Thanks,
@Indy51, for posting the comment by Bret Scher. These were also some of my initial thoughts.
In my mind, there are many weaknesses to the study.
First, as
@bulkbiker and others pointed out, this was not really low-carb and especially not ketogenic -- rather most likely higher carb, higher fat, which we know is a problem. However, trust them to extrapolate from 45% carbs to 5% carbs (similar to the famous Havard study linking low-carb to increased mortality (again low-carb here was in the range of 40% of calories from carbs) and come to the conclusion that if 45% carbs increase a-fib than 5% carbs must increase it even more.
Second, the use of the food frequency questionnaire. I don't think people can accurately remember what they ate even yesterday. At least in my personal experience when posting in the "What did you eat?" thread on this forum, I have trouble remembering.
Third, the hazard ratio was 1.18, which is miniscule -- probably close to the hazard ratio of eating red meat. In an observational study, this probably shouldn't even be interpreted. I think the Bradford-Hill criteria suggest a hazard ratio of at least 2 to suggest causation.
I wonder why journalists who write about science don't need any scientific training? Are they really doing their job if they don't understand the subject matter on which they report?
How many T2s will be kept from trying a low carb woe by articles like this, and might eventually go on to suffer from some of the complications of diabetes? Who accepts responsibility for this?
Sorry for this rant, I'll step of my soapbox now.