OBJECTIVE The aim of the study was to assess the occurrence of triggered activity and automaticity in ventricular trabeculae from failing human hearts and normal and failing rabbit hearts during exposure to a normal and altered extracellular environment. METHODS Ventricular trabeculae were harvested from failing human hearts (from patients undergoing cardiac transplantation) and from normal and failing rabbit hearts (combined volume and pressure overload). Trabeculae were superfused with normal Tyrode solution followed by a modified Tyrode solution, which mimicked the extracellular milieu in patients with severe heart failure. Modified Tyrode solution contained low potassium (3.0 mM), low magnesium (0.4 mM), and noradrenaline (1 microM). RESULTS During superfusion with normal Tyrode solution, early afterdepolarisations, delayed afterdepolarisations, and automaticity were not observed in trabeculae from failing hearts. In the modified Tyrode solution, early afterdepolarisations could be induced in 26% of control rabbit and 30% of failing rabbit trabeculae, but never in human trabeculae. During superfusion with the modified solution delayed afterdepolarisations or triggered activity could be induced in 50% of the human failing trabeculae, in 43% of the failing rabbit trabeculae, and in 9% of the normal rabbit trabeculae (p < 0.01); automaticity was observed in 44% of the human trabeculae, and in 7% of the failing rabbit trabeculae, but in none of the control rabbit trabeculae. In failing rabbit myocardium action potential duration was prolonged at cycle lengths > or = 350 ms, but not at shorter cycle lengths. CONCLUSIONS Delayed afterdepolarisations and automaticity, but not early afterdepolarisations, occur more frequently in myocardium from failing hearts, but only during superfusion with a modified Tyrode solution. This emphasises that the extracellular environment is important with respect to arrhythmogenesis in heart failure, apart from the fixed cellular defect due to heart failure per se. Prolongation of the action potential in failing hearts does not occur at physiological and higher heart rates and therefore cannot be regarded as a protective factor in the prevention of reentrant arrhythmias. The rate of triggered and automatic rhythms was slow. Therefore these mechanisms cannot be responsible for clinical ventricular tachycardias or fibrillation, but may serve as triggers for reentrant arrhythmias.