Face: Cluster Phenotyping Predicting Outcomes In A Prospective Multicenter Cohort Study Of Chronic Heart Failure (Chf) Patients With Central Sleep Disorder Breathing (Sdb) Indicated For Adaptive Servo-Ventilation (Asv)

Aims: FACE cohort study was collecting real life clinical data and long-term follow-up in a heterogeneous CHF population with SDB eligible for ASV. Latent class analysis (LCA) was proposed to cluster homogeneous phenogroups that may behaved differently. Methods: Clinical presentation, sleep study, cardiac function and quality of life were assessed at inclusion. Patients were on CHF therapy and were offered to be treated by ASV. Morbidity and mortality were collected at 3-month follow-up. The primary end point in the time-to event analysis was the first event of death from any cause, unplanned hospitalization for worsening heart failure or cardiac transplant. Results: 509 CHF pts were included in the ITT analysis, LCA identified 6 well defined phenogroups. At 3-month follow-up, there was significant difference between clusters for primary outcome incidence p<0.01, and death from any cause or hospitalization for all causes (p<0.01) and hospitalizations for all causes (p<0.02). The worse prognosis occurred in the cluster 1 (low LVEF<35%, male (96%), predominant CSA (90%), stroke (38%), Afib (44%), ASV acceptance (53%)). Discussion: A phenogroup analysis including cardiac functions, SDB and ASV acceptance is promising in predicting outcomes of CHF patients.