The FastDay Forum

http://www.ncbi.nlm.nih.gov/pubmed/22120496

And eggs:

https://www.drfuhrman.com/library/eggs_cancer.aspx

One thing John Briffa does very well in his book "Escape the diet trap" is describe the differences between epidemiological studies and intervention studies.

To paraphrase (hopefully correctly and completely):

Epidemiological studies look at relationships between things and can tell us if two things are associated but not that one is necessarily causing the other. As a necessarily inexact statistical science it cannot prove causality.

Eating meat and eggs "may cause ..." - the operative phrase - because these men shared these two common factors. Notice that the fact that they smoked may also be causal here, maybe even more important, just as easily as hundreds of other things they shared but not identified or studied in this group.

Intervention studies compare their outcomes to a control group not exposed to the intervention. Causality is best demonstrated here.

Group A ate eggs and meat and died from bla-blah at such-n-such a rate. Group B never ate eggs nor meat and lived X-time units longer. Seems likely that they are on to something actionable.

The point is to consider statistical studies as identifying potential causal factors - to narrow down the field of ideas to something that could be tested. And to consider intervention studies as having a much better chance of being able to prove some factor as actually causing the problem.

What ADFnFuel said! Remember to chant when reading nutrition studies: "association is not causation"

My view is that context is important such that a diet high in protein and carbs may increase the risk of cancer, but moderate intake of red meat, dairy, eggs etc will be less harmful in the context of a low to moderate carb intake (and of course, fasting). The research into mechanisms suggests this would be a feasible explanation.