Background Coronary flow responses to endothelium-dependent (acetylcholine [ACh] or 5-hydroxytryptamine [5-HT]) and endothelium-independent (adenosine [ADE] or nitroprusside [NP]) vasodilators may be altered before and after 1-day hypothermia during the perfusion of arginine vasopressin (AVP), D-arginine (D-ARG), L-arginine (L-ARG), or nitro-L-arginine methyl ester (L-NAME). Methods and Results Four groups of guinea pig hearts (37.5 degrees C [warm]) were perfused for 6 hours with AVP, L-ARG, L-NAME, or nothing (control). Five heart groups (cold) were perfused with AVP, D-ARG, L-ARG, L-NAME, or nothing (control), but after 2 hours they were perfused at low flow for 22 hours at 3.7 degrees C and again for 3 hours at 37.5 degrees C. ADE, butanedione monoxime, and NP were given for cardioprotection before, during, and after hypothermia. In warm groups, L-ARG did not alter basal flow or ADE, ACh, 5-HT, or NP responses, whereas L-NAME and AVP reduced basal flow and the ADE response, abolished ACh and 5-HT responses, and increased the NP response. In cold groups after hypothermia, L-ARG did not alter basal flow, but L-NAME, AVP, D-ARG, and control reduced flow. In the postcold L-ARG group, ACh increased peak flow, but NP did not increase flow in other cold groups. Effluent L-ARG and L-CIT in the cold control group fell from 64 +/- 9 and 9 +/- 1 mu g/L at 1 hour to 36 +/- 5 and 5 +/- 1 mu g/L at 25 hours, respectively. Left ventricular pressure and cardiac efficiency improved more in the postcold L-ARG group than in the postcold D-ARG, AVP, and L-NAME groups. Conclusions Endogenous effluent levels of L-ARG and L-CIT decrease after 24 hours in isolated hearts, whereas perfusion of L-ARG improves cardiac performance, basal coronary flow, and vasodilator responses. In contrast, L-NAME, L-ARG, and AVP limit flow and performance but maintain a partial vasodilatory response to NP. Sustained release of NO may account for improved performance after L-ARG after hypothermia.