Today the tabloids in Sweden announced that a new study has found that LCHF (Low Carb, High Fat diet, Atkins, paleo diet etc) can increase risk of stroke. There has been lots of debates about these diets recently, so this was on the news bills. I took the time to read the original scientific paper and found several problems with the news reports:
The connection LCHF-stroke
Even the heading of the original paper mentions high-protein/low carb-diet, but what the researchers actually did was to send a questionnaire to women in a part of Sweden in 1991-92 and then measure the ”outcome” in terms of cardiovascular diseases. They found which of the women that had any such diseases, and used statistical methods to measure if there was any relationship between the variables diet and subsequent disease.
As far as I know there were no talk about Atkins or LCHF at that time in Sweden, I would estimate that approximately 0 % of the women in the sample followed any diet similar to LCHF.
This makes the conclusions dubious at best. As far as my understanding goes the study cannot say very much about the consequences of LCHF or similar diets.
The link high protein-stroke
But, you might say, there was after all a strong statistic correlation between eating little carbohydrates and lots of protein and cardiovascular disease. The problem is that this correlation might be a sign of another confounding variable. To illustrate this in a simple way, let’s say that there is a big ”normal” cluster of women in the sample eating some sort of average food, and then a smaller ”vegetarian” cluster, eating less protein and more vegetables, and then a ”hedonistic” cluster eating more fat, meat, sauces with cream in etc. Then, with my knowledge of the situation in Sweden at that time, I would guess that the ”vegetarian” cluster was most ambitious with exercise and drank the least alcohol and smoke less. And then ”hedonistic” women probably behaved the opposite, smoked more, drank more and exercised less, on average. (I should stress here that they have used a statistical method to standardise the ratios of carbohydrates/fat/protein, so it makes sense to assume that these clusters will really be visible in the study)
The study mentions taking some account for this, looking at smoking and alcohol use. But it does this with variables that are very coarse, dividing the sample into only three groups concerning alcohol intake. And they don’t take exercising into account at all.
This makes it quite possible that what they are really measuring is a correlation between protein/carbohydrate intake and general lifestyle variables like smoking and exercise.
What does this has to do with the themes of this blog?
This subject is not really related to the themes of this blog: psychology, parenting and psychotherapy. But I think it shows how scientific reports in both reputed journals and in mainstream media can be flawed and biased.
Another more important reason that these things might be of interest to the readers of this blog is that as I wrote when reviewing the Paleo-diet-site Marks Daily Apple (MDA), there are new and very interesting research coming about a connection between inflammatory processes in the body and at the same time there are signs that paleo-diet (a diet with similarities to LCHF) can decrease those inflammatory processes. And on MDA there are actually quite interesting reports about depressed people recovering using that diet. There are also other reports saying that a ketogenic diet (with little carbohydrates) can be helpful for depression. I absolutely don’t want to claim that there is any sort of scientific clarity on this, I just want to say that we should keep our eyes open and curious towards this area, and look at different hypotheses with an open mind.
Note 1: For clarity I would like to stress that I do not mix dieting advice into my practice as a psychologist, rather I very much urge people with eating issues to first look at the psychological aspect of their eating.
Note 2: The study is: Low carbohydrate-high protein diet and incidence of cardiovascular diseases in Swedish women: prospective cohort study by Lagiou, Sandin, Lof (Löf), Trichopoulos, Adami, Weiderpass in British Medical Journal.