The present report contains the final results of a Phase I study that evaluated the feasibility, safety, and potential efficacy of intramyocardial injection of autologous bone marrow (BM) in "no-option" patients with refractory angina and myocardial ischemia. Twenty-seven patients underwent electromechanic mapping-guided transendomyocardial injections (n = 12, 0.2 ml each) of unfractionated autologous BM cells directed to ischemic, noninfarcted myocardial territory. Patients were injected with 28 +/- 27 x 10(6)/ml nucleated cells containing 2.2 +/- 1.4% CD34+ cells. The autologous BM injection procedure was successful in all patients and was associated with no adverse events. At 3 months, the Canadian Cardiovascular Society angina score (3.2 +/- 0.5 vs 2.0 +/- 0.91, p = 0.001) and treadmill exercise duration (418 +/- 136 vs 489 +/- 142 seconds, p = 0.017) had improved significantly. The stress-induced ischemia score within the injected territories (118 segments) had also improved (2.2 +/- 0.8 vs 1.7 +/- 1.1, p < 0.001). At 1 year, the clinical improvement was sustained, although 5 patients had undergone revascularization procedures. The number of total injected nucleated cells (CD45+), progenitor cells (CD34+), and the magnitude of secreted vascular endothelial growth factor and macrophage chemoattractant protein-1 by cultured BM cells failed to predict the clinical response. In conclusion, the 3- and 12-month study results have indicated the safety of catheter-based transendocardial delivery of autologous BM cells in patients with advanced symptomatic ischemic heart disease and may suggest sustained potential efficacy. The cellular and humeral characteristics of autologous BM cells did not predict the clinical response, underscoring the advisability of additional mechanistic exploration.