RESUMEN
BACKGROUND: Implementation of universal WBC reduction of blood components means that automated analytical methods may be the only satisfactory way for production laboratories to meet increased testing requirements. STUDY DESIGN AND METHODS: A multicenter study on the performance of a microvolume fluorimeter (IMAGN 2000, Becton Dickinson) was undertaken on 519 RBC, 353 platelet, and 27 fresh plasma units. RESULTS: WBC counts for the RBC samples ranged from 0.02 to 6.94 x 10(6) per unit (mean, 0.57) as determined by FC and from 0.02 to 5.53 x 10(6) per unit (mean, 0.40) as determined by MVF with a mean FC bias of +0.15 x 10(6) WBCs per unit, and discrepancies outside the 95% limits of agreement were mainly associated with higher FC counts. The series of platelet samples showed means of 0.90 (range, 0.06-19.45) and 0.66 (range, 0.01-18.95) x 10(6) WBCs per unit for FC and MVF methods, respectively. FC and MVF results were in good agreement at low counts, although significant discrepancies were noted at higher counts. Overall, for the platelet units, there was a mean FC bias of +0.34 x 10(6) WBCs per unit. The intermethod agreement exceeded 99 percent for both types of blood component when the single (both UK and United States) decision point of 5.0 x 10(6) WBCs per unit was applied. The mean WBC counts for the 27 analyzed fresh plasma units were 61.8, 56.0, and 46.0 per microL by Nageotte hemocytometry, FC, and MVF, respectively. CONCLUSIONS: This evaluation found that the level of intersite consistency for FC was relatively poor compared to that for MVF. The results nevertheless validated the broad equivalence of FC and MVF results for the current Council of Europe and UK/US decision points of <1.0 and <5.0 x 10(6) WBCs per unit.