Bottom line up front: It’s 2019, and we just got word that no mater what kind of soft drink you’re sipping (with or without a straw), you may be affirmatively choosing to shorten your life.
Normally I’d caution against taking action based on headlines like that, but the study that announced this result was based on 400,000 live human beings doing what humans do—in complicated, complex ways. Not in a petri dish; not in a tightly-controlled, double-blind study.
How We Learn About How The Body Works: A Crash Course
You may be surprised to learn how much energy and time this ongoing project is; when you compare what we are learning today to what we “knew” just a few decades ago, you begin to understand we aren’t just discovering fresh facts.
We’re also discovering the need to discard what we thought we understood—in some cases, for decades.
Take, for example, margarine. If you’re old enough, you remember when doctors advised us that butter was deadly. It would clog our arteries with cholesterol. We needed to switch to a form of bread spread and cooking fat that would prevent heart disease!
So we did.
As Harvard Health puts it today, “The truth is, there never was any good evidence that using margarine instead of butter cut the chances of having a heart attack or developing heart disease. Making the switch was a well-intentioned guess (emphasis added), given that margarine had less saturated fat than butter, but it overlooked the dangers of trans fats.”
Medical Research and Advice: The Art of the Well-Intentioned Guess
And now, the refinement in our thinking is comings for “diet soda” (sugar-free or artificially-sweetened soft drinks). Since the link between glucose and metabolic disorders (i.e., diabetes and prediabetes) is clear, it seemed both intuitive and obvious to advise people to steer clear of sugary beverages for their health, and so for years, that’s what the health profession has done. “Stop drinking all of that sugary soda,” we told patients, in good faith, based on what we knew. “It is killing people.”
And they switched, faitfully, to artificially sweetened beverages.
But then evidence started to question that approach. As early as 2013, a large French epidemiological observational study (more about this in a minute) compared the diabetes risk of three groups of women over the course of 14 years, based on their own self-reports: Those who consumed soft drinks versus those who drank no soft drinks, but did consume 100% fruit juice (which also contains a great deal of sugar—but naturally-occurring fruit sugar).
- Type 2 diabetes risk was significantly elevated (at the 95% confidence interval) among women in the highest 25 percentile of soft drink consumers (whether these were sugar- or artificially-sweetened).
- The fruit juice consumers saw no correlated risk in diabetes incidence.
Something was going on here. Something strange. And that’s when researchers decided to start feeding both sugar and artificial sweeteners to rats—which is backwards, in terms of how science is usually done, but valid in this case.
Back to the Drawing Board with Artificial Sweeteners: It’s Messy
For those who aren’t in the medical or scientific research fields, I’d like to quickly explain the difference among types of studies we talk about when we describe “a new study,” because not all studies are created equal.
Are you ready to ride along with me through the rough terrain of How Medicine Learns Things? Great!
The next time somebody says “A study just revealed,” go to the study and find out what kind of study it was. That will tell you a great deal about how much you should take away from the study.
In ascending order, from small to large, (from least-applicable to the whole human species to potentially most-applicable), here we go.
Test Tube Lab Study: Somebody tried something in a highly controlled lab setting as a proof-of-concept. Most drug development trials begin this way. It’s the earliest form of research, and a “success” means many more studies lie ahead.
Animal Research Studies: Based on work in a test tube, lab, or petri dish and a hypothesis, animals (generally animals with genetic similarities to humans) are introduced to a variable (a drug, surgery, a behavioral treatment) and their response is studied. (Animal studies may or may not include a “control group” who does not receive the treatment, so researchers can be sure the effects of the thing they’re studying were actually caused by that thing—not something else unrelated. Control groups are also used in other study formats).
Case Reports: Based on single or small groups of patients, these are the smallest types of human “studies” and simply report an interesting constellation of symptoms, a syndrome, a novel disease or treatment to an existing condition that produced an outcome not previously known in the literature.
Cross-Sectional Study: These studies select a specific group of people at one point in time or at a specific interval and study a condition, set of conditions, or outcomes. For example: The Short-Term Effects of Doritos, Donuts, and Mountain Dew on 25 Dental School Students Prepping For Final Exams: A Cross-Sectional Analysis. (While this study doesn’t exist, to the best of my knowledge, it could.)
Cohort (Prospective Observational) Study: These long-term studies group many people with a similar health condition or treatment status, then compare them over a long period of time with others who do not belong to the same group. Studies of smokers, cancer patients, etc. are by definition cohort studies, as are long-term studies of groups of people with diagnoses such as depression, PCOS—the possibilities are broad, but the results may not be generalizable to the entire human population. (Epidemiological cohort studies, like the French study that released this news about soda and the Nurses’ Health Study, seek to broaden their large demographic samples to represent the larger populations.)
Randomized Controlled Study: The “gold standard” for trials of novel treatments, interventions, protocols, and substances, in these studies subjects are assigned to more than one group. They receive either the treatment bing studied, a placebo, “sham” treatment, or perhaps differing doses or treatment at differing intervals; then results are compared.
Systematic Review and Meta-Analysis: These studies are retroactive studies of other studies. And while that may sound like a waste of time, it isn’t. (Going back and looking again is how we discovered our guesses about butter and margarine and cholesterol and trans-fats was wrong.) These studies now use the power of big data and unimaginable processing, as well as our own critical thinking, to go back through mountains of literature about diseases, conditions, drugs, DNA—even more. Then researchers look for new patterns and the opportunity to reinterpret what we know.
The End of the Study is Just the Beginning of the Mystery
All of this was a long way of getting here: A study of more than 400,000 European adults over sixteen years showed that the risk of premature death is higher among those who consumed more than 2 glasses per day of soft drinks—whether those were sweetened with sugar or artificial sweeteners.
But that correlation may not be driven by soda itself.
As the study’s author pointed out, “in these types of studies (observational epidemiology) there are other factors which may be behind the association we observed…For instance, high soft drink consumption may be a marker of overall unhealthy diet.”
Get into the habit of reading beyond the headline, as a consumer of health information (and as a consumer of food, beverages, and supplements). Understand what kinds of “studies” underlie the big “story” of the day. Look for the number of patients; look at how long they were studied; ask, “Could anything else account for these effects?”
And congratulations if you just set aside your soda for a tall, refreshing glass of water—maybe with a spritz of healthy (we think) fruit juice.