RESUMO
In metabolomics, nuclear magnetic resonance (NMR) spectroscopy allows to identify and quantify compounds in biological samples. The sample preparation generally requires only few steps; however, an indispensable factor is the addition of a locking substance into the biofluid sample, such as deuterium oxide (D2O). While creatinine loss in pure D2O is well-described, the effects of different D2O concentrations on the signal profile of biological samples are unknown. In this work, we investigated the effect of D2O levels in the NMR buffer system in urine samples, in dependence on dwell time and temperature exposition. We reveal a decrease of the urinary creatinine peak area up to 35% after 24 h of dwell time at room temperature (RT) using 25% (v/v) D2O, but only 4% loss using 2.5% D2O. 1H, inverse-gated (IG) 13C, DEPT-HSQC NMR, and mass spectrometry (MS) experiments confirmed a proton-deuterium (H/D) exchange at the CH2. This leads to underestimation of creatinine levels and has an extensive effect when creatinine is used for normalization. This work offers a sample stability examination, depending on the D2O concentration, dwell time, and temperature and enables a method to correct for the successive loss. We propose an equation to correct the creatinine loss for samples prepared with various D2O concentrations and storage temperatures for dwell times up to 24 h. The correction function was validated against an external data set with n = 26 samples. To ensure sufficient creatinine stability in future studies, we suggest that a maximum of 10% D2O should be used at 4 °C or 2.5% D2O at RT, respectively.
