S16 Detection of sleep-disordered breathing in chronic heart failure patients: utility of heart rate variability versus pulse oximetry?

Abstract

Introduction and Objectives:Sleep-disordered breathing (SDB) is a frequent comorbidity in chronic heart failure (CHF). Patients are often asymptomatic and sleep studies may be required for SDB diagnosis. Our department has previously reported that %VLFI component of Heart Rate Variability (HRV) is correlated with apnoea-hypopnoea index (AHI) in CHF (r=0.52). Thus, we tested the hypotheses that %VLFI component of HRV, or pulse oximetry, can be used to rule out SDB in patients with CHF. Methods Stable CHF patients attending cardiology clinics were enrolled, irrespective of cause or severity of CHF. Patients were studied using polysomnography, simultaneous ambulatory electrocardiography and pulse oximetry. SDB was defined as AHI (greater-than or equal to) 15.0/h, measured by polysomnogram. Fourier analysis of the electrocardiogram was used to measure %VLFI component of HRV, with a cutoff (greater-than or equal to) 2.23% to indicate SDB. The oxygen desaturation index (ODI) (greater-than or equal to) 3% was measured by pulse oximeter, with a cutoff >7.5 desaturations/ h to indicate SDB. Diagnostic performance of %VLFI and ODI(greater-than or equal to) 3% were calculated, with the polysomnogram as reference standard for SDB diagnosis. Results: 180 CHF patients were studied, seven were excluded due to insufficient sleep (<200 min). In 173 CHF patients (mean (SD) age 66.9 (13.0) years; 86% male; Epworth Sleepiness Scale 7.6 (4.3); NYHA 2.1 (0.6); median (IQR) BNP 118 (55e239) pg/ml), SDB was present in 77 (45%) patients with mean AHI 32.4 (18.2)/h. %VLFI was measured in 77 (45%) patients: in CHF patients with SDB (n=36), mean %VLFI was 3.13% (2.4) compared to 3.25% (2.6) in patients without SDB (n=41). Cardiac pacing, atrial fibrillation and frequent ectopy prevented %VLFI measurement in the remainder. ODI (greater-than or equal to) 3% was measured in 171 patients: in CHF patients with SDB (n=76), mean ODI (greater-than or equal to) 3% was 29.2 (17.2)/h compared to 10.2 (6.4)/h in patients without SDB (n=95). Conclusion: The %VLFI component of HRV has no utility to screen for SDB in patients with CHF. Moreover, it could not be measured in more than half of this cohort of patients. In contrast, the high sensitivity and negative predictive value of the ODI (greater-than or equal to) 3% suggest pulse oximetry is a valuable tool to rule out SDB in CHF patients.(Table presented).