Advertisement

Modern Psychopharmacology and Psychiatric Treatment

  • Ross J. Baldessarini
Chapter

Abstract

The need for effective treatments of psychiatric disorders is indicated by the high prevalence of many of these disorders [47] , particularly substance-use, anxiety, and mood disorders, as well as their high burden of direct and indirect costs to society (Tables 1.1 and 1.2).

Keywords

Psychotropic Drug Major Psychiatric Disorder Biological Hypothesis Statistical Modeling Technique Major Mood Disorder 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

In the fields of observation chance favors only the prepared mind Louis Pasteur

The need for effective treatments of psychiatric disorders is indicated by the high prevalence of many of these disorders [47], particularly substance-use, anxiety, and mood disorders, as well as their high burden of direct and indirect costs to society (Tables 1.1 and 1.2). Psychiatric, substance-abuse, and primary brain disorders account for approximately 13% of the global disease burden (years of healthy life lost due to early death or disability). Depression alone, even excluding bipolar disorders, is the third leading contributor to the worldwide disability. Associated suicides number 1.5 million/year, and attempts are estimated at over 20 million/year [24]. Throughout the recorded history of medicine, efforts have been made to utilize chemical or medicinal means to modify abnormal behavior and emotional pain. Alcohol and opiates have been used for centuries not only by physicians and healers but also spontaneously for their soothing or mind-altering effects. Stimulant and hallucinogenic plant products also have been a part of folk practices for centuries. More recently, man has applied modern technology, first to “rediscovering” and purifying many natural products, later to synthesizing and manufacturing their active principles or structural variants with desired properties. Throughout the discussion that follows, the classes of chemicals used for their psychotropic effects (altering feelings, thinking, or behavior) are referred to by the somewhat arbitrary terms antipsychotic, antimanic or mood-stabilizing, antidepressant, and antianxiety agents. This system of terminology grows out of the allopathic tradition of modern scientific medicine, which treats with drugs producing effects opposite, or antagonistic to the signs and symptoms of a given illness.
Table 1.1

Needs and markets for psychotropic drugs

Prevalence of psychiatric illnesses (United States)

Substance abuse (alcohol and drug) (27%)

Anxiety disorders (25%)

Major depression and dysthymia (24%)

Bipolar disorders (5%)

Antisocial personality (3.5%)

Schizophrenia and related psychoses (1%)

Two or more disorders (27%)

Professionally treated (42%)

Illness costs

Mental illness + dementias (30%)

Cardiovascular (18%)

Brain + spinal injuries (13%)

Cancers (12%)

Mood disorders (11%)

Acquired immunodeficiency (HIV/AIDS) (8%)

Arthritis (5%)

Schizophrenia (4%)

Ranking by sales income

1. Psychotropics

2. Cancer chemotherapies

3. Anticholesterol agents

4. Diabetes drugs

5. Gastric proton-pump inhibitors

6. Antihypertensives

7. Analgesics

Table 1.2

Prevalence of psychiatric and neurological disorders, sex risk, and disease burden statistics (Europe, 2010 [87])

Disorder

Prevalence (%) (median (IQR))

Female/male risk ratio

Disease burden (DALY/10,000)

Anxiety disorders

14.6 (6.80–20.4)

2.5

23.0

Sleep disorders

10.8 (8.10–16.8)

1.2

9.4

Juvenile behavioral disorders

9.60 (1.70–27.0)

1/2.4

Major depression

6.30 (3.54–8.18)

2.3

104

Somatoform disorders

5.60 (1.87–6.93)

2.1

Dementias

5.40 (0.30–1.00)

1.6

53.7

Substance abuse (alcohol + drug)

4.38 (0.90–6.05)

1/3.3

67.1

Personality disorders

1.30 (1.30–0.14)

1/2.9

Psychotic disorders

1.10 (0.22–1.21)

1/1.2

15.3

Mental retardation

1.00 (0.40–1.40)

1/1.2

0.10

Bipolar disorder

0.80 (0.80–1.03)

1.2

17.5

Eating disorders

0.62 (0.01–0.63)

6.2

Neurological disorders

1.2

71.5

DALY: disease-associated life-years lost

Reciprocal sex risk ratios are greater in males

The modern era of psychopharmacology can be dated from 1949, when the antimanic effects of the lithium ion were discovered, or 1952, when the psychotropic and antiadrenergic effects of reserpine were investigated, and when the special properties of chlorpromazine began to be recognized. The antidepressant monoamineoxidase (MAO) inhibitor iproniazid also was introduced in the early 1950s, and in the late 1950s, the “tricyclic” antidepressant imipramine was introduced. Use of the anxiolytic-sedative meprobamate began in 1954, and the first benzodiazepine, chlordiazepoxide, was being developed before 1960 as an antianxiety agent. By the end of the 1950s, general medicine and psychiatry had therapeutic agents available for the psychotic and major mood disorders—including schizophrenia, mania, and severe depression—and the anxiety disorders (formerly “neuroses”). Remarkably few fundamentally new kinds of psychiatrically therapeutic agents have been developed since that time. Instead, the past half-century has been marked by an accumulation of structural analogues of earlier agents or chemically dissimilar drugs with similar actions and clinical effects—all with adverse effects, some similar and others new. Nevertheless, there have been important advances in understanding the biological and clinical actions of the available psychotropic drugs and their appropriate clinical use. Currently available psychotropic drugs are the leading pharmaceutical products of all kinds, based on their annual sales (Table 1.1).

Psychotropic drugs that are currently available are summarized by generic and corresponding original or prominent brand names in Table 1.3. A recent market analysis of leading psychotropic drugs based on annual sales in the United States [30] indicated the following ranking: (1) alprazolam (Xanax ®), (2) zolpidem (Ambien ®), (3) S-citalopram (Lexapro ®), (4) lorazapam (Ativan ®), (5) gabapentin (Neurontin ®), (6) clonazepam (Klonopin ®), (7) sertraline (Zoloft ®), (8) duloxetine (Cymbalta ®), (9) amphetamines (Adderall ®), (10) venlafaxine (Effexor ®), (11) quetiapine (Seroquel ®), (12) trazodone (Desyrel ®), (13) diazepam (Valium ®), (14) R,S-citalopram (Celexa ®), and (15) fluoxetine (Prozac ®).
Table 1.3

Common psychotropic drugs: generic–brand names

Antidepressants (23)

Antipsychotics (19)

Mood-stabilizers (5)

Atomoxetine (Strattera)

Aripiprazole–Abilify

Carbamazepine–Tegretol

Amitriptyline–Elavil

Asenapine–Saphris

Divalproex–Depakote

Bupropion–Wellbutrin

Chlorpromazine–Thorazine

Oxcarbamazepine–Trileptal

Citalopram–Celexa

Clozapine–Clozaril

Lamotrigine–Lamictal

S-Citalopram–Lexapro

Fluphenazine–Prolixin

Lithium carbonate–Lithobid

Clomipramine–Anafranil

Haloperidol–Haldol

 

Desipramine–Norpramin

Iloperidone–Fanapt

Desvenlafaxine–Pristiq

Loxapine–Loxitane

Doxepin–Sinequan

Lurasidone–Latuda

Duloxetine–Cymbalta

Mesoridazine–Serentil

Fluoxetine–Prozac

Olanzapine–Zyprexa

Fluvoxamine–Luvox

Paliperidone–Invega

Imipramine–Tofranil

Perphenazine–Trilafon

Mirtazapine–Remeron

Quetiapine–Seroquel

Nortriptyline–Pamelor

Risperidone–Risperidal

Paroxetine–Paxil

Thiothixene–Navane

Phenelzine–Nardil

Thioridazine–Mellaril

Selegiline–Emsam

Trifluperazine–Stelazine

Sertraline–Zoloft

Ziprasidone–Geodon

Tranylcypromine–Parnate

  

Trazodone–Desyrel

  

Venlafaxine–Effexor

  

Vilazodone–Viibryd

  

Agents commonly employed in the United States. Atomoxetine is approved for attention disorders; oxcarbamazepine is used off-label for bipolar disorder

The impact of modern psychopharmaceuticals on the practice of psychiatry since the 1950s has been compared to the impact of the antibiotics on general medicine since the 1940s. Quantitatively, the utilization of chlorpromazine compares well with that of penicillin: in the first decade of its availability this antipsychotic drug was given to approximately 50 million patients throughout the world, and some 10,000 scientific papers were written about it [72]. At the present time, psychotropic drugs not only are among the leading pharmaceuticals of all types, but several command markets of several billion US dollars/year. These facts underscore the revolutionary impact of these drugs on modern psychiatry.

Prior to the 1950s, most severely disturbed psychiatric patients were managed in relatively secluded public or private institutions, usually with locked doors, barred windows, and other physical restraints. The few medical means of managing their symptoms included use of barbiturates, bromides, opioids, and anticholinergic drugs such as scopolamine for sedation. Other treatments included soothing baths and wet-packs, as well as “shock” treatments with insulin, atropine, or convulsant drugs, and later electrically induced convulsions, along with neurosurgical techniques including prefrontal leucotomy. Since then, most of these forms of treatment, except for electroconvulsive treatment (ECT) have virtually disappeared. Many locked doors have opened, except for severely disturbed, aggressive or suicidal patients, and both patients and psychiatric facilities have been returned to “the community,” to general hospitals, open-door day-treatment centers, and to hospital-based or free-standing outpatient clinics and private offices. However, to conclude that modern psychotropic drugs have been solely responsible for these revolutionary changes would be a gross exaggeration. In the same period, partly independent changes in the clinical management of psychiatric patients also were beginning. These included use of group and milieu techniques to complement individual psychotherapy, greater appreciation of the untoward regressive effects of institutions on behavior, and a strongly increased social consciousness throughout medicine and particularly in community-based psychiatry. A fair conclusion would be that social and administrative changes and the new drugs had mutually facilitating and enabling interactions, which resulted in a melioristic trend toward progress and change.

Observations that underscore the important impact on hospital practice associated with the new antipsychotic, antimanic, antidepressant, and antianxiety drugs introduced in and following the 1950s include the observation that in the United States the number of patients hospitalized in public mental institutions reached a peak of approximately 500,000 in the 1950s, with an initially rapid and now slower downward trend to less than 30,000 currently, despite a steady increase in the general population. This change has resulted not only from the beneficial effects of modern psychotropic drugs but also from policy decisions to alter the pattern of mental health-care delivery, notably including decisions to reduce the number of available beds in most public psychiatric institutions. Ironically, rates of new admissions and of readmissions have increased over time, particularly among the very young and very old, despite a misleading decline in the prevalence of psychiatric hospitalization, as length-of-stay has declined markedly, largely through administrative demands driven by hopes of cost-containment [46]. In addition, there has been a major shift in the proportion of hospitalized seriously mentally ill persons in hospitals vs. in jails and prisons, whose numbers currently exceed those of the hospitalized mentally ill in the 1950s ([38], Fig. 1.1). A large proportion of patients formerly held in hospitals for many months now attain outpatient status within weeks or even days, owing to current philosophies and systems of care combined with the effects of modern chemotherapy. However, their aftercare, quality-of-life, and functional levels vary markedly among locales, diagnoses, and individuals [31].
Fig. 1.1

Institutional trends in the United States, 1930–2000. Hospitalization in public mental hospitals peaked in the 1940s and 1950s and declined thereafter as numbers of persons in prisons increased, many of whom were severely mentally ill. Adapted from Harcourt [38]

In general, a number of serious problems remain despite striking improvements in the clinical treatment of patients with severe psychiatric disorders. Whereas many acute episodes of such disorders can be interrupted and shortened with modern treatments, and highly disturbed behavior is relatively infrequent, even in public mental institutions, the available chemotherapies have severe shortcomings, including general limitations of efficacy, particularly incomplete evidence of long-term preventive or prophylactic effectiveness, and incomplete or intermittent adherence to prescribed regimens, as well as sometimes medically significant adverse or toxic effects. Many patients with severe psychotic, mood, or anxiety disorders respond only marginally to available treatments, despite a tendency to continue their use or even to add increasingly complex combinations of drugs of largely untested benefit and safety.

An additional complication for the medical treatment of psychiatric disorders is the uncertainty or lack of specificity of diagnosis and classification that is characteristic of psychiatric illnesses. Diagnosis has become more important than ever, especially in order to optimize matching of clinical conditions, choice of treatment, and chances of a beneficial response with tolerable side effects. The advent of modern psychopharmacologic treatments, perhaps more than any other single factor, contributed to a vigorous renewal of interest in nosology and its ally, psychiatric epidemiology. There is also renewed interest in classical, descriptive psychopathology, which American psychiatry has tended to ignore more than other cultures. Nevertheless, etiology, whether biological or psychological, remains unknown for most psychiatric disorders, and classification or diagnosis rests on descriptions of clinical features of syndromes, as well as on their course or natural history and outcomes, familial associations, and to some extent, responses to treatment. Despite the fundamentally unsatisfactory basis of psychiatric nosology and the inescapable contributions of individuality even to the most classic syndromes, diagnosis continues to gain objectivity, coherence, and reliability. Associations between specific clinical syndromes and predictable responses to psychotropic agents continue to support efforts to improve psychiatric classifications. Indeed, examples of clinical syndromes that became much more widely recognized with the development of effective treatments for them include major depression, bipolar disorder, panic-agoraphobia, obsessive-compulsive disorder, and others [3, 36, 69].

The current clinical and academic standard for psychiatric diagnosis in the United States and many other countries is the Diagnostic and Statistical Manual (DSM) of the American Psychiatric Association, now entering its fifth revision. The World Health Organization sponsors a psychiatric component of its International Classification of Diseases (ICD); now in its tenth revision, it is widely used internationally as well as by many government agencies and insurance companies in the United States. Additional systems of diagnosis and gathering of clinical information have been developed for specialized research purposes, and clinical testing of novel psychotropic drugs depends heavily on rating scales that aim to capture the severity of characteristic clinical symptoms of common psychiatric disorders. However, even the best available systems of diagnostic classification and symptom rating for clinical or research applications require gathering and interpretation of data by and from individual persons, so that subjective and idiosyncratic, as well as culture-bound, elements of their application can scarcely be avoided.

Development of New Psychotropic Drugs

Drug Discovery

Most available psychopharmaceuticals have been discovered and introduced in one of three basic ways: (a) rediscovering and exploiting folk usage of natural products, usually with isolation of active principles and synthesis of similar molecules with comparable effects (e.g., reserpine, opioids, and centrally active sympathomimetics); (b) serendipitous observation that an agent developed for another purpose has a desirable but unexpected clinical effect (examples include chlorpromazine, haloperidol, iproniazid, imipramine, meprobamate, and trazodone); or (c) synthesis and functional screening of structural analogues of known drugs or of novel compounds in search of behavioral or molecular effects similar to those of known agents (examples include the piperidine phenothiazines, thioxanthenes, butyrophenones and diphenylbutylpiperidines, modern antipsychotic drugs modeled after the structure or activities of clozapine, and the serotonin-reuptake inhibitors and other modern antidepressants) [13]. These methods are summarized in Table 1.4, and examples of serendipity in the discovery of novel psychotropic drugs are provided in Table 1.5 [57].
Table 1.4

Characteristics of psychotropic drug development

Use and later purification and synthesis of natural products (e.g., alcohol, amphetamines, cocaine, opiates, reserpine)

Serendipity combined with research uncovers leads (e.g., buspirone, chlordiazepoxide, chlorpromazine, clozapine, haloperidol, improniazid, imipramine)

Synthesis and screening of chemical analogs (e.g., benzodiazepines, most tricyclic antidepressants and antipsychotics)

Partial knowledge of pharmacodynamics supports functional screening (e.g., haloperidol, serotonin-reuptake inhibitors)

Cloning of genes for novel proteins and rational, computer-assisted, design of small molecules (functions not well defined; limited success for psychotropics)

Overall progress severely limited by lack of pathophysiology or etiology

Table 1.5

Examples of serendipity in psychopharmacological drug discovery

Agent

Background

Lithium carbonate

Not so good for gout

Amphetamines

Derived from ephedrine in Chinese herbal ma huang

Reserpine

Derived from Vedic herbal (Rauwolfia) for snakebite and madness

Chlorpromazine

Antihistaminic preoperative sedative

Imipramine

Putative antipsychotic (“analogue” of chlorpromazine)

Iproniazid

Antituberculous but also MAO inhibiting and mood elevating

Chlordiazepoxide

Possibly sedative muscle relaxant

Haloperidol

Nonanalgesic meperidine analog

Clozapine

Surprising imipramine analog

Carbamazepine

Not just another anticonvulsant

Valproic acid

A noninert solvent

Fluoxetine

Antidepressant and anxiolytic

Buspirone

An early atypical antipsychotic

An important reality that underlies the process of drug development in the psychopharmaceutical industry is the profit motive. Psychotropic drugs have gradually risen to the top of all classes of drugs in annual prescription counts and income from sales. As noted, individual psychotropic compounds represent tens of millions of prescriptions and several have generated several billions of dollars in sales annually. In a recent international drug market survey, drugs acting on the central nervous system (CNS) were the leading category of all types (Table 1.1), with annual international sales of nearly $120 billion, or more than 15% of the total world drug market, and four of the ten best-selling drugs recently were psychotropics (antipsychotics and antidepressants), all returning 4–5 billion dollars/year [54]. However, there is an emerging tendency for expected growth annual sales of psychotropic drugs to level off, with concern about a growing disparity between massively rising costs of research and development vs. a decline in the rate of new products marketed among all pharmaceuticals (Fig. 1.2). These trends are indications of a highly successful, but maturing, market and they reflect the enormous difficulties in developing new, and especially, innovative or superior drugs, and especially psychotropics.
Fig. 1.2

Pharmaceutical research and development costs vs. new-drug approvals. Research and development (R&D) costs are in billions of US dollars/year and new-drug application (NDA) approvals by FDA are for all types of drugs (1994–2010). Adapted from Harris [39]

The process of new-drug development in psychopharmacology is a fundamentally conservative and empirical process that appears to overvalue principles of drug action established or proposed for known agents. This process results in searches for more drugs with similar effects and limitations. For example, it remains hard to imagine investing tens or hundreds of millions of dollars in developing a potential antipsychotic agent that has no antagonistic action on central dopamine or serotonin receptors, or an antidepressant that does not limit the inactivation of serotonin or norepinephrine. Following such conservative models derived from the pharmacology of older, successful, agents may be an effective business model, but is hardly likely to provide highly innovative or truly unique means of achieving desired clinical ends. More fundamentally, the process of psychopharmaceutical drug development over the past half-century reflects the severely limiting effect of a lack of knowledge of etiology of psychiatric disorders, and only fragmentary and unconvincing notions about their possible pathophysiology. Even the pathophysiological hypotheses that have been proposed are logically circular and based largely on known actions of available treatments. In short, drug development for psychiatry has been empirically effective, if basically repetitious, for several decades but true innovation remains extraordinarily elusive.

Current procedures for developing, testing, and seeking regulatory approval of new drugs in the United States have arisen by tradition as well as by the regulatory requirements of the US Food and Drug Administration (FDA), with similar procedures followed elsewhere by the activities of increasingly merged, international pharmaceutical corporations. In this process, once the potential clinical usefulness of a new molecule is suspected, initial animal experimentation is conducted to establish its apparent spectrum of pharmacological activities, to evaluate its metabolism and disposition, as well as its potential toxicity, and to estimate likely clinical doses and the ratio of its median toxic or lethal doses to its median effective doses (therapeutic index) or margin of safety. The typical course of discovery and development of new psychotropic drugs and the standard phases of drug development and typical times involved are summarized in Tables 1.6 and 1.7.
Table 1.6

Historical evolution of modern psychopharmacology

Before 1950s

Long use of sedatives (antihistamines, barbiturates and newer nonbarbiturates, bromides, cannabis, chloral hydrate, ethanol, opiates, reserpine) as well as stimulants (amphetamines, camphor), antiparesis (cerebral syphilis) agents (mercurials, penicillin), and physical methods (baths, restraints, insulin, chemical and electroconvulsive treatments (ECT), leucotomy)

1950s

Serendipitous discoveries of lithium, phenothiazines and haloperidol, antidepressants, benzodiazepines

1960s

Gradual and contentious acceptance vs. psychotherapy; commercial potential emerges; regulatory requirements standardized; randomized-controlled trials become standard; rating-scales and simple statistics develop; initial efforts in pathophysiology inspired by early pharmacodynamics

1970s

Molecular and receptor-based preclinical screening compete with animal behavioral modeling; intensive pharmacocentric research aimed at pathophysiology; drugs and syndromes encourage each other as new diagnoses proliferate with new treatments (e.g., attention, bipolar, depressive, obsessive-compulsive, panic, phobic disorders)

1980s

High throughput, automatic synthesis and drug screening established; nosology expands to over 300 diagnoses, driven partly by drugs and insurability; few novel and not-better agents emerge; improved methods for analysis of therapeutic trials

1990s

“Decade of the brain”; genetics and brain imaging emerge; new molecular gene-product targets identified (often without clear functions); research and development costs rise; billion-dollar/year drugs; interest in “effectiveness” (practical utility) vs. efficacy emerges; markets expand to general medicine; adverse drug effects: different but unavoidable; effect-size declines as placebo-rates rise and multisite, off-shore trials become popular

2000s

Innovation limited despite continued advances in basic and clinical neuroscience; ratio of research and development costs-to-novel-products highly unfavorable; many patents run out and insurers encourage use of generics; markets begin to saturate; clinical practice dominated by limited contact and medicating; growing concern about adverse metabolic and behavioral effects; growth of postmarketing research; efforts at evidence-based therapeutics emerge

Table 1.7

Typical phases of drug development in the United States

Phase

Tasks

Years

Preclinical research

Identify leads, preliminary pharmacology; patents (delayed as long as possible)

2–5

Initial FDA review

IND, enter clinical testing

ca. 5

Clinical testing

Phase I

Human tolerability, pharmacokinetics

4–6

Phase II (A and B)

Pilot and later rigorous small trials, dose-finding

Phase III

Large controlled, pivotal trials

FDA review

Review of all data, licensing (NDA)

2–3

Postmarketing (“Phase IV”)

Refine uses, doses, adverse effects

Open-ended

End of patents

Generics appear, prices fall

20

Phases of Drug Development

The early, preclinical steps in drug development and their refinement have become an increasingly sophisticated subspecialty within the field of psychopharmacology. They involve such technologies as molecular targeting of suspected primary sites of action (such as hormone or neurotransmitter receptor proteins, or cell membrane transporter proteins) in addition to more traditional modeling based on effects on the behavior of laboratory animals (discussed in more detail in the second edition of this book; [4]). It is now commonplace to design selective small molecules with high affinity for defined macromolecular target sites [25]. Moreover, many steps in the chemical synthesis of variants of a desired compound (to seek more ideal candidate molecules or potential follow-on agents for later development), as well as molecular screening procedures that seem to be highly efficient, are often automated and robotically controlled. These modern procedures involve high-capacity or combinatorial chemistry and high-throughput screening methods that are technically impressive but expensive and liable to produce large numbers of uninteresting candidate molecules [26, 27, 48, 49, 84]. In addition, pharmacokinetic modeling can identify oral bioavailability, elimination half-life and clearance rates as well as likely enzymatic routes of metabolic conversion, based usually on animal models [59]. Advanced techniques include the introduction and expression of human genes of interest into test animals so as to provide modeling that is more likely to represent human subjects [21]. Another increasingly employed technique is the early application in test animals and in human subjects of in vivo labeling of brain target sites of interest with radiolabeled tracer molecules (such as for positron-emission tomography [PET]) as an approach to estimating potency, dosing requirements, and pharmacokinetic measures [66, 83].

Following initial identification of a promising candidate molecule, the process of drug development usually splits into further pursuit of basic mechanisms of action and preclinical pharmacological characterization vs. initiation of human and clinical studies, as are summarized in Table 1.7. Basic pharmacology occurs in both academic and industrial laboratories, but clinical testing for safety and efficacy, as required by regulatory agencies, is largely the province of the pharmaceutical industry, owing mainly to the extraordinary costs involved, and its timing is driven by the limited patent-life of new drugs (typically 20 years in the United States since 1995). Developmental costs for a single new drug that acts on the CNS can run to hundreds of millions of dollars, although the actual costs vary with definitions of what is included in research and development costs as opposed to marketing activities [1, 2]. In turn, such costs drive efforts to develop “blockbuster” drugs with sufficiently large markets as to guarantee recovery of investment costs as well as substantial profit.

The first phase of human experimentation (Phase I) involves toxicological and pharmacokinetic studies in healthy human volunteers. Such subjects are increasingly difficult to access due to ethical constraints against enrolling persons whose voluntary status is questionable (such as prisoners, other institutionalized or cognitively compromised persons, or the poor anticipating financial compensation). Initial Phase I studies typically are carried out with small numbers of closely monitored subjects under clinical-laboratory conditions. If this phase of drug development is encouraging, preliminary clinical trials are undertaken with regulatory review and approval (typically, with an investigational new drug (IND) permit from the FDA).

These Phase II trials are typically relatively small and involve carefully diagnosed and closely clinically evaluated subjects under rigorous investigative conditions. It is also at this point that the drug development sponsor will have decided on a plausible target indication or diagnostic group of particular interest. Such planning is not always sustained, since candidate drugs sometimes have unexpected clinical effects that have not been anticipated, and may lead to alternative developmental paths: examples include the dopamine partial agonist pramipexole, which proved to be successful as an antiparkinsonism agent rather than an atypical antipsychotic drug, and the serendipitous discovery of the erectile-dysfunction activity of sildenafil [17, 68]. Phase II trials may or may not include placebo controls or comparisons with established treatments, and aim primarily to develop evidence of efficacy under highly controlled conditions (“proof of concept”). If the Phase II trials are encouraging, the next step are larger and broader Phase III trials, typically in more clinically representative samples and often involving multiple sites.

Such Phase III trials are sometimes considered early (IIIA) or late (IIIB, typically following submission of preliminary findings to a regulatory agency). Formerly, they were often conducted within academic medical centers, but have increasingly been shifting toward management by contract research organizations (CROs) working with individual clinicians, as well as toward large, international, multisite, collaborative trials, often in dissimilar cultures. In such settings, methods of diagnosis, clinical assessment, and symptom ratings may not be well standardized and validated. Placebo-controlled, randomized trials have long been considered optimal for testing new treatments. However, ethical and clinical controversy about their use in some clinical circumstances is now coupled with growing reluctance of potential patient-subjects to accept enrollment in trials involving a placebo condition when known treatments are available. Nevertheless, FDA approval typically requires at least two “pivotal” trials of substantial size, involving randomized, blinded, comparisons of the test drug against a placebo, and showing statistical superiority of the active agent. A major scientific reason to continue to require some placebo-controlled trials is that comparisons of new vs. established drugs are likely to yield little or no apparent difference, whereas clear superiority of a new agent over an established drug is rare in psychiatry. This circumstance can lead to the logically highly risky conclusion that a finding of “not different from” is equivalent to “about as good as.” Such potentially false conclusions are especially likely if a trial is poorly designed or conducted, with high levels of random variation, or if the standard comparator does not show expected efficacy, as happens more often than one might expect of an established treatment.

If all three phases of clinical investigation are successful, the drug can be a candidate for licensing by a regulatory agency, such as by an FDA new-drug application (NDA). Licensing requirements and procedures in other countries differ in some details, including requirements for placebo-controlled trials. Drug monitoring and regulatory processes exist in many individual countries, including Health Canada, the British Medicines and Healthcare products Regulatory Agency (MHRA), and the scientifically-oriented British National Institute for Health and Clinical Excellents (NICE), and the European Medicines Agency (EMA). Approval processes also vary widely in their linking of the licensing of new drugs for clinical use with their pricing. Price regulation is not done in the United States, where brand-name drugs typically command high prices during their patent-life, and subsequent generic products are much less costly.

An important additional phase of drug development is “after-marketing” (“Phase IV”), in which the optimal clinical use and dosing of newer drugs is clarified, and perhaps extended to new indications with additional controlled trials. Of particular importance, this phase of drug development can provide much more information about potential adverse effects than are detected or quantified during Phases II or III, especially involving events of low prevalence, whose detection and quantification may require large numbers of exposed persons. Nevertheless, in general, current means of monitoring FDA-required safety of new drugs remain less well developed than tests of clinical efficacy. They continue to rely heavily on passive and incidental observation or reporting of adverse effects by patients to investigators or clinicians, with inconsistent reporting to regulatory authorities after licensing. There are growing efforts to incorporate explicit assessments of suspected areas of risk into clinical trials as well as in after-marketing monitoring programs, with the aim of improving timely detection and quantification of adverse effects and increasing the safety of marketed drugs. Another component of after-marketing investigations involves practical or clinical effectiveness trials that compare the therapeutic value, relative tolerability, and acceptability of licensed treatments in large samples of patients aiming to depict drug performance in broad clinical practice. Sometimes, such effectiveness trials incorporate the methods of randomization and blinding that are more typical of Phase II and III trials.

An important movement in recent decades involves efforts to place therapeutics on a more sound, scientific basis by pooling and comparing findings from available clinical trials, in what is referred to as evidence-based therapeutics [64, 70]. The aim is to rank specific treatments by evidence of their relative efficacy, safety, and tolerability. Such efforts have obvious potential clinical value, but are also encouraged by administrators, policy makers, and insurers, in order to maximize not only the clinical value of treatment choices, but also to limit costs. For most classes of psychotropic drugs, this effort has yielded very limited success, at least in determining compelling and consistent rankings of specific drugs within a class by efficacy and safety [50, 51, 77, 82].

Nature of Clinical Trials

It is important to have some appreciation of the nature of clinical therapeutic trials, their design, limitations, and interpretation, in order to evaluate new findings critically. This need is particularly important since treatment trials in general, and particularly for psychiatric disorders, vary markedly in quality. Basic requirements for a credible therapeutic trial include clearly defined and clinically reasonably similar, but hopefully representative, patient-subjects in substantial numbers (N), and with substantial morbidity, evaluated by clinically relevant, sensitive, and reliable measures of clinical change.

Effects of Trial Size

Sometimes, a large N can work against a successful outcome (support for the hypothesis that a drug is effective), particularly in trials involving not only large numbers of subjects, but also multiple collaborating sites, each of which may contribute only a few patients. In large, complex trials, it becomes very challenging to establish and maintain reliable and consistent diagnoses, clinical assessments, and quantitative ratings on standard symptom scales. Complexity surely increases when multiple regional, national, and cultural differences may affect the meaning of diagnostic terms and criteria and of items on standard symptom-rating scales, even after translation into local languages. In turn, compromises in quality control and inconsistent interpretations can increase heterogeneity and may result in high inter-site variability, which is rarely reported. These circumstances also can make pooling of data across sites risky, and commercial application of pooled data to all sites (such as for licensing across countries) questionable. Trial complexity and heterogeneity appear to have greater impact on responses during placebo treatments than with active test drugs, possibly as a reflection of the phenomenon of regression to the mean or chance outcomes, which are more likely with placebo [77, 82, 83]. It is also likely that heterogeneity and compromised control of the conduct of trials have contributed to a noteworthy trend in recent years toward falling drug–placebo contrasts in trials of various types of psychotropic drugs [77, 83]. In turn, it is tempting to combat this trend by use of ever-larger and more complex multisite trials in order to increase statistical power (N) as effect-size (drug–placebo contrast) diminishes, in a basically circular process [77, 78].

Subject Recruitment

As clinical psychopharmacology has evolved over recent decades, it has become increasingly difficult to enroll patients who are diagnostically and clinically typical, at least moderately symptomatic, and not already treated more or less optimally. This quest has become particularly difficult in academic medical centers of North America and Europe, where teaching and research clinics tend to accumulate ever-more difficult and treatment-unresponsive, but already more-or-less optimally treated, patients who may be not only unrepresentative of broader clinical samples, but also unlikely to improve sufficiently for plausible inclusion in clinical trials of new treatments, though often willing to participate in hopes of encountering a better treatment. Of note, in well-established academic centers, treatments tested earlier in the life of a research clinic may have shown superior outcomes compared to those tested later, as more and more difficult, less treatment-responsive, patients tend to be retained over time. Since minor clinical changes are now routinely encountered, outcome measures other than indices of clinical improvement are sometimes employed. These may, for example, include time to a decision to change treatment, usually assumed to reflect poor response or limited tolerability, but typically complex and incompletely defined [53]. In addition, the efficacy of many psychotropic drug treatments is modest, and drug–placebo differences sometimes quite small [6, 77]. In part for this reason, and because major clinical improvements may require several months, it is a common practice to define outcomes in terms of proportions of patient-subjects reaching a defined level of clinical change (commonly ≥50% improvement in a rating-scale measure).

Difficulties in recruiting appropriate trial subjects contribute to pressure to rely increasingly on CRO-organized trials based in private clinics and offices, or to seek far less extensively treated or studied patients in less developed countries—again with an uncertain impact on control of diagnosis and clinical assessments and on the reliability and generalizability of the results obtained. The trend toward larger, multinational trials is also motivated by the quest for less costly opportunities for carrying out work in underdeveloped countries with lower labor costs. However, risks involved, in addition to matters of heterogeneity and quality control already mentioned, include variable levels of required infrastructure and training of research clinicians in many sites, as well as risk of inadequate supervision and temptations to distort data collection and reporting to earn money. Countering such potential problems can become very expensive, and tend to limit anticipated cost-savings. Another emerging problem with the international quest for low-cost alternatives to developing drugs in developed countries is that the quality of drug products manufactured in underdeveloped countries can be highly variable, unreliable, and sometimes unsafe, perhaps particularly with generic drugs developed after original patents have expired [39, 73]. Again, interventions required to assure product quality can rapidly erode expected savings in drug production.

Clinical Heterogeneity, Randomization, and Blinding

A major source of potentially misleading outcomes in modern therapeutic trials is virtually irreducible clinical heterogeneity. Even with rigorous application of modern diagnostic criteria and formal systems of clinical assessment involving standardized and structured interviewing techniques, most clinical syndromes in psychiatry involve large inter-individual differences and change over time. Variability arises from individual differences in the quality and severity of illnesses, as well as in the timing of interventions in the course of acute, recurring, or chronic illnesses. Moreover, some disorders (e.g., acute psychoses and apparently unipolar major depression) may meet even rigorous diagnostic criteria at one time, but not later [65]. To a large extent, such variability can be managed more or less adequately by inclusion of substantial numbers of patients and effective randomization. Nevertheless, even with randomization, heterogeneity within and among trials yields average outcomes that may not differ markedly between treatments within individual trials, or after pooling of findings across trials. Currently, it is fashionable to pool data across trials by the methods of meta-analysis [34]. Meta-analytic summaries are descriptive clinical exercises, in which randomization among trials does not obtain, and statistical power is limited by the number of trials pooled, not by total numbers of subjects. Pooling data from trials involving psychiatric patients often yields average outcomes that fail clearly to distinguish one treatment from another, and so frustrate the aims of evidence-based therapeutics [28, 77, 82].

“Blinding” is an important method of controlling bias arising from observers or patients (“single-blind”) or both (“double-blind”) to keep those involved unaware of the identify of a presumably randomly assigned treatment option. Success of the technique varies, and can be compromised when characteristic or obvious drug side effects make it clear which treatment is involved. Some studies include a debriefing phase at the end of the trial, but before the blind is broken, in which both patients and investigators are asked to guess at the treatment given to individual patients. However, standards by which to use such information to assess the resulting primary outcome findings are rarely defined or applied.

Even with randomization and blinding, therapeutic trials are at risk of additional artifacts that are not always readily controlled. A common problem is subject-retention. The longer a trial continues, and the more prevalent are adverse effects of a test agent that limit its tolerability, or the greater the lack of obvious benefit (as with a placebo or an ineffective drug), the less likely a patient-subject will continue to the designed end of a trial. Premature “dropouts” in such nonrandom circumstances can introduce major sources of bias and limit statistical power through attrition, as well as losing potentially valuable information concerning reasons for early dropping out. Moreover, if an ineffective treatment or placebo condition leads to earlier discontinuation than with a test drug, with higher levels of symptoms, it is likely that drug–placebo differences will be exaggerated. Conversely, if dropouts occur earlier with an active drug than a placebo, the opposite conclusion may be drawn, and the experimental treatment be considered ineffective, particularly if some degree of spontaneous improvement occurs over time to favor placebo treatment.

Analysis of Trial Outcomes

Various analytical methods have been developed to limit such risks [41]. They include statistical modeling techniques that make use of all data available at all time points, such as mixed-effects modeling, generalized estimating equation (GEE) regression modeling, and survival analysis. Such methods are far more powerful than older, simpler, but potentially-biased methods of comparing outcomes between subjects who complete a trial (“completer” analyses) or by making contrasts of initial and final ratings obtained at any time (as in “intent-to-treat” analyses), as well as based on considering a last observation as an end-point (last-observation-carried forward method). Such methods risk introducing biases that can have unpredictable effects on conclusions drawn [88], but even modern statistical modeling techniques involve assumptions and simplifications. A summary of methodological concepts associated with the design and analysis of clinical trials is provided in Table 1.8.
Table 1.8

Methodological concepts for clinical psychopharmacology

Term

Definition

RCT

Randomized, controlled trial vs. placebo, standard agent, or other controls; usually under double-blind conditions

Efficacy

Typically, superiority to placebo at p < 0.05, but effect-size can be large or small (commonly based on “response” as <50% reduction of severity ratings)

Blinding

Single: patient unaware of treatment provided; double: neither patient nor clinician or rater is aware of treatment; double-blinding is standard but not always guaranteed

Enrichment

Selecting subjects for a second-phase of a trial based on responsiveness in a preliminary phase (typically in acute illness); popular for efficiency, but can bias and be misleading

Effectiveness

Beneficial and tolerable in clinical settings, typically long-term

Probability ( p)

Likelihood of statistical superiority vs. control (typically null or no-difference) is <5% likely (smaller p-value is generally better)

Continuation

Treatment continued after initial response in acute illness, ideally (but not always) to full clinical recovery

Prophylaxis or maintenance

Reduction of risk of future morbidity or recurrences of acute episodes with long-term treatment; proof of such benefits remain inadequate in much of psychopharmacology

Discontinuation

A trial design that removes an initially effective treatment; sometimes proposed to indicate prophylaxis but at great risk of discontinuation-associated stress, especially when discontinuation is rapid and patients incompletely recovered

Power

Ability to identify difference between treatments; greater with larger effect-size, greater precision, and larger sample (N) to yield lower p-value

Effect-size (d )

Response-difference ([treatment–control]/[measurement variance])

RR and RD

Relative response rate (RR), typically proportion attaining a criterion outcome (typically 50% reduction in symptom-rating scores) with drug/control, or the difference (RD) in responses or improvements between drug and placebo

CI

Confidence interval (typically reported 95% CI includes 95% of measurement variance; nonoverlap of 95% CIs indicates p < 0.05, by inspection)

Meta-analysis

Statistical method for pooling results across trials (N = trials, not cases); results can be expressed as RR (response rate-ratio), odds ratio (OR), or RD (response difference), usually weighted by trial size and variance measurement

NNT or NNH

Number of cases needed to treat or harm, for superior benefit or risk of an adverse effect vs. a control (ideally NNT < NNH); computed as reciprocals of RD; NNT >10 indicates limited benefit

Noninferiority

Similar outcome with two treatments compared head-to-head with sufficient statistical power to rule out defined levels of potential difference

EBM or EBT

Evidence-based medicine or therapeutics: research-supported treatment, commonly arising from meta-analyses + clinical experience + commonsense (risks over-interpretation); difficult to achieve in psychopharmacology due to typically similar efficacy assessments within drug-classes

Survival analysis has become particularly popular for use in evaluating outcomes in clinical trials, by considering the latency to a defined outcome, such as the time from the start of randomized treatment to a particular level of clinical improvement or to a first relapse, although it was originally introduced to compare times to death under different clinical conditions or treatments [23]. The technique assumes that latency to a defined outcome is a suitable proxy measure for a general clinical outcome. For example, in the long-term treatment of a recurrent major mood disorder, a longer delay of a next recurrence, following initial recovery, is often considered evidence of a superior prophylactic treatment. A tacit assumption involved is that delay of an initial recurrence is a fair surrogate for long-term wellness, but this assumption has rarely been tested empirically [10, 11].

The use of randomization to an active test drug vs. an inactive but similar-appearing dummy pill (placebo) continues to be considered about the best research design for experimental therapeutic trials. This conclusion is probably valid scientifically, but use of placebos can lead to complicated sociological and even ethical dilemmas. Moreover, different cultures vary in their insistence on data from placebo-controlled trials for licensing of new drugs. Use of placebo controls continues to be a policy of the U.S. FDA in its requirement for a minimum number (usually two) of so-called “pivotal” trials with placebo controls. In large part, the controversy involved reflects the tension between scientific aims and clinical considerations. Now that effective and tolerated medicinal agents are available for most serious mental illnesses, some observers and many patients and their families argue that depriving a suffering or potentially suicidal patient of an established treatment is clinically very questionable and perhaps unethical, and would prefer to rely on comparisons of older vs. newer drugs. However, as noted above, such comparisons may mislead by failing to distinguish between active treatments and encouraging the risky conclusion that they are similar in efficacy. Moreover, testing aimed at supporting equal efficacy is challenging and typically requires large numbers of subjects.

Ideally, it would be good to seek innovative treatments that are demonstrably superior in efficacy or safety to older, standard, treatments. However, at least in clinical psychopharmacology, such aims are usually excessively idealistic, as clear outcomes favoring one treatment over another are uncommon or rare, as already noted. Most often, outcome involves similar final clinical status or levels of improvement with any two active treatments of similar kind. However, even well-established standard treatments sometimes fail to perform in any given trial for a variety of reasons, and leave the logical conundrum of deciding whether “similar to” implies “about as good as, or not.” Without an inactive treatment condition (placebo), it is hard to know if the new drug is indeed about as good as an older standard agent, or if the trial simply failed. Some trial designers attempt to limit the clinical and possible ethical liability involved by arranging for smaller numbers of subjects to be randomized to placebo than to an active treatment. If “quantitative ethics” is a legitimate concept, it can also be argued that trials not involving a placebo typically require much larger numbers of participants since differences between active drugs of similar type are typically small (i.e., effect-size is small), so that larger N is needed to demonstrate a statistical difference. Conversely, drug–placebo contrasts are usually easier to detect with smaller numbers of subjects, so that overall exposure to potential risks is more limited.

A further dilemma in interpreting trial outcomes arises from the interplay among effect-size (difference between treatments assigned randomly), N, and variance (variation that reflects the precision of measurements) is whether to make more of effect-size or the probability of difference from a null condition ( p-value indicating that a null [typically, no-difference] can be rejected). Many clinicians and some investigators are highly impressed by tiny p-values and pay less attention to effect-size. Indeed, some regulatory agencies consider a novel drug to be “effective” if it outperforms a comparator or placebo at the 5% probability level ( p < 0.05). However, it is not uncommon in clinical psychopharmacology to encounter levels of effect-size that strain credulity as to clinical value. For example, differences between antipsychotic drugs and placebo in schizophrenia patients or between antidepressants and placebo in depressed patients may be on the order of 10% or less, even with probabilities below 5%. Moreover, although it can be expensive to do so, it is usually possible to “force” a low p-value by increasing N, even with very modest effect-sizes. That is, probability values ( p) are smaller with (as desired, or inversely proportional to): larger effect-size (difference between treatments), larger N, and smaller measurement variance. Some regulatory agencies also set standards of efficacy or safety of new drugs based on differences in observed outcomes that meet the statistical criterion of p < 0.05, despite sometimes clinically trivial effect-size. Exaggerating the significance of probabilities can also lead to pharmacologically illogical, and clinically potentially dangerous conclusions, including concepts such as “safe vs. risky” doses, or “effective vs. ineffective” doses. Notably, if an adverse effect occurs at p > 0.05 at a particular dosing level, it is not reasonable to conclude that the observed adverse effect is clinically unimportant. Instead, adverse effect risks, as well as benefits relate to drug dose in a continuous, logarithmic, statistical manner, such that beneficial and adverse effects are almost universally nearly linearly proportional to the logarithm of dose. Moreover, risks or benefits at particular doses and tissue concentrations of a drug can vary markedly among individuals or groups (such as juvenile vs. geriatric patients).

Drug Dosing

Scientifically sound information on relationships of doses and beneficial and adverse effects of most psychotropic drugs remains notoriously meager. Very few prospective, randomized trials seeking to assess dose-effects are reported, and estimates of clinically appropriate or approximately equivalent doses of specific drugs remain largely clinical, informal, and impressionistic [32]. Doses that have regulatory approval and appear in product information bulletins appear to balance the need to be large enough to assure substantial efficacy in typical patients, and yet to limit risks of unpleasant or toxic adverse effects, both of which are usually more or less linearly related to the logarithm of daily dose of most drugs. Given such limited information, dosing is largely left to clinical, empirical trial-and-error assessments with individual patients, starting at the low end of the recommended dosing range, and gradually increasing to gain benefits and avoid intolerable side effects. Exceeding recommended dosing limits, clinical indications or diagnoses in individual patients, as well as use of drug combinations that are not formally recognized by regulatory agencies are common practices. They are most safely carried out with the understanding that they represent informal N = 1 trials for individual patients for whom more standard treatments have proved to be inadequate or poorly tolerated, ideally with some indication to the patient of the “off-label” status of a recommendation, and with extra documentation in the clinical record to support such decisions. Assessments of relationships among dose, circulating serum concentrations, and clinical effects of drugs are best carried out with due consideration of typical pharmacokinetic characteristics, such that stable or steady-state tissue levels are reached in multiples of five or six times elimination half-life.

Carryover and Withdrawal Artifacts

A final consideration about the design of experimental therapeutic trials is the importance of time, both regarding duration of treatment, the timing of outcome assessments with respect to previous treatments, and the impact of treatment discontinuation. Duration of trials is typically decided in efforts to balance chances of finding robust clinical improvement while limiting early dropouts and the cost of the trial. Some conditions show beneficial effects surprisingly quickly, for example within days in cases of juvenile attention-deficit/hyperactivity disorder, and within a few weeks in acute mania, whereas major depression tends to show more improvement and better separation of drug and placebo with trials that last more than a month [74].

A particularly important aspect of the timing of therapeutic trials is the relationship of what occurs during the trial to what has preceded it. That is, many trials are affected by carry-over effects due to gradual elimination of previous treatments, especially when these are long-acting, such as with highly lipophilic agents (whose elimination half-life may be as long as weeks rather than days) or injected prodrugs, which are hydrolyzed in the body to liberate the active parent compound and can require months to be cleared from tissue pools [18, 86]. Another commonly encountered situation is rapid discontinuation of ongoing treatment to enter an experimental trial or to shift from a short-term to long-term phase of an extended trial (treatment discontinuation trial). Both situations can lead to markedly increased symptomatic expression of the illness being treated or major relapses that may be delayed for some weeks, with significant adverse clinical effects and ethical implications as well as profoundly interfering with sound interpretation of findings from trials. Such treatment discontinuation-associated clinical worsening has been reported with mood-stabilizing [29], antipsychotic [80], and antidepressant drugs [10], and is to be distinguished from physiological withdrawal reactions that usually occur within days of discontinuing some psychotropic drugs, including benzodiazepines [62] and short-acting serotonin-reuptake inhibitors [43]. All of these reactions can be markedly limited by gradual discontinuation of psychotropic drugs, usually over several weeks, especially after long-term use of short-acting agents in relatively high doses.

Long-Term or Prophylactic Trials

Much more research is available to support relatively short-term treatment of acute phases of illness than for long-term treatment, even though most major psychiatric disorders are recurrent-episodic, or chronic with variable levels of symptomatic intensity. In part, this disparity reflects the major complexities encountered in attempting to prove long-term, preventive, or prophylactic effectiveness. These include high costs, recruiting and retaining research subjects, especially when a placebo condition is involved and as dropout rates rise with longer trial duration, maintaining blinding to treatment conditions, inclusion or control of multiple treatments or not, long-term burdens of adverse effects, and many logistical difficulties. More fundamentally, there is a lack of agreement on standards by which such trials should be designed, conducted, and interpreted [7, 35, 61, 71]. An older tradition was to enter patients in any phase of complex illnesses and to follow them with various treatment options, often including placebos, and for very long periods that usually exceeded a year. Outcome measures were often quite simple: counting recurrences or time-ill between treatments. More recently, long-term trials have relied on the methods of survival analysis, largely as a means of limiting trial duration by measuring time to a first recurrence of illness with vs. without active treatment. As noted above, the assumption that such latencies are robust indications of long-term morbidity or wellness remains poorly tested [11].

It has become common to require passing through a short-term trial for treatment of acute illness (often with a placebo control) and then to continue observations for at least several months, often after discontinuing active treatment at some point. Many acute psychiatric illnesses or exacerbations or recurrences of chronic or episodic disorders require many weeks or even months for full clinical recovery, even though major symptomatic improvement can often be documented within several weeks [75]. However, there is striking lack of agreement about when it is clinically and ethically appropriate to discontinue treatment for experimental purposes, or how to provide adequate information to participants as to qualify them for informed consent. Drug-discontinuation-associated recurrences are especially likely during incomplete clinical recovery from an acute illness, but may occur after discontinuing treatment even many months or years after full remission and sustained clinical stability [9, 81]. It is also not clear to what extent even very slow and gradual removal of a drug can limit such risks [12].

The preceding considerations indicate that incomplete clinical recovery and even late risk of reaction to relatively rapid treatment discontinuation severely complicate the design of long-term trials or extensions of short-term trials. Moreover, extension trials are often biased by requiring short-term responsiveness to a particular treatment (usually one produced by the study-sponsor’s product, so-called “enrichment” designs) before being continued or re-randomized into a continuation phase that involves treatment discontinuation. Discontinuation trials of this kind may add support for short-term efficacy of a particular treatment by documenting the impact of its removal, but not provide evidence of true, long-term prophylaxis. In short, scientifically optimal and clinically appropriate, ethical, and feasible design and analysis of trials aiming to test for long-term preventive effects remain major challenges for experimental psychiatric therapeutics.

Reporting and Pooling of Trials Data

In recent years, corporate-sponsored therapeutic trials of psychotropic agents have come to dominate the research literature. Even reports of trials nominally arising from academic centers often have corporate sponsorship. In such circumstances, it would not be surprising if purely objective and scientific purposes of testing efficacy and safety of new agents were not entirely isolated from commercial aims, including licensing and product placement in markets [2, 40]. Given the close monitoring by regulatory agencies, it is unlikely that overtly misleading information is reported in corporate-sponsored findings. It is, however, likely that the design, analysis, and reporting of trials by manufacturers who know it best can be crafted so as to favor a particular product. In recent years, there have been vigorous efforts to create systematic and publicly accessible registries of treatment trials that are planned, in progress, or recently completed (e.g., [22, 89]) as well as standards accepted by journal editors for the publication of reports of trials (e.g., [45]). These procedures are designed to improve access to information by healthcare professionals and, to some extent, the general public and to limit selective and incomplete reporting [44]. Nevertheless, there is growing evidence of selective reporting of some trial findings in the peer-reviewed research literature, often in directions that tend to favor pharmaceutical manufacturers’ products, even in head-to-head comparisons, which are potentially valuable but uncommon and put particular products at risk of inferior performance [42, 58, 63, 76, 79].

Biological Hypotheses in Psychiatry

The introduction of relatively effective and somewhat selective medicinal agents for the treatment of psychotic, manic-depressive, depressed, and anxious patients in the 1950s and 1960s, as well as other agents that could mimic or worsen symptoms of some illnesses encouraged formulation of biological hypotheses about the possible pathophysiology of various major mental illnesses. Such biomedical hypothesizing was further encouraged as knowledge of drug actions increased in subsequent years. For example, an association of depressed mood in some vulnerable persons given reserpine or other antiadrenergic agents to treat hypertension and in others exposed to diets deficient in aromatic amino acids that are precursors of monoamine neurotransmitters [60], as well as beneficial effects of exposure to antidepressants or stimulants, encouraged speculation that depressive illness may involve a deficiency of neurotransmission in the brain mediated by monoamines including norepinephrine and serotonin (5-hydroxytryptamine). In addition elevation of mood or worsening of mania or psychotic illness with dopaminergic stimulants, and the opposite effects of antipsychotic–antimanic drugs with potent central antidopaminergic actions encouraged a dopamine-excess hypothesis concerning mania and schizophrenia. Additional alternatives, also strongly encouraged by psychopharmacological findings, included searches for endogenous “psychotoxins” or hallucinogens in psychotic disorder patients, encouraged by the striking psychotropic actions of hallucinogens found in nature or synthesized in laboratories, including LSD (d-lysergic acid diethylamide) and alkylated tryptamines [8].

Such theorizing seemed plausible and rational, and it strongly encouraged a generation of intensive investigations of what might be considered basically “pharmacocentric” theories in a modern phase of biological psychiatry [5]. Years of efforts were made to test the implications of such hypotheses, especially by seeking evidence of corresponding biological abnormalities in psychiatric patients. Nevertheless, to date, the results have not, on balance, led to a compelling and coherent body of evidence for specific metabolic, pathophysiological changes in patients with a range of major psychiatric disorders. Many changes that have been identified may represent state-dependent conditions secondary to physiological changes associated with acute or chronic psychiatric illnesses—that is, effects, not causes. Alternative strategies as well as failures to find strong support have largely led to the abandonment of the pharmacocentric approach. Nevertheless, applications of modern metabolic, genetic, and brain imaging strategies also have led to few clear conclusions, despite growing evidence of major genetic contributions to the risk of several psychiatric disorders [14, 15, 16, 20, 33, 56].

There are several difficulties in efforts to support biological hypotheses of the pathophysiology of mental illnesses. One is that actions of psychotropic drugs which are both necessary and sufficient to account for their clinical benefits remain elusive. In addition, immediate pharmacodynamic actions of such drugs are only a small part of increasingly complex, and later-emerging actions that involve deep alterations in cellular functioning, even at the level of gene expression and regulation [19, 37, 52, 55]. There is also a risk that treatments themselves induce metabolic and physiological changes in the nervous system that can lead to artifactual research findings in patients who have been treated, especially for prolonged periods with multiple psychotropic drugs.

Moreover, many drugs in general medicine are effective but act at sites far removed and only indirectly associated with the pathophysiology of diseases. Examples include the effects of antipyretics in infections, or of diuretics in congestive heart failure. It is also an uncomfortable fact that most psychotropic drugs are not highly disorder-specific, but instead tend to be broadly useful palliative agents that produce beneficial effects in a variety of syndromes. Examples include both antipsychotic and antimanic and sometimes antidepressant effects of antipsychotic drugs, as well as antidepressant, anxiolytic, and even analgesic effects of many antidepressants. Such nonspecificity is not ideal for spinning biological hypotheses based on limited knowledge of drug actions. Moreover, decades of essentially repetitious development of increasingly chemically diverse drugs in a particular class (such as antidepressants or antipsychotics), based on conservative and limited understanding of their actions, had tended to set up and reinforce a logically circular conceptual system. This conservative cycle (Fig. 1.3) can lead not only to misleading biomedical speculation, but also impede progress in developing innovative drugs based on novel principles.
Fig. 1.3

Pharmacocentric cycle of biological psychiatry. Interrelationships among drug treatments, diagnosis, partial understanding of drug actions (pharmacodynamics), and speculations about putative pathophysiology

Despite the lack of compelling evidence for pathophysiological factors in major psychiatric disorders derived from speculations concerning clinical and pharmacological effects of drugs that either worsen or ameliorate symptoms of the disorders, major advances have been made in methods and technology, and in advancing basic understanding of the brain, as well as encouraging much deeper interest in brain and behavioral neurosciences, and the training of a growing cadre of scientifically competent and imaginative investigators with an interest in psychiatric disorders. Additional secondary benefits from the explosive upsurge in interest in biomedical aspects of psychiatry, strongly encouraged by the beneficial effects of psychotropic drugs, is renewed interest in descriptive psychiatry and nosology, in efforts to address the “phenotype problem” of needing more secure and credible diagnostic groups with whom to search for anatomical, metabolic or genetic anomalies. This fundamental problem is far from resolved by currently available diagnostic systems, and requires much closer collaboration among biological scientists and clinical investigators. As a final and repeated comment, it is likely that the lack of solid clinical findings to support pharmacocentric biological theorizing arises not from technical shortcomings of modern medical biology, but from oversimplifications and false expectations arising from pharmacology, particularly as applied to disorders of essentially unknown etiology.

Conclusions

The development of modern psychopharmacology has had a striking impact on psychiatry in the past half-century. Effective and relatively safe medicinal treatments are available for most of the major psychiatric disorders. These treatments have had beneficial interactions with changes in the philosophy and administration of healthcare delivery programs, and have contributed to markedly decreased importance of prolonged psychiatric hospitalization. In addition, the development of psychopharmacology has contributed to a heightened awareness of the medical and scientific traditions of psychiatry and encouraged greater reliance on clinical studies of diagnosis, psychopathology, and epidemiology as well as of pathophysiology and treatment, especially for the psychotic and major mood disorders. The scientific design and conduct of objective, controlled clinical trials of new drugs in psychiatry have become increasingly sophisticated and serve as a model for other medical specialties. An important limitation to the discovery of new psychiatric medicines is that many of the preclinical methods for screening compounds with potentially useful effects have led to more compounds with previously known effects as well as adverse actions. Psychotropic drug development remains highly empirical and remarkably conservative conceptually. Progress remains hampered by a lack of knowledge of biological causes or of precise pathophysiological bases of the major mental disorders, most of which remain idiopathic. Indeed, it remains an act of faith to believe that such cause-related biological mechanisms will be discovered. The partial success of older treatments, paradoxically, contributes further to the problem of testing new agents, in part because patients available for study are increasingly treatment-resistant, complex, and otherwise atypical, as more of the broader spectrum of patients is successfully managed with available treatments.

References

  1. 1.
    Adams C, Brantner V. Estimating the cost of new drug development: is it really 802 million dollars? Health Aff. 2006;25:420–8.CrossRefGoogle Scholar
  2. 2.
    Angell M. The Truth About the Drug Companies. New York: Random House; 2004.Google Scholar
  3. 3.
    Baldessarini RJ. Frequency of diagnoses of schizophrenia vs. affective psychoses from 1944-1968. Am J Psychiatry. 1970;127:759–63.PubMedGoogle Scholar
  4. 4.
    Baldessarini RJ. Chemotherapy in psychiatry. 2nd ed. Cambridge, MA: Harvard University Press; 1985.Google Scholar
  5. 5.
    Baldessarini RJ. American biological psychiatry and psychopharmacology 1944–1994 (Chap. 16). In: Menninger RW, Nemiah JC, editors. American psychiatry After World War II (1944–1994), a Volume Celebrating the 150th Anniversary of the Founding of the American Psychiatric Association. Washington, DC: APA; 2000. p. 371–412.Google Scholar
  6. 6.
    Baldessarini RJ. Drug therapy of depression and anxiety disorders (Chap. 17). In: Brunton LL, Lazo JS, Parker KL, editors. Goodman and Gilman’s the pharmacological basis of therapeutics. 11th ed. New York: McGraw-Hill; 2005. p. 429–59.Google Scholar
  7. 7.
    Baldessarini RJ. Long-term drug efficacy in major psychiatric disorders: problems and opportunities. In: Proceedings of an International Society for Clinical Trials and Methodology (ISCTM) Annual Meeting, Chicago, IL. 2006. http://www.isctm.org/public_access/r_baldessarini-jul06-session1present.pdf.
  8. 8.
    Baldessarini RJ, Stramentinoli G, Lipinski JF. Methylation hypothesis. Arch Gen Psychiatry. 1979;36:303–7.PubMedCrossRefGoogle Scholar
  9. 9.
    Baldessarini RJ, Suppes T, Tondo L. Lithium withdrawal in bipolar disorder: Implications for clinical practice and experimental therapeutics research. Am J Ther. 1996;3:492–6.PubMedCrossRefGoogle Scholar
  10. 10.
    Baldessarini RJ, Tondo L, Ghiani C, Lepri B. Discontinuation rate vs. recurrence risk following long-term antidepressant treatment in major depressive disorder patients. Am J Psychiatry. 2010;167:934–41.PubMedCrossRefGoogle Scholar
  11. 11.
    Baldessarini RJ, Tondo L, Lepri B. Empirical test of a basic assumption underlying survival analysis as applied to experimental therapeutics in psychopharmacology. J Clin Psychopharmacol. 2010;30:72–5.PubMedCrossRefGoogle Scholar
  12. 12.
    Baldessarini RJ, Tondo L, Viguera AC. Discontinuing psychotropic agents. J Psychopharmacol. 1999;13:292–3.PubMedCrossRefGoogle Scholar
  13. 13.
    Ban TA. The role of serendipity in drug discovery. Dialogues Clin Neurosci. 2006;8:335–44.PubMedGoogle Scholar
  14. 14.
    Bondy B. Genetics in psychiatry: are the promises met? World J Biol Psychiatry. 2011;12:81–8.PubMedCrossRefGoogle Scholar
  15. 15.
    Brown GG, Thompson WK. Functional brain imaging in schizophrenia: selected results and methods. Curr Topics Behav Neurosci. 2010;4:181–214.CrossRefGoogle Scholar
  16. 16.
    Burmeister M, McInnis MG, Zöllner S. Psychiatric genetics: progress amid controversy. Nat Rev Genet. 2008;9:527–40.PubMedCrossRefGoogle Scholar
  17. 17.
    Campbell SF. Science, art and drug discovery: a personal perspective. Clin Sci (Lond). 2000;99:255–60.CrossRefGoogle Scholar
  18. 18.
    Campbell A, Baldessarini RJ, Teicher MH, Kula NS. Prolonged antidopaminergic actions of single doses of butyrophenones in the rat. Psychopharmacology (Berl). 1985;87:161–6.CrossRefGoogle Scholar
  19. 19.
    Chen G, Huang LD, Zeng WZ, Manji HK. Mood stabilizers regulate cytoprotective and mRNA-binding proteins in the brain: long-term effects on cell survival and transcript stability. Int J Neuropsychopharmacol. 2001;4:47–64.PubMedGoogle Scholar
  20. 20.
    Chen CH, Suckling J, Lennox BR, Ooi C, Bullmore ET. Quantitative meta-analysis of fMRI studies in bipolar disorder. Bipolar Disord. 2011;13:1–15.PubMedCrossRefGoogle Scholar
  21. 21.
    Chen AA, Thomas DK, Ong LL, Schwartz RE, Golub TR, Bhatia SN. Humanized mice with ectopic artificial liver tissues. Proc Natl Acad Sci USA. 2011;108:11842–7.PubMedCrossRefGoogle Scholar
  22. 22.
    Clinicaltrials.gov. Clinical trials registration. 2011. http://www.clinicaltrials.gov/ct2/info.
  23. 23.
    Collett D. Modeling survival data in medical research. 2nd ed. Boca Raton, FL: Chapman & Hall/CRC; 2003.Google Scholar
  24. 24.
    Collins PY, Patel V, Joestl SS, March D, Insel TR, Daar AS. Grand challenges in global mental health. Nature. 2011;475:27–30.PubMedCrossRefGoogle Scholar
  25. 25.
    Dahl SG, Sylte I. Molecular modeling of drug targets: the past, the present and the future. Basic Clin Pharmacol Toxicol. 2005;96:151–5.PubMedCrossRefGoogle Scholar
  26. 26.
    Dedeurwaerdere S, Wintmolders C, Straetemans R, Pemberton D, Langlois X. Memantine-induced brain activation as a model for the rapid screening of potential novel antipsychotic compounds: exemplified by activity of an mGlu2/3 receptor agonist. Psychopharmacology (Berl). 2011;214:505–14.CrossRefGoogle Scholar
  27. 27.
    Dolle RE, Bourdonnec BL, Worm K, Morales GA, Thomas CJ, Zhang W. Comprehensive survey of chemical libraries for drug discovery and chemical biology: 2009. J Comb Chem. 2010;12:765–806.PubMedCrossRefGoogle Scholar
  28. 28.
    Eddy DM. Evidence-based medicine: a unified approach. Health Aff. 2005;24:9–17.CrossRefGoogle Scholar
  29. 29.
    Faedda GL, Tondo L, Baldessarini RJ, Suppes T, Tohen M. Outcome after rapid vs. gradual discontinuation of lithium treatment in bipolar mood disorders. Arch Gen Psychiatry. 1993;50:448–55.PubMedCrossRefGoogle Scholar
  30. 30.
  31. 31.
    Fransen V, editor. Mental health services in the United States and England: struggling for change. Princeton, NJ: Robert Wood Johnson Foundation; 1991.Google Scholar
  32. 32.
    Gardner DM, Murphy AL, O’Donnell H, Centorrino F, Baldessarini RJ. International consensus study of antipsychotic dosing. Am J Psychiatry. 2010;167:686–93.PubMedCrossRefGoogle Scholar
  33. 33.
    Gershon ES, Alliey-Rodriguez N, Liu C. After GWAS: searching for genetic risk for schizophrenia and bipolar disorder. Am J Psychiatry. 2011;168:253–6.PubMedCrossRefGoogle Scholar
  34. 34.
    Glass GV. Integrating findings: the meta-analysis of research. Rev Res Educ. 1977;5:351–79.Google Scholar
  35. 35.
    Glick ID. Understanding the results of CATIE in the context of the field. CNS Spectr. 2006;11 Suppl 7:40–7.PubMedGoogle Scholar
  36. 36.
    Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER. Depression: a neglected major illness. J Clin Psychiatry. 1993;54:419–24.PubMedGoogle Scholar
  37. 37.
    Haenisch B, Bönisch H. Depression and antidepressants: insights from knockout of dopamine, serotonin or noradrenaline re-uptake transporters. Pharmacol Ther. 2011;129:352–68.PubMedCrossRefGoogle Scholar
  38. 38.
    Harcourt BE. The mentally ill, behind bars. New York Times, 15 Jan 2007, p. A19.Google Scholar
  39. 39.
    Harris G. Deal in place for inspecting foreign drugs. New York, NY: New York Times; 13 Aug 2011. http://www.nytimes.com/2011/08/13/science/13drug.html?scp=2&sq=Gardiner%20Harris&st=cse.
  40. 40.
    Healy D. The creation of psychopharmacology. Cambridge, MA: Harvard University Press; 2002.Google Scholar
  41. 41.
    Hennen J. Statistical methods for longitudinal research on bipolar disorders. Bipolar Disord. 2003;5:156–68.PubMedCrossRefGoogle Scholar
  42. 42.
    Heres S, Davis JM, Maino K, Jetzinger E, Kissling W, Leucht S. Why olanzapine beats risperidone, risperidone beats quetiapine, and quetiapine beats olanzapine: exploratory analysis of head-to-head comparison studies of second-generation antipsychotics. Am J Psychiatry. 2006;163:184–94.CrossRefGoogle Scholar
  43. 43.
    Himei A, Okamura T. Discontinuation syndrome associated with paroxetine in depressed patients: a retrospective analysis of factors involved in the occurrence of the syndrome. CNS Drugs. 2006;20:665–72.PubMedCrossRefGoogle Scholar
  44. 44.
    Hirsch L. Trial registration and results disclosure: impact of US legislation on sponsors, investigators, and medical journal editors. Curr Med Res Opin. 2008;24:1683–9.PubMedCrossRefGoogle Scholar
  45. 45.
    International Committee of Medical Journal Editors (ICJME). Uniform requirements for manuscripts. 2011. http://www.icmje.org/urm–main.html.
  46. 46.
    Isaac RJ, Armat VC. Madness in the streets. New York: Simon & Schuster; 1990.Google Scholar
  47. 47.
    Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, et al. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Arch Gen Psychiatry. 1994;51:8–19.PubMedCrossRefGoogle Scholar
  48. 48.
    Kiselyuk A, Farber-Katz S, Cohen T, Lee SH, Geron I, Azimi B, et al. Phenothiazine neuroleptics signal to the human insulin promoter as revealed by a novel high-throughput screen. J Biomol Screen. 2010;15:663–70.PubMedCrossRefGoogle Scholar
  49. 49.
    Lazo JS, Wipf P. Combinatorial chemistry and contemporary pharmacology. J Pharmacol Exp Ther. 2000;293:705–9.PubMedGoogle Scholar
  50. 50.
    Leucht S, Corves C, Arbter D, Engel RR, Li C, Davis JM. Second-generation versus first-generation antipsychotic drugs for schizophrenia: meta-analysis. Lancet. 2009;373:31–41.PubMedCrossRefGoogle Scholar
  51. 51.
    Leucht S, Heres S, Kissling W, Davis JM. Evidence-based pharmacotherapy of schizophrenia. Int J Neuropsychopharmacol. 2011;14:269–84.PubMedCrossRefGoogle Scholar
  52. 52.
    Lesch KP. Serotonergic gene expression and depression: implications for developing novel antidepressants. J Affect Disord. 2001;62:57–76.PubMedCrossRefGoogle Scholar
  53. 53.
    Lieberman JA, Stroup TS, McEvoy JP, Swartz MS, Rosenheck RA, Perkins DO, et al. Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) Investigators. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med. 2005;53:1209–23.CrossRefGoogle Scholar
  54. 54.
    Maggon K. Pharmaceutical market intelligence monograph. 2010. http://knol.google.com/k/global-pharmaceutical-market-intelligence-monograph.
  55. 55.
    Manji HK, Moore GJ, Chen G. Lithium up-regulates the cytoprotective protein Bcl-2 in the CNS in vivo: a role for neurotrophic and neuroprotective effects in manic depressive illness. J Clin Psychiatry. 2000;61 Suppl 9:82–96.PubMedGoogle Scholar
  56. 56.
    Mantripragada KK, Carroll LS, Williams NM. Experimental approaches for identifying schizophrenia risk genes. Curr Topics Behav Neurosci. 2010;4:587–610.CrossRefGoogle Scholar
  57. 57.
    Meyers M. Happy accidents: serendipity in modern medical breakthroughs. New York: Arcade; 2007.Google Scholar
  58. 58.
    Moreno SG, Sutton AJ, Turner H, Abrams KR, Cooper NJ, Palmer TM, et al. Novel methods to deal with publication biases: secondary analysis of FDA trial registry database and related journal publications. BMJ. 2009;339:2981–6.CrossRefGoogle Scholar
  59. 59.
    Pang KS, Durk MR. Physiologically-based pharmacokinetic modeling for absorption, transport, metabolism and excretion. J Pharmacokinet Pharmacodyn. 2010;37:591–615.PubMedCrossRefGoogle Scholar
  60. 60.
    Parker G, Brotchie H. Mood effects of the amino acids tryptophan and tyrosine. Acta Psychiatr Scand. 2011;124:417–26.PubMedCrossRefGoogle Scholar
  61. 61.
    Pecenak J. Relapse prevention in schizophrenia: evidence from long-term, randomized, double-blind clinical trials. NeuroEndocrinol Lett. 2007;28 Suppl 1:49–70.PubMedGoogle Scholar
  62. 62.
    Podhorna J. Experimental pharmacotherapy of benzodiazepine withdrawal. Curr Pharm Des. 2002;8:23–43.PubMedCrossRefGoogle Scholar
  63. 63.
    Rief W, Nestoriuc Y, von Lilienfeld-Toal A, Dogan I, Schreiber F, Hofmann SG, et al. Differences in adverse effect reporting in placebo groups in SSRI and tricylcic antidepressant trials: systematic review and meta-analysis. Drug Saf. 2009;32:1041–56.PubMedCrossRefGoogle Scholar
  64. 64.
    Sackett DL, Rosenberg WM. The need for evidence-based medicine. J R Soc Med. 1995;88:620–4.PubMedGoogle Scholar
  65. 65.
    Salvatore P, Baldessarini RJ, Tohen M, Khalsa HMK, Perez Sanchez-Toledo J, Zarate Jr CA, et al. McLean-Harvard International First-Episode Project: Two-year stability of DSM-IV diagnoses in 500 first-episode psychotic disorder patients. J Clin Psychiatry. 2009;70:458–66.PubMedCrossRefGoogle Scholar
  66. 66.
    Schou M, Sóvágó J, Pike VW, Gulyás B, Bøgesø KP, Farde L, et al. Synthesis and positron emission tomography evaluation of three norepinephrine transporter radioligands: [11C]desipramine, [11C]talopram and [11C]talsupram. Mol Imaging Biol. 2006;8:1–8.PubMedCrossRefGoogle Scholar
  67. 67.
    Shakespeare W. Macbeth (published ca. 1603–1611). In: Boyce (editor). Encyclopedia of Shakespeare. New York: Roundtable Press.Google Scholar
  68. 68.
    Sit SY. Dopamine agonists in the treatment of Parkinson’s disease past, present and future. Curr Pharmaceut Design. 2000;6:1211–48.CrossRefGoogle Scholar
  69. 69.
    Stoll AL, Tohen M, Baldessarini RJ. Increasing diagnostic frequency of obsessive-compulsive disorder. Am J Psychiatry. 1992;149:38–640.Google Scholar
  70. 70.
    Straus SE, Richardson WS, Glasziou P, Haynes RB. Evidence-based medicine. 3rd ed. Edinburgh: Churchill-Livingstone; 2005.Google Scholar
  71. 71.
    Swann AC. Long-term treatment in bipolar disorder. J Clin Psychiatry. 2005;66(Suppl):7–12.PubMedGoogle Scholar
  72. 72.
    Swazey J. Chlorpromazine in psychiatry: a study in therapeutic innovation. Cambridge, MA: MIT; 1974.Google Scholar
  73. 73.
    TalentMash. Generic drugs: yes or no? 2011. http://talentmash.com/generic-drugs-yes-or-no.
  74. 74.
    Tedeschini E, Fava M, Papakostas GI. Placebo-controlled, antidepressant clinical trials cannot be shortened to less than 4 weeks’ duration: a pooled analysis of randomized clinical trials employing a diagnostic odds ratio-based approach. J Clin Psychiatry. 2011;72:98–118.PubMedCrossRefGoogle Scholar
  75. 75.
    Tohen M, Zarate Jr CA, Hennen J, Kaur Khalsa HM, Strakowski SM, Gebre-Medhin P, et al. The McLean-Harvard First-Episode Mania Study: prediction of recovery and first recurrence. Am J Psychiatry. 2003;160:2099–107.PubMedCrossRefGoogle Scholar
  76. 76.
    Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selective publication of antidepressant trials and its influence on apparent efficacy. N Eng J Med. 2008;358:252–90.CrossRefGoogle Scholar
  77. 77.
    Undurraga J, Baldessarini RJ. A 30-year meta-analytic review of antidepressant efficacy in acute major depression. Neuropsychopharmacology. 2012;37:851–64.PubMedCrossRefGoogle Scholar
  78. 78.
    Vázquez G, Baldessarini RJ, Yildiz A, Tamayo J, Tondo L, Salvatore P. Multi-site international collaborative clinical trials in mania: commentary. Int J Neuropsychopharmacol. 2011;14:1013–6.PubMedCrossRefGoogle Scholar
  79. 79.
    Vedula SS, Bero L, Scherer RW, Dickersin K. Outcome reporting in industry-sponsored trials of gabapentin. N Engl J Med. 2009;361:1963–71.PubMedCrossRefGoogle Scholar
  80. 80.
    Viguera AC, Baldessarini RJ, Hegarty JD, van Kammen DP, Tohen M. Clinical risk following abrupt and gradual withdrawal of maintenance neuroleptic treatment. Arch Gen Psychiatry. 1997;54:49–55.PubMedCrossRefGoogle Scholar
  81. 81.
    Viguera AC, Baldessarini RJ, Friedberg J. Risks of interrupting continuation or maintenance treatment with antidepressants in major depressive disorders. Harv Rev Psychiatry. 1998;5:293–306.PubMedCrossRefGoogle Scholar
  82. 82.
    Yildiz A, Vieta E, Leucht S, Baldessarini RJ. Efficacy of antimanic treatments: meta-analysis of randomized, controlled trials. Neuropsychopharmacology. 2011;36:375–89.PubMedCrossRefGoogle Scholar
  83. 83.
    Yildiz A, Vieta E, Tohen M, Baldessarini RJ. Factors modifying drug and placebo responses in randomized trials for bipolar mania. Int J Neuropsychopharmacol. 2011;14:863–75.PubMedCrossRefGoogle Scholar
  84. 84.
    Wang J, Maurer L. Positron emission tomography: applications in drug discovery and drug development. Curr Topics Med Chem. 2005;5:1053–75.CrossRefGoogle Scholar
  85. 85.
    Weaver DF, Weaver CA. Exploring neurotherapeutic space: how many neurological drugs exist (or could exist)? J Pharm Pharmacol. 2011;63:136–9.PubMedCrossRefGoogle Scholar
  86. 86.
    Wistedt B, Wiles D, Kolakowska T. Slow decline of plasma drug and prolactin levels after discontinuation of chronic treatment with depot neuroleptics. Lancet. 1981;1(8230):1163–4.PubMedCrossRefGoogle Scholar
  87. 87.
    Wittchen HU, Jacobi F, Rehm J, Gustavsson A, Svensson M, Jönsson B, Olesen J. Size and burden of mental disorders and other disorders of the brain in Europe in 2010. Eur Neuropsychopharmacol. 2011;21:655–79.PubMedCrossRefGoogle Scholar
  88. 88.
    Woolley SB, Cardoni AA, Goethe JW. Last-observation-carried-forward imputation method in clinical efficacy trials: review of 352 antidepressant studies. Pharmacotherapy. 2009;29: 1408–16.PubMedCrossRefGoogle Scholar
  89. 89.
    World Health Organization (WHO). International clinical trials registry platform (ICTRP). 2011. http://www.who.int/ictrp/en/.

Copyright information

© Springer Science+Business Media New York 2013

Authors and Affiliations

  • Ross J. Baldessarini
    • 1
  1. 1.Harvard Medical SchoolMcLean Hospital Psychopharmacology ProgramBelmontUSA

Personalised recommendations