Abstract 3893: Endpoint Comparisons in an Advanced Heart Failure Trial: Data From the ESCAPE Trial
Background: The selection of endpoints in HF trials can affect the outcome of the trial, and optimal endpoints have been widely debated. The use of a composite endpoint may have potential disadvantages. It has been suggested that a global ranking endpoint may avoid the limitations of a standard composite endpoint. We sought to investigate whether or not the results of the ESCAPE trial would have shown a benefit of pulmonary artery catheter-guided therapy (PAC) vs usual care (UC) with different endpoints.
Methods: Results of ESCAPE were re-calculated with different endpoints for the entire cohort and subgroups. The Global Rank Endpoint was determined by the time to death, time to rehospitalization, and change in quality of life (QOL) from baseline. Patients with the worst event (death) were assigned the worst rank in order of time to event (patients who died first assigned worst rank, second patient who died second worse rank, etc); followed by the second worst event (time to rehosp) then change in QOL. Patients who did not have an event or missing follow-up QOL were assigned worse rank based on QOL change from baseline.
Results: shown below
Conclusion: There was a nonsignificant, consistent trend towards benefit of PAC when the global rank endpoint was used. Standard composite endpoints showed no benefit of treatment regardless of component specificity. Discharge sodium was the strongest predictor of patients in whom PAC-guided therapy showed no benefit on any endpoint, and harm based on the original endpoint. Selection of endpoints for clinical trials of acute heart failure should integrate cause-specific morbidity/mortality endpoints using novel techniques.