Abstract: TH-PO734
Prognostic Risk Score for Kidney Disease Progression in African Americans Without Type 2 Diabetes
Session Information
- Diversity and Equity in Kidney Health - I
November 03, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
Abstract Time: 10:00 AM - 12:00 PM
Category: Diversity and Equity in Kidney Health
- 800 Diversity and Equity in Kidney Health
Authors
- Nadkarni, Girish N., Icahn School of Medicine at Mount Sinai, New York, New York, United States
- Takale, Dipti, Persistent Systems Ltd, Pune, Maharashtra, India
- Stapleton, Sharon, Renalytix, New York, New York, United States
- Fleming, Fergus, Renalytix, New York, New York, United States
- Coca, Steven G., Icahn School of Medicine at Mount Sinai, New York, New York, United States
Background
African Americans (AAs), even without type 2 diabetes have a higher incidence of adverse kidney outcomes, compared to other populations. This is driven in part by genetic (APOL1 risk genotype) and multiple non-genetic factors. We sought to develop a composite prognostic risk score combining clinical data elements and biomarkers using machine learning for identification of AAs with non-diabetic CKD at highest risk of progression.
Methods
We conducted an observational study in a biobank that linked banked plasma samples and genetic information to longitudinal electronic health record (EHR) data in AAs with impaired kidney function. We measured plasma levels of soluble tumor necrosis factor receptor (sTNFR)1, sTNFR2, and kidney injury molecule-1 (KIM1) using previously validated assays. We then trained a random forest model using a 75:25 split for predicting a composite outcome of eGFR decline of ≥5 ml/min per year, ≥40% sustained decline in eGFR, or kidney failure within 5 years.
Results
In 472 AAs with non-diabetic CKD, the median age was 62 years, 62% were female, the baseline eGFR was 66 ml/min/1.73 m2, and 14% had the APOL1 risk genotype. Over 5 years, 7% experienced the composite endpoint. The composite risk score had an AUC of 0.78 (95% CI 0.72, 0.86). On applying a risk cutoff that considered 10% of the population as high-risk, the positive predictive value (PPV) for the outcome was 51%, and the NPV was 91% in the bottom 90% of the population. In a sub-analysis of AAs with the APOL1 risk genotype, the PPV and NPVs were also 50% and 91%, respectively. The hazard ratio (HR) was 31 (95% CI 15 to 68) for the top decile of risk score vs. the rest of the population.
Conclusion
A composite risk score accurately risk-stratified AAs with and without APOL1 risk genotype for kidney outcomes. With further validation, this is a valuable tool for population health management, clinical trial enrichment, and ensuring health equity.
Funding
- Commercial Support – Renalytix