Gd(III) toxicity is such that, for in vivo use, the ion must be incorporated into stable, ionically bound complexes with organic chelating agents. Although metal-chelate-based contrast agents have been demonstrated safe in phase I-III studies, their potential toxicities must be considered, particularly in the development of new contrast media. The potential for adverse effects is thought to arise from in vivo dissociation into constituent metal ion and ligand. Release of Gd(III) from ligand may also be facilitated by endogenous metals [e.g., Zn(II)] that compete with Gd(III) for ligand-binding sites. In vivo complex stability toward dissociation depends on several closely balanced characteristics of the metal and ligand that maintain metal-chelate integrity during residence in the body. Among design strategies explored to minimize the potential for in vivo Gd(III) release are the use of ligands that: form Gd(III) complexes with high thermodynamic stabilities; have binding selectivity for Gd(III) over endogenous metal ions; incorporate rigid cyclic structures and form Gd(III) complexes with greater kinetic stabilities than their open-chain homologs; or incorporate slight excess ligand to "scavenge" loosely bound or free endogenous metal ions that might otherwise compete with Gd(III) for ligand binding.