Erythroblasts count by ADVIA 2120i in patients with thalassemia and sickle cell disease and in newborns.

Background. The presence of nucleated red blood cells (NRBC) in peripheral blood of patients with beta-thalassemia or sickle-cell disease in newborns, reveals abnormal activity in bone marrow erythropoiesis. While it is related to transfusional needs in patients with beta-thalassemia or sickle-cell disease, in the latter case instead it is strictly related to mortality index and perinatal complications. This study aims to evaluate NRBC counting performances of the Siemens automated analyzer ADVIA2120i in the above-mentioned three categories of subjects, as they have in common a potential alteration in their red blood cells (RBC) lysis sensitivity and a need for accurate NRBC count. The authors have also defined and verified a validation rule of NRBC count in automation useful for the screening of positive samples while improving the appropriateness of microscopic revisions. The rule has been defined on the performance of ADVIA2120i NRBC standard counts and partial counts (PEROX analysis and BASO analysis channels), hemoglobin (Hb) and the ratio of microcytic hypochromic red blood cells (M/I).Methods. In total, 187 samples of peripheral blood collected in K3EDTA-coated tubes (50 normal controls, 26 patients with sickle-cell disease, 55 patients with beta-thalassemia and 56 newborns) were analyzed by ADVIA2120i. This NRBC ADVIA2120i standard count and their different instrumental parameters (%HISTO, %RESID, %GAUSS, ºROX) were compared to the same performed by visual microscopy reference method (MO). Agreement between the average value of the two methods was estimated by means of Pearson’s correlation, Passing-Bablok regression and bias analysis, whereas diagnostic accuracy was estimated through ROC curves analysis. The proposed screening rule has been defined according to: NRCB counts standard method and evaluation of analytical performances of different instrumental parameters (%HISTO, %RESID, %GAUSS, ºROX), Hb value and M/I.Results. The comparison between the two described methods provided the following results: Pearson’s correlation: r=0.98 (with CI 95% from 0.97 to 0.98 p<0.0001); Passing-Bablok regression: y=1.20x+0.0 (with CI 95% from 0.95 to 1.55 and from 0.0 to 0.0 respectively); bias −0.907 (CI 95% from −2.47 to 0.66). NRBC ADVIA2120i standard count in all the samples shows an area under curve (AUC) of 0.80, in newborns and sickle-cell disease patients AUC is 0.66, while in thalassemic patients AUC is 0.73. The ºROX index shows better diagnostic performance than ADVIA2120i standard NRBC count in the group of thalassemic and sickle-cell disease patients, with an AUC of 0.83 and 0.76 respectively. The proposed screening rule shows a significative match with the microscopic examination (Pearson’s Chi-squared test, p<0.001).Conclusions. In the three categories of subjects which were examined, we suggest that the NRBC automated count by ADVIA2120i should be confirmed every time by means of visual microscopy, especially in the case of patients with hemoglobinopathy. However, such diagnostic instrumental performance can be improved with the use of more complex algorithms which also should be considered: calculation parameters (ºROX), Hb value and M/I index.