Jitschak Storosum, M.D.
Bibliography

Thomas P. Laughren, M.D.
Bibliography

Fouzia Laghrissi-Thode, M.D.
Bibliography

Regulatory, Scientific and Clinical Issues in Drug Development

Earl L. Giller, Jr., M.D., Ph.D.
Bibliography

Abstract
Regulatory, scientific and clinical perspectives agree in principle about the critical information that a drug development program must deliver. In acute treatment, the efficacy objective is to demonstrate how close the medication can come to rapid and complete removal of symptoms, with sustained absence of these target symptoms without the appearance of other symptoms that may be medication-related. For bipolar disorder, in particular, this means treatment of manic symptoms without exacerbation or precipitation of depressive symptoms, and/or treatment of depression without development of mania. For any new medication, potentially drug-related serious and significant adverse events must be at an acceptable level as balanced against the benefit of the drug. For bipolar disorder in particular, because of the frequent use of combination therapy, information about pharmacokinetic and pharmacodynamic drug interactions in acute and long-term treatment is critical.

Scientists, regulators and clinicians, however, often differ in emphasis. The scientific community, through frequent meetings and peer-reviewed literature and grant support, has perhaps the most robust method for ongoing global consensus development. Some informative experiments (i.e., clinical trials) from a scientific perspective, however, may not be ethical or feasible. Standard clinical trials evaluate medications by group rather than individual comparisons. Regulatory agencies have a public health perspective to evaluate the benefit/risk ratio of a new medication. They tend to be country or region specific rather than global, although ICH is moving towards global standards. Regulators are primarily responsible to legislative oversight, which may reinforce a risk aversive stance, rather than being guided by peer review, scientists or clinicians.

These differences contribute to the conservative nature of current clinical trial designs.

• Identical duplicate trials are usually done to replicate efficacy and safety data rather than risking some variation that could provide new clinically useful information. This usually means two placebo-controlled trials instead of one placebo-controlled and an alternative demonstration of efficacy.

• Subjects are as “clean” as possible, i.e., without the usual psychiatric and medical comorbidity seen in practice.

• Trial lengths are not informed by time to establish efficacy.

• The same efficacy standard is used for staged or add-on combination therapy as is used for simultaneous start of combination therapy. The former design will likely show a greater treatment effect but with incomplete responders. The latter design is closer to clinical practice, but likely to show a smaller treatment effect.

A global development program is complicated by different regional regulatory requirements. The European regulatory environment is in flux. Different terms are used for treatment length, and relapse and recurrence are not as easily defined for bipolar disorder as for unipolar depression. There is a perceived, although perhaps dated, perception that the treatment of acute mania is different from long-term treatment in Europe. This difference, in combination with an interest in efficacy and safety beyond the acute treatment of mania, has led to a strong requirement for a 12-week haloperidol-controlled trial for registering a novel antipsychotic for the acute treatment of mania.

Meetings and sessions such as this one have helped us all to develop better strategies for the development of new medications for bipolar disorder. I would submit that similar sessions across multiple venues, hopefully with a tripartite panel to provide continuity and refine guidelines, would help develop these strategies even further.

Alan Metz, M.D.
Bibliography

Abstract
The past decade has seen a resurgence in the interest in studying bipolar disorder, particularly from within the pharmaceutical industry. Following four decades in which only two treatments, lithium and carbamazepine, were studied systematically and received regulatory approvals, there are now well-controlled studies of several potential new treatments for bipolar disorder from a variety of drug classes, especially the anticonvulsants (e.g. valproate, lamotrigine, gabapentin, topiramate) and the atypical antipsychotics (e.g. risperidone, olanzapine, ziprasidone). Furthermore, interest has increased in studying phases of the illness that go beyond the acute treatment of mania, into other areas including bipolar depression and maintenance treatment. Until recently, some of these illness phases had not been well-studied in controlled clinical trials, and there was little clinical or regulatory precedent from which to take guidance in the design and implementation of such studies. In the US, FDA guidelines pertaining to the study of bipolar disorder are more than 20 years old - current regulatory guidance may best be sought from consultation with FDA and/or study of recent approval documents available via the U.S. Freedom of Information Act. However, over the course of even a few years, it has been our company's experience that the FDA stance on regulatory issues may change dramatically, as evidenced by recent communication addressing the nature of bipolar depression and the issue of "pseudospecificity" potentially raised with respect to unipolar depression. On the whole, the FDA approach might be summarized as requiring acute efficacy data from multiple studies along with adequate dosing and safety data for the population and duration of use intended, with maintenance efficacy data being considered useful but not essential for granting an initial indication. In Europe, the situation is quite different. EMEA is in the process of finalizing guidelines for the study of bipolar disorder which would make it essential for sponsors to provide data on maintenance treatment for 3-6 months depending on illness phase. In many cases comparative data against existing treatments are also required. Reconciling these differing requirements poses a tremendous challenge to the pharmaceutical industry sponsor seeking global access for novel treatment strategies. This presentation will cover examples of these challenges, such as the historical unwillingness of the FDA to consider maintenance treatment indications in the absence of acute efficacy claims in psychiatry, and examples of how regulators and industry could cooperate in the attempt to expedite the availability of novel potential treatments needed for the current and future generations of bipolar disorder sufferers.

The Ethical Dilemma in the Development of a Mood Stabilizer, a Need for Alternative Designs

Willem A. Nolen, M.D., Ph.D.
Bibliography

Abstract
Scientific clinical research in psychiatry has to fulfill same ethical and legal requirements as research in somatic medicine: studies have to be valid; studied subjects have to give informed consent; and studies have to be approved by an investigational review board. With respect to the criterion of validity, there is no motive to give research in psychiatry a special place: it has to fulfill similar methodological requirements of science. In addition, although some psychiatric patients can not express their free will, most psychiatric patients are capable to judge their participation in scientific clinical research and to give informed consent as well as most patients with somatic diseases. Nevertheless, performing treatment studies in psychiatry is difficult, especially in severe disorders such as bipolar disorder with its chronic and remitting course. This is shown by the fact that since lithium no drug has really been shown an effective mood stabilizer. This despite the fact that a number of drugs, especially anticonvulsants such as carbamazepine, valproate and lamotrogine, have been found efficacious in some phases of this disorder.

What is an effective mood stabilizer?
To be considered a mood stabilizer a drug should be effective in treating acute mania without causing depression; effective in treating acute bipolar depression without causing mania; effective in preventing recurrences of both mania and depression (i.e. in prophylaxis); and preferentially also effective in rapid cycling. From a scientific point of view, efficacy of a new drug can only be shown in randomized, controlled, double-blind studies (RCTs), either versus placebo to proof that it is efficacious (i.e. more effective than placebo) or versus a standard treatment to proof that it has advantages over the standard treatment (i.e. better efficacy). Another option would be that a new drug is studied in order to show that it is non-inferior to a standard treatment, which means that it has to be shown that it is no worse in terms of efficacy. Besides that such studies should have large number of patients to be included, many other factors could also obscure a possible difference, such as inclusion of not sick enough patients, choice of wrong outcome criteria and insufficient inter-rater reliability on rating scales between investigators. Therefore, non-inferiority studies are not recommended by the European Agency for the Evaluation of Medicinal Products (EMEA) .

European registration requirements
Overall, the European requirements are more stringent than the US requirements of the Food & Drug Association (FDA) who has accepted several drugs for the treatment of bipolar disorder during the last years. So far, the EMEA requires three armed RCTs in which a new drug is compared both with placebo and with a standard treatment. For registration as a mood stabilizer EMEA requires that efficacy has been shown in acute episodes (mania and bipolar depression) as well as in prophylaxis, although recently they also intend to accept registration of drugs for the treatment of acute mania. Moreover it is required by the EMEA that efficacy has been shown in monotherapy. Efficacy in acute episodes means that the drug leads to response, i.e. a clinical relevant reduction of manic or depressive symptoms as measured with rating scales. Moreover with continuation of treatment no relapse and no worsening of the other pole may occur. In acute mania a new drug should be compared with placebo and lithium or an antipsychotic as standard treatment. The acute phase of the study should last 3-4 weeks and the continuation phase at least 12 weeks.

In acute bipolar depression the standard treatment should be lithium, but an antidepressant could also be considered although such a drug implies the risk of inducing switches into mania. The acute phase of the study should last 6-8 weeks, and the continuation phase at least 3-6 months.  Prophylactic efficacy means that the drug is effective in the prevention of recurrences of both manic and depressive episodes. This has to be shown in patients after recovery from both manic and depressive index episodes. The duration of treatment in such a study should depend on the expected recurrence rate in the population studied, but at least one year or until outcome is reached (e.g. a new episode).

Problems
There are several problems with these requirements of the EMEA. First, eligible patients for such studies hardly exist. One would like to include patients with well defined bipolar disorder, preferentially without severe co-morbidity and severe suicidality. However many if not most patients with bipolar disorder have co-morbid disorders, such as anxiety disorder or substance abuse, or have been suicidal before. In addition, their illness often has been complicated with psychotic episodes or a more or less chaotic course pattern with mixed episodes or rapid cycling. Thus, inclusion of “pure” bipolar patients in RCTs limits generalizability of the results to the large group of “real” bipolar patients . Second, a number of methodological problems are associated with placebo-controlled studies, questioning their validity . Third, it is very difficult if not impossible to find eligible patients who can be randomized to placebo. This especially is a problem as lithium is available as an effective standard treatment. According to the latest version of the Declaration of Helsinki (amendment Edinburgh 2000) there is a restriction on the use of placebo: “[….] a new method [i.e. a new drug; WN], should be tested against those of the best current prophylactic [….] or therapeutic methods”. As lithium is an effective treatment, patients should consequently be offered a study is which a new drug is tested against lithium and not against placebo. One possible solution to overcome this problem would be to select patients who are refractory or intolerant to lithium or who refuse lithium, but results from such a study can not be generalized to the total population of bipolar patients. Another option would be a placebo-controlled add-on studies, in which either a new drug or placebo is added to an ongoing treatment in patients who have not responded satisfactory to that treatment. We and others have discussed several alternative designs for such studies 2 . However, besides the problem associated with concurrent medications, i.e. pharmacokinetic and pharmacodynamic interactions, generalizability of such a studies is also limited. So far the EMEA does not accept such designs in the pivotal registration studies 1. Moreover, it could only lead to restricted registrations, i.e. for the treatment of refractory patients, or as add-on treatment respectively.

Ethical dilemma
As a consequence, and despite the Declaration of Helsinki, placebo-controlled, monotherapy studies are still asked for by registration authorities and still proposed by pharmaceutical companies, in order to proof that a drug is a mood stabilizer. Such studies include acute treatment of mania and bipolar depression, and prophylaxis. However, many patients and/or their treating physicians decline such studies because of the risks associated with the use of placebo. A major risk could be suicide associated with lack of response. So far, in acute unipolar depression rates of suicides have not been found different between placebo and active compounds in RCT’s involving placebo, probably by exclusion of patient with serious suicide risk . However, there are other risks associated with the use of placebo: besides the chance that they do not recover while participating in an acute study, they have the chance that they develop a new episode in a prophylactic study.

Especially the latter may be a major problem, as new manic or depressive episodes carry a serious risk for patients, not only the direct and often large psychological and social consequences from an episode, but also the chance that a new episode increases the vulnerability for further episodes, as suggested by the kindling hypothesis. It is clear that there exist an ethical dilemma: the demand from a scientific point of view and supported by the registration authorities, to perform placebo-controlled monotherapy studies versus the problem that such studies can be harmful for patients. How big this dilemma is, is shown by the fact that since the seventies no successful placebo-controlled monotherapy prophylactic study has been performed, i.e. no study leading to a positive result on the primary chosen outcome measure. One suggested solution to overcome this dilemma of not being able to perform rigorous placebo-controlled RCTs, is the option to exclude very severe patients. This point has probably led to the negative result (no efficacy on the primary outcome measure) of the long-term study with valproate versus lithium and placebo by Bowden et al. . Another option, especially for prophylactic studies, is to broaden the outcome criterium: e.g. not to wait until a new full episode has to developed, but to include also prodromal symptoms as early signs of an episode. A possible outcome could then be either a new episode or prodromal signs leading to the decision of the physician to intervene, e.g. with extra medication. However, this approach may have contributed to the negative result of the long-term study with lamotrigine versus placebo in bipolar patients with rapid cycling by Calabrese et al. .

Are placebo-controlled studies permissible?
Placebo-controlled monotherapy studies in bipolar disorder are necessary to proof that a drug is efficacious. According to Miller the use of placebo is only justified when risks are minimized and patients are offered short-term treatment optimization after completion of the study 4. To my opinion, the use of placebo is also ethically permissible, but only in acute studies in mania or bipolar depression, and involving not too severely ill patients. In such populations it is ethically justified to compare a new drug with placebo (and a standard treatment) under the condition the new drug or standard treatment can be given to all non-responders immediately after the study period of 3(-4) weeks (in mania) or 6(-8) weeks (in depression) and that patients who worsen during the study can stop prematurely. In contrast however, long-term prophylactic studies are not permissible, while from a practical point of view such studies have appeared almost impossible to perform, which even increases the ethical dilemma. Registration authorities should acknowledge that the dilemma exists, by accepting alternative designs and weakening their requirements for registration of drugs intended for the treatment of bipolar disorder.

References
1.  European Agency for the Evaluation of Medicinal Products. Note for guidance on clinical investigation of medicinal products for the treatment and prevention of bipolar disorder. EMEA, London, 2001.

2.  Rush AJ, Post RM, Nolen WA, Keck PE, Suppes T, Altshuler L, McElroy SL. Methodological issues in developing new acute treatments for patients with bipolar illness. Biol Psychiatry 2000;48: 615-24.

3.  Lavori PW. Placebo control groups in randomized treatment trials: a statistician's perspective. Biol Psychiatry 2000; 47: 717-23.

4.  Miller FG. Placebo-controlled trials in psychiatric research: an ethical perspective. Biol Psychiatry 2000; 47:707-16.

5.  Kahn A, Warner HA, Brown WA. Symptom reduction and suicide risk in patients treated with placebo in antidepressant clinical trails: an analysis of the Food and Drug Administration database. Arch Gen Psychiatry 2000; 57: 311-7.

6.  Bowden CL, Calabrese JR, McElroy SL, Gyulai L, Wassef A, Petty F, Pope HG, Chou JC, Keck PE, Rhodes LJ, Swann AC, Hirschfeld RM, Wozniak PJ. A randomized, placebo-controlled 12-month trial of divalproex and lithium in treatment of outpatients with bipolar I disorder. Divalproex Maintenance Study Group. Arch Gen Psychiatry 2000;57:481-9.

7.  Calabrese JR, Suppes T, Bowden CL, Sachs GS, Swann AC, McElroy SL, Kusumakar V, Ascher JA, Earl NL, Greene PL, Monaghan ET. A double-blind, placebo-controlled, prophylaxis study of lamotrigine in rapid-cycling bipolar disorder. Lamictal 614 Study Group. J Clin Psychiatry 2000; 61:841-50.
 

Top of Page