TY - JOUR
T1 - Identification and enumeration of nucleated red blood cells on the coulter LH 750
AU - Chin-Yee, Ian
AU - Brown, Wendy
AU - Johnson, Kathleen
AU - Keeney, Michael
AU - Steele, Bernard
AU - Wolfe, Nancy
AU - Kaplan, Sandra
PY - 2002/12/1
Y1 - 2002/12/1
N2 - Identification of nucleated red blood cells (NRBCs) is an important and challenging aspect of the automated complete blood count. To date, automated methods have focused on either identifying NRBCs on the basis of volume (events hovering at the white blood cell [WBC] threshold) or by employing nucleus-staining dyes. The new Beckman Coulter LH 750 complete blood cell counter uses advances in computer algorithms coupled with impedance, differential volume, conductivity, and light-scatter information to define a "signature position" occupied by NRBCs. In this multicenter study, we compared NRBC and WBC enumeration by the LH 750 with flow cytometry (FCM) and manual slide review. From 3 sites, 314 samples, including 51 cord blood samples, were selected on the basis of either flagging on a predicate analyzer or detection of NRBCs by blood-film review. Correlation determinations (r2) and a signed rank test for difference between medians (P) were performed on each data set. Correlation of the NRBC pooled data from all sites showed an r2 of 0.710 for LH 750 versus FCM. There appeared to be a significant difference between the LH 750 and the FCM procedure (P < .001). Closer examination of the peripheral blood samples (n = 263) revealed that the r2 of the LH 750 versus FCM was 0.781 with no significant difference between the data sets (P = .22). Comparison of cord blood NRBC data (n = 51), showed an r2 of 0.710 for the LH 750 versus FCM, with a statistically significant underestimation of NRBCs by the LH 750 (P < .001). For each data set, similar results were obtained when compared with the manual NRBC method. Comparison of the WBC pooled data from 2 sites in the range of 0.41 to 41.0 × 109/L (n = 162) showed an r2 of 0.978 for LH 750 versus FCM. On a subset of 67 samples, the Abbott Cell-Dyn 4000, which uses a laser-excited nuclear dye to identify NRBCs, was compared with the LH 750. Flow cytometry was used as the comparative procedure. The r2 of the Cell-Dyn 4000 to FCM was 0.631 (P = .36) and the r2 of the LH 750 to FCM was 0.620 (P = .82). Looking specifically at those samples with a diagnosis of sickle cell disease (n = 19), we found the LH 750 results to correlate better to the FCM procedure than the Cell-Dyn 4000 did (r2 = 0.874 and 0.567, respectively). The false positive/false negative rate for the complete data set (n = 67) was 1.5%/16.4% on the Cell-Dyn 4000 and 3.0%/6.0% on the LH 750. The LH 750 represents a significant improvement over previous impedance analyzers in the detection and quantitation of NRBCs. It provides an extremely accurate WBC count in the presence of these cells. When compared with the Cell-Dyn 4000, the LH 750 showed improved sensitivity to the presence of NRBCs. Importantly, these data are provided on the LH 750 at no additional cost with every sample analyzed for a complete blood cell count and differential.
AB - Identification of nucleated red blood cells (NRBCs) is an important and challenging aspect of the automated complete blood count. To date, automated methods have focused on either identifying NRBCs on the basis of volume (events hovering at the white blood cell [WBC] threshold) or by employing nucleus-staining dyes. The new Beckman Coulter LH 750 complete blood cell counter uses advances in computer algorithms coupled with impedance, differential volume, conductivity, and light-scatter information to define a "signature position" occupied by NRBCs. In this multicenter study, we compared NRBC and WBC enumeration by the LH 750 with flow cytometry (FCM) and manual slide review. From 3 sites, 314 samples, including 51 cord blood samples, were selected on the basis of either flagging on a predicate analyzer or detection of NRBCs by blood-film review. Correlation determinations (r2) and a signed rank test for difference between medians (P) were performed on each data set. Correlation of the NRBC pooled data from all sites showed an r2 of 0.710 for LH 750 versus FCM. There appeared to be a significant difference between the LH 750 and the FCM procedure (P < .001). Closer examination of the peripheral blood samples (n = 263) revealed that the r2 of the LH 750 versus FCM was 0.781 with no significant difference between the data sets (P = .22). Comparison of cord blood NRBC data (n = 51), showed an r2 of 0.710 for the LH 750 versus FCM, with a statistically significant underestimation of NRBCs by the LH 750 (P < .001). For each data set, similar results were obtained when compared with the manual NRBC method. Comparison of the WBC pooled data from 2 sites in the range of 0.41 to 41.0 × 109/L (n = 162) showed an r2 of 0.978 for LH 750 versus FCM. On a subset of 67 samples, the Abbott Cell-Dyn 4000, which uses a laser-excited nuclear dye to identify NRBCs, was compared with the LH 750. Flow cytometry was used as the comparative procedure. The r2 of the Cell-Dyn 4000 to FCM was 0.631 (P = .36) and the r2 of the LH 750 to FCM was 0.620 (P = .82). Looking specifically at those samples with a diagnosis of sickle cell disease (n = 19), we found the LH 750 results to correlate better to the FCM procedure than the Cell-Dyn 4000 did (r2 = 0.874 and 0.567, respectively). The false positive/false negative rate for the complete data set (n = 67) was 1.5%/16.4% on the Cell-Dyn 4000 and 3.0%/6.0% on the LH 750. The LH 750 represents a significant improvement over previous impedance analyzers in the detection and quantitation of NRBCs. It provides an extremely accurate WBC count in the presence of these cells. When compared with the Cell-Dyn 4000, the LH 750 showed improved sensitivity to the presence of NRBCs. Importantly, these data are provided on the LH 750 at no additional cost with every sample analyzed for a complete blood cell count and differential.
KW - Cell-Dyn 4000
KW - Coulter LH 750 analyzer
KW - Flow cytometry
KW - NRBC count
KW - WBC count
UR - http://www.scopus.com/inward/record.url?scp=0036965448&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=0036965448&partnerID=8YFLogxK
M3 - Article
AN - SCOPUS:0036965448
VL - 8
SP - 210
EP - 217
JO - Laboratory Hematology
JF - Laboratory Hematology
SN - 1080-2924
IS - 4
ER -