Restenosis after percutaneous coronary balloon angioplasty remains a significant problem. Despite success with a variety of agents in animal models, no agent has proved clearly successful in reducing restenosis in humans. There are many potential reasons for this, but one possibility is that because of our incomplete understanding of the restenotic process, therapy has been directed at the wrong target. Arterial remodeling (changes in total vessel area or changes in area circumscribed by the internal elastic lamina) is well described in de novo atherosclerosis, and there is increasing evidence that this process occurs after angioplasty. Thus, restenosis can be thought of not merely as neointimal formation in response to balloon injury, but as arterial remodeling in response to balloon injury and neointimal formation. Arterial remodeling may consist of actual constriction of the artery, as has been described in some animal models and in preliminary fashion in humans, or of compensatory enlargement as has been described in de novo atherosclerosis and in the hypercholesterolemic rabbit iliac artery model. Arterial constriction can result in restenosis with minimal neointimal formation. Compensatory enlargement accommodates significant amounts of neointimal formation, with preservation of lumen area despite an increase in neointimal area adequate to cause restenosis in a noncompensated artery. This expanded paradigm of arterial remodeling and intimal formation may in part account for the lack of success in clinical trials to date, and therapy directed at arterial remodeling as well as intimal formation may be required to reduce restenosis after coronary interventions.