Kidney Week

Abstract: PUB138

Predictors of Intradialytic Hypotension in Critically Ill Patients Undergoing Kidney Replacement Therapy: A Systematic Review

Session Information

• Publication Only

Category: Dialysis

• 801 Dialysis: Hemodialysis and Frequent Dialysis

Authors

• da Cunha Lyrio, Rafaella Maria, Universidade Salvador, Salvador, BA, Brazil

Rafaella Maria da Cunha Lyrio, MD

• Calcagno, Tess Moore, Cleveland Clinic, Cleveland, Ohio, United States

Tess Moore Calcagno, MD

• Sampaio, Estevão Farias, Hospital Geral Ernesto Simões Filho, Salvador, BA, Brazil

Estevão Farias Sampaio, MD

• Hennemann Sassi, Rafael, Hospital de Clinicas de Porto Alegre, Porto Alegre, RS, Brazil

Rafael Hennemann Sassi, MD

• Passos, Rogerio, Sociedade Beneficente Israelita Brasileira Albert Einstein, São Paulo, SP, Brazil

Rogerio Passos, MD

Background

This systematic review aims to identify predictors of intradialytic hypotension (IDH) in critically ill patients undergoing renal replacement therapy (RRT) for acute kidney injury (AKI).

Methods

A comprehensive search of PubMed was conducted from 2002 to April 2024. Inclusion criteria comprised studies involving critically ill adults undergoing RRT for AKI. Exclusion criteria included pediatric patients, non-critically ill individuals, those with chronic kidney disease, and those not undergoing RRT. The primary outcome was identifying predictive tools for hypotensive episodes during RRT sessions.

Results

The review analyzed data from 8 studies involving 2,873 patients. Various machine learning models were assessed for their predictive accuracy. The Extreme Gradient Boosting Machine (XGB) model emerged as the top performer, achieving an area under the receiver operating characteristic curve (AUROC) value of 0.828 (95% CI: 0.796–0.861). It was closely followed by the deep neural network (DNN) with an AUROC of 0.822 (95% CI: 0.789–0.856). Notably, all AUROC values from machine learning models surpassed those of other predictors. The SOCRATE score, incorporating cardiovascular SOFA score, index capillary refill, and lactate level, showed increasing IDH incidence rates with the number of parameters, yielding an AUROC of 0.79 (95% CI: 0.69–0.89, P < 0.0001), indicating robust predictive accuracy. Peripheral perfusion index (PPI) and heart rate variability (HRV) exhibited AUROCs of 0.721 (95% CI: 0.547–0.857) and 0.761 (95% CI: 0.59–0.887), respectively. Additionally, pulmonary vascular permeability index (PVPI) and mechanical ventilation displayed significant diagnostic performance, with an area under the curve (AUC) of 0.68 (95% CI: 0.53–0.83) and 0.69 (95% CI: 0.54–0.85), respectively. A PVPI ≥ 1.6 at the onset of intermittent hemodialysis (IHD) sessions predicted IDH associated with preload dependence with a sensitivity of 91% (95% CI: 59–100%) and specificity of 53% (95% CI: 42–63%).

Conclusion

Machine learning models and clinical parameters offer promise in predicting IDH in critically ill AKI patients undergoing RRT. Further research is needed to manage hypotensive episodes effectively and improve patient outcomes.