When is a high fat diet not a high fat diet? Stuff via Fanatic Cook

Here's an interesting paper brought my way by Dave Lull who browses Fanatic Cook, a place I'm not likely to visit in my travels. Though perhaps I should, to keep an eye on what the carbophiliac people are up to...

This paper is another in which a Western/Cafeteria diet (45% fat, 40% carbs) is described as high in fat and, wait for it, low in carbohydrate. Assuming 2400kcal per day, that's 1000kcal of carbs or 250g/d. You can appreciate the level of intellectual honesty here, so of course the temptation is to bin the paper. No, WAIT.

There's some interesting stuff. Never mind the shifts in cholesterol levels (it was a cross over study with a low-fat high-carb period), these are exactly what you would expect. The study was careful not to check the LDL particle sizes and numbers otherwise they might have found the small dense LDL which would probably have predominated on the HC period. That is along side the lower HDL figure they did report on the HC phase. They were also careful not to look at glycosylation levels (HbA1c would be useless in a short study like this, but fructosamine looks at two weeks of glycosylation and would have been interesting) to see what was happening in the "eating" world, rather than during a glucose tolerance test. You don't put this sort of control in to nutrition studies because it might not get you the answers you want. High fat diets DO tend to produce higher fasting glucose levels and can even produce higher fasting insulin levels, as here, but 24h AUC for insulin and glucose would be lower. Fructosamine would have shown this. That's how come I can sit here with an HbA1c of 4.4% and a fasting glucose of 5.5mmol/l, higher some days. No suggestion anywhere that Vidon et al or Fanatic Cook see things this way. I do. Shrug.

So, not my type of study. But look at this, this is what grabbed me:

"Lower cholesterol concentrations occurred despite a higher cholesterol synthesis rate (P less than 0.05) and higher HMG-CoA reductase mRNA concentrations (P less than 0.05). LDL receptor mRNA concentrations were unchanged, LRP mRNA concentrations were lower (P less than 0.01)"

The molecular techniques are a bit beyond me to comment on, so I'm just going to believe Vidon et al, that cholesterol synthesis went UP on the high carbohydrate phase diet, plasma levels went DOWN and gene expression for the LRP went DOWN.

Now, they were looking at HMG-CoA reductase in white blood cells, not the liver, but they seem to think the same thing happens in both places:

"Because synthesis and secretion in the plasma pool increased, it appears that the lower plasma cholesterol concentrations were related to increased removal from plasma. However, the mRNA concentrations of 2 main lipoprotein receptors that control the clearance of cholesterol from plasma were unchanged (LDL receptors) or decreased (LRP)"

Now I find this very interesting. More cholesterol produced, less is taken up, yet plasma level drops.

Here's an ad hoc hypothesis as good as anything in the paper:


Some receptor other than the LDLr or the LRP is taking up the LDL cholesterol.


Lets have a guess at which one, let's guess at the oxidised cholesterol receptor. There is at least one short term intervention study showing that small reductions in total fat calories in the diet increase the level of oxidised LDL in the blood, markedly. There was a 15% difference in the fat content of the two diets in Vidon's study. Oxidised LDL is eaten by macrophages, using the oxLDL receptor. They become foam cells in fatty streaks.

Perhaps the cholesterol levels drop as it's all now in foam cells on the arterial walls????

Never forget the American paradox. According the Vidon et al there must a sharp reversal of those benefits associated with increasing fat from 18% of calories through to 32% (p less than 0.001 for trend by quartiles) if you extend it to 45%. Duh.

Maybe, maybe not.

Peter