Total Pageviews

Thursday, February 2, 2017

Lustig on processed food: a critical review

In a recent opinion piece in JAMA, Robert Lustig has written a ‘viewpoint’ entitled: “Processed Food – An Experiment That Failed”[1].  This blog attempts to provide a critique of that article. I apologise for the length but I’m afraid it is necessary.
The article begins with a definition of processed food. Not one drawn up by the internationally recognised organisations such as the FDA[2], the FAO[3] or EFSA[4]. Instead he choses the definition established by the prolific food writer Joanna Blythman[5], whose 7 popular food and health books range from: “Swallow this - Serving up the food industry’s darkest secrets” to “How to avoid GM foods”.  Blythman has 7 criteria for defining processed foods:

Þ   Mass produced
Þ   Consistent batch to batch
Þ   Consistent country by country
Þ   Uses specialised ingredients from specialised companies
Þ   Consists of pre-frozen macronutrients,
Þ   Stays emulsified
Þ   Long shelf life or freezer life

It isn’t clear from Lustig’s article (I haven’t read nor do I intend to read Blythman’s books) if a processed food must meet all of the above in which case the vast majority does not (Coca Cola isn’t emulsified and doesn’t have pre-frozen macronutrients; bread is mass produced and consistent batch to batch but not country to country). It isn’t clear whether a processed food must meet a given number of these criteria, some of which are quite subjective. For example, what is a ‘specialised ingredient’ and a ‘specialised supplier’? And what is meant by the term ‘pre-frozen macronutrients’. Who would pre-freeze sugar or fat? In effect, the definition chosen by Lustig is one that suits his general argument: It is vague in detail and all embracing in its scope.

He then goes on to cite his own work to illustrate the nutritional properties of processed foods, of which he raises 11 issues.

1. Fibre. We eat too little fibre for good health, a challenge that we have identified half a century ago. Lustig sees processed food as a poor source of fibre and advises us to seek “real food”, a term that I find a tad challenging as a nutritionist. I published an analysis of strategies to improve the fibre intakes of adult Irish women based on a national dietary survey data[6]. There were four areas, which if taken together, would increase fibre intake in this group from 16 to 23 g/d:  increase whole meal bread intake (from 20% of bread intake as whole meal to 60%), increase % of women consuming breakfast cereals from 40 to 70% and increase intakes of both fruit and pulses from 1 to 3 servings per week. So three processed foods help us achieve the dietary guideline for fibre in women: whole meal bread, breakfast cereals and also pulses, which in Ireland will be mainly tinned or frozen.
2 & 3. Too few omega 3 and too much omega 6 fatty acids. The focus here is on the inflammatory system (pro and anti; fostered by omega 6 and 3 respectively) and completely ignores the role of long chain omega 3 fats in brain function, blood clotting, arterial wall function, vision and reproduction. Lustig puts the omega 6 to 3 ratio at 25:1 (no references supplied) whereas US literature puts the figure at 10:1[7]. No mention is made of increasing oily fish intake (the main source of long chain omega 3 fats) to modify this imbalance to a more favourable level.
4. Too few micronutrients. Lustig talks about antioxidants with vague terms such as ‘cellular damage’. He ignores food fortification, which is one route favoured by the WHO to redress micronutrient deficiencies[8]. Thus, recent studies have shown that infantile beriberi (thiamine deficiency) can be eliminated in rural Cambodia by giving pregnant women a local processed fish sauce, fortified with thiamine[9]. Among Irish women of childbearing age, the level of the B-vitamin, folic acid in their red blood cells, increases as they consume more folic acid fortified foods, mainly in the form of breakfast cereals[10]. Again, we see processed foods as part of the solution and not necessarily the problem
5. Too many trans fats. These were regulated in the US in 2014 and in the EU a decade earlier. In Ireland, the Food Safety Authority in 2008 noted “the low levels of TFA observed in the current survey corroborate industry commitments to reduce levels of TFA in fast-foods in Ireland” but the report went on to warn about replacing trans fats with saturated fats[11].
6. Too many branched chain amino acids. Lustig says that when they are consumed in excess, they are metabolised (deaminated to be technical) in the liver, causing fat accumulation. This to me was mind blowing. In all the years I’ve spent in nutrition research, I had never associated branched chain amino acids (BCAA) with processed foods and never associated them with liver fat accumulation. Lustig references his own published paper here. It is not a BCAA study but yet another of his reviews with a catchy title: “Sickeningly sweet: does sugar cause diabetes? Yes”. One systematic review in the literature concludes that plasma BCAA levels may be a  “useful biomarkers for early detection and diagnosis of insulin resistance in non diabetic patients with obesity.” However, they add that this is strongly influenced by ethnicity, genes, gender and diet[12]. A Japanese group (The Takayama Study) followed 13,525 adults for 10 years and concluded: “Data suggest that a high intake of BCAAs may be associated with a decrease in the risk of diabetes[13]”. So this somewhat limited literature on BCAA and diabetes is quite inconclusive. Lustig maybe is mixing up BCAA intake and plasma BCAA profile, which are horses of quite different colours. It is beyond me as to how BCAA intake in the diet might be linked to processed foods. I would imagine it is linked primarily with protein intake.

7. Too many emulsifiers. According to Dr Lustig, the consumption of too many emulsifiers (they allow fat and water to mix) in processed food, “may strip away the mucin layer that protects intestinal epithelial cells, predisposing individuals to intestinal disease or food allergy”.  The reference cited is to a web site on super weeds where the word “emulsifier” isn’t mentioned. So, this says to me that a sloppy editorial approach was taken to this viewpoint. Leaving aside the absence of any meaningful reference to back up this bizarre statement, Lustig ignores the largest source of human gut exposure to a very powerful emulsifying agent: human bile which daily secretes about 10-12 grams of bile acids into our guts. This bile allows ingested fat to mix with the watery nature of the gut contents so that enzymes can digest the fat. Nothing but nothing that we have in processed foods comes near that. It is quite simply an absurd point for Lustig to make.

8. Too many nitrates. The author associates cured meats with nitrates and increased risk of colon cancer. In fact, cured meats are rich in nitrites but vegetables and fruits are rich in nitrates. For example, 100 grams of cured ham contains about 1 milligram of nitrite. Of plant based foods, the next highest are 100 grams of French fries with about 5 times less (2 milligrams) nitrites[14]. When it comes to nitrates, spinach has a staggering 125 times more nitrates than any processed meat. So what is the dietary problem: Nitrates or nitrites? In fact the problem is overall meat intake of which fresh unprocessed red meat is the biggest culprit, at least according to the World Cancer Research Fund[15] and its link to cancer has nothing to do with nitrates or nitrites. US red meat intakes accounts for about 55% of total meat intake with processed meat contributing just 25%[16]. Processed meat intake rich in nitrites should be reduced but more so, we should focus on unprocessed red meat.

9 & 10. Too much salt and alcohol. Few can argue with this and it’s nice to see wine, beer and spirits included as processed foods.

11. Too much fructose. This is Dr Lustig’s signature tune. However, while his headline is fructose, the text refers to sugar, which is 50% fructose and 50% glucose. Overconsumption of fructose in alone can only be achieved in humans under experimental conditions and most of those studies exceed the existing 95th percentile of normal fructose intake (from sugar) in the US diet[17].  Lustig shares the same principles of Gary Taubes who argues that not all calories are equal and that sugar is more fattening per calorie than fat, a fact refuted by the recent experiments of Kevin Hall[18] of the NIH Diabetes group, incidentally funded by the Taubes-Lustig Nutrition Science Initiative (

Based on all of the above, the case now moves to four concluding points. The first is on the US diet and involves a series of figures on changing nutrient intake in the US, and as expected, a conclusion is reached that says the US caloric intake has risen dramatically in last few decades and this is dominated by sugar and processed foods. Not a single reference appears in this section so I can’t argue with the exact figures.  The changing pattern of food intake in the US has two opposing schools of thought. The difference between the two is the estimate of food wastage in the US[19] and you can believe whichever you like. As I have come to expect in such arguments, the compelling evidence of physical activity in health protection[20] and the decline in work-based physical activity[21] are utterly ignored. The next overarching point is that all this change in food intake leads to ill health. Here, would you believe, the focus is on sugar. The level and composition of dietary fats don’t seem important to mention nor do issues such as minerals (good and bad) nor vitamins. Thirdly we move to the environment and again the focus is on sugar where it is argued that sugar production and corn monoculture leads to soil erosion and loss of arable land which is associated with herbicide resistance and super weeds. Nothing of ruminant methane production and its contribution to greenhouse gas loads, notwithstanding that these dominate the climate change agenda. A report of the FAO on Global Food and Climate change concluded: “However, for many crop and country combinations there was not enough evidence to draw any general conclusions; thus, the authors concluded that the evidence was either inconclusive, absent or contradictory for rice, cassava and sugar cane[22]”. The final conclusion that Lustig draws is in relation to the top 10 global food companies. However, he makes the common mistake of thinking that these are the dominant suppliers of food in general or processed food in particular. One major review concluded: “Although the top ten soft drink companies account for half of global sales, the top ten packaged food companies account for only a small proportion of market share with most individual companies contributing less than 3.3% each. Major multinational companies need to be joined by the myriad of small- and medium-sized enterprises in developing and implementing programs to improve the health of the public, globally[23]

And so, back to the beginning, where Lustig starts off by saying that: “Those of who have participated in science know that 9 out of 10 experiments are failures”. Having spent 4 decades as an experimental biologist in human nutrition, I can say that based on my experience, the good doctor is talking through his hat. I did have one experiment that failed and was binned forever when a fish oil encapsulating company mixed up the placebo and fish oil capsules half way through a large human dietary intervention study. Scientists start off with hypotheses, which they do not try to prove true but rather, they try to prove false. Whatever the result yields it is more knowledge than we had before allowing a new aspect of the hypothesis to be examined and on and on.

 Now the idea that food processing was an experiment that failed is absurd. Food processing took the wheat from flour to make bread, cakes, biscuits; it took grapes to make us wine and milk to make us butter, cheese and yogurt. It allowed us to preserve food through pickling or smoking or drying. And all of these are millennia old. Popcorn was used as a garland for young Aztec women in ceremonial dances. Pizzas were sold in Venice 800 years ago and every city in Europe had its unique patented sausage: Chorizo, Bratwurst, Biroldo, and Salami etc. Canned food came centuries ago and frozen foods a 100 years ago. High sugar foods such as cakes or biscuits were common in ancient Rome such as Cantuci or Savillum. Fundamentally, from the advent of fire, man has transformed the food chain to make it safer and tastier and modern times have done little to add to this except to make food cheaper and more available. I have argued this area at length in my recent book: “Ever seen a fat fox ~ Human obesity explored”[24]

All in all, this ‘Viewpoint’ would be failed if an undergraduate student were to present it, at least in my time. The editor of JAMA paediatrics should be ashamed of his or herself for gross negligence of the peer review process.

[1] Lustig RH (2017) JAMA Pediatrics, January23
[6] Gibney m ( ) Br J Nutr
[7] Kris-Etherton PM (2000) Am J Clin Nutr;71 (suppl):179S–88S
[9] Whitfield KC (2016) JAMA Pediatr. Oct 3: 170
[10] Hopkins S (2015) Am J Clin Nutr.  Jun: 101 (6):1163-72
[12] Xue Zhao (2016) Journal of Diabetes Research, Article ID 2794591
[13] Nagata C et al (2013) ,Am J Epidemiol.;178(8):1226–1232
[14] Hord NG et al (2009)  Am J Clin Nutr;90:1–10.
[16] Daniel CR et al (2011) Public Health Nutr. Apr; 14(4): 575–583.

[17] Fructose, High Fructose Corn Syrup, Sucrose and Health Ed: James M. Rippe
[18] Hall K et al (2016) Am J Clin Nutr. Aug; 104(2):324-33
[19]Hall K et al (2009)  PLoS One. Nov 25;4 (11):e7940
[20] Ekelund, U et al (2005) American Journal of Clinical Nutrition, 103 (3): 613–21
[21] Church TS et al (2011 ) PLOS One | Volume 6 | Issue 5 | e19657

[23] Alexander E et al (2011) Global Health. Aug 1;7:26.

Sunday, January 1, 2017

Think sweet ~ stay slim

Think sweet ~ stay slim

Notwithstanding the overwhelming evidence that an excess of calorie intake per se, irrespective of the source of calories, causes obesity, there remains a populist and a policy view that sugar, rather than fat is the primary culprit. Although the biological argument that not all calories are equal and that carbohydrate calories are more fattening has been debunked by sound experimental evidence[1], the case moves more to appetite as the causative factor. Since people don’t eat what they don’t like, an understanding of the “liking” of particular foods becomes important.

To study this phenomenon, two approaches are possible. One is to use a ‘cross sectional’ approach which seeks to compare obese persons and lean persons for their food preferences at a single moment in time. The problem here is that the obese persons may have acquired a view of what putatively ‘causes’ obesity and in so doing, they may bias their response to any study questionnaires. Thus obese subjects at a given point in time might express a lower liking for sugar-sweetened beverages compared to diet beverages because that is what they are conditioned by society to believe. Moreover, obese subjects may in fact use sugar free beverages to manage their weight. The alternative to this approach is to complete a ‘prospective’ study in which people’s likings for food are measured at a single point in time and the group subsequently followed over several years to see who gets obese and who stays slim. That way, the presence of obesity is independent of the answers to the original questionnaire.

A recent paper from the French research consortium “NutriSanté” has examined the effect of liking for fat, sugar and salt on subsequent development of obesity[2]. The NutriSanté study is an internet-based study, initiated in 2009, with a current enrolment of some 65,683 volunteers. Data is gathered at the point of recruitment on many aspects of health such as diet, physical activity, weight, smoking and alcohol habits, weight management and so on and is repeated annually.  Each month, volunteers are invited to complete questionnaires related among other things to food behaviour.  The questionnaire covered salt (11 questions) and sweet (21 questions) tastes plus preferences of fat-and-salt (31 questions) and fat and sweet sensations (20 items).  Some 49,066 subjects agreed to answer these questions. However, when the data were cleaned up to ensure that all respondents had completed every aspect of the overall NutriSanté study, there were full data on 24,776 subjects.

In the 4 years of follow up, 24,112 subjects remained non-obese while 664 subjects became obese. The key sensory findings are thus:

Sensory liking scores
Non-obese subjects
Obese subjects
Statistical conclusions
Liking for fat
Highly different
Liking for sweet
No difference
Liking for salt
Highly different

Those who gained weight had a higher preference for fat and salt but not for sugar.  Another approach to the data analysis is to look across the “liking” score data and to divide subjects according to their ranking, starting at the “low liking” (lowest quarter or ‘quartile’) right up to the “highest liking”  (highest quartile). These data are controlled statistically for all known confounding variables.

Statistical difference in risk of developing obesity between lowest and highest quartiles of “liking”

Fat liking
Yes, up 85%
Yes, up 49%
Fat + salt liking
Yes, up 106%
Fat + sweet liking
Yes, up 37%
Sweet liking
Yes, down 59%
Natural sweetness liking
Yes, down 69%

The authors also looked at baseline food intake (as opposed to liking) and observed no association between sugary food intake and obesity. Indeed, the evidence was that those who did not develop obesity had higher sugar intakes at baseline. The results are summarised below.

Food group intake g/d
Non-obese subjects
Obese subjects
Statistical conclusions
Highly different
Highly different
Processed meat
Highly different
Milk and yogurts
Highly different
Not different
Highly different
Whole grain products
Highly different
Sugar and products
Highly different
Fatty sweet products
Not different
Sugar-sweetened soft drinks
Not different

The authors conclude thus: “This prospective study reinforced results from cross sectional studies, by highlighting that fat liking was prospectively associated with an increased risk of obesity and diet appeared to substantially explain this relationship. Results have also shown that sweet liking is associated with a decreased risk of obesity, and there is no significant association between salt liking and obesity risk”.

Once again, we have a fully public funded comprehensive and prospective study from a highly reputable group, with the study suitably powered for optimal statistical analysis, which challenges conventional wisdom. It will be ignored but, you see, it won’t go away and one day, when the false gods of the demonic sugar sect are reduced to clay, these papers will come back to haunt the high priests of public health nutrition. I wanna be there!!!

By the way, the blog stands now at 231,325 views. Thank you all and a happy New Year to all of you.

[1] See blogs of November 21st 2016 and of August 14th, 2015.
[2] Lampuré et al. International Journal of Behavioral Nutrition and Physical Activity (2016) 13:74