New Evidence in Support of Tapering Protocols for First-Episode Psychosis

Treatment guidelines generally support trying to discontinue antipsychotics in patients diagnosed with “first-episode psychosis” after one or two years of initial use, but these guidelines also advocate preventing relapses. Consequently, the guidelines often lead to first-episode patients being maintained on antipsychotics indefinitely, as discontinuation is associated with an increased risk of relapse. However, a number of long-term studies have reported higher recovery rates for patients with first-episode psychosis (FEP) who are not on antipsychotic medication, which has prompted at least a few efforts to assess risks and benefits of supported tapering efforts, particularly in populations of first-episode psychotic patients.

A recently published study in JAMA Psychiatry of a randomized clinical trial of 347 first-episode patients in the Netherlands adds to the evidence for providing such tapering support. Those in the “drug-reduction or discontinuation” (DRD) compared to the treatment-as-usual group, had improved functioning and less-intense psychotic symptoms at the end of four years of follow-up.

“In the short term, those [in the DRD group] had higher risk of relapse and lower quality of life,” the researchers wrote. “From 3 years onwards, DRD showed better researcher-rated functioning, with a similar trend for symptom severity.”

Yet, this occurred even though the “mean daily dose” of antipsychotic medication was the same in both groups from years one to four, at the equivalent of 4 mg. of olanzapine. This led the researchers to conclude that “better long-term functioning” of the tapering group “could reflect the learning experience of a guided tapering attempt, rather than direct medication effects.”

The researchers noted that their outcomes were somewhat similar to an earlier drug-discontinuation RCT conducted in the Netherlands by Lex Wunderink. He randomized 128 patients who had recovered from a first episode of psychosis to treatment as usual with antipsychotics or to a drug-tapering program, and while the relapse rate was higher in the tapering group during the first year, by the end of seven years, the recovery rate was more than twice as high for those who had been in the drug-tapering group (40% versus 18%), and cumulative relapse rates were slightly higher in the drug-maintained group.

In this latest Dutch study, the researchers noted that studying discontinuation efforts in first-episode patients is important because the vast majority of people dislike the adverse effects of antipsychotics so much that 70% quit taking antipsychotics within the first year, often without any support from their psychiatrists. For instance, in one study of the drugs’ adverse effects, users reported “struggling to maintain daily functioning due to the side effects . . . weight gain, sleepiness, chronic diarrhea, confusion, dizziness, being cut off from feelings, loss of will-power” and so forth.

Much like in the Wunderink study, the 347 first-episode patients enrolled in this study had been stabilized on an antipsychotic for three to six months, and thus this was a study in voluntary, initial good responders to an antipsychotic.  The 168 people assigned to the DRD group were provided guidance for a six-month hyperbolic tapering course, and encouraged to engage in a collaborative tapering process, one that would lead to at least a 25% dose reduction, with the tapering continued down to either a zero dose or a return of symptoms. The 179 people in the maintenance group would remain on the drug, with a maximum reduction of 25% from their baseline dose.

This intervention lasted six months, and at the end 66% of the DRD group were completely off the medication, with the mean daily dose for this group having dropped from an equivalent olanzapine dose of 9 mg. to 3.8 mg. However, 27% of the maintenance group broke protocol and had completely discontinued antipsychotics, and another 12% had reduced their dosage by more than the stated maximum reduction in the protocol. As a result, the mean daily dose for the maintenance group also dropped—from an olanzapine equivalent of 9 mg. to 6.9 mg.

At the end of this six-month intervention, there was little difference between the two groups on various outcome measures, including relapse. The maintenance group did report a slightly better quality of life (based on the EuroQOL questionnaire), but otherwise there was little to distinguish the two groups. Fewer than 20% in either group relapsed during this intervention period.

After that intervention, researchers conducted a follow-up assessment at the end of one year (six months after the intervention period), and then annually for three more years. During this 42-month follow-up, many in the DRD group who had gone off the medication went back on low doses, while many in the maintenance group reduced their doses. As a result, by the end of year one, medication use was similar in both groups (mean daily olanzapine equivalent of around 5 mg.), and medication use in the two groups stayed more or less at those levels for the next three years.

Yet, in spite of this dose parity, the outcomes for the two groups varied over time, which begs the question of why this would be so.

Four-year Outcomes

The study’s primary outcome was patient-rated functioning, using the World Health Organization Disability Assessment Schedule. The authors found no significant differences between the groups at any time during the four-year study. However, significant differences did emerge on some secondary outcome measures.

As mentioned above, the maintenance group reported a better quality of life at the end of the six-month intervention period. And at the end of one year, the relapse rate in the maintenance group was lower than in the DRD group (22% to 40%). Those were the two main data points that told of an early-stage better outcome in the maintenance group. The quality of life in the two groups had evened out by the end of year one and stayed that way through the rest of the follow-up, and the cumulative relapse rate at the end of four years was actually slightly higher for those randomized to maintenance treatment. Serious adverse events were similar in both groups during the four-year study, although there were three suicides in the DRD group compared to the maintenance group.

Strikingly, though, after the end of two years, outcomes for the two groups began to diverge on scales that measured functional outcomes and psychotic symptoms (the Global Assessment of Functioning and PANSS, respectively). At the third- and fourth-year assessments, the DRD group was functioning better and had better PANSS scores than the maintenance group. While the overall differences on these two scales were not large, the differences were consistent and trending wider. (See GAF and PANSS graphics below. The orange dots are the DRD group, and the blue dots are the maintenance group.)

This result aligns somewhat with Wunderink’s randomized tapering study, in that it found a better long-term outcome for those randomized to the tapering arm of the study. However, Wunderink’s 2013 report also made it possible to calculate recovery rates for those off medication or down to a very low dose at the end of seven years compared to those on a standard dose, regardless of which randomization group they had been in, and in that comparison the medication effect was even stronger. The recovery rate for the off-med/low dose group was 53% versus 17% for those on standard doses.

The authors of this latest study have not yet sorted through their data to see if there is this type of long-term medication effect, with a comparison of outcomes by long-term medication use regardless of their initial randomization group. Yet, in the comparison of outcomes by randomized group, the mean daily dose for the two groups was basically the same at the end of year one and stayed that way through the next three years, and thus this study doesn’t show any direct medication effects. This led the researchers, in their discussion, to speculate that the difference in the two groups’ GAF and PANSS scores at the end of three and four years “may reflect a learning experience to use antipsychotics to better handle psychotic vulnerability.”

The Power of Agency

In an interview with Mad in America, psychiatrist Iris Sommer from the University Medical Center Groningen, lead author on the study, elaborated on this idea that early tapering provided a lasting “learning experience.”

Part of her inspiration for doing such a study, she said, was her own FEP patients’ desires for more autonomy and empowerment over their treatment. “In my clinical experience, almost all patients really want to discontinue medication. That’s really a heartfelt desire for many of them.”

Her study, conducted at 26 specialized psychosis units in the Netherlands, shows that most patients can be empowered to pursue these wishes without too much risk. “The first year, not much good came out of tapering,” she said. “But then things were much different at the longer term. And if we look at two, at three and four years, we actually saw benefits—and we saw only benefits then.”

However, given that medication usage was similar during those last years, there was no obvious reason for the better outcomes for the DRD group. “For me, it was very interesting to think about what exactly was the cause of these benefits,” Sommer said. “We cannot simply say ‘just use as few medications as you can, that will be better.’ That’s not the explanation.”

The research team surmised that people in the DRD group got some enduring benefits from collaborating with their clinicians on tapering. “A first psychotic episode is quite chaotic,” Sommer explained. “It can be very frightening, very stressful, and people often do not really realize what’s happening to them. And medication is started before they actually realize it, and before they really understand what medication it is and what it does and how it will benefit them. And then people are told to keep on using this medication.”

In such circumstances, Sommer added, patients often appreciate slowing down and having deeper discussions with clinicians about their experiences and the pros and cons of medications, and experimenting to learn about what, if any, level of medication they personally find most helpful and tolerable. This study, she said, suggests that people who have such collaborative tapering and learning opportunities will gain knowledge and “autonomy” and will become better managers of their own overall situations over the long term, with reduced symptoms and improved functionality, even if they’re still taking a low dose of antipsychotics.

“They had learned a lot from tapering medication, and I think they were better in balance with both their own psychotic vulnerability and how to cope,” Sommer said.

Tapering Doesn’t Have to Be Scary

Vermont psychiatrist Sandra Steingard, who conducted her own tapering study some years ago with chronic patients, isn’t so sure of Sommer’s explanation for the difference in long-term outcomes . . . so much information still needs to be gleaned from the study data that has been collected. Still, she describes the study as  “terrific” and “valuable.”

Steingard applauded the very existence of a tapering study of this kind. “It’s really hard to do a four-year study. This is a feat,” she said. “They deserve a lot of commendation.”

Steingard called the long-term GAF and PANSS differences between the DRD and maintenance groups “modest,” but noted that, at the very least, the study shows persuasively that tapering can be successfully and relatively safely done by most people after first-episode psychosis. This finding may help prevent clinicians from putting people unnecessarily on high doses of antipsychotics indefinitely—and potentially prevent some patients from experiencing long-term debilitating impacts from those high doses. “What I think the study does is give clinicians permission to offer [tapering] to people, to engage with people in a somewhat different way, and reasonably and conscientiously offer them a different approach. And that’s important… If you can keep people from developing a long-term disability, that’s a hugely valuable contribution.”

Even the higher rates of relapse in the DRD group in the first year, suggested Steingard, are not necessarily cause for undue alarm.

“A lot of times, when people talk about a recurrence of psychosis, they’ll talk about the most dramatic, awful things that happen,” said Steingard. Consequently, many clinical guidelines emphasize trying to avoid relapses at all costs. “But in my experience, a lot of people will have what will meet the definition for a relapse or recurrence, and it’s not devastating.” Sometimes, she said, a “relapse” in a study of this kind can be as simple as someone contacting their psychiatrist and saying they’re having some symptoms and would like to try a higher dose again.

Indeed, Sommer clarified that relapses in their study were often of a more “subtle” nature, and the fact that everyone in the Netherlands has health insurance that provides “intensive” mental health and community-based supports provided a “soft landing” for even more intense relapses.

Furthermore, a myopic focus on preventing relapses, Steingard said, can obscure other important outcome factors,  such as people’s long-term functioning and quality of life. “There’s this whole, old narrative, ‘Psychosis is bad for the brain. You don’t want to take the risk of even one relapse.’ And I just don’t agree with that, and I think this study supports this other side.”

Steingard qualified her remarks by wondering how “generalizable” the study is for other common types of patients, since the researchers excluded anyone who’d been coercively drugged or been violent, and even then, it was only a subset of first-episode patients who agreed to be in the study. Sommer said hundreds—more than half of the people they approached—declined to participate even before the study’s exclusion process began, and so the generalizability is unknown.

Another potential influence on the generalizability of the findings to a North American context is that the starting doses before tapering in this study were already below the lower end of the American Psychiatric Association’s recommended dose range for FEP patients.

Still, Steingard appreciated the researchers’ speculations about the positive impacts of “therapeutic alliance,” informed consent, and shared learning and collaboration between practitioners and clients. “This idea of engaging people in the conversation and helping with their agency has been a value that I’ve had, and when they hypothesize that this may be why people are doing better, it was very reinforcing of these core values,” said Steingard. But, she added, she isn’t entirely convinced that’s the only beneficial factor at work. “How much of it is the social interventions, how much is the [lower doses of] meds—that’s still, to me, an open question.”

The Patient Voices are Missing

Indeed, the brevity of the six-page study report left many open, interesting questions.

For example, the overall cumulative drug burden in the first year was about a third higher for the maintenance group than for the DRD group—and during that period, overall dosages were at their highest. Could this difference have caused lasting impacts? The researchers themselves noted that tapering provided “fewer disadvantages and more benefits” over the long term for women—which they believed “could reflect relative overdosing of female patients.”

In addition, if the maintenance group was indeed less instructed and supported in safe tapering, then many in this group could have been experimenting on their own and experiencing more hazards of intermittent or interdose drug withdrawal. Along these same lines, it would be helpful to hear from the patients themselves about why they continued taking antipsychotics—were these actually autonomous decisions, or did some experience difficulty withdrawing or coercive pressures from families, clinicians, or housing and community service providers?

Perhaps most important of all, how do the people in both groups who’ve now stayed off antipsychotics for several years compare to everyone else? Wunderink provided some insight into that question, and provided even more compelling evidence that, over the long-term, schizophrenia and other psychotic patients off medication have much higher recovery rates, are much more likely to work, and have fewer relapses.

These and other questions may yet get addressed. Sommer estimates that about 15% of the trial participants were still completely off antipsychotics at the four-year mark, and analyzing that subgroup as well as other subgroups and issues are slotted for follow-up papers in coming years. The current paper is part of an ongoing program across 26 specialized psychosis units in the Netherlands that began in 2017, called the Handling Antipsychotic Medication Long-Term Evaluation of Targeted Treatment (HAMLETT) study.

Other papers that have already been published about some of the same patients include studies about cognition during psychosis and treatment, patients’ views on collaborating with treatment providers, the withdrawal effects of different classes of antipsychotics, and dangerous dopamine hypersensitivity and other problematic effects of antipsychotics.

****

See MIA commentary on article: Does First Year Treatment With an Antipsychotic Lead to Long-term Harm?

Sommer IE, de Beer F, Gangadin S, de Haan L, Veling W, van Beveren N, Boonstra N, Rosema BS, van Os J, Kikkert M, Koops S, Noorman J, Thielen F, Wijnen B, Begemann M; HAMLETT-OPHELIA Consortium. Early Dose Reduction or Discontinuation vs Maintenance Antipsychotics After First Psychotic Episode Remission: A Randomized Clinical Trial. JAMA Psychiatry. 2025 Oct 1:e252525. doi: 10.1001/jamapsychiatry.2025.2525. Epub ahead of print. PMID: 41032294; PMCID: PMC12489793. https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2839607