Effects of adaptive servoventilation to treat central sleep apnea in heart failure patients with reduced ejection fraction on exercise capacity and mortality: the Bad Oeynhausen ASV Registry

Background and purpose: Central sleep apnea (CSA) and Cheyne-Stokes respiration (CSR) are highly prevalent in heart failure patients with reduced left ventricular ejection fractions (HF-REF). Adaptive servoventilation (ASV) therapy was designed to treat CSA-CSR and intended to improve quality of life and survival in HF-REF patients. In contrast, the randomized controlled multi-center, multi-national SERVE-HF study demonstrates an increase in all cause and cardiovascular mortality in these patients. The Bad Oeynhausen ASV-registry prospectively enrolled all patients with comparable inclusion and exclusion criteria, and aimed to investigate changes in exercise capacity and all-cause mortality in these patients.Methods: From 01/2004 to 10/2013 we enrolled a total of 550 HF-REF patients (NYHA ≥ II, LV-EF ≤45%) with moderate to severe predominant CSA (AHI ≥15/h) into our prospective registry (NCT01657188). Exercise capacity was determined by standardized 6-minute walking tests (6MWT) and cardiopulmonary exercise testing at 3, 6, 12 and 24 months, in addition all-cause mortality by routine follow-up, clinical records, telephone calls to physicians and/or relatives.Results: Mean age was 67.7±10 years, 90% males, NYHA ≥III 52%, LV-EF 29.9±8%; AHI was 35.4±13.6/h, T<90% 58±73min. In 224 patients ASV was prescribed, 326 patients did not receive ASV. 6 MWT in controls was 342 m [95% CI 323–361] at baseline and 348 m [95% CI 320–376] at 24 months, and 347 [95% CI 330–364] at baseline and 343 m [95% CI 321–364] at 24 months in ASV treated patients (p=n.s.), respectively. Peak VO2 at baseline was 13.7 ml/kg/min [13.2–14.2] in controls and 13.8 ml/kg/min [13.2–14.4] in ASV-treated patients with no change to 24 months FU, 13.5 ml/kg/min [12.7–14.4] and 13.4 ml/kg/min [12.6–14.2] in controls and ASV treated patients, respectively. During median follow-up of 6.6 years, 109 (48.7%) ASV treated patients and 191 (58.6%) control patients died. Adjusted Cox modelling for ASV effects on mortality revealed a hazard ratio of 1.0 (95% confidence interval 0.77–1.29), p=0.996, whereas higher age, lower LV-EF and more hypoxemia (T<90%) at night were negative predictors for survival.Conclusion: Long-term ASV treatment of CSA-CSR in HF-REF patients showed to have no impact on exercise capacity and survival in our prospective single-academic centre registry.