Abstract: SA-OR011
Hierarchical Composite End Points in the LIBERATE-D Trial: A Post Hoc Analysis
Session Information
- AKI Trials, Therapeutics, and Risk Stratification
October 24, 2026 | Location: Room 603, Convention Center
Abstract Time: 04:40 PM - 04:50 PM
Category: Acute Kidney Injury
- 102 AKI: Clinical, Outcomes, and Trials
Authors
- Siew, Edward D., Vanderbilt University School of Medicine, Nashville, Tennessee, United States
- Hsu, Chi-yuan, University of California San Francisco, San Francisco, California, United States
- Tuot, Delphine S., University of California San Francisco, San Francisco, California, United States
- Vijayan, Anitha, Intermountain Health, Salt Lake City, Utah, United States
- Matzumura Umemoto, Gonzalo, Washington University in St Louis, St. Louis, Missouri, United States
- Guide, Andrew, Vanderbilt University School of Medicine, Nashville, Tennessee, United States
- Greevy, Robert, Vanderbilt University School of Medicine, Nashville, Tennessee, United States
- McCoy, Ian Ellis, University of California San Francisco, San Francisco, California, United States
- Kwong, Yuenting Diana, University of California San Francisco, San Francisco, California, United States
- Birkelo, Bethany, University of Minnesota Twin Cities, Minneapolis, Minnesota, United States
- Zhuo, Hanjing, University of California San Francisco, San Francisco, California, United States
- Lee, Benjamin J., Houston Methodist, Houston, Texas, United States
- Chan, Ming-Jen, Chang Gung University, Taoyuan, Taoyuan City, Taiwan
- Liu, Kathleen D., University of California San Francisco, San Francisco, California, United States
Background
Hierarchical composite endpoints (HCEs) have been proposed as an alternative to single or composite endpoints to better inform the net effect of an intervention. In a post-hoc analysis of LIBERATE-D (NCT04218370), which compared two dialysis weaning strategies in patients with AKI-D, we examined the impact of the intervention within a HCE framework.
Methods
We compared the conservative dialysis intervention to a conventional 3x/week approach on the following HCEs: 1) Mortality at 28 days, 2) Dialysis dependence at 28 days, 3) Days with dialysis (calculated as 28 minus dialysis-free days). The Win Ratio (WR) was calculated as (# Wins (intervention)/# Losses (intervention), and the Win Odds (WO) was calculated as [Probwin + (0.5*Probtie)]/[Probloss + (0.5*Probtie)]. Sensitivity analyses included examining different thresholds for # dialysis days difference (1-7), classifying a total of <= 5 dialysis days as ties and examining outcomes to 90 days.
Results
The incidence of HCE outcomes by study arm are shown (Fig). Death within 28 days was higher with conservative dialysis, although the number was small and no case upon adjudication was related to lack of KRT. KRT dependence and dialysis days after randomization were less common with conservative dialysis. The WR and WO for the intervention were 1.70 (1.23, 2.45) and 1.58 (1.19, 2.17), respectively (the original primary outcome logistic regression OR was 1.76 (1.02-3.03) for being alive and not receiving dialysis at hospital discharge). Sensitivity analyses across 1-7 dialysis days difference, classifying <=5 dialysis days as ties, and at 90 days favored the conservative strategy. Subgroup analysis by eGFR and race were similar but limited by power.
Conclusion
When LIBERATE-D was re-analyzed using HCE, the superiority of the conservative dialysis approach remains evident. Observed WRs and WOs were robust to the threshold applied for dialysis days.
Acknowledgment
We would like to thank all of the patients who volunteered to participate in the LIBERATE-D Study. We would also like to thank the clinicians, staff, and coordinators who supported the study.
HCE Event Rates Stratified by Study Arm
Funding
- NIDDK Support