The comet assay has become a popular method for the assessment of DNA damage in biomonitoring studies and genetic toxicology. However, few studies have addressed the issue of the noted inter-laboratory variability of DNA damage measured by the comet assay. In this study, 12 laboratories analysed the level of DNA damage in monocyte-derived THP-1 cells by either visual classification or computer-aided image analysis of pre-made slides, coded cryopreserved samples of cells and reference standard cells (calibration curve samples). The reference standard samples were irradiated with ionizing radiation (0-10 Gy) and used to construct a calibration curve to calculate the number of lesions per 10(6) base pair. All laboratories detected dose-response relationships in the coded samples irradiated with ionizing radiation (1.5-7 Gy), but there were overt differences in the level of DNA damage reported by the different laboratories as evidenced by an inter-laboratory coefficient of variation (CV) of 47%. Adjustment of the primary comet assay end points by a calibration curve prepared in each laboratory reduced the CV to 28%, a statistically significant reduction (P <0.05, Levene's test). A large fraction of the inter-laboratory variation originated from differences in image analysis, whereas the intra-laboratory variation was considerably smaller than the variation between laboratories. In summary, adjustment of primary comet assay results by reference standards reduces inter-laboratory variation in the level of DNA damage measured by the alkaline version of the comet assay.