HOW WELL WERE HEART FAILURE CLINICAL TRIAL ENDPOINTS SELECTED? EVIDENCE FROM A REVIEW OF CLINICALTRIALS.GOV AND OBSERVATIONAL STUDIES

Purpose:  Value of Information (VOI) analysis has received increasing attention as a framework for prioritizing clinical trial investment. Value calculations in VOI analysis require direct links from clinical endpoints to comprehensive outcomes such as mortality and quality of life. The purpose of this study is to systematically review surrogate and comprehensive endpoints used in heart failure clinical trials and document evidence on the links between surrogate endpoints and mortality.

Method:  We first searched the ClinicalTrials.gov database from the inception of the database to 2015 for all interventional heart failure trials. We then extracted the primary outcomes and broke down composite outcomes into individual outcomes. We then conducted a systematic review of long-term (>=5 years) observational studies in heart failure in PubMed published during the past 20 years. We examined if the surrogate endpoints used in heart failure trials were proven to be associated with long-term mortality.

Result:  738 heart failure clinical trials that were registered on ClinicalTrials.gov reported a wide variety of primary outcomes. A composite endpoint that included mortality was the most frequently used primary endpoint, reported in 13% of all heart failure trials. 3.1% included mortality as an individual primary outcome. Other frequently used primary outcomes included left ventricular ejection fraction (LVEF) (6.5%), exercise capacity (5.6%), brain natriuretic peptide (3.0%), hospitalization (1.9%), etc. Hospitalization was the most frequently used outcome in a composite outcome alongside mortality, reported in 86% of such studies, followed by heart failure (13%), myocardial infarction (13%), stroke (12%), etc. A search of PubMed yielded 80 eligible studies. Most of the frequently used trial endpoints had been examined in at least one long-term observational study for their associations with mortality. LVEF was linked to long-term mortality in 29% of all studies. Blood pressure, New York Heart Association status, heart rate, and serum creatinine made up the remaining four of the top five endpoints linked to long-term mortality and were reported in 26%, 19%, 14% and 7.5% of studies, respectively.

Conclusion:   The use of surrogate endpoints in heart failure clinical trials is extensive. While many of the frequently used surrogate endpoints were found to be associated with long-term mortality, the strength of evidence varied significantly. This should be considered when selecting trial endpoints, in order to answer the question of VOI.