INTRODUCTIONPortal hypertension is a major complication of chronic liver disease. In its pathophysiology, increased hepatic resistance is followed by a hyperdynamic circulatory state [1] . This hyperdynamic state, in which nitric oxide (NO) and prosatcyclin (PGI2) are important vasoactive substances, induces a decreased platelet activity, even in the absence of hepatic damage. Although NO plays an important role in modifying platelet adhesion in portal hypertension [2] , PGI2, a powerful vasodilator prostanoid with antithrombotic properties, was also found increased in mesenteric vascular bed [3] . Ultra low dose aspirin (ULDA) has shown prothrombotic activity when analysed in humans and in the interface platelet-endothelial cell [4,5] . Previous papers have shown the potentially beneficial effect of ULDA in portal hypertensive animals normalizing altered platelet activity and induced hemorrhagic time [6] . Further experiments were performed to clarify if this effect was mediated mostly by a cyclooxigenase (COX) pathway or by modifying NO synthesis, the two aspirin mechanisms of action [7] , although a previous study with a different model suggested that ULDA could decrease PGI2 synthesis [8] . The same in vivo Laser induced thrombus formation, was used in portal hypertensive rats to investigate the effects of Indomethacin, a non selective COX inhibitor, and L-Nitro Arginin Methyl Ester (NAME), a non selective inhibitor of NO production. The results suggested that the effects of ULDA were more influenced by COX pathway than by NO synthesis inhibition [9] . Addition of ULDA in portal hypertensive rats, when previously inhibiting COX with Indomethacin, increased number of emboli and duration Abstract AIM: To study the mechanism involved in the potentially beneficial effect of ultra low dose aspirin (ULDA) in prehepatic portal hypertension, rats were pretreated with selective COX 1 or 2 inhibitors (SC-560 or NS-398 respectively), and subsequently injected with ULDA or placebo.
METHODS:Portal hypertension was induced by portal vein ligation. Platelet activity was investigated with an in-vivo model of laser induced thrombus production in mesenteric circulation and induced hemorrhagic time (IHT). Platelet aggregation induced by ADP and dosing of prostanoid products 6-keto-PGF1α, TXB2, PGE2 and LTB4 were also performed.
RESULTS:The portal hypertensive group receiving a placebo showed a decreased in vivo platelet activity with prolonged IHT, an effect that was normalized by ULDA. SC-560 induced a mild antithrombotic effect in the normal rats, and an unmodified effect of ULDA. NS-398 had a mild prothrombotic action in portal hypertensive rats, similar to ULDA, but inhibited a further effect when ULDA was added. An increased 6-keto-PGF1α was observed in portal hypertensive group that was normalised after ULDA administration. TXA2 level after ULDA, remained unchanged.
CONCLUSION:These results suggest that the effect of ULDA on platelet activity in portal hypertensive rats, could act through a COX 2