A survey of the results of haematological parameters, using a common rat blood sample in Japanese laboratories

A survey was conducted to determine the accuracy and quality control of automated haematology analysers used in non-clinical toxicity studies. Pooled blood samples from male Sprague-Dawley rats were distributed to 98 laboratory facilities throughout Japan, the samples being delivered under refrigeration to each facility within 18 h of sample preparation. At each facility, the samples were analysed within 4 h of receipt. Commercially available normal human blood samples from a single lot were also analysed at the same time. Most haematological results were within the mean ± 3SD (standard deviation), but some facilities gave either high or low values consistently for both human and rat samples. No facility gave high or low values to certain parameters sporadically, which suggests no problem with the accuracy of the equipment. However, it was suspected that there would be some problem in comparing analytical values determined in a unique way by specific equipment design. The use of certain equipment resulted, in rat haematocrit values in particular, being either too high or too low. In these cases, it was deemed necessary to make some adjustments or calibration changes. There were also platelet values with a 'plus drift' which was apparently due to contamination with, or failure to identify small red blood cells (RBC). There was no deviation in values which could be attributed solely to the mechanical operation of any of the analytical equipment. Non-standard, initial setting up of the equipment (originally intended for human use, but now used for a variety of animal species) has been recognised as the main cause for a wider range of the analytical values seen. The results of this survey suggest that it may be necessary to review equipment calibration at each facility, and to re-establish the historical background data.