Abstract: SA-PO797
A Metabolomics Approach to CKD Prediction After RCC Nephrectomy
Session Information
- CKD: Epidemiology, Risk Factors, Prevention - III
October 27, 2018 | Location: Exhibit Hall, San Diego Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 1901 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention
Authors
- Jaimes, Edgar A., Memorial Sloan-Kettering Cancer Center, New York, New York, United States
- Ostrovnaya, Irina, Memorial Sloan Kettering Cancer Center, New York, New York, United States
- Martin, Axel S., Memorial Sloan Kettering, New York, New York, United States
- Weiss, Robert H., UC Davis, Nephrology, Davis, California, United States
Background
Patients with RCC who undergo nephrectomy can develop CKD as result of reduction in renal mass and given common risk factors between CKD and RCC it is likely that RCC patients already have parenchymal kidney disease . Thus it is imperative to obtain additional information in order to identify those at higher risk for CKD and/or its progression.
Methods
We hypothesized that there is a metabolomics signature for future renal functional decline in tumor-adjacent “normal” kidney tissue. We utilized a non-targeted metabolomics approach coupled with longitudinal patient data from post-operative visits to identify metabolites in non-malignant renal tissue linked to future eGFR decline. We used two statistical approaches: (1) univariate analysis adjusting for age, BMI, and nephrectomy type, and (2) cross-validated penalized regression (elastic net) methodology. Metabolomic analysis was performed by Metabolon (Durham, NC) using non-targeted metabolomic gas and liquid chromatography coupled to a mass spectrometry approach.
Results
We studied 138 samples from nephrectomies performed at MSKCC. 79 were radical (57.2 %), 100 subjects were male (72.5%), 122 were white (88.4%), average age was 61.9 years and 23 were diabetic (16.7 %). 476 metabolites were included in our analysis. Using univariate linear regression, we found that the most significant associations with eGFR slope were by the metabolites N6-trimethyl-lysine, , guanosine and tyrosyl-leucine, as well as two unidentified metabolites, X13871 and X15168, with p<=0.012, although none reached significance when adjusted for multiple comparisons using FDR. We also created composite metabolite score utilizing all of the metabolites using elastic net modelling fitted on cross-validation samples. This score was univariately associated with the eGFR slope with p<0.001 and remained significant after adjusting for clinical variables. A similarly developed score based on tumor tissue metabolites was not significantly associated with the slope (p=0.18).
Conclusion
In this study we identified metabolites found in the adjacent non-malignant kidney tissue that are associated with eGFR decline post-nephrectomy. Prospective evaluation of surgical specimens using a targeted metabolomics approach to validate our findings will lead to a new paradigm by which RCC patients can be evaluated, and potentially treated, for unrecognized and incipient CKD.
Funding
- NIDDK Support