Cell transplantation is emerging as a new treatment designed to improve the poor outcome of patients with cardiac failure. Its rationale is that implantation of contractile cells into postinfarction scars could functionally rejuvenate these areas. Primarily for practical reasons, autologous skeletal myoblasts have been the first to be considered for a clinical use. A large number of experimental studies have consistently documented a robust engraftment of myoblasts, their in-scar differentiation into myotubes, and an associated improvement in left ventricular function. The early results of phase I clinical trials have then established both the feasibility and safety of this procedure with the caveat of arrhythmic events. Efficacy data are equally encouraging but definitely need to be validated by large prospective placebo-controlled, double-blind randomized trials such as the Myoblast Autologous Grafting in Ischemic Cardiomyopathy (MAGIC) study, the results of which are now pending. In addition to assessing the effect of myoblast transplantation on regional and global heart function, these results will also provide comprehensive safety data and thus allow a more objective assessment of the risk-benefit ratio. However, it is already apparent that the outcome of myoblast transfer could most likely be improved by optimizing the purity of the cell yield (by selecting muscle-derived progenitors less lineage-committed than the myoblasts), the mode of delivery (by increasing the accuracy of cell injections while decreasing their invasiveness), and the survival of the engrafted cells (by concomitant graft vascularization and incorporation of cells in three-dimensional matrices). Most, if not all, of these changes will have to be incorporated before skeletal myoblasts can acquire the status of therapeutic agents. Furthermore, there is increasing evidence that myoblasts may act by attenuating left ventricular remodeling or paracrinally affecting the surrounding myocardium but not by generating new cardiomyocytes because of their strict commitment to a myogenic lineage. Thus, improvement of function is not tantamount of myocardial regeneration, and if such a regeneration remains the primary objective, it is worth considering alternate cell types able to generate new cardiac cells that will be electromechanically coupled with the host cardiomyocytes. In the setting of this second generation of cells, human cardiac-specified embryonic stem cells may hold the greatest promise. © 2007 Elsevier Inc. All rights reserved.