Differential Diagnosis of Glycosuria Using Raman Spectroscopy

The aim of this research was to detect spectral differences in glycemic components. Urine samples were collected from 40 patients who were divided into a control group and a diabetic and hypertensive group. The samples were obtained in the morning, fasting, and stored a freezer at − 80 °C until spectral analysis. Spectral data collection was performed using a dispersive Raman spectrometer (Dimension P-1 model, Lambda Solutions, Inc., MA, USA). The equipment uses a stabilized multimode diode laser operating at 830 nm, with about 300 mW power output, and time integration to collect the Raman signal was adjusted to 5 s. The mean Raman spectra displaced from the urine of patients in the study groups (CT and DM & HBP) were identified at the range of 516 and 1127 cm⁻¹. Comparative analysis of mean urine spectra showed a significant difference (p < 0.05) between the groups, the Student's t-test was used to compare the mean Raman spectra of the groups. The comparative analysis of peak intensities at 516 and 1127 cm⁻¹ in the urine of diabetic control and hypertensive patients revealed that it was higher in the DM & HBP group than in the CT group, however, with no significant difference (p > 0.05). To quantify the glucose in urine and discriminate the groups, a model was developed to estimate the concentration using a quantitative regression model based on partial least squares (PLS). According to the data obtained, there was an excellent correlation (r = 0.98) between the concentrations estimated by the model and the concentrations determined by colorimetric analysis. Discriminant analysis (DA) based on a regression model (PLS) proved to be promising as it discriminated the control group without errors, and the rate in the DM & HBP group was 89.1%. Raman spectroscopy can be a potentially useful tool for testing glucose in urine.