In a number of adverse drug reactions leading to hepatotoxicity drug metabolism is thought to be involved by generation of reactive metabolites from nontoxic drugs. In this study, an in vitro assay was developed for measurement of the impact of metabolic activation of compound on the cytotoxicity toward a human hepatic cell line. HepG2 cells were treated for 6 h with compound in the presence or absence of rat liver S9-mix, and the viability was measured using the MTT test. The cytotoxicity of cyclophosphamide was substantially increased by S9-mix in the presence of NADPH. Three NADPH sources were tested: NADPH (1 mmol/L) or NADPH regenerating system with either NADP+/glucose 6-phosphate (G6P) or NADP+/isocitrate. All three NADPH sources increased the cytotoxicity of cyclophosphamide to a similar extent. Eight test compounds known to cause hepatotoxicity were tested. For these, only the cytotoxicity of diclofenac was increased by S9 enzymes when an NADPH regenerating system was used. The increased toxicity was NADPH dependent. Reactive drug metabolites of diclofenac, formed by NADPH-dependent metabolism, were identified by LC-MS. Furthermore, an increase in toxicity, not related to enzymatic activity but to G6P, was observed for diclofenac and minocycline. Tacrine and amodiaquine displayed decreased toxicity with S9-mix, and carbamazepine, phenytoin, bromfenac and troglitazone were nontoxic at all tested concentrations, with or without S9-mix. The results show that this method, with measurement of the cytotoxicity of a compound in the presence of an extracellular metabolizing system, may be useful in the study of cytotoxicity of drug metabolites.