Tuesday, July 24, 2007

Will Diet Coke make you fat?

Splashed over headlines and sound bites today will be articles and reports telling you that a new study suggests that if you drink any soft drinks, even diet ones, that your risk of developing high blood pressure, high cholesterol, diabetes and obesity rises dramatically.

Scary stuff, maybe you should immediately pitch your diet coke stash.

Don't throw out it out yet.

So this study, published yesterday in Circulation, looked at over 6,000 folks and their soft drink consumption with the authors reporting that they controlled for age, sex, physical activity index, smoking, saturated fat consumption, trans fat consumption, fiber, magnesium, total calories and glycemic index in asking whether or not consuming soft drinks (diet or otherwise) raised the risk of developing metabolic syndrome (the constellation of symptoms listed above).

The authors concluded,

"In middle-aged adults, soft drink consumption is associated with a higher prevalence and incidence of multiple metabolic risk factors"
and they found this to be true regardless of whether the soft drinks were of the regular or diet variety.

Problem is, there are too many inconsistencies and omissions in this study for it to have any value. Let's go through some of them:

  • They don't tell us how they tracked calories. They mention controlling for calories in their abstract but then fail to mention in their methodology how exactly they determined how many calories folks in this study were consuming. That's quite an important methodologic piece as if it turns out that folks who drink soft drinks (diet or otherwise) tend to consume more calories, it'll matter how they tracked calories since it's calories that lead to weight gain and weight that leads to metabolic syndrome.

  • The calories they did track are inaccurate. They reported that folks who drank no soft drinks consumed on average 1,800 calories daily and those drinking 2 or more daily 2000 or so Calories. In what alternate universe is the average per capita consumption of calories in America 1,800-2,000 calories daily? According to the USDA, the average American consumes around 2,700 calories daily and even back in 1970 was still higher than the calories reported in this study. The Food and Agriculture Organization of the United Nations actually thinks the average American consumes around 3,770 calories daily. Clearly if this study's calorie models are flawed, with calorie models based off of reported dietary intakes than dietary intakes are flawed in which case the conclusions aren't valid.

  • The authors did not control for type of carbohydrates. They kind of did, indirectly, by controlling for glycemic index, but frankly a more valuable control would be for whole grain consumption as its consumption has been shown quite conclusively to minimize the risk of developing metabolic syndrome while the converse is true from refined carbohydrates (the white stuff)

  • The authors did not control for sodium intake which of course would impact on the development of hypertension (one of the criteria above). Here actually is an argument that might be applicable as many diet beverages do still have salt in them

  • The authors did not control for meals out. If would certainly seem plausible to me that the folks drinking more soft drinks (diet or otherwise) are the folks eating out more regularly. Meals out have more calories and as a whole more dietary salt, less healthy fats and more refined carbohydrates - all of which contribute to metabolic syndrome development.

  • The authors could not come up with any commonality to explain the fact that diet soft drink consumption and regular soft drink consumption led to the same outcomes. Unfortunately, without commonality what they've failed to prove is causality and instead have proved association. They did mention "sweetness" of the beverages as a possibility leading people to consume more, but if you remember back to my criticisms on the study's calorie models, they concluded that in fact calories were not statistically different between groups. As well, if they did want to blame sweetness, another control would have to be juice consumption, because if both high fructose corn syrup and aspartame can have the same effect, so too should plain old fructose.

    So to sum up, the study's calorie models are flawed at best, the authors didn't control for known confounding variables that have in fact been proven to be contributors to metabolic syndrome and they admit that they can't think of any similarities between diet soft drinks and regular soft drinks that would explain their results.

    The ABA (the American Beverage Association) came out with a statement of their own and amazingly, I agree with their conclusions more than this study's.

    The ABA states,
    "This study doesn't prove any link between soft drinks and increased risk of heart disease. Its assertions defy the existing body of scientific evidence as well as common sense. Even the researchers acknowledge that their study can't support a link."
    I couldn't (and didn't) say it better myself.

    You may now enjoy your Diet Cokes.