NMR is a central tool in the field of metabolomics, thanks to its ability to provide valuable structural and quantitative information with high precision. Most NMR-based metabolomics studies rely on 1D 1H detection, which is heavily limited by strong peak overlap. 13C NMR benefits from a wider spectral dispersion and narrower signal line width but is barely used in metabolomics due to its low sensitivity. Dissolution dynamic nuclear polarization (d-DNP) offers an opportunity to improve 13C NMR sensitivity by several orders of magnitude. Here, we show that this emerging hyperpolarized metabolomics approach can provide meaningful information about clinical samples. Achieving sub-mM limits of detection with 13C at natural abundance in urine samples was made possible by a meticulous design of the experimental workflow. The analysis of human urine samples from patients with different stages of chronic kidney disease (CKD) was performed using 13C d-DNP NMR and benchmarked to conventional 1H NMR metabolomics at a high magnetic field to explore the complementarity between the two methods. Multivariate analysis of the d-DNP 13C NMR dataset provided a statistical model able to distinguish patients with CKD from control patients. Moreover, 13C d-DNP NMR spectra highlighted several biomarkers known to be biologically relevant. Some of them were in agreement with those obtained with conventional 1H NMR, and the results also highlighted the complementarity of biomarker coverage between hyperpolarized and conventional NMR metabolomics. In particular, 13C hyperpolarized NMR allowed the annotation of two biomarkers that could not be detected by 1H NMR because of peak overlap (i.e., guanine and guanidoacetate).