Influence factors and a predictive scoring model for measuring the biochemical response of primary biliary cholangitis to ursodeoxycholic acid treatment

Abstract

Aims: The biochemical response after ursodeoxycholic acid (UDCA) treatment contributes toward predicting the prognosis for primary biliary cholangitis (PBC) patients. This study aimed to establish a score model that can be used for predicting the biochemical response.

Patients and methods: A total of 218 patients in the derivation group and 66 patients in the verification group were enrolled. Response endpoints were based on the Barcelona criteria combined with the Paris I criteria. We determined independent factors of the biochemical response by univariate and multivariate analyses. Then, we established a predictive score model on the basis of regression coefficients after adjusted multivariate analyses.

Results: The median follow-up duration in the derivation and the verification group was 12.9 and 12.2 months, respectively. Multivariate logistic regression analysis after adjusting for sex and age indicated that First-UDCA treatment [odds ratio (OR)=2.543, 95% confidence interval (CI): 1.234-5.240, P=0.011], baseline alanine aminotransferase level (OR=1.265, 95% CI: 1.089-1.471, P=0.002), and baseline total bilirubin level (OR=0.571, 95% CI: 0.420-0.776, P<0.001) were independent factors that influenced the biochemical response in PBC patients after 1 year of UDCA treatment. Therefore, the resulting biochemical response prediction score model represented the sum of the points corresponding to these three variables. The area under the receiver operating characteristic curve of the score model in the derivation group and the verification group was 0.763 (95% CI: 0.701-0.817, P<0.001) and 0.798 (95% CI: 0.681-0.887, P<0.001), respectively.

Conclusion: We developed and verified an easy-to-use scoring model for the first time, which showed excellent predictive value for the biochemical response in PBC patients.