Correlation is not causation, especially not in nutrition where it's so hard to control for hidden factors. There exist randomized controlled trials that show the exact opposite of this association, see https://nutritionj.biomedcentral.com/articles/10.1186/1475-2.... I consider these RCTs to be much higher quality evidence than other types of study designs such as the in the OP.