For years, methods have been available for the predictive identification of chemicals that possess the intrinsic potential to cause skin sensitization. However, many have proven less suitable for the determination of relative sensitizing potency. In this respect, the local lymph node assay (LLNA) has been shown to have a number of important advantages. Through interpolation of LLNA dose-response data, the concentration of a chemical required to produce a threshold positive response (a 3-fold increase in activity compared with concurrent vehicle controls, the EC3 value) can be measured. The robustness of this parameter has been demonstrated rigorously in terms of inter- and intralaboratory reproducibility. Additionally, the relationship between potency estimates from the LLNA and an appreciation of human potency based on clinical experience has been reported previously. In the present investigations, we have sought to consolidate further our understanding of the association between EC3 values and human skin-sensitization potency by undertaking a thorough and extensive analysis of existing human predictive assays, particularly where dose-response information is available, from historical human repeated insult patch tests (HRIPTs). From these human data, information on the approximate threshold for the induction of skin sensitization in the HRIPT was determined for 26 skin-sensitizing chemicals. These data were then compared with LLNA-derived EC3 values. The results from each assay, expressed as dose per unit area (microg/cm(2)), revealed a clear linear relationship between the 2 values, thereby substantiating further the utility of LLNA EC3 values for prediction of the relative human sensitizing potency of newly identified skin sensitizers.