Elafibranor in primary biliary cholangitis: no significant extra adverse events, but most safety questions are underpowered

Design

Systematic review and meta-analysis of three randomised trials with GRADE and trial sequential analysis

Population

274 patients with primary biliary cholangitis and inadequate response to ursodeoxycholic acid (183 elafibranor, 91 placebo)

Primary outcome

Any adverse event, with nine further safety outcomes

Effect

Any adverse event RR 1.06 (95% CI 0.97 to 1.16); pruritus RR 0.77; fatigue RR 0.71 (not significant)

This meta-analysis pooled three placebo-controlled randomised trials with 274 patients with primary biliary cholangitis who responded inadequately to ursodeoxycholic acid (183 on elafibranor, 91 on placebo). All trials were at low risk of bias. Ten safety outcomes were pooled with GRADE and trial sequential analysis.

There was no significant difference from placebo for any outcome. Any adverse event had a risk ratio of 1.06 (95% CI 0.97 to 1.16; I-squared 0%). Pruritus (RR 0.77) and fatigue (RR 0.71) leaned in favour of elafibranor, and nausea and treatment-related events leaned slightly against it, none significantly. Certainty was moderate, downgraded for imprecision. Sequential analysis confirmed futility for any adverse event but showed the other nine outcomes had only 2.2% to 9.7% of the required information.

In plain terms, the pooled data are reassuring but cannot exclude rare or long-term harm. Elafibranor is approved as a second-line option by the FDA and EMA (2024), and its status and cost in India are not stated here. Monitoring should follow the product label.

• Consider second-line therapy for PBC patients with an inadequate response to ursodeoxycholic acid, in a specialist clinic.
• Review ursodeoxycholic acid response at 12 months with ALP-based criteria.
• Monitor patients on second-line therapy as the product label directs.
• Check availability and cost locally; approval in India is not stated in this report.

Why it matters

It separates safety outcomes that look truly null from ones that are simply underpowered.

Don't overread it

Only three trials with 274 patients were pooled; rare and long-term harms cannot be excluded, and the review addresses safety, not efficacy.

The statistics, in plain English

A risk ratio near 1.06 with an interval of 0.97 to 1.16 means any-adverse-event rates were about the same as placebo. Trial sequential analysis estimates how much data are needed for a reliable answer; here only a small fraction (2.2% to 9.7%) had accrued for nine outcomes, so 'no difference' mostly means 'not enough data yet'.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.