[Accuracy, precision and clinically acceptable level of complete blood cell count by an automated multichannel hematology analyzer].

Measurement of complete blood cell count (CBC) is one of the essential laboratory tests. In this study, the accuracy of CBC was high, when measured by six different automated blood cell counters, including SE-9000/RAM-1, CELL-DYN 4000, ADVIA 120, VEGA RETIC-LC141 or GEN*S. The correlations of CBC counts among these instruments were also good. In contrast, the accuracy in abnormal samples depended on matching the instrument and type of clinical conditions. Internal quality control of automated multichannel hematology analyzers was recommended by the NCCLS H26-A in 1996. However, actual external quality control of the CBC count was poorly understood. We surveyed three samples from healthy volunteers for CBC values in 1997. Among six different instruments, the inter-assay of the red blood cell count (RBC), hemoglobin, hematocrit and MCV was fairly good, shown as < 4.2%, < 3.0%, < 4.4% and < 3.4%, respectively. In contrast, the inter-assay of the white blood cell count (WBC) and platelet (PLT) was not good, shown as < 11.4% and < 9.6%, respectively. The clinically acceptable levels for the blood cell count were reported to be 4% for RBC, 3% for hemoglobin, 4% for MCV, 5% for WBC and 7% for PLT, by JCCLS in 1994. The clinically acceptable CBC by hematologists at Keio University resembled those by councilors of The Japanese Society of Clinical Hematology, whereas, residents at Keio University and general physicians required more precise clinical CBC counts. These results indicate that a larger study is needed to clarify the accuracy, clinically acceptable level, and performance of different automated blood cell counters.