There's nothing capable of showing causal effects.

All of that data supports the current toxicity metrics.

What we tend to see are underpowered studies with little replication...if any, non-standard techniques, unsubstantiated deviations from established protocols, and of course, passing off of molecular fishing expositions, as being able to accurately determine treatment effects.

Consider that the OP posted a study where the authors state that their results shouldn't be extrapolated to represent normal pregnancies.

Having the entire study population comprised of high risk pregnancies is a major issue.