Commentary on “Is sugar Toxic” by Gary Taubes
New York Times April 13 2011
Toxicity as a concept
The first point to make about this article is that the title itself reveals the depth of ignorance of the writer. That all substances are capable of being toxic has been known for over 500 years (Paracelsus 1493-1541). Even water and air are both toxic when the body is exposed to either in amounts only slightly higher than normal intakes. Drinking too much water is an unfortunately common cause of death among certain mentally disturbed patients; and the adverse effects of increasing the intake of air has a limiting influence on the use of positive pressure ventilation to save the lives of patients with severe breathing difficulties. It is the amount ingested that determines whether harm results from exposure to any substance.
Thus to ask “Is sugar toxic” is to address the subject from naive stand point. The legitimate question is “does sugar exert toxic effects when consumed in amounts likely to be seen among the general population?” The remarkable observation with regard to sugar is that it is among the very few substances that are so non-toxic that it has not proved possible to identify an intake level that results in harm, even from experiments in which subjects have consumed vastly more than is ever likely to be seen among the general population (Food and Nutrition Board 2003). So the short answer to Taubes rhetorical question is that there is no credible evidence that sugar has proved toxic under any circumstances. With typical journalese sleight of hand, Taubes acknowledges that this has been the conclusion of several serious reviews of the evidence, but only well buried in the middle of the article. But he goes on to assert, incorrectly, that the evidence does not refute his accusations of toxicity. That the evidence does show him to be incorrect is not appreciated by Taubes because he is asking the wrong question. Sugar has been shown not to be toxic at doses likely to be consumed by any sane member of the public. Thus, in all practical meaning of the term, it is non-toxic and has been so described by the ruling of the US Federal Drug Administration that sugar is “Generally Recognised as Safe” (FDA 1986).
In theory there must be a level of intake of sugar at which some adverse effects would be observed in people of normal health. Unfortunately for Taubes, we do not yet know what that level of intake might be. What we do know is that the harmful level of intake lies outside the vast range of intakes so far studied. Clearly a similar question might be posed in relation to people with a pre-existing medical condition, such as diabetes. But that represents an entirely different question from that implied by the appearance of Taubes polemic in a general circulation newspaper, and a question that will not be addressed here.
Science vs. moral panic
The second point that needs to be made is that Taubes’ article, like the primary source he cites so extensively (viz. a YouTube copy of a presentation from Professor Robert Lustig, an expert not in toxicology, or even nutrition, but in paediatric endocrinology) is that the method of argument is hardly scientific, and therefore lacks objectivity. Sugar is described pejoratively as “toxic” and “evil” within the first paragraph. Even allowing for the fashions of mass journalism, this style of argument is, to speak plainly, biased. It is insufficient for the journalist to claim in his defence, that he is merely quoting his source’s views. It is too well known a trick for journalists, unable to produce solid evidence to back a story, to abdicate responsibility for the production of proof of their assertions, simply by finding someone prepared to make a baseless statement that can then be cited as a valid opinion. Where matters of fact are concerned this lazy form of journalism is wholly unacceptable. As a form of argument, Taubes and Lustig provide examples of a moral panic rather than a scientific revelation. Facts are amenable to proof. A belief that something is true has remarkably little relevance to whether it actually is true (Haldane).
Harms attributed to sugar
The first substantive accusation levelled against sugar is that it is “our excessive consumption of sugar (which he equates with HFCS) is the primary reason that the numbers of obese and diabetic Americans have (sic) skyrocketed in the past 30 years”. He then adds heart disease, hypertension and many common cancers to the list of indictments. This fantastic mono-causal theory of most of the common chronic diseases of our time is familiar from many spurious internets web sites. It is however, entirely without foundation. In addition, it is an example of a common failing of assuming simplistic solutions to complex and multi-faceted problems. Or as the redoubtable H.L.Mencken put it “for every complex problem, there is a solution, neat, plausible, and wrong”. In the case of the theory that sugar causes ill-health, however, the hypothesis is not even plausible.
A cursory examination of the available evidence shows that the basis of the theory is flawed. He displays a parochial attitude to evidence that suit his argument, by only considering the increase reported, over the last 30 years, in sugar (and HFCS) consumption that has occurred in the USA. He can then point to a roughly parallel increase in obesity to be a result of this change in sugar consumption. The fact that a similar increase in both obesity and diabetes has occurred in Europe against a background of falling sugar consumption (and a conspicuous absence of HFCS) is ignored. In parenthesis, he also incorrectly asserts that the estimates of sugars “consumption” provided by the US Department of Agriculture are “typically considered the most reliable”, when these are in fact estimates of supply, and thus exceed actually intake by a factor of at least two.
Circumstantial evidence
If we are obliged to take circumstantial evidence such as that posited by Taubes seriously, then let us at least get our facts right. He is claiming that there has been an increase in consumption of sugar over the last 30 years or so, and that this has caused an increase in obesity, diabetes, heart disease, hypertension, and several cancers. And yet, it is a matter of public record that the age standardized mortality rate in the USA (the most reliable proxy measure of prevalence of these diseases) from heart disease has fallen dramatically, and that from stroke and most cancers has also fallen (Rogers et al., 2011). Perhaps Taubes has made the common, but elementary, mistake of taking crude event numbers for these diseases and forgetting that the increase in average life expectancy inevitably exaggerates the crude (non-age-standardized) event rates for any disease that is more common in older people.
Taubes acknowledges that comparison of recent data on the (national average) intake of sugars with population prevalence of obesity and diabetes is only circumstantial evidence. He then sets out to bolster the persuasiveness of this evidence by claiming that a similar surge in both sugar consumption and diabetes occurred in the early part of the 20th century. Please note that by a further sleight of hand he has conveniently forgotten the obesity part of the argument, since the early twentieth century is inconveniently not associated with any great concern about corpulence in the general population (indeed the reverse). But to claim that data from that historical period on either sugar consumption or the prevalence of diabetes was sufficiently reliable to draw any meaningful conclusions is simply risible. Equally, his rehearsal of the arguments played out around this issue at that time are simply irrelevant.
Post-hoc rationalizations
Taubes rewrites the history of twentieth century nutrition in order to portray his ideas as plausible. He claims that Yudkin’s conviction that sugar causes heart disease was discounted because he also argued that fat was not important, at a time when the lipid theory was in the ascendancy. This is, at best, a half-truth. Yudkin’s ideas were certainly taken seriously for some time after the lipid theory of heart disease became generally accepted. Only in 1989 (COMA 1989) were they finally discarded by a UK Government committee that thoroughly examined his evidence quite separately from any prejudice as to alternative theories. This committee took the unusual step of specifically rejecting Yudkin’s theories while naming their author. This would hardly have been necessary if his ideas were not widely taken seriously. Needless to say the principal reason for the rejection of Yudkin’s ideas on sugar as a cause of heart disease is not mentioned by Taubes since they relate to the fact that carbohydrate metabolism in rodents, the subject of Yudkin’s experiments, is not comparable to that in humans. Thus conclusions drawn by Yudkin from rat experiments were not interpretable in terms of human risk. Thus Taubes is simply incorrect in asserting that data on the metabolism of sugars in rats gives useful indications of the likely effect in humans. This is a fact he acknowledges later in the article, but only after the impression that Yudkin (and certain like-minded American researchers such as Reiser) were right all along, has been firmly planted in the minds of the reader.
Metabolic syndrome and cancer
Taubes article wanders over a number of issues culminating in metabolic syndrome and cancer. He quotes lots of opinions; blames fructose; and finishes by saying how worried he is that sugar might be very bad for you indeed. He cites Havel’s work on fructose as evidence of its ill-effects without mentioning that Havel has not shown any ill effects when fructose is consumed at levels seen in the American population. Moreover, the ill effects reported by Havel (Havel 2005) following excessively high consumption of fructose were not metabolic syndrome, obesity, diabetes, heart disease or cancer but a simple and hardly world shattering increase in circulating triglycerides. An increase, incidentally, that is of unknown pathogenic significance, since it was not a result of a high fat diet, the usual cause of this biochemical response (Parks and Hellerstein, 2000).
Conclusions
Taubes article is a familiar, if exhaustive, mixture of selected and exaggerated evidence taken out of context, coupled with a large number of unverifiable (and no doubt carefully selected) quotations from numerous researchers. At no point does he produce a “smoking gun”. He is dismissive of all serious analyses of the evidence from scientific sources, other than the few that suit his case. The reputable, and exhaustive, analysis of all the evidence by the Institute of Medicine for the US Government (Food and Nutrition Board, 2005) concluded that it was not possible to identify a level of sugar intake that actually displayed any toxic effects. Even the controversial review of the data by the WHO Expert Consultation in 2003 (WHO, 2003) did not draw the scaremongering conclusions that Taubes finds so “worrying”. It found no convincing evidence for the involvement of sugar consumption in general in any disease except tooth decay – curiously both a report and a condition overlooked by Taubes altogether.
Taubes exhibits a fundamental misunderstanding of how science works that is not uncommon among journalists. Scientists often have ideas, which they aggrandize with the description “working hypotheses”. These form the basis of the designs of their experiments (assuming they have the resources to carry out any experiments). When conducted these experiments often show that the hypothesis was wrong. At that point the scientist should abandon their idea and seek other hypotheses to test. However plausible, a hypothesis that does not fit the experimental observations is as useful as a chocolate teapot. Unfortunately, some scientists hold on to their pet ideas long after the evidence has accumulated to show that they are wrong. This is called bigotry and is a vice not unknown either in academia or the media. The public deserve a health warning when it is around.
References
Department of Health (1989) Dietary Sugars and Human Disease. Committee on Medical Aspects of Food Policy. Report on Health and Social Subjects No 37. HMSO, London.
Food and Nutrition Board, Institute of Medicine of the National Academies (2005) Dietary Reference Intakes for energy, carbohydrate, fiber, fat, fatty acids, cholesterol, protein, amino acids (macronutrients). The National Academies Press. Washington, DC.
Glinsmann WH, Irausquin H, Park YK. (1986) Evaluation of health aspects of sugars contained in carbohydrate sweeteners. Report of Sugars Task Force, 1986. J Nutr. 116(11 Suppl):S1-216.
Havel PJ (2005) Dietary fructose: implications for dysregulation of energy homeostasis and lipid/carbohydrate metabolism. Nutr Rev. 63:133-57.
Parks EJ and Hellerstein MK (2000) Carbohydrate-induced hypertriacylglycerolaemia: historical perspective and review of mechanisms. American Journal of Clinical Nutrition 71:412-433
Roger VL et al., (2011) Heart disease and stroke statistics – 2011 Update: a report from the American Heart Association. Circulation. 123:e18-e209
World Health Organization / Food and Agriculture Organization. “Diet Nutrition and the Prevention of Chronic Disease”. Technical report series 916. WHO, Geneva. 2003
