Sugary drinks may kill 184,000 adults per year

Lethal habit?

Lethal habit?

Drinking sugary drinks – like soft drinks and juice – may kill as many as 184,000 adults per year. This according to a new study published in Circulation.

Dariush Mozaffarian, from Tufts University, Boston, and a senior author of the study, said the focus should be on cutting the drinks out of diets in order to save lives.

‘It should be a global priority to substantially reduce or eliminate sugar-sweetened beverages from the diet,’ he said.

‘Some population dietary changes, such as increasing fruits and vegetables, can be challenging due to agriculture, costs, storage, and other complexities. This is not complicated.

‘There are no health benefits from sugar-sweetened beverages, and the potential impact of reducing consumption is saving tens of thousands of deaths each year.’

The study only looks at statistics and makes an estimation of the effect. This can’t prove what causes what. So you can’t know for sure if 184,000 adults are killed every year. It could be significantly fewer, perhaps even no people… or it could be much worse.

But no matter the exact number I think Dr. Mozaffarian is right. Doing nothing while tens or hundreds of thousands of people could be dying every year is not acceptable. This looks like the tobacco fight all over again.


  1. palo
    Observational study using dietary surveys for data = inconclusive.

    The same type of study used to impose the meat and fat phobia that has resulted in the biggest explosion of obesity and diabetes in the history of mankind.

    Replies: #2, #3, #4
  2. Dr. Andreas Eenfeldt, MD Team Diet Doctor
    Certainly observational data (statistics) is normally inconclusive and I said so in the post above.

    Difference between this and the fat phobia is that humans have always been eating natural fat – like saturated fat – during millions of years of evolution. We've never been drinking soft drinks loaded with sugar before, let alone multiple times a day every day.

  3. FrankG
    You are right, and as Dr Andreas already pointed out in the post this is only observational, BUT observations can suggest an hypothesis, which then requires clinical trials to take to the next stage.

    The hypothesis suggested by observational trials can be proven wrong in clinical trials... as has happened with those testing your "meat and fat phobias" or with HRT but that does not mean they are all wrong.

    Unlike the meat and fat phobia you mention (which incidentally were never even shown to be harmful in observational studies, let alone anything further), there ARE now solid clinical trials demonstrating the harm of so much added sugar.

    If you need further discussion, consider that the "evidence" for smoking causing cancer was largely derived from such observational studies; as human clinical trials would be highly unethical. Do you think we should dismiss the observations because they do not meet your standards and assume that smoking cigarettes, or added sugar is safe?

    Reply: #5
  4. murray
    Professor Tim Noakes presents an excellent lecture on causation and the use of epidemiological data, reviewing the Bradford Hill criteria for establishing probable causation. Notably, one gets more insight into what a study may show by viewing the study in conjunction with other known data.

    There are major differences between survey meat studies and survey soda studies. First, diet recall is poor, but it is much easier to say how many sodas I had last week than how many grams of meat. Second, meat correlation to cancer has been very weak. Survey data with weak associations mean little on their own. Third, there are well-understood metabolic and biochemical pathways that demonstrate how sugar can cause the types of damage observed in survey studies. The survey data in this case confirms that the observed metabolic effects make a practical difference over time. There are no such identified cancer-causing biochemical pathways for meat. The meat-cancer correlation is so weak that it is far more plausibly explained as correlation to other causal factors. For example, one might explain the data by observing that people have been told red meat is bad for health for a long time now, so a greater proportion of the health-indifferent have eaten red meat than vice versa. Fourth, when you eat meat, you are ingesting numerous things: methione-rich protein, perhaps collagen protein, different types of fats, various levels of byproducts of Maillard browning, etc., plus accompaniments, such as potatoes and an alcoholic beverage. When you drink a can of soda, it is far more plausible to isolate sugar as the dominant causal influence from drinking the beverage.

    So when one examines studies in broader context, two studies that share a formal characteristic (both are survey studies) may have quite different substantive implications in terms of enhancing the plausibility of hypothesized causalities.

  5. palo
    Frank G said "Do you think we should dismiss the observations because they do not meet your standards and assume that smoking cigarettes, or added sugar is safe?"

    Research standards - not my standards, as you falsely and maliciously accused me - dictate that an observational study hypothesis has to be confirmed with a randomized controlled trial to be valid.

    Furthermore, results from a well constructed observational study - certainly not one relying on surveys which by nature can be full of errors and irregularities - may be acceptable if the relative risk (as it was with tobacco) is 2,300% and the absolute risk is 5-10%.

    Sugar and tobacco are poisons. To suggest that I assume that smoking cigarettes, or added sugar is safe is a vicious lie!

    Do you think we should accept observations that we like even though they are not even close to meeting research standards, thus jeopardizing our credibility in these matters and hurting our noble cause?

    Reply: #6
  6. murray
    Good points, but I don't agree we have to have randomized control trials to confirm causation. To be sure, they are useful, but not necessary. A good example is the brilliant synthesis of clinical experience and epidemiological data of Dr. John Snow to discover the cause of cholera in London in the 1800s, which is well chronicled in The Ghost Map: The Story of London's Most Terrifying Epidemic--and How It Changed Science, Cities, and the Modern World, by Steven Johnson. There never was a randomized control trial with cholera-contaminated water, yet today no one advocates the miasma theory.
  7. palo
    Reply: #8
  8. Mike S.
    Not particularly impressed by that Mayo Clinic article. They assumed that, for reported diets, any rEI/BMR values (i.e., Energy Intake as calculated from reported food intake, divided by Basal Metabolic Rate as predicted from the Schofield equations) of less than 1.35 were "physiologically implausible values" and said that such values "may be considered 'incompatible with life'".

    If that's true, I should be dead, as a couple of years back I did several weeks of dietary recording - i.e., taking down everything I ate, checking food weights with a diet scale, recording calories and macronutrient values - to get a baseline before starting a new weight-loss program. I deliberately tried not to modify my diet in any way, and during the recording period I actually gained a small amount of weight (not much more than my average daily fluctuations, though, so probably not significant). But by my calculations my average rEI/BMR was around 1.25-1.30 - "physiologically implausible". Yeah, right. My physiology says otherwise.

    Not that I'm inclined to think much of the Circulation study. I tend to be highly skeptical of these kinds of mortality studies anyway, as they are often poor at accounting for and dealing with confounding factors. And this one's heavy reliance on BMI correlations I find troubling.

    We know enough about the problems of SSBs to advocate that people abstain from them; this study really doesn't bring anything new and significant to the table.

Leave a reply

Reply to comment #0 by

Older posts