The Liver Signal Europe Just Sent About Cenobamate Will Reshape How Sponsors Design Post-Approval Safety Plans

Pull up the EMA’s April 10, 2026 PRAC meeting highlights. The lead item is not a label revision buried in a safety variation procedure. It is a direct healthcare professional communication — the regulatory equivalent of a flare gun — about cases of severe liver injury associated with cenobamate, the sodium channel and GABA-A receptor modulator approved in Europe for focal-onset seizures in adults. PRAC doesn’t reach for that tool lightly. When it does, the downstream consequences run well past the pharmacy counter and into every active trial that touches hepatic endpoints, every DSMB charter governing a CNS drug with a similar mechanism, and every sponsor whose post-approval pharmacovigilance plan was written to meet the minimum, not the moment.

This is the trend most clinical operations teams haven’t named yet: Pharmacovigilance Signal Escalation as Protocol Design Driver. Three converging pressures — accelerated approval timelines compressing post-market data maturity, increasing PRAC willingness to mandate direct professional communications rather than defer to label updates, and the structural lag between regulatory safety signals and investigational protocol amendments — are forcing a reckoning with how sponsors build safety monitoring infrastructure from Day 1 of a trial, not after the CRL arrives.

What PRAC Just Did — And Why It’s Bigger Than One Drug

Cenobamate was authorized in Europe under the brand name Ontozry. Its mechanism is distinctive — it reduces epileptic seizure frequency through a dual action on sodium channels and positive allosteric modulation of GABA-A receptors — and its efficacy data from the pivotal Phase 3 program, sponsored by SK Life Science, was strong enough to support approval. The post-market period is where the hepatic signal emerged with enough severity and specificity to trigger a PRAC recommendation for a direct healthcare professional communication in April 2026.

PRAC’s decision to mandate communication directly to prescribers — rather than relying on a label amendment to propagate safety information through the standard channels — reflects a regulatory philosophy that has been hardening across multiple therapeutic areas. The EMA’s pharmacovigilance framework, anchored in Regulation (EU) No 1235/2010 and its implementing legislation, gives PRAC the authority to recommend urgent safety restrictions and direct professional communications when the committee determines that routine label updates are insufficient to protect patients in real time. PRAC used that authority here. The subtext is that they did not trust the standard signal diffusion pathway to reach prescribers fast enough.

For anyone running a trial in the epilepsy space — or in any CNS indication where hepatic safety is a known liability — this decision is a signal about regulatory expectations, not just about cenobamate.

The Structural Gap Sponsors Keep Ignoring

Here is the counterintuitive reading of this situation that most regulatory affairs teams will miss: the cenobamate liver signal is not primarily a story about a drug that surprised regulators. It is a story about the distance between the safety monitoring architecture built for a Phase 3 trial and the pharmacovigilance infrastructure required to catch a low-frequency, severe hepatic adverse event in a heterogeneous post-approval population.

Consider the operational mechanics. A pivotal epilepsy trial typically enrolls a controlled population — adults with refractory focal seizures, screened for baseline hepatic function, excluded if they carry comorbidities that confound liver enzyme interpretation. The trial’s data safety monitoring board reviews hepatic laboratory data on a schedule defined in the DSMB charter, typically quarterly, with stopping rules calibrated to the trial’s own baseline incidence. That infrastructure is purpose-built for the trial. It was never designed to detect what happens when the drug reaches patients with undiagnosed hepatic vulnerabilities, polypharmacy interactions that weren’t represented in the trial population, or monitoring frequencies that a community neurologist manages very differently than a clinical trial site coordinator.

The FDA recognized this structural gap explicitly in its guidance on safety reporting requirements for INDs and bioavailability and bioequivalence studies, finalized in 2012, and the EMA addressed post-authorization safety study requirements under the PASS framework in its 2023 guideline on good pharmacovigilance practices. Both frameworks acknowledge the discontinuity between trial-era safety data and post-approval signal detection. But acknowledgment is not architecture. Sponsors routinely submit pharmacovigilance plans that satisfy regulatory review and fail operational stress-testing the moment a real-world signal begins to aggregate in VigiBase or the EMA’s EudraVigilance database.

The cenobamate PRAC action is evidence that regulators are no longer willing to wait for that architecture to catch up on its own.

The Operational Reckoning Coming for Sponsors and Sites

For sponsors in active CNS development — particularly those running trials on novel anticonvulsants, GABA-modulators, or any compound with a hepatic metabolism profile — the April 2026 PRAC action on cenobamate should function as an operational audit trigger. Not because their compound is necessarily at risk, but because the regulatory posture it reveals will migrate into IND-stage expectations faster than most development teams anticipate.

The FDA’s own emerging framework around real-world pharmacovigilance integration, signaled through its 2023 discussion paper on using real-world data to support safety evaluations for approved drugs, points in the same direction PRAC just walked. Regulators on both sides of the Atlantic are converging on a position: the pharmacovigilance plan you file at approval needs to function as a living clinical safety infrastructure, not a document. That means pre-specified hepatic monitoring visit schedules written into post-approval commitments with enough operational specificity that a community prescriber can execute them. It means signal detection thresholds in EudraVigilance or FDA’s Sentinel System that are calibrated to the drug’s known mechanism, not just generic reference ranges.

For CROs managing post-approval safety studies and PASS obligations in Europe, the PRAC cenobamate decision raises an immediate question about how PSURs — Periodic Safety Update Reports — are being written and reviewed. If a severe liver injury signal was accumulating in EudraVigilance with enough density to prompt a direct healthcare professional communication in April 2026, the question of when that signal first crossed a reviewable threshold in earlier PSUR cycles will be answered in the next round of GVP Module VII compliance assessments. CROs that manage PSUR authorship for CNS sponsors should be reviewing their signal triage documentation now.

For clinical trial sites running investigational CNS studies, the operational implication is less abstract: expect sponsor protocol amendments that add hepatic monitoring visit windows, lower the threshold for unscheduled laboratory assessments, and introduce DSMB reporting triggers for transaminase elevations that would have previously been managed as protocol-specified events without expedited escalation. The cenobamate signal will be cited in those amendments by name. I’ve submitted these packages myself — the precedent-citing language matters to ethics committees and IRBs, and a named PRAC action carries more weight in a safety amendment justification than a theoretical risk narrative.

Technology vendors in the pharmacovigilance and risk-based monitoring space face a different pressure. The expectation that post-approval safety signals be detected, triaged, and escalated on timelines that support direct professional communications — rather than the 60-to-90 day PSUR cycle — demands signal detection tooling that most mid-tier sponsors are not currently running. Platforms integrating EudraVigilance feeds with real-time disproportionality analysis and automated case narrative generation are no longer competitive differentiators for large pharma. They are baseline infrastructure for any sponsor with a post-approval asset in a high-sensitivity therapeutic area. The cenobamate action makes that concrete.

In the next 12 to 18 months, expect PRAC to issue direct healthcare professional communications at a rate that begins to function as a parallel regulatory channel — one that bypasses the standard label variation timeline precisely because the standard timeline is too slow for severe, low-frequency signals in therapeutic areas where patients and prescribers make daily clinical decisions. Sponsors who have not rebuilt their pharmacovigilance plans around that operating assumption will find themselves in reactive amendment mode, fielding IRB inquiries about why their investigational protocol’s safety monitoring section does not reflect the regulatory posture their own approved compounds are generating in the post-market environment. The liver is the first organ signaling that the old model is failing. It will not be the last.

References

1. European Medicines Agency — “Meeting highlights from the Pharmacovigilance Risk Assessment Committee (PRAC) 7–10 April 2026” (April 2026)
2. European Medicines Agency — Regulation (EU) No 1235/2010 of the European Parliament and of the Council on pharmacovigilance of medicinal products for human use
3. European Medicines Agency — Guideline on Good Pharmacovigilance Practices (GVP) Module VII: Periodic Safety Update Report (Revised, 2023)
4. U.S. Food and Drug Administration — Guidance for Industry: Safety Reporting Requirements for INDs and BA/BE Studies (2012)
5. U.S. Food and Drug Administration — Discussion Paper: Using Real-World Data and Real-World Evidence to Support Regulatory Decision-Making for Drug Safety (2023)
6. SK Life Science / Angelini Pharma — Cenobamate (Ontozry/Xcopri) Phase 3 pivotal trial program, focal-onset seizures in adults
7. European Medicines Agency — EudraVigilance pharmacovigilance database, signal detection and management framework