Drug metabolism studies, together with pathologic and histologic evaluation, provide critical data sets to help understand mechanisms underlying drug-related hepatotoxicity. A common practice is to trace morphologic changes resulting from liver injury back to perturbation of biochemical processes and to identify drug metabolites that affect those processes as possible culprits. This strategy can be illustrated in efforts of elucidating the cause of acetaminophen-, troglitazone- and valproic acid-induced hepatic necrosis, microvesicular steatosis and cholestasis with the aid of information from qualitative and quantitative analysis of metabolites. From a pharmaceutical research perspective, metabolite profiling represents an important function because a structure-activity relationship is essential to rational drug design. In addition, drugs are known to induce idiosyncratic hepatotoxicity, which usually escapes the detection by preclinical safety assessment and clinical trials. This issue is addressed, at present, by eliminating those molecules that are prone to metabolic bioactivation, based on the concept that formation of electrophilic metabolites triggers covalent protein modification and subsequent organ toxicity. Although pragmatic, such an approach has its limitations as a linear correlation does not exist between toxicity and the extent of bioactivation. It may be possible in the future that the advance of proteomics, metabonomics and genomics would pave the way leading to personalized medication in which beneficial effect of a drug is maximized, whereas toxicity risk is minimized.