>>512883527
in situ human biochemical nutritional data is probably the lowest quality data in the history of science. whats fine for one person, even healthy, could kill another.
for instance, one of the essential (not manufactured by the body, must be consumed) amino acids, phenyl alanine, is found commonly in chicken and beef products. rarely, people are diagnosed with a disease called โphenylketonuriaโ in which ANY consumption of phenylalanine will kill them fucking DEAD in no time.
then thereโs the matter of basal metabolic rate, personal satiety (subjective), etc. damn near fuckin impossible to isolate a specific food but standardize the rest of the food intake, over time, for a large enough sample to make population-wide inferences.
think about how certain foods affect health. people with epileptic conditions may find that cutting carbs entirely (leading to the body reorganizing its processes to prioritize fats as the primary energy source - the metabolites of that are known to reduce epileptic events when paired with a lack of glucose metabolites), whereas that dietary decision would wreck the system of someone who relies on glycogen stores (think runners, swimmers, etc).
body composition plays an outsized role in what the ideal macromolecular makeup of a diet should look like. take two people at 6โ3โ 260, one who works out daily and the other who is an oily bag of bones and fat. if both pursue a 200g protein daily intake, itll benefit the athletic one while probably half of that will just convert to sugar (gluconeogenesis) in the sedentary fat fuck.
tl;dr dont trust any studies that make population-wide inferences about the effect of common foods on the body