Improving Primary Biliary Cholangitis Outcomes in the Evolving Landscape of Second-Line Therapies

Abstract

Primary biliary cholangitis (PBC) is an autoimmune liver disease characterised by T-cell–mediated destruction of intrahepatic bile ducts, leading to progressive cholestasis and, if untreated, end-stage liver disease with the attendant need for liver transplantation (LT). In addition, symptoms like fatigue and pruritus affect up to 70% of patients, significantly impairing patients' quality of life [1]. The introduction of ursodeoxycholic acid (UDCA) nearly 30 years ago revolutionised PBC management and changed the natural history of the disease by improving cholestasis [2] and transplant-free survival [3]. However, UDCA is ineffective in up to 40% of patients and does generally not alleviate symptoms, highlighting the need for effective second-line therapies [4]. To date, second-line therapies for PBC include the farnesoid X receptor (FXR) agonist obeticholic acid (Ocaliva, OCA) and four peroxisome proliferator-activated receptor (PPAR) agonists: fenofibrate and bezafibrate (off-label) and elafibranor and seladelpar (licensed).

OCA, approved in 2016, was supported by the POISE trial, where 47% of patients (vs. 10% placebo) achieved the primary endpoint, defined as alkaline phosphatase (ALP) reduction < 1.67 times the upper limit of the normal range (ULN) and ≥ 15% reduction from baseline, and normal bilirubin at 12 months [5]. Although the COBALT confirmatory trial was terminated due to recruitment issues [6], the POISE long-term safety extension (LTSE) study showed that PBC patients treated with OCA for 6 years had a 2.4% rate of clinical events (LT or death), significantly lower than rates in non-OCA treated real-world external controls from the Global PBC (10.0%) and UK-PBC (13.2%) cohorts [7]. The major safety concerns that emerged in the trial and post-approval studies were worsening of pruritus and the occurrence of severe liver injury in patients with decompensated cirrhosis [5, 8]. In a decision that, while surprising to the community, may have been anticipated by regulators, the EMA Committee for Medicinal Products for Human Use (CHMP) has recommended the revocation of the marketing authorization for OCA. The recommendation stems from the conclusion that the benefits of OCA no longer outweigh its associated risks. Following a thorough review of the available evidence, the committee determined that the clinical benefits of OCA have not been substantiated [9].

PPAR agonists exert anticholestatic effects through activating PPARs with different specificities. Among them, the pan-PPAR agonist bezafibrate has been extensively used as off-label treatment in PBC patients with no response or intolerant to UDCA based on the results of the phase 3 BEZURSO and FITCH trials. In the BEZURSO trial [10], 31% of bezafibrate-treated patients (vs. 0% in the placebo group) met the primary endpoint, defined as normalisation of ALP, transaminases, total bilirubin, albumin, and international normalised ratio at 24 months. Additionally, in a large real-world Japanese study, bezafibrate was associated with a significant improvement in transplant-free survival [11]. Furthermore, in the FITCH trial [12] bezafibrate demonstrated a beneficial effect on patients with moderate-to-severe pruritus. The major safety concerns that emerged in the trial and real-world studies are myalgias, drug-induced liver injury (DILI) and creatinine increase [10-12]. Recently, the FDA granted accelerated approval for two novel PPAR agonists, elafibranor (PPARα/δ agonist) and seladelpar (PPARδ agonist) based on the ELATIVE [13] and RESPONSE [14] trials, respectively. The primary endpoint, identical to the POISE trial, was met by 51% of elafibranor-treated patients (vs. 4% in the placebo group) and 61.7% of seladelpar-treated patients (vs. 20% in the placebo group). Only seladelpar significantly reduced the pruritus numerical rating scale in patients with baseline moderate-to-severe pruritus. DILI and myalgias were rare in both trials.

The emergence of novel therapies for PBC, each with distinct mechanisms and varying efficacy and safety profiles (Table 1), has highlighted the need for comparative data.

In the current issue of Liver International, Giannini E. G. and Pasta A. et al. [15] attempt to address this gap through a systematic review and network meta-analysis based on 570 patients enrolled in three phase III RCTs (POISE [5], ELATIVE [13], and RESPONSE [14] trials). The authors identified three common comparators to make indirect comparisons among the three drugs: the biochemical response at 12 months as defined in the POISE trial, the incidence of de novo pruritus, and the occurrence of serious adverse events (SAE). The analysis confirmed that all three agents demonstrated greater efficacy than placebo in achieving biochemical response, but OCA was uniquely linked to an increased risk of pruritus and a higher incidence of SAE. Seladelpar notably reduced the risk of de novo pruritus, while elafibranor showed no difference compared with placebo. The results of indirect comparisons showed that elafibranor outperformed seladelpar in achieving biochemical response (risk ratio 4.37, 95% CI 1.01–18.87), whereas no significant differences emerged between seladelpar and OCA or elafibranor and OCA. Both seladelpar and elafibranor were associated with a lower risk of de novo pruritus compared to OCA, while no differences were observed between the two newly approved PPAR agonists. No significant differences in the incidence of SAE were observed among the three therapies.

These findings must be interpreted in light of the study's limitations. The lack of direct, head-to-head comparisons and individual-level data, necessitates reliance on indirect comparisons, which are based on assumptions of transitivity and consistency. More importantly, the limited number of studies available in the literature hugely affect the precision and reliability of the network estimates. Additionally, the exclusion of the Phase III BEZURSO trial, while ensuring consistency due to differences in inclusion criteria and primary endpoint definitions, reduces the generalisability of the results. Another important consideration is the absence of long-term safety and survival data for the newer agents, particularly in comparison to OCA (and bezafibrate, which was excluded from the analysis). Such data will require years of observation to emerge. Finally, focusing solely on the onset of new pruritus may not fully capture the broader impact of these therapies on patients' overall symptom burden.

Even with the use of the best treatment option according to comparative data, the possibility that adequate biochemical response will not be achieved is still considerable. Indeed, another important readout of this study is that an overall 53.0% of patients treated with an active therapy (vs. 15.2% of placebo group) met the biochemical response criteria. Multiple strategies could serve to address this “therapeutic ceiling”, such as identifying predictive tools of response to improve patient selection for different therapies and adopting an “early escalation” combination treatment strategies [16]. Ongoing studies testing triple therapy with UDCA, OCA, and bezafibrate suggest a synergistic effect and improved safety profile [17, 18].

On a different note, the major challenge faced in PBC is the regulatory approval model, which demands evidence on hard endpoints for full marketing authorisation. Post-approval, retaining patients in placebo-controlled trials becomes harder, and the novel therapies will face additional hurdles with expanding treatment options to demonstrate their efficacy through both trial data and real-world evidence accepted by regulators [19].

In conclusion, the therapeutic landscape for PBC is expanding, yet challenges remain to achieve treatment personalisation and improve patients' outcomes. High-quality head-to-head studies are needed to directly compare short-term efficacy and safety profiles, and real-world data will provide insights into long-term outcomes.

M.C. has received grant/research support from Genetic Spa and Intercept and has served on an advisory board and/or been a consultant (paid or unpaid) for Advanz, Albireo, AMAF, Calliditas, CymaBay, Echosens, EpaC, Falk, Genetic Spa, GSK, Ipsen, Kowa, Mayoly, Mirum, Perspectum, PSC Paediatric Foundation, and Zydus. P.I. has received grant/research support from Bruschettini and Intercept and has served on an advisory board and/or been a consultant (paid or unpaid) for Advanz, AMAF, Calliditas, Cymabay, EpaC, Falk, Gilead, GSK, Ipsen, Mirum, and Zydus. The remaining authors have no conflicts to report.