Tuesday, March 23, 2010

New Review of Controlled Trials Replacing Saturated fat with Industrial Seed Oils

Readers Stanley and JBG just informed me of a new review paper by Dr. Dariush Mozaffarian and colleagues. Dr. Mozaffarian is one of the Harvard epidemiologists responsible for the Nurse's Health study. The authors claim that overall, the controlled trials show that replacing saturated fat with polyunsaturated fat from industrial seed oils, but not carbohydrate or monounsaturated fat (as in olive oil), slightly reduces the risk of having a heart attack:
These findings provide evidence that consuming PUFA in place of SFA reduces CHD events in RCTs. This suggests that rather than trying to lower PUFA consumption, a shift toward greater population PUFA consumption in place of SFA would significantly reduce rates of CHD.
Looking at the studies they included in their analysis (and at those they excluded), it looks like they did a nice job cherry picking. For example:
  • They included the Finnish Mental Hospital trial, which is a terrible trial for a number of reasons. It wasn't randomized, properly controlled, or blinded*. Thus, it doesn't fit the authors' stated inclusion criteria, but they included it in their analysis anyway**. Besides, the magnitude of the result has never been replicated by better trials-- not even close.
  • They included two trials that changed more than just the proportion of SFA to PUFA. For example, the Oslo Diet-heart trial replaced animal fat with seed oils, but also increased fruit, nut, vegetable and fish intake, while reducing trans fat margarine intake. The STARS trial increased both omega-6 and omega-3, reduced processed food intake, and increased fruit and vegetable intake. These obviously aren't controlled trials isolating the issue of dietary fat substitution. If you subtract the four inappropriate trials from their analysis, which is half the studies they analyzed, the significant result disappears. Those four just happened to show the largest reduction in heart attack mortality...
  • They excluded the Rose et al. corn oil trial and the Sydney Diet-heart trial. Both found a large increase in total mortality from replacing animal fat with seed oils, and the Rose trial found a large increase in heart attack deaths (the Sydney trial reported total mortality but not CHD deaths).
The authors claim, based on their analysis, that replacing 5% of calories as saturated fat with polyunsaturated fat would reduce the risk of having a heart attack by 10%. Take a minute to think about the implications of that statement. For the average American, that means cutting saturated fat nearly in half to 6% of energy, which is a challenge if you want to eat a normal diet. It also means nearly doubling PUFA intake, which will come mostly from seed oils if you follow the authors' advice.

So basically, even if the authors' conclusion were correct, you overhaul your whole diet and replace natural foods with industrial foods, and...? You reduce your 10-year risk of having a heart attack from 10 percent to 9 percent. Without affecting your overall risk of dying. The paper states that the interventions didn't affect overall mortality.


* Not even single blinded.  Autopsies were not conducted in a blinded manner. Physicians knew which hospital the cadavers came from, because autopsies were done on-site. There is some confusion about this point because the second paper states that physicians interpreted the autopsy reports in a blinded manner. But that doesn't make it blinded, since the autopsies weren't blinded. The patients were also not blinded, though this is hard to accomplish with a study like this.

** They refer to it as "cluster randomized", which I feel is a misuse of that term.  The investigators definitely didn't randomize the individual patients: whichever hospital a person was being treated in, that's the food he/she ate. There were only two hospitals, so "cluster randomization" in this case would just refer to deciding which hospital got the intervention first. I don't think this counts as cluster randomization.  An example of cluster randomization would be if you had 10 hospitals, and you randomized which hospital received which treatment first.  It's analogous to individual randomization but on a group scale.

43 comments:

Robert Andrew Brown said...
This comment has been removed by the author.
Robert Andrew Brown said...
This comment has been removed by the author.
Robert Andrew Brown said...

Thanks Stephan

(I apologise for the deletions above, I was in a rush and it is not easy to spot formatting changes from a cut and paste)

I still have not found the time to more than skim this paper. What get me most is the certainty with which their conclusions are trumpeted, when they go against the biology and so are at risk of being flawed, and there is a lot of evidence that throws their conclusion into questions.

This is the sort of comment that is made, and given the tone of certainty in their paper, there is no wonder the media are going with it.


"Thus, the evidence is most consistent and robust for CHD benefits when SFA is replaced with PUFA, rather than with MUFA or carbohydrate, suggesting that lower risk may be more strongly related to increased PUFA rather than decreased SFA consumption."


To make the point that there is evidence that takes a counter position, here are refs to a paper and editorial suggesting saturated fats improve the risk of cardiac disease in women.

http://www.ajcn.org/cgi/reprint/80/5/1102

Dietary fats, carbohydrate, and progression of coronary atherosclerosis in postmenopausal women

"In postmenopausal women with relatively low total fat intake, a greater saturated fat intake is associated with less progression of coronary atherosclerosis, whereas carbohydrate intake is associated with a greater progression."



http://www.ajcn.org/cgi/content/full/80/5/1102



Saturated fat prevents coronary artery disease? An American paradox

"These effects include the paradox that a high-fat, high–saturated fat diet is associated with diminished coronary artery disease progression in women with the metabolic syndrome, a condition that is epidemic in the United States"

Robert Andrew Brown said...

^

And the following comment was made in the results section

Dietary fats, carbohydrate, and progression of coronary atherosclerosis in postmenopausal women

"Quantitative coronary angiography was performed at baseline and after a mean follow-up of 3.1 y in 2243 coronary segments in 235 postmenopausal women with established coronary heart disease"

"Polyunsaturated fat intake was positively associated with progression when replacing other fats (P = 0.04) but not when replacing carbohydrate or protein. Monounsaturated and total fat intakes were not associated with progression."

Joel said...

Actually they do, albeit very quietly and without comment, sneak in data about total mortality on page 5. The pooled relative risk was 0.98. As you predicted, the sum of adding PUFAs and making the diets healthier overall (adding whole foods, fruits, veggies, taking out trans fat, etc) is pretty much neutral.

Thomas said...

Stephan-a basic question here-why do you think these researchers do this-Do you think they actually have an agenda and why would they stake their reputation as good researchers on an agenda that, from what I can tell, doesn't advance their careers much? In fact, I would think turning out studies like this would earn them some disrespect. Are they blinded by bias? It is eye opening (Ive seen a lot of this kind of stuff) and makes anyone using studies to back up a claim no more convincing (at least on the surface) than someone stating opinion only. I hate having to dig into each and every study to see if it was good science or not (too time consuming).

Matt Stone said...

Crap. They recommend eating 15% of your dietary calories as PUFA. I'm only eating 1%. I feel a heart attack coming on.

Glenn said...

Like Thomas, I find this very disheartening. It really undermines whatever credibility "science" has when speaking on any controversial topic. Rather than bias, though, I wonder if there is any accountability at Harvard.
Stephan wrote:
"This paper is a breathtaking example of how far a person will go to confirm his own pre-existing ideas. I can't imagine this kind of paper in any field other than diet-health science."
Does that not bother any of the other faculty there?

fraz said...

Stephan,

Would you mind if a link to this blog post were posted to the PLoS Medicine comments page?

Chris Kresser said...

A perfect example of how myopic and dishonest medical research can be. A mentor of mine likes to say, "If you look for something hard enough, you'll find it." So true in this case.

It's because of stunts like this, which are all too common, that Ioannadis argues that Most Published Research Findings are False. And as Tom Siegfried explains in Odds Are, It's Wrong, science fails to face the shortcomings of statistics.

Statistics and RCTs are like any kind of technology. Their value completely depends upon who employs them, how they're employed, and to what end.

Ned Kock said...

Another thing that looks suspicious about this meta-analysis is that they say that: “Statistical evidence for substantial between-study heterogeneity was not present (Q-statistic p = 0.13; I2 = 37%).”

Now, we know that p values go down with sample size, and are usually high with small samples unless the effect measured by the statistic is very strong, regardless of the statistic used.

Well, with a sample of only 8 studies, their p value is close to significant at p < .05!

If the sample of sourced studies was a little higher (say, 20), there would be significant between-study heterogeneity, which would call the meta-analysis into question.

Moreover, to the best of my knowledge, the Q statistic is not very reliable with small samples, due to its low power as a test of heterogeneity.

Finally, the sourced study with the largest sample (n = 9,057), indicated as Minesota CS on Figure 2, found increased risk of heart disease associated with increased consumption of PUFAs and reduced consumption of SFAs.

Dr. B G said...

I like your title Stephan!

Looks like Dariush is influenced by those at Harvard like Walter Willet, the pro-PUFA, anti-sat fat devil incarnate... (Ancel Keys spirit?)

Thoroughly disappointing.

Aaron Blaisdell said...

I think Harvard, in particular, is housing a lot of scoundrels. I personally know a Harvard research scientist who has been accused of publishing fabricated data. It's disgraceful to science and to academia. Most of the scientists I know personally are honest, hard working folk. But some institutions appear to be prone to acquiring the dishonest types. I think Harvard falls squarely in that category.

Thanks for the thorough analysis Stephen. I skimmed the paper but couldn't make heads-nor-tails of it given I lack a background in the studies they included (and excluded) from their analysis.

Stephan said...

Hi Robert,

I agree, these guys have very little consideration for data coming from fields outside their own.

Hi Joel,

Thanks, I've edited the post to reflect that.

Hi Thomas,

I continue to maintain that these researchers are trying to do good, they just aren't very effective because they're stuck in an old paradigm. Although Harvard diet-health research does have a history of catering to the food industry.

Hi Fraz,

Please do.

Hi Ned,

I agree. There wasn't much heterogeneity because they cherry-picked only studies that agreed with their hypothesis.

Hi Aaron,

I agree. There's a lot of great research coming from Harvard, but it's a very high-pressure environment. Any time you have that kind of pressure, you're going to have people cutting corners.

Beth said...

Thanks for posting this analysis. I saw the news reports & decided that I don't much believe their conclusions, but I didn't know where they were off.

They seemed to be saying that adding omega 3 PUFA is good, and that other foods (e.g. carbs) are not, but why pick on SFA? So the order of goodness is omega-3, SFA, carbs, omega-6, etc, why not say that SFA is more healthy than a bunch of other stuff. :-) That was my naive analysis, along with the idea that there is a vendetta against SFA.

Beth

LynneC said...

From The Heart.org
'Mozaffarian pointed to a recent meta-analysis of prospective epidemiologic studies showing that there was no significant evidence that dietary saturated fat is associated with an increased risk of coronary heart disease or cardiovascular disease'

Mozaffarian recognizes that sat fats are not associated with a greater risk of heart disease, (he surmises because we have replaced sat fats with refined carbs),
so he cherry picks the studies that will support his point.

I did see the mention of Unilever on his disclosures...hmmm.

John said...

I can only imagine having to leave the country to find saturated fat foods at some point down the road...

Sara said...

Stephan--
Thanks for taking the time to comment on this study. It always concerns me when a study recommends consuming more PUFA in the place of saturated fat. Some doctors read these studies and take them to heart without exploring them in depth. As a woman, this advice will be more likely to lead you in the direction of breast cancer. Scary thought.

Jamie Scott said...
This comment has been removed by the author.
Jamie Scott said...

I guess if we are going to say one positive regarding this study, they are suggesting replacing a fat with a fat - albeit somewhat misguidedly. I note the headline that has come through this morning on my Medscape alert;

Replacing Saturated Fat With Polyunsaturated Fats, NOT CARBS, Reduces CHD Risk.

I see that as a small consolation at least.

Thanks for your insight into this Stephan.

Jamie Scott said...
This comment has been removed by the author.
WL said...

Hi Stephan,

"Cluster randomized" actually is a standard statistical term, although I'm not familiar with the Finnish trial and don't know if it should be called cluster randomized.

Just a footnote. Thanks for your always thoughtful posts!

Stephan said...

Hi WL,

Thanks. I've changed my footnote a bit. Let me know what you think if you get a chance.

Stanley said...

The PLoS site won't allow me to leave a comment. Hopefully someone else here will have better luck, before we kill more people with corn oil.

http://www.plosmedicine.org/article/comments/info:doi/10.1371/journal.pmed.1000252

Stephan,

Thanks for your insightful and socially responsible review. It's really too bad that the BBC site doesn't seem to allow comments. I shudder to think how many more CHD deaths the article will end up causing.

Stanley said...

P.S. Blogger truncated my link. Just search PLoS for the study title if you want to comment.

Stephan said...

Tried to comment on it. We'll see if they post it.

Monica said...

"This paper is a breathtaking example of how far a person will go to confirm his own pre-existing ideas. I can't imagine this kind of paper in any field other than diet-health science."

Unfortunately, I think it's more prevalent in other areas of science that you may realize. The worst fields are generally the ones that are politicized, however.

Helen said...

Jamie,

I don't know if it's really better to replace saturated fat with PUFAs than with carbs. In fact, I doubt it. I see where you're coming from, conceptually, though.

Thanks, Stephan, for dissecting this! I was hoping you would.

Has anyone been able to track (or can you remember) the mainstream coverage of this compared to the Krauss meta-analysis showing no link between saturated fat and CVD? Just wondering if it's getting more play.

Mark said...

Monica,

Indeed. And this field is heavily politicized. I sometimes wonder if articles like this are not part of some conspiracy to ensure that Americans continue to consume low-quality, high-profit-margin foods from Agribusiness.

And how would Big Pharma's antidepressant drug sales be without our enormous consumption of toxic vegetable oils?

Jamie Scott said...

Helen,

There was some coverage of this study on the BBC. Like you I wondered about whether the Krauss paper had received any publicity on the same site. I spent an hour on the BBC site the other night and could not find a single report that referred to the Krauss paper or for any support in that direction. In fact, without exception, every report on fat/saturated was on the negative health benefits of consuming fat.

Austin said...

Stephan you tore this "meta-analysis" to shreds. With that said I am sure you read who funded the study. How can you honestly still think their intentions are good? Any educated scientist would realize that from an evolutionary standpoint, PUFAs would not be more advantageous than SFAs. To naturally obtain even 1 tablespoon of w-6 would take eating 7 bunches of broccoli or roughly 10 pounds of broccoli, or a less extreme example would be 3 pounds of a high starch species of corn!

Jeffrey of Troy said...

Thomas said:
"...a basic question here-why do you think these researchers do this-Do you think they actually have an agenda and why would they stake their reputation as good researchers on an agenda that, from what I can tell, doesn't advance their careers much?"

Beth said:
"They seemed to be saying that adding omega 3 PUFA is good, and that other foods (e.g. carbs) are not, but why pick on SFA?"

Yeah, I think that's the point of this paper. Ever since Taubes came out with GCBC in 07, the clock's been ticking on the high carb recommendation. So they're fighting a rear-guard action to say " we were right about the sat fat tho!", thereby salvaging their careers.

Chris Kresser said...

It takes a very strong person to admit they've been wrong for the past 20 years, even to themselves. Can you imagine spending the better part of your career investigating a hypothesis that later turned out to be incorrect? Talk about a waste of time.

That's why I have some empathy for these folks. But that only goes so far, because public health is at stake. At some point they need to get over their own disappointment and start doing what's right.

Stephan said...

Hi Austin,

I agree. From first principles, the idea that it's healthy to eat a lot of PUFAs is absurd.

I don't necessarily dismiss studies because of their funding sources though. Sometimes controversial research gets funded by industry because the NIH won't touch it. Ronald Krauss's research is funded in part by the dairy industry. He's the guy who's been researching LDL particle subtypes for decades. But it does make you have to examine the research with a bit more skepticism.

Hi Chris,

True. Once you box yourself in with decades of research, it's super hard to change your tune. Especially if your research has altered the course of public health efforts. I know it's hard enough for me to admit I've changed my mind on something, and I've only been putting my opinions on the internet for 2 years on this blog.

WL said...

Hi Stephan,

Your revised footnote looks good to me. With only two hospitals, there isn't any real benefit from randomizing which hospital gets the intervention first, so I do think it's misleading to call the trial cluster randomized.

Fjoraro said...

The study Robert posted, that suggests saturated fats improve the risk of cardiac disease in women, was also carried out by Dariush Mozaffarian. In Men's Health november 2007, Mozaffarian commented on that study where he said that before the paper was published in the American Journal of Clinical Nutrition, he encountered formidable politics from other journals.

Quoted to Men's Health: "In the nutrition field, it's very difficult to get something published that goes against established dogma," said Dr. Dariush Mozaffarian MD MPH, assistant professor, Harvard. "The dogma says that saturated fat is harmful, but that is not based, to me, on unequivocal evidence." Mozaffarian says he believes it's critical that scientists remain open minded. "Our finding was surprising to us. And when there's a discovery that goes against what's established, it shouldn't be suppressed but rather disseminated and explored as much as possible."
http://drbganimalpharm.blogspot.com/2009/06/benefits-of-high-saturated-fat-diets-in.html

It's a pity Mozaffarian and his group have given way to the pressure and now publish this study where they come up with a conclusion that is very likely to cause harm to anyone's health if followed.

The numbers they use in the study are certainly speculative and not very reliable. For example it is usual in this kind of studies to classify lard as a saturated fat, although such fat is only 40% saturated. The control groups, that are supposed to be high in saturated fat, might actually be eating less saturated fat and more unsaturated fat than what is shown. For the specific individuals that had the heart attack, we cannot be sure how much PUFA they really could have been consuming.

However, under the right circumstances I think it could be possible that replacing saturated fat food with PUFA food might show a reduction in the risk of heart disease. As Stephan has shown in older posts, arachidonic acid production maxes out around 4% of calories from linoleic acid. So if the omega 6 intake is not reduced below this threshold, the risk of getting a heart attack might not be expected to be significantly lowered. From Table 1 from the study, we can see that none of the control groups had an intake less than 4% from PUFA. On average, the control groups had an energy intake from PUFA of 5%. This is how it has been for almost every study of this kind; the people that eat a “saturated fat diet” also eat a high PUFA diet. When the PUFA intake is not reduced to a low enough level, other factors might have more influence than a change in the saturated fat to unsaturated fat ratio.

An important point is that in real life it is not possible to only replace saturated fat with polyunsaturated fat. Foods with a high PUFA content usually also have a higher antioxidant content than foods that are high in saturated fat and low in PUFA. Animal fats contain cholesterol which during heating and processing result in oxidized products, which some studies suggest might contribute to the development of atherosclerosis. Animal fats also have some preformed arachidonic acid which can be directly incorporated into the cells, but I am not sure about the significance of this and a proportion of dietary LC-PUFAs tend to be lost during cooking. But on the other hand is red meat high in absorbable iron, and reducing the iron level in the body should reduce the risk of having a heart attack, especially when PUFA are present in the diet.

I think that the cholesterol lowering effect, like in the Oslo study, occurring after the diet change perhaps could lower the risk of heart attack. Since the participants in both the saturated and the PUFA group probably had about the same amount of PUFA or oxidized lipids in their LDL particles, having less LDL particles at a time in circulation would perhaps be less detrimental in this case.

Fjoraro said...

Another way to explain the result from this study could be due to the statement that the average PUFA consumption in the intervention groups accounted for as much as 14.9% of total energy intake. In Modern Nutrition in Health and Disease (http://www.scribd.com/doc/27915730/Modern-nutrition-in-Health-and-Disease-9th-ed) it is stated that when intakes of linoleic acid are above 12%, the tissue concentration of arachidonic acid may actually decrease, since the delta 6 desaturase enzyme would be inhibited. If this is correct, the people in the intervention groups might have lowered their concentration of arachidonic acid. This alone should reduce their risk of having a heart attack, as is shown in some studies: http://jn.nutrition.org/cgi/content/full/134/11/3095

So when the authors of the study say that; “The findings also suggest that an upper limit of 10% energy consumption from PUFA may be too low, as the participants in these trials who reduced their risk were consuming about 15% energy from PUFA”, maybe they are not entirely incorrect (when the right conditions happen to occur) that increasing an already high intake of PUFA to an even higher intake can slightly reduce the risk of having a heart attack. Of course, this would anyway be meaningless since it would not lower their overall mortality risk and since the heart attack risk can be way more reduced by keeping the PUFA intake as low as feasible.

Andy said...

But isn't it true that oxLDL levels increase even after PUFAs reach 4% of calories? In the study Stephan writes about in the oxLDL part 2 post, the monounsaturated fat diet had 4% of energy as PUFA, the n-6 rich diet had over 12% as PUFA. In the n-6 rich diet, oxLDL increased by around 30% compared to the MONO diet.

And Stephan makes a compelling case oxLDL has a lot to do with CHD, so shouldn't it have played it's role in the trials replacing SFA with PUFA?

john said...

whats up stephan-

have you had any experience with, or run across any research around, rice bran oil? i just read the following blog entry and was curious what you thought.

http://blog.ideasinfood.com/ideas_in_food/2010/04/rice-bran-oil.html

thanks

john (in cville)

DaisyPatch said...

I noticed it was mentioned that people had trouble commenting on the PLoS controlled study, or intended to try commenting in the future. But when I look, I do not see their comments - only one by spiritosl about a possible math error. I left a test comment and it was accepted without any problem. So maybe I could encourage the rest of you to try again?

trinkwasser said...

Curious that Mozaffarian, who has done much better work in the past, would publish this just as Frank Hu came onside with Krauss. Maybe Harvard are desperately trying to achieve "balance" in order to avoid upsetting their funding sources.

Elizabeth said...

Another point on the Finnish mental health hospital study: if Dr. Natalie McBride is correct in her hypothesis that severely damage gut biota results in damage to the brain and nervous system--that disease of the is actually the root cause of many mental illnesses--then a nutritional study of a mental health population is meaningless. The population is pre-selected for poor health, including conditions that lead to cardiovascular disease as well as mental illness.

Tom Master said...

"Fears have recently been expressed, both within the profession
and outside it, that clinical trials may sometimes operate
against the best interests of the patients. We would like to record our own experience of the reverse-namely, that the
necessity in this trial for careful supervision and continuity of
care resulted in unusually good doctor-patient relations and
patient morale."

Ah yes, more people in the experimental groups died than in the control groups, but they had good morale! Finally took the time to read the Rose et al Corn oil study, and found that a little comical.