We thank Drs Rodrigues and Banales for the interest in our study.1, 2 We aimed to examine the histological effect of the combination of ursodeoxycholic acid (UDCA) and bezafibrate (BZF) in difficult to treat patients with primary biliary cholangitis (PBC). We agree with Dr Rodrigues et al that late-stage PBC is an entity underrepresented in clinical trials, so recruitment of greater number of patients with the cirrhotic disease may be difficult to obtain in future trials. PBC is increasingly diagnosed at a much earlier stage, and more than 60% of patients are asymptomatic at presentation with the majority not having cirrhosis3, 4; this was observed in our study. Thus a key therapeutic challenge in early non-cirrhotic PBC is to prevent progression; and/or if fibrosis exists to promote regression to earlier stages. The early identification of progression based on biomarkers and predictive scores is important for this group of difficult-to-treat patients with PBC who may benefit from therapies beyond UDCA, including obeticholic acid, off-label bezafibrate, as well as clinical trial agents in development.5

Our results reinforced previous data on patients with PBC for whom UDCA response is insufficient, where combined UDCA-BZF therapy is associated with significant decreases or even normalization of biochemical markers of cholestasis.6 We observed that normalization of ALP at 12 months was associated with histological benefit at 5 years, but not in those patients without ALP normalization. Moreover, we observed a significant improvement in GLOBE and UK-PBC risk score under UDCA-BZF therapy. However, these results should be considered with caution since these predictive score needs to be validated in combination therapy; furthermore, the impact of BZF on AST values (a component of APRI and the UK-PBC risk score) needs better understanding, recognising extra-hepatic sources for the AST may include fibrate induced myopathy.

We want to point out that moderate or severe lymphocytic interface hepatitis (LIH) has been identified as an independent predictive factor of cirrhosis development as well as liver-related morbidity and mortality.7, 8 Notably, LIH evaluation still needs liver biopsy since non-invasive biomarkers are not validated for this histologic finding. As for the POISE trial population where obeticholic acid was studied,9 we observed a high prevalence of LIH in our cohort, which was a marker of low probability of UDCA response.10 However, in our study, a significant reduction in LIH was associated with UDCA-BZF therapy.

Finally, we agree with the editorial that the benefit of UDCA-BZF combination treatment needs further studies, especially in long-term controlled studies, in comparison with, and/or in conjunction with other therapies such as obeticholic acid, and from a cost-effectiveness perspective.

The authors' declarations of personal and financial interests are unchanged from those in the original article.2