One key issue in the pharmaceutical development of new compounds is knowledge on metabolism, the enzymes involved and the potential hepatotoxicity of a drug. Primary cultured hepatocytes are a valuable in vitro model for drug metabolism studies. However, human hepatocytes show phenotypic instability and have restricted accessibility and high batch-to-batch functional variability, which seriously complicates their use in routine testing. Therefore, several liver-derived cell models have been developed for drug metabolism and hepatotoxicity screening to circumvent these drawbacks. Hepatoma cell lines offer important advantages, availability, an unlimited life span and a stable phenotype, thus rendering them suitable models for such studies. However, currently available human hepatoma cell lines are not a good alternative to cultured hepatocytes as they show very limited expression for most drug-metabolising enzymes. Other approaches have been developed to generate immortalised hepatic cells with metabolic competence (use of plasmids encoding immortalising genes to transform human hepatocytes, cell lines obtained from transgenic animals, hepatocytomes or hydrid cells). Recombinant models heterologously expressing cytochrome P450 enzymes in hepatoma cells have also been generated, and are widely used in drug metabolism and toxicity evaluations. In recent years, new approaches to up-regulate the expression of drug-biotransformation enzymes in human cell lines (i.e., transfection with the expression vectors encoding key hepatic transcription factors) have also been investigated. This paper reviews the features of liver-derived cell lines, their suitability for drug metabolism and hepatotoxicity studies, and the state-of-the-art strategies pursued to generate metabolically competent hepatic cell lines.