[Value of serum alpha-fetoprotein for the prognostic evaluation of hepatitis B virus-related acute-on-chronic liver failure treated with artificial liver].

To investigate the value of alpha-fetoprotein (AFP) level on survived hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) patients treated with artificial liver. Clinical indicators of HBV-ACLF patients who were previously treated with plasma exchange-based artificial liver at our department were retrospectively collected. The difference of serum AFP level between the survival and the death group was compared at 30, 90 and 180 days after artificial liver treatment. The ROC curves of the subjects were plotted, and the sensitivity and specificity of AFP for the survival prediction of the patients at 30, 90 and 180 days after artificial liver surgery were calculated. AFP was divided into a high AFP group and a low AFP group using median value. AFP and postoperative survival predictive value at 30, 90, and 180 days were analyzed. A total of 93 cases were included in this study. The AFP of the survival group at 30, 90, and 180 days was (231.0 ± 286.2) ng / ml, (237.69 ± 297) ng / ml, (229.44 ± 286.46) ng/ml, and the death group was (76.4 ± 104.7) ng/ml, (103.13 ± 116.99) ng / ml, (136.34 ± 2.9.29) ng/ml, respectively. AFP of the death group was significantly lower than the corresponding survival group ( < 0.05). Receiver operating characteristic (ROC) curve analyses indicated that the area under the curve (AUC) and its 95% confidence interval at 30, 90, and 180 days after artificial liver surgery were 0.739 (0.611 ~ 0.867), 0.675 (0.550 ~ 0.80), 0.653 (0.524 ~ 0.781), respectively. The median serum AFP value was 110 ng/ml, and the survival analysis showed that the survival time of the high AFP group was significantly higher than the low AFP group at 30 d ( = 0.01), 90 d ( = 0.04) and 180 d ( = 0.03) after artificial liver surgery. Serum AFP can be used as a predictor of survival for HBV-ACLF patients after artificial liver therapy and its clinical value needs to be further verified by the larger sample size.