Fig 1.
Cell differential description for the four hematology analyzers (median with interquartile (first to third quartile) range (IQR) with whiskers up to largest observation that falls within 1.5 times the IQR and down to smallest observation falling within this distance.
The observations outside the whiskers boundary are plotted as outliers) A. Platelet count (PLT), B. Neutrophils, C. Lymphocytes, D. Eosinophils, E. Monocytes, and F. Basophils. The platelet count was significantly higher in the CELL-DYN Sapphire platform than in Siemens ADVIA 2120i and Beckman Coulter DxH900 (p = 0.027 and p = 0.002, respectively). Lymphocytes were significantly higher in the Sapphire and XN-1000V platforms vs. the 2120i (p = 0.006 and p = 0.024, respectively), and basophils were significantly lower in the XN-1000V platform vs. DxH900 (p = 0.038). Otherwise, there were no other statistically significant differences between 3-hour baseline cell counts among the 4 hematology analyzers tested.
Fig 2.
Ability of the four hematology analyzers to give stable cell differential assessments up to 72 hours in whole blood samples kept at 4°C (LS means±95%CI per platform.
Baseline Mean is the average of the 2120i, DxH900, and XN-1000V platforms, n = 18). Results of the linear mixed models (differences in LS means versus 3-hour 20°C baseline) for each cell by time point and temperature are presented in S4 Table. A. Platelets, B. Neutrophils, C. Lymphocytes, D. Eosinophils, E. Monocytes, and F. Basophils.
Fig 3.
Ability of the 4 hematology analyzers to give stable cell differential assessments up to 72 hours in whole blood samples kept at 20°C (LS means±95%CI per platform.
Baseline Mean is the average of the 2120i, DxH900, and XN-1000V platforms, n = 18). Results of the linear mixed models (differences in LS means versus 3-hour 20°C baseline) for each cell by time point and temperature are presented in S4 Table. A. Platelets, B. Neutrophils, C. Lymphocytes, D. Eosinophils, E. Monocytes, and F. Basophils.
Fig 4.
Ability of the 4 hematology analyzers to give stable cell differential assessments up to 72 hours in whole blood samples kept at 30°C (LS means±95%CI per platform.
Baseline Mean is the average of the 2120i, DxH900 and XN-1000V platforms, n = 18). Results of the linear mixed models (differences in LS means versus 3-hour 20°C baseline) for each cell by time point and temperature are presented in S4 Table. A. Platelets, B. Neutrophils, C. Lymphocytes, D. Eosinophils, E. Monocytes, and F. Basophils.
Fig 5.
Comparison of absolute eosinophil count for those with eosinophilic asthma (dashed line), healthy (with atopy (dashed/dotted line) and without atopy (solid line) participants by the 4 hematology analyzers by time and temperatures.
A. Samples stored at 4°C, B. stored at 20°C, C. stored at 30°C and D. stored at 37°C. The baseline is shown in the 20°C (B) plot and mean eosinophil counts for those with eosinophilic asthma ranged between 0.552–0.579 cells x109 /L, healthy participants with atopy at 0.190–0.232 cells x109/L, and healthy without atopy at 0.106–0.142 cells x109/L, depending on analyzer (S6 Table). Larger absolute drops are observed in the Sapphire and 2120i eosinophil counts for participants with eosinophilic asthma after 24h at 20°C and 30°C for both analyzers and also at 37°C for the Sapphire; moving them into the healthy range. More extensive variability (outliers) is observed at higher temperatures for the XN-1000V after 48h. No platform maintains reliability to measure eosinophil count across all storage times and temperatures.
Fig 6.
Activation/maturation markers measured on eosinophils using flow cytometry show activation and death with time and temperature (LS Means±95%CI, n = 12 (6 with eosinophilic asthma, 3+3 healthy with and without atopy)) A. live/dead aqua staining, B. CD62L (l-selectin), C. CD66b, D. CD123, E. FcɛR1, and F. CD11b. Results of the linear mixed models (differences in LS means versus 3-hour and 20°C baseline) for each marker by time point and temperature are presented in S10 Table. Live-dead aqua, CD123 and FcɛR1 increased dramatically at 30°C by 72h. CD66b and CD11b increased at 4°C and decreased at 30°C over time. CD62L decreased over time at both 20°C and 30°C.
Fig 7.
Eosinophil-derived neurotoxin (EDN) in plasma.
A. EDN release with time and temperature, data merged from healthy (n = 12) and asthmatic patients (n = 6), shaken, and non-shaken samples. B. EDN normalized to eosinophil count per donor at baseline. Each data point represents geometric mean±SD, n = 12 per time point 6-72h (shaken and non-shaken), n = 4–6 at 3h.