Good Calories, Bad Calories: A Critical Review; Chapter 8 – The Science of the Carbohydrate Hypothesis

coverIntroduction

This is another post in my ongoing series of posts on Gary Taubes’s Good Calories, Bad Calories (GCBC). Several of the publications that Taubes cites for various claims are amazingly obscure. I’m working on getting a couple of them right now, so perhaps this post will be updated in the future when I receive and look through them. In the meantime this post is the fact-checking I have done of chapter eight so far.

Surprisingly it doesn’t contain that much “science of the carbohydrate hypothesis.” It’s really more of an introduction to the similarly-named chapters toward the end of the book. He spends some time with Tokelau (which I will get to soon), discusses what homeostasis is, and then reworks one of his earlier articles about salt toward the end of the chapter.

Not the Introduction

God, I love the beach. I bet it's paradise over there.

Ian Prior in Tokelau, 1971

For the first several pages of chapter eight Taubes discusses a rather interesting cohort study involving the people of the Tokelau Islands in the south pacific. Briefly, Ian Prior, an epidemiologist from New Zealand, decided to study the residents of Tokelau while the New Zealand government wanted to organize a large voluntary migration effort of the peoples from the islands to NZ. Some opted to stay on the Tokelau Islands rather than migrate, which provided Prior an excellent opportunity to study the two populations and compare their lifestyles, diet, and health. As you might guess, the islanders that “immigrated” to NZ ended up having more health problems than those that stayed behind.

Why? What happened? According to the University of Minnesota synopsis of the study, it is due to “the tendency for migrants at all ages to be heavier than non-migrants. Migrants had more diabetes and smoked more, drank more alcohol, and exercised less.1 These factors make the poorer health among the migrants pretty easy to explain. Taubes, however, doesn’t want you to think that any of this has to do with the study results so he straight-up lies to you on page 138:

A number of factors combined to make this higher disease incidence among the migrants difficult to explain. For one thing, the Tokelauans who emigrated smoked fewer cigarettes than those who remained on the atolls, so tobacco was unlikely to explain this pattern of disease.

He goes on to state that the migrants were far more physically active and had a much more “rigorous” lifestyle than the non-migrants because they worked in sewing factories and had to walk to shops. I’m not even kidding. Again, this directly contradicts the actual study results.

So how does Taubes explain the health discrepancy? Why, it’s exclusively due to saturated fat intake, of course! Yes, as it turns out the native diet of the Tokelauans ate a diet largely composed of fish, coconuts, and fruit. According to Taubes it was the coconut oil in the Tokelauan diet that was apparently uniquely protective because coconut oil has saturated fats. Moreover, Taubes also mentions that the migrants to NZ started eating bread and potatoes which contributed to their health problems. And, of course, he has to drag Keys into this saying “If Keys’s hypothesis was correct, the migrants should have manifested less evidence of heart disease, not more.”

Actually, no. The non-migrants should have manifested less heart disease. Taubes is conflating the fatty acids in coconuts with all saturated fatty acids. The fatty acids in coconuts are quite a bit different from that fatty acids found in the foods that Taubes promotes throughout the book (beef tallow, bacon, cheese, etc.) It has been demonstrated to the satisfaction of most people that coconut oil does not have a huge impact on serum cholesterol levels and by extension heart disease. Matter of fact, the first person to demonstrate this was none other than Ancel “Beelzebub” Keys himself!2,3

Aside from alcohol consumption, diet was not even mentioned in the study synopsis as being a potential factor in the health decline. Nevertheless, Taubes credits the saturated coconut oil for the good health of the non-migrants, and blames the bread and potatoes on the poor health of the migrants. What Taubes glosses over are other dietary components introduced to the migrants which include salt, eggs, dairy, and red meats.4

* * *

One of the main goals of GCBC is to upend everything we all thought we knew about nutrition: Saturated fats from animals are actually good for you! Carbohydrates from fruits, vegetables, or grains are actually the enemy! Calories don’t matter! You can gorge yourself of whatever you like, and as long as it doesn’t include carbohydrates you’re in the clear! Dietary fiber is meaningless! I suspect Taubes does this not because these things are true or even that he believes these things are true, but rather because it sells more books. It’s a shame people have to get wildly inaccurate information about how to take charge of their own health for the financial gain of one individual, but that seems to be the case.

Another one of these bits of “conventional wisdom” that he upends is the idea that salt has anything to do with high blood pressure. Page 146:

[P]ublic-health authorities for the past thirty years have insisted that salt is the dietary cause of hypertension and the increase in blood pressure that accompanies aging. […] That’s the hypothesis. But in fact it has always been remarkably difficult to generate any reasonably unambiguous evidence that it’s correct.

He goes on to state that even if by some miracle salt really does somehow influence blood pressure, it has such a negligible effect that cutting salt intake “makes little difference” to our overall health. As evidence he cites a 2004 Cochrane Review by He and MacGregor. Old versions of Cochrane reviews are difficult to find. It seems they usually replace the old ones every time the data gets updated. In this case I was only able to find the 2013 version of the same title by He and MacGregor.5 Let me quote from the end of the review (emphasis mine):

Our meta-analysis of randomised trials of longer-term modest reductions in salt intake demonstrates a significant effect on blood pressure in individuals with both elevated and normal blood pressure. The blood pressure fell, on average, by 5/3 mmHg in hypertensives and 2/1 mmHg in normotensives. These falls in blood pressure would have an immediate and significant impact on population blood pressure and would, therefore, be predicted to reduce stroke deaths by approximately 14% and ischaemic heart disease (IHD) deaths by 9% in individuals with elevated blood pressure, and in individuals with normal blood pressure reduce stroke and IHD deaths by approximately 6% and 4% respectively. It is important to note that these reductions in stroke and IHD deaths were estimated from a previous meta-analysis of prospective observational studies. A recent meta-analysis of 1 million adults in 61 prospective studies demonstrates that the relationship between blood pressure and cardiovascular risk is much stronger than previously estimated. Therefore, the reductions in stroke and IHD with the modest reductions in salt intake might be even greater. Since raised blood pressure is also a major risk factor for heart failure, a reduction in salt intake would likely reduce the incidence of heart failure.

I’ll leave it up to you if you think if a modest reduction in salt intake in fact “makes little difference.”6

* * *

Taubes tries to make the case that CHOs are the real culprit leading to insulin resistance which leads to hypertension and ultimately atheroslcerosis. Page 150 of the hardback Taubes states:

Finally, by the mid-1990s, diabetes textbooks, such as Joslin’s Diabetes Mellitus, contemplated the likelihood that chronically elevated levels of insulin were “the major pathogenetic defect initiating the hypertensive process” in patients with Type 2 diabetes. But such speculations rarely extended to the potential implications for the nondiabetic public.

Taubes is quoting out of context here. The actual quote states7:

Should hyperinsulinemia be the major pathogenetic defect in initiating the hypertensive process in patients with NIDDM, it is unlikely that it sustains the hypertension indefinitely. With increasing insulin resistance, the pancreas must secrete higher levels of insulin. Eventually, its reserve capacity is exhausted and plasma insulin levels fall. Other mechanisms must then sustain the hypertension.

Interestingly, immediately before the cherry-picked quote from Taubes is a statement that contradicts Taubes’s argument:

[H]ypertension and insulin resistance may be different clinical manifestations of a common underlying cellular defect by which the level of cytosolic free calcium is increased and the level of intracellular free magnesium is decreased.

As a matter of fact the entire paper makes the argument that the cause of hypertension is clearly multifactorial.

* * *

Page 150:

Though this carbohydrate-induced water retention and the hypertensive effect of insulin were occasionally discussed in nutrition and dietetics textbooks-Modem Nutrition in Health and Disease, for example, which was published in 1951 and was in its fifth edition by the 1970s-they would appear solely in the technical context of water and electrolyte balance (sodium is an electrolyte), whereas the discussion of hypertension prevention would focus exclusively on the salt hypothesis.

Nope. Not even close. Modern Nutrition in Health and Disease describes a variety of factors associated with hypertension. And when it comes to prevention does not “focus exclusively on the salt hypothesis.” Example from the 1999 edition that I have:

In hypertensives whose blood pressures have been controlled with medications, weight loss or NaCl restriction more than doubles the likelihood of maintaining normal blood pressure after withdrawal of drug therapy. The following lifestyle modifications have been recommended as adjunctive or definitive therapy for hypertension: weight reduction if overweight; aerobic exercise; limited NaCl and alcohol intake; maintenance of adequate dietary potassium, calcium, and magnesium intake; smoking cessation and reduced dietary saturated fat and cholesterol intake for overall cardiovascular health.

It’s a shame Taubes never reads what he cites.

cloud

Refs

Good Calories, Bad Calories: A Critical Review; Chapter 12 – Sugar

Introduction

This post is another in a long line of self-abusive posts in which I spend way too much time combing Gary Taubes’s Good Calories, Bad Calories (GCBC), and usually end up disappointed at the publishing industry and feel depressed for the masses of people that actually trusted Taubes’s “research.”

This chapter is not as bad as many of the others, though. It happens to be about sugar, and the fact is that you don’t need to heavily distort scientific publications to indict sugar. Taubes still manages it, however. If you’d like to read other chapter reviews of GCBC by yours truly, visit my Book Reviews page.

As you read this post you might leave with the impression that I advocate sugar consumption. Nothing could be further from the truth: it is clearly an ingredient that Americans (and perhaps any country with a Pepsi presence) consume far too much. It would only benefit your health to reduce your consumption of it. What I do advocate is intellectual honesty which is why I come to sugar’s defense when journalists like Taubes make up or distort evidence against it.

Not the Introduction

On page 200 Taubes states 

The more fructose in the diet, the higher the subsequent triglyceride levels in the blood. For this reason, fructose is referred to as the most lipogenic carbohydrate.*

*Credit for this observation dates to 1916, to Harold Higgins of the Nutrition Laboratory of the Carnegie Institution.

I’m not sure that Taubes can really state this. The Higgins paper he cites does not really measure lipogenesis. It measures respiratory quotient. In the paper Higgins does say that one might say that levulose (fructose) and galactose, judging by the RQ, could preferentially turn into fat.1 However, Higgins actually observed the sugars being burned:

A study of figure 1 shows clearly that levulose [fructose], sucrose, probably lactose and possibly maltose give indications by the respiratory quotients of being metabolized, and in all probability burned by the fourth to the seventh minute after ingestion “on an empty stomach,” which is quite as rapidly as alcohol in the same subject and much quicker than alcohol in four other subjects.

* * *

On page 200, Taubes states

Although sugar also seemed to raise cholesterol levels, particularly LDL, as would be expected for any nutrient that increased triglyceride synthesis in the liver. In 1992, John Bantle reported that LDL cholesterol in diabetic patients was elevated more than 10 percent on a high-fructose diet after a month, which is comparable to what can be achieved by saturated fats.

For this claim he cites two references. One is a study co-authored by Bantle (although he is not the lead author, Joyce Swanson is).2 That study does show that at a couple of points in the study LDL was elevated compared to the controls, as was triglycerides at one point and total cholesterol at one point. (See pic below) However, contrary to Taubes’s claim this study used normal, healthy patients – not diabetics.

Metabolic effects of dietary fructose in healthy subjects

Dietary fructose effects on lipoprotein metabolism and risk for coronary artery disease

The other citation for his claim that fructose elevates LDL is a review article that makes more of a point to state that the evidence on this issue is quite conflicting.3 (see pic above) From the author’s conclusion:

Perhaps the most general conclusion that could be drawn from this review of the effects of dietary fructose on lipoprotein metabolism is how little we actually know.

Does Taubes even read the studies he cites? And shockingly Taubes now admits that sat fat can raise LDL cholesterol levels!

Also from the conclusion:

[I]t would seem unlikely, based on available data, that dietary fructose at quantities obtainable from natural sources provided in a well-balanced diet would result in any deleterious metabolic effects.

* * *

On page 201 Taubes claims:

[F]ructose elevates blood pressure more than an equivalent amount of glucose does, a phenomenon called fructose-induced hypertension.

What you might think after reading this statement is that if you consume fructose this might lead to some degree of hypertension, but unless you’re a rat you would be wrong. Taubes cites two publications for this claim: one by Hodges and Rebello and one by Hwang et al.4,5 Both mention that fructose consumption leads to hypertension in rats. In fact the title of the Hwang paper is “Fructose-Induced Insulin Resistance and Hypertension in Rats.” However, you will find the opposite is true in humans. The other paper makes it clear that fructose has no effect on blood pressure in humans. From the text: “Fructose ingestion produced no significant changes in blood pressure […]” The authors also mention that lactose and galactose also showed no increase in blood pressure. Although to be fair to Taubes the study does mention that both sucrose and glucose increased blood pressure in humans so he’s not way off base here. Nevertheless, he seems to give all the studies he cites a mere cursory glance before writing about their results and very often getting them wrong.

* * *

Page 202, Taubes claims: “The American Journal of Clinical Nutrition dedicated an entire issue to the deleterious effects of dietary fructose.” Not exactly true. The AJCN did do a supplement of one issue that focused on fructose, and it wasn’t focused on the deleterious effects, either. Some of the reports had nothing to do with fructose and health. Just look at some titles like “Manufacturing, composition, and applications of fructose” and “Worldwide production of high-fructose syrup and crystalline fructose.” Other reports indicated neutral to favorable data regarding fructose and health. For example, one report on diabetes concluded6:

In summary, the side effects of fructose supplementation do not seem at this time to be of particular concern when fructose is ingested in modest amounts.

Another report on the public health implication of fructose states7:

On the basis of currently available information, as reviewed in this monograph, fructose is a valuable, traditional source of food energy, and there is no basis for recommending increases or decreases in its use in the general food supply on in special dietary use products.

Another report says that fructose aids in mineral absorption.8 Another says that fructose may increase physical performance.9 These are all from the same supplement that Taubes mentions. I know this because Taubes even cites yet another report on fructose and thermogenesis that concludes the following10:

The greater thermic effect of fructose, its nondependence on insulin for its metabolism, and its greater sweetening potency compared with glucose are factors that may speak in favor of fructose as a valuable carbohydrate for the dietary management of obesity and NIDDM [non insulin-dependent diabetes mellitus].

But of course there was no way in hell Taubes was going to mention this part of the conclusion. Instead he mentions the part that says more research is needed – a sentiment you can find at the end of literally all nutrition studies and likely any scientific study period.

* * *

Page 203:

In 2002, the Institute of Medicine of the National Academies of Science released its two-volume report on Dietary Reference Intakes (subtitled Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids), and spent twenty pages discussing the possible adverse effects of sucrose and high-fructose corn syrup. It then concluded that there was still “insufficient evidence” to set up an upper limit for sugar consumption in the healthy diet.

Despite Taubes’s attempt to make this seem like a huge scandal because of a wildly incompetent Institute of Medicine, this is pretty standard stuff. No tolerable upper limits (UL) were set for any macronutrient. ULs are set for things like vitamin A and iron; micronutrients at which there is a point where they start to become acutely toxic.

Refs

1. Higgins, H. L. The Rapidity with Which Alcohol and Some Sugars May Serve as Nutriment. Am. J. Physiol. 41, 258–265 (1916).

2. Swanson, J. E., Laine, D. C., Thomas, W. & Bantle, J. P. Metabolic effects of dietary fructose in healthy subjects. Am. J. Clin. Nutr. 55, 851–856 (1992).

3. Hollenbeck, C. B. Dietary fructose effects on lipoprotein metabolism and risk for coronary artery disease. Am. J. Clin. Nutr. 58, 800S–809S (1993).

4. Hodges, R. E. & Rebello, T. Carbohydrates and blood pressure. Ann. Intern. Med. 98, 838–841 (1983).

5. Hwang, I. S., Ho, H., Hoffman, B. B. & Reaven, G. M. Fructose-induced insulin resistance and hypertension in rats. Hypertension 10, 512–516 (1987).

6. Gerrits, P. M. & Tsalikian, E. Diabetes and fructose metabolism. Am. J. Clin. Nutr. 58, 796S–799S (1993).

7. Glinsmann, W. H. & Bowman, B. A. The public health significance of dietary fructose. Am. J. Clin. Nutr. 58, 820S–823S (1993).

8. O’Dell, B. L. Fructose and mineral metabolism. Am. J. Clin. Nutr. 58, 771S–778S (1993).

9. Craig, B. W. The influence of fructose feeding on physical performance. Am. J. Clin. Nutr. 58, 815S–819S (1993).

10. Tappy, L. & Jéquier, E. Fructose and dietary thermogenesis. Am. J. Clin. Nutr. 58, 766S–770S (1993).