A reader recently sent me a reference to an interesting paper titled "Dietary Fat Saturation Effects on Low-density-lipoprotein Concentrations and Metabolism in Various Animal Models". It's a review of animal studies that have looked at the effect of different fats on LDL concentration as of 1997. They nail their colors to the mast in the first sentence of the abstract:
Saturated vegetable oils (coconut, palm, and palm kernel oil) and fats (butter and lard) are hypercholesterolemic [raise cholesterol] relative to monounsaturated and polyunsaturated vegetable oils.But don't let this fool you; the actual data they present are much more interesting. First of all, they expressly exclude studies on models that have an "abnormal degree of response to a hypercholesterolemic diet". In other words, they attempt to create a self-fulfilling prophecy by excluding models that don't support their hypothesis. Even after stacking the deck, the data they present still fail to support their position.
When an investigator wants to study diet-induced atherosclerosis, first he selects a species that's susceptible to it. These are generally herbivorous or nearly herbivorous species such as rabbits, guinea pigs, hamsters, and several species of monkey. Then, he feeds it an "atherogenic diet". This is typically a combination of 0.1 to 1% cholesterol by weight, plus 20-40% of calories as fat. The fat can come from a variety of sources, but animal fats or saturated vegetable fats are typical. The remainder of the diet is processed grains, vitamin and mineral supplements, and often casein for protein.
Let's put that amount of cholesterol into human context. Assuming the average person eats about 2 pounds dry weight of food per day, 0.5% cholesterol would be 4.5 grams. That's the equivalent of:
- 17.5 pounds of beef steak, or
- 3.8 pounds of beef liver, or
- 22.5 eggs
It's important to remember that this hamster experiment has little to do with the situation in humans. No one is claiming that reducing saturated fat and cholesterol will reduce a human's LDL by 7-fold. Long-term dietary interventions that reduce SFA and dietary cholesterol without increasing PUFA have little to no effect on LDL cholesterol, and can in fact increase LDL oxidation. Furthermore, humans are very resistant to blood cholesterol changes in response to dietary cholesterol, suggesting that we have an evolutionary history metabolizing it. Finally, as I've discussed in a previous post, saturated fat does not influence total blood cholesterol or LDL in humans in the long term, and the effects are modest even in the short term.
But let's get back to the animal models. The hypothesis the paper is attempting to support is that saturated fat raises LDL in a variety of (herbivorous) animal models. If that were true, it should be able to raise LDL even in the absence of added cholesterol. So let's consider only the studies that didn't add extra cholesterol to the diets. And if saturated fat raises LDL, it should also do it relative to monounsaturated fat (MUFA- like olive oil), rather than only in comparison to PUFA. So let's narrow the studies further to those that compared SFA-rich fats, MUFA-rich fats and PUFA-rich fats. In Fernandez et al. (1989), investigators fed guinea pigs 35% of calories from corn oil (PUFA), olive oil (MUFA) or lard (MUFA-SFA). Here's what their LDL looked like:
The same investigators published two more studies showing similar results over the next five years. The next study was published by Khosla et al. in 1992. They fed cebus and rhesus monkeys cholesterol-free diets containing 40% of calories from safflower oil (PUFA), high-oleic safflower oil (MUFA) or palm oil (SFA-MUFA). How was their LDL?
None of the differences were statistically significant. Khosla and colleagues published another study with the same result in 1993. This is hardly supportive of the idea that saturated fat raises LDL in animal models. The most you can say is that PUFA lowers LDL in some, but not all, species. There is no indication from these studies that SFA raises LDL in the absence of excessive dietary cholesterol. I didn't cherry pick studies here, I mentioned every study in the review paper that met my two criteria of no added cholesterol and a MUFA comparison group. The bottom line is that experimental models of atherosclerosis rely on overloading herbivorous species with dietary cholesterol that they are not equipped to clear from their bodies. SFA does exacerbate the increase in LDL caused by cholesterol overload. But in the absence of excess cholesterol, it does not generally raise LDL even in species ill-equipped to digest these types of fats. Dietary cholesterol has little or no influence on LDL in humans. So there is no cholesterol overload for saturated fat to exacerbate. Consistent with this, saturated fat does not influence LDL in humans in the long term. This is contrary to the mainstream consensus, but is an inevitable conclusion if you carefully consider the evidence from controlled trials and observational studies.
PUFA vegetable oils do lower LDL in humans, and the effect appears to persist for at least a few years. But this is a Pyrrhic victory, as omega-6 PUFA increase LDL oxidation and exacerbate chronic inflammatory processes. Vegetable oils are not a solution to the coronary heart disease epidemic, to the contrary.