Wednesday, February 09, 2005

No, fruit juice will not make your children fat

A recent study that got a lot of media coverage was billed as showing a link between fruit juice and obesity in children. The study, by Welsh et al. (2004), was published in Pediatrics.

A typical article on the study was the CNN (via AP) article, "Study links juice, chubby children" (Feb. 7, 2005). The article starts off by summarizing the findings of the study:
"Sweet drinks -- whether Kool-Aid with sugar or all-natural apple juice -- seem to raise the risk of pudgy preschoolers getting fatter, new research suggests.

"That may come as a surprise to parents who pride themselves on seeking out fruit drinks with no added sugar.

"'Juice is definitely a part of this,' said lead researcher Jean Welsh of the Centers for Disease Control and Prevention.

"While fruit juice does have vitamins, nutritionists say it's inferior to fresh fruit. The new U.S. dietary guidelines, for example, urge consumers away from juice, suggesting they eat whole fruit instead.

"The bottom line, though, is that 'children need very few calories in their day,' Welsh said.

"'Sweet drinks are a source of added sugar in the diet.'

"She said preschoolers were better off snacking on fruit or drinking water or milk.
The article goes on to discuss some of the study's results in more detail, but it is this introduction that is particularly worthy of note. After reading this, it sounds as though juice has been shown to be just about one of the worst things you could give your child to drink, as it's equivalent to Kool-Aid (functionally uncarbonated soda), and is just added sugar that will cause your child to gain weight.

The study's results are quite different. In their study Welsh et al. acquired height, weight, and diet data on more than 10,000 2-3 year olds, re-sampled the same children a year later, and then looked to see if there was a correlation between weight gain and sweet drink consumption. Note the terminology - the primary correlate that Welsh et al. examined in their study was not juice consumption, but instead sweet drink consumption, which they define as
"all sugar-sweetened and naturally sweet drinks listed on the HFFQ: 'vitamin C juice (orange juice or juice with vitamin C added),' 'other juices,' 'fruit drinks (Hi-C, Kool-Aid, lemonade),' and 'soda (soda, soft drink, pop [not sugar-free]).'"
It is true that Welsh et al. also examined the correlation between weight gain and sweet drinks excluding soda, and just pure fruit juice, but the CNN article focuses on the sweet drink results.

One of the methodological problems in Welsh et al.'s study is that children who consume more juice are likely to consume more calories, and thus any weight gain observed might be due solely to the excess calorie intake, not due to juice intake by itself. Welsh et al. did, in fact, find that children who consumed more juice consumed more calories overall in their diets: "Energy intake increased as the consumption of sweet drinks increased with mean calorie consumption for those who consumed 0 to <1, 1 to <2, 2 to <3, and ≥3 drinks 1425, 1596, 1771, and 2005, respectively (data not shown)." To control for this possible confounding effect they used a logistic regression analysis that included many possible variables, including (in their most controlled analysis): "age; gender; race/ethnicity; birth weight; and intake of sweet foods, high-fat foods, and total energy." With this analysis they found:
"Normal or underweight children who consumed 1 or more sweet drinks daily were 1.3 to 1.5 times as likely to become overweight as the referent group (<1 drink daily), but these results were not statistically significant. Children who were at risk for overweight at baseline and consumed 1 to ≥3 sweet drinks daily, however, were significantly more likely to become overweight than the referent. ... Similarly, overweight children who consumed 1 to ≥3 sweet drinks daily were more likely to remain overweight"
As I've mentioned before, when a result is not statistically significant you must say that there is no effect of the treatment. So, a translation of the above paragraph would read, "Sweet drinks had no significant effect on the liklihood of normal or underweight children to become overweight, but children at risk for overweight and overweight children were more likely to become or remain overweight if they consumed sweet drinks." So, only if your child is overweight or at risk of being overweight does consumption of sweet drinks have a demonstrable effect on their weight.

The CNN article does include this result, saying,
"The link between sweet drinks and being overweight showed up for all three weight categories, although it wasn't statistically significant for the normal and underweight children."
However, the CNN article does not translate the statistics for their readers anywhere nearby in the article. Thus, anyone not well-versed in statistics will likely not understand the last portion of that sentence, and may erroneously come to the conclusion that sweet drinks cause weight gain in all weight categories of children.

Remember, however, that the authors define sweet drinks as being any sugary drink, including sodas, Kool-Aid, and the like. Thus, this result has no bearing on the consumption of fruit juice alone. So, what did the authors find regarding fruit juice?
"With fruit juice only, we found no significant associations for at-risk or normal/underweight children (odds range: 0.8–1.2). Among children who were overweight at baseline, the association with overweight was positive, although the strength was diminished (odds range: 1.3–1.5), and the results were of only borderline significance."
Generally, when something is statistically significant, authors say so, so "borderline significance" can probably be translated as "not significant" for our purposes. Thus, the results of Welsh et al.'s study when examining solely fruit juice consumption are exceedingly simple to summarize: consumption of fruit juice was unrelated to the risk of children of any weight becoming or remaining overweight. Alternately, if we wanted to be generous with our statistics, we could say that fruit juice may have had a difficult to detect effect on the weight of already overweight children, but no effect on any other weight class of child. In either case, pure fruit juice has not been shown to be correlated with children becoming overweight.

The CNN article ignores this distinction, however, and goes on to discuss how a school banned fruit juice (implying that it was a good thing), and to discuss why drinking sweet drinks might cause overweight:
"Sweet drinks are high in calories and low in fiber. Nutritionists believe that calorie-dense, low-fiber foods may lead to overeating because those foods are quickly consumed but less filling than foods higher in fiber.

"The authors suggest that limiting sweet drinks may help solve the growing problem of childhood obesity. One in five American children is overweight, according to the National Institutes of Health.
To be fair, by this point in the article, the article's author has mentioned that sweet drinks were defined to include things other than juice, but the article never clarifies that when examining solely juice there were no significant effects. So, while what the article says here is true, by failing to clarify that these results don't apply to juice alone, they leave parents with the implication that even drinking pure fruit juice is bad for their children. Then the article reinforces this erroneous conclusion with its ending anecdote:
"But Dr. Rebecca Unger, who evaluates overweight children in private practice and at Children's Memorial Hospital in Chicago, said the study backs up what she sees in the real world.

"'We do see kids do well when we cut out juice,' she said. 'Sometimes that's all they need to do.'
So, even though most of the CNN article is technically correct, the implications that permeate it regarding drinking pure fruit juice are entirely without basis in the published paper. Children have to drink something when they're thirsty, and this study shows that 100% fruit juice is likely a better alternative than many of the other drink choices out there (<100% juice, lemonade, Kool-Aid, soda, etc.). I doubt many parents will get that take-home message from this article, though.

As a final note, I find the nutritional proclamation to eat a piece of fruit instead of drinking fruit juice to be nonsensical. When I'm thirsty, I want a tall glass of something liquid, not a crunchy, solid piece of fruit.


Welsh, JA, ME Cogswell, S Rogers, H Rockett, Z Mei, and LM Grummer-Strawn. 2004. Overweight Among Low-Income Preschool Children Associated With the Consumption of Sweet Drinks: Missouri, 1999–2002. Pediatrics 115: e223-e229. (Abstract, PDF)

No comments: