Biomarkers of chronic kidney disease in the urine of diabetic/hypertensive patients by means of Raman spectroscopy

摘要

Diabetes mellitus (DM) and arterial hypertension (AH) are common diseases that, if untreated, predispose the patient to renal failure. This study aimed to evaluate possible biomarkers in the urine of patients with DM and AH capable to predict the chronic renal disease, by means of Raman spectroscopy. Urines were obtained from patients with DM and AH, and separated into four groups: no symptoms of diseases related to DM and AH (G1), with low clinical complications (G2), with severe clinical complications (G3), and with chronic kidney disease (G4) arised from DM and AH. It has been used a dispersive Raman spectrometer (830nm, 250mW, 20s accumulation). In the spectra of urine it was identified Raman peaks at 680cm~(-1) (creatinine), 1004cm~(-1) (urea) and 1128cm~(-1) (glucose). The results revealed that G2, G3 and G4 presented the creatinine peak with lower intensity than G1 (p ＜ 0.05). It was observed that G2, G3 and G4 showed lower intensity of the urea peak compared to G1 (p ＜ 0.05) and G4 showed lower intensity compared to G2 and G3 (p ＜ 0.05). Despite not significant, the glucose peak showed lower intensity in G1 when compared to the other groups. A model for classification of groups according to clinical criteria, using Sparse Multinomial Logistic Regression, taking as inputs the intensities of creatine, urea and glucose peaks allowed correct classification of 88.9% for G1, 36.8% for G2, 43.8% for G3 and 84.2% for G4. These results demonstrated the possibility of obtaining diagnostic information for complications of kidney disease associated to DM and AH, particularly the renal failure.