Thanks, Bob.

I just checked the links to the phenol explorer (on my desktop); both links work, the first one leading directly to caffeic acid, the second one to the database homepage. It also works on my iphone SE. So, maybe it's your connection or browser, that's acting up. But thanks for the heads-up.

You are pointing out an important issue: the self-reporting of nutrition data. The unavoidable inaccuracies plague all studies that use those devices. And thanks for the hint at Mendelian randomization. I have it on my radart since a while, just want to package it in a subject that serves to highlight its advantages and drawbacks.