SummaryA spontaneously immortalized human keratinocyte line, HaCaT, was used as anin vitromodel to predict the cutaneous irritation of anionic surfactants. For this purpose, a number of sodium salts of N‐alkyl sulphates with hydrocarbon chain lengths varying between C8and C16were studied for possible cytotoxic effects. The endpoints used to assess toxicity were uptake of the vital dye neutral red (NR) and cell morphology criteria 24 h after dosing. A linear proportionality between keratinocyte number and NR uptake was established. All tested surfactants had cytotoxic effects as demonstrated by a decreased NR uptake, which showed a clear dose–response relationship. Concentrations resulting in 50% inhibition of NR uptake (IC‐50) ranged from 0·15 mmol (sodium lauryl sulphate, C12) to 1·23 mmol (sodium octyl sulphate, C8). Thein vitrocytotoxicity data were highly reproducible when the test was repeated after several weeks. The cytotoxicity data from these assays were compared with the irritant responses (as evaluated by measurement of erythema and transepidermal water loss) obtained after 24 h application of the same compounds (300 μ1 of 20 mmol aqueous solution) to the volar forearm of human volunteers. There were significant linear correlations between the IC‐50 values and both barrier damage (transepidermal water loss) and erythema (as evaluated by skin colour reflectance measurements). For the test substances, however, the sensitivity of thein vitrosystem was between 10 and 100 times higher than that observed in human skinin vivo.This quantitative difference can largely be ascribed to the effective permeability barrier of normal human skinin vivo,which protects living keratinocytes from the cytotoxic effects of surfactant molecules.The results indicate that normal human keratinocytes in culture are a promising screening method for predicting the irritation potential of anionic surfactants. Confirmation, however, has still to be obtained by appropriatein vivotesting in human