I hate to side with the American Beverage Association in any argument, infamous as they are for their dubious claims about the health effects of soda, but in this case they're right: this study is nothing more than number crunching topped with a frightening headline to generate attention. The authors "... are at best estimating effects of sugar-sweetened beverage consumption."
But we don't have to take the sugary drink lobby's word for it; let's look at the Circulation study itself:
We modeled global, regional, and national burdens of disease associated with [sugar sweetened beverage] consumption by age/sex in 2010... We computed cause-specific population-attributable fractions for SSB consumption, which were multiplied by cause-specific mortality/morbidity to compute estimates of SSB-attributable death/disability. Analyses were done by country/age/sex; uncertainties of all input data were propagated into final estimates.Where did the data for these estimates come from? "Data on SSB consumption levels were pooled from national dietary surveys worldwide." Ah. Food surveys, those most unreliable sources of information about what people eat and drink. If the numbers going into the models are junk, then the results coming out of them will be useless. It doesn't matter how many times you check your work.
The authors of the study also repeat the questionable claim that sugary drinks "are a single, modifiable component of diet, that can impact preventable death/disability ..." But all high-calorie foods can be linked to morbidity and mortality the way this paper linked them with soda. If you eat poorly, simply eliminating soda will not do much for your health. It's a start, but losing weight and preventing diabetes is not that easy, I'm sorry.
Everybody knows that sugar in great enough quantities is harmful. We have good science to back up that claim. Statistical studies like this one tell us absolutely nothing useful, they only give the soda lobby an easy argument to knock down.