{"id":"e81cc95daa37","type":"article","url":"https://hartvaat.nl/2026/06/26/machine-learning-voorspelt-diureticumrespons-bij-acuut-hartfalen-drc-ahf/","title":"Machine learning voorspelt diureticumrespons bij acuut hartfalen — DRC-AHF","title_en":"A Machine-Learning Model for Accurate Diuresis Prediction in Acute Heart Failure (DRC-AHF).","category":"hartfalen","category_label":"Hartfalen","professions":["cardioloog","internist"],"tags":[],"journal":"ESC heart failure","doi":"10.1093/eschf/xvag185","source_url":"https://doi.org/10.1093/eschf/xvag185","authors":["Gracjan Iwanek","Mateusz Guzik","Amadeusz Oleszczak","Iga Karbowiak","Dominik Marciniak","Matteo Pagnesi","Julio Nuñez","Robert Zymliński","Piotr Ponikowski","Jan Biegus"],"significance":5,"published":"2026-08-06","source_date":"2026-06-26","image":"","kennis":["https://hartvaat.nl/kennis/hartfalen/acuut-hartfalen/","https://hartvaat.nl/kennis/hartfalen/hartfalen-stadiumindeling-nyha/"],"congress":"","summary_en":"A prospective single-centre study developed a machine learning model to predict 6-hour urine output after intravenous furosemide in acute heart failure, using eGFR and spot urine sodium/creatinine measured two hours post-dose. In an external validation cohort (n=50), the model showed strong correlation with observed output (r=0.90) and correctly classified diuresis response in 88% of patients, outperforming the standard NRPE equation (58%). While the tool offers a promising approach to personalizing diuretic therapy in diuretic-resistant AHF, its single-centre derivation and small validation sample warrant multicentre external validation before routine clinical adoption.","created":"2026-07-30T01:35:07Z","updated":"2026-08-10T10:36:49Z","licence":"Citeer vrij, met bronvermelding en een link naar hartvaat.nl (de url van het record). Samenvattingen zijn redactioneel werk van HartVaat; de oorspronkelijke publicaties blijven van hun uitgevers (doi). Geen medisch advies.","body_markdown":"Een prospectieve single-centre studie ontwikkelde een machine learning-model dat de urineproductie na 6 uur intraveneuze furosemide bij acuut hartfalen nauwkeurig voorspelt op basis van eGFR en urine-elektrolyten twee uur na toediening. In het validatiecohort (n=50) correleerde de voorspelde met de werkelijke urineproductie sterk (r=0,90) en classificeerde het model 88% van de patiënten correct in diurese-categorieën, tegen 58% voor de bestaande NRPE-formule. Het model biedt een potentieel nuttig hulpmiddel voor het individualiseren van diureticumtherapie bij diureticumresistentie, hoewel de single-centre opzet en kleine validatiegroep nog beperkingen vormen voor directe klinische implementatie.","abstract_original":"AIMS: We aimed to develop a machine learning-based tool for accurate quantitative prediction of diuretic response in acute heart failure (AHF). METHODS: We prospectively enrolled 296 AHF patients (20% female, mean age 68 ± 13 years, mean eGFR 62 ± 30 ml/min/1.73m2, median NT-proBNP 7112 [4082-14400] pg/ml) in a single-centre derivation/validation study. A Random Forest regression model was developed using derivation (n=50) and optimization (n=246) cohorts and externally validated in an independent cohort (n=50; mean age 69 ± 14 years, mean eGFR 56 ± 27 ml/min/1.73m2). Predictors were eGFR, spot urine sodium (uNa), and urine creatinine (uCr) obtained 2 hours after intravenous furosemide. RESULTS: In the validation cohort (n=50), the median observed 6-hour urine output was 1220 [950-1750] ml vs. model-estimated 1446 [957-1749] ml (p=0.480). Observed and estimated values showed a strong positive correlation (r=0.90, 95% CI: 0.83-0.94, p<0.0001), with a mean absolute error (MAE) of 413 ml and a mean bias of 67 ± 383 ml on Bland-Altman analysis. The model correctly classified 44/50 patients (88%) into predefined diuresis categories (≤900 ml, 900-1800 ml, ≥1800 ml), compared with 58% for the reference Natriuretic Response Prediction Equation (NRPE). MAE decreased progressively by approximately 200 ml per 100 additional patients, confirming data-driven performance improvement. CONCLUSIONS: A three-variable machine learning model (eGFR, uNa, uCr at 2 hours post-diuretic) predicted 6-hour urine output with high accuracy in AHF and demonstrated adaptive improvement with expanding training data. The calculator is freely available at https://diuresis.umw.edu.pl."}