Abstract: TH-PO1082
Composite Primary Outcome Variables in Nephrology Clinical Trials
Session Information
- CKD: Therapeutic Advances
October 24, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 2302 CKD (Non-Dialysis): Clinical, Outcomes, and Trials
Authors
- Fishbane, Steven, Northwell Health Division of Nephrology, Great Neck, New York, United States
- Upadrista, Pratap Kumar, Northwell Health Division of Nephrology, Great Neck, New York, United States
- Shah, Hitesh H., Northwell Health Division of Nephrology, Great Neck, New York, United States
Background
Clinical trials use one or more primary outcome variables to assess results. The type of outcome selected can affect sample size needed and interpretation of results. Composite endpoints have been increasingly popular in nephrology trials. Although potentially beneficial, composites have also been criticized for leading to potentially misleading results. In this analysis we studied nephrology trials with a special emphasis on composite outcomes and the component variables from which the composites were constructed.
Methods
We searched nephrology trials published in major medical and nephrology journals between 1/1/2019 – 12/31/2023. Only articles based on primary trials were used; post hoc or other secondary analyses were excluded. For each publication a reviewer assessed the primary outcome as well as study characteristics and results. For each study that used composite primary outcomes (CPO), the individual component results were analyzed.
Results
183 studies were analyzed, of which the primary endpoint was a single outcome variable in 123 (67.2%), co-primary variables in 28 (15.4%) and CPO in 32 (17.4%) (Table). With CPOs, the mean number of individual component variables was 3.4±1.6. Studies with CPOs had larger sample sizes (<0.0001) and were significantly more likely to have positive study results (p=0.03, Table), with 50% of the CPO component variables having positive results. We found no association between the number of component variables and likelihood of positivity of the overall CPO. Trials with CPO were more likely to be published in medical than nephrology journals 25/70 (35.7%) vs. 7/113 (6.2%), p<0.0001. Reporting of CPO was frequently associated with failure to provide results for the composite’s individual component variables (42.9%).
Conclusion
Nephrology trials commonly utilize composite primary endpoints. In this study we describe the impact on study results, the quality of reporting of CPOs and note how these endpoints can alter the interpretation of studies.
| Composite (17.4%) | Single or Co-Primary (82.6%) | P Value | |
| Journal Type | |||
| Medical | 25/70 (35.7%) | 45/70 (64.3%) | 0.0001 |
| Nephrology | 7/113 (6.2%) | 106/113 (93.8%) | <0.0001 |
| <0.0001 | |||
| Mean # Patients | 2,957.3±3765.4 | 494.2±1070.6 | <0.0001 |
| Positive Study Results | 20/32 (62.5%) | 62/151 (41.1%) | 0.03 |
| Composite Outcomes | |||
| Mean # of Components | 3.4±1.6 | ||
| Positive Studies with 3 or less Components | 13/21 (61.9%) | ||
| Studies with Insufficient Description of Composite Components | 9/21 (42.9%) | ||
Funding
- Clinical Revenue Support