Effects of preconditioning and Ginkgo biloba extract (EGb 761) were studied in isolated nondiabetic and diabetic ischaemic and re-perfused rat hearts. Hearts were randomly divided into five groups in both the age-matched non-diabetic and the 8-week streptozotocin-induced diabetic groups: Group I, hearts were subjected to 30 min of global ischaemia followed by 30 min of re-perfusion; Group II, one cycle of preconditioning consisting of 5 min ischaemia and 10 min re-perfusion before the induction of 30 min of ischaemia and 30 min of re-perfusion; Group III, two cycles of preconditioning; Group IV, three cycles; and Group V, four cycles before the onset of 30 min ischaemia followed by 30 min of re-perfusion. Four cycles of ischaemic preconditioning resulted in a reduction of arrhythmias in non-diabetic rats. Thus, in non-diabetics, the incidence of ventricular fibrillation and tachycardia fell from 92% and 100% (no preconditioning) to 33% (p < 0.05) and 42% (p < 0.05), respectively. Four cycles of preconditioning failed to reduce the incidence of re-perfusion arrhythmias in diabetic subjects. Preconditioning reduced the formation of oxygen free radicals measured by electron spin resonance spectroscopy, but the recovery of cardiac function was low in all non-diabetic and diabetic preconditioned groups. EGb 761 at 25 and 50 mg/kg improved cardiac function in non-preconditioned and preconditioned non-diabetic and diabetic hearts. During re-perfusion in the four-cycle preconditioned non-diabetic and diabetic groups, the amount of free radicals was reduced approximately by 50 and 70% using 25 and 50 mg/kg of EGb 761, respectively. EGb 761 improved cardiac function after ischaemia in both non-preconditioned and preconditioned non-diabetic and diabetic rats. Our data suggest that diabetes could abolish the precondition-induced protection.