Mortality Estimation in Renal Disease (MERD Score): A Model Predicting Waitlist Mortality in Kidney Transplant Candidates.

BACKGROUND

Predicting waitlist mortality is important for prioritizing organ allocation and selecting candidates for extended criteria donors. Currently, there is no widely adopted and reliable index for predicting early mortality among kidney transplant candidates. In this study, we aim to develop an index score utilizing variables from the OPTN database to predict mortality among adult kidney transplant candidates within 3 years of being on the waitlist.

METHODS

This study utilized data from 147,307 adult kidney transplant candidates listed in the OPTN database from 2018 to 2023. The cohort was randomly divided into training and validation groups. Sixteen variables were analyzed using univariate logistic regression, with significant factors incorporated into a multivariable analysis to develop the MERD (Mortality Estimation in Renal Disease) score. Predictive performance was assessed through ROC analysis in both cohorts.

RESULTS

Ten variables, age, ABO blood type, ethnicity, dialysis duration, presence of peripheral vascular disease, albumin level, functional status, Previous kidney malignancy, primary etiologies of kidney disease, and insurance type were identified as significant predictors and used to formulate the MERD score. The AUC was 0.6657 in the training cohort and 0.6580 in the validation cohort.

CONCLUSION

The MERD score provides proof of concept for short-term mortality prediction for kidney transplant waitlist candidates. Further prospective validation and model refinement are warranted.