Speaker
Description
Identification of chronic kidney disease patients at risk of progressing to end-stage renal disease (ESRD) is essential for treatment decision-making and clinical trial design. Here, we show that proton nuclear magnetic resonance (NMR) spectroscopy of blood plasma specimens together with techniques from machine learning improves the currently best performing kidney failure risk equation, the so-called Tangri score. Our NMR study cohort comprises 4640 participants from the German Chronic Kidney Disease (GCKD) study, of whom 185 (3.99%) progressed over a mean observation time of 3.70 ± 0.88 years to ESRD requiring either dialysis or transplantation. In this context, we also introduce mixed graphical models to reveal important associations between NMR derived metabolic features, demographic and drug features and variables measured by standard clinical chemistry. Results show important associations between chronic kidney disease and for example gout. In summary, we demonstrate that NMR substantially improves the analysis of chronic kidney disease.