The paper starts off by attacking the idea that saturated fat consumption raises blood cholesterol levels which in turn raises cardiovascular risk. This hypothesis does not have empirical support, they argue, but you can formulate your own opinion by going through the references the authors cite.

They then go on to take down the idea that LDL-C causes heart disease. They go into the history of why this is thought to be, starting in 1985 where it was found that people with Familial Hypocholerestemia (FH), and the elevated LDL-C that comes with it, did tend to die prematurely of cardiovascular disease (CVD). But there are problems with assuming that this link is causal.

Elderly people with FH don’t die of heart disease at a higher rate than normal people, however, which calls into question whether exposure to LDL-C over long time periods really is the problem. Overall, people with FH also don’t have higher all-cause mortality than the general population. In fact, they may even have lower all-cause mortalities, suggesting there are important protective effects of LDL-C.

LDL-C is also a relatively poor marker of cardiovascular risk, argue the authors. They put forth a number of factors that are superior in their predictive power of CVD, including their favourite which appears to be Coronary Artery Calcium scoring.

There are also different types of LDL-C, and by generalizing on that marker we are missing a lot of context. Small, dense LDL particles do show correlation with CVD while large, less dense LDL particles appear to be harmless.

Insulin resistance is the real risk factor that we know of that causes CVD, the authors argue.

The authors then go into some detail on the low-carb diet, and how various relevant markers change when a patient uses this diet. They go on to show that people with biomarkers consistent with LCDs show insignificant gains from statins in controlled trials, providing evidence that people on LCDs should not take statins.