Of all the research that has been done in the past 20 years on psychotic disorders, Martin Harrow’s ongoing study of long-term outcomes in such patients is, in my opinion, the most important work. He and his colleagues have now published their 20-year results. Given the 15-year outcomes data he published in 2007, his latest findings should not be surprising. The schizophrenia and schizoaffective patients who took antipsychotics regularly during the 20 years, compared to those who quit taking the medications (usually within the first two years), experienced more psychosis, more anxiety, and markedly fewer periods of “sustained recovery.” They were also more cognitively impaired.
The same dramatic difference in outcomes was seen in patients diagnosed with a mood disorder with psychotic features: those who stayed on antipsychotics fared markedly worse over the long-term.
The psychiatric establishment mostly ignored Harrow’s 2007 report, or tried to explain away his findings. But with the publication of the 20-year results, I think the time has come for psychiatry—and our society—to take a close look at his research, and to try to honestly assess what is going on. A full-bodied inquiry is essential for another reason too: We are now prescribing antipsychotics to an ever larger number of children, and to many non-psychotic adults as well, and if antipsychotics are worsening the long-term outcomes of people with a psychotic disorder, which is the obvious concern raised by Harrow’s findings, then we really need to rethink the use of these medications in those other populations.
Here is a review of the study’s design, and the findings from the two papers.
From 1975 to 1983, Harrow, who is a psychologist at the University of Illinois, enrolled 200 people with a diagnosis of schizophrenia or a milder psychotic disorder into his NIMH-funded study, recruiting the patients from two Chicago hospitals. One was private and the other public, as this ensured that the patient group would be economically and ethnically diverse. The enrolled patients had a median age of 22.9 years, and for 46%, this was their first hospitalization. Another 21% of the patients had had one previous hospitalization. Thus, this study largely charted long-term outcomes in patients newly diagnosed, or recently diagnosed, with a psychotic disorder. (See chart).
All of the patients were treated conventionally in the hospital, and then Harrow followed them as their lives unfolded, periodically assessing how well they were doing. Were they symptomatic? In recovery? Employed? Were they taking antipsychotic medications? Or other psychotic medications? He conducted such reviews at 2 years, 4.5 years, 7.5 years, 10 years, 15 years, and 20 years. At the end of 15 years, he had successfully followed 145 of the 200 patients enrolled into the study, and at the end of 20 years, he had outcomes data for 139 patients. For a long-term study, this is a very good retention rate.
The 15-Year Results
In 2007, Harrow published the 15-year results for the 145 people in the study. Of the 145, 64 were diagnosed with schizophrenia, and the remaining 81 with milder psychotic disorders (schizoaffective disorder, bipolar with psychotic features, and unipolar depression with psychotic features.) Here were his findings:
1. Recovery rates in schizophrenia group
At the end of two years, those who had stopped taking antipsychotics were doing slightly better on a “global assessment scale” than those taking an antipsychotic. Then, over the next 30 months, the collective fates of the two groups began to dramatically diverge. The off-med group began to improve significantly, and by the end of 4.5 years, 39% were in recovery. In contrast, outcomes for the medicated group worsened during this 30-month period. As a group, their global functioning declined slightly, and at the 4.5-year mark, only six percent were in recovery, and few were working.
That stark divergence in outcomes remained for the next ten years. At the 15-year followup, 40 percent of those off antipsychotics (25 of the 64 patients) were in recovery, compared to five percent of those taking antipsychotics. (To be in recovery, a person had to have no positive or negative symptoms; couldn’t have been hospitalized in the previous year; and adequate work and social functioning.) See charts for recovery rates and for global adjustment.
2. Spectrum of outcomes in schizophrenia group
Harrow divided long-term outcomes for the 64 schizophrenia patients into three categories: recovered, fair, and uniformly poor. Of the 25 patients who stopped taking antipsychotics, 10 recovered (40%), 11 had fair outcomes (44%), and 4 (16%) had uniformly poor outcomes. In contrast, only 2 of the 39 patients who stayed on antipsychotics recovered (5%); 18 had fair outcomes (46%), and 19 (49%) had uniformly pair outcomes. (See chart.) In sum, medicated patients had one-eighth the recovery rate of unmedicated patients, and a threefold higher rate of faring miserably over the long term.
3. Psychotic symptoms in the schizophrenia group
At the 10-year follow-up, 23% of the patients off antipsychotics were experiencing psychotic symptoms, versus 79% of those still on the drugs. At the 15-year followup, 28% of those off antipsychotics had psychotic symptoms, versus 64% of those on the medications. (See chart).
4. Global adjustment of those with milder psychotic disorders
At the end of two years, those with milder psychotic disorders who got off antipsychotics were doing somewhat better than those on the drugs. This difference in global outcomes became pronounced by the end of 4.5 years, with those off medications doing markedly better, and remained so throughout the 15 years. (See chart.)
5. Global outcomes of all 145 patients
Harrow provided global adjustment data for all four groups in his study: schizophrenia on meds, schizophrenia off psychiatric medications, milder disorders on psychiatric meds, milder disorders off. At the end of 15 years, the global outcomes for the four groups lined up like this, from best to worst: Milder disorders off meds, schizophrenia off meds, milder disorders on meds, and schizophrenia on meds. (See chart.)
6. Global outcomes for schizophrenia patients by prognostic type.
At the start of the study, Harrow grouped his schizophrenia patients into two subgroups: those with a good prognosis and those with a bad prognosis. Although he didn’t provide the global data for these two subtypes, he did report this finding: “In addition, global outcome for the group of patients with schizophrenia who were on antipsychotics were compared with the off-medication schizophrenia patients with similar prognostic status. Starting with the 4.5-year followup and extending to the 15-year follow-up, the off-medication subgroup tended to show better global outcomes at each follow-up.”
Interpreting the 15-year findings: Harrow’s explanation
In his discussion, Harrow noted that it was those with a good prognosis (which was characterized by a stronger internal sense of self,) who were more likely to stop taking antipsychotics, and thus stated that his study had simply identified a subset of schizophrenia patients who could fare well off medication. He did not attribute the poor outcomes in the medicated patients to possible iatrogenic effects of antipsychotics.
Another possibility: The drugs are to blame.
In order to explore whether the antipsychotics may be to blame for the poor outcomes, these questions need to be examined:
a) Is there evidence in the scientific literature that antipsychotics might increase a person’s biological vulnerability to psychosis over the long-term? (The dopamine supersensitivity theory.) If so, is this problem seen in the data that shows the medicated patients were much more likely to still be psychotic at the 10-year and 15-year follow-ups?
b) Nancy Andreasen has reported that antipsychotic usage is associated with a decrease in brain volumes over time, and that this decrease in brain volumes is associated with an increase in negative symptoms and cognitive impairment. Is this effect of antipsychotics showing up in the poor recovery rates for the medicated patients over the long-term? Is this why so few worked?
c) Those with milder psychotic disorders could be expected to have a better long-term course than those diagnosed with schizophrenia. Yet, the schizophrenia patients off meds fared better over the long-term than those with milder disorders on the medications. If the drugs have long-term iatrogenic effects, wouldn’t that explain this surprising outcome?
d) In the study, there are several subsets of patients that can be identified, including good-prognosis and bad-prognosis schizophrenia patients. Why, in every case, did the off-med group have better global outcomes over the 15-year followup?
The 20-Year Results
In his current paper in Psychological Medicine (published online), Harrow has grouped his patients in a slightly different manner. He grouped those with a schizoaffective diagnosis, who in the 15-year study had been in the “other psychotic disorders category,” with the schizophrenia patients. There are 70 in this SZ (schizophrenia spectrum) category in his new paper, and 69 now in a “mood disorders with psychotic features” diagnostic group.
Here are the relevant findings from his 20-year report:
1. Psychotic symptoms in SZ group
At the two-year follow-up, about 35% of the SZ group were off antipsychotics, and that percentage remained fairly stable throughout the next 15 years. There was no significant differences in severity of psychotic symptoms between the on-med and off-med groups at two years, but starting with the 4.5-year followup and continuing through year 20, those “who were not on antipsychotic medications were significantly less psychotic than those on antipsychotics.”
2. High anxiety in SZ group
At the two-year followup, about 50% of those on antipsychotics and a similar percentage of those off medications were experiencing “high anxiety.” However, over the next 30 months, high anxiety symptoms soared in the on-antipsychotics group, such that nearly 75% were experiencing this disress by year 4.5%, whereas anxiety markedly declined for those off antipsychotics, such that only about 20% were experiencing this distress by year 4.5. This dramatic difference in anxiety symptoms remained throughout the 15 years, with more than half of those on antipsychotics still suffering from high anxiety at the end of 20 years.
3) Cognitive function in the SZ group
The researchers assessed cognitive function at each followup, with one test assessing ability to access general information, and the other abstract thinking. At three of the six follow-ups, those off antipsychotics showed significantly better cognitive functioning, and in the other three follow-ups, there was a general trend favoring those off antipsychotics.
4) Relapse rates in the SZ group
Harrow and his colleagues assessed whether patients who were not psychotic at a followup then relapsed in the next study interval. The relapse rate was much higher in each instance for the medicated patients. Between the 7.5 and 10-year assessments, the relapse rate was 33% for those on medications, 0% for those off the drugs. Between 10 and 15 years, the relapse rate was 67% for the medicated group and 0% for those off medication. Between 15 and 20 years, it was 25% for those on antipsychotics and 11% for those off medication. This finding revealed that those who remitted off medication were very likely to stay well.
5) Sustained periods of recovery in the SZ group
Of the 70 SZ patients, 24 remained continuously on antipsychotics throughout the 20 years. This was the patient group that were fully medication compliant, yet only 4 of the 24 patients (17%) “ever entered into a period meeting the operational definition of recovery during any of the six follow-ups.” The reasons they failed to do so was either because they were psychotic or not working, Harrow noted.
In contrast, there were 15 in the group of 70 who were off antipsychotics by the two-year follow-up and remained off the drugs throughout the remaining 18 years. Thirteen of these 15 patients (87%) “experienced two or more periods of recovery,” which meant they were both asymptomatic and working more than 50% of the time.
Outcomes in the mood disorders group
At every follow-up, significantly more of the unmedicated patients than those on antipsychotics experienced a period of recovery. Those with mood disorders who stayed on antipsychotics had poor long-term functioning, which was “consistent with the data on the poorer long-term functioning of the schizophrenia (group) who were on antipsychotic medications,” Harrow wrote.
Is it the Drugs?
Thus, we see in these two reports, which arise from the best longitudinal study we have today of long-term outcomes of schizophrenia patients, a consistent story: At every turn, those on medications—as a group--have worse long-term results. They suffer more psychotic symptoms, they are more anxious, they relapse more, they don’t do as well on cognitive tests, and very few enjoy a period of recovery (when the definition of recovery includes returning to work at least part-time.)
This is outcomes data that needs to be closely examined. As we do so, we need to ask whether we should continue with our current paradigm of care, which emphasizes continual use of antipsychotics, or whether that paradigm of care needs to be rethought. Indeed, in this new paper, Harrow and his co-authors do broach the essential question: “Is very long-term treatment with antipsychotic medications undesirable?”
To address that question, we need to look at other research that bears on this question. Is there reason to believe that antipsychotics induce a dopamine supersensitivity, that leads to more chronic symptoms over the long-term? If antipsychotics are associated with a shrinkage of brain volumes, and if studies have shown that as this shrinkage occurs, there is an increase in negative symptoms and cognitive impairment, does this explain why patients on the drugs have poor long-term functional outcomes? We need to assess whether this larger body science tells us that Harrow’s findings are to be expected, as they are consistent with what we have learned about the long-term effects of antipsychotics, or whether there is another plausible explanation—other than the drugs are to blame—for the markedly worse outcomes in the medicated patients.
And I would think that this is a question, given our society’s current widespread use of antipsychotics, of extreme moral urgency.