Considerations for successful clinical development for orphan indications

Abstract

Clinical trials (programs) for orphan indications present unique challenges for the drug developer. The fact that the disease is often poorly characterised and under-researched makes it difficult to define suitable clinical parameters; the paucity of patients and specialist centres create logistical issues; the fact that the regulatory framework is not always well suited to novel therapies in a small and unstudied patient group requires innovative thinking and interaction with the regulators to achieve an appropriate study design. These factors all need to be given adequate consideration when developing treatments for orphan indications.

Keywords::

• clinical trial
• orphan drug
• rare disease

1. Introduction

Successful design of clinical trials is as much an art as a science. Even with extensive knowledge and experience, people still get it wrong. A poorly designed clinical trial may be impractical and fail to recruit patients, it may select the wrong end points, it may be underpowered and fail to show an effect even though one exists or it may not address the aspects which the regulators require and therefore be useless for marketing authorisation (MA), which is the ultimate goal of every drug development program.

In addition to the usual challenges of designing a successful clinical trial, orphan drug studies contain their own unique challenges. This article will outline some of the main issues to consider when designing a clinical trial or clinical program for an orphan drug.

2. Animal models of the disease

Animal models are very often lacking for rare diseases, creating problems of how to study potential candidates for therapy. It may also impact the design of clinical trials, when selecting dose and outcome measures and may preclude early application for orphan designation, since some evidence of effect (preclinical or clinical) is generally required. Developing an animal model early in the process may prove valuable later.

3. Knowledge of the natural history of the disease

Many rare diseases have not been well studied and there may be very little in the literature on the natural course of the disease; much of the ‘evidence' may be anecdotal or case studies. This can make it extremely difficult to estimate the expected effect size of a therapy and to decide on the most appropriate duration for study. Oftentimes, one would be learning about the clinical outcomes of the disease during the conduct of the clinical study. It may be useful to conduct a brief observational study prior to the clinical study to better understand the natural history of the disease, determine potential clinical outcomes of the study and provide data to create historical control groups. However, this may not always be feasible. A detailed discussion of this issue can be found in the summary of a recent workshop on natural history studies of rare diseases [1].

4. Regulatory paradigm

The regulatory pathway determines the amount of work that one would have to do. There is a common belief, both within and outside the pharmaceutical industry, that it is quicker, easier and cheaper to develop an orphan drug compared with other drugs. While there may, in some cases, be some truth to this, it is certainly not a given. This issue is discussed in detail in a recent book on rare diseases, but it will be mentioned here briefly [2].

First, contrary to common belief, regulators will not automatically accept one small pivotal trial for registration of an orphan drug. In fact, the regulators officially require the same level of proof of safety and efficacy for orphan drugs as for other drugs.

The European Medicines Agency (EMA) has produced a guidance document that outlines the circumstances under which a single pivotal clinical trial may be acceptable for MA [3]. It is a stand-alone guidance document, in principle applicable to any drug. In cases where the sponsor provides evidence from only one pivotal study, the study must be exceptionally compelling with respect to internal and external validity, clinical relevance, degree of statistical significance, data quality and internal consistency. The Food and Drug Administration (FDA) makes similar caveats in its guidance documents.

Therefore, even in a rare disease, multiple confirmatory studies might be necessary and it is crucial to take this into account when embarking on the clinical trial program. The regulators do exercise some degree of flexibility in the assessment of orphan drugs, as evidenced by the results of a recent study by National Organization for Rare Disorders, which looked at 135 approved non-cancer orphan therapies and found that 45 (33%) were classified as ‘Conventional' approval, 32 (24%) as ‘Administrative Flexibility' and 58 (43%) as ‘Case-by-Case Flexibility' [4]. The EMA has also adopted a somewhat flexible approach to the granting of MA in many cases, always providing that the risk–benefit assessment is positive.

Second, orphan drugs are not automatically eligible for expedited approval procedures. Both the EMA and FDA have special procedures designed to speed up the availability of new drugs for patients with serious diseases and unmet medical need, but the procedures are neither specific nor automatic for orphan drugs. However, it is clear that orphan drugs are likely to qualify more often than drugs for common indications where many treatments are already available.

Third, orphan drugs are not automatically eligible for conditional approval – a misconception which can lead a company to assume that it does not need to provide a complete and robust set of clinical data prior to MA. Once again, the procedure is neither specific nor automatic for orphan drugs but is often particularly applicable to these products, because of the difficulty in collecting sufficient robust clinical data in these populations.

In some cases, when a product is in a truly novel area, the regulators may also be in uncharted territory. The recent example of Glybera illustrates this; Glybera is a gene therapy, developed for the treatment of lipoprotein lipase deficiency, a very rare genetic disorder. It was initially refused a MA by the EMA but the decision was later reversed to grant a MA in a restricted population under exceptional circumstances. The EMA Assessment Report noted [5]:

The Glybera application also needs to be considered in the context of an extremely orphan indication with very limited patients together with a new emerging era of gene therapy medicinal products in clinical practice together with a limited but evolving scientific knowledge. Taking all this into account, and compared to other ‘new concepts' in the past, (for example monoclonal antibodies), it may be unrealistic to expect a similar level of evidence for the demonstration of quality, safety and efficacy and the overall benefit/risk as for ‘classical' medicinal products.

5. Clinical efficacy measures

It can often be difficult to decide on the most appropriate efficacy measures for any clinical trial, but for many (common) conditions, there are guidance documents which provide considerable detail on the expected development program, the size and number of trials and the efficacy end points that are considered acceptable. For orphan drugs, this is not the case and it can be a stumbling block for drug developers. Indeed, a recent study has shown that the two most common characteristics of the clinical trial design which correlate with failure to achieve MA for orphan drugs in the United States are [6]:

• failing to achieve the primary end point in the pivotal trial (odds ratio [OR] = 25.7; 95% CI: 5.27 – 125.1) and

• not identifying the most appropriate target population (OR = 20.0; 95% CI: 2.10 – 190.9).

This example shows how important it is to properly define the research question at the outset and is one of the reasons why it is crucial to have early interaction with the regulators; indeed, the same study has shown this to be a key factor in the success or failure to obtain approval and this is supported by an FDA study on the characteristics of rare disease marketing applications associated with FDA product approvals [7].

In Europe, Scientific Advice (or Protocol Assistance as it is known for orphan drugs) has to be requested and is quite expensive (currently EUR 82,400 for an initial request and 50% that amount again for follow-up requests). However, under the incentives of the orphan legislation, there are significant fee reductions for drugs which have been granted orphan designation. For small and medium-sized enterprises, there is a 100% fee reduction and this is one of the reasons why companies often aim to obtain orphan designation at a fairly early stage of development.

In the United States, interaction with the FDA is standard. Early communication with FDA is encouraged and is particularly important for development of drugs for orphan indications. A pre-investigational (pre-IND) new drug meeting can be a useful forum to discuss and agree on the nature of preclinical studies and early phase clinical studies. Other types of meeting [end-of-Phase I; end-of-Phase II; pre-New Drug Application (pre-NDA) and pre-Biologics License Application (pre-BLA)] are critical for promoting ongoing and interactive discussion with the FDA. Ad-hoc meetings are also encouraged to discuss results from completed trials and potential changes to the clinical development program. Early and often communication with the regulatory agency will provide structure, guidance and understanding of the regulatory expectations and challenges for marketing approval.

6. Surrogate end points

A surrogate end point can be defined as a biomarker intended to substitute for a clinical end point – a direct measure of patient benefit (e.g., survival). ‘Surrogate endpoints may be used as primary endpoints when appropriate (when the surrogate is reasonably likely or well known to predict clinical outcome)' [8].

Surrogate end points can be very useful during a development program, particularly for rare diseases where small available patient numbers prevent large-scale outcome trials and where the natural history of the disease may be poorly understood. But they also have some disadvantages:

• Clinical relevance: Surrogate end points may not be a true predictor of clinical outcome. The preferred clinical outcome may be difficult to study, and post-marketing studies may be needed to further characterise the relationship between the surrogate and clinical end points.

• Risk–benefit: It may be difficult to measure clinical benefit against adverse effects. The clinical benefit may be difficult to define within the scope of early clinical trials, and small trials are likely to be powered to detect only common adverse effects.

• Reliability: The relationship with clinical outcome may vary between drugs/classes.

Potential markers for surrogate end points should be evaluated in the context of the disease process. For example, a promising surrogate end point may be a biological marker influenced by the disease pathway. To be considered acceptable by the regulatory authorities, a surrogate end point needs to be validated; the validation of a surrogate end point is described in International Conference on Harmonisation E8 and in the EMA Guideline on Clinical Trials in Small Populations [9]. The EMA guidance also states that, for rare diseases in certain circumstances, other surrogate markers may be acceptable provided the surrogate marker is ‘reasonably likely – based on epidemiologic, pathophysiologic, or other evidence – to predict benefit'. The FDA also allows for approval based on surrogate end points, but under certain conditions the agency may also require post-approval studies to demonstrate clinical benefit [10].

7. Statistical tools

The statistical challenges when working with small trials often requires creative solutions in both the design and statistical analysis. Unlike common diseases where the sample size can be increased to account for large variability, the ability to conduct a larger trial in an orphan disease is limited by the number of patients who have the disease, who are located in close proximity to a study site and who are eligible for enrolment. It is often difficult, if not impossible for some indications, to control for variability by increasing the sample size. Thus, it is critical to control for variability as much as possible through both the study design and the statistical analysis.

The aforementioned EMA guidance document provides some guidance regarding trial designs and statistical approaches that may be used for studies in small populations. For example, one can reduce variability by choosing design strategies such as stratification in the randomisation, adaptive designs, sequential designs, crossover designs, and N-of-1 trials. Each of these approaches has its own limitations and should be carefully examined to determine whether the approach would be suitable for the specific disease being studied.

Variability can also be controlled to some extent through appropriate analysis methods. The use of analyses for repeated measures, such as a mixed model repeated measures analysis, can be applied to many trials that collect data at several visits during the course of the trial. Enrichment strategies that analyse data according to meaningful prognostic variables and including covariates for important baseline measures can also help in controlling sources of variation within the data. The use of nonparametric analyses and sensitivity analyses are encouraged to assess the robustness of results of underlying model assumptions. Reporting data using confidence intervals is recommended to assist in the clinical interpretation of results.

8. Logistical issues

Owing to the small number of patients available for study and the likely very small number of specialist centres, the protocol needs to be realistic about the number of sites and the recruitment rate.

Here, the involvement of patient groups and foundations can be very helpful; indeed, these groups should be involved from an early stage in the whole protocol development process, as they are often the best ones to define relevant clinical end points, identify specialist centres and disseminate information about the study to patients.

9. Ethical issues

For many rare diseases, there is no approved therapy. In such a situation, placebo-controlled trials may be acceptable. It does, however, raise serious ethical issues when a condition is life-threatening or seriously debilitating, as it is by definition for an orphan disease. Both patients and Ethics Committees may object to the inclusion of a placebo group and the issue becomes even tougher with the inclusion of children.

Placebo-controlled studies that allow for crossover to a treatment intervention after a certain period of time or which provide long-term treatment follow up may alleviate concerns about having access to potentially beneficial treatments.

10. Conclusion

The design of clinical trials (programs) for rare diseases is a complex and specialist area and the above touches only lightly on a few key issues. The developer is often in uncharted waters and will need to consider multiple unique factors and be innovative while navigating the complicated regulatory passageways.

Declaration of interest

AK Hall is the part owner of a CRO specialising in orphan drugs. E Ludington is VP Biometrics at a CRO providing orphan drug services. The authors state no conflict of interest and have received no payment in preparation of this manuscript.