If you still have faith in Public Health England, last week’s episode of More or Less has the antidote. The BBC’s indispensable fact-checking show took a look at the agency’s claims about sugar. In particular, it examined the claim that kids have eaten 18 years’ worth of sugar by the time they are ten years old. This scary-sounding statistic implies that children’s consumption of sugar is at dangerous levels and many people would infer that it has reached dizzying new heights. In fact, as I have said here and said again on the show, people are eating twice as much sugar as the government recommends because the government arbitrarily halved the recommendations back in 2015. Far from rising, sugar consumption has fallen by around a quarter in the UK since its peak in the 1960s.
Asked to comment on these facts, Public Health England’s Dr Louis Levy flatly denied them. At first, he insisted that sugar consumption has remained the same for decades. When confronted with the data, he admitted that it had gone down but only slightly. In the end, he resorted to saying that it didn’t matter whether sugar consumption had gone down because people were still eating too much of it (according to, er, Public Health England).
On the question of whether there was a good reason for halving the sugar guidelines, he claimed that the 2015 report of the Scientific Advisory Committee on Nutrition (SACN) concluded that there is an association between sugar intake and body weight in adults. In fact, as the presenter Tim Harford pointed out, it found that ‘there is a lack of evidence to draw conclusions on the impact of sugars intake on the majority of cardio-metabolic outcomes in adults, including body weight.’
Nevertheless, the SACN report did recommend that the maximum sugar limits be halved – from 10 per cent of daily energy intake to five per cent – and it was this report that led to Public Health England changing the guidelines. And what a change it was. The meagre sugar ration given to civilians during the Second World War would exceed today’s guidelines. There is no chance of any normal people in a market economy sticking within them.
Nor do they need to. The scientific justification for the five per cent guideline was always feeble. In fact, there were three feeble justifications. The first was tooth decay, but that was a non-starter. Drummond and Kirk sum up the mainstream view when they write:
‘Frequent consumption of sugar and poor dental hygiene may promote dental caries but the overall amount is irrelevant’.
Despite pressure from campaigners, the World Health Organisation kept the sugar guideline at 10 per cent of daily calories in 2015. It did, however, offer an extra ‘conditional recommendation’ of five per cent. A conditional recommendation is what the WHO resorts to when it can’t find much evidence of benefit but doesn’t think it will do any harm either. WHO admitted that the five per cent limit was based on ‘very low quality evidence’ and so it was. It came entirely from the observation that tooth decay declined in Japan during the Second World War when privation led to sugar consumption falling almost to zero. The Japanese had little access to fluoride at the time.
The second justification, which Tim Harford mentioned on More or Less, is that people who consume more sugar tend to consume more calories overall and are therefore more likely to become obese. Although you are at no extra risk of obesity from consuming more than five per cent of calories from sugar so long as you abide by the overall calorie guidelines, exceeding the five per cent limit could make it more difficult for you to stay within those guidelines.
To some extent, this is tautological: people who eat more food tend to eat more food. If you’re the kind of person who eats a lot of food, you’re probably the kind of person who eats a lot of sugar. Sugary food tends to be discretionary – sweets, snacks, desserts, etc. – and is often eaten in addition to the core diet. Some people may find this kind of food relatively easy to cut out. Nevertheless, it is strange to portray a sugar intake of more than five per cent (100 calories for a woman) as inherently unsafe when it is clearly possible to consume, say, 250 calories from sugar and 1,750 calories from the rest of the diet.
And whilst the hypothesis is plausible, is it evidence-based? According to SACN, who first proposed the five per cent limit in Britain, it is. Their report cited eleven randomised controlled trials which, they say, ‘demonstrated that relative changes (increases or decreases) in the dietary intake of sugars result in corresponding relative differences in energy intake’. But if you read the studies, you will see that they don’t really show that at all and they certainly do not support lowering the guidelines. Most of them are more damning of fat than sugar and several of them explicit contradict the sugar hypothesis. (I have summarised them all at the bottom of this article.)
What a few of them do show is that if you enrol someone in a randomised controlled trial and tell them to drink sugary drinks for several weeks while eating their normal diet, they will consume more calories overall. No doubt the same would be true if you gave them beer or wine, but drinks do not satisfy appetite and these studies tell us nothing about the effect of sugary food – which is where people get most of their sugar from – nor do they offer any clue as to the optimal level of sugar consumption (which was often not measured in the experiments and which far exceeded 10 per cent, let alone five per cent, of daily energy intake when it was).
The idea that people who consume more than 30 grams of sugar are prone to consuming more calories is flawed in theory and in practice. The empirical evidence (or, at least, the empirical evidence cited by SACN) does not support it, and it seems likely that the mechanism works in the opposite direction, ie. that people who consume more calories are likely to consume more sugar. But even if the hypothesis were true, it would not justify changing the guidelines for how much sugar individuals should consume as a proportion of daily energy intake because the daily energy intake is assumed to be fixed.
That leads us to the third possible justification for halving the guidelines which was also put forward in the SACN report. The committee estimated that a reduction in overall energy intake of 100 calories a day would ‘lead to a moderate degree of weight loss in the majority of individuals’. Seeing this as beneficial to the health of the nation, they noted that:
To achieve an average reduction in energy intakes of 418 kJ (100 kcal/person/day) using this estimated effect size, intake of free sugars would have to be reduced by approximately 5% of total dietary energy
(They also said that ‘the estimated figure of 5 per cent should be treated with caution’ but that warning got lost in the noise.)
SACN admitted that:
Although the dietary recommendation for free sugars has been derived from calculations assuming that the whole calorie reduction would come from the effects of reduced sugars intakes, it is acknowledged that in reality this would not be the case.
And that is the real point. Why focus on calories in sugar specifically? SACN’s flimsy answer was that ‘decreasing the population intake of free sugars is one step that could be taken to help reduce the current UK overconsumption of energy.’
That is trivially true, but with 89 per cent of an adult’s calories coming from sources other than sugar, it is not the obvious place to start. Decreasing population intake of fat or carbs would also help reduce overall calorie intake, but SACN didn’t change the guidelines for either of those. In any case, the issue of ‘overconsumption of energy’ is already covered by the overall calorie guidelines. If people are going to ignore those, why would they abide by the much tougher sugar guidelines?
SACN were clearly keen to send out a message that people should eat less, but the tools at their disposal were quite unsuited to the job. They were tasked with setting guidelines for individual nutrients as a percentage of overall calorie intake and those percentages were always going to add up to 100. If 100 per cent of your daily energy intake adds up to 4,000 calories, it doesn’t matter that you’re only consuming 30 grams of sugar. Conversely, if 100 per cent of your intake adds up to 2,000 calories a day, it doesn’t matter that you’re consuming 70 grams of sugar.
This is the circle that SACN and Public Health England have never managed to square. They have never been able to tell us what harm will come to a person who ignores the sugar guidelines but abides by the calorie guidelines (and brushes their teeth). They never will – because there is none. Insofar as lowering the sugar guidelines had a scientific basis, it was essentially: ‘People eat too much, here’s one thing they could eat less of’.
The general public can be forgiven for not realising that the five per cent target is an evidence-free ‘nudge’. If Public Health England says that it is inherently dangerous to consume more than 30 grams of sugar per day, why would they doubt it? And if the government wants to set about taxing and reformulating the nation’s food supply in a doomed attempt to reach a target that is wholly arbitrary, who are they to complain?
The studies cited in the SACN report:
Drummond and Kirk (1998) advised one group of men to reduce their fat intake and another group of men to reduce both fat and sugar. After six months, the men who had only been advised to reduce fat intake had lost the most weight. The authors conclude that ‘it is surprising that many dietary guidelines recommend a reduction in sugar.’
Drummond et al. (2003) used the same methodology as above but this time neither group succeeded in reducing their fat intake. Although the men who were told to reduce their sugar intake did so, they compensated by consuming more calories from other carbohydrates and therefore did not reduce overall energy consumption.
Reid et al. (2010) gave a litre of sugary drinks to a group of overweight women every day and gave a litre of artificially-sweetened drinks to another group of overweight women. They were not told whether their drinks had sugar in them or not. Those given the sugary drinks increased their overall energy consumption in the first week but, surprisingly, energy intake then fell and within four weeks it was lower than it was when the study began – despite sugar now making up 20 per cent of these women’s daily calorie intake. This was because they reduced their consumption of fat, protein and non-sugar carbohydrates, totally offsetting the extra sugar. They did not gain weight.
Saris et al. (2000) divided 398 obese men into three groups. One group was given a diet low in fat and high in sugar while another was given a diet low in fat and high in complex carbohydrates. The third (control) group ate a normal diet. After six months, both low-fat groups had lost weight, with the high complex carbohydrate group losing the most. The authors note that ‘[i]n the search for dietary factors associated with the development of obesity, sugar intake is commonly proposed. However, epidemiological and experimental data do not support this idea.’
Aeberli et al. (2011) split their subjects into six groups and gave five of them soft drinks with varying degrees of sucrose, fructose and glucose. The control group was advised to reduce fructose intake. The researchers found that ‘total energy intake did not differ significantly between the baseline and any of the 6 interventions.’ It is not clear what relevance this has to the question SACN sought to answer.
Brynes et al. (2003) split seventeen men into four groups and gave each group different diets. Those on the high fat diet spontaneously consumed more calories, those on the high sugar diet did not.
Poppitt et al. (2002) split 39 people into three groups in the same way as Saris et al. (2000) above. Both low-fat groups lost weight, with the high complex carbohydrate group losing the most. The authors conclude: ‘A high sugar intake has been proposed as a causal factor in the etiology of obesity. The results of epidemiologic studies, however, oppose this view and are supported by our current trial. Despite a considerable increase in sugar intake, there was no evidence of weight gain in the LF-SC group.’
Njike et al. (2011) gave one group of men two cups of sugary cocoa a day and gave another group two cups of sugar-free cocoa a day. After six weeks neither group had lost weight compared to a group of men who were given a placebo, although the sugar-free cocoa ‘had a modest, favorable influence on waist circumference’. Overall sugar consumption was not measured.
Raben et al. (2002) split 41 overweight people into two groups. One group was given a very high sugar diet (28 per cent of daily energy), mostly in the form of soft drinks. The other group was given a large quantity of artificially sweetened soft drinks. Both groups were told that their supplements were artificially sweetened and they could eat whatever they wanted. Unsurprisingly, the high sugar group consumed more calories and gained weight, but the additional calorie intake was entirely due to the extra sugar. Energy intake from other food fell, partially offsetting the additional sugar calories.
Reid et al. (2007) gave half their subjects a litre of Irn-Bru a day. The other half were given Diet Irn-Bru. Specially branded bottles were marked ‘sugar’ or ‘diet’ but these were deliberately incorrect 50 per cent of the time. Subjects given the sugary drinks increased their overall energy intake, with calorie consumption rising by around half the amount of that contained in the drinks, ie. they partially offset half of the calories by reducing consumption of fat, protein and non-sugar carbohydrates. There was surprisingly little weight gain, perhaps because all the women were on a low fat diet. ‘Overall,’ the authors conclude, ‘there was no evidence that sucrose was a unique or problematic substance.’