"Pitfalls in assessing platelet activation status in patients with liver disease: authors' reply".

To the Editor: Impaired platelet activation has been considered for long time a typical feature of cirrhosis; however, this has been questioned as laboratory tests could be biased by in vitro artefacts (1). To overcome this issue, we decided to study in vitro intracellular pathways which are implicated in platelet activation. We could demonstrate that platelets from cirrhotic patients overexpress isoprostanes, a family of eicosanoids with pro-aggregating property (2). Such change was associated with systemic increase in biomarkers of platelet activation, namely soluble CD40 Ligand (sCD40L) and sP-selectin, implying that platelet activation can occur also in vivo. We speculated that such increase was dependent on oxidative stress as NOX2, which is implicated in the production of isoprostanes (2), was upregulated and significantly correlated with urinary excretion of isoprostanes. The interpretation of these data is questioned by Lisman and Porte because the systemic increase of platelet biomarkers, as well as soluble NOX2, may merely reflect impaired hepatic clearance; furthermore, they argue that urinary excretion of isoprostanes is not influenced by platelet isoprostanes production. We appreciated the Authors’ contribution to this debate; however, some arguments raised by these authors are essentially speculative. Thus, the role of the liver in the clearance of platelet molecules released upon activation is still uncertain. The clearance of sCD40L, for instance, is essentially renal (3, 4), and the role of liver, if any, is still unclear. Concerning NOX2, the method we developed is very recent; therefore, we cannot estimate the role of liver on its clearance. As far as urinary excretion of isoprostanes is concerned, Lisman and Porte should consider that isoprostanes stem predominantly from platelet and monocyte activation (5); therefore it is plausible that platelets contribute to the urinary excretion of isoprostanes. Consistent with this hypothesis is the parallel reduction in platelet isoprostanes and urinary isoprostanes in patients with hereditary deficiency of NOX2 (2). Therefore, the coexistence of platelet overproduction of isoprostanes in vitro and the enhanced urinary excretion of isoprostanes is consistent with platelet over-activation in cirrhosis; in accordance with this suggestion, urinary excretion of 11-dehydro-thromboxane B2 (11-dehydro-TxB2), a marker of in vivo platelet activation (6, 7), is elevated in cirrhosis (8). On this basis, we believe that urinary excretion of isoprostanes and 11-dehydro-TxB2 represents, so far, the only reliable method to further explore in vivo the mechanism(s) accounting for platelet hyperactivation in cirrhosis.