Back
For best output, select "Paper Size" as "A4" and "Margin" as "0" or "None".
To save or print to PDF, please select Print Destination > Save as PDF, enable Background Graphics under "More Settings", then click "Save".
During the congress, E-Posters will be accessible to all participants on the congress website 24/7, as well as in the E-poster stations in the congress center.
Preparing your E-Poster
Please review the E-Poster format requirements carefully when preparing your E-Poster. Should your E-Poster not meet the mentioned requirements, it may not be displayed as described above.
E-Poster Submission Deadline
Please prepare and upload your E-Poster no later than March 14, 2026 11.59PM CET. After this date, you will no longer be able to prepare and upload your E-poster and it will not be displayed and accessible on the congress website.
Please follow the instructions below to input your abstract title.
Abstract titles should be brief and reflect the content of the abstract.
Fluid overload contributes to hypertension, heart failure, and increased mortality among dialysis patients. The Body Composition Monitor (BCM, Fresenius Medical Care) is a validated, non-invasive tool for assessing fluid status and body composition in this population. However, the cost and limited availability of BCM may restrict its widespread use in some dialysis facilities. We developed an AI-based tool-- with and without the use of BCM-- to measure overhydration in dialysis patients, which has been previously presented (reference 1). Building on our previously validated AI-based hydration model, we aimed to identify the most influential predictors of overhydration and propose a clinically actionable framework for its integration into dialysis practice.
Adult patients from Fresenius Medical Care clinics in Czech Republic, Italy, Portugal, Slovakia, Spain, and the United Kingdom were included. Data were obtained from the EuCliD database, covering January 2016 through September 2023. Variables included patient characteristics, comorbidities, laboratory markers, treatment parameters, medications, hospitalizations, and BCM measurements. The target variable was the overhydration (OH) value assessed by BCM. The final dataset comprised 36,803 patients with 920,182 target events. Patients were split per country into 80% training and 20% validation subsets. eXtreme Gradient Boosting (XGBoost) regression was used for model building. Two models were developed: one including the average of prior BCM assessments (w/ BCM) from the previous year, and another without BCM measurements (w/o BCM). Shapley additive explanations (SHAP) were applied to interpret the contribution of individual features to model predictions.
The w/BCM model achieved superior predictive performance (R² = 0.57 vs 0.36 w/o BCM; mean absolute error= 0.74 vs 0.93 w/o BCM) (reference 1), demonstrating that even non-BCM variables can meaningfully estimate hydration status. Figure 1 compared global feature importance with local explanations. For the w/ BCM model, the top global feature importance was the BCM-measured overhydration from the previous year, while another BCM measurement, total body water, did not appear among the top ten features. The four most important features in the w/o BCM model were effective OCM clearance, diabetes, effective OCM Kt/V, and dialysis vintage.
Using the overhydration prediction models, we proposed a possible clinical integration protocol. A dashboard displays predictions at each session to track OH values longitudinally. OH>2.5 L for three sessions or a consistent upward trend, triggers a clinical review. If fluid overload is confirmed based on patient symptoms, signs, and relative blood volume slopes, the recommended response includes stepwise dry weight reduction with ongoing monitoring of patient response and symptoms.
By identifying interpretable predictors of fluid overload, our model supports clinician trust and enables targeted, early intervention. Ongoing work will evaluate real-world integration of this tool across European dialysis centers.
