Cheers 😂

Going for top read 2025 too

#episky #ispe #pharmacoepi

#episky #sibsky

@danielberglind.bsky.social @pmadleydowd.bsky.social and more @ki.se @au.dk

Thank you, Magda! 🎉🎉 This will enable even more synergies between us!

Thanks mate ! Would not have been possible without your help and support!

Hey hey hey. All for the CV mate

Loool! Perhaps they are relying on scholar rather than their own publisher link? 😂

as would larger cohorts with deep phenotyping, ideally including genetics; I am deeply concerned about genetic confounding, considering the known genetic overlap between both psychiatric and epileptic indications and neurodevelopmental outcomes!

And yes, I definitely agree that there is likely residual confounding despite controlling for observed indications and SES. Better data on siblings would be incredibly useful, ...

Yes, they were independent of the original safety signals. Lamotrigine has been estimated to not be associated with malformations or neurodevelopmental outcomes in a series of cohorts.

It’s all quite challenging. However, we believe we’ve identified some consistent signals for valproate (expected), topiramate (verification of prior findings), and carbamazepine (novel). Whether these hold up under further causal scrutiny remains to be seen...!

We tried to adjust for many of these imbalances (inc bipolar indication), but the coverage & quality of the registry data have limitations. This is partly why sibling comparisons are valuable, assuming indications remain somewhat stable across pregnancies (this may not hold for episodic conditions).

However, it’s not ideal. As you noted, lamotrigine is also used for bipolar (and on/off-label elsewhere). Also, individuals with more 'severe indications' (e.g., generalized epilepsies) are often treated with e.g. topiramate, introducing differential indication severity even within antiseizure meds.

Re Lamotrigine: It is first-line treatment for epilepsy and is generally considered the safest option during pregnancy (not linked to malformations). Using it as a reference, rather than ‘no use,’ has clear advantages – and this approach has been used in the teratology literature.

Thanks for the interest, @sebatlab.bsky.social ! Agree that the sibling analysis shows (much) attenuation. However, we’re cautious about overinterpreting the statistical null given our limited power. Still, we believe this highlights the value of such analyses and the promise as samples grow larger.