Antipsychotics and Long-Term Functioning in Psychotic Disorders

Tackling the “superior first-level compellingness” of the antipsychotics-make-schizophrenia-worse thesis

In a post from 2023, A Rationalist Approach to Psychiatric Conspiracy Theories, I discussed how many ideas in antipsychiatric circles can have “superior first-level compellingness” (borrowing the phrase from Scott Alexander’s description of conspiracy theories). The idea that antipsychotics worsen the course of schizophrenia, popularized by journalist Robert Whitaker in Anatomy of an Epidemic (2010), is one of those provocative theses that is supported by superficial facts that can seem compelling to many smart people but that have failed to persuade the mainstream psychiatric community.

I wrote in the post:

“In the case of psychiatric-medications-make-mental-illness-worse arguments, I had to do a lot more figuring out on my own. Again, I found the bulk of mainstream response to be inadequate. Becoming deeply familiar with the research literature myself, engaging with some specific instances of it, discussing these issues with close colleagues, triangulating research with clinical experience, and then later on teaching trainees on these issues was all quite helpful. Part of the solution was appreciating the ways in which psychiatric medications can and do negatively affect the functioning of many individuals (the grain of truth in the argument), and distinguishing that question from the question of the worsening of illness itself. And part of it was appreciating the methodological issues on a gut level that can produce a very compelling impression of worsening outcomes in longitudinal data.”

Regular readers of Psychiatry at the Margins will remember that I detailed my own understanding of the literature around antipsychotics and long-term functioning in schizophrenia in a post in June 2024. That piece has now been published in the journal Psychological Medicine as an academic article:

Making sense of the literature on antipsychotics and long-term functioning: taking natural history and personalization seriously (open access)

The published article is updated, includes a discussion of the RADAR study in addition to the OPUS study, Chicago Follow-up study, and Mesifos trial, and incorporates other revisions made in response to comments by peer-reviewers.

I do encourage you to read the full article to appreciate the finer details. For the benefit of the readers, I am posting the introduction, clinical discussion, and conclusion sections below.

---

Introduction

At first glance, the literature on the relationship between long-term antipsychotic use and individual functioning may suggest that antipsychotic medications worsen outcomes (Harrow & Jobe, 2018; Whitaker, 2011). However, this perceived association is deeply confounded by the natural course of psychotic disorders and by individual characteristics that influence the likelihood of continued medication use. The academic debate on this topic has been intense, and in communities critical of mainstream psychiatry – such as Mad in America and Mad in the UK – the belief that antipsychotics negatively impact functioning is widely accepted and promoted. In contrast, responses from the psychiatric community have often focused on highlighting the methodological flaws in these studies (Aftab, 2023; Leucht & Davis, 2017; Pierre, Zito, Yang, & Marder, 2023), leaving clinicians uncertain with regards to how these findings should inform clinical practice (with notable exceptions, such as Wunderink, 2019 and Wunderink, 2024).

This review aims to clarify the complicated literature on antipsychotic use and long-term functioning from the vantage point of clinical practice, considering the heterogeneity of natural course of psychotic disorders and its implications for personalized care. Rather than offering a comprehensive or systematic review of research studies, the focus of this paper will be on a few key studies that exemplify broader trends in the literature and will illustrate how they can be clinically interpreted with attention to the natural history of these disorders and the need for individualized treatment.

Summary of methodological and interpretative issues

To the extent that we have been able to study the effects of antipsychotics under randomization, antipsychotics show beneficial effects on psychotic symptoms, relapse prevention, quality of life, and social functioning. Randomized trials, however, have generally been of less than 2-year duration. All studies longer than 2 years are either observational or quasi-randomized. With long-term observational studies of antipsychotics, we are forced to tackle the natural history of psychotic disorders and diverging clinical trajectories. We are also forced to confront the fact that those who successfully discontinue antipsychotics are systematically different from those who are unable to discontinue or unsuccessfully discontinue and then resume. They are likely to be either folks who are naturally on the long-term remission trajectory, or those with a recurrent course experiencing a transient periods of remission or those with chronic symptoms who have the psychological and social resources to manage symptoms without antipsychotic medications.

Clinical implications

There are three situations in which it could be said that a person doesn't ‘need’ antipsychotics

• The person is in the long-term remission trajectory.

• The person can manage with their psychotic symptoms equally well (or reasonably well) via psychological and social interventions alone (this is a poorly characterized subset, but we know such patients exist).

• A third category would be the situation where antipsychotics are warranted but have proven to be ineffective or not tolerated (i.e. with adequate trials of multiple antipsychotics, including clozapine). Almost everyone is able to tolerate at least some of the available antipsychotics, but the ones tolerated might not be the ones effective.

Considering the effects of antipsychotics in light of these possibilities generates the following scenarios with regards to the net effect of antipsychotics on functioning:

Scenario #1: If a person stays on antipsychotics when they could've safely gotten off them (because they happen to be among those in a long-term remission trajectory or because they are in the poorly characterized subset of folks who can manage equally well via psychological and social interventions), the net effect on functioning is likely negative.

Scenario #2: If a person goes off antipsychotics that were effective for them, and they caused minimal side effects and the person tolerated them well, and if they have a recurrent or chronic form of psychotic illness, then they are at a higher risk of experiencing future episodes or exacerbations of psychosis, and their functioning is likely negatively affected by discontinuation (i.e. the net effect of antipsychotics on functioning is positive).

Scenario #3: If a person goes off antipsychotics that were effective for them and they caused substantial side effects, and if they have a recurrent or chronic form of psychotic illness, their functioning may improve or worsen or be unchanged depending on which – psychosis or medication adverse effects – had a bigger negative impact on their functioning.

Scenario #4: If a person goes off antipsychotics that were largely ineffective for them and they have a chronic persistent form of psychosis, their functioning may either stay the same (if they had minimal side effects) or may improve (if they had substantial side effects)

As reviewed in the initial section of this paper, there is considerable heterogeneity in reports from longitudinal studies. Scenario #1 probably applies to around 20–30% of people with first-episode psychosis (and to a substantially lower percentage of people with multi-episode schizophrenia spectrum psychosis) but this will vary from sample to sample (Chakraborty, 2021; Wunderink, 2024). Patients in this scenario possibly discontinue medications on their own, stay well, and don't stay in contact with clinical services. But many of them will be in clinical care, and clinicians have to actively consider dose-reduction and discontinuation as an option after a period of stability following psychosis, especially first-episode.

For people with recurrent psychosis who try dose reduction and discontinuation, they either find that they cannot successfully discontinue because psychotic symptoms soon return or they end up restarting antipsychotics for acute treatment at a later point when they experience psychosis again. Some people in scenario #3 may find it preferable to use antipsychotics only for acute treatment, but this strategy does come with accompanying risks, as many lose insight during acute states of psychosis, and there are serious legal, financial, and social risks to consider (incarceration, homelessness, alienating family, etc.) in addition to the disruption that hospitalization brings.

An important factor with regards to functioning and/or successful antipsychotic discontinuation is the availability of competent psychotherapists who are skilled in working with people with psychotic disorders. This is especially the case for those with recurrent or chronic symptoms. In my clinical experience, the people who successfully discontinue antipsychotics tend to be people with high premorbid functioning, who are highly motivated, engaged in psychotherapy, have good psychological insight, and strong social support.

If patients are actively experiencing medication effects such as feeling mentally dull, numb, ‘like a zombie,’ sedated, clinical experience suggests that a dose reduction or change in medication does usually improve functioning. It is also reasonable to think that those who experience metabolic adverse effects such as weight gain or pre-diabetes/diabetes can also experience decline in functioning, and dose reduction or medication switching can help with that. If a patient is tolerating a dose of the antipsychotic medication with minimal adverse effects and minimal impact on the patient's functioning (self-reported or otherwise), available literature doesn't provide strong evidence to suggest that dose reduction or discontinuation will be of functional benefit. The clinical goal, in my opinion, is to maintain patients on a medication dose, ideally near the low end of the therapeutic dose range, that they tolerate and that doesn't interfere with their functioning. Some, such as Wunderink, 2019, have proposed to gradually taper the antipsychotic dosage after first episode stabilization to half the daily defined dosage, ‘if increased relapse risks are acceptable.’ This is not unreasonable and will indeed benefit some people, but if a person is tolerating standard therapeutic dose without functional impairment, a higher risk of relapse in pursuit of an uncertain long-term functional benefit will not be acceptable to many patients and clinicians.

The problem arises when the medication or the dose of the medication required for clinical stabilization is one that actively produces adverse effects. E.g., when patients require high doses, or polypharmacy, or when they don't respond to first-line medications and require clozapine, which has more adverse effects and is less well tolerated. In such cases, we are forced to balance active symptom control or reducing the risk of relapse with active adverse effects. Ideally, the decision should be an informed one by the patient, but in practice, the culture when it comes to the treatment of schizophrenia spectrum disorders is extremely paternalistic. Clinicians want to minimize relapse, and other considerations are often secondary to them, even if those other considerations are more important to the patient. Resultantly, many patients are pressurized to stay on medication doses that they do not tolerate well, and patients are offered little to no support by the clinicians if they decide to reduce or discontinue the medication. Clinical decision-making here is made more complex when patients have poor insight or a poor appreciation of the risks, when there are concerns by family members about decompensation, and when, especially in places such as the US, there is a real risk of becoming homeless or being arrested. There are good reasons why clinicians who have skin in the game are so conservative in the management of psychotic disorders. While this clinical attitude of prioritizing relapse risk over functioning and tolerability keeps many patients stable, it also harms those who would've benefited from a personalized reduction, change, or discontinuation of medication.

An important consideration here is that there is a scarcity of established guidelines and practical ‘know-how’ when it comes to dose-reduction and discontinuation of antipsychotics (see Horowitz, Jauhar, Natesan, Murray, & Taylor, 2021, for a proposed method of tapering antipsychotics). For clinicians to be more responsive to patient preferences for dose reduction and discontinuation, the psychiatric community must address existing knowledge gaps with a research program that informs us about optimal tapering speed, accompanying psychosocial supports for successful tapering, and strategies for monitoring early warning signs.

All this demands personalization of antipsychotic treatment. This is rather obvious in many ways, and commentators on both sides of the debate are likely to agree on the need for personalization even when they disagree on what the ideal dose target is. Unfortunately, most clinical services continue to insist on long-term maintenance treatment for everyone and fail to support (or adequately guide) dose-reduction and discontinuation – driven by conservatism, paternalism, risk aversion, and outdated knowledge. On the other hand, in many critical and survivor spaces hostile to psychopharmacology, antipsychotics are demonized and everyone is encouraged to discontinue. Any current attempt at personalization of antipsychotic treatment will be crude and uncertain, but it is what patients deserve. Lex Wunderink puts it well: ‘Even small steps forward will easily outperform the current one-size-fits-all approach.’ (Wunderink, 2019)

Conclusion

The clinician faces a difficult task when it comes to pharmacological treatment of psychosis. She has to be mindful of and take into account:

• Variable trajectory of psychotic disorders, with the most common trajectory in schizophrenia spectrum being no remission and no recovery, but a subset showing remission and recovery for extends periods of time.

• Antipsychotics are beneficial for the average patient in the acute treatment of psychosis and over 1–2 years, however, there are subsets of patient who do not respond adequately to antipsychotics (including clozapine) and for whom the adverse effects exceed benefits offered.

• Working with patients who lack insight into their condition, or who at high risk of suicide, violence, homelessness, or incarceration.

• A paternalistic medical and societal culture that prioritizes relapse prevention over patient functioning and quality of life.

• Long-term studies of antipsychotics are largely observational and do not adequately control for natural history or individual characteristics, while randomized trials of maintenance treatment show benefits with regards to symptom control, relapse prevention and functioning.

• A binary insistence on being on or off medications in communities critical of psychiatry, while ignoring dose reduction as a viable harm reduction strategy.

Psychiatric treatment of psychotic disorders has history suffered from a lack of recognition of natural history variation, and neglect of iatrogenic harm and patient preference driven by a culture of paternalism and genuine difficulties that arise in working with impaired insight, resulting in a lack of personalized treatment. The body of literature on the association of long-term antipsychotic use and functioning isn't sufficient to favor antipsychotic discontinuation or dose reduction below standard doses for most patients, but it is sufficient to highlight the necessity of personalization of clinical treatment.

---