There are few stable and reproducible large-animal models of chronic heart failure produced by ischaemic damage to the myocardium. Here we characterize a novel method of inducing myocardial damage in closed-chest sheep by catheter delivery of thrombogenic coils, and compare this with a newly described open-artery model of cardiac injury in sheep. Sham controls were compared with animals subjected to (a) 90 min of coronary artery occlusion/reperfusion by PTCA (percutaneous transluminal coronary angioplasty) balloon, and (b) permanent coronary artery occlusion induced by catheter delivery of thrombogenic coils (seven sheep/group). Both balloon occlusion/reperfusion and permanent coil occlusion resulted in well-defined anteroapical infarcts, as documented by ECG changes, significant rises in creatine kinase (both groups P<0.001) and troponin-T (both groups P<0.05), and post-mortem examination. Washout of enzymes was much more rapid in the reperfused group (P<0. 01). Infarction resulted in significant reductions in left ventricular (LV) ejection fraction (both groups P<0.01) and regional wall abnormalities. Ejection fraction 7 days post-coil (21.3+/-4.2%) was significantly lower (P<0.01) than that 7 days post-balloon (38. 8+/-4.5%). Coil-induced infarction was associated with acutely reduced arterial pressure (P<0.05), and increases in heart rate (P<0. 05), atrial pressures (P<0.05), plasma brain natriuretic peptide levels (P<0.05) and adrenaline levels (P<0.05). Rises seen in plasma endothelin levels in sham controls were blunted in the coil group (P<0.001). Haemodynamic changes were less marked in the balloon group. In conclusion, restriction of coronary artery occlusion to 90 min results in infarction, but less LV dysfunction with reduced early remodelling, compared with permanent occlusion. Acute changes in biochemical markers, haemodynamics, neurohormones and LV function confirm that these are excellent models of open- and closed-artery myocardial infarction leading to asymptomatic LV dysfunction.