Commentary: Does First Year Treatment With an Antipsychotic Lead to Long-Term Harm?

Today, we published an article about a Dutch study that strengthens the evidence for providing first-episode psychotic patients with an opportunity to taper from their antipsychotic medication. The study was a randomized clinical trial, and at the end of four years, the tapering group had better functional outcomes than the group randomized to standard care.

It’s heartening to see that a study of this type was done. In clinical studies that chart long-term outcomes for patients diagnosed with schizophrenia and other psychotic disorders, the recovery rate is regularly higher for those who manage to stop taking antipsychotic medication. Those findings provide a compelling reason to provide tapering support to first-episode patients, with such support hopefully maximizing the percentage of patients who are able to successfully get off the medication and stay well.

In this Dutch study, the researchers didn’t attribute the tapering group’s better outcomes at four years to a difference in the long-term use of antipsychotic medication. The intervention period ended after six months, and after that it became a naturalistic study, with patients in both groups able to change their usage of the drugs, and somewhat surprisingly, after the end of one year, the use of antipsychotics was quite similar in both groups throughout the next three years, with a mean daily dose in olanzapine equivalents of 4 to 5 mg.

However, there is an “evidence-based” reason for the researchers to further investigate this medication question. There was a difference in exposure to antipsychotic medication during the first year of this study, and the charts in this study beg this question: Does a first year of antipsychotic medication at a standard dose foretell a poor long-term course?

Here is the chart that shows the difference in antipsychotic exposure during the first year of the study. Both groups started at a mean daily dose equivalent to 9.3 mg of olanzapine, and then for the next 12 months the maintenance group had exposure to significantly higher doses of an antipsychotic than the tapering group. (The blue dot in the graphic is the maintenance group; the orange dot is the tapering group.)

Now, it is well established that regular use of antipsychotics can induce brain changes that impair long-term outcomes, and in two distinct ways.

First, antipsychotics block dopamine receptors in the brain, and this triggers a compensatory response marked by an increase in dopamine receptors. In a series of articles that date back to the late 1970s, Canadian researchers told of how this drug-induced dopamine supersensitivity could make patients more biologically vulnerable to psychosis. This dopamine supersensitivity, they explained, increases the risk of relapse when patients try to come off the medication, and yet can lead to chronic, more severe  symptoms when patients stay on an antipsychotic.

There is reason to think that a higher dose of an antipsychotic will increase the strength of this compensatory response, and there is also reason to worry that the longer a person is exposed to the drug, the more likely that this supersensitivity will persist, even when someone tapers from the medication. This raises the possibility that in this Dutch study a dopamine supersensitivity may be present in the maintenance patients at the end of one year that is less pronounced or absent in the drug tapering group, and that this difference will persist in the following years. If that is the case, then one might expect that symptom scores for the two groups would diverge somewhat after the first year, even though daily medication after that point normalized for both groups at a low daily dose.

That is, in fact, what you see in the Positive and Negative Syndrome Scale (PANSS) scores for the two groups. After the end 12 months, the PANSS scores for the maintenance group worsened, while they improved for the tapering group.

The second iatrogenic element with antipsychotics is that they have been found to shrink brain volumes, with this shrinkage mostly occurring during the first year of exposure. It is also dose-related. Moreover, as Nancy Andreasen and colleagues reported, this shrinkage is associated with a worsening of negative symptoms, functional impairments and cognitive abilities. These brain-shrinkage findings suggests that in this study, the reduction in brain volumes may be significantly greater in the maintenance group than in the tapering group at the end of one year, and that this difference in brain volumes would lead to a difference in global functioning between the two groups after that period.

That is what shows up in their chart on “global assessment of functioning.” The global functioning of the maintenance group steadily worsens after the first year, whereas global functioning in the tapering group improves after the first year, such that there is a significant difference in global functioning between the two groups at the third-year and fourth-year assessments.

While this is best described as a speculative explanation for the difference in outcomes, nevertheless it’s easy to see the concern that does arise from this study. Does regular exposure to antipsychotics during the first year take a long-term toll, even if the patients subsequently get down to a low daily dose? Do these two iatrogenic effects settle somewhat permanently into the brain after one year of regular exposure?

That question in turn begs another “what if” question. What long-term outcomes might emerge from a selective-use protocol that allowed for a sorting of first-episode patients into three groups: those who could recover without exposure to antipsychotics; those who needed the medication for a period of time but could then successfully taper from the drugs; and those who, for whatever reason, needed to remain on the drugs long-term?

Although the drug remembered today as the first antipsychotic, chlorpromazine, was introduced into asylum medicine in 1955, there have been very few efforts to assess treatment of first-episode patients with this selective use model. However, three notable efforts of this kind all found that a significant percentage of first-episode patients recovered without exposure to antipsychotics, and that it was this never-exposed group that had the best long-term outcomes.

The first such study was conducted by Maurice Rappaport and his colleagues at the University of California, San Francisco in the 1970s. They randomized 80 young male schizophrenics admitted to Agnews State Hospital to drug and non-drug groups. After the patients were discharged from the hospital, they were free to change their medication usage, and thus at the end of three years Rappaport ended up with four groups:

A) Drug-free group in the hospital and stayed off the drugs following discharge.

B) Drug-free group in the hospital and started taking the medication following discharge.

C) Medicated group in the hospital and stopped taking the medication following discharge.

D) Medicated group in the hospital that remained medication compliant following discharge.

Here is how the outcomes of the four groups stacked up:

As the table shows, two-thirds of those randomized to the non-drug group in the hospital (24 of 41) never took antipsychotics during the three years, and it was this never exposed group that had by far the best outcomes. Here is what Rappaport and his collaborators concluded:

During this same period, Loren Mosher, who was then head of schizophrenia studies at the NIMH, conducted his famous Soteria experiment (which ran for 10 years). Patients were randomized either to conventional care in a hospital setting or to care in a residential setting that used antipsychotics in the selective way described above. At the end of two years, 42% of the Soteria patients had never been exposed to antipsychotics, 39% had used the drugs for a shorter period of time, and 19% ended up taking them continually. Moreover, in comparison to the hospitalized patients treated conventionally with antipsychotics, at the end of two years the Soteria patients had “lower psychopathology scores, fewer [hospital] readmissions, and better global adjustment.”

Mosher wrote:

Mosher’s fellow psychiatrists did not appreciate this finding, and he was ousted from his position at the NIMH as head of schizophrenia studies. Since that time, it is difficult to find any U.S. study that involved treating newly diagnosed psychotic patients without the immediate use of antipsychotics. However, that was the protocol adopted in Tornio, Finland beginning in 1992.

In their “Open Dialogue” practice, first-episode psychotic patients were not immediately put on antipsychotics. Instead, they relied on regular “dialogical” meetings with the patients and their families as their first-line treatment, and as long as the patient’s “grip on life” improved during the initial weeks—as evidenced by a willingness to engage in the dialogical meetings, or by taking better hygienic care, and so forth—they continued to forgo the use of antipsychotics. However, if patients didn’t improve during this initial period, then they offered them the choice of using lower doses of antipsychotics for a period of time (six months or so), and then at the end of that time, they would see who could successfully taper from the medication.

This was the standard of care for all psychotic patients treated in this region of Finland for at least 20 years, and it produced extraordinary results. At the end of five years, 79% of their patients were asymptomatic; 73% were working or in school; 7% were unemployed but looking for work; and only 20% were on government disability. As for antipsychotic usage, 67% of their patients had never been exposed to the drugs, 13% had used them for a time, and the remaining 20% took them continually.

As I mentioned at the start of this comments, it is heartening to learn of studies assessing the impact of antipsychotic-tapering protocols on long-term outcomes. But even as those studies are conducted, the assumption is that immediate use of the drugs is so well-established that it would be medically negligent to conduct a study that didn’t involve immediate use of antipsychotics. However, that is not an evidence-based assumption.

In 2017, Leucht and colleagues conducted a meta-analysis of all RCTs of antipsychotic trials since the introduce of chlorpromazine in 1955. They found 167 studies that met their inclusion criteria, with 28,102 patients in these studies. Their report included two extraordinary findings:

a) The first was that the antipsychotic-placebo difference on the 210-point PANSS scale (used to measure positive and negative symptoms) was only 9.6 points. This difference doesn’t rise to the level of a “minimum clinically important difference.”

b) The second was that they couldn’t find any good-quality RCTs of antipsychotics in first-episode or early-episode schizophrenia. The results from their meta-analysis, they wrote, were “representative only for chronically ill, often previously treated patients.”

In short, there is no good evidence that antipsychotics provide a benefit to first-episode or early-episode psychotic patients. Furthermore, the limited evidence that does exist on this question, as reviewed above, suggests that 40% to 67% of such patients could recover without exposure to the drugs, and that this group would, by far, have the best long-term outcomes.

So yes, it is heartening to see the drug-tapering studies. But what is also needed, and needed badly, are studies that assess long-term outcomes with selective-use models of care that avoid immediate use of antipsychotics. This Dutch study reinforces that need, for it does raise a question of whether a first year of treatment with antipsychotics, even at a moderate dose, induces iatrogenic changes in the brain that persist and take a long-term toll.