Development and Validation of a Scoring System to Predict Outcomes of Patients With Primary Biliary Cirrhosis Receiving Ursodeoxycholic Acid Therapy

Background & Aims Approaches to risk stratification for patients with primary biliary cirrhosis (PBC) are limited, single-center based, and often dichotomous. We aimed to develop and validate a better model for determining prognoses of patients with PBC. Methods We performed an international, multicenter meta-analysis of 4119 patients with PBC treated with ursodeoxycholic acid at liver centers in 8 European and North American countries. Patients were randomly assigned to derivation (n = 2488 [60%]) and validation cohorts (n = 1631 [40%]). A risk score (GLOBE score) to predict transplantation-free survival was developed and validated with univariate and multivariable Cox regression analyses using clinical and biochemical variables obtained after 1 year of ursodeoxycholic acid therapy. Risk score outcomes were compared with the survival of age-, sex-, and calendar time-matched members of the general population. The prognostic ability of the GLOBE score was evaluated alongside those of the Barcelona, Paris-1, Rotterdam, Toronto, and Paris-2 criteria. Results Age (hazard ratio = 1.05; 95% confidence interval [CI]: 1.04-1.06; P <.0001); levels of bilirubin (hazard ratio = 2.56; 95% CI: 2.22-2.95; P <.0001), albumin (hazard ratio = 0.10; 95% CI: 0.05-0.24; P <.0001), and alkaline phosphatase (hazard ratio = 1.40; 95% CI: 1.18-1.67; P =.0002); and platelet count (hazard ratio/10 units decrease = 0.97; 95% CI: 0.96-0.99; P <.0001) were all independently associated with death or liver transplantation (C-statistic derivation, 0.81; 95% CI: 0.79-0.83, and validation cohort, 0.82; 95% CI: 0.79-0.84). Patients with risk scores >0.30 had significantly shorter times of transplant-free survival than matched healthy individuals (P <.0001). The GLOBE score identified patients who would survive for 5 years and 10 years (responders) with positive predictive values of 98% and 88%, respectively. Up to 22% and 21% of events and nonevents, respectively, 10 years after initiation of treatment were correctly reclassified in comparison with earlier proposed criteria. In subgroups of patients aged <45, 45-52, 52-58, 58-66, and ≤66 years, age-specific GLOBE-score thresholds beyond which survival significantly deviated from matched healthy individuals were -0.52, 0.01, 0.60, 1.01 and 1.69, respectively. Transplant-free survival could still be accurately calculated by the GLOBE score with laboratory values collected at 2-5 years after treatment. Conclusions We developed and validated scoring system (the GLOBE score) to predict transplant-free survival of ursodeoxycholic acid-treated patients with PBC. This score might be used to select strategies for treatment and care.