Abstract: TH-PO223
Hypotension at Initiation of Continuous Kidney Replacement Therapy: A Harbinger of Poor Outcomes
Session Information
- Hemodialysis, Hemodiafiltration, and Frequent Dialysis
October 24, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 801 Dialysis: Hemodialysis and Frequent Dialysis
Authors
- Thadani, Sameer, Baylor College of Medicine, Houston, Texas, United States
- Silos, Christin N., Baylor College of Medicine, Houston, Texas, United States
- Dolan, Kristin J., Baylor College of Medicine, Houston, Texas, United States
- Srivaths, Poyyapakkam, Baylor College of Medicine, Houston, Texas, United States
- Akcan Arikan, Ayse, Baylor College of Medicine, Houston, Texas, United States
Background
Hypotension occurs in 27-50% of connections to continuous renal replacement therapy (CRRT) yet its impact on outcomes in children has not been thoroughly investigated. Using machine learning techniques and clinical expertise, we aim to build upon prior research by examining the impact of hypotension at CRRT initiation on dialysis liberation.
Methods
Single center retrospective study of critically ill children who received CRRT between 9/2016-10/2018 at Texas Children’s Hospital. We utilized high resolution hemodynamics (1-minute intervals) to diagnose hypotension which was defined by a 20% decrease from baseline mean arterial pressure (MAP). Baseline MAP was the average MAP one hour prior to CRRT initiation. Illness severity was assessed using the PELOD-2 score. Liberation was defined as discontinuation of all dialytic modalities 30 days after CRRT initiation. We applied the LASSO method and clinical expertise to construct our final adjusted model.
Results
Our cohort included 59 patients with a median age and weight of 59 months (12-152) and 18 kgs (9-41) respectively. The median PELOD-2 score was 8 (6-10) and a total of 25 (42%) of patients required vasoactive medications at CRRT initiation. A total of 16 (27%) of patients had hypotension at CRRT initiation and 9 (15%) did not liberate. LASSO regression identified cardiac and gastrointestinal indications for ICU admission, history of organ transplantation, locations of hemodialysis catheter, and hypotension after initiation as significant variables. With addition of clinically relevant variables (PELOD-2, vasoactive inotropic score, age, and weight) and utilizing backwards selection, hypotension at CRRT initiation (aOR 0.21 (CI 0.04-0.98) remained negatively associated with liberation
Conclusion
Our model, which integrated machine learning and medical expertise, revealed that hypotension during CRRT initiation was associated with decreased odds of dialysis liberation. Our findings require further external validation but have identified a potentially modifiable factor associated with morbidity. Additional studies should investigate the variation in hypotension during CRRT and how its timing and duration impact patient outcomes.