Another thing to keep in mind in this context: Publishing studies is not mandatory. It very well could be the case that a company keeps ordering studies while not publishing them when they get a result they don't like. Eventually randomness / chance will seem to show a statistically significant result. We'd have to know how many studies they did as compared to how many they published.

Not directly related to this particular question, but just to point out how truly flawed nutrional studies can be: Some of the famous "studies" (IIRC some of Harvard's Walter Willet's studies are of this kind) purporting to show that red meat causes various kinds of disease count things like macdonalds, pizza, lasagna etc as "red meat" because some parts of those foods may consist of meat. Sloppy and misleading "studies" of this kind are commonplace in nutrirtion studies. Unstated asasumptions (e.g. of all the things in there the meat surely is the ingredient causing the problems) and dearth of controls (e.g. healthy user bias. mgiht the people avoiding the junk food be doing other healthy behaviours which are responsible for better health outcomes?) seems to be the norm rather than the exception in these so-called studies which might more appropriately be called propaganda pieces. It really largely seems to be a junk science / pseudo-science field, especially nutritional epedemiology.

Poorly controlled correlation studies and + spurious theoretical mechanism does not a sturdy reliable scientific argument make.,