Background: Sublethal periods of ischemia preceding a prolonged interval of ischaemia protect the myocardium. This myocardial preconditioning (PC) appears to be effected by K_ATP channels. These channels occur both in the sarcolemma and the mitochondrial membrane. We investigated whether mitochondrial K_ATP channels are the end-effector of PC in the human myocardium. Methods: Right atrium specimens obtained from patients undergoing cardiac surgery were prepared and incubated in buffer solution at 37°C. After 30-min stabilisation, the muscles were made ischemic for 90 min and then reperfused for 120 min. The preparations were randomised into eight experimental groups (n=6/group): (1) Aerobic control — incubated in oxygenated buffer for 210 min, (2) ischemia alone — 90 min ischemia followed by 120 min reperfusion, (3) PC — preconditioned with 5 min ischemia/5 min reperfusion, (4) Glibenclamide (10 μM) in the incubation media for 10 min before PC, (5) 5-hydroxydecanoate (5-HD, MitoK_ATP blocker, 1 mM) in the incubation media for 10 min before PC, (6) HMR 1883 (SarcK_ATP blocker, 10 μM) in the incubation media for 10 min before PC, (7) Pinacidil (0.5 mM) in the incubation media for 10 min before ischemia, and (8) Diazoxide (MitoK_ATP opener, 0.1 mM) in the incubation media for 10 min before ischemia. Creatinine kinase leakage into the medium (CK, IU/g wet wt) and MTT reduction (OD/mg wet wt.), an index of cell viability, were assessed at the end of the experiment. Results: Ischemia alone resulted in a significant increase in CK leakage (8.01±0.35) and decrease in MTT (0.15±0.01) from the values seen in the aerobic control (2.24±0.52 and 0.78±0.10 respectively, P<0.05 in both instances). PC fully reversed the effect of ischemia (CK=2.97±0.31 and MTT=0.61±0.05; P<0.05 vs. ischemia alone group but P=NS vs. aerobic control group). Both Glibenclamide and 5-HD abolished the protection induced by PC (CK=6.23±0.5 and 7.84±0.64; MTT=0.18±0.03 and 0.13±0.02, respectively, P<0.05 vs. PC), but interestingly, the protective effect of PC was not abolished by HMR 1883 (CK=2.85±0.24 and MTT=0.58±0.05, P=NS vs. PC). Diazoxide mimicked the protective effect of PC (CK=3.56±0.32 and MTT=0.58±0.02, P=NS vs. PC), however pinacidil exhibited less protection than PC (CK=4.02±0.16 and MTT=0.30±0.02, P<0.05 vs. PC). Conclusions: These studies demonstrate that K_ATP channels are the end-effectors of ischemic preconditioning and that protection is mediated by mitochondrial K_ATP channels in human right atrial myocardium.