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.
Despite widespread erythropoietin (EPO) use in chronic kidney disease anemia, the FDA declared in 2011 that no dosing strategy has been identified that doesn’t increase cardiovascular risks. Current management lacks validated methods to predict treatment response prospectively, leading to futile dose escalation in non-responders. We developed an integrated framework combining machine learning prediction with cardiovascular safety assessment to optimize EPO therapy.
We conducted a retrospective cohort study using MIMIC-IV database (v3.1), analyzing 4,539 CKD anemia patients across 7,929 admissions. We evaluated dose-response relationships and developed a Random Forest classifier trained on 2,166 patients with complete EPO-hemoglobin response data to predict treatment response. Primary outcomes were hemoglobin response (achievement of 10-12 g/dL target) and cardiovascular events stratified by EPO dose.
Nearly half (48.5%) of patients showed no hemoglobin response to EPO therapy, while only 11.8% achieved target hemoglobin levels. Patients with baseline hemoglobin <7 g/dL had 72% non-response rate. The Random Forest model achieved 82.1% accuracy (AUC 0.832) in predicting treatment response, with baseline hemoglobin as the strongest predictor (42.4% feature importance). Analysis of 60,418 cardiovascular events revealed dose-dependent risk: 18% at low doses (<4,000 units/day) increasing to 45% at very high doses (>20,000 units/day). Medium doses (4,000-10,000 units/day) optimised the efficacy-safety balance, achieving a 42% response rate with only a 22% cardiovascular event rate.
Figure 1. Machine learning prediction of EPO response in CKD anemia. (A) Random Forest model performance showing 82.1% accuracy and AUC of 0.832, outperforming logistic regression. (B) Feature importance analysis identifies baseline hemoglobin as the strongest predictor (42.4%) of EPO response. (C) Distribution of EPO response categories demonstrating that 48.5% of patients showed no hemoglobin response to therapy.
This study provides the first integrated framework addressing the FDA’s longstanding challenge by enabling prospective identification of EPO non-responders. Implementation could avoid futile dose escalation in nearly half of patients while minimizing cardiovascular risk through evidence-based dosing strategies.
