, n (%) 7 (13) 63 (35) 70 (29.7) Hypertension (%) 28 (50) 122 (67.8) 150 (63.6) Atrial fibrillation, n (%) 5 (9) 21 (12) 26 (11) Diabetes, n (%) 8 (14) 34 (19) 42 (17.8) Chronic obstructive 5 (9) 16 (9) 21 (8.9) lung disease, n (%) Medication b-blockers, n (%) 13 (23) 94 (52) 107 (45.3) ACE inhibitor, n (%) 12 (21) 93 (52) 105 (44.5) Calcium channel blockers, n (%) 6 (11) 16 (9) 22 (9.3) Digitalis, n (%) 1 (2) 6 (3.3) 7 (3.0) Diuretics, n (%) 8 (14) 32 (18) 40 (17.0) ASS, n (%) 12 (21) 81 (45) 93 (39.4) CSE inhibitors, n (%) 11 (20) 71 (39) 82 (34.7) Nitrate, n (%) 7 (12.5) 23 (13) 30 (12.7) Coumadine, n (%) 2 (3.6) 59 (33) 61 (25.8) TTIR, transthoracic impedance recording. PSG-RDI >10h )1, n PSG-RDI < 10 h )1, n Total, n (a) TTIR-RDI > 10 h )1, n TTIR-RDI < 10 h )1, n Total, n PG-RDI > 10 h )1,n PG-RDI < 10 h )1,n Total, n Table 2 (a) TTIR versus polysomnography (PSG) during in-laboratory sleep study; (b) TTIR versus polygraphy (PG) during unattended sleep study (b) TTIR-RDI < 10 h )1, n TTIR-RDI > 10 h )1, n Total, n TTIR, transthoracic impedance recording; RDI, respiratory disturbance index. Comparison of TTIR versus ambulatory monitoring Of 180 patients monitored, 46 revealed a PG-RDI > 10 h )1 (Table 2b). In this cohort, mean PG-RDI was 25.9 (min 10.2 h )1, max h )1, STD 25.9 h )1 ). In the group with pathological PG-RDI, TTIR-RDI was >10 h )1 in 37 patients. Only 10 out of 134 patients with a PG- RDI < 10 h )1 demonstrated a TTIR-RDI > 10 h )1. Sensitivity was accordingly 80.4%, specificity 92.5%, positive predictive value 73%, and negative predictive value 97%. Correlation analysis revealed a correlation coefficient of 0.92 (Fig. 6). Figs 7 and 8 depict ROC (AUC 0.962, 95% CI ) and Bland Altman results. DISCUSSION This is the first study to report that TTIR incorporated into a Holter ECG represents a useful tool for the screening of nocturnal apneas and hypopneas. In a head-to-head comparison with full in-laboratory polysomnography and ambulatory

5 Screening for sleep-related breathing disorders 459 Figure 3. Correlation analysis polysomnographic respiratory disturbance index (PSG-RDI) versus transthoracic impedance recording respiratory disturbance index (TTIR-RDI) (log axes). Figure 6. Correlation analysis polygraphic respiratory disturbance index (PG-RDI) versus transthoracic impedance recording respiratory disturbance index (TTIR-RDI) (log axes). 100 TTIR-RDI 100 TTIR-RDI Sensitivity (%) AUC: Sensitivity (%) AUC: Specificity (%) specificity (%) Figure 4. Receiver operating characteristic for transthoracic impedance recording compared with in-hospital polysomnography. Figure 7. Receiver operating characteristic for transthoracic impedance recording compared with ambulatory polygraphy. Figure 5. Bland Altman plot polysomnographic respiratory disturbance index (PSG-RDI) versus transthoracic impedance recording respiratory disturbance index (TTIR-RDI). Figure 8. Bland Altman plot polygraphic respiratory disturbance index (PG-RDI) versus transthoracic impedance recording respiratory disturbance index (TTIR-RDI).

6 460 A. Mueller et al. monitoring, TTIR demonstrates high levels of sensitivity (95% and 80.4% respectively), specificity (97.2% and 92.5%), as well as positive (95% and 73%) and negative (97% and 97%) predictive values for the detection of SRBD. Simultaneous investigation of cardiac arrhythmias and screening for SRBD are feasible. Transthoracic impedance recording Until now, TTIR has been applied for various purposes. Noninvasive monitoring of cardiac output by impedance recording, for example, has been favorably compared with standard methods such as pulmonary artery thermodilution techniques, with achievement of equivalent results (Sageman et al., 2002). Frerichs et al. (2002) and Hahn et al. (1995) validated the ability of electrical impedance tomography to detect local changes in pulmonary air content, and found significant correlation to spirometry and electron-beam computed tomography. Rateadaptive pacemakers utilize impedance recording for the detection of minute ventilation. As early as , Mond et al. (1988) and Nappholtz et al. (1986) used impedance recording between the tip of a pacemaker electrode and the pulse generator case to provide parameters for rate control. Numerous studies have confirmed high correlations between impedance recording and actual minute ventilation (Cole et al., 2001; Jordaens et al., 1989; Kay et al., 1989; Lau et al., 1989). Most recently, highresolution impedance recording integrated into a cardiac pacemaker was used to detect pulmonary congestion (Yu et al., 2005). In one of the cases cited by this study, impedance fell by 13% (P < ) from baseline during 2 weeks prior to hospital admission for worsening heart failure. Whole-body impedance recording for diagnosis of sleep apnea has been evaluated in a study by Saarelainen et al. (2003). In 12 patients studied here, sensitivity of impedance cardiography for the detection of sleep apnea was 89% and specificity, 80%. As the relevance of SRBD for patients with cardiac pacemakers has been discovered during recent years, transthoracic impedance, measured between the tip of the ventricular electrode and the pacemaker can, has been employed with satisfactory results for the diagnosis of SRBD (Defaye et al., 2004). Scharf et al. (2004) revealed the high accuracy of pacemaker transthoracic impedance signals for the detection of sleep apnea. In a cohort of 22 patients, 13 showed a PSG-RDI > 10. Of these 13 patients, 12 also demonstrated derived pacemaker impedance of apnea hypopnea index, AHI > 10. These results, obtained in a small patient cohort, are comparable with our in-laboratory findings, and are somewhat more exact than our ambulatory results. Further studies will be required to investigate the influence of pulmonary diseases, thoracic malformations, and body position on TTIR. Comparison of TTIR and portable monitoring devices in the investigation of suspected sleep apnea The American Sleep Disorders Association (ASDA) has classified sleep study systems into four categories: in-laboratory attended standard polysomnography (level 1), unattended home sleep study with comprehensive portable devices incorporating the same channels as the in-laboratory standard polysomnography (level 2), unattended devices that measure at least four cardiorespiratory parameters (level 3), and unattended devices recording one or two parameters (level 4) (Flemons et al., 2003; ATS/ACCP/AASM Taskforce Steering Committee, 2004). In an early study with MESAM 4 Ò (Madaus Inc., Freiburg, Germany), a level 3 device, Stoohs and Guilleminault (1992) reported a high correlation between the oxygen desaturation index from MESAM 4 and the RDI from PSG (r ¼ 0.89). For the same device, Esnaola et al. (1996) (ERJ 1996) revealed sensitivity of 98% and specificity of 78%, using an AHI threshold of 10 h )1. The PolyG Ò device (CNS Inc., Chantassen, MN, USA) which records oronasal airflow (thermistor), chest and abdominal excursion, oxygen saturation, ECG, and body position was tested by Man and Kang (1995). They disclosed a sensitivity of 82.6% and a specificity of 94.9%, with the use of an AHI threshold of 5h )1, and 85.7% and 94.7%, respectively, when the threshold was set to 15 h )1. Using the same thresholds, Fietze et al. (2002) reported, for the MERLIN Ò device (Healthdyne Technologies, Marietta, GA, USA), sensitivity of 94.4 and specificity of 83.3, as well as sensitivity of 88.5% and specificity of 97.5% respectively. In a study using a level 4 device, based on the peripheral arterial tone, Bar et al. (2003) reported high correlation (0.88; P < ) with results obtained from poysomnography. The results for TTIR for apnea and hypopnea detection are within the range of results reported for level 3 and level 4 devices (Bar et al., 2003; Ferber et al., 1994; Levy et al., 1992). Comparison of TTIR- and Holter ECG-based parameters in the investigation of suspected sleep apnea Standard Holter ECG-based systems for SRBD screening have been intensively investigated. Roche et al. (1999) reported a sensitivity of 83% with the use of standard deviations of the normal-to-normal beat interval (SDNN) index, and a specificity of 96.5% using SDNN. The same author recently presented findings on the application of data on heart rate increment as an approach for early detection of SRBD (Roche et al., 2004). Using an appropriate threshold, other authors have disclosed sensitivity of 73.9% and specificity of 76.6%. Stein et al. (2003b) applied simple heart rate tachograms derived from ECG monitoring, a technique first described by Guilleminault et al. (1984). Longer phases with cyclic variations in heart rate showed positive predictive accuracy of 86% for significant sleep apnea syndrome. These studies were limited by the small number of patients (n ¼ ) and by the exclusion of patients with atrial fibrillation, autonomic dysfunction (e.g. diabetes mellitus), paced rhythms, and/or frequent ventricular arrhythmias. In contrast, our study included a total of 236 patients. Furthermore, our use of TTIR-enabled investigation of patients with arrhythmias: a

7 Screening for sleep-related breathing disorders 461 crucial aspect, as this is the primary purpose of any Holter ECG. Rhythm disorders and SRBD Evidence exists that arrhythmias are strongly related to SRBD (Becker et al., 1998; Grimm et al., 2000). Compared with general cardiology patients (32%), the proportion of patients with obstructive sleep apnea (OSA) is significantly higher in patients with atrial fibrillation (49%) (Gami et al., 2004). In addition, patients with OSA who undergo DC cardioversion for atrial fibrillation carry a higher risk of recurrence of atrial fibrillation (82% versus 42%, P ¼ 0.013), if OSA is not treated with CPAP (Kanagala et al., 2003). Patients with ventricular arrhythmias who were at significant risk of sudden cardiac death, and who were implanted with an ICD, showed a prevalence of 40% for SRBD (Fries et al., 1999). An appreciable share of pacemaker patients ranging in studies from 30% (Fietze et al., 2000) to 61.9% (Defayse et al., 2004) also suffers from SRBD. TTIR offers the unique opportunity of studying arrhythmias together with SRBD simultaneously in a Holter ECG. No restrictions are necessary for patients with supraventricular or ventricular arrhythmias. Patients who have undergone Holter ECG with impedance recording and who have demonstrated positive TTIR-RDI (above 10 h )1 ) should undergo further diagnosis in a sleep laboratory. Therapeutic options should be discussed only after definitive diagnosis in a sleep laboratory. Transthoracic impedance recording enables screening of a large group of patients, e.g. patients with cardiac arrhythmias. The widespread use of this method could help in establishing the diagnosis of SRBD in patients who would perhaps have not been recognized otherwise. Study limitations Analysis of TTIR was performed by a single investigator by means of digital signal-processing software. The calculation of interscorer variability was consequently not possible. An automatic tool has yet to be tested prospectively in a large trial. We applied a cut-off value of 50% TTI amplitude reduction to define apneas/hypopneas. This was a slightly arbitrary number derived from a small number of unblinded synchronized recordings. In this study, we did not differentiate among central, obstructive, or mixed apneas or hypopneas, or Cheyne Stokes respiration, as clinical consequences are similar. This correlates with the study by Scharf et al. (2004). Our definition for apneas and hypopneas did not include oxygen saturation criteria. We did not investigate the influence of body position on TTIR. In addition, a measurement of event-by-event agreement between impedance and conventional scoring was not performed. Our subjects were, in general, relatively lean (mean BMI ± SD ¼ 28.8 ± 5). Further study will therefore be necessary to disclose whether TTIR is as accurate in fatter patients. CONCLUSION For the first time, TTIR integrated in a Holter ECG was tested in a large patient cohort (n ¼ 236) for the detection of SRBD. Sensitivity and specificity were sufficiently high to effectively enable the application of TTIR parameters for screening tools. In contrast to heart rate-based parameters, TTIR can also be employed in patients with atrial fibrillation and ventricular rhythm disturbances. No exclusion criteria have been applied. The combination of TTIR and ECG recording also allows simultaneous analysis of SRBD and arrhythmias, and may provide new insights into the mechanism of arrhythmias in this special patient group. ACKNOWLEDGMENTS/DISCLOSURE OF CONFLICTS This study was supported in part by the company getemed AG (Teltow, Germany). getemed AG did not participate in the analysis or interpretation of the data. We thank getemed for their support. We thank all our colleagues from the Achenbach Krankenhaus, in Ko nigs Wusterhausen, Germany, for their helpful assistance in data acquisition. REFERENCES American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep, 1999, 22: ATS/ACCP/AASM Taskforce Steering Committee. Executive summary on the systematic review and practice parameters for portable monitoring in the investigation of suspected sleep apnea in adults. Am. J. Respir. Crit. Care Med., 2004, 169: Bar, A., Pillar, G., Dvir, I., Sheffy, J., Schnall, R. P. and Lavie, P. Evaluation of a portable device based on peripheral arterial tone for unattended home sleep studies. Chest, 2003, 123: Becker, H. F., Koehler, U., Stammnitz, A. and Peter, J. H. Heart block in patients with sleep apnoea. Thorax, 1998, 53: 29S 32S. Becker, H. F., Jerrentrup, A., Ploch, T., Grote, L., Penzel, T., Sullivan, C. E. and Peter, J. H. Effect of nasal continuous positive airway pressure treatment on blood pressure in patients with obstructive sleep apnea. Circulation, 2003, 107: Bradley, T. D. and Floras, J. S. Sleep apnea and heart failure. Part II. Central sleep apnea. Circulation, 2003, 107: Bradley, T. D., Logan, A. G., Kimoff, R. J., Series, F., Morrison, D., Ferguson, K., Belenkie, I., Pfeifer, M., Fleetham, J., Hanly, P., Smilovitch, M., Tomlinson, G. and Floras, J. S. Continuous positive airway pressure for central sleep apnea and heart failure. N. Engl. J. Med., 2005, 353: Cole, C. R., Jensen, D. N., Cho, Y., Portzline, G., Candinas, R., Duru, F., Adler, S., Nelson, L., Condie, C. and Wilkoff, B. L. Correlation of impedance minute ventilation with measured minute ventilation in a rate responsive pacemaker. Pacing Clin. Electrophysiol., 2001, 24: Davies, C. W., Crosby, J. H., Mullins, R. L., Barbour, C., Davies, R. J. and Stradling, J. R. Case control study of 24 hour ambulatory blood pressure in patients with obstructive sleep apnoea and normal matched control subjects. Thorax, 2000, 55:

8 462 A. Mueller et al. Defaye, P., Pepin, J. L., Poezevara, Y., Mabo, P., Murgatroyd, F., Levy, P. and Garrigue, S. Automatic recognition of abnormal respiratory events during sleep by a pacemaker transthoracic impedance sensor. J. Cardiovasc. Electrophysiol., 2004, 15: Defayse, P., Pepin, J. L., Poezevara, Y., Mabo, P., Murgatroyd, F., Levy, P. and Garrigue, S. Automatic recognition of abnormal respiratory events during sleep by a pacemaker transthoracic impedance sensor. J. Cardiovasc. Electrophysiol., 2004, 15: Esnaola, S., Duran, J., Infante-Rivard, C., Rubio, R. and Fernandez, A. Diagnostic accuracy of a portable recording device (MESAM IV) in suspected obstructive sleep apnea. Eur. Respir. J., 1996, 9: Ferber, R., Millman, R., Coppola, M., Fleetham, J., Murray, C. F., Iber, C., McCall, V., Nino-Murcia, G., Pressman, M. and Sanders, M. Portable recording in the assessment of obstructive sleep apnea. ASDA standards of practice. Sleep, 1994, 17: Fietze, I., Rottig, J., Quispe-Bravo, S., Riedel, F., Witte, J., Baumann, G. and Witt, C. Sleep apnea syndrome in patients with cardiac pacemaker. Respiration, 2000, 67: Fietze, I., Glos, M., Rottig, J. and Witt, C. Automated analysis of data is inferior to visual analysis of ambulatory sleep apnea monitoring. Respiration, 2002, 69: Flemons, W. W., Littner, M. R., Rowley, J. A., Gay, P., Anderson, W. M., Hudgel, D. W., McEvoy, R. D. and Loube, D. I. Home diagnosis of sleep apnea: a systematic review of the literature. An evidence review cosponsored by the American Academy of Sleep Medicine, the American College of Chest Physicians, and the American Thoracic Society. Chest, 2003, 124: Frerichs, I., Hinz, J., Herrmann, P., Weisser, G., Hahn, G., Dudykevych, T., Quintel, M. and Hellige, G. Detection of local lung air content by electrical impedance tomography compared with electron beam CT. J. Appl. Physiol., 2002, 93: Fries, R., Bauer, D., Heisel, A., Juhasz, J., Fichter, J., Schieffer, H. and Sybrecht, G. W. Clinical significance of sleep-related breathing disorders in patients with implantable cardioverter defibrillators. Pacing Clin. Electrophysiol., 1999, 22: Gami, A. S., Pressman, G., Caples, S. M., Kanagala, R., Gard, J. J., Davison, D. E., Malouf, J. F., Ammash, N. M., Friedman, P. A. and Somers, V. K. Association of atrial fibrillation and obstructive sleep apnea. Circulation, 2004, 110: Grimm, W., Koehler, U., Fus, E., Hoffmann, J., Menz, V., Funck, R., Peter, J. H. and Maisch, B. Outcome of patients with sleep apneaassociated severe bradyarrhythmias after continuous positive airway pressure therapy. Am. J. Cardiol., 2000, 86: Guilleminault, C., Connolly, S., Winkle, R., Melvin, K. and Tilkian, A. Cyclical variation of the heart rate in sleep apnoea syndrome. Mechanisms, and usefulness of 24 h electrocardiography as a screening technique. Lancet, 1984, 1: Hahn, G., Sipinkova, I., Baisch, F. and Hellige, G. Changes in the thoracic impedance distribution under different ventilatory conditions. Physiol. Meas., 1995, 16: A161 A173. Jordaens, L., Berghmans, L., Van Wassenhove, E. and Clement, D. L. Behavior of a respiratory driven pacemaker and direct respiratory measurements. Pacing Clin. Electrophysiol., 1989, 12: Kanagala, R., Murali, N. S., Friedman, P. A., Ammash, N. M., Gersh, B. J., Ballman, K. V., Shamsuzzaman, A. S. and Somers, V. K. Obstructive sleep apnea and the recurrence of atrial fibrillation. Circulation, 2003, 107: Kay, G. N., Bubien, R. S., Epstein, A. E. and Plumb, V. J. Ratemodulated cardiac pacing based on transthoracic impedance measurements of minute ventilation: correlation with exercise gas exchange. J. Am. Coll. Cardiol., 1989, 14: Kunst, P. W., Vazquez, da., Bohm, S. H., Faes, T. J., Lachmann, B., Postmus, P. E. and de Vries, P. M. Monitoring of recruitment and derecruitment by electrical impedance tomography in a model of acute lung injury. Crit. Care Med., 2000, 28: Lau, C. P., Antoniou, A., Ward, D. E. and Camm, A. J. Reliability of minute ventilation as a parameter for rate responsive pacing. Pacing Clin. Electrophysiol., 1989, 12: Levy, R., Cosio, M., Gibbons, L., Macklem, P. and Martin, J. Induction of sleep apnoea with negative pressure ventilation in patients with chronic obstructive lung disease. Thorax, 1992, 47: Man, G. C. and Kang, B. V. Validation of a portable sleep apnea monitoring device. Chest, 1995, 108: Mond, H., Strathmore, N., Kertes, P., Hunt, D. and Baker, G. Rate responsive pacing using a minute ventilation sensor. Pacing Clin. Electrophysiol., 1988, 11: Nappholtz, T., Valenta, H., Maloney, J. and Simmons, T. Electrode configurations for a respiratory impedance measurement suitable for rate responsive pacing. Pacing Clin. Electrophysiol., 1986, 9: Pepperell, J. C., Ramdassingh-Dow, S., Crosthwaite, N., Mullins, R., Jenkinson, C., Stradling, J. R. and Davies, R. J. Ambulatory blood pressure after therapeutic and subtherapeutic nasal continuous positive airway pressure for obstructive sleep apnoea: a randomised parallel trial. Lancet, 2002, 359: Roche, F., Gaspoz, J. M., Court-Fortune, I., Minini, P., Pichot, V., Duverney, D., Costes, F., Lacour, J. R. and Barthelemy, J. C. Screening of obstructive sleep apnea syndrome by heart rate variability analysis. Circulation, 1999, 100: Roche, F., Duverney, D., Court-Fortune, I., Pichot, V., Costes, F., Lacour, J. R., Antoniadis, J. A., Gaspoz, J. M. and Barthelemy, J. C. Cardiac interbeat interval increment for the identification of obstructive sleep apnea. Pacing Clin. Electrophysiol., 2002, 25: Roche, F., Sforza, E., Duverney, D., Borderies, J. R., Pichot, V., Bigaignon, O., Ascher, G. and Barthelemy, J. C. Heart rate increment: an electrocardiological approach for the early detection of obstructive sleep apnoea/hypopnoea syndrome. Clin. Sci. (Lond.), 2004, 107: Saarelainen, S., Himanen, S. L., Hasan, J., Virkkala, J. and Koobi, T. Whole-body impedance recording a practical method for the diagnosis of sleep apnoea. Clin. Physiol. Funct. Imag., 2003, 23: Sageman, W. S., Riffenburgh, R. H. and Spiess, B. D. Equivalence of bioimpedance and thermodilution in measuring cardiac index after cardiac surgery. J. Cardiothorac. Vasc. Anesth., 2002, 16: Scharf, C., Cho, Y. K., Bloch, K. E., Brunckhorst, C., Duru, F., Balaban, K., Foldvary, N., Liu, L., Burgess, R. C., Candinas, R. and Wilkoff, B. L. Diagnosis of sleep-related breathing disorders by visual analysis of transthoracic impedance signals in pacemakers. Circulation, 2004, 110: Sin, D. D., Logan, A. G., Fitzgerald, F. S., Liu, P. P. and Bradley, T. D. Effects of continuous positive airway pressure on cardiovascular outcomes in heart failure patients with and without Cheyne Stokes respiration. Circulation, 2000, 102: Stein, P. K., Duntley, S. P., Domitrovich, P. P., Nishith, P. and Carney, R. M. A simple method to identify sleep apnea using holter recordings. J. Cardiovasc. Electrophysiol., 2003a, 14: Stein, P. K., Duntley, S. P., Domitrovich, P. P., Nishith, P. and Carney, R. M. A simple method to identify sleep apnea using Holter recordings. J. Cardiovasc. Electrophysiol., 2003b, 14: Stoohs, R. and Guilleminault, C. MESAM 4: an ambulatory device for the detection of patients at risk for obstructive sleep apnea syndrome (OSAS). Chest, 1992, 101: Young, T., Palta, M. and Dempsey, J. The occurrence of sleepdisordered breathing among middle aged adults. N. Engl. J. Med., 1993, 328: Yu, C. M., Wang, L., Chau, E., Chan, R. H., Kong, S. L., Tang, M. O., Christensen, J., Stadler, R. W. and Lau, C. P. Intrathoracic impedance monitoring in patients with heart failure: correlation with fluid status and feasibility of early warning preceding hospitalization. Circulation, 2005, 112: