There has been no cogent hypothesis to explain the molecular basis of analgesic and non-steroidal anti-inflammatory drug (NSAID) associated renal papillary necrosis (RPN) and upper urothelial carcinoma (UUC). The microsomal cytochrome P-450 enzyme system may generate reactive intermediates which promote pathophysiological effects in the lung, liver and renal cortex, but the absence of P-450 activity in the medulla suggests that it is unlikely that similar events lead to RPN and UUC. Other enzymes (eg. peroxidases) convert substituted aromatics into benzoquinoneimines (an intermediate that has previously been defined in P-450-mediated toxicity). The medulla is rich in fatty acid peroxidases involved in the metabolism of arachidonic acid. NSAID and analgesics interact with key enzymes in this pathway, which could lead to the co-oxygenation of exogenous and endogenous compounds via the peroxidase, lipoxygenase, or prostaglandin hydroperoxidase enzymes. The generation of reactive molecules in the medulla could explain both RPN and UUC via the alkylation of macromolecules. The formation of free radicals would give rise to extensive lipid peroxidation, (there are large quantities of free polyunsaturated fatty acids in the medullary interstitial cells), an event of major potential importance to local cell destruction and genotoxic effects. At present this proposed mechanism of co-oxygenation offers the most attractive working hypothesis to explain the molecular pathogenesis of both RPN and UUC.