JAMA Internal Medicine - Sugar Industry

A few folks reached out to me to ask if I'd seen the recent JAMA Internal Medicine piece on the Sugar Industry's influence of nutrition science back in the 1960's. Several publications have covered the topic, some fair, some oversimplifying the situation, few adding sufficient context - for a roundup of prominent links, see the following: 1 , 2 , 3 , 4 , 5 , 6. Most of the links and published commentaries either directly or implicitly state that the Sugar Research Foundation (SRF) paid researchers to think in some way and write research for them - alas, the nature of the files uncovered can't confirm such a relationship; the review's authors even note this in their limitation section: "There is no direct evidence that the sugar industry wrote or changed the NEJM review manuscript; the evidence that the industry shaped the review’s conclusions is circumstantial".

Before I mention any context that I think is relevant, let's start off by that this was a massive failure of transparency on the part of the authors, McGandy, Hegsted and Stare, and the SRF, regardless of the times and NEJM's position on conflicts of interest. It's an unacceptable lack of transparency.

While the JAMA-IM article states that they looked at other reports from the time to contextualize their discussion, there is an extensive history that it glosses over regarding sugar, saturated fat, and cardiovascular disease, and the competition between researchers supporting different hypotheses at the time.  I think the context is pretty relevant to the JAMA-IM piece at hand, given the focus only on the potential for industry bias and little discussion of the potential for non-funding related biases. I'd urge anyone whose interested in the heated history of sugar/saturated fat/cardiovascular disease to read my piece responding to the Sugar Conspiracy and documenting the history of the Keys/Yudkin interaction.

While likely not as well known to the public as Keys, one of the authors highlighted in the JAMA Internal Medicine paper, Mark Hegsted, is a well-published nutrition researcher and played a significant role in literature characterizing the impact of dietary fats/cholesterol on serum cholesterol and its implication for CVD - there's even an equation named after him. While authors of many media publications interpret this new JAMA paper as clear evidence that the SRF paid Hegsted and colleagues to 'downplay' sugar's role in CVD, that's far from clear - similar to Keys at the time, its alternatively very possible that Hegsted was defending his hypothesis and, similar to Keys, saw many flaws in the sucrose hypothesis due to limited evidence (most of the evidence from this time is highly limited). Not surprisingly, this was convenient for the sugar industry. Hegsted's student, Walter Willett, now head of the Harvard School of Public Health, defended this idea in StatNews:

“He was a very hard nosed, data driven person, who had a record for standing up to industry interests,” including losing a job at the USDA for standing up to the beef industry, Willett wrote in an email. “I very much doubt that he changed what he believed or would conclude based on industry funding.”

The JAMA-IM paper and subsequent media outlets further failed to note that Hegsted wasn't some sort of sugar-pushing researcher. While he downplayed its role in the etiology of cardiovascular disease, his own testimony to the Senate Select Committee on Nutrition and Human Needs that led to the 1977 Dietary Goals for the U.S. highlights the excess intakes of sugar in the American diet - you can find the text of his testimony in Hegsted's 2000 Annual Review of Nutrition Report titled "From chick nutrition to nutrition policy". The totality of Hegsted's work, to me, suggests less that he was bought out and more that he quite literally profited from vigorously defending a hypothesis (saturated fat/CVD) that was called into question by those vigorously pushing a sugar/CVD narrative at the time (in case you've forgotten, Yudkin, a sugar-CVD advocate, wrote a book titled "Pure White and Deadly" - it wasn't about cocaine).

The JAMA Internal Medicine paper would've benefited from looking at other reviews at the time as a sort of control group. Keys is an obvious option of a researcher who even more vigorously attacked the sugar issue, without known sugar industry funding. The JAMA-IM authors could've looked even to a review written by researcher more sympathetic to the sucrose hypothesis and noted many of the same limitations argued by the NEJM authors.

As I mentioned in my response to The Sugar Conspiracy, the human evidence for sugar's role in CVD focused largely on its ability to increase triglycerides; the JAMA-IM paper gives a small discussion of this, but fails to mention an important point. Much of the evidence put forth at these times were from statistically weak ecological/case-control studies linking sugar intake and CVD, and triglycerides and CVD. It wasn't until 1972 that triglycerides were independently associated with coronary heart disease in multivariate analyses controlling for other established risk factors - after the time of Keys' publications discussing sugar/Yudkin and after the NEJM pieces by Hegsted and colleagues. Even with an independent association of triglycerides and CHD, there was significant debate over whether changes in triglycerides due to carbohydrate content/source were transient. This evidence surrounding triglycerides is important context, given that the JAMA-IM authors, in their discussion of  levels of evidence used to influence policy, are critical of the NEJM authors for focusing on RCTs that use serum cholesterol as an outcome and not other biomarkers. At the time, I'm not aware of other biomarkers that would've been appropriate to use - hindsight is 20/20.

The JAMA-IM authors conclude with:"Policymaking committees should consider giving less weight to food industry–funded studies, and include mechanistic and animal studies as well as studies appraising the effect of added sugars on multiple CHD biomarkers and disease development". The call for using mechanistic and animal evidence is a highly questionable claim, especially in the context of what was known at the time about carbohydrates, sugar and CVD - a review by Grande at the time specifically highlighted the limitations of translating animal evidence to humans within this arena, and further noted that the choice of animal species (e.g. dog vs rodent) had a major impact on conclusions made. Animal evidence has arguably led dietary recommendations astray, overemphasizing the impact of dietary cholesterol on CHD (thanks rabbits!). In an era of medicine where even surrogate outcomes, such as LDL cholesterol, are being critiqued, I'm not sure why these authors are arguing for inclusion of even weaker lines of evidence. An even remotely familiar understanding of the animal nutrition world would demonstrate clear examples where dogs, rabbits, mice, and non-human primates can lead to drastically different conclusions about human nutrition.

As a final note about the content of the JAMA-IM piece, the authors state that this review by Hegsted and colleagues emphasizes the importance of having scientific reviews being written by individuals without conflicts of interest. While this sounds fine, it brings up a couple of discussion points: 1) - there isn't just one type of 'review': there are narrative reviews and systematic reviews. Narrative reviews are notoriously filled with the potential for bias, as they rely on a researcher's perspective of the field; I think they're important in the field of nutrition because the limitations in our lines of evidence (we don't have many double blind randomized controlled trials) require synthesis of different weaker forms of evidence; alas, different eyes and brains find different perspectives on the data. This Hegsted and colleagues NEJM piece was a narrative review, and not surprisingly, one could accuse it (and many other reviews at the time, regardless of funding) of bias. Systematic reviews can be much more robust, and synthesize meta-data - they are not without their flaws and can be highly limited due to inclusion criteria, but they take a more rigorous approach to identify relevant publications and meta-analyze the data. Systematic reviews are a creation of modern science, and their ability to be influenced by different funding sources remains to be thoroughly analyzed. 2) - the authors keep referring to 'conflicts of interest' as though it only means financial conflicts of interest. There are numerous sources of bias that can influence the interpretation of data, and nutritional sciences, with its lack of definitive trial evidence, is the perfect medium for fostering biased looks at data. In my own discussions with folks in the field, there's even an ongoing debate about whether primary researchers in a field should be performing reviews on that field, because being so close to the data (and having generated some of it) can bias your perspective. While it may be convenient to have a conversation about the 'independent' researcher and the industry-funded researcher, this dichotomy is both naive and unhelpful, unless you sell books about the evils of industry-funding.

The media would do well to start having an adult conversation about the issue of funding and bias, presenting the multiple possible interpretations of the hazy associations between researchers and industry. The major take away from this JAMA-IM analysis, for me, is that industry is going to interact with and sometimes fund researchers with whom the industry believes will view the evidence in a way that can bolster their bottom line; this isn't too surprising. This happens now, and it's happened in the past; there's even evidence that our 'The Sugar Conspiracy' hero/JAMA-IM cited Yudkin was a bit cozy with the egg industry. As long as nutritionists remain interested in studying food and nutrients, there will always be an industry marketing department.eager to give some research money. The academic question that remains is the threat that this presents to the integrity of the questions asked, the data generated, and its interpretation by individual researchers and ultimately, policymakers - all needing to be balanced against the potential for unexplored research questions and limited collaborative expertise if industry funding is cast aside.

I would like to see the media begin to become critical of zealous researchers and alternative sources of bias. Industry is going to remain especially interested in funding primary research and reviews as zealous scientists and clinicians continue to make hyperbolic claims beyond the causal evidence. Yudkin writing a book titled 'Pure White and Deadly' probably didn't exactly discourage the sugar industry from wanting to throw money at dietary fat researchers who held alternative hypotheses to his hyperbole. In a similar light, the modern day Yudkin -Robert Lustig - penning articles titled 'The Toxic Truth About Sugar' and publishing uncontrolled trials of fructose reduction with large associated PR efforts probably isn't discouraging the sugar/sugar sweetened beverage industry from funding researchers to call out the obvious hyperbole and perform highly controlled trials that not surprisingly don't recapitulate weaker lines of evidence. The media and those invested in the conversation need to present all of the questions at hand, not just a narrative of 'big industry buys scientists to distort the irrefutable evidence': What are we to do about industry funding researchers with messages some aspects of public health don't like? Does that warrant ending all industry funding? How can industry and academics interact? What balance of industry and private funding should exist if it even should? How do we report on perceived vs actual influence? How do we encourage researchers to take a less biased approach to looking at the evidence, especially when we're surrounded by constant alarms ringing about ending obesity, diabetes, CVD and cancer epidemics? All of these questions are part of a larger conversation that we need to be having and presenting to the tax-paying public who are ultimately going to play a big role in research funding. Calling for less weight to be given to all industry-funded research, and making claims about the modern research environment based on a past lack of transparency to push a narrative of outright industry influence probably isn't exactly my idea of a strong start to this conversation.

It might also help if the world of nutrition could get itself together enough to decide when and how we synthesize our multiple lines of imperfect evidence. We can't have Harvard colleagues vigorously defending recommendations based on observational evidence then have the media writing stories about how BigCereal sold us the idea that breakfast is the most important meal of the day with observational evidence.