Do Carbs Make You Crazy?
Evidence that blood glucose and dietary carbohydrate affect mood.
Posted Mar 01, 2012
I think pretty much everyone has heard of these ideas. Most recently I saw them in an article, "How I Overcame Bipolar II (and Saved My Own Life)" by Michael Ellsberg. Here is a quote from the article, when he saw an orthomolecular physician specializing in diet and health:
Dr. Hoffman told me there is mounting clinical evidence linking mood swings to blood sugar issues, and that in his experience bipolar patients respond well to cutting out refined sugar, and coffee and alcohol (which affect blood sugar) from their diets. "You should stop eating refined sugar altogether, and stop drinking alcohol and coffee," he told me.
Apparently, Mr. Ellsberg went to a psychiatrist also:
I asked the psychiatrist I was seeing at the time whether he thought there was any link between nutrition and mental health. He looked at me as though I had just asked whether there was any link between mental health and UFO rectal probes. "There is absolutely no evidence of any link whatsoever between dietary choices and mental health," he said curtly, and changed the subject.
Given a couple of years of intense research into the subject, I would (perhaps not so) respectfully disagree. Dietary choices can make a large impact on mental health. But what about the more specific question of whether we eat a lot of starches and sugars or not?
A well-known author and mental health clinician who believes in an ancestral health approach to general well-being is Nora Gedgaudas, who has just released an updated version of Primal Body, Primal Mind. In my reading of her book, I found she advocates a very low carbohydrate diet approach, that in effect we are either fat burners or sugar burners, and fat burners are steady and serene, while sugar burners tend to be moody and cranky. Nora herself clarified her position which can be paraphrased thusly: her clinical experience has shown her that limiting starches and sugars for most people has been very helpful (among many other maneuvers).
Then we have the opposite side of the argument, that starches and sugar are helpful, in fact a requirement for good mood. This argument is personified by rodent and women's health researcher Judith Wurtman at MIT who argues in her book The Serotonin Solution that carbohydrate improves and is in fact essential for good mood, particularly in the case of PMS, atypical depression, and seasonal affective disorder.
What did I learn in psychiatry residency? Well, that the ideas about sugar causing ADHD were unfounded, and that carbohydrates in general wouldn't be an issue plus or minus unless one had uncontrolled diabetes, in which case, of course, uncontrolled glucose swings could cause all sorts of craziness. Diabetes is one of those conditions we are supposed to rule out before spending 20 hours talking to someone about his or her mother, or prescribing Prozac.
So whom do we believe? Orthomolecular Dr. Hoffman? My psychiatry residency training program? Nora Gedgaudas? Parents who have observed their kids after a party with CAKE and ICE CREAM and FACE-PAINTING FAIRY PRINCESSES? (I'm going to tell you right now that the kids totally spaz out for several hours and then collapse in a twitching exhausted heap of sugar/princess withdrawal. The confounders in such an experiment are many, I must admit).
What have I learned in my years of combing the evidence? Don't believe anyone. Look it up your own self, and see if it makes sense in the context of physiology and evolution. Most hunter gatherer societies that I'm aware of ate starch when they could get their hands on it, so it doesn't make sense to me that carbohydrates alone would cause craziness. (We will also leave out alcohol and caffeine which Dr. Hoffman wisely recommends bipolar folks to avoid - both of which will affect sleep and unnaturally increase blood sugar falls).
So what does the literature say? Let me begin with a very pleasantly retro article from the halcyon days of 1966, when they believed in hypoglycemia outside of insulin tumors and diabetes: Relative Hypoglycemia as a Cause of Neuropsychiatric Illness.
This paper defines "relative hypoglycemia" as a drop in blood sugar that occurs after eating unwisely, say lots of caffeine or sugar. Apparently the people of the early 20th century suffered from this condition, which was misdiagnosed sometimes as a brain tumor, diabetes, or cerebrovascular problems, but when you stop eating like a jerk and consume some protein, fat, and get rid of the caffeine and alcohol, your symptoms go away.
In 1966 Dr. Salzer found that:
...patients with relative hypoglycemia have been diagnosed as having psychoneurotic anxiety, psychoneurotic depression, depressive reactions, schizophrenia, manic-depressive psychosis, psychopathic personality, chronic alcoholism, convulsive disorders, migraine, idiopathic cephalalgia, second cervical root syndrome, neurodermatitis, and even hypertensive cardiovascular disease... Major symptoms from a psychiatric standpoint are depression, insomnia, anxiety, irritability, lack of concentration, crying spells, phobias, forgetfulness, confusion... antisocial behavior, and suicidal tendencies... The major neurological symptoms are headache, dizziness, inward and external tremulousness, numbness, blurred vision, staggering, fainting or blackouts, and muscular twitching...There are also extensive somatic symptoms as follows: exhaustion, fatigue, sweating, anorexia, tachycardia, cold hands and feet, obesity, chronic indigestion, bloating, abdomnial spasm, muscle and joint pains, backache, muscle cramps, colitis, and convulsions.
Wow. That sounds really bad. It sure does sound like sugar (and caffeine) can make you pretty crazy. In the modern emergency room we have dextrose IV drips, ginger ale, and instant glucometer measures. I think perhaps that put the end to the mysterious hypoglycemia masking as a brain tumor era of medicine.
Somewhere between 1966 and the early 1990s, it became very unfashionable for medical professionals to endorse the idea of "hypoglycemia" outside of a super low measured blood glucose level of around 40 or less. "Relative (or reactive) hypoglycemia" had been consigned to a psychosomatic diagnosis (meaning it is "all in your head.") In fact, the year before I was born, the American Diabetes Association, the Endocrine Society, and the American Medical Association issued a joint statement on hypoglycemia, stating that publicity "has led the public to believe that there is a widespread and unrecognized occurrence of hypoglycemia in this country... These claims are not supported by the medical evidence." The American Dietetic Association stated: "Valid evidence is lacking to support the hypothesis that reactive hypoglycemia is common cause of violent behavior."
However, when blood glucose falls to quite low levels, the body does tend to react rather aggressively in order to prevent death. Stress hormones, including catecholamines and glucocorticoids, growth hormone, and glucagon are all released. This high-powered combo would no doubt cause sweating, palpitations, anxiety, weakness, crankiness, and other symptoms traditionally associated with hypoglycemia but can also occur whenever high levels of these hormones are released for any cause (1). If the blood glucose falls low enough that the brain is impaired, certainly very bizarre behavior (such as psychosis) can be observed.
How common is such low glucose outside very rare insulin-producing tumors or accidental overdoses of diabetic medications? Let's consider the gold standard oral glucose tolerance test. After an overnight fast, you are given 50 grams of dextrose (which is merely glucose). Then blood glucose is measured continuously for up to six hours. Often the lowest glucose level is missed if the glucose is not measured continuously, but rather at half hour intervals, as is also common. There is a wide range of glucose nadirs in the blood levels of those tested. The average level is about 65, but 10% of people fell below 47, with 2.5% of people below 39. That means that 1 in 40 people could expect to have "true" hypoglycemia in the context of a pure carbohydrate snack after a fast. This effect can be accentuated by caffeine and alcohol.
However, when one goes around measuring glucose levels of typical human beings eating mixed meals (not participating in oral glucose tolerance tests) levels almost never fell below 70. In these healthy subjects, levels pretty much stayed between 70-100 all day long, before and after meals, and overnight. It was a combination of these findings and studies of so-called hypoglycemics whose blood glucose levels did not correlate to symptoms that consigned "relative" or "reactive" hypoglycemia to conservative medical never-never land. However, many people (including myself, in full disclosure) have predictable "sugar crash" feelings about 90-120 minutes after a pure sugary snack or after drinking soda. Adding protein and fat to meals and snacks eliminates this problem which I would find odd if it were entirely psychosomatic.
There are other very interesting correlates between hypoglycemia and behavior and mood. A researcher named Bolton studied the Peruvian Quolla tribe he characterized as "perhaps the meanest and most unlikeable people on earth." (2) These people, particularly the men, seemed to act out in irrational acts of violence, and also seemed to have very strong sugar cravings. Bolton did glucose tolerance tests (GTTs) in many of the men, and he noticed a statistically significant correlation between the most aggressive subjects and those who had the lowest blood glucose values during the GTT (3).
Another researcher, Virkkunen, studied men who had committed serious violent assaults. Their GTTs were also remarkable for higher peak glucose values and subsequently lower glucose nadirs than controls (4). Folks with antisocial personalities have also been noted to have low blood glucose levels, and, once again, it is important to note that many of these men perpetrated the crimes under the influence of alcohol, which will accentuate the hypoglycemic effect. Similar studies in the general population show that both men and women whose blood glucose fall more rapidly during a glucose tolerance test will tend to have higher ratings of aggression (5)(6). In these studies, a glucose level of 63 or lower was strongly associated with the aggressive tendencies. This is still outside a normal healthy human's blood glucose range eating normal food, but much, much closer to the normal range than 40.
Wow. I guess rapidly fluctuating blood glucose does cause behavioral changes...wait a minute, there, buckaroos--as I have noticed in my perusal of the rather badly reasoned behavior/blood glucose/insulin resistance literature (which I will go over in a future post), these are mere correlations. We can't be sure of one causes the other or if there is another factor (or factors) that causes both low blood sugar and violent tendencies. Why is the blood glucose changing more rapidly in certain people than others to the same food or glucose stimuli? Could an underlying mechanism explain both aggression and hypoglycemia?
Virkkunen's more violent reactive hypoglycemics had enhanced insulin response to the oral GGT. And it has been found that impulsive offenders who act aggressively, particularly when intoxicated, have lower levels of serotonin turnover (measured low levels of the metabolite of the serotonin metabolite 5-HIAA in the cerebrospinal fluid (7). Low levels of brain serotonin are associated with enhanced insulin secretion and a tendency to develop low blood glucose levels. The data is a bit tenuous, but at least this mechanism makes some sense in the big picture.
So it seems that a diet that predisposes us to violent blood sugar swings (meaning high in sugar and low in protein and fat, made worse by caffeine and alcohol) can affect moods negatively and perhaps even cause violent behavior in some. In the next post I'll review the evidence that carbohydrates are essential to good mood.
Copyright Emily Deans, M.D.