This makes the RAIL very volatile:
- 24 July 2023: 100 ng/day based on CPCA category 2
- 23 October 2023: 1500 ng/day based on EAT (following earlier acceptance of EAT by TGA and EMA)
- 26 July 2024: 100 ng/day based on CPCA category 2
It is unfortunate that no short motivation is listed (to allow assessment of future impact on other cases) or transitional limits are instated.
An EAT positive/equivocal motivation with no experimental root cause for the outcome would be very strange indeed.
I find it noteworthy that the EAT negative on N-nitroso-dorzolamide (structural resemblance and shared CPCA category) remains intact, possibly making the data to build the decision on very specific.
Overall the July publications from HC, TGA and EMA still give confidence in the endorsement of the EAT (the later TGA update leaves the EAT on nitroso-sertraline intact). (So far, EMA/HC/TGA congruence on limits was around 80%).
Could this be a reflection of in vivo taking priority over in vitro but regulators not being ready yet to deal with proposed AI derivatisation based on positive in vivo mutagenicity data?
(Though it seems to me exceptional to not be able to support AIs above 1500 ng/day, making a temporary go-back to 100 ng/day probably unnecessary burdensome, especially considering that also the CPCA explanatory paper of NITWG shows that the alpha-benzylic feature is overvalued. Maybe other regulators feel more comfortable with 1500 ng/day in case of an in vivo positive or equivocal EAT that is being investigated).
Or mutagenicity warnings from further in vitro studies (which can be overalarming as seen by the fluoxetine case)?
It is still early to see and explain a definite pattern, but I couldn’t help but notice that the current volatility in potency assignment (in literature and by regulators; by looking at new in vitro and in vivo data) seems to be bigger for NDSRIs for which the API itself is known to inhibit CYP enzymes with specifically an at least weak inhibitory effect on CYP2C19, so far one of the most commonly identified bioactivators for NDSRIs based on in vitro studies.
(Sertraline is known to inhibit several CYP enzymes, including a weak inhibitory effect on CYP2C19. This property is shared with fluoxetine for example. Dorzolamide is not known for its inhibitory effect on CYP2C19 I believe.)
Another property that NDSRIs with volatility in potency assignment often have, is the presence of aromatic rings. Aromatic hydroxylated metabolites can sometimes form reversible adducts with macromolecules, which limits the ultimate biological impact in vivo, but forms a possible explanation for in vitro mutagenicity should this be the case.