In particular, table 2 summarizes what we know about each type of food (sodium, saturated fat, trans fat, carbs, sugar, ...) and what type of study we know it from (e.g., randomized trial or association study?): http://circ.ahajournals.org/content/123/24/2870/T2.expansion...)
The best type of evidence you could have is a randomized trial of clinical end points--divide people randomly into (say) low fat and low carb diets, run the study for a few years, and measure what percentage of people have a heart attack. That's column 4, and it's nearly empty. The only things we seem to know are that total fat seems to be neutral, and substituting polyunsaturated fat for saturated fat reduces heart attacks.
In the absence of that gold standard evidence, we're relying on weaker but more convenient types of study designs -- randomized trials that measure surrogate metrics (e.g., blood cholesterol and blood pressure), or various types of association studies.
As of now, those seem to be telling us that alcohol, dietary fiber, and unsaturated fats help our risk factors for a heart attack, and trans fats, sodium, carbs, and cholesterol make our risk factors worse. But those conclusions can easily change since they're based on a surrogate, and indeed, we've already seen the recommendation against eating cholesterol reversed: http://well.blogs.nytimes.com/2015/02/19/nutrition-panel-cal...