Online Program

Return to main conference page
Wednesday, September 27
Wed, Sep 27, 10:45 AM - 12:00 PM
Lincoln 6
Parallel Session: Challenges and Utility of Analyses of Actigraphy Data

Use of Accelerometer Data to Evaluate Physical Activity as a Surrogate Endpoint in Heart Failure Clinical Trials (300544)

*Tracy Bergemann, Medtronic, Plc. 

Background: Recent research of physical activity levels in heart failure patients shows a strong correlation with neurocognition and other more traditional functional assessments of heart failure, such as the exercise capacity test. Accelerometer technology provides the opportunity to collect objective data on physical activity continuously over time as a heart failure patient goes about their daily activities. Accelerometer data from current implantable devices such implantable cardiac defibrillators continuously record total daily active time (single-axis accelerometer). While increases in physical activity have face validity as a study objective in clinical trials, changes in summaries of the single-axis accelerometer recordings may function as a surrogate endpoint for traditional heart failure clinical outcomes. Methods: Patient level accelerometer, baseline, and clinical outcomes data from seven different randomized clinical trials of implantable cardiac defibrillators with or without cardiac resynchronization therapy were compiled. All trials had statistically significant treatment effects on primary objectives. Thresholds from accelerometer recordings are applied to translate data to activity in minutes per day. Summaries of accelerometer data over time were assessed as potential surrogate endpoints. The relationship between treatment effect and summaries of accelerometer data, and the relationship between summarized accelerometer data and clinical outcomes were tested. Results: Average activity levels collected during trial follow-up were strongly associated with traditional heart failure clinical outcomes such as time to first heart failure hospitalization or death. No difference was found, however, in average activity levels between randomized groups. The challenges inherit in the data analyzed, for e.g., missing data, variability, timing of collection, and their impact will be discussed. Several hypotheses shall be suggested that could improve the probability that accelerometer data may be used for surrogate endpoints in future studies.