Objective: There are controversial reports in conscious animals regarding the role of cyclooxygenase-2 in late preconditioning (LP). This study analyzed the effect of COX-2 involvement in non-preconditioned hearts (NP) and in mediation of LP protection against stunning in conscious sheep submitted to a prolonged reversible ischemia. Methods: Six groups were considered: NP: 12 min ischemia and 120 min reperfusion; LP consisting of six periods of 5 min-ischemia-5 min reperfusion 24 h before the 12 min ischemia; NP and LP with either the non-selective COX-1 and COX-2 inhibitor, aspirin (20 mg/kg), or the specific COX-2 inhibitor, celecoxib (3 mg/kg) before the 12 min ischemic period. Results: Mean postischemic wall thickening fraction (as % of preischemic values) improved from 49.6 ± 4.0% in NP to 72.5 ± 3.5% in LP (p < 0.01) and a similar protection was obtained with aspirin and celecoxib in NP hearts (p < 0.01). Neither aspirin nor celecoxib administration prior to the prolonged ischemia on day 2 abrogated LP improvement of postischemic dysfunction. Moreover, LP with aspirin improved the protective response (80.7 ± 2.6%) over that obtained with aspirin in NP hearts (66.6 ± 4.7%, p < 0.05). This effect was not obtained with celecoxib. Conclusions: Aspirin and celecoxib showed that COX-2 has a detrimental effect on mechanical cardioprotection in NP hearts of conscious sheep submitted to a prolonged reversible ischemia, and does not seem to participate as mediator of LP. Aspirin revealed a similar COX-1 deleterious action, since only when both COX-1 and COX-2 were inhibited, LP was put in evidence adding functional improvement over that obtained in NP hearts treated with aspirin.