Currently, drugs have been synthesised that can significantly delay the course of several viral infections, including those provoked by HBV, HCV or HIV, but that display consistent side effects, including toxicity for organelles such as mitochondria. Several in vitro models and techniques have been developed to analyse the effects of such compounds. HepG2 cells (from human hepatoma) are an excellent model to investigate mitochondrial (mt) toxicity because of their high content of organelles and mtDNA, and actually different investigators are indeed using such cells. Studies in vitro on cell lines are relatively easy, but it is necessary to be careful in the interpretation of data, which are usually obtained on continuously growing, tumour cells, quite different from normal, resting, non-neoplastic cells collected from a patient. Direct analysis of drug-induced mt damage in patients is extremely more complex than that performed using in vitro models because of the difficulty to obtain adequate cells or to have discrete amounts of biological material, the status of the patient at the moment of cell collection, the use of an adequate assay and its correct execution, and finally the possibility to find sex- and age-matched healthy controls as source of reference parameters.