Voltammetric electronic tongue combined with chemometric techniques for direct identification of creatinine level in human urine

Urinary creatinine content (UCC) has been investigated as a promising approach for health states monitoring. Most of the techniques used for UCC detection are non-portable and require dilution with other reagents. This study aims to investigate the ability of a Voltammetric Electronic tongue (VE-tongue) combined with chemometrics to discriminate and identify UCC patterns of 59 volunteers without any dilution. The UCC measured by VE-tongue system were related with those found by Jaffe's reaction as a reference method. Loading analysis was carried out to examine the electrodes and features contribution. Principal Component Analysis (PCA) and Support Vector Machines (SVMs) were implemented to analyze VE-tongue data. Furthermore, Partial Least Squares-regression (PLS-regression) was trained to correlate VE-tongue data with the reference measurement in order to build and validate a prediction model for UCC determination. The obtained results demonstrate that PCA and SVMs could classified urine samples into three classes according to low, medium and high creatinine levels (CLs). Promising findings were also obtained in the identification of a new measurement set. Moreover, high correlation coefficients were obtained between the electrochemical technique and spectroscopic Jaffe's method using PLS-regression. This demonstrates a well promising tool for simple alternative process for UCC determination without dilution.