Win ratio analysis of transvenous phrenic nerve stimulation to treat central sleep apnoea in heart failure

Aims: Central sleep apnoea (CSA) is present in 20–40% of heart failure (HF) patients and is associated with poor clinical outcomes and health status. Transvenous phrenic nerve stimulation (TPNS) is an available treatment for CSA in HF patients. The impact on HF outcomes is incompletely understood. The win ratio (WR) allows inclusion of multiple endpoint components, considers the relative severity of each component, and permits assessment of recurrent events in evaluation of clinical benefit. Methods and results: A WR hierarchy was pre-defined for analysis of the HF subgroup of the remedē® System Pivotal Trial. The analysis used three hierarchical components to compare all treated to all control subjects: longest survival, lowest HF hospitalization rate, and ≥2-category difference in Patient Global Assessment at 6 months. Sensitivity analyses were performed substituting Epworth Sleepiness Scale and 4% oxygen desaturation index for the third component, and a 4-component WR hierarchy was also evaluated. Ninety-one HF subjects, 43 receiving TPNS and 48 in the control group, provided 2064 pairwise comparisons. More patients treated with TPNS experienced clinical benefit compared with control (WR 4.92, 95% confidence interval 2.27–10.63, P < 0.0001). There were 1111 (53.83%) winning pairwise comparisons for the treatment group and 226 (10.95%) for the control group. Similarly, large WRs were observed for all additional WR hierarchies. Conclusions: This WR analysis of the remedē® System Pivotal Trial suggests that TPNS may be superior to untreated CSA in HF patients with CSA using a hierarchical clinical benefit endpoint composed of mortality, HF hospitalization, and health status.