Feb 10 2010
Study Exonerating Saturated Fat Has Potential Conflict of Interest
Photo by markhillary
Last month The American Journal of Clinical Nutrition published a meta-analysis of 21 studies concluding that saturated fat is not linked to heart disease–a finding that flies in the face of 20th century nutrition dogma.
Despite the tremendous impact such information would have on the field of nutrition however, the research was largely ignored by mainstream media outlets.
The meta-analysis performed by scientists at the Children’s Hospital Oakland Research Institute was funded by the National Dairy Council and the Unilever corporation, two institutions with a vested interest in selling food products to consumers. The study was also funded by the National Institute of Health.
Technically funding source should not impact scientific outcomes, but analyses have repeatedly shown that industry sponsorship can strongly predict research conclusions that favor the source of funding.
I asked Marion Nestle, professor in the Department of Nutrition, Food Studies and Public Health at NYU, about the specific impact of industry funding on a meta-analysis of this nature.
“Meta-analyses are particularly subject to investigator bias (conscious or unconscious) because of the selection criteria for inclusion.”
So while this study may provide another clue about the relationship between saturated fat and heart disease, caution should be taken when interpreting the results.
Conflict of interest is a serious concern in science that is used to guide health policy and consumer behavior. In this case the influence of industry sponsorship would be especially unfortunate since an impartial evaluation of the link between saturated fat and heart disease is greatly needed.
Saturated fat is most commonly associated with animal products like beef and pork, and is known to raise cholesterol. Since high cholesterol is considered a biomarker of heart disease, it has long been assumed that saturated fat contributes directly to America’s number one killer.
But while this argument seems to make intuitive sense, a direct causal relationship between saturated fat and heart disease has never been established. Moreover, evidence is accumulating that cholesterol in general is not the best predictor of heart disease and that refined carbohydrates are a bigger problem.
Thus there is a real need for rigorous science regarding the role of saturated fat in heart health and other diseases.
“My take on this one is that it is one more piece of evidence that saturated fat may not be AS important a determinant of heart disease risk as is sometimes believed-at least in the kinds of studies included in the analysis. These, of course, do not paint a complete picture of the situation.”
Saturated fat is unlikely to make or break any diet in terms of overall health, let alone heart disease in particular. General dietary patterns consistently prove to be better predictors of long-term health than any single food or nutrient, and anything we discover about saturated fat is unlikely to change this.
“In any case, this is another example of what happens when you look at single nutrients outside of their dietary context.”
While we’re waiting for science and industry to battle it out in the lab, focus on eating real, unprocessed foods most of the time and don’t get too hung up on the details.
What do you think of industry sponsorship in nutrition studies?
Update: After speaking with readers and colleagues I have made some revisions to this article to clear up my stand on this research. I do not intend to imply that the investigators of this study were influenced by their funding source, only that caution should be used when interpreting the results of any meta-analysis where there is a potential conflict of interest.
If you enjoyed this post, you might want to subscribe to Summer Tomato and follow me on Twitter!
Want more? Check out Summer Tomato Live!
Tweet
Tags: American Journal of Clinical Nutrition, dietary pattern, food industry, heart disease, Marion Nestle, saturated fat
Print This
•
Loading ...
No scientist is unbiased, regardless of funding. You only have to read about Ancel Kyes in Good Calories Bad Calories. Keys was anti-fat and was ready to cook the data in the six nations study to prove his point. On the way, he inspired the Standard American Diet and, in my opinion, caused more harm than a few wars put together. He was not funded. He only had a huge ego to satisfy.
It took Krauss many years to come out with this study and try to exonerate saturated fat. Many years in which he witnessed saturated fat’s positive effects on beneficial sub-fractions of HDL and LDL that he studied in his lab.
Krauss may be biased, as any person might be, but his reputation in the scientific community is impeccable.
Great points. And since industry funded science is known to be even more biased than independently funded science, it is even harder to believe.
Darya, on what basis do you say that industry funded research is known to be more biased than independent research? And, for that matter, how do you define “independent research”? I’m pretty sure that anything that’s government funded is at least as likely to suffer “investigator bias” as something that is corporate funded. And foundations that fund studies have as strong a point of view as corporations. The problem is that the proliferation of studies has produced little but noise.
Hi Francis, I linked to a study that specifically looked at outcomes of studies funded by industry on the same nutrition topic as studies funded by traditional government funding agencies. The industry sponsored studies were significantly more likely to result in a positive outcome. In intervention studies reviewed in the article, 0% of industry sponsored studies were negative for the funding agencies. This is not an isolated case, either. Similar bias has been found in pharmaceuticals, cigarettes and other industries with strong financial interest in research outcomes. It is very unfortunate.
There is always bias, but it has been demonstrated to be worse when industry funding is involved. At least for certain industries like food, pharma and cigarettes.
OK, so the bias study you link to shows a statistically-significant skew in conclusions unfavorable to the sponsor. But that’s why published, peer-reviewed studies include their supporting data: to figure out if the conclusions are warranted from the data presented. Sure, I’d grant that funding like this means you probably want to more closely examine the study, but to say that a study is “tainted” merely based on the conclusion being favorable to one of the sponsors is possibly indicative of bias on your part.
Do you believe *this* study to actually be biased, or are you merely engaging in ad hominem argument? If you believe it to actually be biased, based on what *data*?
Actually I think the findings of this study are probably correct. When I say tainted I mean that it cannot be definitive until it is repeated by an unbiased source. I’m not actually against industry funded research, I just question how much we can conclude from it.
As Nestle points out, a meta-analysis is particularly susceptible to bias because you get to pick what studies are included and which are excluded. It’s like getting to pick your own jury in a court of law.
“It’s like getting to pick your own jury in a court of law.”
This does not seem to be a fair statement to me. This would only be true if the study investigators (and/or their sponsors) selected the scientists that peer-reviewed this paper. They are the closest thing to a jury and as far as we know they are unbiased.
Fair enough. Though the American Journal of Clinical Nutrition is run by The American Society of Nutrition, which has definitely been friendly to industry interests (http://www.nytimes.com/2009/09/05/business/05smart.html).
Remember that this study is not primary research, it is a meta-analysis that requires setting selection criteria for what to include. I’m not saying it is wrong, but I can’t tell how valid it is without doing my own search through Medline and basically redoing it myself.
the nyt-link is horrible.
“Eileen T. Kennedy said the program’s criteria were based on government dietary guidelines and widely accepted nutritional standards.”
so froot loops are better than oatmeal/spelt and is acceptable because gov. dietary guidelines allow a product that is 42% sugar? maybe one should start there first and only after that investigate how much influence the sponsor had on the outcome.
One would also question “conventional” data – where did the funding come from with that? For example, we are told that saturated fats are seriously wrong for us… that started with the CSPI. They made restaurants switch to TRANSFAT laden oils. Ironically, they are turning around and looking to sue based on the dangers of transfats.
All in all, I state that the whole food question is easy: eat REAL foodas much as you can. That includes foods that have minimal refinement. We survived for millenia using saturated fats, vs margarine. Butter is real, margarine is not.
Milk fat is real – today’s Danone is not.
HFCS is killer, sugar is not.
I lost 45 lbs eating real foods, that is it. I am living proof that real foods work.
uhm no. Can’t remember the book/paper but the traditional finnish diet is/was one of the most unhealthiest diets, mainly because it was based on (lots of) butter, cheese, (all) meat and fish. heart attack rates were extremely high. maybe darya can cite the paper.
“I lost 45 lbs eating real foods, that is it. I am living proof that real foods work.”
i don’t know why but everybody seems to think that because they lose weight, their preferred lifestyle/diet works, that’s just nonsense.
I think what Jason means is that you can just concentrate on feeding yourself real food, prepared at home, and in moderate amounts. You don’t have to worry about all the studies, media sensationalism, or which specific nutrients you need and in what amounts each day. The body will crave what it needs. No reason to subscribe to a Named diet, especially one that’s trying to sell their own products.
Read Summer Tomato, Michael Pollan, and Well Done Chef. I’ve done exactly that, lost about 40 pounds, and created a healthstyle that works for me (and that I’m constantly refining).
i obviously read ST, read MP, haven’t heard of WDC.
don’t get me wrong, i agree with you and jason on how to eat and go further and say… a company like Nestle shouldn’t even be around (and i say this as a european)
but if you eat like WDC or ST suggests, the question is not: is it bad for me? but rather: is it the best diet for archiving my personal optimum, mentally and physically?
and claiming that (a diet which fulfills that constraint) should not be based on whether or not you lost weight.
I am not convinced that the drive to publish positive results is differentially biased by the nature of funding. Publish or Perish may never have been a greater driving force than it is now given the economic climate, and so there is always a drive to get interesting results published regardless of funding source. I have funding from all three major sources (government, foundation and industry) and don’t feel a modicum of difference in influence on the publication process. Seems that the finding that “analyses have repeatedly shown that industry sponsorship can strongly predict research conclusions that favor the source of funding” may be influenced by the fact that all publications are biased towards positive results (caveat: I have not reviewed this study).
In my opinion, collaboration between industry and science should be encouraged, not scowled at. The integrity of research should be based on integrity of the lab and the quality of the report. There is not enough money being invested in science by our government and foundations to eliminate a valuable source of funding.
Below is the executive summary of the American Council on Science and Health’s report, “Scrutinizing Industry-Funded Science: The Crusade Against Conflicts of Interest.”
ACSH receives no strings attached support from industry, foundations and individuals.
We encourage people to critically evaluate all science with the utmost skepticism, regardless of who funds it. But evaluate the SCIENCE.
The full report is at
http://www.acsh.org/publications/pubID.1687/pub_detail.asp (click on “View PDF Version” on the right of the screen.
Executive Summary
For approximately a century, industry has been a powerful motivating force in the creation of new technology and the underwriting of scientific research. Yet the last two decades have seen the development of a sweeping conflicts of interest movement aimed squarely at curtailing academic/industry biomedical research collaborations and restricting membership on government scientific advisory boards to researchers associated with industry.
Conflicts of interest activists assert that ties between researchers and industry are harming patients and consumers, undermining public trust in research, food safety and environmental regulation and boosting the costs of medicine and other products. As evidence, they repeatedly point to the same handful of research “scandals” and have produced a number of studies correlating industry sponsorship with favorable results in clinical research. In addition, the anti-industry activist groups are trying to exclude academic researchers who have any ties whatsoever to industry from government scientific advisory boards. However, even the activists’ own flawed studies can’t demonstrate that industry “influence” is distorting the decisions made by those boards. The campaign to purge any experts with industry ties—no matter how slender—from advisory panels is chilling scientific debate and depriving regulators and the public of valuable insights.
These conflicts of interest activists focus almost entirely on the alleged baleful effects of financial conflicts of interest while ignoring how other conflicts can bias scientific research and advice to government agencies. People are influenced by all sorts of interests besides money. Why should having once consulted with Pfizer or DuPont disqualify a scientist from serving on a government advisory board or writing a review article in a scientific journal, while being a lifelong member of Greenpeace does not? And if owning $10,000 in Dow stock represents a potential conflict of interest, surely $10,000 in funding from the Union of Concerned Scientists does too.
Contrary to the claims of conflicts of interest activists, the overwhelming majority of patients and research subjects are not being harmed, public trust in scientists and scientific research remains extremely high, and new drugs not only save lives but money. CenterWatch, which tracks 59,000 clinical trials in the United States, found that industry-sponsored drug trials are in fact safer than those at academic institutions funded by government. Polls regularly show that physicians and scientists are two of the most trusted professions. In fact, a recent poll found that three-quarters of cancer patients would have no qualms about enrolling in a study of a treatment sponsored by a company in which a researcher owns stock or from which he/she receives royalties. And finally, econometric research shows that newer drugs, rather than increasing overall medical costs (and thus arguably being foisted in a fraudulent fashion on a cash-strapped public), reduce other medical expenses by a factor of five.1
The current obsession with conflicts of interest is not harmless. The activists have provoked the development of unnecessary and complex academic regulations and restrictions that are interfering with the speedy translation of scientific discoveries into effective treatments and new products and technologies. Instead of improving public health or making our environment safer and cleaner, the activities of conflict of interest activists are harming them. Researchers are abandoning universities and some are even leaving the country for locales in which academic-industry collaboration is encouraged rather than denigrated and penalized. Government agencies are being denied access to sound scientific advice, which distorts regulatory priorities, risks lives and raises costs.
When abuses have been uncovered, private entities including journals, universities and scientific professional societies have adequately addressed them. Such private solutions include the advent of permanent online peer-review of scientific studies and the requirement by scientific journals that all clinical trials be registered. These private efforts are undercutting the campaign by activists to have Congress enact onerous conflicts of interest regulations.
The plain fact of the matter is that there is very little evidence that alleged conflicts of interests are significantly distorting scientific research, harming consumers and patients or misleading public policy. Most conflicts of interest activists clearly have prior strong ideological commitments against markets and corporations. They view the conflicts of interest campaign as another tool to attack an enterprise which they already despise on other grounds.
Do you assert the same inference toward government funded research?
The government provides the bulk of scientific research funding, so no I do not think the bias is the same. The financial incentives are not the same.
You really think so?
I think that view is quite naive; paint me cynical.
(sorry I hit the “Submit Comment” button too soon)
While industry funded studies may be suspect, they should not be any more suspect then any other study. There is no reason to believe that the interests of any other funder (including the governments, or even academia itself) introduces less bias. Maybe the article you mentioned that showed more positive outcomes was because the other studies were funded by other interests that influenced them, which is why those studies were more negative, not the other way around.
I believe it would be blindly optimistic to expect an industry to fund studies that may cause financial loss to the only audience they really care about — their investors.
I have not looked up the specific numbers, but I believe if I did I would find that the vast majority of scientific studies in the US are funded by government and industry…not independents. Lets get to the heart of it …industry funds government. Just look at the pharmaceutical industry and the FDA…it’s an abomination.
Darya,
you write, ” When I say tainted I mean that it cannot be definitive until it is repeated by an unbiased source.”
Wouldn’t you say the same about a study funded by the tooth fairy? That it should be replicable?
Do you think Nestle has ANY biases? What about a meta-analysis that she did, might it be susceptible to her non-financial bias? (Or perhaps a financial bias, i.e. to sell more books promote her career, etc.) I’m not saying she IS bias, I’m just saying she’s susceptible to bias, just as anyone else is.
That’s why it is best to critically analyze all studies, regardless of any potential for bias (of any kind). Take a good hard look at this study you say is tainted, but take just as hard a look at a study you think may NOT be subject of bias. Doing any less is being a shoddy consumer of science.
Jeff Stier
Twitter: @JeffACSH
I don’t disagree with you. Maybe the word tainted is too strong? The study makes bold claims about saturated fat and heart disease and I do not feel I can trust the findings without re-running the search (basically the entire study) myself. To me that makes it not very useful as a resource.
It’s not always necessary to repeat a study to figure out its limitations. While I haven’t read the article in depth, I noticed that their inclusion criteria required prospective studies only and the endpoints needed to be actual cardiovascular disease (CVD) (i.e. a cardiac arrest or stroke and not merely sky-high cholesterol levels). While they get more precision using these criteria, they narrow their pool of 600+ articles to 21. Also, the likelihood of a random healthy adult in a prospective study getting a CVD during the period of the study is fairly low. These two factors combine to greatly lower the chance of finding a statistically significant result. Retrospective studies obviously have issues, but the pool of people with CVD as endpoints would be much higher.
I’m sure many of these studies also reported CVD risk factors too, but this data was ignored. What does it mean if saturated fats lead to significant increases in CVD risk factors, but not CVD events? I don’t know, but it would be really interesting/important if this was the case. Instead we have a meta-analysis where the criteria were only the most restrictive and gave a null result.
Great points Dan. There is actually a body of literature decoupling traditional heart disease markers from CVD events, and I think that is why this study did it this way. But yes, it narrows the field quite a bit and makes it difficult to interpret.
Thanks for your input.
I agree that industry funded research should be taken with a grain of salt and I don’t think this one meta-analysis holds a lot of weight. However, in this particular situation the dairy industry doesn’t lose anything if people switch to low fat cheese or milk, as long as they’re still eating dairy.
Yeah, I agree.
You are right— “caution should be used when interpreting the results of any meta-analysis where there is a potential conflict of interest.”
I agree that industry funding of a study IS a potential source of bias. It is one of many potential sources.
Some of my thoughts on this at a talk at CME conference last year:
http://www.youtube.com/watch?v=mGE4mCVrpKI
Darya:
I think you should read the follow-up opinion article the same group published:
http://www.ajcn.org/cgi/content/abstract/ajcn.2008.26285v1
More recently, however, stimulation of Tolllike
receptor 4 has been attributed to liposaccharide and lipopeptide
contamination of bovine serum albumin, a commonly
used reagent in cell culture preparations, rather than saturated fatty
acids (121).Nonetheless, there is a growing body of evidence from
cellular and animal studies that supports the proinflammatory effects
of saturated fat, as reviewed extensively elsewhere (108).
So we’re getting inflammation from bacterial sources and not the saturated fat itself perhaps? That’s not a very favourable thing for products of cow milk.
The Framingham study evidence underlying the “lipid hypothesis” was never strong to start with. Since then a massive lipid lowering campaign has shown no effect on heart disease rates. While an elegant and seemingly intuitive hypothesis, more and more openly people are rightly questioning the wisdom of the cholesterol lowering campaign.
Cholesterol is an essential component of every cell membrane and important for myriad physiologic functions. When Dr. Uffe Ravnskov, MD PhD looked at the medical literature he found something quite surprising had been documented there. On average people with higher cholesterol live longer. Cholesterol is a mediator in heart disease but blood cholesterol levels have next to no effect on heart disease rates again heart disease rates mostly unchanged since the advent of the massive cholesterol lowering campaign. Here is something else to consider, as any chemist will tell you, cholesterol is a single molecule. How then are there “good” and “bad” cholesterol molecules. It is at best scientifically imprecise and at worst a crass marketing ploy to talk about the levels of high and low denisty lipoprotein (say it again lipoprotein i.e. a protein – they are carrier proteins) as implying different cholesterol molecules. Then again the statin cholesterol lowering drug class alone is a 30 billion dollar a year industry. Of course none of this even begins to cover the serious side effects of statins.
http://healthjournalclub.blogspot.com/