Asenapine in the treatment of negative symptoms of schizophrenia: clinical trial design and rationale
Статьи Опубликовано в журнале:«PSYCHOPHARMACOLOGY BULLETIN»; Vol. 40; No. 2; стр. 41-53.
By Larry Alphs, MD, PhD, John Panagides, PhD, and Scott Lancaster, MS
Abstract ~ Although the positive symptoms of schizophrenia are more likely than the negative symptoms to result in a patient's hospitalization, positive symptoms tend to respond more completely to antipsychotic drugs. When positive symptoms are controlled, residual negative symptoms may remain. If these negative symptoms persist, they can have a considerable impact on a patient's ability to function in society. Current therapies have only a limited effect on negative symptoms. Consequently, broad-spectrum agents that effectively treat both positive and negative symptoms are needed. One obstacle to the regulatory approval of an agent for treating negative symptoms is the difficulty of designing a trial to demonstrate efficacy for these symptoms. Agreeing on a definition of negative symptoms, establishing patient inclusion criteria, and determining how to account for confounding factors represent only a few of the challenges to study design.
How these challenges can be met is illustrated in the design of a series of clinical trials to assess the efficacy of asenapine, a psychopharmacologic agent being developed for the treatment of schizophrenia and, in particular, the treatment of negative symptoms associated with schizophrenia. These trials, the protocols for which are described in this paper, will not only determine the efficacy of asenapine but will add to our knowledge of patients with predominant, persistent negative symptoms, an understudied and inadequately treated patient population.
Key Words: asenapine, clinical trial, study design, negative symptoms, schizophrenia
Introduction
Schizophrenia encompasses various symptoms, including both overt positive symptoms and the less obtrusive, but often more psychosocially devastating, negative symptoms. Negative symptoms are a loss of normal function, with patients displaying decreased emotional expressivity and perception, reduced fluency and productivity of thought and speech, reduced desire for social involvement and social interaction, and a loss or lack of goal-directed behavior,1 all of which have consequences for a patient's ability to function in society.
Negative symptoms can occur in schizophrenia regardless of the stage of illness. Indeed, they may be an early sign of schizophrenia, with social withdrawal and decreased motivation preceding a first psychotic episode.2 Negative and positive symptoms can coexist in the same patient, but it is the positive symptoms that are more likely to result in hospitalization. When positive symptoms are controlled with antipsychotic medication, persistent negative symptoms may predominate as a key area of dysfunction.1
Currently available antipsychotics, both conventional and atypical agents, are effective for treating the positive symptoms of schizophrenia. Conventional antipsychotics, however, appear to have an insufficient effect on negative symptoms, and although some atypical agents may have a modest therapeutic effect on these symptoms,3 it remains unclear whether the effect is direct or indirect. Thus, despite the therapeutic effects of our best treatments, treatment of negative symptoms remains a significant unmet need.
In this report, we discuss some of the challenges encountered when designing clinical trials to assess a drug's efficacy for negative symptoms and describe how these challenges have been met in clinical trials of asenapine, a psychopharmacologic agent in development for the treatment of schizophrenia and bipolar disorder. These asenapine studies, referred to collectively as the Aphrodite trials, consist of two 26-week, multicenter, double-blind, flexible-dose trials that compare asenapine with olanzapine in stable patients with predominant, persistent negative symptoms of schizophrenia. Each 26-week study is followed by a 26-week blinded extension trial, for a total study duration of 52 weeks.
Conceptual Issues in Clinical Trial Design
Defining Negative Symptoms
A key aspect of designing large multinational clinical trials that assess a drug's efficacy for negative symptoms is to precisely define these symptoms in a way that is understood by both investigators and supporting clinicians and that is accepted by drug regulatory agencies worldwide. To allow a comparison of results across studies, a single precisely articulated definition of negative symptoms is needed.4
Distinguishing between primary and secondary negative symptoms also is important for identifying potential treatments. Primary negative symptoms are a feature of schizophrenia intrinsic to the disease itself and not secondary to other medical, cognitive, psychologic, or environmental causes.1,5 These symptoms have been recognized since the early 1900s as a fundamental component of schizophrenia.6 Secondary negative symptoms result from other factors, such as adverse drug effects, particularly extrapyramidal symptoms; depression; cognitive deficits; hospitalization; or environmental deprivation (Figure 1).1,5-7 In some cases, negative symptoms represent a response to positive symptoms, as in the case of severe paranoia leading to social withdrawal or mutism.
FIGURE 1
Designing Trials for Broad-Spectrum Antipsychotic Agents
A broad-spectrum agent for the treatment of schizophrenia is one that is effective for both positive symptoms and negative symptoms.8 One challenge in simultaneously gaining regulatory approval for treatment of positive and negative symptoms is the need to demonstrate, through convincing scientific data from well-designed studies, that a drug is effective for both types of symptoms.8
Assessing the efficacy of a broad-spectrum drug for both positive and negative symptoms in a single set of studies seems efficient. However, this approach requires that two distinct primary endpoints be identified and met in a single study. Statistical methods would need to be incorporated to adequately power the study to demonstrate efficacy for both types of symptoms. Furthermore, the characteristics of the study population for demonstrating efficacy for positive symptoms are likely to differ from those required for demonstrating efficacy for negative symptoms. As previously indicated, an individual patient may exhibit positive and negative symptoms in any combination or severity. Patients who have sufficient symptoms to anticipate change in one endpoint may not have sufficient symptoms in the other. Requiring patients to meet requirements for both endpoints may lead to identification of a small subpopulation whose response does not generalize to the entire population. Also, given the nature of negative symptoms, treatment effects would need to be established in patients whose negative symptoms are well established and relatively stable for a number of months. On the other hand, effects on positive symptoms are generally demonstrated in patients with an acute exacerbation of schizophrenia. Thus, patients appropriate for the positive symptoms endpoint may be inappropriate for the negative symptoms endpoint.
To appropriately label the treatment effect, the nature of the response must be correctly identified. Concerns have been raised that improvement in positive symptoms by a broad-spectrum drug might indirectly improve negative symptoms. Similarly, improvements in cognition, depression, or extrapyramidal symptoms or a lessening of sedation might also be incorrectly identified as an effect on primary negative symptoms. Careful design and clinical assessments are needed to avoid these confounds to interpretation.
Consequently, to demonstrate the efficacy of a broad-spectrum agent for both positive and negative symptoms, separate trials with carefully defined populations are desirable so that data can be adequately interpreted. In particular, the effects of treatment on primary negative symptoms must be distinguished from those on secondary negative symptoms.9 In such studies, sources of secondary negative symptoms need to remain stable for the duration of the trial.9
To demonstrate the efficacy of a broad-spectrum agent for treating negative symptoms, a placebo control is not possible because, to be of interest, any effect must be enduring. However, long-term monotherapy with a placebo is likely to result in recurrence of the full disease state with exacerbation of positive symptoms in many patients. Such outcomes have consequences both for the ethical conduct of the trial and for data interpretation. For these reasons, the investigational drug must be evaluated against an active comparator.
The choice of the active comparator and the dose to be used require careful consideration. Clearly, the active comparator must have a minimal likelihood of producing secondary negative symptoms and should not have a differential treatment effect on other variables, such as depressive symptoms or cognitive deficits that might confound interpretation of effects on negative symptoms. In addition, dosing and titration schedules for both the broad-spectrum agent and the comparator drug should be designed to minimize the occurrence of adverse effects that could affect the interpretation of results. Conventional antipsychotics, such as haloperidol, are poor choices for comparator agents because of the likelihood of producing extrapyramidal symptoms.
To address concerns that observed differences may result from worsening of negative symptoms during treatment with the active comparator, rather than from improvement following treatment with the novel broad-spectrum agent, results should be compared with stable baseline evaluations of negative symptoms.
Covariate and multivariate statistical analyses can be used to identify and mitigate some of the effects of confounding factors. However, unless confounding variables are minimized, a clear interpretation of results will not be possible.9,10
Possible Trial Designs for Broad-Spectrum Drug
There are several possible design approaches for evaluating the efficacy of a broad-spectrum agent for schizophrenia. One approach is to conduct a randomized, double-blind trial in a "real-world" patient population. With this study design, patients are randomly assigned to start treatment with the new broad-spectrum drug or to continue their current treatment. The advantages of this type of study include rapid patient recruitment, because few patients would be excluded from enrollment, and the results would be broadly generalizable, because data would be gathered from patients seen in a "typical" clinical practice. However, these advantages are offset by two important disadvantages. First, comparing the new treatment to "standard care" would provide a reference population in which the potential confounding effects of these treatments on secondary negative symptoms are not adequately controlled. Second, conducting a double-blind trial in this population is difficult because it would be logistically impossible to blind the patient and prescribing physician to the treatment the patients are receiving; particularly for patients in the comparator group, who would be taking various drugs and various doses. Consequently, this type of study would likely require open-label treatment with blinded-rater assessments.
Another design approach is a study in which all eligible patients are initially switched from their current medication to treatment with a single antipsychotic drug. After a period of stable treatment on this antipsychotic, patients whose negative symptoms remained stable would be randomly assigned to treatment with the broad-spectrum drug or a comparator drug. In this way, a double-blind design is possible. This randomized, double-blind design also provides a consistent approach to achieving baseline values for all patients, but at the expense of slower recruitment and possibly higher withdrawal rates during the prerandomization antipsychotic treatment period. Nonetheless, with such a design, even a small effect size11 is likely to be clinically meaningful because the effect of the broad-spectrum agent would be compared with that of a drug already proven to be clinically effective in treating general symptoms of schizophrenia, although not specifically for treating negative symptoms.
For this study design to be optimal, a large sample size may be needed to demonstrate superiority of the broad-spectrum agent because even a small effect size relative to an active comparator may be relevant to clinical treatment and because the comparator drug may have a limited, but incomplete, effect on negative symptoms. Another limitation of this design is that if both drugs improved primary negative symptoms equivalently, the effects would not be demonstrated by the primary endpoint.
Operational Issues in Clinical Trial Design
Patient Selection Criteria
To demonstrate a drug's specific efficacy for negative symptoms, the study population must be clearly defined and selected, with investigators having specific criteria for patients to be recruited and regulatory agencies agreeing that the indication being sought for the drug is appropriately demonstrated in the patient population being studied. Defining the patient inclusion and exclusion criteria necessitates discussion of whether a trial should include only patients with primary negative symptoms. At issue here is a concern that in a clinical trial situation, it may not be feasible to absolutely distinguish patients with primary negative symptoms from those with secondary negative symptoms. Yet limiting the number of patients with secondary negative symptoms is important because their inclusion will increase the difficulty of identifying treatment effects specific for primary negative symptoms. Given the inadequacy of current treatment, it is of interest to demonstrate efficacy of a broad-spectrum drug in patients with severe predominant, persistent negative symptoms, as well as in those with less severe primary negative symptoms.
When establishing the inclusion criteria, consideration must be given to setting a minimum level for negative symptoms and a maximum level for positive symptoms to limit confounds of data interpretation. Setting the minimum level of negative symptoms too low could limit the ability to measure any change in these symptoms,12 but setting the minimum too high may limit the generalizability of the results. Similarly, if the minimum level of positive symptoms is set too low, generalizability of results may be limited, and, if set too high, it may be difficult to determine whether the drug had a direct or indirect effect on negative symptoms.
Trial Design and Duration
A trial designed to evaluate the effect of a broad-spectrum agent on negative symptoms should include a prerandomization and a postrandomization treatment period. The prerandomization phase establishes the baseline stability and severity of symptoms and confirms the absence of clinical benefit from current treatment. The duration of this phase should be long enough to permit demonstration of stable underlying clinical symptoms while keeping the total study duration at a reasonable length. The postrandomization phase is designed to determine the extent of the drug's effect on negative symptoms and whether the effect, if any, is maintained. The duration of this phase could be weeks, months, or years. To establish initial efficacy, the duration of active treatment should range from 4 to 12 weeks. To demonstrate longerterm efficacy, the study duration should be 6 months or longer.8
Analysis
The type of study design, choice of comparator, inclusion and exclusion criteria, study length, and primary audience for the data (e.g., regulators, clinicians, public health officials) should be considered when developing an analysis plan. Any analysis plan should also address how to handle potential confounding of secondary effects. If randomization or the effects of the comparator drug on the secondary negative symptoms do not adequately distribute the potential confounders, then appropriate covariate adjustment must be made.
Withdrawal rates and patterns also need to be considered. This is true for any clinical trial, but given the length of the study and the type of patient enrolled in a negative symptoms trial, the withdrawal pattern is especially critical. If the study is for a regulatory submission, agreement on the primary statistical model, imputation strategy, or sensitivity analyses should be sought before the trial begins.
The traditional approach to addressing patient withdrawals is to analyze the endpoint using the imputation method called last observation carried forward (LOCF). The LOCF method tends to be favored by regulators because it is conservative (biased toward the null hypothesis), particularly in placebo-controlled studies and when the therapeutic response is monotonic. With a study using an active comparator as the control, one must consider if LOCF is still appropriate. If early withdrawals are more likely with the comparator than with the broad-spectrum agent, then LOCF would not be appropriate, because the test statistic would be biased in favor of the investigational drug.
Imputation methods may not be necessary if an appropriate active comparator is chosen. The ideal comparator would have a withdrawal pattern similar to that of the new broad-spectrum drug so that the withdrawals could be considered missing at random (MAR).13 With MAR data, a repeated measures analysis or a random effects model might be the most appropriate analysis method. If MAR cannot be achieved, a mixture mode should be considered, at least as a secondary analysis.
Asenapine for Negative Symptoms—The Aphrodite Trials
The Aphrodite series of clinical trials have been designed to determine the efficacy and safety of asenapine, a broad-spectrum psychopharmacologic agent, compared with that of the atypical antipsychotic olanzapine, the active comparator, in patients with predominant, persistent negative symptoms of schizophrenia. Few other studies have been specifically designed for this subpopulation of patients with schizophrenia. Studies of amisulpride and other atypical antipsychotics have been conducted in patients with predominant negative symptoms.14-22 However, the Aphrodite trials are unique in that only patients whose symptoms were prospectively determined to be stable and persistent during a pretreatment period were enrolled.
Study Design
The Aphrodite studies comprise two phase III, randomized, doubleblind, flexible-dose trials, each followed by an extension study, that were conducted at centers in North and South America, Europe, South Africa, and Australia. Approximately 440 outpatients (220 patients in each treatment group) were enrolled in each 26-week study.
The 26-week trials included a screening visit, a 30-day stable observation period during which patients continued on their current antipsychotic drug, a baseline visit, and an active treatment period consisting of up to a 4-week antipsychotic switch period, during which current antipsychotic medication was tapered off and study drug started. This was followed by 22 weeks of monotherapy (Figure 2). For patients willing to continue longer, the initial 26-week study was extended for an additional 26 weeks of double-blind treatment during which the longterm safety and efficacy of asenapine for treatment of negative symptoms were evaluated. These extension studies continued until all patients either completed 52 weeks of treatment or withdrew from the study.
FIGURE 2 
Inclusion criteria were designed to recruit patients with predominant primary negative symptoms. To this end, patients were eligible to enroll in these trials if they had a documented current diagnosis of schizophrenia; a minimum Positive and Negative Syndrome Scale (PANSS) negative symptom subscale score of 20 at screening and baseline, with a minimum score of 4 (moderate) on at least three of the Marder factors for negative symptoms (blunted affect, emotional withdrawal, poor rapport, passive social withdrawal, lack of spontaneity, motor retardation, active social avoidance)23, and a PANSS positive symptom subscale score (delusions, conceptual disorganization, hallucinations, excitement, grandiosity, suspiciousness/persecution, hostility) less than the PANSS negative symptom subscale score (Marder factors)23 at screening and baseline. At screening, patients were to have demonstrated clinical symptom stability for both positive and negative symptoms for the preceding 5 months.
In line with the inclusion criteria, exclusion criteria were designed to exclude patients whose negative symptoms were likely to be secondary to other causes or for whom these secondary causes might have contributed substantially to their negative symptom pathology. Thus, patients were excluded from the trials if they had a score of 3 or more on the global Parkinsonian item of the Extrapyramidal Symptom Rating Scale-Abbreviated (ESRS-A), a score of 9 or higher on the Calgary Depression Scale for Schizophrenia (CDSS), a substance-induced psychotic disorder or current substance abuse, or a concurrent psychiatric disorder other than schizophrenia.
To prevent treatment bias, patients receiving olanzapine within the preceding 5 months whose negative symptoms did not respond adequately were excluded, as were patients receiving clozapine or antidepressants or mood stabilizers within the preceding 5 months.
The stability and persistence of negative symptoms were determined retrospectively by thorough review of available medical records, interviews with caregivers and medical professionals, and acquisition of medical and psychiatric history records. To ensure that patients enrolled in the trial had predominant negative symptoms, patients were excluded if they had a score of 4 or higher on 2 or more items on the PANSS positive symptom subscale. Enrollment of patients with secondary negative symptoms was limited by the ESRS-A, CDSS, and positive symptom inclusion and exclusion criteria, but no formal diagnosis of primary or secondary negative symptoms was made.
Patients received either asenapine 5 mg b.i.d. during the first week, then 5 or 10 mg b.i.d. thereafter, or olanzapine 10 mg daily during the first week, then 5-20 mg daily thereafter. Clinicians were allowed to increase, decrease, or hold prescribed doses constant at their discretion based on considerations of efficacy and safety. Because asenapine was administered as a fast-dissolving sublingual tablet and olanzapine as a film-coated tablet, a double-dummy technique was used to maintain the double blind, with active and placebo tablets being identical in appearance.
Olanzapine was chosen as the active comparator because it is effective for the positive symptoms of schizophrenia and causes few associated extrapyramidal symptoms. A placebo-only treatment group was not included in these trials because of the clinical and ethical considerations of treating patients with schizophrenia with placebo for long periods. A flexible dosing schedule for both asenapine and olanzapine was used to mimic clinical practice and minimize the occurrence of adverse effects in individual patients who might jeopardize the blinded study. In the extension studies, which also were double blind, patients continued on the medication (asenapine or olanzapine) taken during the initial 26-week studies, and dosages were maintained at the level used at the end of the first trial.
Efficacy and Safety Assessment Endpoints
The primary efficacy endpoint in the Aphrodite trials was the total score on the 16-item Negative Symptom Assessment (NSA-16) scale.24 The NSA-16 provides a comprehensive evaluation of negative symptoms, and raters from different cultures who speak different languages can be trained quickly to use it reliably.25 Using the NSA-16 in a clinical trial, rather than a measure such as the Brief Psychiatric Rating Scale withdrawal/retardation subscale, which has less sensitivity, permits a decrease in the sample size needed to demonstrate a clinically meaningful effect and increases the level of information available about the nature of the effect on negative symptoms.11 NSA-16 scores have been correlated with scores on functional outcome scales,26 providing support that this scale measures a clinically important functional construct for the patient.
A number of secondary rating scales, including the PANSS total score, PANSS factor scores,23 and PANSS subscale scores (positive, negative, and general psychopathology), were used in the Aphrodite trials to further interpret the primary results from the NSA-16.
Other scales were used to help clarify whether the effects observed with the primary ratings were caused by primary or secondary effects on negative symptoms and to help establish whether changes in negative symptoms corresponded with changes in measures of functioning and long-term outcome. The Clinical Global Impression-Severity of Illness and Clinical Global Impression-Improvement scales were used to provide additional clinical efficacy support for the NSA-16 results. The CDSS and the Hamilton Anxiety Scale were used to assess potential changes in depression and anxiety, respectively, symptoms that might confound interpretation of treatment effects on negative symptoms. The subscales of the Central Nervous System Vital Signs Neurocognitive Test Battery were used to assess changes in cognitive symptoms, and the ESRS-A was used to assess changes in extrapyramidal symptoms, two other potential confounders to data interpretation.
Several measures of clinical outcome were also used, including the patient-rated Quality of Life Enjoyment and Satisfaction Questionnaire and the Personal Evaluations of Transitions and Treatment.
To assess safety, all observed or volunteered adverse events, as well as abnormal laboratory results, were recorded.
Statistical Analysis
The primary analysis used data from the intent-to-treat population, which included all patients randomly assigned to treatment who had at least one post-baseline NSA-16 score. Missing observations were imputed using the LOCF method. The primary efficacy endpoint, the change in NSA-16 scores from baseline, was evaluated by fitting a fixed-effects analysis of covariance to the data, allowing for variability caused by center and treatment, with covariates of baseline score and duration of predominant negative symptoms. Descriptive statistics were used to report adverse effects and changes in extrapyramidal symptoms, laboratory values, and clinical monitoring results.
The study employed an adaptive design in which an interim analysis was used to reassess the assumptions of effect size and variance, correcting the sample size if needed. The adaptive design used the methods of Cui et al.27 to prevent inflation of type I error should the sample size need adjustment. The interim analyses were conducted by an external statistician who informed the study team whether enrollment needed to be increased to identify a treatment effect. This interim analysis did not include a recommendation to stop for futility. Therefore, a signal to the study team that the sample size was adequate could have been given if the study was suggesting the hypothesized efficacy or if the study suggested futility.
Conclusions
Although progress has been made in managing the negative symptoms of schizophrenia, effective treatment for these symptoms remains a tremendous unmet need. Additional work is needed to develop drugs with a direct effect on negative symptoms and to design trials that can accurately assess the effect of new broad-spectrum agents on these symptoms. Although it may not be possible to completely distinguish drug-related improvements in negative symptoms from improvements in positive symptoms, depression, or adverse effects, careful attention to designing clinical trials for evaluating treatments for negative symptoms may lessen the impact of such confounding factors.
The Aphrodite trials, which were designed to evaluate the efficacy and safety of asenapine in patients with predominant, persistent negative symptoms of schizophrenia, address some of the difficulties inherent in assessing drug therapy for negative symptoms. Results from these trials will not only determine the efficacy and safety of asenapine in patients with negative symptoms but will provide valuable data on this understudied patient population, as well as contribute to our knowledge of good study design.
Acknowledgments
Ideas presented in this paper were summarized at a presentation at the International Society for CNS Clinical Trials in February 2006. Special thanks are given to Jacquelyn Wilson and Pilar Carzola and the many colleagues at Pfizer, Organon, and Quintiles who helped conduct these trials and to the hundreds of patients who agreed to participate in them.
REFERENCES
1. Andreasen NC. Improvement of negative symptoms: concepts, definition and assessment. Int Clin Psychopharmacol. 1997;12(Suppl 2):S7-S10.
2. Carpenter WT, Jr., Heinrichs DW, Wagman AM. Deficit and nondeficit forms of schizophrenia: the concept. Am J Psychiatry. 1988;145:578-583.
3. Erhart SM, Marder SR, Carpenter WT. Treatment of schizophrenia negative symptoms: future prospects. Schizophr Bull. 2006;32:234-237.
4. Alphs L. An industry perspective on the NIMH consensus statement on negative symptoms. Schizophr Bull. 2006;32:225-230.
5. Javitt DC. Management of negative symptoms of schizophrenia. Curr Psychiatry Rep. 2001;3:413-417.
6. Flaum M, Andreasen N. The reliability of distinguishing primary versus secondary negative symptoms. Compr Psychiatry. 1995;36:421-427.
7. Moller HJ. Management of the negative symptoms of schizophrenia: new treatment options. CNS Drugs. 2003;17:793-823.
8. Kirkpatrick B, Fenton WS, Carpenter WT, Jr., Marder SR. The NIMH-MATRICS consensus statement on negative symptoms. Schizophr Bull. 2006;32:214-219.
9. Arango C, Buchanan RW, Kirkpatrick B, Carpenter WT. The deficit syndrome in schizophrenia: implications for the treatment of negative symptoms. Eur Psychiatry. 2004;19:21-26.
10. Moller HJ. Atypical neuroleptics: a new approach in the treatment of negative symptoms. Eur Arch Psychiatry Clin Neurosci. 1999;249(Suppl 4):99-107.
11. Eckert SL, Diamond PM, Miller AL, et al. A comparison of instrument sensitivity to negative symptom change. Psychiatry Res. 1996;63:67-75.
12. Buchanan RW, Davis M, Goff D, et al. A summary of the FDA-NIMH-MATRICS workshop on clinical trial design for neurocognitive drugs for schizophrenia. Schizophr Bull. 2005;31(1):5-19.
13. Rubin DB. Multiple Imputations for Non-response in Surveys. New York, NY: John Wiley; 1987.
14. Olie JP,Spina E, Murray S, Yang R. Ziprasidone and amisulpride effectively treat negative symptoms of schizophrenia: results of a 12-week, double-blind study. Int Clin Psychopharmacol. 2006;21:143-151.
15. Moller HJ. Amisulpride: efficacy in the management of chronic patients with predominant negative symptoms of schizophrenia. Eur Arch Psychiatry Clin Neurosci. 2001;251:217-224.
16. Danion JM, Rein W, Fleurot O. Improvement of schizophrenic patients with primary negative symptoms treated with amisulpride. Amisulpride Study Group. Am J Psychiatry. 1999;156:610-616.
17. Paillere-Martinot ML, Lecrubier Y, Martinot JL, Aubin F. Improvement of some schizophrenic deficit symptoms with low doses of amisulpride. Am J Psychiatry. 1995;152:130-134.
18. Alvarez E, Ciudad A, Olivares JM, Bousono M, Gomez JC. A randomized, 1-year follow-up study of olanzapine and risperidone in the treatment of negative symptoms in outpatients with schizophrenia. J Clin Psychopharmacol. 2006;26:238-249.
19. Riedel M, Muller N, Strassnig M, et al. Quetiapine has equivalent efficacy and superior tolerability to risperidone in the treatment of schizophrenia with predominantly negative symptoms. Eur Arch Psychiatry Clin Neurosci. 2005;255:432-437.
20. Speller JC, Barnes TR, Curson DA, Pantelis C, Alberts JL. One-year, low-dose neuroleptic study of in-patients with chronic schizophrenia characterised by persistent negative symptoms. Amisulpride v. haloperidol. Br J Psychiatry. 1997;171:564-568.
21. Loo H, Poirier-Littre MF, Theron M, Rein W, Fleurot O. Amisulpride versus placebo in the medium-term treatment of the negative symptoms of schizophrenia. Br J Psychiatry. 1997;170:18-22.
22. Boyer P, Lecrubier Y, Puech AJ, Dewailly J, Aubin F. Treatment of negative symptoms in schizophrenia with amisulpride. Br J Psychiatry. 1995;166:68-72.
23. Marder SR, Davis JM, Chouinard G. The effects of risperidone on the five dimensions of schizophrenia derived by factor analysis: combined results of the North American trials. J Clin Psychiatry. 1997;58:538-546.
24. Alphs LD, Summerfelt A, Lann H, Muller RJ. The negative symptom assessment: a new instrument to assess negative symptoms of schizophrenia. Psychopharmacol Bull. 1989;25:159-163.
25. Alphs L, Daniel DG, Velligan D, et al. Training for assessment of negative symptoms of schizophrenia across languages and cultures utilizing the Negative Symptoms Assessment Scale (NSA). Presented at American Psychiatric Association 159th Annual Meeting, May 20—25, 2006,Toronto, ON, Canada.
26. Morlock R, Velligan D, Wang M, et al. Association between changes on the Negative Symptom Assessment-16 scale (NSA-16) and functional outcomes in schizophrenia. Presented at European College of Neuropsychopharmacology, September 16—20, 2006, Paris, France.
27. Cui L, Hung HM,Wang SJ. Modification of sample size in group sequential clinical trials. Biometrics. 1999;55:853-857.
