Background: The variable course of autosomal dominant polycystic kidney disease (ADPKD), and the advent of renoprotective treatment require early risk stratification. We applied urinary metabolomics to explore differences associated with estimated glomerular filtration rate (eGFR; CKD-EPI equation) and future eGFR decline.
Methods: Targeted, quantitative metabolic profiling (1H NMR-spectroscopy) was performed on baseline spot urine samples obtained from 501 patients with ADPKD. The discovery cohort consisted of 338 patients (56% female, median values for age 46 [IQR 38 to 52] years, eGFR 62 [IQR 45 to 85] ml/min/1.73m2, follow-up time 2.5 [range 1 to 3] years, and annual eGFR slope -3.3 [IQR -5.3 to -1.3] ml/min/1.73m2/year). An independent cohort (n = 163) was used for validation. Multivariate modelling and linear regression were used to analyze the associations between urinary metabolites and eGFR, and eGFR decline over time.
Results: Twenty-nine known urinary metabolites were quantified from the spectra using a semi-automatic quantification routine. The model optimization routine resulted in four metabolites that most strongly associated with actual eGFR in the discovery cohort (F = 128.9, P = 7×10-54, R2 = 0.724). A model using the ratio of two other metabolites, urinary alanine/citrate, showed the best association with future annual change in eGFR (F = 51.07, P = 7.26×10-12, R2 = 0.150). This association remained significant after adjustment for clinical risk markers including height-adjusted total kidney volume (htTKV). Results were confirmed in the validation cohort.
Conclusions: Quantitative NMR profiling identified urinary metabolic markers that associated with actual eGFR and future rate of eGFR decline. The urinary alanine/citrate ratio showed additional value beyond conventional risk markers.