Drug-induced liver injury (DILI) encompasses a spectrum of clinical disease ranging from mild biochemical abnormalities to acute liver failure. The majority of adverse liver reactions are idiosyncratic, occurring in most instances 5-90 days after the causative medication was last taken. The diagnosis of DILI is clinical, based on history, probability of the suspect medication as a cause of liver injury and exclusion of other hepatic disease. DILI can be hepatocellular (predominant rise in alanine transaminase), cholestatic (predominant rise in alkaline phosphatase) or mixed liver injury. An elevated bilirubin level more than twice the upper limit of normal in patients with hepatocellular liver injury implies severe DILI, with a mortality of approximately 10% and with an incidence rate of 0.7-1.3 per 100,000. Although acute liver failure is rare, 13-17% of all acute liver failure cases are attributed to idiosyncratic drug reactions. Response to drug withdrawal may be delayed up to 1 year with cholestatic liver injury with occasional subsequent progressive cholestasis known as the vanishing bile duct syndrome. Overall, chronic disease may occur in up to 6% even if the offending drug is withdrawn. Antibiotics and NSAIDs are the most common cause of DILI. Statins rarely cause significant liver injury whereas antiretroviral therapy is associated with hepatotoxicity in 10% of treated patients. Multiple mechanisms of DILI have been implicated, including TNF-alpha-activated apoptosis, inhibition of mitochondrial function and neoantigen formation. Risk factors for DILI include age, sex and genetic polymorphisms of drug-metabolising enzymes such as cytochrome P450. In patients with human immunodeficiency virus, the presence of chronic viral hepatitis increases the risk of antiretroviral therapy hepatotoxicity. Over the next decade, the combination of accurate case ascertainment of DILI via clinical networks and the application of genomics and proteomics will hopefully lead to accurate prediction of risk of DILI, so that pharmacotherapy can be optimised with avoidance of adverse hepatic events.