Fezolinetant is a neurokinin-3 receptor antagonist licensed for vasomotor symptoms of menopause. An imbalance in neoplasm events in its phase 3 SKYLIGHT programme was raised in 2023 and treated by regulators as a chance finding without a plausible mechanism. This paper in Drug Safety, by two authors, re-examines that reasoning. It is a commentary and not a new analysis of patient-level data.
They report non-benign neoplasm rates of 0.2, 0.7 and 1.2 per 100 subject-years for placebo, 30 mg and 45 mg, a relative risk of 6.6 at the licensed dose (95% CI 0.9 to 51.3). The imbalance rises with dose but is not statistically significant. They also report that of 15 references used in the European regulatory review to argue against a mechanism, 11 concern the neurokinin-1 receptor or substance P and not the neurokinin-3 receptor, and propose four possible mechanisms. They call for structured post-authorisation surveillance and a periodic independent safety review.
The numbers are very small, with an interval that includes no effect. The paper does not show that fezolinetant causes cancer and the regulators have not changed their position.
- Non-benign neoplasms were 0.2, 0.7 and 1.2 per 100 subject-years for placebo, 30 mg and 45 mg.
- The relative risk at the licensed dose was 6.6 (95% CI 0.9 to 51.3), which is not statistically significant.
- The authors argue that most references used to dismiss a mechanism concern a different receptor.
- Regulators had set the finding aside as chance, and the authors call for post-authorisation surveillance.
- Keep screening up to date, report suspected neoplasms as adverse events, and follow the current label.
Why it matters
It challenges the reasoning, not the data, used to set aside a dose-related imbalance and shows why post-marketing surveillance exists.
Don't overread it
This is a commentary based on trial data that is not statistically significant, and regulators have not changed their assessment.
The statistics, in plain English
A relative risk of 6.6 with an interval from 0.9 to 51.3 means the data are consistent with almost no difference and also with a very large increase. With so few events, the estimate is unreliable in both directions, which is why the authors argue for surveillance and not a conclusion.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for clinical pharmacology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free