Antipsychotic discontinuation in schizophrenia: rational or reckless? Authors: Robert B. Zipursky, Ofer Agid, Gary Remington Published: 2025 Our understanding of schizophrenia has changed dramatically since the introduction of antipsychotic medication in the early 1950s. Multiple avenues of research have established that schizophrenia is highly treatable, relapses are largely preventable, and those affected can lead lives that are happy and satisfying1,2. Moncrieff and Horowitz, in their paper, “Helping people to discontinue antipsychotics: if, when and how” challenge the consensus that lifelong treatment of psychotic disorders with antipsychotic medication is required after multiple episodes. It is important to first clarify two aspects of their argument: 1) whether it relates to specifically to individuals with schizophrenia versus any psychotic illness, and 2) whether the argument is about discontinuation of antipsychotic medication versus reduction of dose. It has been common practice for programs specializing in the treatment of first episode psychosis to speak with patients and family members about having a “first episode of psychosis” and to limit discussion of specific diagnoses such as schizophrenia and bipolar disorder. While this approach has its benefits when considerable diagnostic uncertainty exists, it is unhelpful when providing recommendations regarding the risks of discontinuing antipsychotic treatment3. Treatment guidelines have rightly focused on the question of antipsychotic discontinuation following a first episode of schizophrenia, for which there is now considerable evidence to support the importance of maintenance treatment4. There are certainly other causes of a first episode of psychosis for which a recommendation of long-term treatment may not be appropriate including brief psychotic disorders and substance-induced psychoses. Many first psychotic episodes are due to bipolar disorder. While almost all people with bipolar disorder require maintenance treatment, mood stabilizing medication is often an option and many can be successfully managed without long-term antipsychotic treatment5. For the sake of clarity, our critique of the argument put forth by Moncrieff and Horowitz relates to the issue of antipsychotic discontinuation in schizophrenia. We will then address the issue of dose reduction. Antipsychotic discontinuation in schizophrenia The argument against long-term antipsychotic treatment for schizophrenia put forth by Moncrieff and Horowitz is based on the premises outlined below, none of which is supported by the weight of current evidence. We strongly disagree with their conclusions and believe that they will put many people with schizophrenia at needless risk of harm.Antipsychotics cause a high burden of debilitating and harmful side effects. While the adverse effects of antipsychotics have been well established, we do not agree that they need be debilitating. There are undoubtedly individuals who feel that antipsychotic medication makes them, “feel ‘braindead’, ‘like a zombie’ or of creating a ‘living hell’”. These severe adverse experiences may be due to pseudoparkinsonism, neuroleptic dysphoria or akathisia. It is unreasonable to expect patients to accept these side effects for the long-term. Instead, they should be and can be managed effectively in the vast majority of cases through treatment optimization, including reducing the dose or switching medications.Antipsychotic treatment interferes with achieving other aspects of recovery that are highly valued, such as functional capacity and the fulfillment of personally meaningful goals. There is no robust evidence to support the inference that the protection from relapse associated with maintenance antipsychotic medication comes at the expense of other functional aspects of recovery. Psychiatrists are not alone in focusing on the importance of preventing recurrence of psychotic symptoms; achieving control of their psychotic symptoms is usually among the highest priorities for people with schizophrenia, ranking higher than work and eliminating side effects6. There are many factors that contribute to the challenges that people with schizophrenia experience in functioning including problems with cognition, social cognition and negative symptoms together with other social determinants of health7. Maintenance antipsychotic treatment ought not be one of them. If there are concerns that medications are contributing to difficulties in functioning, then patients and their psychiatrists can work together to make necessary medication changes.Those who take ongoing antipsychotic medication have worse long-term outcomes than those who do not. On the contrary, the most recent Cochrane systematic review and meta-analysis of maintenance antipsychotic treatment in schizophrenia found that it not only prevented relapses and rehospitalizations but was also associated with improved functioning and quality of life8. As Moncrieff and Horowitz acknowledge, the naturalistic studies that have described better functioning in those who have discontinued antipsychotic medication are confounded by severity and indication, as well as by reverse causation (i.e. those with more severe and debilitating illness are more likely to continue antipsychotic treatment)9.Many individuals will not relapse when antipsychotics are discontinued. Moncrieff and Horowitz suggest that naturalistic studies, “provide confirmatory evidence that some patients can remain well without ongoing antipsychotic medication” and that up to 40% of people do not relapse after antipsychotic discontinuation in randomized controlled trials10. As we have previously described in our systematic review of relapse risk following a first episode of non-affective psychosis, the 40% estimate is limited by the inclusion of studies dating back as far as 1959 when very different diagnostic systems and clinical treatment practices were in use11. The three-year risk of relapse following a first episode of non-affective psychosis in the two open non-randomized antipsychotic discontinuation studies included in our review, published in 2001 and 2012, were estimated at 98% and 97%, respectively11. Both studies involved discontinuation of long-acting injectable antipsychotics, which Moncrieff and Horowitz suggest may be associated with a lower risk of relapse. Lappin et el., in their 10-year follow-up of individuals treated for a first episode of psychosis, reported that only 6% of those with a presenting diagnosis of schizophrenia achieved an Early Sustained Recovery defined as having achieved remission within 6 months of treatment and no further psychotic episodes over 10-year follow-up12. Studies reporting that a significant number of patients diagnosed with schizophrenia successfully discontinue antipsychotic medications have for the most part used the presenting diagnosis rather the diagnosis at the time medications are discontinued3. When individuals with an established diagnosis of schizophrenia discontinue antipsychotic medication, it is rare that they will remain well3.Prolonged antipsychotic treatment contributes to the high risk of psychotic relapse. Moncrieff and Horowitz assert that increasing the duration of treatment increases the risk of relapse when medications are discontinued. Such a relationship has never been established. While Tiihonen et al. reported that the risk of treatment failure (psychiatric hospitalization or death) increased with the duration of antipsychotic treatment before discontinuation13, they were careful to point out that the direction of causation is uncertain. They used prospectively collected nationwide register data to investigate the question of how long antipsychotic medication should be continued following a first episode of schizophrenia. In particular, they were interested in determining whether there was a duration of treatment beyond which the risk of psychotic relapse would be reduced. Their conclusion was that once antipsychotic treatment is initiated for a first episode of schizophrenia, there is no duration of treatment beyond which it is safe to discontinue antipsychotic medication at least during the first eight years, casting doubt on those treatment guidelines that suggest that discontinuation after one to two years may be considered. The authors suggest that the high risk of treatment failure seen in those who were treated the longest with antipsychotic medication is likely explained by greater illness severity rather than D2 receptor supersensitivity. As individuals in this study had not been randomly allocated to different durations of antipsychotic treatment, one cannot conclude that longer durations of antipsychotic treatment led to higher rates of relapse.Rates of relapse may be exaggerated by the adverse effects of the withdrawal process itself. Moncrieff and Horowitz propose that antipsychotic exposure may lower the threshold for psychotic relapse after discontinuation. Psychotic relapses occur over a wide range of times following discontinuation and sometimes occur years later, hardly a timeframe one would expect with withdrawal. Many individuals with schizophrenia are known to frequently miss multiple consecutive daily doses of medication and they do not typically go into a rapid withdrawal state14. The findings of the HAMLETT study, which reported that patients treated with partial agonists showed half the risk of relapse of those treated with high affinity D2 antagonists, is cited as evidence that dopaminergic hypersensitivity plays a key role in determining relapse15. Interpretation of the HAMLETT results are confounded by both indication and severity as participants had a range of psychotic diagnoses and were not randomly assigned to different antipsychotics. There is no conclusive evidence that psychotic relapses reflect a withdrawal phenomenon. Rather, relapses occur when treatment that has been very effective at controlling the symptoms of schizophrenia is discontinued or reduced to subtherapeutic levels.Psychotic relapses do not result in worse outcomes. Moncrieff and Horowitz cast doubt on the theory that psychosis itself is toxic to the brain, that antipsychotics thereby prevent disease progression, and that relapses in themselves lead to worse clinical outcomes. While there is much room for ongoing academic debate about the premises of this theory1, it does not follow that psychotic relapses do not lead to severe negative outcomes. We find it disconcerting that Moncrieff and Horowitz would state that, in the RADAR study, severe relapses requiring hospitalizations occurred in “only” 27.7% of those who stopped medication completely and in “only” 29.4% who reduced their dose by at least 50%16. A severe psychotic relapse requiring hospitalization is a sentinel event in any clinical trial. That psychotic relapses have not been definitively proven to impact on long-term outcome specifically is hardly reassuring. The relevant outcomes from a relapse cannot be limited to long-term measures of symptom severity and functioning. The traumatic journey through a psychotic episode is not eliminated if one is able to return to one’s pre-relapse level of symptoms and functioning. Psychotic relapses have profound and well-documented negative effects. They are associated with increased risks of homelessness, unemployment, violence, incarceration, impaired ability to live independently, reduced quality of life, and greater occupational dysfunction17,18.The risks of long-term treatment may outweigh the benefit for many. Maintenance treatment with antipsychotic medication has a profound and established impact on preventing psychotic relapses. It has been consistently associated with lower all-cause mortality compared to antipsychotic discontinuation9. Our view is that psychiatrists must a take an assertive stance in educating their patients about the risks of relapse and untreated psychosis. Relapses can lead to severe and catastrophic consequences. It is critical that patients understand the importance of ongoing treatment to prevent these outcomes. Shared decision-making should support patients to make the decisions that are in their best interest; some may well decide against long-term antipsychotic treatment based on their personal balance of risks and benefits. Given the extremely high risk of psychotic relapse in schizophrenia and the severe sequalae that often result, the mental health and physical health benefits of long-term antipsychotic treatment should be expected to outweigh the risks for the large majority of people with schizophrenia9,13.Gradual hyperbolic tapering of antipsychotics will allow many to remain off of antipsychotic medication. While accepting that antipsychotic medication can bring about remission of the psychotic symptoms of schizophrenia, Moncrieff and Horowitz would have us believe that, were it not for the D2 receptor supersensitive caused by these medications, most people with schizophrenia could remain free of psychosis off of antipsychotic medication. In this view, it is the antipsychotic medication that have transformed what should be an acute illness into a chronic one. How ironic that the medications that now allow people with schizophrenia to live lives that are not overwhelmed by psychotic symptoms are cast as the iatrogenic cause of chronic illness and disability. Their conclusion that continuing medication for one or two years after remission, as many clinical guidelines recommend, may increase the risk of withdrawal when the medication is discontinued, is not justified. Likewise, their proposal that cautious dose reduction could be appropriate in most patients who demonstrate symptom remission from a first episode of psychosis or following multiple episodes has no foundation. On the contrary, it may well be that those who have the most robust response to antipsychotic medication are most likely to relapse when they are discontinued19. Antipsychotics cause a high burden of debilitating and harmful side effects. While the adverse effects of antipsychotics have been well established, we do not agree that they need be debilitating. There are undoubtedly individuals who feel that antipsychotic medication makes them, “feel ‘braindead’, ‘like a zombie’ or of creating a ‘living hell’”. These severe adverse experiences may be due to pseudoparkinsonism, neuroleptic dysphoria or akathisia. It is unreasonable to expect patients to accept these side effects for the long-term. Instead, they should be and can be managed effectively in the vast majority of cases through treatment optimization, including reducing the dose or switching medications. Antipsychotics cause a high burden of debilitating and harmful side effects. While the adverse effects of antipsychotics have been well established, we do not agree that they need be debilitating. There are undoubtedly individuals who feel that antipsychotic medication makes them, “feel ‘braindead’, ‘like a zombie’ or of creating a ‘living hell’”. These severe adverse experiences may be due to pseudoparkinsonism, neuroleptic dysphoria or akathisia. It is unreasonable to expect patients to accept these side effects for the long-term. Instead, they should be and can be managed effectively in the vast majority of cases through treatment optimization, including reducing the dose or switching medications. Antipsychotic treatment interferes with achieving other aspects of recovery that are highly valued, such as functional capacity and the fulfillment of personally meaningful goals. There is no robust evidence to support the inference that the protection from relapse associated with maintenance antipsychotic medication comes at the expense of other functional aspects of recovery. Psychiatrists are not alone in focusing on the importance of preventing recurrence of psychotic symptoms; achieving control of their psychotic symptoms is usually among the highest priorities for people with schizophrenia, ranking higher than work and eliminating side effects6. There are many factors that contribute to the challenges that people with schizophrenia experience in functioning including problems with cognition, social cognition and negative symptoms together with other social determinants of health7. Maintenance antipsychotic treatment ought not be one of them. If there are concerns that medications are contributing to difficulties in functioning, then patients and their psychiatrists can work together to make necessary medication changes. Antipsychotic treatment interferes with achieving other aspects of recovery that are highly valued, such as functional capacity and the fulfillment of personally meaningful goals. There is no robust evidence to support the inference that the protection from relapse associated with maintenance antipsychotic medication comes at the expense of other functional aspects of recovery. Psychiatrists are not alone in focusing on the importance of preventing recurrence of psychotic symptoms; achieving control of their psychotic symptoms is usually among the highest priorities for people with schizophrenia, ranking higher than work and eliminating side effects6. There are many factors that contribute to the challenges that people with schizophrenia experience in functioning including problems with cognition, social cognition and negative symptoms together with other social determinants of health7. Maintenance antipsychotic treatment ought not be one of them. If there are concerns that medications are contributing to difficulties in functioning, then patients and their psychiatrists can work together to make necessary medication changes. Those who take ongoing antipsychotic medication have worse long-term outcomes than those who do not. On the contrary, the most recent Cochrane systematic review and meta-analysis of maintenance antipsychotic treatment in schizophrenia found that it not only prevented relapses and rehospitalizations but was also associated with improved functioning and quality of life8. As Moncrieff and Horowitz acknowledge, the naturalistic studies that have described better functioning in those who have discontinued antipsychotic medication are confounded by severity and indication, as well as by reverse causation (i.e. those with more severe and debilitating illness are more likely to continue antipsychotic treatment)9. Those who take ongoing antipsychotic medication have worse long-term outcomes than those who do not. On the contrary, the most recent Cochrane systematic review and meta-analysis of maintenance antipsychotic treatment in schizophrenia found that it not only prevented relapses and rehospitalizations but was also associated with improved functioning and quality of life8. As Moncrieff and Horowitz acknowledge, the naturalistic studies that have described better functioning in those who have discontinued antipsychotic medication are confounded by severity and indication, as well as by reverse causation (i.e. those with more severe and debilitating illness are more likely to continue antipsychotic treatment)9. Many individuals will not relapse when antipsychotics are discontinued. Moncrieff and Horowitz suggest that naturalistic studies, “provide confirmatory evidence that some patients can remain well without ongoing antipsychotic medication” and that up to 40% of people do not relapse after antipsychotic discontinuation in randomized controlled trials10. As we have previously described in our systematic review of relapse risk following a first episode of non-affective psychosis, the 40% estimate is limited by the inclusion of studies dating back as far as 1959 when very different diagnostic systems and clinical treatment practices were in use11. The three-year risk of relapse following a first episode of non-affective psychosis in the two open non-randomized antipsychotic discontinuation studies included in our review, published in 2001 and 2012, were estimated at 98% and 97%, respectively11. Both studies involved discontinuation of long-acting injectable antipsychotics, which Moncrieff and Horowitz suggest may be associated with a lower risk of relapse. Lappin et el., in their 10-year follow-up of individuals treated for a first episode of psychosis, reported that only 6% of those with a presenting diagnosis of schizophrenia achieved an Early Sustained Recovery defined as having achieved remission within 6 months of treatment and no further psychotic episodes over 10-year follow-up12. Studies reporting that a significant number of patients diagnosed with schizophrenia successfully discontinue antipsychotic medications have for the most part used the presenting diagnosis rather the diagnosis at the time medications are discontinued3. When individuals with an established diagnosis of schizophrenia discontinue antipsychotic medication, it is rare that they will remain well3. Many individuals will not relapse when antipsychotics are discontinued. Moncrieff and Horowitz suggest that naturalistic studies, “provide confirmatory evidence that some patients can remain well without ongoing antipsychotic medication” and that up to 40% of people do not relapse after antipsychotic discontinuation in randomized controlled trials10. As we have previously described in our systematic review of relapse risk following a first episode of non-affective psychosis, the 40% estimate is limited by the inclusion of studies dating back as far as 1959 when very different diagnostic systems and clinical treatment practices were in use11. The three-year risk of relapse following a first episode of non-affective psychosis in the two open non-randomized antipsychotic discontinuation studies included in our review, published in 2001 and 2012, were estimated at 98% and 97%, respectively11. Both studies involved discontinuation of long-acting injectable antipsychotics, which Moncrieff and Horowitz suggest may be associated with a lower risk of relapse. Lappin et el., in their 10-year follow-up of individuals treated for a first episode of psychosis, reported that only 6% of those with a presenting diagnosis of schizophrenia achieved an Early Sustained Recovery defined as having achieved remission within 6 months of treatment and no further psychotic episodes over 10-year follow-up12. Studies reporting that a significant number of patients diagnosed with schizophrenia successfully discontinue antipsychotic medications have for the most part used the presenting diagnosis rather the diagnosis at the time medications are discontinued3. When individuals with an established diagnosis of schizophrenia discontinue antipsychotic medication, it is rare that they will remain well3. Prolonged antipsychotic treatment contributes to the high risk of psychotic relapse. Moncrieff and Horowitz assert that increasing the duration of treatment increases the risk of relapse when medications are discontinued. Such a relationship has never been established. While Tiihonen et al. reported that the risk of treatment failure (psychiatric hospitalization or death) increased with the duration of antipsychotic treatment before discontinuation13, they were careful to point out that the direction of causation is uncertain. They used prospectively collected nationwide register data to investigate the question of how long antipsychotic medication should be continued following a first episode of schizophrenia. In particular, they were interested in determining whether there was a duration of treatment beyond which the risk of psychotic relapse would be reduced. Their conclusion was that once antipsychotic treatment is initiated for a first episode of schizophrenia, there is no duration of treatment beyond which it is safe to discontinue antipsychotic medication at least during the first eight years, casting doubt on those treatment guidelines that suggest that discontinuation after one to two years may be considered. The authors suggest that the high risk of treatment failure seen in those who were treated the longest with antipsychotic medication is likely explained by greater illness severity rather than D2 receptor supersensitivity. As individuals in this study had not been randomly allocated to different durations of antipsychotic treatment, one cannot conclude that longer durations of antipsychotic treatment led to higher rates of relapse. Prolonged antipsychotic treatment contributes to the high risk of psychotic relapse. Moncrieff and Horowitz assert that increasing the duration of treatment increases the risk of relapse when medications are discontinued. Such a relationship has never been established. While Tiihonen et al. reported that the risk of treatment failure (psychiatric hospitalization or death) increased with the duration of antipsychotic treatment before discontinuation13, they were careful to point out that the direction of causation is uncertain. They used prospectively collected nationwide register data to investigate the question of how long antipsychotic medication should be continued following a first episode of schizophrenia. In particular, they were interested in determining whether there was a duration of treatment beyond which the risk of psychotic relapse would be reduced. Their conclusion was that once antipsychotic treatment is initiated for a first episode of schizophrenia, there is no duration of treatment beyond which it is safe to discontinue antipsychotic medication at least during the first eight years, casting doubt on those treatment guidelines that suggest that discontinuation after one to two years may be considered. The authors suggest that the high risk of treatment failure seen in those who were treated the longest with antipsychotic medication is likely explained by greater illness severity rather than D2 receptor supersensitivity. As individuals in this study had not been randomly allocated to different durations of antipsychotic treatment, one cannot conclude that longer durations of antipsychotic treatment led to higher rates of relapse. Rates of relapse may be exaggerated by the adverse effects of the withdrawal process itself. Moncrieff and Horowitz propose that antipsychotic exposure may lower the threshold for psychotic relapse after discontinuation. Psychotic relapses occur over a wide range of times following discontinuation and sometimes occur years later, hardly a timeframe one would expect with withdrawal. Many individuals with schizophrenia are known to frequently miss multiple consecutive daily doses of medication and they do not typically go into a rapid withdrawal state14. The findings of the HAMLETT study, which reported that patients treated with partial agonists showed half the risk of relapse of those treated with high affinity D2 antagonists, is cited as evidence that dopaminergic hypersensitivity plays a key role in determining relapse15. Interpretation of the HAMLETT results are confounded by both indication and severity as participants had a range of psychotic diagnoses and were not randomly assigned to different antipsychotics. There is no conclusive evidence that psychotic relapses reflect a withdrawal phenomenon. Rather, relapses occur when treatment that has been very effective at controlling the symptoms of schizophrenia is discontinued or reduced to subtherapeutic levels. Rates of relapse may be exaggerated by the adverse effects of the withdrawal process itself. Moncrieff and Horowitz propose that antipsychotic exposure may lower the threshold for psychotic relapse after discontinuation. Psychotic relapses occur over a wide range of times following discontinuation and sometimes occur years later, hardly a timeframe one would expect with withdrawal. Many individuals with schizophrenia are known to frequently miss multiple consecutive daily doses of medication and they do not typically go into a rapid withdrawal state14. The findings of the HAMLETT study, which reported that patients treated with partial agonists showed half the risk of relapse of those treated with high affinity D2 antagonists, is cited as evidence that dopaminergic hypersensitivity plays a key role in determining relapse15. Interpretation of the HAMLETT results are confounded by both indication and severity as participants had a range of psychotic diagnoses and were not randomly assigned to different antipsychotics. There is no conclusive evidence that psychotic relapses reflect a withdrawal phenomenon. Rather, relapses occur when treatment that has been very effective at controlling the symptoms of schizophrenia is discontinued or reduced to subtherapeutic levels. Psychotic relapses do not result in worse outcomes. Moncrieff and Horowitz cast doubt on the theory that psychosis itself is toxic to the brain, that antipsychotics thereby prevent disease progression, and that relapses in themselves lead to worse clinical outcomes. While there is much room for ongoing academic debate about the premises of this theory1, it does not follow that psychotic relapses do not lead to severe negative outcomes. We find it disconcerting that Moncrieff and Horowitz would state that, in the RADAR study, severe relapses requiring hospitalizations occurred in “only” 27.7% of those who stopped medication completely and in “only” 29.4% who reduced their dose by at least 50%16. A severe psychotic relapse requiring hospitalization is a sentinel event in any clinical trial. That psychotic relapses have not been definitively proven to impact on long-term outcome specifically is hardly reassuring. The relevant outcomes from a relapse cannot be limited to long-term measures of symptom severity and functioning. The traumatic journey through a psychotic episode is not eliminated if one is able to return to one’s pre-relapse level of symptoms and functioning. Psychotic relapses have profound and well-documented negative effects. They are associated with increased risks of homelessness, unemployment, violence, incarceration, impaired ability to live independently, reduced quality of life, and greater occupational dysfunction17,18. Psychotic relapses do not result in worse outcomes. Moncrieff and Horowitz cast doubt on the theory that psychosis itself is toxic to the brain, that antipsychotics thereby prevent disease progression, and that relapses in themselves lead to worse clinical outcomes. While there is much room for ongoing academic debate about the premises of this theory1, it does not follow that psychotic relapses do not lead to severe negative outcomes. We find it disconcerting that Moncrieff and Horowitz would state that, in the RADAR study, severe relapses requiring hospitalizations occurred in “only” 27.7% of those who stopped medication completely and in “only” 29.4% who reduced their dose by at least 50%16. A severe psychotic relapse requiring hospitalization is a sentinel event in any clinical trial. That psychotic relapses have not been definitively proven to impact on long-term outcome specifically is hardly reassuring. The relevant outcomes from a relapse cannot be limited to long-term measures of symptom severity and functioning. The traumatic journey through a psychotic episode is not eliminated if one is able to return to one’s pre-relapse level of symptoms and functioning. Psychotic relapses have profound and well-documented negative effects. They are associated with increased risks of homelessness, unemployment, violence, incarceration, impaired ability to live independently, reduced quality of life, and greater occupational dysfunction17,18. The risks of long-term treatment may outweigh the benefit for many. Maintenance treatment with antipsychotic medication has a profound and established impact on preventing psychotic relapses. It has been consistently associated with lower all-cause mortality compared to antipsychotic discontinuation9. Our view is that psychiatrists must a take an assertive stance in educating their patients about the risks of relapse and untreated psychosis. Relapses can lead to severe and catastrophic consequences. It is critical that patients understand the importance of ongoing treatment to prevent these outcomes. Shared decision-making should support patients to make the decisions that are in their best interest; some may well decide against long-term antipsychotic treatment based on their personal balance of risks and benefits. Given the extremely high risk of psychotic relapse in schizophrenia and the severe sequalae that often result, the mental health and physical health benefits of long-term antipsychotic treatment should be expected to outweigh the risks for the large majority of people with schizophrenia9,13. The risks of long-term treatment may outweigh the benefit for many. Maintenance treatment with antipsychotic medication has a profound and established impact on preventing psychotic relapses. It has been consistently associated with lower all-cause mortality compared to antipsychotic discontinuation9. Our view is that psychiatrists must a take an assertive stance in educating their patients about the risks of relapse and untreated psychosis. Relapses can lead to severe and catastrophic consequences. It is critical that patients understand the importance of ongoing treatment to prevent these outcomes. Shared decision-making should support patients to make the decisions that are in their best interest; some may well decide against long-term antipsychotic treatment based on their personal balance of risks and benefits. Given the extremely high risk of psychotic relapse in schizophrenia and the severe sequalae that often result, the mental health and physical health benefits of long-term antipsychotic treatment should be expected to outweigh the risks for the large majority of people with schizophrenia9,13. Gradual hyperbolic tapering of antipsychotics will allow many to remain off of antipsychotic medication. While accepting that antipsychotic medication can bring about remission of the psychotic symptoms of schizophrenia, Moncrieff and Horowitz would have us believe that, were it not for the D2 receptor supersensitive caused by these medications, most people with schizophrenia could remain free of psychosis off of antipsychotic medication. In this view, it is the antipsychotic medication that have transformed what should be an acute illness into a chronic one. How ironic that the medications that now allow people with schizophrenia to live lives that are not overwhelmed by psychotic symptoms are cast as the iatrogenic cause of chronic illness and disability. Their conclusion that continuing medication for one or two years after remission, as many clinical guidelines recommend, may increase the risk of withdrawal when the medication is discontinued, is not justified. Likewise, their proposal that cautious dose reduction could be appropriate in most patients who demonstrate symptom remission from a first episode of psychosis or following multiple episodes has no foundation. On the contrary, it may well be that those who have the most robust response to antipsychotic medication are most likely to relapse when they are discontinued19. Gradual hyperbolic tapering of antipsychotics will allow many to remain off of antipsychotic medication. While accepting that antipsychotic medication can bring about remission of the psychotic symptoms of schizophrenia, Moncrieff and Horowitz would have us believe that, were it not for the D2 receptor supersensitive caused by these medications, most people with schizophrenia could remain free of psychosis off of antipsychotic medication. In this view, it is the antipsychotic medication that have transformed what should be an acute illness into a chronic one. How ironic that the medications that now allow people with schizophrenia to live lives that are not overwhelmed by psychotic symptoms are cast as the iatrogenic cause of chronic illness and disability. Their conclusion that continuing medication for one or two years after remission, as many clinical guidelines recommend, may increase the risk of withdrawal when the medication is discontinued, is not justified. Likewise, their proposal that cautious dose reduction could be appropriate in most patients who demonstrate symptom remission from a first episode of psychosis or following multiple episodes has no foundation. On the contrary, it may well be that those who have the most robust response to antipsychotic medication are most likely to relapse when they are discontinued19. Antipsychotic dose reduction Dose reduction is conceptualized as the means to the end of antipsychotic discontinuation by Moncrieff and Horowitz. Many individuals with schizophrenia that we see are interested in the idea of reducing the dose, even in absence of troublesome dose-related side effects, as a step towards discontinuation. For others, wishing to reduce their dose may be viewed as an affirmation that they are now much improved and no longer need as much medication as was required when they were unwell. If the goal of discontinuation is not achievable for the overwhelming majority of individuals with schizophrenia, then what is the goal of dose reduction? The concept of antipsychotic treatment optimization needs to be distinguished from antipsychotic dose reduction. Optimizing antipsychotic treatment is a multifaceted process involving various strategies, including adjusting dose, switching to a different antipsychotic, introducing medications to address side effects and addressing polypharmacy. The goal of optimizing treatment is to ensure effective symptom management while minimizing side effects. This is a central responsibility of clinicians. In contrast, dose reduction as a standalone strategy is a high-risk approach that can lead to relapse. Certainly, many individuals are treated for their acute episodes of psychotic symptoms with doses of antipsychotic medication that at steady state result in side effects that are not tolerable on an ongoing basis. We commonly see individuals discharged from inpatient care with intolerable akathisia, parkinsonism, sexual dysfunction and sedation. These side effects can usually be resolved by dose adjustment though they sometimes require prescription of additional medications in the short-term to ensure they are resolved as quickly as possible. It is less clear whether weight gain and adverse metabolic effects are dose-related at therapeutic doses20. The extent to which dose reduction is appropriate has already shifted considerably with the introduction of second and third generation antipsychotics. First generation antipsychotic medication were prescribed at doses as much as an order of magnitude higher than is now considered appropriate. As recently as the 1990s, standard treatment for a first episode of schizophrenia involved recommended doses of haloperidol in the range of 20-40 mg/day21. The gold standard dose of haloperidol for clinical trials investigating the efficacy of second generation antipsychotics was 20 mg/day, or 1000 mg chlorpromazine equivalents/day. Our group and others have demonstrated that patients with a first episode of schizophrenia are likely to respond robustly to doses of antipsychotic medication equivalent to haloperidol 2-4 mg/day22,23 or doses of second generation antipsychotics that result in levels of D2 receptor occupancy of 65% or greater24. Similar doses are now considered to be effective for clinical response for multi-episode patients who do not have treatment resistant schizophrenia. Given the shift towards using the lowest effective dose for acute treatment, the potential for making substantial reductions in antipsychotic dose for maintenance treatment has diminished. Moncrieff and Horowitz make the case that very gradual tapering using a hyperbolic strategy is most prudent and may facilitate successful reductions in maintenance doses and antipsychotic discontinuation for some. The difficulty that arises with this approach is that there is often an extended delay between achieving a new steady state antipsychotic level and relapse. Discontinuation studies suggest that approximately 80% of relapses will occur within the first year and most of the remainder within the second year11. We would expect a similar timeframe for relapses that occur when the dose is reduced to levels that are no longer effective for maintenance treatment. As a result, it becomes very challenging to know when dose reduction has been successful and the next step down in dose can be initiated. Using a hyperbolic tapering approach might well be of value in helping patients to reach a dose that is tolerable for them. For many antipsychotic medications, dose adjustment can also be carried out in conjunction with therapeutic drug monitoring of antipsychotic plasma levels to minimize the likelihood that the dose has been reduced to a level that is ineffective for relapse prevention. We agree that ongoing efforts to minimize the adverse effects of antipsychotic medications including weight gain and metabolic side effects, tardive dyskinesia and sexual side effects continue to be an important priority. However, evidence to date suggests that antipsychotic discontinuation and intermittent treatment strategies are associated with a high risk of relapse including severe relapses leading to hospitalization25. More recently, and drawing upon other lines of evidence such as neuroimaging, we have proposed “extended antipsychotic dosing”, a strategy that involves intermittent, but regular, antipsychotic exposure26. This work is premised on the finding that continuous occupancy of dopamine D2 receptors is unlikely to be required for ongoing antipsychotic efficacy. Reducing dosing frequency may be a more promising strategy for reducing antipsychotic side effects than reducing dose. To date, this work has largely centered on alternate day dosing, which affords a 50% reduction in antipsychotic dose. Results have been favourable and support studies that are currently underway27. In conclusion, we have argued that the rationale for discontinuing antipsychotic medication in people with schizophrenia is flawed at multiple levels. “Helping” people with schizophrenia to discontinue antipsychotics has the potential to cause a great deal of harm with little chance of significant benefit. There is an overwhelming body of evidence supporting the need for continued antipsychotic treatment in schizophrenia. The most responsible course of action based on evidence to date is to maintain antipsychotic treatment at the lowest effective dose to maximize the individual patient’s long-term stability, functioning, and quality of life.