Epicardial antiarrhythmic drug administration was studied as a therapeutic approach for experimental ventricular tachycardia (VT) in an open-chest dog model. Lidocaine-polyurethane matrices (28%, w/w) were formulated as a model system. Matrices were placed on the left ventricular epicardium in each of 23 anesthetized open-chest dogs with ouabain-induced VT, to evaluate effectiveness in restoring sinus rhythm. Conversion occurred in all animals treated with matrices containing 300 mg or more of lidocaine after 1.5 to 7.0 min. The matrix lidocaine content correlated linearly with the time required for conversion to sinus rhythm (r = 0.75, P = 0.0002); irrespective of matrix size the myocardial/plasma lidocaine ratio was 20.1 +/- 4.2 (mean +/- SD) at the time of conversion. In a separate series of five dogs without ventricular tachycardia, systolic wall thickening measured with sonomicrometers after 5 min of controlled-release lidocaine administration (500- to 1000-mg matrix lidocaine content, 7.48 +/- 3.49-mg/kg dose) was only minimally diminished (-14.1%) and this effect was observed only at the site of matrix placement on the anterior-apical epicardium. In contrast, intracoronary injection of 0.3 or 1.0 mg/kg of lidocaine-HCl resulted in complete elimination of wall thickening or replacement by systolic thinning. Thus epicardial administration of lidocaine from polyurethane matrices was an effective means of treating ouabain-induced ventricular tachycardia. Regional myocardial function in the vicinity of the matrices was modified to a very limited degree, supporting the view that the matrices can be used safely, without serious risk to ventricular contractile performance.