Wheat and Schizophrenia

A number of people, knowing that my partner Sally has schizophrenia, have suggested that we look into the relationship between schizophrenia and wheat gluten. I decided to do a quick survey of the literature. Here's a recap of what I've found so far.

Wheat: Bringer of disease?

First of all, know that I'm a skeptic. Whenever somebody tells me wheat gluten is evil for this or that reason, I nod knowingly but politely, the way I nod knowingly but politely at Jehovah's Witnesses who come to my door; then I excuse myself with one of the rejoinders listed here.

But I'm also open-minded enough, and just plain curious enough, to look into such things before I dismiss them totally. It turns out a growing body of evidence has implicated the gliadin family of proteins (present in grains of the Triticum family) in a variety of atopic and inflammatory disease processes. The gliadins (of which there are three main types) stimulate production of the human protein zonulin, which mediates permeability of the intestines, through mechanisms that are just now being elucidated. "Leaky gut," in turn, is known to be a factor in autoimmune diseases. (As one researcher puts it: "When the zonulin pathway is deregulated in genetically susceptible individuals, autoimmune disorders can occur.") See this 2009 paper in PNAS as well as this 2011 paper for examples of recent work in this fast-moving area. For more citations, go here.

Zonulin research has direct application to such diverse ailments as diabetes, celiac disease, multiple sclerosis, and rheumatoid arthritis, and there is reason to believe autism and schizophrenia may in some way, in some individuals, be influenced by defects in zonulin metabolism. For example, we know that schizophrenia sufferers are at increased risk for celiac disease, and indeed in a 2011 study involving 1401 schizophrenia patients (and 900 controls) it was found: 

Among schizophrenia patients, 23.1% had moderate to high levels of IgA-AGA compared with 3.1% of the comparison group (χ2 = 1885, df = 2, P < .001.)

English translation: IgA, or immunoglobulin A, plays a critical role in mucosal immunity and is secreted in large amounts in the intestinal tract (in normal individuals). AGA refers to gliadin proteins. Elevated production of IgA antibodies in response to gliadin is characteristic of celiac disease (in fact, the gliadin antibody test is a mainline diagnostic tool in detecting the condition).

So where do we stand with regard to wheat and schizophrenia? The study cited more than any other on this subject is Dohan and Grasberger, "Relapsed schizophrenics: earlier discharge from the hospital after cereal-free, milk-free diet," Am J Psychiatry, 1973 Jun;130(6):685-8. You won't find the full PDF online, but I obtained a copy of it, so I'll summarize. The essence of it (as stated in the Abstract) is:

Routinely treated schizophrenics, who on admission were randomly assigned to a diet free of cereal grains and milk while on the locked ward, were discharged from the hospital about twice as rapidly as control patients assigned to a high-cereal diet. Wheat gluten secretly added to the cereal-free diet abolished this effect. These and previous findings suggest that cereal grains may be pathogenic for those hereditarily predisposed to schizophrenia just as they are for celiac disease, a disorder that may be genetically related.

The 115 study subjects were male veterans at the Coatesville, PA, Veterans Administration Hospital (average age 38) who met the description of schizophrenia in the original 1952 DSM-I. (Exactly why the VA hospital was using an out-of-date version of the DSM, I'm not sure. DSM-II was available at the time of the study.) Patients were randomly assigned either to a diet free of all cereal grains and milk products, or to the usual milk-containing hospital diet, which (for the experiment) "was made somewhat higher in cereal grains (e.g., spaghetti instead of potatoes)." Milk was omitted from from the cereal-free diet "because it produces a relapse in some celiac patients and thus, we hypothesized, may be harmful to some schizophnenics."

The study tracked the number of patients discharged from the hospital each 30 days for a period of 360 days. Discharge readiness of patients was apparently determined "after review of their cases at a staff conference." Unfortunately, the vagueness of this part of the protocol does raise serious questions about blinding and possible researcher bias. There's no indication of whether the authors of the paper themselves had a hand in determining discharge readiness. Another serious weakness of the study is that we don't know exactly which patients got how much of what medicines. All the authors say is: "Chlorpromazine and thionidazine were the most frequently administered antischizophrenic medications." We also learn, in the Discussion portion of the paper, that patients received the cereal-free/milk-free (CFMF) diet only during the portion of their stay when they were confined to a locked ward, which was "usually" 15 days. (According to the authors, 76% of the CFMF group got the restricted diet for 15 days, total, before going back on an unrestricted diet.)

I took the paper's tabular results and made a graph (below) from the numbers in Table 2. The graph shows days until release at the bottom and percent of patients released on the y-axis. You can see that the difference in release rates was highest in the first 90 days, then the curves for the two diet groups start to come together.

Data from: Dohan and Grasberger, "Relapsed schizophrenics:
earlier discharge from the hospital after cereal-free,
milk-free diet," Am J Psychiatry, 1973 Jun;130(6):685-8

I've looked at a number of other studies on gluten restriction and schizophrenia. They all suck in one way or another, whether in study design or (more often) the number of study subjects. The most interesting study I found was a controlled double-blind study of 24 patients involving the use of a gluten-free diet for 14 weeks. It found "there were beneficial changes in the whole group of [schizophrenic] patients between pre-trial and gluten-free period in five dimensions of the PIP (Psychotic In-Patient profile), maintained during the gluten challenge period," but the authors also admitted that the "changes could be attributed to the attention the patients received." (I'm not sure, but I think that last part means the study might have been invalid.)

One of the studies I looked at (that reported remarkable results) was basically a case study of a single patient. Another study, with the lofty title "Effects of Gluten on Schizophrenics," reported: "Thirteen schizophrenics were given gluten-free peanut-flour supplementary cookies and 13 were given virtually identical cookies with gluten added. Tests and rating scales before and after the ten-day study period showed no greater improvement for those receiving the gluten-free cookies than for those receiving the gluten-added cookies." From this, the authors of course concluded that gluten restriction is not at all beneficial, completely overlooking the fact that peanut cookies are not exactly hypo-allergenic. Recipe considerations aside, are we to decide an issue as complex as the role of gliadins in schizophrenia on the basis of just 13 people eating some cookies? I for one consider this kind of "study" a pathetic waste of research dollars.

Bottom line, research on the role of wheat in schizophrenia is hardly what you'd consider strong. The studies are sloppily designed, poorly controlled, use ridiculously small numbers of study subjects, employ crude methods of monitoring patient progress, etc. (yet of course they apply sound statistical methods to the numbers to make them look like solid science). The papers I saw wouldn't make good science-fair projects. All of them were laughably poor in one respect or another.

On the other hand, I can't claim to have read every single study that's ever been done in this area. Perhaps you'd like to dig into this yourself? If so, start with some of the references mentioned here. (And leave a comment below if you find something interesting.)

In the meantime, what I feel comfortable saying at this point is that it's way too early to rule out any role for gluten, zonulin, intestinal permeability, etc., in the pathogenesis of schizophrenia. I remain hopeful that good work in this area will come along sooner or later, and we'll eventually see some interesting results. For now, though, the science is pathetically weak. I don't think we dare draw any firm conclusions yet.