Junkfood Science: Sweet kids and hyperactivity

June 04, 2008

Sweet kids and hyperactivity

Decades of parents have feared that the very foods kids love the most could harm them. To each generation of parents, worries surrounding sucrose (refined sugar, high fructose corn syrup, or any of its many names) are new and frightening. But their fears aren’t new. Sweeteners have been studied more than any other food ingredient in history and the very same scares surface over and over again, often more elaborate with each telling... and the science re-examines those fears over and over again, too. Sadly, the reassuring science rarely seems to reach parents.

A unique paper was just published that is one of the most thorough, calm and careful examinations of the scientific evidence surrounding one of the biggest and oldest concerns about sugar: that it might contribute to hyperactivity, ADHD or behavioral problems in children.

What makes this paper especially valuable for today’s parents, is that its author looks at each of the mechanisms that has been proposed for this possible link and reviews the body of evidence and biological plausibility, scientific principles and logic or fallacies surrounding each one. In doing so, he patiently teaches fundamental scientific principles of reasoning. The lengthy paper, published in the current issue of Critical Reviews in Food Science & Nutrition, was authored by Dr. David Benton, BSc, Ph.D., DSc, a food and cognition/behavior specialist at the University of Wales Swansea in the UK. Here is a summary of his review.

Correlations versus fair tests for causation

Dr. Benton opens by explaining that even if observational studies appear to find an association between a higher consumption of sugar or sodas and hyperactivity or behavioral problems, we must be careful to avoid viewing them as proof of a causal relationship. As he repeatedly shows throughout this paper, in which he reviews 109 studies and reviews, we can be led astray by studies that aren’t fair tests. So, while studies observing correlations are often a source of scares, those fears repeatedly don’t hold up when they’re examined in fair tests.

Correlations. Especially in the United States, it’s widely believed that sugar causes behavioral problems in children, he says. Since the 1960s, popular authors have even described sugar as deadly for as many as one in ten people. Early suspicions that sugar causes hyperactivity were generated by reports showing hyperactive children consuming more sugar, but as Dr. Benton describes, those correlations were typically explained by the fact that hyperactive children tend to need, and eat, more energy supplied in sugar; or represented parenting styles that didn’t control sugar intake or behavior. In other words, higher sugar consumption was a reflection of the disorder, not responsible for causing it.

Blinded and placebo-control. Throughout the 1980s and 1990s, to test these correlations, the effects of sucrose were repeatedly examined in carefully designed double-blind, placebo-controlled trials giving children sugary drinks after overnight fasts alone or as part of a meal. They came up null. In other words, while parents reported changes in their child’s behavior when they knew their child had consumed sugar, in blinded trials when they didn’t know, they reported no difference in behaviors when their child was given sucrose or a placebo. Nor were behavioral changes demonstrated when objective measures were used by blinded observers.

In an analysis of 16 double-blind, placebo-controlled randomized clinical trials that had examined sugar and behavior and cognitive performance in 23 groups children, for instance, researchers at Vanderbilt University Child Development Center in Nashville, TN, found no adverse affects of sugar. The authors concluded that “the strong belief of parents may be due to expectancy and common association.” Confirmation bias can get the best of us.

[This has been described by other researchers as the birthday party expectation: put a group of kids together in a party, add energy, and “it must be the sugar in the birthday cake causing them to bounce off the walls” — not the balloons, presents, music, games, pony and clown!]

Real life. Some people have suggested that clinical trials may have been done under controlled laboratory conditions that cannot be generalized to real life. Dr. Benton explains that even in studies where the families lived in their own homes and had all of their food provided to them for weeks, there have been no differences in behavioral effects demonstrated with sucrose or artificially sweetened foods or beverages.

Specially sensitive children. Others have suggested that perhaps certain sub-groups of children are especially sensitive to sugar, such as those with ADHD, sugar reactivity or younger children. There have been a number of studies specifically examining these possibilities and in each, using a variety of behavioral measures, sugar was found to have no statistical effect, he reveals.

Sugar high and crash. Parents who believe their child reacts to sucrose usually report a reaction in less than an hour due to a “sugar rush.” However, if there is a tendency for a rapid increase in blood sugar followed by a rapid fall, explains Dr. Benton, then there would be behavioral effects and differences evident over time. Studies measuring behavior one to three hours after consuming sucrose, however, have found no evidence of an effect at any time point.

Artificial sweeteners. Finally, he says, some have proposed that the failure to find a significant adverse response to sucrose could reflect the artificial sweetener used as the placebo. But, again, studies have tried different non-sucrose sweeteners and found behavior after varying artificial sweeteners didn’t differ, either.

Not all studies are created equal. As much as we want to believe anecdotal reports from parents or what we think we experience ourselves, to scientifically test our perceptions requires studies using double-blinded, placebo-controlled designs. Because of the enormous placebo-nocebo effect and subjectivity of symptoms surrounding things like behavior and mood, conducting well-designed trials that follow the fair test rules are essential for making credible conclusions.

The replication of studies by other researchers is also important before we accept conclusions uncritically. A string of early studies in the 1980s led by the same investigator had all reported that sucrose stimulated anti-social behavior, spurring this concern. While his studies had been originally described as double-blind, says Dr. Benton, the investigator later confessed that they hadn’t been blinded and had been “open trials,” and that “none of the studies used proper control groups, random selection, nutritional assessment...”

Over the decades, several main theories attempting to explain how sucrose might influence behavior have been proposed. Dr. Benton conducted detailed examinations of each one and goes on to describe for readers where the science stands.

Food intolerance

Immunologically mediated adverse reactions are appropriately called food allergies or hypersensitivities, however, in the case of sucrose there is little evidence that such an immunological response occurs, explains Dr. Benton.

Food intolerances are caused by mechanisms other than allergies. There is a rare congenital condition identified in infancy that results in an inability to break down sucrose due to the absence of the enzyme sucrase isomaltase in the small bowel, resulting in watery diarrhea. Therefore, there is a plausible mechanism for food intolerances to have physical effects, but whether sucrose intolerance actually occurs in sufficient frequency among children to explain behavioral problems is another question.

A meta-analysis of well-designed trials that included elimination diets followed by double-blind interventions looking for a role of food intolerances in ADHD and other behavioral disorders found small effects from a wide range of foods, but sugar wasn’t one of them. In studies of children who exhibited behavioral problems that parents believed were due to foods, among some four dozen foods identified to which children could potentially respond to, sugar is way down on the list. “No two children responded in a similar manner — that is the pattern of problem foods was individual. These foods could not be described as highly processed, for example, wheat, diary products, grapes, and oranges were common problems. The picture was of a highly idiosyncratic response,” writes Dr. Benton.

When examining the prevalence of food intolerance being reported as related to behavioral problems, various researchers have found rates of 0.5% to 1.4%. “Thus, it is a minority of children whose parents relate adverse behavior to diet and when this happens there are many foods that are more likely to be a source of problems than sucrose,” he writes.

Reactive hypoglycemia and hyperinsulinism — rapid blood sugar changes

The idea of blood sugar highs, followed by rapid drops, being responsible for adverse symptoms was first described in 1924 by Seale Harris, who characterized hyperinsulinism in tumors of the pancreas. This syndrome of low blood sugar and the epinephrine response (often misinterpreted as anxiety, explains Dr. Benton) includes: weakness, hunger, shakiness, sweating, palpitations, confusion or erratic behavior, and other symptoms. Harris first described this among nondiabetics who responded by eating regular high-protein meals. How often does this occur in the general population?

As many as one-third of women attribute their symptoms to hypoglycemia, according to research reviewed by Dr. Benton, but to accurately assess the body’s ability to control blood sugar levels, oral glucose tolerance tests are used. When 135 patients claiming to suffer from serious reactive hypoglycemia were actually examined, only four proved to really have this disorder. In the general population, hypoglycemia is very rare and even among those who believe they respond adversely to sucrose, very few people actually have physiological problems, explains Dr. Benton.

As Dr. Benton goes on to review, many people believe that high sucrose diets predisposes people to more irritability or aggression and this has been rigorously studied, especially among violent criminals. To cut to the chase, certain violent offenders with impulsive aggressiveness were shown to have low levels of the serotonin metabolite 5-hydroxyindoleacetic acid, a condition predisposing to low blood sugars, meaning the tendency to develop hypoglycemia may merely be a marker for low levels of brain serotonin, not the cause. In fact, a number of double-blind studies on children found “significantly less irritable behavior was displayed in those who had drunk a glucose containing drink rather than a placebo.” Sugar generally decreases irritability, not heighten it.

Most behind blood sugar theories of hyperactivity and behavior or mood changes, he says, is the assumption that sugar generates large swings in blood sugars and that it can cause effects even if the actual blood sugars don’t reach low enough levels to make a clinical diagnosis of hypoglycemia. But in nondiabetics, the fluctuations after eating are normal and harmless. Most illogical is that refined sugar is singled out as being uniquely harmful.

“Given our understanding of the factors that influence the glycemic response to diet, it is surprising that sucrose continues to be selectively singled out as the food largely responsible for swings in blood glucose,” says Dr. Benton.

The change in blood sugars after consuming various foods has been studied and compared to the changes after consuming pure glucose (the glycemic index). Sucrose (sugar) falls in the middle of the scale. Fructose has a very low glycemic index — meaning sucrose (remember sugar science: sucrose is made up of about half fructose and half glucose) has less of an effect on blood levels than glucose. As Dr. Benton explains:

[T]he view that sucrose is uniquely involved in generating rapid swings in blood glucose, or even plays a major cause, is simplistic. There are many sources of carbohydrate that have a greater impact on blood glucose than sucrose. Similarly as sucrose is usually consumed as a sweetener for food items containing protein and fat its impact on blood glucose can be expected to be ameliorated by these macro-nutrients.

In normal people eating as people typically do, a hypoglycemic response is rare.

The underlying assumptions behind the serotonin mechanism has also been questioned by scientists, as Dr. Benton goes on to explain. Some people have proposed that high-carbohydrate meals and sugars raise blood sugar and the release of insulin, causing tryptophan to bind to albumin and other long-chain neutral amino acids (LNAA) to be taken up into muscle — meaning more tryptophan might cross the blood brain barrier to be metabolized into serotonin and raise moods. This is behind the popular perspective that depressed people self-medicate by eating sugar, implying that they seek out sugars. Thus, in this case, correlations between mood problems and sucrose intake is not the cause of the mood problems, but an attempted solution. Reverse causation to observed correlations.

But even the basic assumptions behind this proposed mechanism has scientific problems, he explains. For this effect to be attributed to sucrose would require the sugar or soda to be consumed by itself and so long after the previous meal that no protein remains in the gut. That’s because even high carbohydrate foods such as bread, potatoes and rice contain enough protein to block the mechanism. However:

It is unreasonable to expect that very frequently canned drinks will be consumed to the exclusion of all other food for many hours. From choice nobody is going to exclusively consume a diet of carbohydrate, as it would be unpalatable and unbalanced. In fact, a chronic lack of protein in the diet would decrease rather than increase levels of tryptophan, as protein is the only source of tryptophan.

It also takes a ratio of tryptophan-to-LNAA of at least double before increases in serotonin synthesis can be demonstrated, even in rat studies. But, in humans, even a meal of 100 grams of carbohydrate fails to significantly elevate the level of tryptophan in cerebral spinal fluid, he explains.

He also cautions against making generalizations about reactive hypoglycemia from studies which examine pure glucose taken alone after fast periods. They aren’t accurate tests of real life. “There is surprisingly little evidence of marked changes in blood glucose concentrations in healthy subjects during their everyday life,” he explains. “A meal tolerance test is a better reflection of everyday life.” Typical meals rarely stimulate reactive hypoglycemia because fats and proteins mitigate the response.

As he concludes:

In normal individuals, fed in a usual manner, a hypoglycemic response is so uncommon that a number of professional bodies have issued public statements. The American Diabetes Association, The Endocrine Society, and The American Medical Association jointly issued a statement (Statement, 1973). They stated that widespread publicity has: “led the public to believe that there is a widespread and unrecognized occurrence of hypoglycemia in this country. Furthermore, it had been suggested repeatedly that the condition is causing many of the common symptoms that affect the American population. These claims are not supported by medical evidence.” The American Dietetic Association endorsed the view that, “Valid evidence is lacking to support the hypothesis that reactive hypoglycemia is a common cause of violent behavior...”

The “empty calorie” argument

Perhaps, the most feared effect of sucrose and other refined foods, says Dr. Benton, is that they provide calories stripped of micro-nutrients.

The role of vitamin supplementation in reducing violence, cognition or anti-social behaviors has been the focus of research, but the findings remain inconclusive and controversial. It is biologically conceivable that certain deficiencies could affect neurotransmitters in poorly nourished populations and that there might be a role for supplements, but research to date suggests this would affect a minority of poorly nourished children.

But “even if you are convinced that micro-nutrient deficiencies exist,” he emphasizes, that doesn’t mean that sugar is responsible for lower-quality diets or that removing sugar will improve the quality of the diet. Such “advice is simplistic,” he adds.

It is well-recognized by nutritionists that we do not consume foods, rather we consume diets. Much of the population, rather than seeing such a view as a cliché, function using a child-like classification: There are two types of food, good and bad.

“Adequate nutrition demands the intake of a wide range of nutrients,” he reminds us. The entire diet is what must be examined, not single foods in isolation. While logically, at some point, consuming too much of any food would come at the expense of other foods and make a diet too unbalanced. But, among populations in the real world setting, there is no evidence of micro-nutrient deficiencies due to excessive sugar.

Although it might be thought mathematically inevitable that for a given energy intake a higher consumption of sugar will be associated with a lower intake of micro-nutrients, such a view depends on an assumption that may not stand examination. A high proportion of the energy provided by the diet comes in the form of fats that, with the exception of fat soluble vitamins, are a poor source of micro-nutrients. Given the frequently reported inverse relationship between fat intake and sugar intake it is likely that an increased consumption of sugar will be at least partially at the expense of the fat content of the diet... An assumption that the removal of sugar from the diet will result entirely in its replacement with micro-nutrient rich items may not be valid.

Reviews of the quality of children’s diets and associations with sugar have found variations with age, with both positives and negatives. But the bottom line is that the effects of sugar on micro-nutrients are “always so small as to be of no clinical significance.” In fact, “added sugars have little or no association with the diet quality of individuals over the age of two years, children or adolescents. In epidemiological studies the use of large sample sizes can produce statistically significant findings of little clinical significance.”

He reviews study after study finding that sugar intake among children in real life is directly related to an adequacy of micro-nutrients. Even among children at the highest amounts of sugar, nutritional quality of children’s diets appears to be adequate. There are simply “few grounds for concern,” he says. “After considering obesity, behavior, and a range of other disorders, an American expert report found that there was insufficient evidence to set a tolerable upper limit for the intake of total or added sugars (Food and Nutrition Board, 2005).”

Reviews of the evidence, both here and in Europe, have repeatedly concluded that current levels of sugar intake do not compromise micro-nutrient status, he says, even among those eating the highest amounts of sugar. There is little credible support for the view that a diet high in added sugar causes deficiencies, he explains. Why? “Those eating a lot of sugar also tend to eat more of all nutrients, and total energy consumption is a better predictor of micro-nutrient status than the level of sugar intake.”

The greatest risk for nutritional deficiencies is among those not eating enough or eating restrictive diets. The research consistently finds that “in those who are consuming an adequate number of calories the level of micro-nutrient intake generally achieves recommended levels, irrespective of the sugar intake,” he explains. [Popular concerns of widespread vitamin and nutritional deficiencies was reviewed here.]

It’s easy to worry that children with poor appetites or who are picky eaters and eat only a limited variety of foods are at higher risk for deficiencies. But children have naturally greater preferences for sweet tastes than adults and removing sugar from their diets could not only reduce vital energy, it could actually worsen their diets, he explains.

We have a genetically determined predisposition to like sweet tastes, sugar may be more likely to be a larger proportion of a limited diet. The risk is that if sweet tasting items are removed from an already inadequate diet, in those prepared to eat a limited range of foods, then the diet may become worse... In fact a pleasant sweet taste could be part of the solution to the extent that it can be used to encourage the eating of foods that would not otherwise be consumed. For example it is the sweet sauce that makes baked beans acceptable, a food item that many children would otherwise find unattractive.

It’s the Mary Popins adage that a spoonful of sugar helps...

Theories abound that minor, asymptomatic deficiencies could have a cumulative effect on the brain and behavior. While such arguments may be fine in theory, he said, “hard evidence is required before it can be taken seriously.” The body of evidence continues to suggest that in a diet with sufficient energy (calories), any dilution of micro-nutrient intake associated with sugar consumption is so limited as to be of little concern. With micro-nutrient deficiencies most related to diets low in overall energy, if we’re worried about nutritional adequacy, the most likely successful intervention would be to encourage the consumption of more food in general.

As Dr. Benton sums up:

A plausible mechanism must be established before concluding that the correlation between the consumption of sucrose and the behavior of adolescents reflects a causal relationship. Although parents often report that within an hour of the consumption of sucrose their child reacts adversely, a series of well designed studies have resulted in negative findings suggesting that this is in most cases not a physiological reaction. Similarly the evidence does not support the view that sugar is a major cause of low blood glucose... there [is] little reason to associate sugar intake and micro-nutrient deficiency, in those with an overall adequate intake of energy. As there are no obvious mechanisms by which sucrose can influence behavior...

The science continues to offer positive reassurance to parents. There is no credible evidence to support worries that children today are eating excessive amounts of sugar, or that sugar in their diets is contributing to hyperactivity or ADHD or that is a causative factor in any disease. Sadly, the evidence and conclusions of pediatric and mental health professionals often doesn’t break through popular discourse, even after decades.

While it’s understandable that parents of children diagnosed with ADHD would search for explanations, ADHD isn’t caused by sugar or children’s diets. ADHD has a strong genetic root, said Dr. Benton, with half of children of parents with ADHD likely to develop the disorder themselves. Sugar is a source of energy, not the source of ADHD.

“Diet, especially sugar, is not a cause of ADHD,” reiterated Dr. Stephen Soreff, M.D., of Metropolitan College of Boston University in Boston, in a new review for e-medicine. Children of parents with ADHD are 2 to 8 times more likely to develop ADHD than the general population, he also explains.

On a lighter note, after all of this science stuff, it’s easy to forget that we’re talking about a food that children through the ages have had very different relationships with than grownups. If you missed the sweetest essay on sugar from confectionary historian, Tim Richardson, it’s a fun perspective of sweets, just in time for summer.

Bookmark and Share