Results Samples were obtained from 22 children aged from eight months to 16-4 years, median 5-5 years, and 10 adults aged between 26 and 44 years. Data from adults and children were combined for analysis. The numbers of the different cell types were higher in skin puncture blood (4-8-6-4% higher for subsets other than CD3+ CD16+ CD56+ natural killer cells, which were 10-7% higher), the difference being statistically significant for all cell types except CD3+ CD8+ T lymphocytes (table). Significance was lost for all pairings when total lymphocytes and subpopulations were expressed in proportional terms (data not shown). Paired values were highly correlated (r=0 90 for CD3 + CD16 + CD56 + natural killer cells, and 0-96-0.98 for total white cells, lymphocytes and other subsets, p<00001 for all pairings). The differences between paired CD3 + CD4 + T cell counts is shown in the figure.
DiscussionCollection of blood by skin puncture provides a rapid and simple altemative to venepuncture when small volumes of blood are required, and is useful when dealing with small children and subjects with difficult veins. In this study no undue difficulty was found in obtaining the required volume ofskin puncture blood; should problems arise it may be possible to reduce test volumes by 25-50%, at least when the total lymphocyte count is within normal limits. Previous studies have shown clinically acceptable comparability in white cell and differential counts between the two samples types,2 3 although in the earlier study neutrophil values were significantly higher in skin puncture blood. Similar results were obtained in the present study, the usually slightly higher total white cell count of skin puncture samples being reflected proportionally throughout the range of lymphocyte subsets. Although of borderline statistical significance for most subpopulations, the magnitude of the difference was small and unlikely to be of clinical importance.The precision of both replicate haemoglobin and platelet estimations from skin puncture samples has been reported to be poorer than that from venous blood,45 although the latter claim could not be verified using the same sample collection method as that of the present study.2 While we did not count lymphocytes in sequential paired samples there appears to be no reason why their numbers should fluctuate to a greater degree in skin puncture blood, a contention supported by the highly significant correlations found for each lymphocyte subpopulation between the two sample types.Our results indicate that skin puncture blood can be safely used as an alternative to venous blood for the enumeration of CD4 and other lymphocyte subsets. It would be prudent when monitoring cell numbers over time, however, to use one sample type alone in order to minimise the variables affecting the results obtained.1 Calvelli T, Denny TN, Paxton H, Galman R, Kagan J.Guidelines for flow cytometric immunophenotyping: a report from the National Institute of Allergy and Infectious Diseases, Division of AIDS. Cytometry 1993;14:...