Background  Budd–Chiari syndrome carries significant mortality, but factors predicting this outcome are uncertain.

Aim  To determine factors associated with 3-month mortality and compare outcomes after surgical shunting or liver transplantation.

Methods  From 1985 to 2008, 51 patients with Budd–Chiari syndrome were identified.

Results  By logistic regression analysis, features associated with higher risk of 3-month mortality were Rotterdam class III, Clichy >6.6, model for end-stage liver disease (MELD) >20 and Child–Pugh C. Rotterdam class III had the best performance to discriminate 3-month mortality with sensitivity of 0.89 and specificity of 0.63, whereas Clichy >6.60 had sensitivity of 0.78 and specificity of 0.69; MELD >20 had sensitivity of 0.78 and specificity of 0.75 and Child–Pugh C had sensitivity of 0.67 and specificity of 0.72. Eighteen patients underwent surgical shunts and 14 received liver transplantation with no significant differences in survival (median survival 10 ± 3 vs. 8 ± 2 years; log-rank, = 0.9).

Conclusions  Rotterdam score is the best discrimination index for 3-month mortality in Budd–Chiari syndrome and should be used preferentially to determine treatment urgency. Surgical shunts constitute an important therapeutic modality that may help save liver grafts and prolong transplantation-free survival in a selected group of patients with Budd–Chiari syndrome.