Résumé : Objective: Stem cell therapy can facilitate cardiac repair in infarcted myocardium, but the optimal cell type remains uncertain. We conducted a randomized, blind, and placebo-controlled comparison of autologous bone marrow mononuclear cell and mesenchymal stem cell therapy in a large-animal model of chronic myocardial infarction. Methods: Eleven weeks after coronary ligation, 24 dogs received intramyocardial injections of mononuclear cells (227.106 ± 32.106 cells), mesenchymal stem cells (232.106 ± 40.106 cells), or placebo (n = 8 per group). Cardiac performance and remodeling were assessed up to 16 weeks' follow-up. Results: At echocardiographic analysis, the wall motion score index showed a sustained improvement after mononuclear cell transfer (from 1.8 ± 0.1 to 1.5 ± 0.07) and a moderate late improvement after mesenchymal stem cell transfer (from 1.9 ± 0.08 to 1.7 ± 0.1). After mononuclear cell transfer, end-systolic elastance increased (from 2.23 ± 0.25 to 4.42 ± 0.55 mm Hg/mL), infarct size decreased (from 13% ± 0.67% to 10% ± 1.17%), N-terminal B-type natriuretic propeptide level decreased (from 608 ± 146 to 353 ± 118 pmol/L), and relative wall area and arterial density increased. Vascular endothelial growth factor receptor 2 expression was upregulated in the border zone. No change in cardiac contractility or histologic parameters was noted in the mesenchymal stem cell group. Conclusion: In a canine model of chronic myocardial infarction, bone marrow mononuclear cell transfer is superior to mesenchymal stem cell transfer in improvement of cardiac contractility and regional systolic function and reduction in infarct size and plasma N-terminal B-type natriuretic propeptide level. Functional improvement is associated with a favorable angiogenic environment and neovascularization. © 2009 The American Association for Thoracic Surgery.