CAIXETA, DOUGLAS C.LIMA, CASSIOXU, YUNGUEVARA-VEGA, MARCOESPINDOLA, FOUED S.GOODACRE, ROYSTONZEZELL, DENISE M.SABINO-SILVA, ROBINSON2023-04-262023-04-262023CAIXETA, DOUGLAS C.; LIMA, CASSIO; XU, YUN; GUEVARA-VEGA, MARCO; ESPINDOLA, FOUED S.; GOODACRE, ROYSTON; ZEZELL, DENISE M.; SABINO-SILVA, ROBINSON. Monitoring glucose levels in urine using FTIR spectroscopy combined with univariate and multivariate statistical methods. <b>Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy</b>, v. 290, p. 1-9, 2023. DOI: <a href="https://dx.doi.org/10.1016/j.saa.2022.122259">10.1016/j.saa.2022.122259</a>. Disponível em: http://repositorio.ipen.br/handle/123456789/34004.1386-1425http://repositorio.ipen.br/handle/123456789/34004The development of novel platforms for non-invasive continuous glucose monitoring applied in the screening and monitoring of diabetes is crucial to improve diabetes surveillance systems. Attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy of urine can be an alternative as a sustainable, label-free, fast, non-invasive, and highly sensitive analysis to detect changes in urine promoted by diabetes and insulin treatment. In this study, we used ATR-FTIR to evaluate the urinary components of non-diabetic (ND), diabetic (D), and diabetic insulin-treated (D + I) rats. As expected, insulin treatment was capable to revert changes in glycemia, 24-h urine collection volume, urine creatinine, urea, and glucose excretion promoted by diabetes. Several differences in the urine spectra of ND, D, and D + I were observed, with urea, creatinine, and glucose analytes being related to these changes. Principal components analysis (PCA) scores plots allowed for the discrimination of ND and D + I from D with an accuracy of ∼ 99 %. The PCA loadings associated with PC1 confirmed the importance of urea and glucose vibrational modes for this discrimination. Univariate analysis of second derivative spectra showed a high correlation (r: 0.865, p < 0.0001) between the height of 1074 cm-1 vibrational mode with urinary glucose concentration. In order to estimate the amount of glucose present in the infrared spectra from urine, multivariate curve resolution-alternating least square (MCR-ALS) was applied and a higher predicted concentration of glucose in the urine was observed with a correlation of 78.9 % compared to urinary glucose concentration assessed using enzyme assays. In summary, ATR-FTIR combined with univariate and multivariate chemometric analyses provides an innovative, non-invasive, and sustainable approach to diabetes surveillance.1-9openAccessdiabetes mellitusfourier transform spectrometersinfrared spectraurinebiological markersquantitative chemical analysisglucoseArtigo de periódico29010.1016/j.saa.2022.1222590000-0001-7404-9606https://orcid.org/0000-0001-7404-960689.085.75