Correlation Not Causation: The Health Risks of Diet Sodas

Yesterday some Big News hit the internet: Diet Sodas Linked With Health Risks (this link courtesy of Reuters Health). Now according to the press release, diet sodas are linked with increased risk for heart disease and diabetes; specifically that adults who drank one or more sodas a day had a 50% increased risk of 'metabolic syndrome.' This syndrome is characterized by excess fat around the waist, high 'bad' cholesterol and low 'good' cholesterol, and high blood pressure among other things. Furthermore, this syndrome places a person at risk for heart disease and diabetes.

The study, published in the most recent issue of the journal Circulation, included over 6000 middle-aged women and men who were observed over the span of four years. Findings indicated that those who drank one or more normal (not diet) soft drink a day had:
- A 31% greater risk of becoming obese
- A 25% greater risk of developing high blood sugar and triglycerides
- A 32% greater risk of having low good and high bad cholesterol

Those who drank one or more diet soda per days had
- A 50-60% greater risk of developing metabolic syndrome

So why would diet soda consumption, even though diet sodas are sugar-free and very low-calorie, contribute to a constellation of symptoms usually associated with obesity? The study illustrates the crucial difference between correlation and causation when considering the significance of data.

"On average, soda drinkers tend to eat more calories, consume more saturated fat and trans fat, eat less fiber, exercise less and be more sedentary," Vasan [the lead author] said in a telephone interview.

That is, drinking sodas, and diet sodas, may be correlated with other life choices that are less than healthy. The drinking of diet soda, while unlikely to largely contribute to (cause) the development of metabolic syndrome, is a behavioral trait that is associated with others that are negative. So if you DO exercise and otherwise eat a balanced diet, you shouldn't feel awful about drinking a diet soda every now and then: its the behavioral patterns which are reflected in constant diet soda drinking rather that the soda itself which may be harmful.

More like this

drinking a diet soda every now and then:

For me, every now and then means that I have a soda about one out of every three 30-minute periods during a given day....

I'm also an underexcercised fat guy. That's the real reason I'm an unhealthy mess who will probably die early. But, now, perhaps I can participate in a class action suit against the American Coalition For Pushing Diet Soda Down Our Throats. Yay!

-Rob

Those are some interesting findings. It's probably true that lifestyle has a much bigger impact than the diet sodas themselves. In a way this finding is disappointing to me because it obfuscates any findings as to whether the diet sodas themselves are harmful. People have decried for years that the sugar substitutes are harmful and should be avoided. I've seen some compelling theories as to how this may occur such as the organic hydrolysis of the aspartame molecule into methanol in the stomach. However, I've never seen any good studies to suggest that acute conditions can occur from consumption of diet sweeteners.

I suspect that a lot of the diet soda consumed in the U.S. is drunk along with fast food or with other high-fat, high-sugar, high-calorie meals. It's heavily marketed in a way that implies it's a weight-loss product, when at best diet soda is merely a "no-weight-gain-from-this-substance-alone" product. It's also very easy to fall into the trap of telling yourself "I'm drinking diet soda, so I don't need to be careful about the other things that I eat and drink."

Yeah, this one's kind of a no-brainer for a number of reasons. The most obvious to me is this: Who drinks diet soda? People who are concerned about their weight. Who's concerned about their weight? Overweight people. Who has metabolic syndrome? Overweight people.

so I drink one regular soda and one diet soda every day. Does this mean I have a 80% chance of developing metabolic syndrome?

"eh": Statistics doesn't work that way (maybe you're joking?) :) For one thing, its the average across all people who participate in a certain behavior not the individual risk. Also risk doesn't compound that way in statistics. Only you and your doctor can determine your actual risk.

Who drinks diet soda? People who are concerned about their weight. Who's concerned about their weight? Overweight people. Who has metabolic syndrome? Overweight people.

Yup. Someone I knew used to jokingly say that 'diet coke is fattening cause everyone I see drinking diet coke is large.' You assessment is correct, although the press release didn't spin it as quite the no-brainer (sadly). The headline alone is misleading enough.

"Only you and your doctor can determine your actual risk"

After seeing how biostat is taught in med schools, even that seems a little too generous.

By Biomed Tim (not verified) on 24 Jul 2007 #permalink

Yes, this sounds like yet another study conducted by people who have no clue of even the most basic facts of statistics. And it probably is.

But just because you confuse correlation and causation doesn't mean that you couldn't hit upon something interesting every once in a while. The point that I am trying to raise is that the sweeteners used in diet sodas are also used in animal food. I'm not talking about cats and dogs here, but about pigs. Yes, pigs get Aspertam so they will grow big and fat.

The no-brainer here is that as soon as we taste something sweet, our digestive system kicks in and makes itself ready for all the sweet sugar. But since its been fooled in the case of diet soda, the result will be hunger and more often than not a craving for sweets. And that's just what they found (although it's on page 2 of the cited article):

"On average, soda drinkers tend to eat more calories, consume more saturated fat and trans fat, eat less fiber, exercise less and be more sedentary," Vasan said in a telephone interview.

Sadly, the researchers were not able to explain this: "Maybe drinking something sweet conditions you in such a way that you develop a preference for sweet things,"

After seeing how biostat is taught in med schools, even that seems a little too generous.

Yikes!

But just because you confuse correlation and causation doesn't mean that you couldn't hit upon something interesting every once in a while. The point that I am trying to raise is that the sweeteners used in diet sodas are also used in animal food. I'm not talking about cats and dogs here, but about pigs. Yes, pigs get Aspertam so they will grow big and fat.

Thats fascinating. Do you remember who did that research? I'd love to blog it!

Just covered this on my meal blog. In my view, this is the age old problem of dieting. We have to realize that it's not a zero sum game. The void of sugary sodas can't be filled by reeses cups, cheesesteak pizzas.

What about the aspartame etc. that many diet sodas have? It's been linked to cancer, no?

This isn't even a issue of causation. This is an issue with the direction of the arrow of correlation. '

If you shifted the study to check the % of people with "metabolic syndrome" (which sounds contrived to cover general obesity) who drink diet soda, you'll find a large correlation which by the same grade of logic indicates that the syndrome causes people to drink diet soda.

The reason large people are drinking diet soda is that they don't want the empty calories from the HFCS. Sure, they get lots of empty calories elsewhere, but that's one spot they can steer around.

By wildcardjack (not verified) on 28 Jul 2007 #permalink

jvarisco: It seems likely that most of the anti-Splenda, anti-Nutrasweet health information out there is propaganda excreted by the sugar industry. They've already polluted the environment (especially the Everglades) for many years thanksto phosphate-heavy runoff and stuff, so they might as well start in on polluting minds as well.

Shelley -- If the BFBers get wind of this, how do you think they'll spin it? Will it be "See? People who drink diet soda get fat, so what people eat and how much they weigh are unrelated?" or "There go the sloppy scientists again, scrambling for anything they can find to demonize so-called 'junk food' by saying it leads to (gasp! The horrors!) fatness!" I say a mixture of both.

In any case, someone will point out that she has seen thin people drink not only diet soda but regular soda too, and that her scrawny aunt has diabetes and high cholesterol, and that'll be the end of it -- being fat is NOT bad for you, ever, nope, and quit trying to pull my fingers outta my ears!