You can't base any conclusions on observational studies. Their only purpose is to highlight areas that may be worthy of randomised controlled studies. However, long term randomised controlled studies are simply impossible to conduct on diet, especially considering most changes that we suspect may be due to diet are due to diet over the long term. For instance, I attempted (and frequently failed) to follow the advised low fat calorie counted diet promoted by the various agencies for, maybe 35 or more years. Prior to that I ate what my mother cooked, I used to eat pork chops with a nice thick layer of fat, and crispy crackling, with a couple of potatoes and green veggies. Or maybe a slice of gammon, fried in lard, again with crispy fat, new potatoes and cauliflower. I remember the 'sauce' was the drippings from the pan, poured over. My mother used to say 'to grease your little engine!
Then, following the suddenly changed advice, I stopped having butter, Vitalite was the fat of choice, with vegetable oil or corn oil used to cook and latterly, Flora. Then, for the last 6 or so years, I have eaten only natural fats, butter, lard, fatty cuts of meat again, crispy skin on chicken.
So what will impact most on my health? The 20 years of good old fashioned cooking, the 35 years eating faux foods or the six years eating only real foods, as fresh as possible and only natural fats and very low carbs, of course.
Short RCTs have been tried on diet and whilst they do seem to show no disadvantage to the LCHF diet, they are difficult to randomise and impossible to blind. Certainly it is impossible to blind a diet that is full of fresh natural produce, so everybody knows which diet they are on. And the scientists analysing the results know which diet the participants are on. And if the two diets are equivalent in calorific value, it negates the 'benefit' that many of us find eating LCHF, in that it is a natural and subconscious appetite suppressant.
The Nurses Health study has thrown up anomalies with regard to drug usage. It was that study that seemed to indicate that HRT was heart protective. It was only when a randomised control trials were done (where people were carefully paired, such that the only difference was the HRT) and it was a double blinded, and lo and behold, HRT actually increases the risk of heart disease. What skewed the results that the women who chose to take HRT were more health conscious, were likely to be slimmer, were likely to exercise more, were likely to not smoke or to smoke less or to not drink or drink less. In exactly the same way it may be throwing up anomalies with regard to diet.
Millions and millions of pounds (or dollars) are spent every year on these observational studies. Thousands of able people must be employed undertaking this work. I am not at all convinced that this is not a phenomenal waste of the world's resources! Just eat real food, the stuff we've been eating for centuries and if you are diabetic or insulin resistant, also make it low carb. Job done!