Aims The role of liver biopsy in primary biliary cholangitis (PBC) is controversial, as is the optimal method of histological assessment. We compared the Ludwig and Ishak systems and three components of the Japanese (Nakanuma) staging system to evaluate their clinical and biochemical correlations and prognostic value.

Methods We reviewed biopsies from 106 patients with PBC, derived from a previous trial of colchicine therapy with 24–34 years’ follow-up, following which five clinical outcomes were evaluated: hepatic decompensation, cholestatic PBC death/liver transplant, portal hypertensive PBC death, all PBC deaths and overall survival.

Results Ludwig and Ishak stages correlated well with prognostically significant parameters, including serum bilirubin, and both Mayo and Child Scores. Serum aspartate aminotransferase correlated with interface hepatitis (IFH), and alkaline phosphatase with orcein deposition, bile duct (BD) loss and cholestasis. Ludwig correlated with all five clinical outcomes, while Ishak stage was only significantly correlated with two. While sinusoidal fibrosis, orcein deposition, BD loss and cholestasis all predicted hepatic death/transplant, after correction for Mayo Score, the only histological parameters predictive of clinical outcomes were IFH (associated with two) and sinusoidal fibrosis (associated with all five).

Conclusion Liver biopsy is required in the diagnosis of around 20% of patients with PBC. The Ludwig system is of more prognostic value than both Ishak and any of the three individual components of the Nakanuma staging system, but the major histological parameter providing independent prognostic value beyond the Mayo Score is sinusoidal fibrosis.