Its not a hard analysis to do but then they wouldn't be able to publish as easily. The scientific process is not about progress, its about publish or perish. D is perfect for this type of non-sense science as it is correlated with so many things and each one can get dozens of papers out of it.
I wrote this whole long comment talking about all the issues in trying to design a good trial to figure this out, but here's one trial I think will give us some more info, it seems better designed than the other ones I've seen (sample size remains a problem):
It can’t be that hard right? It’s stuff that you can buy over the counter and that you’d have to use crazy amounts of to overdose.