The American Journal of Clinical Nutrition just published the results of a major Japanese study on saturated fat intake and cardiovascular disease (1). Investigators measured dietary habits, then followed 58,453 men and women for 14.1 years. They found that people who ate the most saturated fat had the same heart attack risk as those who ate the least*. Furthermore, people who ate the most saturated fat had a lower risk of stroke than those who ate the least. It's notable that stroke is a larger public health threat in Japan than heart attacks.
This is broadly consistent with the rest of the observational studies examining saturated fat intake and cardiovascular disease risk. A recent review paper by Dr. Ronald Krauss's group summed up what is obvious to any unbiased person who is familiar with the literature, that saturated fat consumption doesn't associate with heart attack risk (2). In a series of editorials, some of his colleagues attempted to discredit and intimidate him after its publication (3, 4). No meta-analysis is perfect, but their criticisms were largely unfounded (5, 6).
*Actually, people who ate the most saturated fat had a lower risk but it wasn't statistically significant.
17 comments:
The japanese have very low obesity rates, so this is probably more applicable to those of us who are at a healthy weight than to the population as a whole.
At the high end, the Japanese in the study were consuming about 20g/d of sat fat - prob. at the low end for most Americans. Also, sat fat consumption was highly correlated with wealth, fruit and veggie consumption, and a bunch of other factors so there is probably residual confounding.
I hadn't looked at August's AJCN before this morning. Much more entertaining than usual! Not quite as entertaining as the back and forth between de Lorgeril and Ridker about JUPITER... but good. Thanks for the links!
Does the article suggest where the saturated fat came from. Saturated fat in yellowfin tuna might be a different animal tha saturated fat in pastrami.
Of the two editorials on the study by Krauss's research group, the first comment on adding too many controls to a regression model and a worry about measurement error struck me as reasonable, although many studies have those flaws.
The second editorial alleging among other things that Krauss was paid off by the dairy was more, in Stephan's words, "hysterical."
Krauss's group wrote, "The second study cited by Katan et al was published only after our report appeared (4), but it, too, provides evidence that concurs with the postulation by us and others that the replacement of saturated fat with polyunsaturated fat, and not carbohydrate, is associated with CVD benefit. "
Most paleo / WAPF / anti-inflammation diet advocates demonize both excess polyunsaturated fat and carbohydrate (in some cases focusing on sugars) as leading to cardiovascular disease. Certainly I try to avoid both polyunsaturated fat and carbohydrate in excess.
I didn't check the other studies cited, but if Krauss, his research group and others are going to start promoting increasing omega 6 polyunsaturated fat intake, based on the theoretical analyses of Stephan and others, I am not sure this advice will lower heart disease in the population.
Hi David,
I agree that the study looks susceptible to residual confounding. However, the fact that the overall result is consistent with nearly all the other observational studies regardless of location increases my confidence in the result.
Hi Peter,
SFAs probably came from eggs, fish and land animal fat mostly.
Hi Jeremy,
Did you read Krauss's response to Scarborough et al's overcorrection criticism? He points out that when you compare studies that corrected for serum cholesterol with those that didn't, they both give the same result. Scarborough et al. must have known that when they wrote their criticism, because all you have to do is glance at the tables in Krauss's study to figure it out.
Chris - I wouldn't limit the value of this information to only those who are non-obese. Given the information available, there's simply no reason to believe that sat fat comsumption is a significant part of the disease/aging/obesity issue. This study is another piece of the accumulating evidence. There's not even a serious model of how sat fat COULD be the cause of disease/aging/obesity. The sat fat scare was a scam from day one.
David, I would concur with the residual confounding. Nonetheless, what you are suggesting is 'sat fat could still be bad for you above the threshold of this study.' What evidence exists, and by what model would anyone explain, the sat fat conjecture? I know of none. Certainly the inuit didn't suffer from 'sat fat toxicity', nor any of the other paleolithic peoples who ate massive quantities of coconut oil or animal blood or pastured raw milk products (well over 20g/day of sat fat). The sat fat conjecture was tested to the tune of nearly a billion dollars by the US govt, and the tests were negative. The case against sat fat is non-existent. This is another piece of that relatively complete puzzle (relative to what little we really know about diet science). The shocker of this study is that some people will be surprised by the results! We should all, had we been exposed to the evidence vice the propaganda, be saying "ho hum, of course not."
And while I'm feeling chatty, any one of us could decide for ourselves by testing whether or not sat fat is an issue. Get a baseline measure of your own lipid profile, fasting glucose, abdominal circumference, insulin, A1C, or whatever it is that you find to be significant. Replace the high density carbs in your diet with a ton of sat fat for a period of time, and remeasure. If you held the other variables to a reasonable minimum (didn't add 10 shots/day of corn oil to the routine), this kind of test answers a much more important question than 'what should we scientists/doctors be telling people to eat?' It tells you what you should/shouldn't be eating, and how to help others who are in your circle of influence figure out for themselves what they should be eating. Technology like glucose monitors and a relatively cheap fasting lipid profile offers the means to get past "Guru X's theory of a good diet" to "I ate this way and the measured result was ...".
Stephan,
I interpreted the reply of Siri-Tirano et al to the editorial of Scarborough et al to be a new set of estimates that the original authors did in response to the editorial, not merely a restatement of results in the original meta analysis. I spent 30 seconds flipping through the original paper and did see the numbers in the response to the editorial.
Jeremy
No, I do not say that saturated fat is 'the cause of disease/aging/obesity'
But, omega-six does have some evidence of improving insulin sensitivity, probably related to nfkb, jnk, cox, etc signaling as well as fatty acid metabolites like ceramide. I do not think this is a good way to improve one's insulin sensitivity, but to see context for this study.
Hi Jeremy,
What I mean is that if you already know which studies corrected for serum cholesterol and which didn't, you could look through Krauss's tables and see that correcting for cholesterol made no difference. It would have been easy for Scarborough et al. to do that before publishing their editorial. The fact that they didn't suggests to me either that they're lazy, or they wanted to publish a critique of Krauss's paper regardless of its speciousness.
Hi Stephan,
Re: ...(3, 4). No meta-analysis is perfect, but their criticisms were largely unfounded (5, 6).
Those comments you have linked as 3..6 require a subscription, could you post the quotes, if you have an acess and if you can? Thanks
Stan (Heretic)
As Jeremy mentioned, it seems to me that they are Krauss and his group is agreeing that the replacement of saturated fat by polyunsaturated fat will lower CVD. The only thing they seem to be saying is that you shouldn't tell people that saturated fat is bad because people will replace it with sugar and other refined carbohydrates (rather than PUFA).
What do you think of this, Stephan? Should we start replacing our butter with vegetable oils?
p.s. great blog!
What happens when you reduce fat consumption and replace saturated fats with PUFAs? Your HDL goes down and your oxidized LDL goes up, that's what happens. Not very good for CVD.
- JLL
Total fats, saturated fat and cholesterol are not associated to stroke - http://www.canibaisereis.com/2010/12/06/como-reduzir-o-seu-risco-de-acidente-vascular-cerebral/
Stephan, any ideas on whether it is significant that some hunter gatherers (namely East African ones) consumed more PUFAs and MUFAs combined than saturated fats? Would this at all reflect on the flexibility of the cell membrane (wild shot in the dark here)?
Post a Comment