February 19, 2016
Statement by Nina Teicholz, science journalist and member, The Nutrition Coalition
Thank you very much for this opportunity. I am speaking today on behalf of The Nutrition Coalition, a fledgling group of doctors, PhDs, and others, not backed by any industry, who want to be sure that nutrition policy is based on a rigorous and comprehensive review of the science, in the interest of the public health. We do not endorse any particular diet.
The problem we might agree upon today is that the Guidelines have they been unable, over the past 35 years, to effectively combat obesity, diabetes, and other nutrition-related diseases.
Several explanations are usually given for this conundrum. One is that Americans don’t adequately follow the DGAs, yet this explanation is contradicted by USDA data, showing that the public has, on the whole, been compliant. From 1970 to 2005, the consumption of grains rose by 41%; vegetable oils by 91%; fish and shellfish by 37%; vegetables by 23%; and fruits by 13%. Eggs and red meat each fell by 17%, and whole milk by 73%. These changes represent overall compliance with the Guidelines.
A second explanation for why the Guidelines have failed is that Americans simply eat too much. Calorie consumption has risen. This is a possible explanation, yet the evidence is unclear. A large body of evidence now shows that the human body does not respond to all types of calories equally and that the overconsumption of carbohydrates (because they stimulate insulin production) might be uniquely fattening, as well as contribute to diabetes. In the US, because the increase in per capita calorie consumption in recent decades comes almost entirely from carbohydrates, it is impossible to disentangle whether it’s the increased calories or the increased carbohydrates that might have been responsible for growing obesity.
A third explanation for obesity and diabetes rates today is that junk food alone is to blame, yet in the last 15 years, Americans have cut their consumption of sugar and refined carbs, which are good proxies for junk food consumption, yet rates of obesity and diabetes have not abated.
An alternative explanation for the problem of nutrition-related diseases is that the DGA recommendations themselves are somehow flawed or that the advice only works in a minority of the population.
There is only time to touch on a few issues today. An important one is the Nutrition Evidence Library, or NEL, which is used as a basis for the scientific reviews that undergird the Guidelines. It is problematic that the NEL is not consistently used by the committee; on several key reviews in 2015, including the one on saturated fats, the NEL was not consulted. A larger problem, however, is that the NEL is incomplete. In fact, it’s safe to say that it does not include a majority of the scientific literature on nutrition and disease. Remember that the National Heart Lung and Blood Institute, and its predecessor agency, since 1948, have been virtually obsessed with the hypothesis that fat of some kind or amount causes heart disease and has spent billions of dollars trying to get an answer. What’s happened to that research?
On saturated fats, for instance, there have been huge clinical trials, several funded by the NIH, conducted on altogether 75,655 men and women, in experiments lasting 1-12 years. None of these are in the NEL. The major epidemiological trials on saturated fats from the 1960s and 70s, on tens of thousands of people, are also missing. Moreover, in the last five years, there’s been a sea-change in the thinking on saturated fats, with at least 13 major meta-analyses and systematic reviews, several of which received NIH-funding, and most of these are missing from the NEL. The 2015 review of saturated fats, conducted without the NEL, also missed about half of these review papers. Indeed, although limits on saturated fat have been the central pillar of dietary advice since the first Guidelines in 1980, no dietary guidelines advisory committee has ever directly reviewed the clinical trial or epidemiological evidence on these fats, which is astonishing and ought to be rectified as soon as possible.
The low-fat diet is another recommendation that has been studied extensively, in a half-dozen NIH-funded, multi-center clinical trials, on altogether more than 57,000 men and women, in experiments lasting 1-8 years. From these studies, 13 publications reported the major results, only one of which is in the NEL.
Possibly because those trials did not show any benefit of the low-fat diet for health, Guidelines advisory committees since 2010 have backed off from any “low-fat” language, but its “Dietary Patterns” are all still modeled at 32-24% of energy as fat, which is still “low fat” according to historical definitions, and the Guidelines still recommend “low-fat” dairy and “lean” meat. It would be good to reconcile these apparent contradictions and also to make it clear to the public how much fat is advisable to eat for good health. We’ve been advised to eat a low-fat diet for decades now, and if that’s no longer the recommendation, then funds should be spent to reeducate the public.
Furthermore, after only a brief search, we found that there are more than 58 papers on low-carbohydrate diets that are not in the NEL. No doubt there are more. This has been a burgeoning area of research over the past 10 years, yielding an overall conclusion that carbohydrate restriction may very well be the best hope in in helping people combat the worsening metabolic state that leads to diabetes, obesity, and heart disease. Yet the 2015 Committee did not review this literature, saying that it couldn’t find any studies. Very likely, that’s because they weren’t in the NEL.
It seems worth questioning, therefore, if there should be a NEL. Why should reviews not take place simply by searching the major electronic data bases, such as Pub Med, directly, which is the standard method for scientific reviews? The process of admitting studies into the NEL introduces the possibility of bias and clearly has lead to major deficiencies.
There is the additional problem in the 2015 committee’s report, namely that it made recommendations based on “weak” or “inconclusive” evidence. The vegetarian diet is one such recommendation. And although the committee recommended eating “lean meat” and reducing “red and processed meats,” it did not do any comprehensive review of the science on those topics. Actually, there are NIH-funded clinical trials where red meat was significantly reduced. Yet these studies are not in the NEL. The committee looked instead at epidemiological data in which the category of meat was mixed together with eggs and dairy.
This over-reliance by the Guidelines on epidemiological data, which can only show association but not causation is also a crucial issue. Unless the associations found in these studies are very strong, with relative risks of at least 5 or 6, they are considered highly unreliable. (In the field of nutrition, these numbers are usually less than 1.3). This weak science has been the cause of quite a few mistaken ideas: that fat causes cancer, that olive oil is healthiest, that dietary cholesterol causes heart disease, and that vitamin anti-oxidants prevent disease. All this advice had to be reversed. Indeed, when epidemiological claims have been tested in clinical trials—which can demonstrate cause and effect—these claims are shown to be right only 0-20% of the time. These are poor odds on which to bet the public health
The case for using epidemiology has been made in instances where clinical-trial data are lacking, but as described above, in this case, there are numerous clinical trials that have been conducted—they’ve simply been ignored or forgotten. These studies clearly need to be entered into the record and properly prioritized as a more rigorous kind of evidence.
The lack of nutritional sufficiency is also urgent problem for the Guidelines. Preliminary modeling has shown that this this problem could be eliminated by lifting the caps on saturated fats, since many of the borderline or inadequate nutrients, such as iron, and folate, are found uniquely and in their most bioavailable form, in animal foods.
Finally, there’s the question of the target population of the Guidelines. If they’re only for healthy people, then they aren’t for a majority of Americans who are now overweight or obese, with pre-diabetes or diabetes. These and other nutrition-related diseases develop along a continuous spectrum. They are all part of an increasingly poor metabolic state, known as “metabolic syndrome,” which evolves progressively, with formal diagnosis an arbitrary marker (e.g., a person is increasingly overweight until obese, increasingly insulin resistant until formally “diabetic,” and suffering from worsening cardiovascular risk factors until experiencing a heart attack.) The distinction between disease prevention and treatment is therefore somewhat artificial. Nutritional advice to prevent these diseases is very likely similar to advice to successfully treat them. Or, perhaps there should be separate sets of Guidelines: one for healthy people, one for the majority who are metabolically unhealthy, and another set for children who have different nutritional needs. The current one-size-fits-all Guidelines clearly hasn’t worked. Thank you, again, for allowing me to share these views with you.
 Wells, Hordan F. and Buzby, Jean C., “Dietary Assessment of Major Trends in U.S. Food Consumption, 1970-2005,” Economic Information Bulletin, Number 33, USDA Economic Research Service, March 2008.
 Cohen, Evan et al., “Statistical review of US macronutrient consumption data, 1965–2011: Americans have been following dietary guidelines, coincident with the rise in obesity.” Nutrition, Volume 31, Issue 5, 727 – 732.