Fuzzy and Fizzy

The contested science behind Bloomberg’s ban on large-sized sodas.

NY Mayor Michael Bloomberg as a nanny
The Center for Consumer Freedom ran an ad of New York Mayor Michael Bloomberg as a nanny in protest of his proposed plan to forbid the sale of sugary drinks larger than 16 ounces

Also in Slate, William Saletan looks at whether it’s hypocritical for Mayor Bloomberg to extol the virtues of eating contests. Daniel Engber has previously written on obesity and health care reform, the concept of “fat taxes,” whether fresh produce makes a difference in fighting obesity, and the problem with efforts to fight childhood obesity.

 ”What’s next?” asks the industry-funded Center for Consumer Freedom, in response to a New York City plan to forbid the sale of sugary drinks larger than 16 ounces in restaurants, sports venues, and movie theaters. “Limits on the width of a pizza slice, size of a hamburger or amount of cream cheese on a bagel?” Mayor Michael Bloomberg defended his idea Friday, assuring Matt Lauer of the Today show that “we’re not banning you from getting the stuff. … It’s just if you want 32 ounces, the restaurant has to serve it in two glasses. That’s not exactly taking away your freedoms.”

It didn’t take long for the details of this putative public-health measure to dissolve into a super-sized debate between Libertarians and Liberals, soda jerks and health nuts. But before we all start pouring it on about the costs of obesity and an apartheid of pleasure, let’s consider the scientific facts—contested or otherwise—that led Bloomberg to propose his soda-size control in the first place. Would the pint-cup cutoff reduce obesity, or not?

There are two questions here, both of them well-studied by nutritionists and epidemiologists. First, would people eat or drink less if they were served in smaller portions? Second, would they lose weight if they consumed less sugary liquid?

The answers seem obvious to most people: yes and YES! That’s part of what makes it so hard to study the questions. Since we already think that both super-sizing and soda make us fat, the people who avoid super-sized portions and nondiet soda tend to be the same ones who join the gym, eat fish twice per week, and natter on about the joys of kale. These health-conscious people also tend to have money, and given that obesity and poverty are related in all kinds of ways, it’s tricky to determine which parts of an obesogenic lifestyle might count the most.

Tricky, but not impossible. There are plenty of data, for example, to show that portion sizes have increased over the last 30 years while childhood obesity rates tripled. For a 2011 paper in the Journal of Nutrition, Carmen Piernas and Barry Popkin used national surveys to study the diets of more than 30,000 kids and measure how their junk-food habits changed from 1977 to 2006. They found that soft drink portions had increased by almost one-third. (Once upon a time, a “king-size soft drink” was just 12 ounces.) But sodas and fruit drinks weren’t the only foods that swelled: The energy content of cheeseburgers increased by almost one-quarter, pizza slices jumped by 35 percent, and portions of Mexican food by even more than that.

We know that kids got their junk food in bigger portions in the 1980s and 1990s, and that they got fatter overall. But what’s the rationale for singling out the jumbo soft drinks for government oversight, as opposed to the double-wide pizza and 10-inch burritos? That’s where things get interesting—provided you’re interested in contentious science, slippery datasets, and the connivings of shady technical consultants.

The case against soda starts with the idea that the body’s response to beverages differs from its response to solid food. People regulate their caloric intake without thinking. If we eat more at one meal, we’ll go lighter on the next. But for some reason (the mechanism is still unknown), drinks throw this compensatory mechanism out of whack. A sugar-sweetened beverage might give you as much energy as a candy bar or a heap of mashed potatoes, but you’ll be less inclined to account for it with tempered eating down the line.

Beverages may be less filling, but that doesn’t mean they’re not filling at all. Soda-tax advocates like to talk about the “added calories” from sugar-sweetened drinks, as if drinking a 16-ounce cup of Coke would have no effect whatsoever on your appetite. In 2009, the New York City health department released a PSA that warned consumers, “Drinking 1 can of soda a day can make you 10 pounds fatter a year”—a calculation that only works if you imagine people adding an additional soft drink without adjusting their diets in any other way. A New York Times investigation later discovered that city officials knew they were fudging the facts on soda to make a point.

Still, the evidence of reduced compensation for high-calorie drinks makes them look like a special case, and a good target for public-health research. Quite a few epidemiological studies have been carried out to see if consumption patterns correlate with changes in body mass index. A 2004 paper in JAMA, for example, checked the dietary habits and BMI of more than 50,000 middle-aged women over a span of a decade. After adjusting for smoking status, cereal fiber intake, levels of physical activity, and many other factors, the researchers found that women who starting drinking more sugary beverages midway through the study gained more weight than women who did the opposite. (Strangely, those whose consumption remained steady—high or low—tended to do just fine.) Yet other observational studies have found no evidence that sugar-sweetened drinks are especially fattening.

To resolve these discrepancies, scientists produced a series of meta-analyses that looked for patterns across the whole research literature. In 2006, Vasanti Malik and Frank Hu of the Harvard School of Public Health summed the results from 30 different studies, and found that half of them showed a significant effect—which was enough for them to conclude that “the weight of epidemiologic and experimental evidence” supports a link between sugar-sweetened beverages and obesity. As was the case with the observation studies, though, meta-analyses from other groups came up with different results. A review from last year, by Richard Mattes and David Allison, looked at the same corpus as Malik and Hu but found the evidence inconclusive. Another review from 2007 concluded from 88 studies that there were “clear associations” between soft drinks and body weight, while a review from 2008 combined a dozen studies of children and adolescents and found an association that was “near zero.”

The inevitable next step came in 2011, with the publication of a paper that tried to review the quality of all these reviews—a meta-meta-analysis. The authors of this study, led by Douglas Weed, gathered 17 omnibus accounts of sugar-sweetened beverage research and tested each one according to a standard template. On average, they said, the papers used just “4 of 11 best practices for an objective and systematic methodology.” The implication was clear: Research on sugar-sweetened beverages has been so sloppy and vague as to be inconclusive. But Malik and Hu responded with an editorial pointing out the biases in Weed’s own methodology: “Rather than shedding more light on this pressing public health issue,” they wrote, “[it] obscured important relations.” One hopes the Harvard team is about to publish their own meta-meta-analysis, so that another group can compare the two for the world’s first meta-meta-meta-analysis.

Before you throw up your hands in despair, there’s one more set of facts to consider. Weed and his co-authors work as private research consultants, and their 2011 paper was funded by the Coca-Cola Company. (Weed has given seminars for Coca-Cola’s Beverage Institute that are designed to teach doctors “what recommendations should and should not be made to consumers as a result of epidemiological findings.”) The 2008 review that found a “near zero” effect of sugar-sweetened beverages was paid for with a grant from the American Beverage Association, which ended up hiring its senior author to a full-time position. (There were some methodological problems with that paper, too.) Neither Richard Mattes nor David Allison received industry funding for their skeptical meta-analysis, but both have taken grants, honoraria, and consulting fees from numerous food and beverage companies.

The scientists who have lined up against the sugary-drink hypothesis are largely the ones with financial links to the beverage industry. And while that’s not proof of corruption—lots of researchers take private money, and there’s evidence of both black-hat and white-hat bias in the obesity literature—their approach does fit a disturbing pattern: By counseling caution and harping on uncertainty, the industry consultants are appealing to a paranoid style in American science that (intentionally or not) serves to stave off government regulation.

But even if you’re inclined to believe the toxic-soda theory, important questions remain. Can targeted interventions to reduce soft-drink consumption really prevent weight gain? We’ve seen only a few randomized, controlled trials, and their results are not unambiguous. Would drinking less soda make everyone healthier, or just some people? The data suggest that little kids are better at regulating their liquid-calorie intake than are adolescents, and some adults may be more vulnerable to soft drinks than others. Finally, what would happen if we did nothing at all? Carbonated soft-drink sales have been in decline since the late 1990s even without a soda tax or size restriction, and a study from last year found that Americans’ intake of added sugar has dropped by almost 25 percent in recent years. Meanwhile, obesity rates across the whole population seem to be flattening out.

Bloomberg’s public-health officials argue that it’s better to act now, and the beverage industry would rather they waited for more research (and counter-research) to fill out the details. That’s the science. The rest is politics.