I think what he's asking for is long term epidemiological studies, which are hard to come by. I'd have to see the studies you are referring to, but most likely they refer to measurements occurring over a relatively short time period, on a relatively small sample size (or in a lab on animals, or just cells, etc. etc.). This stuff is totally valid, but a lot of it is really hard to detect at a large scale level. You can justify many kinds of diets by looking at the short term biological impact of them, but the reality is that if these diets are so much better it shouldn't be so difficult to detect them in epidemiological studies. The classic example is the widely misinterpreted "China Study" which many people use to justify a vegetarian diet, but in fact did the opposite by proving that the effect size of eating red meat, while statistically significant, was literally 2 orders of magnitude below what would be considered clinically relevant at an epidemiological scale.