Sleep Breath DOI 10.1007/s11325-014-1048-z

ORIGINAL ARTICLE

Accuracy of a novel auto-CPAP device to evaluate the residual apnea-hypopnea index in patients with obstructive sleep apnea Carlos Alberto Nigro & Sergio González & Anabella Arce & María Rosario Aragone & Luciana Nigro

Received: 12 March 2014 / Revised: 29 June 2014 / Accepted: 30 July 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Background Patients under treatment with continuous positive airway pressure (CPAP) may have residual sleep apnea (RSA). Objective The main objective of our study was to evaluate a novel auto-CPAP for the diagnosis of RSA. Methods All patients referred to the sleep laboratory to undergo CPAP polysomnography were evaluated. Patients treated with oxygen or noninvasive ventilation and split-night polysomnography (PSG), PSG with artifacts, or total sleep time less than 180 min were excluded. The PSG was manually analyzed before generating the automatic report from autoCPAP. PSG variables (respiratory disturbance index (RDI), obstructive apnea index, hypopnea index, and central apnea index) were compared with their counterparts from auto-CPAP through Bland–Altman plots and intraclass correlation coefficient. The diagnostic accuracy of autoscoring from auto-CPAP using different cutoff points of RDI (≥5 and 10) was evaluated by the receiver operating characteristics (ROCs) curve. Results The study included 114 patients (24 women; mean age and BMI, 59 years old and 33 kg/m2; RDI and apnea/ hypopnea index (AHI)-auto median, 5 and 2, respectively). The average difference between the AHI-auto and the RDI C. A. Nigro (*) Sleep Laboratory, Hospital Alemán, Pedro Goyena 620 3 B, CP 1424, Buenos Aires, Argentina e-mail: [email protected] S. González : A. Arce Sleep Laboratory, Hospital Alemán, Av. Pueyrredón 1640, CP 1118, Buenos Aires, Argentina M. R. Aragone Central Laboratory, Hospital Italiano, Potosí 4072, CP 1199, Buenos Aires, Argentina L. Nigro Hospital Ramos Mejia, General Urquiza 609, CP 1221, Buenos Aires, Argentina

was −3.5±3.9. The intraclass correlation coefficient (ICC) between the total number of central apneas, obstructive, and hypopneas between the PSG and the auto-CPAP were 0.69, 0.16, and 0.15, respectively. An AHI-auto >2 (RDI≥5) or >4 (RDI≥10) had an area under the ROC curve, sensitivity, specificity, positive likelihood ratio, and negative for diagnosis of residual sleep apnea of 0.84/0.89, 84/81 %, 82/91 %, 4.5/9.5, and 0.22/0.2, respectively. Conclusions The automatic analysis from auto-CPAP (S9 Autoset) showed a good diagnostic accuracy to identify residual sleep apnea. The absolute agreement between PSG and auto-CPAP to classify the respiratory events correctly varied from very low (obstructive apneas, hypopneas) to moderate (central apneas). Keywords Auto-CPAP . Obstructive sleep apnea . Residual sleep apnea . Residual apnea/hypopnea index

Introduction It has been reported that nearly 20 % of the patients with obstructive sleep apnea (OSA) treated with continuous positive airway pressure (CPAP) may have residual sleep apnea. These patients do often not report snoring, apneas referred by someone, daytime sleepiness, and impaired quality of life [1]. The consequences of residual sleep apnea are unknown but could be significant, since patients with even mild OSA may be at increased risk of cardiovascular complications and motor vehicle crashes [2–4]. Also, suboptimal CPAP treatment may be ineffective at lowering blood pressure [5]. These epidemiological data would emphasize the need to identify and properly treat the residual sleep apnea in subjects receiving CPAP. Ideally, the polysomnography (PSG) would be the recommended approach for the diagnosis of residual sleep apnea [6] but it is unusual and probably cost ineffective in routine

Sleep Breath

clinical practice to repeat a PSG after initial CPAP titration for the purpose of determining ongoing CPAP efficacy. Less complex alternatives such as a device based on peripheral arterial tonometry [7] have been used to detect residual obstructive events during CPAP therapy. Currently, several autoCPAP devices designed to detect apneas or hypopneas and provide information about the residual apnea-hypopnea index while patients are on CPAP have been developed and have come into clinical use. Different automatic CPAP devices differ in the analyzed signals, in the definition of respiratory events, in the signal processing, as well as in the algorithm of pressure response. This results in significant differences in the positive pressure behavior in response to bench-simulated or naturally occurring sleep-induced breathing disturbances and in the residual apnea/hypopnea index reported by the automatic analysis of each device [8]. Some publications have compared the automatic residual apnea/hypopnea index that provide the auto-CPAP devices against the manual respiratory disturbance index calculated from the polysomnography [9–12], but there is little evidence of the accuracy of these devices to diagnose residual apnea syndrome [13, 14]. An innovative auto-CPAP machine has been recently developed (S9 Autoset™, ResMed Corp, Sydney, Australia) which adds an algorithm to detect central sleep apnea (CSA) events and differentiate them from obstructive events. The main objective of our study was to evaluate the accuracy of S9 Autoset™ to diagnose residual sleep apnea in patients referred for a polysomnography and CPAP titration.

Materials and methods Subjects A prospective clinical study was conducted in 148 consecutive patients with diagnosis of OSA referred to the Hospital Alemán Sleep Laboratory to undergo CPAP titration by polysomnography (PSG). The recruitment period extended from April 2012 to July 2013. The Hospital Alemán Ethics Committee approved the study protocol. The selection criteria were the following: Inclusion criteria 1. Adult patients, older than 18 years. 2. Written informed consent to participate in the study.

Exclusion criteria 1. Patients who underwent split-night studies, BiPAP titration, use of oxygen, or subjects who did not tolerate or refused CPAP.

2. Patients with significant comorbidities such as chronic obstructive pulmonary disease or hypoventilation syndromes. 3. Polysomnographies with artifacts in the electroencephalogram (EEG) or respiratory channels that did not allow an appropriate scoring, or a total sleep time less than 180 min.

Measurements and analysis Prior to the polysomnography, the patients completed a clinical history. All the subjects underwent CPAP titration by polysomnography with a computerized polysomnographic system (NEUROTRACE or MINI-PC; Akonic, Buenos Aires, Argentina), including electroencephalogram (F4/A2, C4/A2, and O2/A2), bilateral electrooculogram, submental electromyogram, bilateral leg electromyogram, and electrocardiogram. Airflow was measured with a pressure sensor in line with the CPAP mask and an oronasal thermistor; respiratory effort was assessed by thoracic and abdominal

Table 1 Patient characteristics Patient number Age (years)* Men BMI (body mass index, kg/m2)* Apnea/hypopnea index on diagnostic study** Comorbidities - Hypertension - Coronary heart disease - Cerebrovascular ischemia - Cardiac arrhythmias - Diabetes type II - Metabolic syndrome Polysomnographic findings on CPAP titration - TRT (total recording time, min)** - TST (total sleep time, min)* - TWT (total wakefulness time, min)* - SE (sleep efficiency)* -

TNREM (min)* TREM (min)** Residual apnea/hypopnea index (r AHI)** Residual respiratory disturbance index (r RDI)** • r RDI ≥5–10 had a sensitivity of 45 %, a specificity of 97 %, and a positive likelihood ratio of 14.5 for the presence of RSA. Desai et al. [14] evaluated 99 subjects with OSA. The estimated AHI from auto-CPAP was compared w ith the AHI from an overnight polysomnogram on auto-CPAP. RSA was defined as a PSG AHI ≥5. An auto-CPAP cutoff for the AHI of six events per hour was shown to be optimal for differentiating patients with and without RSA with a sensitivity of 92 % and a specificity of 90 % with a positive likelihood ratio of 9.6 and a negative likelihood ratio of 0.085. Our results are not comparable with those of Mulgrew et al. because the criterion for determining residual sleep apnea was different and the number of patients evaluated by them was low. In addition, this study was not designed to validate the auto-CPAP device for the diagnosis of residual sleep apnea but to assess clinical predictors and patterns of residual sleep apnea in patients treated with CPAP. On the other hand, this study is similar to that reported by Desai et al. in terms of sensitivity/specificity, but differs in the criterion used to define residual sleep apnea by the auto-CPAP devices. Possibly, the main reason for this discrepancy is related to technological differences between the auto-CPAP devices and the algorithms used by them to detect respiratory events [8]. Also, we cannot rule out that the interobserver variability in PSG manual analysis, especially in identifying

10

Fig. 5 Bland–Altman plot of automatic auto-CPAP apnea/ hypopnea index (AHI-auto) and respiratory disturbance index from PSG (RDI)

+1.96 SD

5

4,2

AHI-auto - RDI

0 Mean -5

-3,5

-10

-1.96 SD -11,2

-15 -20 -25 0

10

20 RDI

30

40

Sleep Breath 10

Fig. 6 Bland–Altman plot of automatic auto-CPAP hypopnea index (HI-auto) and hypopnea index from PSG (HI-PSG)

+1.96 SD

5

4,1

HI-auto - HI-PSG

0 Mean -5

-4,0

-10

-1.96 SD -12,1

-15 -20

-25 0

5

10

15

20

25

30

HI-PSG

hypopneas and the different criteria used to classify an event as hypopnea, could explain some of the observed differences [19]. Regardless of the auto-CPAP models, published data suggest that the autoscoring of these devices tend to overestimate the AHI compared with PSG at the expense of the hypopnea index (mean difference between PSG and autoCPAP: 1 to 6, limits confidence 0.1 to 31) [10, 11, 14]. In contrast, the auto-CPAP S9 showed a tendency to underestimate the AHI compared to PSG (difference between AHI-auto and RDI - 3.5±3.9). This could be related to technological differences in the auto-CPAP devices to identify respiratory events or also to the criteria used in this study to define hypopnea and the fact that RERAs were included in the analysis. This may have led to the identification of a higher number of respiratory events in the PSG. In fact, the prevalence of RSA in our study was 50 %, a higher value than that reported by other authors [1, 6]. There was a moderate to good agreement (ICC 0.6–0.8) between the total number of apneas or central apneas identified by the auto-CPAP and PSG. Moreover, we observed a very low agreement (ICC2 PSG, RDI≥10 S9 Autoset, AHI>4

0.84 (0.039)

82.5 (70.1–91.3)

81 (68.1–90)

4.3 (2.5–7.4)

0.22 (0.1–0.4)

0.89 (0.04)

81 (63.6–93)

91 (83–96)

9.5 (4.6–19.7)

0.2 (0.1–0.4)

RSA residual sleep apnea, AUC-ROC (SE) area under the receiver operator curve (standard error), +LR positive likelihood ratio, −LR negative likelihood ratio, CI 95% confidence interval 95 %

Sleep Breath

PSG did not take into account the respiratory events during waking periods while the auto-CPAP considered the total recording time to classify the apneas. This study had some limitations. The first limitation is the applicability of these results at home. Due to the design of this study, we cannot draw valid conclusions about the accuracy of the S9 auto-CPAP to detect or exclude RSA outside the sleep laboratory without technical control. Secondly, the performance of this model of auto-CPAP to detect RSA in subjects treated with fixed CPAP is not known. For this purpose, it is necessary to perform a study comparing the autoCPAP versus PSG while patients receive a fixed pressure. In this way, one could evaluate the performance of the device when it is operating in CPAP mode. Finally, as no outcome measure was evaluated, we cannot know the clinical relevance of our findings for the management of RSA patients. In conclusion, the autoscoring of auto-CPAP (S9 Autoset) showed an acceptable diagnostic accuracy to identify residual sleep apnea. The device showed moderate agreement with PSG to identify central apnea, but it was too low for obstructive apneas and hypopneas. Acknowledgments The authors wish to thank Ms. Jaquelina Mastantuono for revising the English text. Conflict of interest This study had no financial support. We declare there were no conflicts of interest related to this investigation.

References 1. Baltzan MA, Kassissia I, Elkholi O, Palayew M, Dabrusin R, Wolkove N (2006) Prevalence of persistent sleep apnea in patients treated with continuous positive airway pressure. Sleep 29:557–663 2. Shahar E, Whitney CW, Redline S, Lee ET, Newman AB, Nieto FJ, O'Connor GT, Boland LL, Schwartz JE, Samet JM (2001) Sleepdisordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med 163:19–25 3. Young T, Blustein J, Finn L, Palta M (1997) Sleep-disordered breathing and motor vehicle accidents in a population-based sample of employed adults. Sleep 20:608–613 4. Young T, Peppard P, Palta M, Hla KM, Finn L, Morgan B, Skatrud J (1997) Population-based study of sleep-disordered breathing as a risk factor for hypertension. Arch Int Med 157:1746–1752 5. Pepperell JC, Ramdassingh-Dow S, Crosthwaite N, Mullins R, Jenkinson C, Stradling JR, Davies RJ (2002) Ambulatory blood pressure after therapeutic and subtherapeutic nasal continuous positive airway pressure for obstructive sleep apnoea: a randomised parallel trial. Lancet 359:204–210 6. Practice parameters for the indications for polysomnography and related procedures. Polysomnography Task Force, American Sleep Disorders Association Standards of Practice Committee (1997) Sleep 20:406–422 7. Pittman SD, Pillar G, Berry RB, Malhotra A, MacDonald MM, White DP (2006) Follow-up assessment of CPAP efficacy in patients

with obstructive sleep apnea using an ambulatory device based on peripheral arterial tonometry. Sleep Breath 10:123–131 8. Sériès F, Plante J, Lacasse Y (2008) Reliability of home CPAP titration with different automatic CPAP devices. Respir Res 9:56 9. Woodson BT, Saurejan A, Brusky LT, Han JK (2003) Nonattended home automated continuous positive airway pressure titration: comparison with polysomnography. Otolaryngol Head Neck Surg 128: 353–357 10. Ueno K, Kasai T, Brewer G, Takaya H, Maeno K, Kasagi S, Kawana F, Ishiwata S, Narui K (2010) Evaluation of the apnea-hypopnea index determined by the S8 auto-CPAP, a continuous positive airway pressure device, in patients with obstructive sleep apnea-hypopnea syndrome. J Clin Sleep Med 6:146–151 11. Ikeda Y, Kasai T, Kawana F, Kasagi S, Takaya H, Ishiwata S, Narui K (2012) Comparison between the apnea-hypopnea indices determined by the REMstar Auto M series and those determined by standard inlaboratory polysomnography in patients with obstructive sleep apnea. Intern Med 51:2877–2885 12. Prasad B, Carley DW, Herdegen JJ (2010) Continuous positive airway pressure device-based automated detection of obstructive sleep apnea compared to standard laboratory polysomnography. Sleep Breath 14:101–107 13. Mulgrew AT, Lawati NA, Ayas NT, Fox N, Hamilton P, Cortes L, Ryan CF (2010) Residual sleep apnea on polysomnography after 3 months of CPAP therapy: clinical implications, predictors and patterns. Sleep Med 11:119–125 14. Desai H, Patel A, Patel P, Grant BJ, Mador MJ (2009) Accuracy of autotitrating CPAP to estimate the residual apnea-hypopnea Index in patients with obstructive sleep apnea on treatment with autotitrating CPAP. Sleep Breath 13:383–390 15. Morgenthaler TI, Aurora RN, Brown T, Zak R, Alessi C, Boehlecke B, Chesson AL Jr, Friedman L, Kapur V, Maganti R, Owens J, Pancer J, Swick TJ (2008) Practice parameters for the use of autotitrating continuous positive airway pressure devices for titrating pressures and treating adult patients with obstructive sleep apnea syndrome: an update for 2007. An American Academy of Sleep Medicine report. Sleep 31:141–147 16. Rechtschafen A, Kales A (1968) A manual of standarized technology, techniques and scoring system for sleep stages of human subjects. Brain information Service, Braininformation institute, University of California, Los Angeles 17. American Sleep Disorders Association (1992) The Atlas Task Force. EEG arousals: scoring rules and examples. Sleep 15:174–184 18. Consenso Nacional sobre el Síndrome Apneas-Hipopneas del Sueño del Grupo Español de Sueño Definición y concepto, fisiopatología, clínica y exploración del SAHS (2005) Arch bronconeumol 41 Extraordinario 4:12–29 19. Ayappa I, Norman RG, Krieger AC, Rosen A, O'malley RL, Rapoport DM (2000) Non-invasive detection of respiratory effortrelated arousals (REras) by a nasal cannula/pressure transducer system. Sleep 23:763–771 20. Masa JF, Corral J, Martín MJ, Riesco JA, Sojo A, Hernández M, Douglas NJ (2003) Assessment of thoracoabdominal bands to detect respiratory effort-related arousal. Eur Respir J 22:661–667 21. Sala H, Nigro C, Rabec C, Guardia AS, Smurra M (2001) Consenso Argentino de Trastornos Respiratorios Vinculados al Sueño. Medicina 61:351–363 22. Guilleminault C, Bassiri A (2005) Clinical features and evaluation of obstructive sleep apnea-hypopnea syndrome and upper airway resistance syndrome. In: Kryger MH, Roth T, Dement WC (eds) Principles and practice of sleep medicine, 4th edn. Elservier Inc, Philadelphia, pp 1043–1052 23. Pita Fernández S, Pértega Díaz S (2003) Pruebas diagnósticas. Cad Aten Prim 10:120–124