[EasyMark]

That first study is -tiny- study which is a good data point but hardly worth changing my diet over. I’ve seen plenty of studies saying that butter in moderate usage is just fine, and the war on saturated fats really should have been limited to hydrogenated oils/margarine

[KempyKolibri]

How about a pooled analysis of 350,000 participants suggesting that for every 5% energy in the form of saturated fat that’s replaced with PUFA (like you find in margarine), the risk of coronary mortality drops by 26%.

Surely that’s both a large enough cohort and a large enough effect size to change one’s diet?

https://www.sciencedirect.com/science/article/pii/S000291652...

[tsimionescu]

This study contradicts another study you were citing in this thread . This one says that replacing SFAs with carbohydrates is a net negative, you have to replace SFAs with PUFAs. The other study was saying that replacing SFAs with either carbohydrates or PUFAs is just as good.

It's almost as if all of these studies are looking at tiny effects that they can't adequately measure, and contradicting each other.

The other study I'm mentioning :

https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD...

[KempyKolibri]

Do you believe there’s a difference in health outcomes between consumption of whole grain carbohydrates and refined carbohydrates?

If yes, do you believe it would be expected to see heterogenous outcomes in studies that don’t disambiguate whole grain and refined carbohydrates when replacing SFA?

If yes, then there’s clearly no contradiction in the above studies. If no for any of the above, I’d love to hear the argument.

[simiones]

I have no idea. I understand there are some a priori reasons to believe whole grains have certain health benefits. From what I quickly found in some basic searches, some studies find an effect, some don't. Those that do are typically population studies, which are often confounded by the correlation-vs-causation issue (are people that eat more whole grain healthier, or are people who live healthier lifestyles in general more likely to also eat whole grain?). Those that don't are typically RCTs, that suffer from the short duration and are unable to capture longer term effects, which are very likely with nutrition.

Also, just as I was mentioning in other comments, I think there is a good chance this reduction of the problem to just whole grain - refined grain is unlikely to tell the full picture. I don't see a priori why eating whole wheat would be exactly as healthy/unhealthy as eating whole rice, or oats, or millet, or barley, or quinoa or any of the many other unrelated plants we call "grains". Maybe we should prefer certain grains and avoid others, regardless of the whole/refined distinction; this difference might also depend on genetic factors, with certain populations perhaps being better suited to certain grains than others. It is very much possible as well that certain grains are better eaten whole, and certain others better eaten refined, say if there are substances in certain husks that are problematic over long time or in certain quantities and so on.

And this is not even going into other factors, like rates of contamination of the grains with pesticides/fertilizers/naturally-occuring substances in certain soils; handling, washing, and preservation; cooking differences; and probably many others that I'm not even thinking of.

And while some of these effects will naturally lead to heterogenous outcomes in studies that don't control for them, this doesn't increase my confidence in those studies. The fact that there are an extreme number of possible confounding variables in everything to do with nutrition is basically why nutrition science is almost hopeless as an entire endeavour: we can only reliably find extremely strong effects ("lack of vitamin C causes scurvy"), and even then we need a bit of luck. The rest is built on a house of cards: every new medical or biological discovery tends to upend nutritional studies and what they control for.

[KempyKolibri]

Ok, then if you have no idea then clearly there’s room for heterogeneity in studies that pool those different types. So there’s no contradiction in the studies I posted, which is the original claim you made.