On a recent PBS television show hosted by Charlie Rose on the "mentally ill brain," Columbia University's Jeffrey Lieberman presented a series of brain scans of a person with schizophrenia, which showed enlarged ventricles and thus, as Lieberman told the audience, "loss of brain gray matter."

(See minute 29:30 of video.) The idea being presented by Lieberman was that schizophrenia is a neurodegenerative disease, characterized by brain tissue loss.

In the discussion, Lieberman also implied that antipsychotic medications, in some way, protect against this neurodegenerative process. The drugs "stabilize the illness," he said. But when people "stop taking the medicines, they get sick again, and when this happens, they have repeated insults to the brain . . . and this leads to a progression, and the progression, like somebody who has multiple little strokes, can lead to some decline in which people are not able to recover to the same level. And if you actually take brain scans over time, you can see the subtle, perceptible loss of brain grey matter."

To the public, this is a new paradigm for understanding why antipsychotics are an essential treatment for schizophrenia. Schizophrenia is a neurodegenerative illness, characterized by brain tissue loss, and antipsychotics are "neuroprotective" agents that thwart that pathological process in some way.

Given that this idea is taking hold, it seems worthwhile to check the literature to see if it is well grounded in science.

MRI Scans

Making sense of MRI studies of schizophrenia patients can be a difficult task, partly because the results can be so inconsistent, and partly because the results are confounded by exposure to antipsychotics. However, earlier this year, Joanna Moncrieff and Jonathan Leo brought new clarity to this topic by analyzing the studies based on the patients' exposure to antipsychotics.

Here is a summary of their findings, which they published in Psychological Medicine.

Studies of chronic patients never exposed to antipsychotics

They identified three studies of schizophrenia patients who were ill for extended periods of time and were never exposed to antipsychotic medications. In comparison to the "controls" in the studies, these never-exposed patients showed "no major differences in global cerebral, grey-matter, ventricular, or CSF (cerebrospinal fluid) volumes."

Studies of first-episode patients with limited exposure to antipsychotics

A number of studies have found brain abnormalities in first-episode schizophrenia patients, and these findings have been seen as proof that the abnormalities must be due to the disease and not the treatment. But as Moncrieff and Leo noted in their paper, many of the patients in the first-episode studies had been on antipsychotics for months, and there is evidence that antipsychotics may induce changes in brain structures in a short period of time. As such, the results from first-admission studies -- as a group -- are confounded by drug exposure.

To counter this problem, Moncrieff and Leo analyzed the results from studies of schizophrenia patients who were either drug naïve or had limited exposure to antipsychotics (on average less than four weeks.) They found 18 such studies (in addition to the three studies discussed above.)

In 13 of them, researchers did not find any differences between the patients and controls in whole-brain volumes, total grey matter, and CSF volumes. Of the five studies that did show differences in brain volumes, one included a subset of patients that had taken antipsychotics for up to 24 weeks, which could have confounded the results. A second study reported differences in an initial sample of 18 patients, but then, in a larger sample of 51 patients, no global differences were found. The remaining three studies found a difference in the size of the "third ventricle only;" a "trend level reduction in whole-brain volume;" and "smaller cerebellar volumes."

Longitudinal studies of patients treated with antipsychotics

In 14 of 26 MRI studies of schizophrenia patients treated with antipsychotics for periods ranging from eight months to 10 years, the patients "showed a greater reduction in whole-brain, cortical or grey-matter volumes, or a greater increase in CSF or ventricular volumes, compared with controls."

Several of the 14 studies quantified the brain-volume loss. In adult patients, "grey matter, whole-brain or cerebral volume showed a decline of 1.2% to 2.9% per year."

Summing Up The MRI Studies

With the MRI literature parsed in this way, it's easy to see that the studies do not provide convincing evidence that schizophrenia patients, if they were not treated with antipsychotics, would regularly suffer "a loss of brain gray matter." A reduction in whole brain volumes is not regularly seen in first-episode patients with limited exposure to neuroleptics, and it was not seen in the three longitudinal studies of never-exposed patients. The fact that it shows up in the majority of studies of drug-treated patients simply leads to this possibility: It could be due to a disease process, or it could be due to the medication, or it could be due to a combination of the two.

The Effect of Antipsychotics on Brain Volumes

To best assess the long-term effects of antipsychotics on brain volumes and to distinguish drug effects from disease processes, you would need to run a long-term study that compared brain volumes in four groups: healthy people on the medications, healthy people off the drugs, schizophrenia patients on the drugs, and schizophrenia patients off the drugs. But you can't put healthy people on antipsychotics and it is considered unethical to withhold antipsychotics from schizophrenia patients for long periods, and so you can't run studies of that type. However, animal research and MRI studies of medicated schizophrenia patients do provide evidence that antipsychotics may cause "brain gray matter loss."

In animal studies, typical antipsychotics have been found to cause neuronal loss and gliosis in the striatum, hypothalamus, brain stem, limbic system and cortex. In a study of non-human primates (macaque monkeys), a daily dose of haloperidol or olanzapine for 18 months led to an 8% to 11% reduction "in mean fresh brain weight compared to controls."

Next, as Moncrieff and Leo reported, a majority of studies of schizophrenia patients on the drugs for longer periods found that they had reduced brain volumes, while in the three longer studies of never-exposed patients, there was no such finding. Furthermore, researchers have found that gray matter loss varies according to whether a typical or atypical antipsychotic is prescribed, and if the drugs did not affect brain volumes, there shouldn't be this variability. Lieberman and others recently conducted a study of this type, mapping the cortical changes seen in 36 first-episode schizophrenia patients treated with either haloperidol or olanzapine for one year. 

Here are their results, which they published in 2009:

• In the haloperidol patients, "a dynamically spreading wave of significant gray matter loss was detected . . . progressive gray matter reduction began in lateral parietal-temporal cortices by three months, spreading into the dorsolateral, medial frontal and prefrontal cortices by six months, and involving most of the frontal cortex by one year after the first psychotic episode." The "total loss is severe by 12 months."

• In the olanzapine patients, "regions of significant progressive gray matter loss were found at all time points, but these were less intense and widespread. These changes also evolved in a distinct antatomical trajectory" compared to the haloperidol-treated patients.

So how should these results be interpreted? Given the effects of haloperidol and olanzapine in macaque monkeys, you might conclude that both drugs cause brain gray matter loss in schizophrenia patients, with olanzapine perhaps less toxic than haloperidol. However, since there was no control group in this study and thus no mapping of gray matter loss in unmedicated schizophrenia patients, the researchers concluded that it was also possible that olanzapine was "neuroprotective," i.e. that it slowed down the loss seen in the haloperidol patients (which presumably was due mostly to the disease.) But, they wrote, without a control group "it is difficult to be sure that the effect is either due to toxic effects of one drug versus protective effects of another (among other interpretations), and it cannot be determined what the normal trajectory of such change is in schizophrenia."

The Bottom Line

While the public may be hearing that schizophrenia is a neurodegenerative disease characterized by a loss of brain gray matter, with atypical antipsychotics neuroprotective against that process, even a quick review of the scientific literature reveals that it is unclear whether the gray matter loss is due to the disease, or to the drug, or to a combination of both.

Here's what Moncrieff and Leo wrote in their conclusion : "Overall, there seems to be enough evidence to suggest that antipsychotic drug treatment may play a role in reducing brain volume and increasing CSF or ventricular spaces . . . although it remains possible that the underlying disease process also causes brain volume changes, we suggest that antipsychotic drug treatment may be responsible for some of the changes that are usually attributed to schizophrenia."

You are reading

Mad in America

The Real Suicide Data from the TADS Study Comes to Light

NIMH hid suicide risk in pediatric study

Can Unethical Research Ever Lead to "Best Evidence?"

Study Was Expected to Make Psychotic Patients Worse