## Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

# Is the ongoing use of placebo in relapse-prevention clinical trials in schizophrenia justified?

Schizophrenia Research, 2-3, 150, pages 427 - 433

### Abstract

#### Background

Placebo-controlled randomised controlled trials (RCTs) continue to be required or recommended by regulatory authorities for the licensing of new drugs for schizophrenia, despite ongoing concerns regarding the risks to trial participants.

#### Methods

In this article we consider the scientific and ethical pros and cons associated with use of placebo in RCTs in schizophrenia, systematically review the published relapse-prevention placebo-controlled RCTs with second generation antipsychotics (SGAs) in schizophrenia, and examine the risks associated with these trials.

#### Results

We identified 12 studies involving 2842 participants of which 968 received placebo. Relapse rates were 56% for placebo and 17.4% for active treatment groups. There is a lack of well-designed longitudinal studies investigating the psychosocial and biological consequences of exposure to placebo, to treatment discontinuation and to relapse in schizophrenia.

#### Conclusion

In the absence of such studies it is risky to assume that patients are not at risk of significant distress and long-term harm, and therefore it is difficult to justify the ongoing use of placebo in relapse-prevention RCTs in schizophrenia.

Keywords: Placebo, Antipsychotic, Relapse-prevention, Maintenance treatment, Schizophrenia.

### 1. Introduction

Antipsychotics are the mainstay of treatment for schizophrenia. Their efficacy and safety for acute and maintenance treatment have been established on the basis of extensive clinical development programs, of which an integral component is the placebo-controlled, randomised, controlled trial (RCT). In maintenance treatment RCTs, the superiority of antipsychotic medications compared to placebo has been clearly demonstrated ( Leucht et al., 2012 ) and relapse-prevention by means of continuous treatment is considered a major treatment goal ( Kane, 2007 ). Despite this, in real-world clinical settings discontinuation of antipsychotic treatment is common, with most patients typically experiencing multiple relapses during the course of their illness ( Robinson et al., 1999 ). In clinical settings the decision to discontinue treatment is largely patient driven, and usually against medical advice ( Perkins et al., 2008 ). Another setting where active treatment is discontinued or withheld is the placebo-controlled RCT. The difference in this case is that discontinuation or withholding of active treatment is pre-planned, with the full participation of clinicians. Given the high risk of relapse associated with antipsychotic treatment discontinuation ( Gilbert et al., 1995 ), it is not surprising that concerns have long been voiced regarding the risks to participants in placebo-controlled RCTs in schizophrenia ( Weijer, 1999 ) and the debate over their continued use when proven treatments exist continues to be lively ( Kim, 2003 ). There is a tension between the need for scientifically valid trials of new psychotropic drugs and concern about the risks associated with conducting placebo-controlled RCTs, when this requires that some patients be denied effective therapy ( Laughren, 2001 ). Consequently, many experienced schizophrenia researchers are unwilling to take part in placebo-controlled RCTs, with many citing the attitude of local ethics committees as the reason for their reluctance ( Fleischhacker and Burns, 2002 ). Risk to trial participants is likely to be greatest in maintenance treatment, or relapse-prevention RCTs where patients, once stabilised, are switched to placebo sometimes for considerable periods, and substantial numbers need to experience a relapse event before a treatment effect can be statistically demonstrated. Despite these concerns, current practice permits the use of placebos in such settings, as long as the benefits are considered to outweigh the risks and burdens. In this critical review we 1) briefly summarise the ethical and scientific pros and cons of placebo in relapse-prevention RCTs in schizophrenia, 2) systematically review the published relapse-prevention RCTs with second-generation antipsychotics (SGAs) in schizophrenia and 3) examine the risks of harm associated with such trials.

### 2. The pros and cons of use of placebo in RCTs

According to the World Medical Association's Declaration of Helsinki, the use of placebo is acceptable in studies where there is no proven intervention, or where compelling and scientifically sound methodological reasons exist for the use of placebo to determine efficacy or safety of an intervention, and where patients who receive placebo will not be exposed to any risk of serious or irreversible harm. The Declaration emphasises the need for extreme care to be taken to avoid abuse of this option, and that the interest of science and society should never take precedence over considerations related to the well-being of the individual patient ( World Medical Association, WMA, 2008 ). Specific criteria are proposed for judging the ethical acceptability of including placebo controls in a RCT, including the likelihood that the experimental intervention will have clinically significant advantages over existing treatments, the existence of compelling reasons for placebo use, a subject selection procedure that minimizes the risk of serious adverse consequences, and a risk-versus-benefit analysis that favours the advantages from placebo use over the risks to subjects ( Carpenter et al., 2003 ).

#### 2.1. The case for placebo

A comparison between the test drug and placebo provides the most powerful method of establishing efficacy. It is impossible to determine in an active control group with non-inferiority design whether both treatments were effective or ineffective. Some regard the use of placebos in RCTs in schizophrenia as an essential component of a comprehensive drug evaluation for new antipsychotic medications, and to be both ethically and scientifically justified ( Addington, 1995 ). Indeed, placebo-controlled relapse prevention studies continue to be required by regulatory authorities in the United States and Canada, and strongly recommended in the European Union, for the licensing of new drugs (). These agencies consider that it is not possible to conduct a valid evaluation of a treatment for schizophrenia without placebo-controlled studies ( Correll et al., 2011 ). The ongoing use of placebo has been justified on the basis of there being no clear evidence of increased risk of persistent morbidity or mortality, and because alternative study designs may not be as good at demonstrating efficacy and tolerability ( Beasley et al., 2003 ). It has been argued that the fundamental reason to favour a test to demonstrate superiority of an active intervention over placebo rather than active control is that the response of a specific population to placebo or the active intervention can be highly variable across studies. Therefore, a non-inferiority design cannot be reliably interpreted ( Beasley et al., 2003 ). A trend towards increasing placebo effects and a decrease in the drug effect size has been reported in RCTs comparing both investigational and approved antipsychotics with placebo in RCTs submitted for new drug applications ( Kemp et al., 2010 ). The argument for continued use of placebo in RCTs is therefore that there is a need to establish assay sensitivity, and there is a moral imperative to guard against ineffectual treatments being approved for use in clinical practice. In one study, the choice of placebo as comparator was considered to be safe and ethical based on three premises: 1) Available empirical evidence at that time suggested no increased risk of severe suffering or long-term morbidity following exposure to placebo; 2) the belief that clinical measures put in place (frequent and careful monitoring for early indicators of worsening and the use of sensitive relapse criteria) would effectively detect early symptoms and prevent serious relapse, and 3) the likelihood that fewer relapses would be necessary to detect a positive outcome with placebo rather than with an active comparator ( Beasley et al., 2003 ).

The European Medicines Agency recently introduced a Guideline on Clinical Investigation of Medicinal Products in the Treatment of Schizophrenia. The document recognises the need for placebo to ensure assay sensitivity, even in well designed and conducted RCTs, but emphasises that these studies need to be conducted in highly controlled settings, with appropriate safeguards. In this context it is considered that the benefits of a placebo arm will generally override ethical reservations in short-term trials. However, long term exposure to placebo is not only ethically problematic but also scientifically unsound due to high rates of withdrawals which make interpretation of the data difficult. Yet, it is stated that, for demonstrating maintenance of effectiveness of treatment over time the inclusion of a placebo arm is possible and appropriate in a randomised withdrawal study as long as it is appropriately designed and conducted. Patients who relapse should receive immediate active treatment, and there should therefore not be ethical problems ( European Medicines Agency Committee for Medicinal Products for Human Use, 2012 ).

A draft guidance addressing enrichment strategies for clinical trials of the United States Food and Drug Administration describes the randomised withdrawal design as a way to establish long-term effectiveness of drugs when protracted use of a placebo would not be acceptable. In this design the study population is on active treatment for an extended period and those who respond enter a blinded, randomised treatment withdrawal phase for a short duration. Patients are withdrawn from the study in the case of symptom recurrence, thereby minimising exposure to placebo treatment ( U.S. Food and Drug Administration Center for Drug Evaluation and Research, 2012 ).

#### 2.2. The case against placebo

On the other hand, there is a concern that, particularly in relapse-prevention RCTs, exposure to placebo is associated with a risk of undue suffering or harm. Inclusion of a placebo-arm appears to be in conflict with clinical equipoise, considered to be the moral foundation of the RCT, which requires the use of best available treatment as the control in RCT. This is consistent with the principle of beneficence which requires that a physician should act in the best interest of each patient. Furthermore, scientific criticisms of the use of an active control with a non-inferiority study design may not present an insurmountable barrier to their use as an alternative to placebo-controlled RCTs (). Meta-analyses indicate that a therapeutic dose of SGA is very likely to be statistically superior to placebo in an adequate trial, and that despite an increasing effect, the average improvement of schizophrenia symptoms in a placebo arm will be small. There are few efficacy differences between SGAs, and expected differences are in their safety profile or their influence on quality of life. Therefore, hypothesis testing is often limited to the problem of confirming that the new drug is not inferior to a comparator antipsychotic with respect to its efficacy ( Fleischhacker et al., 2003 ). An additional point is that high dropout rates have been reported in clinical trials utilizing placebo controls ( Kemmler et al., 2005 ), thereby reducing the statistical power of these studies.

A final point to consider concerns our ability to recognise early signs of recurrence, and the effectiveness of rescue interventions in preventing serious relapse. Such a strategy would comprise frequent monitoring and careful assessment of patients for early warning signs of relapse. While success has been reported in identifying early signs of relapse by means of an instrument specifically designed to detect early warning signs a specifically designed scale ( Birchwood et al., 1989 ), this may not always be the case. Other studies suggest that early warning signs are relatively unreliable predictors of relapse () and it has been reported that in many cases recurrence symptoms do not return gradually, but rather abruptly, with rapid return to levels of previous psychotic episodes (). These findings are consistent with those of a previous study in which it was found that early rescue medication strategies may not effectively prevent relapse in placebo-treated patients ( Schooler et al., 1997 ).

### 3. Systematic review

A recent Cochrane report ( Leucht et al., 2012 ) comprehensively reviewed all RCTs comparing maintenance treatment and placebo. They identified 65 studies published between 1959 and 2011. The majority of studies were conducted with first-generation antipsychotics, often with small samples and poorly reported randomisation, allocation and blinding procedures. Also, the review focused on the benefits and burdens of maintenance antipsychotic treatment rather than the effects of exposure to placebo. We therefore decided to conduct a review of more recent relapse-prevention RCTs — i.e. those comparing SGAs with placebo, and to focus on the outcomes of the placebo-treated patients rather than the antipsychotic treated patients.

#### 3.1. Data extraction methods

To identify the placebo-controlled relapse-prevention RCTs with SGAs in schizophrenia we conducted Medline searches using the terms antipsychotic, placebo, schizophrenia, maintenance and relapse. We then searched Medline for the following SGAs individually, linking them with the search term “placebo”: clozapine, risperidone (oral and depot), olanzapine (oral and depot), quetiapine, ziprasidone, amisulpride, sertindole, aripiprazole (oral and depot), iloperidone, paliperidone (oral and depot), zotepine, bifeprunox, asenapine and lurasidone. We also conducted a search of Cochrane database studies of SGAs versus placebo and the ClinicalTrials.gov database, using the search terms “placebo” and “schizophrenia” combined with each of the above SGAs. References of the identified studies were cross-checked for any additional relevant studies. To identify studies investigating the consequences of placebo treatment, treatment discontinuation and of relapse, we conducted Medline searches using combinations of the following terms linked to schizophrenia, relapse and placebo: Psychosocial, cost, mortality, suicide, outcome, progressive, deterioration.

### 4. Relapse-prevention RCTS comparing second-generation antipsychotics to placebo

We identified 12 studies spanning from 1997 to 2012, involving nine of the 18 investigational drugs. A total of 2842 patients were randomised to the double-blinded arms in these studies, of whom 968 (34%) received placebo. Details of these studies are provided in Table 1 . The following points are of interest: 1) Several of the SGAs received approval for the maintenance treatment of schizophrenia without published placebo-controlled, relapse-prevention RCTs, most notably risperidone (both oral and long-acting injectable), amisulpride and olanzapine depot. 2) The duration of exposure to placebo was considerable, ranging from 26 weeks to 1 year, although 5 studies were terminated early once interim analyses revealed statistically significant group differences. 3) The mean relapse rate in the placebo-treated patients was high — 56%, as opposed to 17.4% in the active treatment arms. 4) Little information is available from these studies on the post-relapse outcome of these patients. Outcomes such as suicide, death and rehospitalisation were inconsistently reported.

Table 1 Published placebo-controlled relapse-prevention randomised controlled trials with second generation antipsychotics in schizophrenia.

Authors Active drug(s) Total sample N Placebo n Study duration Relapse rate active treatment Relapse rate placebo
Dellva et al. (1997) (study 1) Olanzapine 58 13 46 weeks 28.6% 69.9%
Dellva et al. (1997) (study 2) Olanapine 62 14 46 weeks 19.6% 45.5%
Cooper et al. (2000) Zotepine 121 40 26 weeks 8.7% 52.8%
Arato et al. (2002) Ziprasidone 278 71 1 year 38.0% 77.0%
Beasley et al. (2003) Olanzapine 326 102 52 weeks (terminated early) 5.5% 55.2%
Pigott et al. (2003) Aripiprazole 310 85 26 weeks 34.0% 57.0%
Peuskens et al. (2007) Extended release quetiapine fumarate 197 103 1 year (terminated early) 11.7% 48.5%
Kramer et al. (2007) Paliperidone extended release tablets 113 55 Variable, terminated early 25.0% 53.0%
Chen et al. (2010) Quetiapine 178 89 12 months 41.0% 79.0%
Hough et al. (2010) Paliperidone palmitate 410 204 Variable, terminated early 17.4% 47.5%
Kane et al. (2011) Asenapine 386 192 26 weeks 12.1% 47.4%
Kane et al. (2012) Aripiprazole IM depot 403 134 52 weeks (terminated early) 10.0% 39.6%
Total 2842 Total 968   Mean = 21% Mean = 56.0%

There were considerable differences in the study designs. Seven included an initial trial stabilisation period (), two were double-blind extensions of multicentre studies ( Dellva et al., 1997 [two studies]), one selected stable patients but did not have a trial stabilisation phase ( Arato et al., 2002 ), one did not specify selection of stable patients or have an initial stabilisation phase ( Cooper et al., 2000 ) and one selected stable but significantly symptomatic patients and did not have a stabilisation phase ( Pigott et al., 2003 ). All of the studies used relapse as the primary outcome measure, although this was determined in different ways including time to relapse (), time to symptom recurrence ( Cooper et al., 2000 ), time to hospitalisation for psychopathology ( Dellva et al., 1997 [two studies]), time to exacerbation of psychotic symptoms/impending relapse ( Kane et al., 2012 ), estimated probability of relapse at 12 m ( Arato et al., 2002 ) or at 6 months ( Beasley et al., 2003 ), and time to relapse/pending relapse ( Kane et al., 2011 ). Only one of the studies included a post-relapse treatment period ( Hough et al., 2010 ), which allowed a post-hoc assessment of the effect of relapse during placebo treatment on subsequent outcome ( Emsley et al., 2012a ). One study was unique in that it was conducted in a sample of first-episode patients. Interestingly, this study reported the highest relapse rate of 79% at 12 months in the placebo treated patients ( Chen et al., 2010 ). Five of the studies incorporated interim analyses and premature study termination after sufficient relapse events had occurred to demonstrate significant group differences ().

### 5. The consequences of withholding active treatment

The consequences of withholding active treatment in relapse-prevention RCTs are largely related to relapse. Most patients are likely to experience relapse if treatment is discontinued — a comprehensive review ( Gilbert et al., 1995 ) found a mean cumulative relapse rate of 52% (range 0%–100%) over a mean follow-up period of 6.3 months (range 0.5–24 m) for patients withdrawn from antipsychotics compared with 16% for those maintained on antipsychotic treatment. Many patients relapse shortly after active treatment discontinuation, while others remain well for variable periods. The risk of relapse therefore increases as the duration of exposure to non-treatment increases. Relapse rates of over 90% have been reported at 2 years after discontinuation, even after a single-psychotic episode (). In addition to being a frequent occurrence, relapse may have devastating consequences. While many psychosocial consequences are difficult to measure, possible complications include distress to patients and carers, the potential to derail hard-won progress in psychosocial recovery including disruption of relationships, education or employment, suicidal and homicidal behaviour, increased stigma and increased economic burden (). According to the results of an international survey conducted by the World Federation of Mental Health (2006) , caregivers who reported that their family member experienced relapse (N = 838) considered the consequences to be devastating. They reported the following consequences in patients: inability to work (72%); hospitalisation (69%); attempted suicide (22%); and imprisonment (20%). Caregivers reported disruption of their own lives substantially (61%); fearfulness of the patient's well-being (56%); worsening of their own mental health (54%); and worsening of their financial situation (26%).

#### 5.1. Cost of relapse

A study published seventeen years ago estimated the hospitalisation cost of relapse in schizophrenia in the United States by modelling national annual cost of rehospitalisation for multi-episode schizophrenia outpatients. The estimated baseline inpatient cost for the index hospitalizations was $2.3 billion and within 2 years after discharge, the aggregate cost of readmission approached$2 billion. Loss of antipsychotic efficacy was estimated to account for 60% of the rehospitalisation costs and non-adherence for 40% ( Weiden and Olfson, 1995 ). In a cross-sectional analysis conducted in the United Kingdom costs, clinical outcomes and quality of life were compared between patients who had relapsed in the previous 6 months and those who had not. Perhaps surprisingly, there were few differences in clinical and quality of life outcomes between the two groups. However, costs for the patients who relapsed were more than four times higher than those for the non-relapse group ( Almond et al., 2004 ). A Spanish secondary analysis of data obtained from a RCT comparing ziprasidone with placebo over 52 weeks for relapse prevention (n = 218) estimated the cost-effectiveness of active treatment vs. placebo. The active treatment was associated with reduced risk of relapse and the average annual incremental cost per relapse avoided was calculated as 186 euro (mean dose), which was much lower than the minimum cost of a relapse (2830 euro) ( Bernardo et al., 2006 ). Another study investigated the direct cost of relapse during the treatment of patients with schizophrenia in the United States. The sample comprised 1557 participants, of whom 20% relapsed during the 6 months prior to the 1-year study period. Costs for patients with prior relapse were about 3 times that for the patients without prior relapse. Relapse was associated with higher costs for inpatient services as well as for outpatient services and medication ( Ascher-Svanum et al., 2010 ).

#### 5.2. Mortality

As there have been reports that untreated schizophrenia patients have a higher risk of death than those on active treatment ( Tiihonen et al., 2009 ), this issue also needs to be considered in the context of placebo-controlled clinical trials. Mortality rates were assessed for adult patients with schizophrenia assigned to an investigational antipsychotic, active control antipsychotic, or placebo in an analysis of pre-marketing clinical development programs. Safety data were reviewed for 16,791 adult patients with schizophrenia. The mortality rate calculated for patient exposure years for patients assigned to placebo treatment was significantly higher (p < .05) than for either the investigational antipsychotic (OR = 0.23, 95% CI = 0.13 to 0.45) or the active control group (OR = 0.19, 95% CI = 0.08 to 0.45). The causes of death for placebo-treated patients were however not reported. The most frequent causes of death among patients receiving investigational antipsychotics were suicide (28%), cardiovascular events (18%), and respiratory complications (17%) ( Khan et al., 2007 ).

#### 5.3. Risk of suicide

The risk of suicide is significantly increased in schizophrenia, with an estimated 10–13% of individuals with the illness committing suicide. In a 10-year retrospective review of Medical Examiners' cases of suicide in Louisville, Kentucky, USA, cases were aged 20 to 75 (mean 41.6) years. Most were male (62.1%) and Caucasian (86.2%) ( Shields et al., 2007 ). Non-adherence to medication appears to increase the risk for suicide in schizophrenia. Suicide attempts have been linked to non-adherence to antipsychotic medication in a study examining the pharmacy records of 603 patients with schizophrenia aged 15–45 years. Non-adherence to medication for at least 30 days was reported in 33% of the patients, and these patients had a four-fold increased risk for suicide attempts ( Herings and Erkens, 2003 ). However, two studies investigating suicide rates in placebo-controlled RCTs in schizophrenia report no increased risk in patients receiving placebo. In the first study, risks of suicide and suicide attempts in psychotic patients assigned to receive placebo in clinical trials were assessed. The authors searched the Food and Drug Administration database to assess suicides and suicide attempts in clinical trials of three new antipsychotics. Out of a total of 10,118 participants, 26 committed suicide and 51 attempted suicide. Annual rates of suicide and attempted suicide based on patient exposure years were, respectively, 1.8% and 3.3% for placebo; 0.9% and 5.7% for an active control treatment and 0.7% and 5.0% for an investigational antipsychotic ( Khan et al., 2001 ). In another study, all placebo-controlled RCTs that were submitted to the regulatory authority of the Netherlands as part of a registration dossier for the treatment of schizophrenia from 1992 through 2002, were reviewed for suicide and attempted suicide. In total, 7152 patients from 31 studies were included, of whom 1888 were randomised to receive placebo. One suicide occurred in the placebo groups and 1 in the active compound groups. Two attempted suicides occurred in the placebo groups and 11 in the active compound groups. These differences were not statistically significant. On the basis of these findings, the authors concluded that concern about increased risk of suicide should not be an argument against the conduct of placebo-controlled RCTs in schizophrenia, provided that appropriate precautions are taken ( Storosum et al., 2003 ). However, a note of caution has been sounded and it has been pointed out that the overall suicide rates in these RCTs are still at least as high as estimates for those populations in real world settings, despite the routine practice of excluding subjects considered to be at risk for suicide from RCTs. It may be that, because suicide is relatively infrequent, the analyses have limited power to detect differences despite their large numbers ( Kim, 2003 ).

### 6. Studies investigating the outcome after relapse in patients with schizophrenia

A few studies have investigated the impact of relapse on outcome in schizophrenia.

A retrospective, follow-up study of first admission schizophrenic patients who participated in a study more than 50 years ago in which they were randomly assigned to treatment with and without antipsychotic medications was conducted. During the second year following discharge, patients who were initially treated with antipsychotic medications required fewer rehospitalisation days than the initially non-medicated patients, and at 6–7 years following discharge, they were functioning at a higher level in terms of Global Assessment of Functioning scores than patients who did not initially receive treatment with antipsychotic medications. The authors concluded that the results suggest that early treatment with antipsychotic medications decreases the immediate morbidity, and prevented detrimental changes over time ( Wyatt et al., 1997 ). In a 15-yr follow-up of the natural course of schizophrenia in a Dutch incidence cohort, Wiersma et al. (1998) examined the course of the illness after the first four psychotic episodes. Two-thirds of the 82 participants experienced at least one relapse. The authors reported a striking finding, i.e. that on average, 1 in 6 patients failed to remit in terms of persisting psychotic symptoms after each episode, irrespective of which episode it was. Additionally, about 1 in 5 patients developed persistent negative symptoms after each relapse — i.e. an increase from 27% after the first episode to 47% after the 4th episode. Nine patients (11%) committed suicide, four of whom did so during their first or second psychotic episode, while the others occurred during periods of partial remission.

Results from a sub-set of patients from a longitudinal outcome study reported increased times to treatment response in succeeding episodes for patients who had relapsed. Median (SE) times to remission increased from 8.4 weeks (SE = 1.3) for the first episode to 11.9 weeks (SE = 2.3) for the 2nd episode (p = 0.06) for 22 patients who had two episodes, and from 4 weeks (SE = 2.5), to 7 weeks (SE = 1.6) and 24.1 weeks (SE = 2.5) respectively, for 6 patients who had three episodes (p = 0.001) ( Lieberman et al., 1996 ).

In a 7-year follow up study of a small sample of first-episode patients, 31 of 50 participants who had completed a 2-year treatment phase for a first episode of schizophrenia were entered into an intermittent treatment extension phase until recurrence, followed by a further 2-year treatment phase for the recurrence episode. Symptom improvements and functional outcome scores were similar for the 1st episode and 2nd episode for the patients who remained in treatment. However, there was a high discontinuation rate (55%) after relapse, and emergent treatment failure was observed in 5 participants (16%) ( Emsley et al., 2013 ). These findings are consistent with those of Wiersma et al. (1998) , and suggest that the hypothesis that relapse may be biologically harmful ( Lieberman et al., 1996 ) may hold true at least for a subset of patients.

### 7. Studies investigating the outcome of patients who received placebo in RCTs

However, not all evidence points to deterioration after relapse. For example, it has been observed that patients' symptoms rapidly return to baseline if medication is reinstituted shortly after recurrence of psychotic symptoms ( Glovinsky et al., 1992 ), and trajectory analysis of the course of schizophrenia suggests evolution from initial deterioration to a course more consistent with amelioration for most patients ( Levine et al., 2011 ). Also, in a literature review, the clinical and neurobiological evidence for neurotoxicity of psychosis has been examined and found wanting ( McGlashan, 2006 ). Finally, in a recent critical review of the evidence from longitudinal studies of clinical outcomes, MRI brain volumes, and cognitive functioning, the entire concept of schizophrenia as a progressive disorder has been challenged. The authors point out the following: While approximately 25% of people with schizophrenia have a poor long-term outcome, few studies show that the loss of function is incremental, as would be expected with a neurodegenerative illness; reported decreases in brain tissue volumes could be explained by the effects of antipsychotic medication, substance abuse, and other secondary factors; while patients do show cognitive deficits compared with controls, cognitive functioning does not appear to deteriorate over time; finally the fact that some experience deterioration in functioning over time may reflect poor access, or adherence, to treatment, the effects of concurrent conditions, and social and financial impoverishment ( Zipursky et al., in press ).

Furthermore, when considering the risks of treatment discontinuation or placebo-treatment in RCTs, it needs to be remembered that ongoing antipsychotic treatment carries its own risks and burden. Thus, while the risk of relapse is greatest with treatment discontinuation, substantial numbers of patients receiving continuous active treatment experience relapse-events (see Table 1 ). Additionally, there is a substantial safety and tolerability burden associated with antipsychotic medication, and even the possibility that long-term antipsychotic treatment itself may be associated with detrimental brain morphological changes ().

### 8. Potential alternatives to the standard placebo-controlled design

Given the controversies around the traditional placebo-controlled relapse prevention study alternative study designs need to be entertained. The pros and cons of non-inferiority studies have been discussed in a previous paragraph. The most powerful alternative would be a study demonstrating superiority when comparing an experimental treatment to a standard intervention. Admittedly, such a design only holds promise if a large difference between the two interventions, either with regard to efficacy or safety, is to be expected. At the current stage of monotherapy drug development in schizophrenia this appears to be an overambitious goal. Another option, already employed in some more recent studies () is to use a suboptimal dose of an active antipsychotic in lieu of a placebo. Admittedly, this is somewhat of a compromise yet in this case treatment is at least not totally withheld in patients randomized to such a study arm. Another way to reduce the number of patients exposed to placebo is to decrease the number of patients in the placebo group in favour of larger numbers on active treatment(s). This could result in a study in which for instance 200 patients would be randomized to an investigational compound, 200 to standard treatment and 50 to placebo. Needless to say, power calculations based on earlier research will have to guide the choice of sample size in the respective groups. Another alternative is the placebo controlled add-on study, in which patients receive standard treatment augmented by an experimental compound, the effects of which are compared to the standard treatment plus placebo. Such clinical trials are regularly performed in other fields of medicine, for instance in neurology to test the effectiveness of antiepileptic drugs.

A hybrid design incorporating a randomised withdrawal period of 6 months into a long-term parallel group active comparator trial after at least 12 months of treatment is proposed as an optimal approach, as it includes the strengths of both designs — i.e. establishing assay sensitivity as well as obtaining long-term safety and efficacy data ( European Medicines Agency Committee for Medicinal Products for Human Use, 2012 ).

The usefulness of all of the designs briefly outlined above depends on the type of intervention studied and the primary outcome hypothesis. Non-inferiority or superiority designs, for instance, could be used when comparing medications with very distinct pharmacological profiles where safety/tolerability is the main outcome measure. Placebo-controlled add-on studies on the other hand are likely helpful when looking for outcomes beyond mere relapse prevention, as for instance enhanced quality of life or psychosocial integration. The suboptimal dose approach, combined with an unbalanced randomization scheme has already proven useful in classic relapse prevention trials.

### 9. Conclusions

Few studies have investigated the long-term consequences of exposure to placebo in schizophrenia. Also, while the epidemiology of relapse has been studied extensively, only a few studies have prospectively investigated the actual consequences of relapse. Well-designed longitudinal studies are needed, assessing the numerous psychosocial and biological domains that may be affected by relapse. In the absence of such studies it is difficult to assume that patients experiencing relapses are not at risk of severe suffering or lasting harm. In a thoughtful paper considering the risk–benefit analysis of placebo in RCTs in psychiatric research, it is stated that, since no evidence of harm is different from evidence for no harm, the critical issue is which of these carries the burden of proof ( Kim, 2003 ). If the burden is to show evidence for no harm, it is clear that we currently lack the data for such a claim. Indeed, based on the available evidence and common sense we consider that there is sufficient cause for concern that a real risk exists for substantial patient and carer suffering, detrimental effects on social and occupational function, delayed time to treatment response and even the possibility of emergent treatment refractoriness. The argument for inclusion of an acute-treatment placebo controlled study in the developmental program of experimental drugs remains strong both scientifically and ethically. However, we find it difficult to justify the on-going use of placebo in relapse-prevention trials in schizophrenia, particularly given the difficulties in identifying early warning signs and the ineffectualness of rescue medication in preventing full-blown relapse.

None.

### Contributors

Both of the authors met International Council of Medical Journal Editors criteria. Both provided intellectual contributions and approved the final draft submission to the journal.

### Conflicts of interest

AstraZeneca, Janssen, Lilly, Lundbeck, Pfizer, Servier, Otsuka and Wyeth. He has received research funding from Janssen, Lundbeck and AstraZeneca.

Prof. Fleischhacker has received research grants and honoraria from the following companies: AstraZeneca, Bristol Myers Squibb, Eli Lilly, Janssen, Novartis, Organon, Otsuka, Pfizer, Sanofi-Synthelabo.

None.

### References

• D. Addington. The use of placebos in clinical trials for acute schizophrenia. Can. J. Psychiatry. 1995;40:171-176
• S. Almond, M. Knapp, C. Francois, M. Toumi, T. Brugha. Relapse in schizophrenia: costs, clinical outcomes and quality of life. Br. J. Psychiatry. 2004;184:346-351
• N.C. Andreasen, D. Liu, B.A. Ziebell, A. Vora, B.C. Ho. Relapse duration, treatment intensity, and brain tissue loss in schizophrenia: a prospective longitudinal MRI study. Am. J. Psychiatry. 2013;10.1176/appi.ajp.2013.12050674 ([Epub ahead of print])
• M. Arato, R. O'Connor, H.Y. Meltzer. A 1-year, double-blind, placebo-controlled trial of ziprasidone 40, 80 and 160 mg/day in chronic schizophrenia: the Ziprasidone Extended Use in Schizophrenia (ZEUS) study. Int. Clin. Psychopharmacol.. 2002;17:207-215
• H. Ascher-Svanum, B. Zhu, D.E. Faries, D. Salkever, E.P. Slade, X. Peng, R.R. Conley. The cost of relapse and the predictors of relapse in the treatment of schizophrenia. BMC Psychiatry. 2010;10:2
• C.M. Beasley Jr., V.K. Sutton, S.H. Hamilton, D.J. Walker, M. Dossenbach, C.C. Taylor, K.J. Alaka, D. Bykowski, G.D. Tollefson. A double-blind, randomized, placebo-controlled trial of olanzapine in the prevention of psychotic relapse. J. Clin. Psychopharmacol.. 2003;23:582-594
• M. Bernardo, A.J. Ramon, C. Rubio-Terres, J. Rejas. Cost-effectiveness analysis of schizophrenia relapse prevention: an economic evaluation of the ZEUS (Ziprasidone-Extended-Use-In-Schizophrenia) study in Spain. Clin. Drug Investig.. 2006;26:447-457
• M. Birchwood, J. Smith, F. MacMillan, B. Hogg, R. Prasad, C. Harvey, S. Bering. Predicting relapse in schizophrenia: the development and implementation of an early signs monitoring system using patients and families as observers, a preliminary investigation. Psychol. Med.. 1989;19:649-656
• W.T. Carpenter Jr., P.S. Appelbaum, R.J. Levine. The Declaration of Helsinki and clinical trials: a focus on placebo-controlled trials in schizophrenia. Am. J. Psychiatry. 2003;160:356-362
• E.Y. Chen, C.L. Hui, M.M. Lam, C.P. Chiu, C.W. Law, D.W. Chung, S. Tso, E.P. Pang, K.T. Chan, Y.C. Wong, F.Y. Mo, K.P. Chan, T.J. Yao, S.F. Hung, W.G. Honer. Maintenance treatment with quetiapine versus discontinuation after one year of treatment in patients with remitted first episode psychosis: randomised controlled trial. BMJ. 2010;341:c4024
• S.J. Cooper, A. Butler, J. Tweed, C. Welch, J. Raniwalla. Zotepine in the prevention of recurrence: a randomised, double-blind, placebo-controlled study for chronic schizophrenia. Psychopharmacology (Berl). 2000;150:237-243
• C.U. Correll, T. Kishimoto, J.M. Kane. Randomized controlled trials in schizophrenia: opportunities, limitations, and trial design alternatives. Dialogues Clin. Neurosci.. 2011;13:155-172
• D.A. Curson, S.R. Hirsch, S.D. Platt, R.W. Bamber, T.R. Barnes. Does short term placebo treatment of chronic schizophrenia produce long term harm?. Br. Med. J. (Clin. Res. Ed.). 1986;293:726-728
• M.A. Dellva, P. Tran, G.D. Tollefson, A.L. Wentley, C.M. Beasley Jr. Standard olanzapine versus placebo and ineffective-dose olanzapine in the maintenance treatment of schizophrenia. Psychiatr. Serv.. 1997;48:1571-1577
• R. Emsley, I. Nuamah, D. Hough, S. Gopal. Treatment response after relapse in a placebo-controlled maintenance trial in schizophrenia. Schizophr. Res.. 2012;138:29-34
• R. Emsley, P.P. Oosthuizen, L. Koen, D.J. Niehaus, G. Martinez. Symptom recurrence following intermittent treatment in first-episode schizophrenia successfully treated for 2 years: a 3-year open-label clinical study. J. Clin. Psychiatry. 2012;73:e541-e547
• R. Emsley, P. Oosthuizen, L. Koen, D. Niehaus, L. Martinez. Comparison of treatment response in second-episode versus first-episode schizophrenia. J. Clin. Psychopharmacol.. 2013;33:80-83
• European Medicines Agency Committee for Medicinal Products for Human Use. Guideline on clinical investigation of medicinal products, including depot preparations in the treatment of schizophrenia. EMA/CHMP/40072/2010 (, 2012) (9-20)
• W.W. Fleischhacker, T. Burns. Feasibility of placebo-controlled clinical trials of antipsychotic compounds in Europe. Psychopharmacology (Berl). 2002;162:82-84
• W.W. Fleischhacker, P. Czobor, M. Hummer, G. Kemmler, R. Kohnen, J. Volavka. Placebo or active control trials of antipsychotic drugs?. Arch. Gen. Psychiatry. 2003;60:458-464
• W.W. Fleischhacker, R. Sanchez, P.P. Perry, N. Jin, T. Peters-Strickland, B.R. Johnson, R.A. Baker, A. Eramo, R.D. McQuade, W.H. Carson, J.M. Kane. Aripiprazole once-monthly for the treatment of schizophrenia: a double-blind, randomized, non-inferiority study vs. oral aripiprazole. Poster presented at the 51st ACNP Annual Meeting (, 2012)
• W. Gaebel, M. Riesbeck. Revisiting the relapse predictive validity of prodromal symptoms in schizophrenia. Schizophr. Res.. 2007;95:19-29
• W. Gaebel, U. Frick, W. Kopcke, M. Linden, P. Muller, F. Muller-Spahn, A. Pietzcker, J. Tegeler. Early neuroleptic intervention in schizophrenia: are prodromal symptoms valid predictors of relapse?. Br. J. Psychiatry Suppl.. 1993;8–12
• P.L. Gilbert, M.J. Harris, L.A. McAdams, D.V. Jeste. Neuroleptic withdrawal in schizophrenic patients. A review of the literature. Arch. Gen. Psychiatry. 1995;52:173-188
• M. Gitlin, K. Nuechterlein, K.L. Subotnik, J. Ventura, J. Mintz, D.L. Fogelson, G. Bartzokis, M. Aravagiri. Clinical outcome following neuroleptic discontinuation in patients with remitted recent-onset schizophrenia. Am. J. Psychiatry. 2001;158:1835-1842
• Gleeson J.F., S.M. Cotton, Alvarez-Jimenez M., D. Wade, D. Gee, K. Crisp, T. Pearce, D. Spiliotacopoulos, B. Newman, P.D. McGorry. A randomized controlled trial of relapse prevention therapy for first-episode psychosis patients: outcome at 30-month follow-up. Schizophr. Bull.. 2013;39(2):436-448 (Mar)
• D. Glovinsky, D.G. Kirch, R.J. Wyatt. Early antipsychotic response to resumption of neuroleptics in drug-free chronic schizophrenic patients. Biol. Psychiatry. 1992;31:968-970
• R.M. Herings, J.A. Erkens. Increased suicide attempt rate among patients interrupting use of atypical antipsychotics. Pharmacoepidemiol. Drug Saf.. 2003;12:423-424
• D. Hough, S. Gopal, U. Vijapurkar, P. Lim, M. Morozova, M. Eerdekens. Paliperidone palmitate maintenance treatment in delaying the time-to-relapse in patients with schizophrenia: a randomized, double-blind, placebo-controlled study. Schizophr. Res.. 2010;116:107-117
• J.M. Kane. Treatment strategies to prevent relapse and encourage remission. J. Clin. Psychiatry. 2007;68(Suppl. 14):27-30
• J.M. Kane, H.C. Detke, D. Naber, G. Sethuraman, D.Y. Lin, R.F. Bergstrom, D. McDonnell. Olanzapine long-acting injection: a 24-week, randomized, double-blind trial of maintenance treatment in patients with schizophrenia. Am. J. Psychiatry. 2010;167:181-189
• J.M. Kane, M. Mackle, L. Snow-Adami, J. Zhao, A. Szegedi, J. Panagides. A randomized placebo-controlled trial of asenapine for the prevention of relapse of schizophrenia after long-term treatment. J. Clin. Psychiatry. 2011;72:349-355
• J.M. Kane, R. Sanchez, P.P. Perry, N. Jin, B.R. Johnson, R.A. Forbes, R.D. McQuade, W.H. Carson, W.W. Fleischhacker. Aripiprazole intramuscular depot as maintenance treatment in patients with schizophrenia: a 52-week, multicenter, randomized, double-blind, placebo-controlled study. J. Clin. Psychiatry. 2012;73:617-624
• G. Kemmler, M. Hummer, C. Widschwendter, W.W. Fleischhacker. Dropout rates in placebo-controlled and active-control clinical trials of antipsychotic drugs: a meta-analysis. Arch. Gen. Psychiatry. 2005;62:1305-1312
• A.S. Kemp, N.R. Schooler, A.H. Kalali, L. Alphs, R. Anand, G. Awad, M. Davidson, S. Dube, L. Ereshefsky, G. Gharabawi, A.C. Leon, J.P. Lepine, S.G. Potkin, A. Vermeulen. What is causing the reduced drug-placebo difference in recent schizophrenia clinical trials and what can be done about it?. Schizophr. Bull.. 2010;36:504-509
• A. Khan, S.R. Khan, R.M. Leventhal, W.A. Brown. Symptom reduction and suicide risk among patients treated with placebo in antipsychotic clinical trials: an analysis of the food and drug administration database. Am. J. Psychiatry. 2001;158:1449-1454
• A. Khan, K. Schwartz, C. Stern, N. Redding, R.L. Kolts, W.A. Brown, D.S. Robinson. Mortality risk in patients with schizophrenia participating in premarketing atypical antipsychotic clinical trials. J. Clin. Psychiatry. 2007;68:1828-1833
• S.Y. Kim. Benefits and burdens of placebos in psychiatric research. Psychopharmacology (Berl). 2003;171:13-18
• G.T. Konopaske, K.A. Dorph-Petersen, R.A. Sweet, J.N. Pierri, W. Zhang, A.R. Sampson, D.A. Lewis. Effect of chronic antipsychotic exposure on astrocyte and oligodendrocyte numbers in macaque monkeys. Biol. Psychiatry. 2008;63:759-765
• M. Kramer, G. Simpson, V. Maciulis, S. Kushner, U. Vijapurkar, P. Lim, M. Eerdekens. Paliperidone extended-release tablets for prevention of symptom recurrence in patients with schizophrenia: a randomized, double-blind, placebo-controlled study. J. Clin. Psychopharmacol.. 2007;27:6-14
• T.P. Laughren. The scientific and ethical basis for placebo-controlled trials in depression and schizophrenia: an FDA perspective. Eur. Psychiatry. 2001;16:418-423
• S. Leucht, M. Tardy, K. Komossa, S. Heres, W. Kissling, J.M. Davis. Maintenance treatment with antipsychotic drugs for schizophrenia. Cochrane Database Syst. Rev.. 2012;5:CD008016
• S.Z. Levine, I. Lurie, R. Kohn, I. Levav. Trajectories of the course of schizophrenia: from progressive deterioration to amelioration over three decades. Schizophr. Res.. 2011;126:184-191
• J.A. Lieberman, J.M. Alvir, A. Koreen, S. Geisler, M. Chakos, B. Sheitman, M. Woerner. Psychobiologic correlates of treatment response in schizophrenia. Neuropsychopharmacology. 1996;14:13S-21S
• T.H. McGlashan. Is active psychosis neurotoxic?. Schizophr. Bull.. 2006;32:609-613
• R.M. Norman, A.K. Malla. Prodromal symptoms of relapse in schizophrenia: a review. Schizophr. Bull.. 1995;21:527-539
• D.O. Perkins, H. Gu, P.J. Weiden, J.P. McEvoy, R.M. Hamer, J.A. Lieberman. Predictors of treatment discontinuation and medication nonadherence in patients recovering from a first episode of schizophrenia, schizophreniform disorder, or schizoaffective disorder: a randomized, double-blind, flexible-dose, multicenter study. J. Clin. Psychiatry. 2008;69:106-113
• J. Peuskens, J. Trivedi, S. Malyarov, M. Brecher, O. Svensson, F. Miller, I. Persson, D. Meulien. Prevention of schizophrenia relapse with extended release quetiapine fumarate dosed once daily: a randomized, placebo-controlled trial in clinically stable patients. Psychiatry (Edgmont). 2007;4:34-50
• T.A. Pigott, W.H. Carson, A.R. Saha, A.F. Torbeyns, E.G. Stock, G.G. Ingenito. Aripiprazole for the prevention of relapse in stabilized patients with chronic schizophrenia: a placebo-controlled 26-week study. J. Clin. Psychiatry. 2003;64:1048-1056
• D. Robinson, M.G. Woerner, J.M. Alvir, R. Bilder, R. Goldman, S. Geisler, A. Koreen, B. Sheitman, M. Chakos, D. Mayerhoff, J.A. Lieberman. Predictors of relapse following response from a first episode of schizophrenia or schizoaffective disorder. Arch. Gen. Psychiatry. 1999;56:241-247
• N.R. Schooler, S.J. Keith, J.B. Severe, S.M. Matthews, A.S. Bellack, I.D. Glick, W.A. Hargreaves, J.M. Kane, P.T. Ninan, A. Frances, M. Jacobs, J.A. Lieberman, R. Mance, G.M. Simpson, M.G. Woerner. Relapse and rehospitalization during maintenance treatment of schizophrenia. The effects of dose reduction and family treatment. Arch. Gen. Psychiatry. 1997;54:453-463
• L.B. Shields, D.M. Hunsaker, J.C. Hunsaker III. Schizophrenia and suicide: a 10-year review of Kentucky medical examiner cases. J. Forensic Sci.. 2007;52:930-937
• J.G. Storosum, B.J. van Zwieten, T. Wohlfarth, L. de Haan, A. Khan, W. van den Brink. Suicide risk in placebo vs active treatment in placebo-controlled trials for schizophrenia. Arch. Gen. Psychiatry. 2003;60:365-368
• J. Tiihonen, J. Lönnqvist, K. Wahlbeck, T. Klaukka, L. Niskanen, A. Tanskanen, J. Haukka. 11-year follow-up of mortality in patients with schizophrenia: a population-based cohort study (FIN11 study). Lancet. 2009;374:620-627
• U.S. Food and Drug Administration Center for Drug Evaluation and Research. Guidance for industry: enrichment strategies for clinical trials to support approval of human drugs and biological products. Draft guidance. (, 2012) (Dec, www.fda.gov/downloads/Drugs//Guidances/UCM33218 )
• N. van de Laar, I. Henter, J.J. Bartko, R.J. Wyatt. Adverse events during a placebo phase for inpatients with chronic schizophrenia. Biol. Psychiatry. 2001;50:487-492
• P.J. Weiden, M. Olfson. Cost of relapse in schizophrenia. Schizophr. Bull.. 1995;21:419-429
• C. Weijer. Placebo-controlled trials in schizophrenia: are they ethical? Are they necessary?. Schizophr. Res.. 1999;35:211-218
• D. Wiersma, F.J. Nienhuis, C.J. Slooff, R. Giel. Natural course of schizophrenic disorders: a 15-year followup of a Dutch incidence cohort. Schizophr. Bull.. 1998;24:75-85
• World Federation of Mental Health. Keeping Care Complete: Caregivers' Perspectives on Mental Illness and Wellness. An International Survey. (World Federation of Mental Health, 2006)
• World Medical Association (WMA). Declaration of Helsinki. Ethical principles for medical research involving human subjects. Adopted by the 18th WMA General Assembly, Helsinki, Finland, June 1964 and amended by the 59th WMA General Assembly, Seoul, South-Korea, October 2008 (, 2008)
• R.J. Wyatt, M.F. Green, A.H. Tuma. Long-term morbidity associated with delayed treatment of first admission schizophrenic patients: a re-analysis of the Camarillo State Hospital data. Psychol. Med.. 1997;27:261-268
• R.J. Wyatt, I.D. Henter, J.J. Bartko. The long-term effects of placebo in patients with chronic schizophrenia. Biol. Psychiatry. 1999;46:1092-1105
• R.B. Zipursky, T.J. Reilly, R.M. Murray. The myth of schizophrenia as a progressive brain disease. Schizophr Bull.. 2013; (Dec 7, Epub ahead of print, in press)

### Footnotes

a Department of Psychiatry, Faculty of Medicine and Health Sciences, Stellenbosch University, Cape Town, South Africa

b Department of Biological Psychiatry, Innsbruck University Clinics, Austria

Corresponding author: Department of Psychiatry, University of Stellenbosch, Room2027, 2nd Floor Clinical Building, Fransie van Zijl Drive, Tygerberg 7505, Cape Town, South Africa. Tel.: + 27219389227; fax: + 27219389738.