176-38: Comparison of pacemaker respiratory monitoring algorithm with polisomnography in the diagnosis of obstructive sleep apnoea

Abstract

Introduction: Obstructive sleep apnoea (OSA) is often associated with cardiovascular comorbidities, including heart failure, atrial fibrillation and hypertension. Regardless of its high prevalence, OSA is frequently under-diagnosed. It has been suggested that 25% of patients with sinus node disease (SND) or atrioventricular node disease (AVND) requiring pacemaker implantation present severe OSA. New generation pacemakers have respiratory monitoring algorithms based on the minute-ventilation sensor that monitor the respiratory distress index (RDI), thus identifying patients with possible OSA. The aim of this study was to compare the RDI obtained by pacemaker monitoring algorithms (IDR-PM) with the RDI obtained through polisomnography (PSG) and to determine the algorithm diagnostic accuracy. Methods: Unicentric prospective study of consecutive patients submitted to double-chamber pacemaker implantation or generator replacement, using the Reply 200 TM device (LivaNova Group). All patients were submitted to a clinical interview to access OSA symptoms and a PSG overnight study with RDI determination. OSA was diagnosed applying the American Academy of Sleep Medicine criteria. The RDI-PM during the period of the PSG study was registered. The accuracy of the RDI-PM threshold of .20, suggested by the customer for OSA suspicion, was determined and receiver operating characteristic (ROC) curve analysis applying C-statistics was used to explore if other value could improve its diagnostic performance. Summary: A total of 24 patients, aged 75 + 11 years, with SND (54%) or AVND (46%) were submitted to double-chamber pacemaker implantation or generator replacement using the Reply 200 Tm device. Definite OSA was confirmed in 50% of patients (mild: 17%; moderate: 25%; severe: 58%). The RDI-PM during the PSG period was found to be higher in patients with OSA [32 (21-36) vs. 9.5 (5-20) p = 0.008]. Although the correlation between the RDI-PM and the definite RDI obtained through PSG study has been moderate (Pearson R = 0.51; p = 0.011), the IDR-PM presented a high diagnostic accuracy for OSA diagnosis [AUC: 0.813 (95% IC: 0.62-1.0); p = 0.009]. The customer suggested threshold of 20 conferred a diagnostic accuracy of 79%. In this real-life population, the optimal RDIPM cut-off was 17.5 (sensitivity = 92%; specificity = 75%, positive predictive value = 79%, negative predictive value = 90%, overall diagnostic accuracy = 90%). Conclusion: This prospective study confirms the reliability of the respiratory monitoring algorithms based on minute-ventilation sensor available in pacemakers for OSA diagnosis. Our results suggest that RDI-PM diagnostic performance can be improved, but larger studies in real-life populations are required to identify the appropriate threshold. The respiratory monitoring algorithms available in the new generation pacemakers can be valuable tools for timely detection of OSA in clinical practice.