EMR-Based Tool Predicts Graft Failure After Kidney Transplant

Although a successful transplant is expected to last about a decade, nearly one in four kidney grafts fail within five years, often without early warning signs. Clinicians currently lack reliable ways to identify which patients are most at risk of losing a transplanted kidney. A new study now shows that routinely collected kidney function data can be used to dynamically predict graft failure risk years before it happens.

Researchers at Johns Hopkins Medicine (Baltimore, MD, USA) have developed an electronic medical record (EMR)–based prediction model that continuously updates a patient’s risk of graft failure using changes in estimated glomerular filtration rate, a standard blood test that reflects how well a transplanted kidney is functioning. The approach allows risk to be recalculated every time new lab results become available.

The model was built using data from 1,114 deceased-donor kidney transplant recipients and validated across multiple large datasets, including national transplant registries and real-world hospital records. Investigators analyzed repeated kidney function measurements over time and linked these trajectories to graft failure, defined as a return to dialysis or the need for a second transplant within three years.

The results, published in the Clinical Journal of the American Society of Nephrology, showed that the model could effectively distinguish high-risk from low-risk patients. Predictive accuracy reached 0.70 just three months after transplant and improved to 0.90 by 30 months, demonstrating strong performance well before clinical failure occurred.

By identifying patients at elevated risk, the tool could support earlier interventions such as closer monitoring, adjustments to immunosuppressive therapy, and counseling about future transplant needs. At the same time, patients at low risk could safely transition back to routine nephrology care closer to home, freeing transplant center resources.

Researchers now plan to test the model in everyday clinical practice and expand it by incorporating additional health data, including clinical events and other laboratory markers. These efforts aim to further improve early detection of graft injury and long-term transplant outcomes.

“The results from this study can be readily implemented at transplant centers to streamline care of kidney transplant recipients by providing updated risk predictions as new data become available,” said Chirag Parikh, MD, PhD, the study’s senior author. “There could also be future downstream models developed to predict other infections and immunologic complications.”

Related Links:
Johns Hopkins Medicine