ctnebe at t-online.de
Thu Nov 28 17:09:00 EST 1996
in extension of the discussion I want to emphasize that the determination of
absolute counts add to the list of errors.
Cord blood contains more nucleated red cells and it depends on your hematology
system whether you count them or not. Leukapheresis concentrates contain in
addition cell debris where you can argue about counting thresholds. At least
the thresholds should be the same in your hematology and fluorescence flow
Background also depends on the nature of your sample. We found it very useful
to counterstain CD34 (PE) with CD3,16,19 (FITC). The CD34 progenitor cells form
a clear cluster in the FITC vs PE dot plot where as the unspecific debris forms
a diagonal. The backgating shows the clear and typical position in the FCS/SSC
dot plot. The percentage of debris is often in the same range as the true CD34
Theoretically the bead count method should be ideal because it does not combine
the errors of two instruments. However the practical experience with the
commercial product was less satisfying (no good cluster, sticky tubes).
The lysing method influences the result too. FACS lyse reduces the FSC distance
between debris and leukocytes compared to ammonium chloride. The latter however
shows more often incomplete lyse and nucleated reds.
CD45 as the third colour reagent reduces this problem. It excludes erythroblasts
and giant platelets from the analysis. LDS751 or other DNA stains do not
penetrate monocytes well, stain nucleated reds and in part giant platelets and
are therefore unreliable.
I have seen consensus protocols on CD34 but they are either rather general or
cannot substantiate some of their recommendations. However, in good blood
samples a good agreement was seen by Stefan Serke in the INSTAND EQA in Germany
over 35 labs.
Good EQA schemes with problem samples will hopefully help on this issue.
Thomas Nebe, ctnebe at t-online.de
Klinikum Mannheim, D-68135 Mannheim
More information about the Cytometry