The belief that n-3 polyunsaturated fatty acids are in general cytotoxic to tumor cells appears not to be accurate. Of four tumor cell lines exposed to 35 microM docosahexaenoic acid (DHA, 22:6 n-3), we found only one (A-427, lung carcinoma) to be sensitive, whereas three (A-172, A549 and SK-LU-1) in fact were stimulated. A 6-fold higher level of lipid peroxidation in A549 as compared with A-427 cells indicates that cytotoxicity is not determined by the overall level of lipid peroxidation. Moreover, paracetamol (0.1, 0.3 and 1.5 mM), which is known to have both pro- and antioxidant activity, counteracted the cytotoxic effect of DHA on A-427 cells in a dose-dependent manner by a mechanism that does not involve inhibition of overall lipid peroxidation. Although paracetamol (0.1 and 0.3 mM) in the absence of DHA was able to enhance proliferation of all tumor cell lines 1.1-1.4-fold, this was insufficient to explain the ability of the drug to protect against DHA-induced cytotoxicity. Neither did paracetamol cause major changes to the activity of the defense enzyme glutathione peroxidase, known to play a role in the sensitivity of A-427 cells to DHA. Paracetamol could possibly act by reacting with minor, highly toxic, peroxidation products, or alternatively, by altering the substrate for lipid peroxidation, i.e. the fatty acid composition of the membranes, in favor of less toxic products.