The purpose of this study is to evaluate the feasibility of percutaneous antegrade myocardial gene transfer (PAMGT). A consistent and safe technique for in vivo gene transfer is required for clinical application of myocardial gene therapy. PAMGT with concomitant coronary venous blockade was performed in 12 swine. The myocardium was preconditioned with 1 min of occlusion of the left anterior descending and left circumflex arteries. The anterior interventricular vein was occluded during left anterior descending artery delivery, and the great cardiac vein at the entrance of the middle cardiac vein was occluded during left circumflex artery delivery. With arterial and venous balloons inflated (3 min) and after adenosine (25 mug) injection, PAMGT was performed by antegrade injection of an adenoviral solution (1 ml of 10(11) plaque-forming units in each coronary artery) carrying beta-galactosidase or saline through the center lumen of the angioplasty balloon. In one set of animals, PAMGT was performed with selective coronary vein blockade (n = 9); in another set of animals, PAMGT was performed without coronary vein blockade (n = 5). At 1 wk after gene delivery, the animals were killed. Quantitative beta-galactosidase analysis was performed in the left and right ventricular walls. PAMGT was successfully performed in all animals with and without concomitant occlusion of the coronary veins. Quantitative beta-galactosidase analysis showed that PAMGT with coronary blockade was superior to PAMGT without coronary blockade. beta-Galactosidase activity increased significantly in the beta-galactosidase group compared with the saline group: 1.34 +/- 0.18 vs. 0.81 +/- 0.1 ng (P </= 0.01) in the left ventricular wall and 0.91 +/- 0.1 vs. 0.66 +/- 0.07 ng (P </= 0.05) in the right ventricular wall. PAMGT with selective coronary venous blockade is feasible, reproducible, and safely achieved in a large-animal model.