BACKGROUND: An international multicenter study, involving six sites in Nort
h America and six sites in Europe, was undertaken to assess the performance
of standardized methods for the enumeration of CD34+ cells in peripheral b
lood over the dynamic range from 200 cells per mu L to zero. Two commercial
ly available techniques were studied, a flow cytometry method and a microvo
lume fluorimetry method.
STUDY DESIGN AND METHODS: Coded samples were centrally prepared and sent to
test sites by overnight mail. Samples included internal replicates, linear
dilutions, and specimens at the lower limit of detection. In addition, com
mercially available reagent positive control cells were sent to a subset of
laboratories.
RESULTS: Over the sample range studied, the intersite precision among diffe
rent laboratories was good with coefficients of variation ranging from 14 p
ercent to 24 percent for microvolume fluorimetry and from 20 percent to 31
percent for flow cytometry. Intrasite precision ranged from 7 percent to 21
percent. Test linearity was excellent with sites demonstrating a mean r(2)
= 0.992 for microvolume fluorimetry and r(2) = 0.984 for flow cytometry. T
he lower limit of detection was 5 CD34+ cells per mu L for both commercial
assays. Over the range of 5 to 50 CD34+ cells per mu L, the microvolume flu
orimetry assay reported slightly higher values than the flow cytometry assa
y. Preliminary analysis of reagent positive control cells showed very good
precision and accuracy.
CONCLUSIONS: Standardization of CD34+ cells enumeration is improving and co
mmercially available assays provide accurate and precise methods. More inve
stigation of reagent positive control cells is warranted.