Kidney Donor Risk Prediction

About this Application:
This app predicts your risk of hypertension, type 2 diabetes, reduced eGFR, and proteinuria following kidney donation. Additional descriptive statistics on End Stage Renal Disease (ESRD) are included. To get started, please click on the Risk Models tab and enter characteristics at the time of evaluation. All inputs must be filled.

Notes:
The data used to create this app is from the University of Minnesota Living Kidney Donor Program from 1963 through 2017 with a median follow-up time of 22.8 years (range: 0.8-56.4 years). The average age at donation was 39, 43.2% of donors were male, and 94.2% of donors were White. Predictions of hypertension and type 2 diabetes risk may not be valid for non-White donors.

Exclusion criteria included donors with hypertension at donation for hypertension models, donors who donated a partial pancreas for type II diabetes models, and non-white donors for proteinuria and eGFR models.

Key References:

Technical Details:
The accuracy of random forest models and Cox proportional models (both with and without competing risks) in predicting hypertension and diabetes were compared as part of S Vempati's Master's Thesis. The model with the highest predictive accuracy (a random forest model) and the Cox model most similar to the previously published model are presented in the application.

Disclaimer

This app was created as a reference, but does not constitute medical advice. Please consult your healthcare provider for more information.

Author Contributions

Models and content based on work conducted by the authors of Ibrahim et al. (2016), Sanchez et al. (2018), and Ibrahim et al. (2010)
Initial application created by Vempati S for a Master's Thesis
Application was revised and finalized by Northrop EF with assistance from Helgeson E

All applications are created and maintained by the transplant surgery research group comprised of

Matas A, Chinnakotla S, Helgeson E, Vock D, Evans M, and Northrop EF

Kidney Donor Risk Prediction

Risk Models

Select a Hypertension Model:

About the Model:

Select a Type II Diabetes Model:

About the Model:

This plot was created using the same method as used by Ibrahim, et al., 2016. Non-white observations were excluded from this analysis.

for age, BMI, diastolic blood presssure, systolic blood pressure, eGFR at donation, relationship to recipient and if recipient has type II diabetes. eGFR < 45 was adjusted for age, BMI, systolic blood pressure, eGFR at donation, and relationship to recipient. eGFR < 30 was adjusted age, BMI, and systolic blood pressure.

Risk values are presented as cumulative risk in 5-year intervals. For each outcome, Cox regression with the full dataset after excluding non-white donors was used to organize observations into 20 equal-size groups based on the percentile of each donor's predicted survival. Subsequent Cox models were fit for each group and cumulative survival probabilities were recorded at 5-year intervals. For each new data point, predicted survival is estimated by the full-data Cox model and used to classify the new observation into one of the 20 groups. The cumulative survival for the selected group is then displayed on the graph above.

About the Data

The plot below displays the cumulative incidence of ESRD cases over time. This plot is unadjusted, meaning it does not depend on the inputs in the sidebar.

Of the 4297 living donors, 43 developed ESRD after a median time of 34 years. The average (sd) age at donation for these individuals was 35.6 (13.0) years and average eGFR was 91.7 (17.3); 63% were male, 88% white. Of the donors that developed ESRD: 37 donated to a first degree relative, 4 donated to other relative, and 2 were unrelated to the recipient.

Kidney Donor Risk Prediction

Disclaimer

Kidney Donor Risk Prediction

Risk Models

Enter Values at Evaluation:

Donor Characteristics

Recipient Characteristics

About the Model:

About the Model:

About the Model:

About the Data