Page 74 - 2021_05-Haematologica-web

A. D’Alessandro et al.
lating with oxidant hemolysis, including glycolytic
metabolites, metabolites involved in the PPP and GSH
homeostasis (Figure 2), purine metabolism and oxida-
tion, arginine and proline metabolism, fatty acids and nated the 33rd and 66th percentiles for this parameter. carboxylic acids (Online Supplementary Figure S2). Based on these thresholds, we generated Receiver
Based on these data, we identified the distribution of oxidative hemolysis across all the 599 samples tested in this study and determined the thresholds that discrimi-
Figure 3. Receiver Operating Characteristic (ROC) curves of metabolic predictors of low and high (33rd and 66th percentile) oxidative hemolysis across the recalled donor population. Specificity, sensitivity and absolute quantitative threshold for each metabolite are noted. AUC: area under the curve.
1294
haematologica | 2021; 106(5)