OBJECTIVETo compare the predictive value of 5 prognostic models (Child-Pugh scoring, Clichy prognostic index [PI], New Clichy PI, Rotterdam BCS index, and BCS-TIPS PI) for Budd-Chiari syndrome (BCS) in China.

METHODSThe clinical data of 123 patients with BCS were retrospectively analyzed, among whom 99 survived and 24 died. The indices of the 5 prognostic models were respectively calculated, and each index was compared by F-test between the survival and death groups. The area under curve (AUC), sensitivity, and specificity of the models were computed and analyzed by receiver-operator characteristic (ROC) curve.

RESULTSThe indices of Child-Pugh, Clichy, New Clichy and Rotterdam BCS Index models in the death group (8.792∓2.0, 5.924∓0.783, 5.695∓1.81, and 0.615∓1.133, respectively) were significantly higher than those in the survival group (7.141∓1.443, 5.221∓0.834, 3.981∓1.033, and -0.148∓0.896, respectively, P<0.01), and only BCS-TIPS model had no significant difference between the two groups (P>0.05). The AUC of the 5 indices were 0.738, 0.720, 0.776, 0.721, and 0.502, with Youden indices of 0.370, 0.410, 0.439, 0.473, and 0.051, respectively.

CONCLUSIONChild-Pugh scoring, Clichy PI, New Clichy PI, and Rotterdam BCS Index models can distinguish survival from death in BCS patients in China. New Clichy PI has the highest predictive value and is suitable for use in China, whereas the other models have relatively low predictive values, among which BCS-TIPS model is not advisable. Meanwhile effort should be made to establish a prognostic model for BCS in China.