Abstract: SA-OR003
Temporal Kinetics of Sepsis-Associated AKI Subphenotypes
Session Information
- AKI Advances: Biomarkers, Outcomes, and Clinical Trials
November 08, 2025 | Location: Room 320A, Convention Center
Abstract Time: 04:50 PM - 05:00 PM
Category: Acute Kidney Injury
- 102 AKI: Clinical, Outcomes, and Trials
Authors
- Kiernan, Elizabeth, University of Washington, Seattle, Washington, United States
- Zelnick, Leila R., University of Washington, Seattle, Washington, United States
- Mabrey, Frances Linzee, University of Washington, Seattle, Washington, United States
- de Boer, Ian, University of Washington, Seattle, Washington, United States
- Himmelfarb, Jonathan, Mount Sinai Health System, New York, New York, United States
- Kestenbaum, Bryan R., University of Washington, Seattle, Washington, United States
- Bhatraju, Pavan K., University of Washington, Seattle, Washington, United States
Background
Sub-phenotypes may help define AKI pathogenesis, predict outcomes, and enrich clinical trials. Sub-phenotypes may evolve over time, but temporal changes during hospitalization and associations of changing sub-phenotypes with clinical outcomes are not well described.
Methods
Participants enrolled in the Crystalloid Liberal or Vasopressors Early Resuscitation in Sepsis (CLOVERS) clinical trial with AKI at study enrollment who had plasma collected at randomization (V1) and at least one additional time point (24 hrs (V2) and/or 72 hours (V3)) were included. Participants were classified into two sub-phenotypes (SP1 and SP2) using a previously validated 3-biomarker prediction model with SP2 characterized by higher markers of inflammation and endothelial dysfunction. We used Kaplan-Meier 28-day and 90-day mortality point estimates and tested differences by category using a Wald test together with Greenwood’s standard error estimate. To examine the outcome of renal replacement therapy (RRT)-free days, we used linear regression with Huber-White robust standard errors.
Results
Among 374 participants, 240 were classified as SP1 and 134 as SP2 at randomization and 22% of participants changed sub-phenotype classification by V3 (Table 1). Participants classified as SP2 at randomization were more likely to transition phenotypes by V3, with 49% of participants classified as SP2 transitioning to SP1 and 8% of SP1 transitioning to SP2 by V3. 28-day mortality was lowest among those who were SP1 and remained SP1 (5%) and highest among those who were SP2 and remained SP2 (33%). Participants who transitioned from SP1 to SP2 had a 28-day mortality of 28%. Participants who were SP2 and remained SP2 as well as participants who transitioned to SP2 had the fewest days free from renal replacement therapy (RRT).
Conclusion
We identified dynamic changes in AKI sub-phenotype classification associated with clinical outcomes. These results suggest clinically relevant changes of the inflammatory response in sepsis-associated AKI and the utility for repeat sub-phenotype classification during hospitalization.
Funding
- Other NIH Support