To the Editor: We have read with interest the review article by Ge and Runyon on the changing role of non-selective b-blockers (NSBBs) in cirrhosis . Their role in reduction of portal hypertension is accomplished by lowering portal inflow (b1 blockade) and inducing splanchnic vasoconstriction (b2 blockade). Risk of variceal bleeding correlates also with other factors, i.e., infections and severity of liver dysfunction, suggesting absence of a pure mechanical model. For this reason, patients with more severe disease (Child-Pugh Class C) should undergo primary prophylaxis even in the presence of small varices . In this paper, the authors suggested a ‘‘window hypothesis’’, also supported by Krag and co-authors , according to which NSBBs are beneficial only in decompensated patients with medium-large varices but not in patients with early or end-stage cirrhosis with refractory ascites. Their recommendation on patients with refractory ascites is mainly based on data from a study performed by Serstè and collaborators, which showed that mortality in patients with refractory ascites was increased in those taking NSBBs. The same group hypothesized that b-blockade could induce counter-regulatory over-activation of the Renin-AngiotensinAldosterone axis, increasing incidence of paracentesis-induced circulatory dysfunction, which could be associated with impaired renal function and reduced survival [4,5]. These data were not confirmed by another study on patients with ascites taking propranolol ; moreover, by analyzing all published randomized controlled trials on prophylaxis for variceal hemorrhage, bleeding unrelated mortality was similar between patients on NSBBs and those treated with other therapies in primary (277/955 vs. 287/1175; OR 0.91 95% CI 0.73 to 1.15) and secondary prophylaxis (188/1143 vs. 225/1208; OR 0.87 95% CI 0.68 to 1.12) without heterogeneity amongst studies, while causes of death were not different between patients on and off NSBBs therapy. These results were confirmed in the subgroup analysis of studies with higher prevalence of ascites (>50%) . Moreover the dose of propranolol should be further investigated as a potential factor associated with increased mortality; for instance, in the French study, high propranolol doses (mean 132 mg/day) could have contributed to altered haemodynamics . Furthermore, chronotropic and inotropic negative effect by NSBBs may reduce cardiac function in patients with more severe cirrhotic cardiomyopathy or new onset of sepsis. For instance, we admitted a patient with alcoholic cirrhosis on the waiting list for liver transplantation who developed spontaneous bacterial peritonitis and sepsis and required large volume paracentesis (LVP); systemic vascular resistances were severely reduced at baseline in respect of previous procedures, and at the end of LVP there was a further reduction of 31.2%; notwithstanding, a counterbalancing increase of cardiac output to 12.9 L/m (+30.9%) prevented development of hepatorenal syndrome after LVP.