No abstract

Background: A novel method to calculate the 4 °C storage time of red cell concentrates is described to which is based on the analysis of ³¹P nuclear magnetic resonance (³¹P-NMR) spectra or red cell parameters by artificial neural nets. Material and Methods: 142 red cell concentrates, resuspended in SAGM, were used for this study. ³¹P-NMR spectra were obtained by a Bruker AMX 300 system from samples, originating from storage days 0, 7, 14, 28, 21, 35, and 42. Parallel to the ³¹P-NMR spectra, the red cell parameters hematocrit (Hct), mean cell volume (MCV), mean cell hemoglobin concentration (MCHC), free hemoglobin (Hb_e), and K⁺ were determined. Results: Using artificial neural nets, the 4 °C storage time can be determined with an accuracy of ±2.5 days from the ³¹P-NMR spectrum and of about ±4.2 days from the red cell parameters (Hct, MCV, MCHC, Hb_e, K⁺). Conclusion: Because of its simplicity and compatibility with existing devices, artificial neural nets might become a useful tool for quality control of red cell concentrates. ³¹P-NMR spectroscopy could be done noninvasively. Another advantage of neural nets is the possibility to determine additional quality parameters.