Effective local delivery to the heart remains an obstacle to successful therapeutic application of a number of drugs and biological agents. This study was designed to study and optimize the delivery characteristics of transendocardial intramyocardial (IM) administration, determine myocardial deposition and retention over time, and compare it to transepicardial IM injection. Thirty-nine pigs were used for the study (15 for catheter optimization, 15 for transendocardial IM delivery, and 9 for transepicardial IM delivery). (125)I-Fibroblast growth factor-2 (FGF2) (25 microCi) was used as the prototype molecule. Tissue and myocardial distribution was determined at 1 and 24 h and 7 days. Using 1-h (125)I-FGF2 myocardial deposition as a parameter for delivery efficiency, the optimal needle length and delivery volume for transendocardial based delivery were determined to be 6 mm and 0.1 ml, respectively. Using these parameters for endocardial delivery, (125)I-FGF2 cardiac activity was 18.01 +/- 3.84% of delivered activity at 1 h, 11.65 +/- 5.17% at 24 h, and 2.32 +/- 0.87% at 7 days in ischemic animals. Studies in nonischemic animals produced similar results. For transepicardial delivery, (125)I-FGF2 cardiac-specific activity was 23.14 +/- 12.67% for the 6-mm needle, declining to 12.32 +/- 8.50% at 24 h, and did not significantly differ from values obtained following transendocardial delivery. Thus, optimized transendocardial intramyocardial delivery using Biosense guidance results in efficient delivery of FGF2 to the target myocardium that is comparable with transepicardial delivery, both providing markedly higher myocardial deposition and retention and lower systemic recirculation of FGF2 than intracoronary, intrapericardial, or intravenous delivery. However, myocardial distribution is limited to injection sites.