Related Experiment Video
Updated: Jan 18, 2026

Digital Home-Monitoring of Patients after Kidney Transplantation: The MACCS Platform
Published on: April 12, 2021
Predicting Simultaneous Heart Kidney Allocation and Posttransplant Adverse Kidney Outcomes
Mutlu Mete1, Mehmet U S Ayvaci2, Ahmet B Gungor3
1Department of Information Science, University of North Texas, Denton, Texas, USA.
Predicting kidney recovery after heart transplant is crucial for simultaneous heart-kidney transplantation (SHKT) decisions. A machine learning model shows promise in identifying patients at risk for adverse kidney outcomes post-heart transplantation (HT).
Area of Science:
- Cardiology
- Nephrology
- Transplantation Medicine
- Machine Learning in Healthcare
Background:
- Simultaneous heart-kidney transplantation (SHKT) offers improved survival for patients with end-stage heart failure and kidney disease.
- Predicting kidney recovery post-heart transplantation (HT) is challenging, complicating SHKT decision-making.
- Accurate prognostication is essential for optimizing treatment strategies in combined organ failure.
Purpose of the Study:
- To develop and validate a machine learning model to predict adverse kidney outcomes within one year following heart transplantation (HT).
- To assist in the clinical decision-making process for simultaneous heart-kidney transplantation (SHKT) in patients with moderate kidney failure.
- To identify HT recipients who may benefit from or require SHKT based on predicted kidney function.
Main Methods:
- A retrospective cohort of adult HT recipients in the US (October 2018 - December 2020) was analyzed using Organ Procurement and Transplantation Network (OPTN) data.
- A random forest (RF) machine learning algorithm was developed using 15 high-importance variables to predict a composite adverse kidney outcome within one year post-HT.
- Adverse outcomes included need for SHKT, end-stage kidney disease requiring dialysis, severely reduced GFR, or retransplant listing. Model validation was performed internally and externally.
Main Results:
- Of 6579 HT recipients, 13.4% experienced adverse kidney outcomes or received SHKT within one year.
- The RF model achieved high specificity (0.941-0.955) and negative predictive value (0.940-0.955).
- Moderate sensitivity (0.605-0.694) and positive predictive value (0.604-0.680) were observed, with strong class differentiation (c-statistics 0.849-0.899).
Conclusions:
- The developed RF model demonstrates utility in predicting adverse kidney outcomes after heart transplantation.
- This predictive tool can supplement clinical judgment in the complex decision-making process for SHKT.
- Further refinement may enhance its role in patient selection for combined heart and kidney transplantation.
Related Concept Videos
Kidney Transplant I: Introduction
Kidney Transplant II: Surgical Procedure
Kidney Transplant III: Nursing Management
Acute Kidney Injury IV: Diagnostic Studies and Prevention
Acute Kidney Injury V: Interprofessional Care
Drug Dosing in Renal Diseases: Dose Adjustments Based on Drug Clearance and Elimination Rate Constant

