Comparison of performance for leukocyte differential counting of the Technicon H6000 system with a manual reference method using the NCCLS standard.

The National Committee for Clinical Laboratory Standards (NCCLS) has published a tentative standard for leukocyte differential counting, by means of which a manual or automated method for leukocyte differential counting can be compared with a manual reference method. The performance of the Technicon H6000 system was evaluated using the standard at Stamford and Overlook Hospitals. A total of 502 patient samples were analyzed: 315 from Overlook and 187 from Stamford. The H6000 system was found to be approximately four times more precise than the 200-cell manual reference method for each cell type. Correlation of the H6000 system with the manual method was good, with correlation coefficients of 0.98 for neutrophils and lymphocytes, 0.96 for eosinophils, 0.72 for monocytes, and 0.5 for basophils. The clinical sensitivity of the H6000 system, measured in terms of false normals and false abnormals, was similar to that of the manual reference method when measured against itself. There were no clinically significant discrepancies in results from the H6000 system, except for possibly one case where a patient was already on antibiotic therapy. The NCCLS standard was found to be a useful but rather complex and involved method for evaluating the performance of the H6000 system, the major problem being the amount of work needed to count manually the number of cells required for the manual reference method.