You may have seen the recent media storm about coconut oil. “Coconut oil: are the health benefits a big fat lie?” (the Guardian). “Coconut oil ‘as unhealthy as beef fat or butter’.” (the BBC). “Coconut oil isn’t healthy. It’s never been healthy.” (USA Today – yes, I’m an avid follower). My visceral reaction to such headlines is to spin into a uncontrollable rage about how the media intentionally misreport science to attract readers. And that’s what I did this time. Until I read the source material. While predictably sensationalised, the media interpretation, as it turns out, wasn’t far off. In June, the American Heart Association (AHA) published a paper damning saturated fat (again) and advising “against the use of coconut oil.” Recommendations from the AHA carry a lot of weight. As do those from similar institutions both sides of the Atlantic. We trust them. But when it comes to nutrition, we probably shouldn’t.

The AHA paper, “Dietary Fat and Cardiovascular Disease: A Presidential Advisory from the American Heart Association”, is not about coconut oil. It’s about saturated fat. And it’s an attempt by the AHA to reinforce its 1961 diktat that saturated fat causes heart disease, and we should all switch to vegetable oil. Coconut oil merely has a cameo role here. But the paper is damning nonetheless. Coconut oil is high in saturated fat, the story goes, so it must cause heart disease. Hmm, not exactly. Because saturated fat doesn’t cause heart disease. And, in fact, the opposite may be true.

To investigate the effects on cardiovascular disease (CVD) of replacing saturated fat with polyunsaturated (omega-6) vegetable oil, the AHA paper reviews data from randomised controlled trials (RCTs), prospective observational studies, and monkey experiments. And attempts to back it all up with some mechanistic physiology. So far so good. But the reliability of this sort of research relies on an objective, unbiased appraisal of the evidence. Something the AHA fails to deliver.

Most of what follows is based on a superb analysis of the AHA paper by Chris Masterjohn. If you’re interested in the links between nutrition and health, you should follow Chris’ work.

10 RCTs have investigated the effect on CVD of replacing saturated fat with vegetable oil. The AHA research team included only four. And all four just so happen to show a reduction in CVD (30% when combined). Trials showing an increase in CVD, or no effect, were left out. Already I’m suspicious. To be fair, the inclusion criteria look reasonably sensible, but the AHA team didn’t even apply their own criteria consistently.

The Finnish Mental Hospital Study is the cornerstone of the AHA paper, with a 40% reduction in CVD. But it was not a randomised controlled trial – the subjects were not randomised to receive one of the two diets. So, it should not be included as one. The lack of randomisation meant that usage of an antipsychotic drug (which has since been shown to cause CVD) was higher in the saturated fat group. So, the higher incidence of CVD may have been due to the drug, not the saturated fat. The results are invalid. The Oslo Diet-Heart Study was properly randomised. But the vegetable oil diet also included more vegetables, fruit and oily fish. And less margarine (trans fats). These additional (confounding) factors are all protective against CVD. So, it’s impossible to conclude that the 29% reduction in CVD was due to the vegetable oil. “Mixed dietary intervention” was one of the AHA exclusion criteria. And other trials were excluded on that basis. Why not this one? The LA Veterans Administration Hospital Study should be included. But not because of the 20% reduction in CVD. That should be ignored. Because the saturated fat group had way more smokers, and we know smoking causes CVD (confounding again). This trial should be included because it showed a higher incidence of cancer in the vegetable oil group. Despite that group having far fewer smokers. This result is not mentioned in the AHA paper. Anywhere. The fourth included trial, done by the Medical Research Council, found an 18% reduction in CVD. It should be included but interpreted with caution. Firstly, the results were not statistically significant (which means the reduction in CVD could have been due to random chance). And secondly, the vegetable oil used was soybean oil. Which has a higher ratio of omega-3 to omega-6 fats than other vegetable oils, and an unusually high concentration of Co-enzyme Q10. Both of these attributes may be protective against CVD (potential confounding).

The most notable of the six excluded trials is the Sydney Diet Heart Study. It showed a 33% increase in CVD deaths in the vegetable oil group (borderline significant). The AHA team excluded the trial because margarine was a major component of the vegetable oil diet. Margarine contains trans fats, which we now know cause CVD. Fair enough. But, the degree of confounding was unknown, and probably small, because a soft margarine was used (less trans fats than conventional margarine), and the saturated fat group continued to eat conventional margarine. Still, it may have increased CVD in the vegetable oil group, which doesn’t help the AHA’s case. So, the trial was excluded. Which is interesting. Because the LA Veterans Administration Hospital Study also contained some unknown degree of confounding from trans fats. But this time in the saturated fat group (hydrogenated oils were allowed). This may have increased CVD in the saturated fat group, which supports the AHA’s case. So, it was included. Egregious double standards.

Had the Finnish Mental Hospital Study and Oslo Diet-Heart Study been excluded. Which they should have been. And had the CVD results in the LA Veterans Administration Hospital Study been ignored and cancer results highlighted. Which they should have been. And had the Sydney Diet Heart Study been included. Which it should have been. Saturated fat may start to look like the healthier option.

The prospective observational studies, when combined, showed that replacing saturated fat with vegetable oil was associated with a 25% lower risk of CVD. But, we need to be careful. The problem here is that all the studies were carried out in the decades after the AHA’s original 1961 advice. Throughout this time period, the connection between heart health and lifestyle was increasingly embraced by the public. Many people switched to vegetable oil. But those same people also adopted a myriad other heart healthy habits. They stopped smoking. Did more exercise. And ate less junk food. The studies are laughably confounded. It’s impossible to see what effect the different fats had on CVD risk.

The monkey experiments are equally unhelpful. They do show a link between saturated fat intake and CVD. But in these trials, blood cholesterol was elevated to extreme, unrealistic levels. And this was in monkeys. Not humans. It really is a stretch to translate this to the general human population. Indeed, the AHA admits that, “Generalisation of these studies is limited”.

The AHA paper also discusses the mechanism by which replacing saturated with polyunsaturated fat reduces CVD. It’s all to do with cholesterol. Consuming polyunsaturated fat reduces LDL (‘bad’) cholesterol (true). And lower LDL cholesterol is associated with lower CVD risk (OK). But, the paper completely omits the key step in the formation of arterial plaques. Cholesterol, like other fats, is transported around the body in lipoproteins. Low density lipoproteins (LDL) carry it from the liver to the tissues. High density lipoproteins (HDL) carry it back to the liver. It’s the LDL particles that contribute to arterial plaques. But the fatty acids making up the membrane of the LDL particles must first be oxidised. And only polyunsaturated fatty acids are vulnerable to oxidation. The more polyunsaturated fats we eat, the higher the proportion of polyunsaturated fatty acids in the membranes of the LDL, and the more likely the LDL are to be oxidised. So, consuming more vegetable oil simultaneously reduces the risk of CVD (by decreasing LDL in the blood) and increases it (by making the LDL more vulnerable to oxidation). This offsetting mechanism is why the evidence on CVD outcomes appears to be conflicting. But the AHA paper makes no reference to it at all. LDL needs to be oxidised to cause CVD. Saturated fat may actually protect against CVD, because it cannot be oxidised.

Finally we get to coconut oil. The AHA paper admits that “clinical trials that compared direct effects on CVD of coconut oil and other dietary oils have not been reported.” But then advises against using it. The justification? Coconut oil has the highest saturated fat content of any edible oil (true). And saturated fat causes CVD (not true, as we’ve just seen). Coconut oil raises LDL cholesterol (true). Which causes CVD (not quite – remember, the high saturated fat content protects against LDL oxidation, which may protect against CVD). The paper acknowledges the fact that coconut oil raises HDL (‘good’) cholesterol more than it does LDL cholesterol. Raised HDL is associated with lower CVD risk. So, this may completely offset the negative effects of the raised LDL. But the authors decide that, “the LDL cholesterol-raising effect should be considered on its own.” This is completely indefensible.

The coconut oil section makes no reference to studies that investigated the incidence of CVD among populations eating large amounts of coconut. The people of Kitava eat a relatively low fat diet, but get 17% of calories from saturated fat (coconut). A lot more than the 12.6% in the UK. Any CVD on Kitava? Zero. Not one person. The island of Tokelau is even more interesting. The people there get more than 50% of their calories from saturated fat (coconut). That’s around four times the saturated fat intake of the UK. Incidence of CVD? None. No sign of CVD whatsoever. And it’s not due to genetics. Tokelau Islanders who moved to mainland New Zealand ate half as much saturated fat, but suddenly started to develop CVD. Now, of course, Pacific Island populations will have generally healthier lifestyles than we do. More sun. Less stress. Less junk food. But if coconut oil is so individually dangerous, it’s hard to believe that these other factors would completely reverse the negative effects. It’s also worth highlighting the other health benefits of coconut oil here. Seen as the AHA paper makes no mention of them. Coconut oil is potently antibacterial, antiviral and anti fungal. So, it may be useful in treating certain infections. And the medium chain triglycerides (which make up 15% of coconut fat) have been shown to help with weight loss, by simultaneously increasing energy expenditure and suppressing appetite. You can’t just ignore this stuff.

So, there we have it. The AHA’s case against saturated fat is not supported by the evidence. At best, replacing saturated fat with vegetable oil has no effect on CVD, and it may actually increase the risk. It almost certainly promotes cancer development. More potently than smoking. Coconut oil is not going to kill you. And it may well protect you. We will be continuing our daily consumption of coconut oil, and other saturated fats. Vegetable oil? Never.

But coconut oil is a sideshow here. This is about being deceived by the institutions we should be able to trust. The researchers at the AHA are not stupid. They knew exactly what they were doing. They had to reinforce the 1961 recommendation, no matter what the evidence said. Replacing saturated fat with vegetable oil has been at the core of official nutrition guidance in the developed world for nearly 60 years. Reversing it would come at an enormous cost. Credibility would be lost. Reputations ruined. Don’t underestimate the strength of that sort of vested interest. And if the AHA is doing it, what about everyone else? What about nutrition recommendations from the NHS and the British Heart Foundation? From Diabetes UK and Cancer Research? Proceed with caution.

Own your health.

– James.