There is significant room for improvement in the way first-in-man studies of new medicines are designed and described, a working party of the UK’s Royal Statistical Society (RSS) has concluded. It also calls on regulatory authorities to tighten up their mechanisms for collecting, analysing, sharing and communicating data on risks from first-in-man trials.
The Working Party on Statistical Issues in First-in-Man Studies was convened in response to the disastrous volunteer study at London’s Northwick Park Hospital last year. It found that the trial with TeGenero’s monoclonal antibody TGN1412 was “not particularly well-suited to investigating either safety or tolerability, nor even for estimating pharmacokinetic parameters, such as how long TGN1412 stays in the body, its maximum concentration or the time after dosing that this is achieved”.
The group’s remit was to look in detail at any statistical issues that might arise from first-in-man studies, since “a cornerstone of the regulatory framework is the statistical theory and methods that underpin clinical trials”. The RSS report is meant to complement the final recommendations delivered last December by a government-commissioned Expert Scientific Group (ESG) on Phase I trials.
21 recommendations to raise standards
The 21 recommendations made by the Society’s working party are geared particularly to raising standards in risk assessment and communication, preparatory preclinical work and the experimental design of first-in-man studies.
On the risk-assessment front, for example, the report says taking a new compound into humans should be contingent on a quantitative justification of the starting dose and a priori assessment of the risk level for the recommended study dose(s). The sponsors – and MHRA assessors independently – should classify the proposed doses by the degree of assessed risk, in accordance with their confidence in the preclinical data. High a priori risk would rule out the participation of healthy volunteers while demanding a precautionary approach to study design, the working party suggests.
The assessment process would be facilitated by regulatory authorities providing a mechanism for the pharmaceutical industry to collect and share data on serious adverse reactions in first-in man studies, it adds (information-sharing was also a keynote of the ESG report). In the UK specifically, the Medicines and Healthcare products Regulatory Agency (MHRA) should report annually on the designs of, and any serious adverse events occurring in, first-in man studies involving healthy volunteers, both for biological/biotechnology and chemical compounds.
Statistical reporting of preclinical studies needs to be in line with the International Conference on Harmonisation’s requirements for reporting of clinical trials, the report recommends. The objective would be a document that provided “a specific risk assessment for the medicine being trialled, together with a statement of uncertainty”, the RSS comments, adding that such a document should be available to regulators, ethics committees, trial subjects, participating clinicians and insurers.
The report also suggests “crude inter-species scaling” may be an inadequate means of determining the right starting dose for a biological in humans. “In vitro studies using human cells will be necessary to establish the avidity of the ligands for their targets and assist in dose calculation,” it states.
Among its recommendations on the design of first-in-man studies, the working group says study protocols should include justifications of dosing intervals (a much criticised facet of the TeGenero trial) and dose-escalation steps, as well as prior estimates of the number or rate of adverse reactions by dose. If the trial is investigating safety and tolerability, then the sponsor must provide an operational definition of safety.
First-in-man studies should have “the same clarity, care and detail regarding purpose, design and analysis as are currently required for studies supporting drug licensing”, the RSS believes. This means assessments of the trial’s suitability, study management and the intended analyses “will need to be provided in much greater detail than is currently the case”, it warns.