http://youtu.be/fuj6nxCDBZ0

This is something I hadn't heard of before. At 56:30 he talks about a study showing that high Apo B particle numbers hardly raise risk in those with fasting insulin below 12. Over 12, particle number matters... a lot! This would explain the Jimmy Moores of the world with high particle number yet still no disease.

I didn't listen to it. Too lazy. But I can comment on that insulin bit. Taubes talked about some researcher who forgot to shut the insulin drip in a dog's leg, then discovered later that leg had developed extensive atherosclerosis while the other leg was fine. So, if there's an association between insulin level and disease state, it's simply because insulin itself causes the disease, i.e. it directly alters particle size making them more "atherogenic". This is a mischaracterization of lipoprotein particles, since the atherogenic property belongs to insulin, not to the lipoprotein particles, i.e. causality. Lipoproteins are merely the victims of the atherogenic action of insulin. When we combine this with the fact that hyperglycemia causes damage on its own too, we get a double whammy with dietary carbs. It follows that the higher the particle number (the amount as opposed to the size), then the greater the potential number of particles which insulin can act upon, i.e. the more targets the better the chances to hit any 1 of them.

If I'm not mistaken, insulin acts on the production of lipoproteins rather than on lipoproteins directly in the blood, specifically with respect to particle size. Somebody correct me if I'm wrong about that. But anyway, so the more appropriate idea is that the higher insulin is, the greater the chances that smaller - rather than bigger - particles will be produced.

No wait, I think I got it wrong. The insulin drip/dog's leg thing is a local effect.

It's an interesting association, but what's the causation?