The heme oxygenase/biliverdin reductase (HO/BVR) axis catalyzes the degradation of heme, but this system and its by-products, carbon monoxide (CO) and bilirubin, have also been shown to exert cytoprotective effects by activating pro-survival pathways and scavenging free radicals. Naturally occurring substances that upregulate the inducible isoform of HO (HO-1) have therefore been proposed as potential new drugs for the treatment of free radical-induced disease. A number of existing drugs have also been shown to regulate the HO/BVR system, and this capacity is considered an additional mechanism for their therapeutic activity. However, upregulation of the HO/BVR axis is not always beneficial for cells: the heme depletion and accumulation of CO and bilirubin it causes are potentially toxic. Therefore, new pharmacological modulators of HO/BVR activity must act in a dose-dependent manner. This would allow dose titration to achieve a desired pharmacologic effect without producing toxicity. Unfortunately, this goal is more complicated than it seems because toxicity has to be defined in terms of each of the main products of heme metabolism. Furthermore, sensitivity to the therapeutic/toxic effects of these products is likely to be tissue- or cell-type specific. The solution may lie in the use of novel drug-delivery systems that allow targeted delivery of low doses of the HO/BVR modulator to selected tissues.