Journal
SRPSKI ARHIV ZA CELOKUPNO LEKARSTVO
Volume 148, Issue 11-12, Pages 706-710Publisher
SRPSKO LEKARSKO DRUSTVO
DOI: 10.2298/SARH200215090M
Keywords
hemodialysis; machine learning; spent dialysate; VIS-NIR; patient-specific
Categories
Ask authors/readers for more resources
Introduction/Objective Diabetic nephropathy leading to end-stage renal disease is a major health problem worldwide. Hemodialysis (HD) treatment is associated with glycemia variations. Diabetic patients on HD might benefit from a non-invasive online glycemia monitoring system. The aim of this study was to assess the glucose concentration from the matrix of the spent dialysate fluid using near-infrared (NIR) spectroscopy. Methods Blood samples and spent dialysate were collected in the 15th minute of the HD treatment from 15 patients. The spent dialysis fluid was characterized by a NIR spectrometer in the range of 900-1300 nm. In order to apply the artificial neural network (ANN) and train it, the MATLAB NFTOOL program was used. The testing and training of the ANN were executed using the MR spectrum of the spent dialysis fluid as input, and the glucose concentration as output. Results A significant correlation in excess of 93% between the MR spectrum of the spent dialysate and the blood glucose concentration (3-9 mmol/l) was found. Conclusions NIR spectroscopy is a non-invasive and reliable method of glycemia monitoring which can be used in maintaining HD patients.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available