William T. Carpenter Jr: 35 Years of Clinical Trials

Abstract

William T. Carpenter Jr has had a major impact on the design and conduct of clinical trials in schizophrenia. His contributions range from the decisive evaluation of the efficacy of hemodialysis to the development of novel approaches to evaluate new treatments for cognitive impairments and negative symptoms. He has developed innovative dosage reduction strategies. He has led efforts to focus drug development on those illness components that are not responsive to antipsychotic treatment. He has emphasized throughout his career the use of translational science to provide the conceptual framework for clinical trial interventions. This article reviews highlights of his many contributions, with an emphasis on 3 areas: (1) dosage reduction studies; (2) the use of the domains of psychopathology to identify drug development targets; and (3) the use of translational science to guide new drug development.

Introduction

William T. Carpenter has dedicated his research career to the development and evaluation of safer and more effective drugs and treatment approaches for people with schizophrenia. These values are clearly evident in his initial published clinical trial in which he conducted the largest controlled study to date to examine the efficacy of hemodialysis for the treatment of schizophrenia.¹ In an illness in which available treatments offer limited benefit, the initial reports of marked improvement in symptoms and function in people receiving hemodialysis generated substantial enthusiasm. However, the majority of reports were based on open-label, uncontrolled observations. In light of the costs associated with hemodialysis, and, more importantly the unique risk/benefit ratio of the intervention, a well-designed, controlled study was imperative. Carpenter and colleagues conclusively demonstrated the lack of differential efficacy compared with sham dialysis, which saved patients undue hardships associated with an ineffective treatment.¹ The study launched a career of clinical trial research dedicated to the development of more effective and safer interventions. In the current article, 3 areas of treatment development are reviewed: (1) dosage reduction studies; (2) the use of the domains of psychopathology to identify drug development targets; and (3) the use of translational science to guide new drug development.

Dosage Reduction Strategies

The first antipsychotic, chlorpromazine, was discovered in the 1950s. In the 1960s and 1970s, multiple other first generation antipsychotics (FGAs), including fluphenazine and haloperidol, were introduced. FGAs significantly reduce the positive symptoms of schizophrenia, including hallucinations, delusions, and positive formal thought disorder, and are effective for the prevention of relapse and maintenance treatment.² However, the early history of FGAs was characterized by the use of extremely high doses of these agents (eg, rapid neuroleptization). The use of high or excessively high doses of FGAs is associated with severe motor side effects, including tardive dyskinesia. In addition, high dose FGA use may cause secondary cognitive impairments and negative symptoms. The combination of severe motor effects and secondary symptoms are hypothesized to contribute to impaired function and poor quality of life.

In response to the growing appreciation of the magnitude of the problems associated with the use of high dose antipsychotic treatment, a number of strategies were developed to reduce antipsychotic exposure, including drug holidays and reduced dose of depot formulations of fluphenazine.^3,4 Over the course of a 15-year period, Carpenter and colleagues were involved in the development of 3 different approaches to dosage reduction. The first approach was the targeted medication approach for the maintenance treatment of schizophrenia. This approach was based on the following observations: (1) positive symptoms have an episodic nature; (2) in between exacerbations, FGAs serve primarily a prophylactic role; and (3) symptoms and behaviors can be used to identify the onset of symptom exacerbation, which allows for the timely re-introduction of antipsychotics.^5,6 In the Targeted Medication approach, patients are initially stabilized on the lowest possible dose of an antipsychotic. They are then taken off their antipsychotic medication and monitored closely for emerging prodromal symptoms, which have been previously determined to be characteristic of their symptom exacerbations or relapses.⁵ Upon detection of prodromal symptoms, the patients are restarted on their antipsychotic until the resolution of these symptoms: at which time the antipsychotic is discontinued. In the initial, uncontrolled, open-label study of this approach, Herz and colleagues demonstrated the feasibility of the Targeted Medication approach.⁷ In 2 subsequent comparative trials, Carpenter and colleagues examined the efficacy and safety of the approach. In the first study, they found that the Targeted Medication approach achieved significant dosage reduction in the context of comparable outcomes.⁶ However, in the single-blind, controlled replication study, continuous maintenance medication was superior to the Targeted Medication approach for the prevention of relapses and hospitalizations and was associated with better functional outcome.⁸ In the discussion of the results, Carpenter and colleagues noted that the Targeted Medication approach may remain a viable alternative for a select group of patients, who are (1) characterized by an episodic course, with clearcut, gradual onset of prodromal symptoms and (2) able to maintain insight for the need of medications even when experiencing a symptom exacerbation.⁸ In addition, patients whose psychotic symptoms may be due to nonpsychiatric causes (eg, substance abuse; steroid psychosis) or who are not willing to take antipsychotics on a continuous basis may be good candidates for this approach.

The second dosage reduction approach was based on the use of long-acting injectable or depot FGA preparations, especially fluphenazine. In previous studies of depot fluphenazine, dosage reduction was achieved through lowering the dose of the antipsychotic but maintaining the regular schedule of injections.^3,4 This approach achieved dose reduction but was associated with an increased incidence of symptom exacerbations and limited effects on functional outcome. Carpenter and colleagues proposed an alternative approach, which was based on the following observations: The majority of patients who are discontinued from LAIs (1) remain clinically stable for up to 6 weeks following their last injection^9,10; (2) continue to have detectable levels of fluphenazine up to 6 months after their last injection¹⁰; and (3) continue to have significant D2 receptor occupancy for at least 4 months after their last injection.¹¹ These observations led to the hypothesis that significant dose reduction could be achieved through lengthening of the interval between injections from 2 to 6 weeks, which would reduce rates of adverse effects, including tardive dyskinesia. To examine the feasibility of less frequent administration, Carpenter and colleagues conducted a double-blind, randomized clinical trial comparing 54 weeks of standard fluphenazine decanoate treatment (25mg every 2 weeks) with experimental dose reduction (25mg every 6 weeks).¹² The study design incorporated the Targeted Medication approach, in that participants were assessed for characteristic prodromal symptoms, and if they began to exhibit such symptoms, an oral antipsychotic was added to their treatment regimen until the symptom exacerbation subsided. They found that the every 6 week dosage schedule led to significant dose reduction, in the absence of significant differences in exacerbations or hospitalizations. However, there were no significant differences in adverse effects, including extrapyramidal side effects, which may reflect the use of a moderate dose of fluphenazine decanoate (25 mgs) for the 2 arms.

The third dosage reduction approach was also based on the early detection of prodromal symptoms. There are a number of mechanisms through which antipsychotic doses are increased over time. A common reason is the treatment of transient fluctuations in positive symptom intensity with an increase in antipsychotic dose. Positive symptom fluctuations are relatively common in people with schizophrenia, and the repeated use of increased antipsychotic doses, without downward dosage adjustment after re-stabilization, will eventually lead to excessive antipsychotic doses. An alternative approach is to use benzodiazepines to treat these exacerbations. The use of benzodiazepines for the treatment of prodromal symptoms is based on the observations that (1) positive symptom fluctuations are commonly related to acute stressors; (2) mild psychotic symptoms are responsive to antianxiety treatment; and (3) benzodiazepine drugs diminish dopaminergic neurotransmission through gamma-aminobutyric acidergic feedback mechanisms.¹³ In a pilot study, open-label diazepam was found to be effective in the prevention of relapse in antipsychotic-free patients with schizophrenia.¹⁴ In a formal evaluation of the use of diazepam to treat prodromal symptoms, Carpenter and colleagues conducted a double-blind study, in which participants were abruptly withdrawn from their antipsychotic and were randomized to treatment with diazepam, fluphenazine, or placebo.¹⁵ Diazepam was significantly more effective than placebo for the prevention of relapse and numerically but not significantly better than fluphenazine.

In combination, these series of studies have provided the clinician with a number of novel approaches for the treatment of prodromal symptoms and maintenance treatment. These approaches may be modified to meet the specific needs of the individual patient and provide the clinician with an important opportunity to engage in discussions on how best to use antipsychotic and other medications to treat their illness.

The Domains of Psychopathology

In 1974, Strauss, Carpenter, and Bartko proposed that the symptoms of schizophrenia could be separated into 3 complexes: (1) positive symptoms, ie, hallucinations, delusions, and formal thought disorder; (2) negative symptoms; and (3) disorders of relating (reviewed by Buchanan and Carpenter¹⁶). The proposal rekindled interest in negative symptoms and led to multiple negative symptom models, including Negative and Positive schizophrenia, Type 1 vs Type 2 schizophrenia, and Deficit vs Nondeficit schizophrenia (reviewed by Buchanan and Carpenter¹⁶). In 1989, Carpenter and Buchanan introduced the concept of the Domains of Psychopathology in which a domain of psychopathology (1) is an empirically validated group of similar signs and/or symptoms; (2) is a manifestation of a distinct pathophysiological process; and (3) can be distinguished from other groups of signs and/or symptoms occurring in the same syndrome.¹⁶ Five domains were defined for specific study: (1) hallucinations and delusions; (2) cognitive impairment, subdivided to include positive formal thought disorder and attentional/information processing impairments; (3) primary and enduring negative symptoms, ie, deficit symptoms; (4) dissociative affective processes; and (5) neurological impairment. The deconstruction of schizophrenia into individual illness dimensions led to the proposal to use these components to define new clinical targets for drug development targets and was an important impetus for the National Institute of Mental Health–Measurement and Treatment Research to Improve Cognition in Schizophrenia/Treatment Units for Research in Neurocognition in Schizophrenia initiatives.¹⁷

Two domains became a particular focus of Carpenter and colleagues: cognitive impairments and negative symptoms. These 2 domains are critically important determinants of long-term functioning in people with schizophrenia.^18–20 Antipsychotics have limited benefits for these domains, which has led to the investigation of add-on medications for the treatment of cognitive impairments and negative symptoms. A number of small sample size studies suggested that agents that act at the glycine site of the glutamatergic N-methyl-d-aspartate receptor might be effective in treating negative symptoms (reviewed by Buchanan et al²¹). In order to definitively address the potential utility of these agents for negative symptoms and to examine whether they may also have benefit for cognitive impairments, Carpenter and colleagues organized the Cognitive and Negative Symptoms in Schizophrenia Trial, which was a 4-site, placebo-controlled randomized clinical trial of glycine and d-cycloserine.²¹ The study was easily the largest study conducted to date, but, unfortunately, failed to demonstrate a significant benefit of either drug, compared with placebo, for negative symptoms or cognitive impairments.

Translational Science and Drug Development

Cognitive impairments and negative symptoms represent major unmet therapeutic needs in schizophrenia.^22,23 New drug discovery is necessary to address the treatment of these domains, but animal models, with predictive validity, have not been established for cognitive impairments and negative symptoms. Moreover, there may be a shared pathophysiology between these 2 domains: There is a large body of literature that suggests that people with marked negative symptoms are more cognitively impaired than those without these symptoms.²⁴ The observed association between these 2 domains leads to the question: Are cognitive impairments and negative symptoms independent targets for drug development? In order to address these issues, Carpenter and colleagues have developed an innovative approach to evaluate promising new agents. The approach is based on the use of analogous behavioral and biomarker assessments across the following 3 platforms: (1) a preclinical platform based on the maternal stress rat model for schizophrenia; (2) a novel human platform based on biological relatives of schizophrenia probands selected according to phenotype criteria; and (3) a clinical trials platform based on subject selection criteria that assure schizophrenia subjects manifest the relevant pathological domains and phenotypic expression. The preclinical and human platforms have not previously been used in drug development. The evaluation of drug effects across these 3 platforms will enable the determination of whether phenotypic effects in the preclinical rat or human model predict efficacy in people with schizophrenia. Furthermore, the concurrent evaluation of drug effects on cognitive and negative symptom biomarkers and endpoints will enable the determination of whether negative symptom and cognitive impairment pathology represent one or 2 pathways for drug discovery.

Conclusion

William T. Carpenter has had a long and distinguished career in which he has made multiple contributions to the pharmacology of schizophrenia. His work has been characterized by the creative and innovative integration of basic science, clinical science, and treatment research and is an ongoing example of the benefits of translational science conducted in a collaborative, multidisciplinary framework. However, perhaps his greater contribution has been the development of the next generation of clinical trial researchers; individuals with expertise in clinical trial design, neuroimaging, neuropsychology, and psychophysiology. This next generation will be responsible for ensuring his legacy and fulfilling his vision of safer, more effective pharmacological treatments for people with schizophrenia.