RESUMO
Data were obtained from four state-of-the-art automated hematology analyzers and compared with those obtained from microscopic reference procedures. The instruments evaluated were the Technicon's H*1, Sysmex's NE-8000, Coulter's STKR, and Coulter's STKS. Accuracy was assessed by comparing machine-generated white blood cell and red blood cell profiles with those obtained manually by experienced laboratorians. Specimens used were actual clinical samples submitted for routine analysis. The precision of the instruments in counting and sizing blood cells was not significantly different at the clinical decision-making level and consistently exceeded that of the microscopic method. Significant shifts in the leukocyte population were detected with relatively similar sensitivity by all instruments. As expected, the oldest model's clinical efficiency was exceeded by that of the newer analyzers. None of the analyzers performed with an accuracy that permits the laboratory to completely eliminate a microscopic scan of a stained blood film obtained from a patient's initial specimen.