ZAP-70 is a useful prognostic marker in chronic lymphocytic leukaemia, expression showing concordance with unmutated IgVH genes and poor prognosis. Measuring ZAP-70 by flow cytometry is attractive since this is commonly used in CLL diagnosis, but no standard method has yet emerged. Most studies use one of 2 antibodies, the unconjugated 2F3.2 clone (Upstate) and the 1E7.2 clone, conjugated to Alexafluor 488 (Caltag), but several different gating strategies are described. In 72 CLL samples with known IgVH gene mutational status we measured ZAP-70 by flow cytometry using these 2 antibodies and 3 different analysis strategies, and compared these results with the methylation status of the gene at Cytosine +334 (Parker et al, Hematologica, 2005). Effects of sample age and anticoagulant were also assessed. Results were analysed by 3 observers using 1) a quadrant marker on the isotype control to measure percentage ZAP-70 positivity in the tumour population 2) the ratio of geometric means of ZAP-70 to isotype control in the tumour population and 3) measurement of the percentage of the tumour population with ZAP-70 intensity equal to that of sample T cells. Cut-offs for ZAP-70 positivity for analysis methods 1) and 2) and both antibodies were determined by concordance with IgVH gene mutational status. For method 1) these were 10% for the Upstate antibody and 15% for the Caltag antibody; for method 2) the ratios were 1.3 for the Upstate antibody and 1.5 for the Caltag antibody. Interobserver variability was much greater for analysis method 1), with variations in ZAP-70 percentage of >3% in 27% of cases, than for method 2), where the ratios differed by >0.1 in only 5% of cases. Considering discordance between ZAP-70 and IgVH gene mutational status, for the 27 unmutated cases (98–100% germline homology), 2 cases were ZAP-70 negative by both antibodies, 1 case was negative by the Upstate antibody but positive by the Caltag antibody, and 4 cases were negative by the Caltag antibody but positive by the Upstate antibody. Of 9 intermediate cases (96–97.9%homology), 2 were ZAP-70 positive and 7 were ZAP-70 negative by both antibodies. Of the highly mutated (<96% homology) cases, 1 case was positive by the Caltag antibody and 1 different case was positive by the Upstate antibody. Overall, of 8 cases showing discordance between the Upstate and the Caltag antibodies by methods 1) or 2), ZAP-70 gene methylation status concorded with the Upstate result in 7 (88%)and with IgVH gene mutational status in 7. Use of analysis method 3), gating on T cells, changed the ZAP-70 status in 16 cases(22%) using the Upstate antibody and 12 cases(17%) using the Caltag antibody. The change resulted in discordance with IgVH gene status in the majority (86%). 12 CLL cases were studied using both heparin and EDTA anticoagulants, and assayed on days 0,1, and 2 after collection. Day 0 heparin samples gave higher values than EDTA. In 8 EDTA samples, ZAP-70 showed a transient rise on day 1 while in 9 heparin samples there was a slow decline over time. In 3 cases (2 EDTA, 1 heparin) ZAP-70 status changed with sample age. We conclude 1) Analysis by ratio of geometric means, test: isotype control, gives least interobserver variability. 2) Good concordance with IgVH gene mutational status is seen with 1E7.2 or 2F3.2 ZAP-70 clones but more discordant cases are seen with the 1E7.2 clone. 3) Further studies required to resolve issues of sample age and anticoagulant choice 4) Methylation status of the ZAP-70 gene may be useful in resolving issues of discordance.