Summary
I'll go through the papers I used one by one.
Chen et al.
Dr. Honglei Chen et al. conducted a study for the NIH (later presented at a meeting of the American Academy of Neurology). The study noted that people who regularly drink soda (i.e. four cups or more per day) are 30% more likely to develop depression later in life (and the effect was amplified by diet soda). The authors surveyed over 250,000 people.
There seemed to be a very strong correlation at work, and given the size of the survey pool, it didn't seem like a fluke. But the researchers were very quick to note - as were critics - that the study was not proof that drinking soda leads to depression. An association is not the same as proof of causality.
Walton et al.
Several psychologists monitored 40 patients over a short span. Some had a prior history of depression; others did not. Some in each group took aspartame, a common sweetener. The result showed that those with prior depression were more likely to have it exaggerated by consuming aspartame; those without it showed a slight increase in depression-like symptoms.
Ledochowski et al.
This study is for comparison, because fructose is not used in diet sodas but in regular sodas. I was curious as to whether or not the same symptoms would be found - and, surprisingly, they were.
Austrian researchers studied 53 patients over four weeks. These patients had gastroenterological issues which caused them to have fructose malabsorption. Over the study, women with the condition taking fructose developed several symptoms, including signs of depression. Other health issues were noted.
So what?
All three studies seemed to draw a strong conclusion between diet soda/soda (or ingredients in it) and depression, as well as other possible symptoms. There seems to be a strong causal relationship. However, the one downside is that these studies have not been replicated by other groups! The methodology has not been addressed either.
I'm wavering on whether to say yes or no on this. I hate to not draw a definite conclusion, but I have to agree with what many of the articles wrote: More research is needed. But if any of these results are confirmed by another study, then I'd say that there is indeed a good chance that diet soda can cause depression. So I'm leaning towards "yes".
No matter what the answer, it won't hurt you to stop drinking diet soda. Or regular soda for the matter. However, you should always consult a physician before making any (significant) changes to your diet.
Answer
Chen et al.
A study was done by the NIH, which analyzed cases of depression (or lack thereof) in 250,000 adults. georgechalhoub gave quite a nice summary of that study. The media jumped on the story (US News & World report, San Jose Mercury News, The Huffington Post, and, of course, Fox news). Newspapers love stories like this, so I'm not surprised the study made the news. What I was surprised about was that some of the articles said that causality was not truly addressed.
The Huffington post article was written by Dr. Lisa Young, a nutritionist and adjunct professor at NYU.1 Young wrote
So, what do we make of this study?
The study doesn't show that soda causes depression, but rather found an association between the two -- soda drinkers were more likely to be diagnosed with depression. It is important not to confuse the two.
Here is what I told the New York Daily News when called to comment on the study:
It's hard to tell from this single study what the exact connection between sweetened drinks and mood might be -- but either way, you're better off cutting back, says Lisa Young, Ph.D, R.D., an NYC-based nutritionist and adjunct professor in the department of Nutrition, Food Studies and Public Health at NYU.
"I'm not a fan of diet soda, but I don't think it's the soda alone that's causing depression," she told the Daily News. "The thing about soda is, you don't often drink it alone -- you drink it with junk food. You're not getting enough fiber and protein with that sugar, and your blood sugar not being stable could throw off your mood."
Diet soda drinkers who are focused on weight loss might also be blue from feeling like they're struggling with their weight, she said. "You're getting that false sense of sweetness, so you're never really satisfied," which could lead to continued cravings, she said. "[The soda] could be safe, but I think it puts you into this pattern of not really eating healthfully."
An article on Smithsonian.com also discussed the study2:
Does this mean you should stop drinking diet Coke and starting chugging coffee immediately? Probably not. This type of suggested link between two seemingly unrelated factors is an ideal time to bring up the difference between causation and correlation. Do the ingredients in both diet sodas and normally-sweetened drinks trigger changes in brain chemistry that lead to depression? Or are people with the tendency to become depressed simply more likely to drink these beverages in the first place?
Without the full paper, it’s hard to know for sure—we don’t know if the study’s authors controlled for all relevant factors, making sure to compare study participants who were alike in all ways except for their beverage consumption. As a result, a third, unrelated factor may cause people to both drink more soda and become depressed more frequently. Since the study is backward-looking, it’s especially hard to rule this out: The researchers can’t go back to 1996 and make sure to ask the participants every potentially relevant question to ensure that all potentially important factors have been taken into consideration.
Additionally, the fact that an association was found for both regular and diet sodas makes a causative link seem less likely. For that to be the case, either both sugar and aspartame must trigger depression, but at different frequencies, or a third ingredient in both sodas is responsible, but is somehow modulated by the presence of the natural or artificial sweetener.
Also interesting was the article in the San Jose Mercury News. Here's a short excerpt:
The researchers do not say that soda causes depression — they did not determine whether people who are likely to become depressed are also simply more likely to drink soda in the first place, and the study did not control for all relevant factors, such as socioeconomic status. What they did find was that people who drank more than four cans or cups of soda per day were 30 percent more likely to develop depression over the course of the 10-year study than people who drank no soda.
livescience also wrote about it, and touched upon the same things:
The study only found an association, and did not determine whether or not diet soda or fruit drinks caused depression. Although the researchers took into account factors that could affect the results, such as age, sex, education, smoking status, physical activity, body mass index (BMI) and energy intake, it's possible other circumstances, such as a family history of depression or stressful life events, could explain the association.
Now for a meta-analysis of all that.
I just quoted specific excerpts from the articles discussing the study. Aside from the article written by Dr. Young (and the one she was quoted in), all the articles showed the study in a good light. It got a fairly good reaction.
Smithsonian.com commented that it would be difficult to analyze the study, given that it was not yet available. I still can't seem to find it. Chen has done other work on depression and diet (see here; abstract here).
Hold on. Look at part of the livescience quote:
Although the researchers took into account factors that could affect the results, such as age, sex, education, smoking status, physical activity, body mass index (BMI) and energy intake . . .
Look at part of the abstract of the study I listed:
. . . adjusted for age, gender, educational level, marital status, smoking, and coffee drinking.
That sounds pretty similar, if not the same. But the paper's not quite the same. Dang. And it doesn't look like any of the papers that cite it are what we're looking for. Neither are any of the papers that this paper cites (of course, the paper we're looking for was published after this one). So it seems that the situation for us is the same as the situation for Smithsonian.com.
For fun, I'll quote a passage from the article discussing its own methodologys:
Major strengths of our study include the large sample size, long period of exposure assessment, and detailed information on potential confounders and effect modifiers. The analyses were however retrospective in nature. Selection bias cannot be ruled out if depressed individuals with PD were less likely to participate in the follow-up survey compared to depressed individuals without PD; if true the observed odds ratios are likely underestimated. Recall bias could also be a concern as PD patients might be more likely to recall or report depression than the PD-free participants. Although we could not directly evaluate potential impacts from this bias, our additional analysis excluding individuals with poor health status who might be more likely to over report depression, showed similar results. In such a large cohort, we had to rely on the self-report to identify depression and PD diagnoses as well as the years of first diagnoses. Misdiagnosis and under reporting are therefore unavoidable. Our ongoing verification study of PD diagnosis showed that 88% of PD cases were validated with information from their treating neurologists. Although we did not validate self-reported depression, nor did we have data on antidepressant use, in other validation studies self-reported diagnosis of depression was deemed fairly reliable compared to medical record review or specialist interview 16 and self-reports have been used in other longitudinal studies.
So interviewing patients with Parkinson's disease has some issues that can never be adjusted for. How ironic. Some of these problems could crop up in the study we're interested in if similar methods were used, but the issues directly related to PD would not be applicable.
So the study found that depression and drinking diet soda can go hand-in-hand, but causality was not established. That's my conclusion, as I play devil's advocate here.
Walton et. al.
I can't find any other studies that discuss the first study (Young's rebuttal doesn't quite count), so I looked at some others.
This page led me to this page. Which quoted a letter discussing a study. The lengths I go to to get sources for my answers.
At the core of all these links was an analysis of aspartame, approved by the FDA (according to the latter article):
The FDA approved aspartame for use in certain dry foods in 1981 and for soft drinks in 1983. In 1996, it removed all restrictions, allowing use in all food products, including ones exposed to heat, which separates the main ingredients. The FDA has set an acceptable daily limit of 50 mg per kg of body weight, which assumes that aspartame can safely replace all sucrose sweeteners in the diet.
Aspartame is a sweetener, used in, among other things, some diet sodas4. Studies have linked it to a whole bunch of health problems (those poor lab rats!). Dr. Ralph Walton did a study on aspartame in 1993:
Fine, but can those who suffer depression or bipolar disorder be considered part of the general population? In 1993, Dr Walton, who is a psychiatrist, conducted a study of 40 patients with unipolar depression and a similar number without a psychiatric history. The subjects were given 30 mgs per kg of body weight a day of aspartame or a placebo for 20 days (about equal to daily consumption if it completely replaced sugar).
Thirteen individuals completed the study, then an institutional review board called the project to a halt "because of the severity of reactions within the group of patients with a history of depression." In a smaller, shorter crossover design, "again there was a significant difference between aspartame and placebo in number and severity of symptoms for patients with a history of depression, whereas for individuals without such a history there was not."
Accordingly, the author concluded that "individuals with mood disorders are particularly sensitive to this artificial sweetener and its use in this population should be discouraged."
As to further particulars of the study, based on the eight depressed subjects and five healthy subjects who completed it:
Three quarters of the patients with a history of depression taking aspartame reported feeling depressed vs none of the healthy subjects taking aspartame and about 40 percent of both groups taking a placebo. The 40 percent is probably a statistical aberration owing to the small numbers who completed the study. Nevertheless, the figures consistently show the depressed/aspartame group experiencing an array of symptoms in far greater numbers and severity, including: fatigue, nausea, headache, trouble remembering, insomnia, and other symptoms.
Wow. That seems pretty definitive. Yet the article fully admits that peer review is necessary:
Remarkably, Dr Walton’s study is the only one we have related to both mood and aspartame. It would be helpful to get a second opinion, but no one else since, apparently, has tried to either replicate or refute his results. This may be due to the political and funding climate. "The NutraSweet company," Dr Walton told this writer, "clearly tried to block our study."
Ah, sounds like the cigarette companies. But something in me doubts that Walton's explanation is the full story, although it could very well be true. Conclusion: More peer review is needed, but it seems pretty conclusive. If it's right.
There are, of course, other articles discussing the study. This one is a letter written by Dr. Walton. In it, he writes:
That study was published in Biological Psychiatry in 1993 (a copy of the paper is enclosed). It demonstrated that individuals with mood disorders are particularly sensitive to aspartame and experienced an accentuation of depression and multiple physical symptoms. I had expected that the difficulties experienced by patients receiving aspartame would be fairly subtle (the dose of 30mg/kg/day was well below the 50mg/kg/day that the F.D.A. considered "safe"). I was not prepared for the severity of the reactions and for obvious ethical reasons cannot perform any further human studies with aspartame.
Two years after the publication of that study I was contacted by a producer for "60 Minutes" and asked if I would be willing to be interviewed by Mike Wallace for a segment on aspartame. During that interview Mike challenged me on my assertion that there were major problems with this sweetener in view of the fact that the bulk of the medical literature attested to it's safety.
I responded that one had to look carefully at study funding - that virtually all of the studies claiming safety were funded by the industry, whereas independently funded studies invariably identified one or more problems. When he challenged me to prove this I prepared a chart correlating study outcome and funding source. This chart was aired on the 60 Minutes segment, and is enclosed, with further discussion of this entire issue.
That might be an interesting interview to watch. Perhaps I can find a transcript.
Article #2 is also a letter. And a very opinionated one, too. I don't have room (or time, I think) to summarize it, but it gives a lot of different studies regarding the effects of aspartame.
Article #3 is written by the same person (Dr. Betty Martini), this time to Florida State University. It rehashes the same points . . . but it links to Walton's study (co-authored by Robert Hudak and Ruth J. Green-Waite, who seem to have missed out on the excitement)!
Let's look over it, shall we?
The project design called for the recruitment of 40 patients with a history of treatment for recurrent major depression, currently doing wall with a Brief Psychiatric Rating Scale (RPRS) rating of no greater than 6 at the time of the study. Subjects were recruited by word of mouth, and by a posting and distribution of the study protocol and informed-consent statements among patients, attending physicians, administrative and nursing staff, and medical students at the Western Reserve Care System.
Not the best recruiting system. Note also that the study only showed that patients with depression could have that depression worsened by aspartame. That's not quite the same as directly linking aspartame to depression in otherwise happy individuals.
Also, the NutraSweet company didn't do too much to block the study:
NutraSweet Company denied the request from the authors to purchase aspartame.
That's it.
Each participant monitored his own symptoms using a checklist, one of which was provided for each week of the study. Listed symptoms were: headache, nervousness, dizziness, trouble remembering, binge eating, lower back pain, nausea or upset stomach, feeling blue or depressed, insomnia, uncontrollable temper outburst, and other (to be specified by the subject). For each symptom the participant had to assign the following point value on a daily basis: 0 = not present; 1 = mild (symptom occurs but does not disrupt activities); 2 = moderate (symptom occurs but can be controlled whether by medication or other means; 3 = severe (symptom occurs and disrupts daily activities).
Each participant monitored his own symptoms . . . I hope that "he" is not being used literally (but in place of "he or she"). If it's only studying male patients, then this is irrelevant to your question, though still interesting.
Here's a table:
Patients
with history Nondepressed
of depression volunteers
Placebo Aspartame Placebo Aspartame
Headache 63% (5) 88% (7) 80% (4) 20% (1)
Nervousness 25% (2) 63% (5) 0% 0°k
Dizziness 13% (1) 25% (2) 40% (2) 0%
Trouble remembering 0% 63% (5) 0% 20% (1)
Binge eating 13% (1) 13% (1) 0% 0%
Lower back pain 25% (2) 25% (2) 20% (1) 0%
Nausea 25% (2) 100% (8) 40% (2) 20% (1)
Depression 38% (3) 75% (6) 40% (2) 0%
Insomnia 38% (3) 50% (4) 20% (1) 20% (1)
Temper 0% 25% (2) 20% (1) 0%
Other
More energy 0% 25% (2) 20% (1) 20% (I)
Fatigue 0% 25% (2) 0% 20% (1)
Malaise 0% 38% (3) 0% 20% (1)
Weight loss 13% (1) 0% 0% 0%
Pain in eye 13% (1) 0% 0% 0%
Negative thoughts 0% 13% (1) 0% 0%
Bad taste in mouth 0% 13% (1) 0% 0%
Swollen lips 0% 13% (1) 0% 0%
Facial numbness 0% 13% (1) 0% 0%
Conjunctival hemorrhage 0% 13% (1) 0% 0%
Weight gain 0% 13% (1) 0% 0%
Irritability 0% 25% (2) 0% 0%
Less sleep 0% 0% 20% (1) 0%
Diarrhea 0% 0% 20% (1) 20% (1)
Nightmares 0% 0% 0% 40% (2)
More sleep 0% 0% 0% 20% (1)
The numbers in parentheses (e.g. "(1)") indicate the number of people who reported that symptom.
The "nondepressed volunteers" column is what we're interested in, and it's not too convincing.
Here's the study's meta-analysis:
The lack of validation of the extensive anecdotal reports of adverse reactions to aspartame by double-blind studies may reflect the fact that, to date, such studies have not been performed on what we feel is an especially vulnerable population-individuals with mood disorders. In this study even people who believed that they had problems with aspartame (three of the nondepressed volunteers) did not demonstrate significant differences from placebo, whereas patients with a history of depression and no awareness of aspartame intolerance did demonstrate significant adverse reactions. Although there are some long-term studies, (Leone et al 1989) 1- or 2-day challenges, such as those of Schiffman et al (1987), may not be long enough for difficulties to emerge. In this study, patients most often began to report significant symptoms after day 2 or 3. A 1- or 2-day challenge also does not replicate common patterns of daily consumption.
Ledochowski et al.
I found a third study on the website of Psychology Today. It was conducted by researchers in Europe, but it studied the effects of regular soda, not diet soda.
Turns out that in Central Europe, a large percentage (30-50%) of the population suffers from certain types of carbohydrate malabsorption . I asked a friend of mine, an academic gastroenterologist who just left Johns Hopkins, what she thought the percentage of Americans was, and she felt it was 15-20%. In fructose malabsorption(link is external), the GLUT5 transporter in the small intestine doesn't take up fructose as efficiently as it could. That means lots of undigested fructose floats down to the colon, feeding the bacteria there and leading to bloating, cramping, and diarrhea - basically the symptoms of irritable bowel syndrome. It is diagnosed via testing for excess hydrogen in the exhaled breath after a fructose load of 50mg. A similar test can show if someone has lactose intolerance.
Well, the researchers in Central Europe took a hundred or so "otherwise healthy volunteers" who had complained about gastrointestinal distress at a doctor's visit. None were on medication (except oral contraception) or had any signs of chronic or serious illness. They were given a standard scale test for depression (the Beck Depression Inventory) a fructose malabsorption test, then, a week later, a lactose malabsorption test.
Let's cut to the chase - a positive test for fructose malabsorption corresponded to depressed women, but not men. Lactose malabsorption alone didn't matter in either sex, but the 12% of women in the study who were both fructose and lactose malabsorption positive were by far the most depressed. The normal female controls had an average depression score of 7.5 - the combined malabsorption women had an average score of 14.6. That's a huge difference, and the data between the unaffected individuals and the carbohydrate malabsorbers hardly overlap at all.
Aha! Finally, something that says that women are affected by sugars (fructose, in this case). And there's an explanation:
Turns out that fructose (and lactose) can react chemically with tryptophan, the amino acid precursor for our important happy chemical, serotonin. The sugars can degrade tryptophan so that there isn't as much available to be absorbed into the body. And, indeed, fructose malabsorbers have lower levels of tryptophan in the serum than normal controls. And, hey, turns out they have lower serum zinc and folic acid too - both of these findings are associated with depression.
But why would the symptoms of depression be confined to women? The researchers postulated that estrogen made the big difference. Estrogen activates an enzyme called hepatic tryptophan 2,3 dioxygenase that shifts the metabolism of tryptophan from making serotonin (happy) to making kynurenic (not happy). Women already have lower serum levels of tryptophan than men do (which may be part of the reason why we are more vulnerable to depression in the first place), so screwing up whatever available tryptophan in the diet with fructose may lead to even lower levels, and thus depression.
Once again, the lack of peer review is troubling:
Well. All of that is observational data, and correlation does not equal causation. All we have to hang our hats on is a small study from Austria in 2000. In this study, folks with known fructose malabsorption were put on a low-fructose diet, which reduced their depression scores by 65% after 4 weeks. The findings were more pronounced in women than in men.
I looked up the study, whose abstract can be found here. Something interesting is noted in the abstract:
Fifty-three adults (12 males, 41 females)
Now, the number of women in the study shouldn't affect the percentage, but still. That's quite the disparate comparison.
I was able to get the study on a pdf.
Patients
From 100 patients who presented at our office for a medical health check, 53 individuals with gastrointestinal complaints were consecutively chosen to participate in this follow-up study on the basis of the H2-breath test results (see below). Sigmoideoscopy was previously performed in most of these patients because of chronic gastrointestinal complaints, and negative results were reasons for a further checkup in our office. For those who had not had sigmoideoscopy, this examination was prescribed to rule out inflammatory bowel diseases. The otherwise healthy outpatients, aged from 17 to 75 years (mean, 44.8 14.5), gave informed consent to participate in the clinical trial with fructose-reduced diet which was approved by the local ethics committee. There were 12 male (range, 23–75 years; mean 42.3 14.6) and 41 female patients (range, 17–73 years; mean, 45.6 14.6).
No information is given as to how these 100 made it to the office, so the applicant pool cannot be fully analyzed.
Diagnosis of fructose malabsorption was established by H2-breath test after oral fructose load with 50 g fructose given in 250 ml of tap water. All H2-breath tests were performed between 0800h and 0830h after a 12-h overnight fast. Breath-H2 was measured using a Bedfont gastrolizer (Bedfont Ltd., Kent, UK), which had been validated by several authors (11–13). Breath-H2 was monitored before fructose load and in 30 min intervals for at least 2 h after fructose load. Maximum breath-H2 concentrations were registered and the differences from baseline values were calculated yielding the operating parameter DH2.
Here's the paper's meta-analysis:
In this study, no control group could be included, because most of the studied subjects had gastrointestinal complaints, and we could not stay away from dietary interventions for ethical reasons. Thus, it cannot fully be excluded that a placebo effect could have contributed to the observed amelioration of symptoms, e.g. the improvement of BDI after dietary change. However, the first set of data was obtained at the end of the pre-intervention period, and the patients were asked to estimate the parameters of gastrointestinal symptoms using an arbitrary analogous scale retrospectively. During this period it is unlikely that the individuals changed their dietary behavior, therefore one can consider the study presented here as a preliminary follow-up investigation which outlines the consequences of a fructose-reduced diet. Furthermore, a dietary observation period of 4 weeks may compensate a potential bias owing to instantaneous amelioration of symptoms. Because all scores were measured exactly 4 weeks apart, it can be ruled out that in females the menstrual cycle would have had a major influence on the outcome of depression scores or meteorism scores.
1 I say this to emphasize that she's not just a layman. This isn't a proper scientific rebuttal, but it isn't by Joe Schmo, either.
2 The study found that there were benefits to drinking coffee. Fox News wrote, "Coffee and tea contain antioxidants and phytochemicals, which promote health and well-being." livescience also mentioned another study that agrees with this result.
3 "PD" is short for "Parkinson's disease," the target of the study.
4 Yes, in both Pepsi and Coke.