Europe's Thirst for Trial Information

At the end of 2008, everyone conducting clinical trials in Europe will have to switch to a new format for applications for authorization, for introducing substantial amendments, and for declaring the end of a trial. The changes, which are being made by the European Union to take account of recent regulatory developments, underline Europe's insatiable thirst for increased volumes of information about clinical trials.

Peter O'Donnell

One of the prompts for the change is the coming into force of the new rules on pediatric medicines (on which this column has provided extensive coverage). But the concerns over first-in-man trials are also a factor in imposing the new format. Last year, the EU's Medicines Agency in London published a guideline on reducing risk in these trials, and trial authorization applications will have to be adapted accordingly.

There is a third motivation too, which the EU rather primly refers to as "to clarify the submission of information on the results of clinical trials in view of its publication"—another gesture toward the continuing controversy over the comprehensiveness, accessibility, and impartiality of the information available on medicines.

Under the new format, revised during February, the "request for authorization of a clinical trial on a medicinal product for human use to the competent authorities and for opinion of the ethics committees in the Community" will be still more onerous to complete. It will, for instance, no longer be enough to provide the full title of the trial. Additional requirements include "title of the trial for lay people, in easily understood language," and "name or abbreviated title of the trial when used." In the section on general information on the trial, it is no longer enough to "specify the medical condition(s) to be investigated." Now, there is also the obligation to indicate the "medical condition in easily understood language," as well as the therapeutic area.

A new section entitled "Source(s) of monetary or material support" seeks information on the name of the organization and the country, and another new section entitled "Contact point designated by the sponsor for further information on the trial" demands the name of the organization, the contact point (such as "Clinical Trial Information Desk," it gives as an example), and details of address, phone, fax, and email.

Similarly, the application will have to make clear whether the trial is part of a Pediatric Investigation Plan (and if so, supply the European Medicines Agency's decision number for the plan). Under the section on clinical trial sites, there is a new question relating to "networks to be involved in the trial" (for which the example is given of "pediatric networks involved in the trial"), with a request for details of the organization, its activities, and its contact details.

The section on investigative medicinal products also contains some additional questions. "Is it an IMP to be used in a first-in-human clinical trial?" and "If yes, are there risk factors identified, according to the guidance [issued by the EMEA on first in humans]?"

Information has to be given not only on which countries in the European Economic Area (the EEA includes 27 EU member states plus Iceland, Liechtenstein, and Norway) are hosting the trial, but one of the new questions asks: "Is the trial being conducted completely outside of the EEA?" and "If yes, specify the regions in which trial sites are planned."

It will be necessary to supply information on the proposed date of start of recruitment in all the countries concerned, and the person signing the application has also to undertake to "submit the date of inclusion of the first subject." And if a substantial amendment concerns several trials involving the same IMP, then it is now necessary to provide the trials' numbers from the EU's EudraCT database of trials.

The new format is currently available for consultation and comment (

). And the new Annexes 1, 2, and 3 on the Clinical Trial Application Form, the Substantial Amendment Form, and the Declaration of the end of a Clinical Trial Form are due to be implemented "toward the end of 2008," says the EU.

Progression-free survival

Not all the changes that the EU makes are merely bureaucratic. The European Medicines Agency has just produced some thoughtful guidance on using progression-free survival (PFS) as a primary endpoint in clinical efficacy trials for anticancer products. It takes the form of an appendix to existing EMEA guidelines on evaluating anticancer products in man and clarifies issues such as the endpoint definition, frequency and methods of assessment, ascertainment bias, handling of deviations and missing data, and radiology review. It will come into effect this coming August.

Its starting point is that the current approach to definitions may need some modification. PFS is traditionally defined as the time from randomization (or registration, in nonrandomized trials) to objective tumor progression, or death from any cause, whichever first. Disease-free progression (DFS) is defined as the time from randomization to objective recurrence or death from any cause. Disease progression and recurrence are typically assessed based on objective radiological findings, and the definition of progression should follow established response evaluation criteria such as the Response Evaluation Criteria In Solid Tumors or the criteria from the World Health Organization.

However, the guidance acknowledges that modified criteria might in some cases be more appropriate, depending on the type of agent, the site and type of lesion, and the objectives of the trial. For instance, it suggests additional objective clinical and biochemical or radiological criteria may be used to assess progression. The bottom line for the EMEA is that, "In all cases, it is important that the criteria for definition of a progression event are as objective as possible, and that the definitions be clearly and prospectively defined in the protocol."

Data analysis

Some questions for data analysis are also posed by the fact that complete follow-up of all randomized subjects for study outcomes will rarely be available in trials of anticancer products, the guidance recognizes. There may be deviations, and withdrawal from the protocol treatment occurring after randomization may have an impact on the trial data and, consequently, on the trial conclusions, particularly if they are related to treatment assignment. The guidance attempts to deal with the numerous possibilities for such potential distortions. Subjects may receive the wrong study medication, or none at all. They may withdraw from treatment prior to scheduled completion, or may be lost to follow-up. They may change treatment before evidence of progression, or may present with missing evaluations followed by evidence of progression.

Events of withdrawal from study therapy prior to independently adjudicated progression cannot automatically be regarded as noninformative, and the adequacy of censoring these events in the statistical analysis may therefore be questioned, the guidance acknowledges. "There is no way to handle this problem that is optimal for all anticancer studies, but the principles of intention-to-treat provide a reasonable approach and should be followed as far as possible when defining the analysis set for the primary analysis of PFS/DFS," it advises.

Overall, there is a need to predefine and justify methods for handling missing data, including rules of censoring. The methods chosen should minimize bias and loss of information, while being adequate for the aim of the trial. Potential biases should always be addressed and sensitivity analyses should be undertaken using different approaches. The range of sensitivity analyses will depend on the clinical situation and nature of the trial data observed, but should in any case be sufficient to demonstrate that the trial results are robust. "Any differences in conclusions from the range of analyses presented will need to be explained," the guidance stipulates. And for open-label trials, "utmost diligence is required when writing the study protocol and statistical analysis plan, as amendments to important aspects of the analysis made in the knowledge of accruing data would give rise to concern."

Frequency of assessment

The optimal frequency for assessing progression needs to be determined on a trial by trial basis, taking into account the aims of the trial, the treatment schedules, and the patient setting. A balance needs to be found, the guidance stresses, between the need to assess progression as precisely as possible and the need to minimize exposure of patients to invasive and resource-intensive diagnostic procedures.

Nonetheless, evaluation of PFS must start with assessment at baseline of all sites of possible disease specific to the particular tumor type, and subsequently involved sites must be systematically assessed during follow-up, together with other sites, as clinically and radiologically indicated, ideally using the same methods. The frequency of assessment has to be adequate to detect the expected treatment effect. In a randomized trial, the methods and frequency of tumor assessment should be symmetric across study arms, even when treatment cycles are of different lengths. Clinical trials that are not adequately blinded are particularly at risk of ascertainment bias when a change in the clinical status of the patient prompts an unscheduled assessment of disease status. Adherence to protocol defined schedules is essential and deviations should be reported.

Peter O'Donnell is a freelance journalist who specializes in European health affairs and is based in Brussels, Belgium.