Background: We sought to determine the safety and preliminary efficacy of transcatheter intramyocardial administration of myoblasts in patients with heart failure (HF). Methods: MARVEL is a randomized placebo-controlled trial of image-guided, catheter-based intramyocardial injection of placebo or myoblasts (400 or 800 million) in patients with class II to IV HF and ejection fraction <35%. Primary end points were frequency of serious adverse events (safety) and changes in 6-minute walk test and Minnesota Living With HF score (efficacy). Of 330 patients intended for enrollment, 23 were randomized (MARVEL-1) before stopping the study for financial reasons. Results: At 6 months, similar numbers of events occurred in each group: 8 (placebo), 7 (low dose), and 8 (high dose), without deaths. Ventricular tachycardia responsive to amiodarone was more frequent in myoblast-treated patients: 1 (placebo), 3 (low dose), and 4 (high dose). A trend toward improvement in functional capacity was noted in myoblast-treated groups (Δ6-minute walk test of -3.6 vs +95.6 vs +85.5 m [placebo vs low dose vs high dose; P =.50]) without significant changes in Minnesota Living With HF scores. Conclusions: In HF patients with chronic postinfarction cardiomyopathy, transcatheter administration of myoblasts in doses of 400 to 800 million cells is feasible and may lead to important clinical benefits. Ventricular tachycardia may be provoked by myoblast injection but appears to be a transient and treatable problem. A large-scale outcome trial of myoblast administration in HF patients with postinfarction cardiomyopathy is feasible and warranted.