Cholesterol: LDL in Oslo

There is a group of doctors in Oslo who run a coronary care unit and, as part of their day to day work, they do quite a bit of coronary arteriography. They put in the dye and look at the arteries. It's part of their job. I assume they are quite good at it.

They did something very, very strange and then wrote a letter to the editor of the JACC about it, presumably because they couldn't get it published in any other way.

This is what they did.

They simply selected sequential patients with LDL cholesterol scores below 2.7mmol/l. In the old terms that so much cardiac work is published in I think this means an LDL of below 100mg/dl. Quite why anyone with an LDL this low would need a coronary arteriogram is a good question to ask anyone who believes in the lipid hypothesis.

They ignored all people with LDL concentrations from 2.7 to 4.5mmol/l but did enroll all people with an LDL >4.5mmol/l, that is above about 180mg/dl.

So they then had two groups of people, those at catastrophic risk of LDL-blocked-arteries and those with so little LDL they couldn't stick a tail to a donkey, or absorb a bacterial toxin, with it.

They did the scheduled angiography and checked how many patients had >70% blockage of at least two coronary arteries in each group.

Guess what: LDL cholesterol doesn't matter. They recruited 47 patients with low LDL-C, of whom 21 had significant CAD. They got 46 high LDL-C patients, of whom 24 turned out to have CAD.

I know that this is a calculated LDL value, not a particle number etc etc etc but this calculated guestimate is the basis of the lipid hypothesis in its current persona, until it gets its next adhoc makeover.

To a cholesterol sceptic that's pretty much what you would expect, serious heart disease affects people pretty well independently of LDL value and no one should be surprised at this. There is nothing strange in the study so far.

The really weird thing that the Oslo group did was this, wait for it:

They went looking for what might really cause heart disease!


Where do you start looking in a wide open field like this? It's almost like having virgin soil to plough...

It turns out that it's pretty easy to differentiate the groups with heart disease from those without. Here is a list of things which are significantly different between the patient groups with heart disease and those without it, in no particular order:

Ability of arterioles to vasodilate in response to applied acetylcholine

Level of von Willebrand factor (a platelet adhesion factor)

Level of hsCRP


BTW: Remember the JUPITER fiasco? Elevated hsCRP in the Oslo study marked out the groups with at least two severely stenosed coronary arteries, irrespective of LDL level. JUPITER subjects with elevated hsCRP would fit in to the low LDL-C with severe CHD grouping in this study. That is; they quite probably had CHD. Back to Oslo:

Level of TNF alpha

There is an inverse effect of interleukin 10 (it's anti inflammatory)

Enhanced platelet activation as assessed by soluble CD40 ligand

Levels of endothelial and platelet activation as assessed by soluble P selectin

Blood flow response to sodium nitroprusside couldn't distinguish CHD groups from non CHD groups any more than LDL-C could.

Exactly what all of these things mean at the molecular level is not too important, though I'd bet a fair few of them are controlled by NFkappaB (hence by hyperglycaemia and hyperinsulinaemia). What matters to me is that (a) LDL cholesterol doesn't matter and (b) there are at least eight researchers in the cardiology community who are looking for the cause of CHD.

Thank goodness someone is. Good luck to them.

Peter