We analyzed whether
serum albumin is independently associated with portal vein
thrombosis (PVT) in
liver cirrhosis (LC) and if a biologic plausibility exists. This study was divided into three parts. In part 1 (retrospective analysis), 753 consecutive patients with LC with ultrasound-detected PVT were retrospectively analyzed. In part 2, 112 patients with LC and 56 matched controls were entered in the cross-sectional study. In part 3, 5 patients with
cirrhosis were entered in the in vivo study and 4 healthy subjects (HSs) were entered in the in vitro study to explore if
albumin may affect platelet activation by modulating oxidative stress. In the 753 patients with LC, the prevalence of PVT was 16.7%; logistic analysis showed that only age (odds ratio [OR], 1.024; P = 0.012) and
serum albumin (OR, -0.422; P = 0.0001) significantly predicted patients with PVT. Analyzing the 112 patients with LC and controls, soluble clusters of differentiation (CD)40-ligand (P = 0.0238), soluble Nox2-derived
peptide (sNox2-dp; P < 0.0001), and urinary excretion of
isoprostanes (P = 0.0078) were higher in patients with LC. In LC,
albumin was correlated with sCD40L (Spearman's rank correlation coefficient [rs ], -0.33; P < 0.001), sNox2-dp (rs , -0.57; P < 0.0001), and urinary excretion of
isoprostanes (rs, -0.48; P < 0.0001) levels. The in vivo study showed a progressive decrease in platelet aggregation, sNox2-dp, and urinary 8-iso
prostaglandin F2α-III formation 2 hours and 3 days after
albumin infusion. Finally, platelet aggregation, sNox2-dp, and
isoprostane formation significantly decreased in platelets from HSs incubated with scalar concentrations of
albumin. Conclusion: Low
serum albumin in LC is associated with PVT, suggesting that
albumin could be a modulator of the
hemostatic system through interference with mechanisms regulating platelet activation.