By Quanticate Statistical Consultancy Team
Phase I marks a significant milestone in the development of any new medicinal product. A target has been identified, a compound has been discovered that hits the target, and has been refined to ensure it has good properties for development. In-vivo safety pharmacology and toxicology studies have been conducted and the compound has been tested in pre-clinical models of the disease. This process will have taken many years. A multi-disciplinary team has evaluated all the data, and decided to take the plunge and invest in clinical development. And now, the drug is ready to be tested in humans.
The objectives of the Phase I development plan are many. Alongside establishing the safety, tolerability and pharmacokinetics (PK) of single and multiple doses in humans and ascertaining the maximum tolerated dose, it is often important to explore at least some of the following: effect of food on exposure, potential drug-drug interactions, gender and/or age effects, alternative formulations and pharmacodynamic (PD) effects. Traditionally, this has led to sizeable Phase I packages with multiple studies which take many months to complete. However, with careful planning and protocols that are written flexibly, many objectives can be addressed under one protocol (often with many parts) which can improve efficiencies in the drug development process and speed up the time to becoming Phase II ready. This article will illustrate some of the ways in which this can be achieved.
Phase 1 Clinical Development Planning
To maximize efficiency in Phase I development, the clinical development plan must first be considered as a whole, before considering the details of individual studies. What are the objectives of Phase I for the compound in question? What data do the team require before they can proceed to efficacy studies in patients? Are there any potential issues from pre-clinical development that need to be addressed (e.g. QT signal, potential for drug-drug interactions)? Once all the questions are identified, the clinical team can plan which studies need to be conducted in order to answer them, look for specific ways to minimise the time taken and maximise the value of each study.
It is becoming increasingly common to combine single dose ascending (i.e. First-in Human) and repeat dose studies under one protocol1. With a single protocol written in a flexible way, this can save considerable time by cutting out the period between the end of the single dose study and the writing of the repeat dose protocol, and the time for regulatory review of the second protocol. In these combined designs, it is typical for the clinical development team to be unblinded, while the subjects, investigator, all staff at the study site (other than the pharmacist) and study monitors remain blinded to treatment allocation. The clinical development team will carefully review any issues of poor tolerability or adverse events at the individual level, and thus need to be fully unblind in order to make an informed choice of doses for the repeat dose phase whilst the study is ongoing. Provided an efficient process is in place for transfer of samples between the study site and the bioanalysis laboratory, and of electronic data between the study site and the study statistician and pharmacokineticist, statistical summaries of pharmacokinetic and safety data can be provided quickly throughout the course of the trial, to assist the clinical development team at each dose escalation or key decision stage.
Additional objectives can often be easily incorporated into a combined single and repeat dose study. The possibilities here are almost limitless, but a few common examples for an orally administered drug primarily aimed at a broad cross-section of the population might be as follows:
i) Assess the food effect on the PK of a selected dose;
ii) Assess relationship between PK and PD effects (e.g. on QT interval);
iii) Assessment of a potential drug-drug interaction (e.g. effect of drug A on drug B, effect of drug B on drug A, or both);
iv) Assessment of the effect of gender and/or age on PK (if applicable to the target population);
v) Assessment of therapeutic effect, if an appropriate PD measure is available (e.g. fMRI, EEG, blood biomarker, relevant rating scales)
If the protocol is written in a flexible way, some of these may be incorporated in an opportunistic manner. For example, if the dose escalations stop earlier than planned (due to reaching the maximum tolerated dose, or achievement of higher PK exposure than anticipated), one or more of the remaining periods can be utilized to assess, for example, the food effect on a dose that has been established as being well tolerated earlier in that cohort. Optional cohorts may also be included to allow exploration of higher doses where appropriate (e.g. if the planned doses are well tolerated, and there are appropriate safety margins, for example the PK exposures are lower than anticipated). Assessments of PD effects, or drug-drug interactions, can often be built into the design of the planned cohorts, obtaining more information using the same set of subjects.
When planning a combination study with many objectives such as those illustrated above, care must be taken to ensure that all optional aspects are clearly explained. Stopping rules at the individual, cohort and study level and the decision rules for selection of doses and for proceeding (or not) with the optional parts of the study should be outlined. The maximum number of subjects to be included should be stated, and there must be no compromise to subject safety through the flexible nature of the study design. Ethics committees and regulators will be particularly concerned that there is an appropriate data review process in place for each key decision stage.
There may be a greater lead time required between the initiation of protocol development and the first subject first visit (due to study complexity), but provided this time is built in to the development plan it is easily accommodated. In fact, overall there will be a reduction in the development time of the drug (from efficiencies by having only one protocol to review, time savings between cohorts, and in the data management, statistical and reporting processes). Other data analysis activities (e.g. population PK) will benefit too from having data consolidated on to one database.
It is important that the data required for in-stream (i.e. while the study is ongoing) decision-making can be available for review and analysis very quickly, perhaps even within 24 hours of the assessments being performed. Therefore some means of electronic data capture is necessary. With a multi-part study, which may include a mixture of cross-over and parallel group cohorts, the structure of the database must be carefully considered to ensure that the data from any part of the study can be extracted separately as required for reporting. Where possible, data checks should be built in at data-entry, to ensure high quality data are available for in-stream decision making, which will occur prior to database lock.
During the course of the study, there are likely to be several data review meetings to determine the dose level or other design aspect, for the next cohort(s) of subjects. These meetings need to be planned from the outset and incorporated into the study schedule. Key points to identify upfront are:
i) which data are required for decision making (PK, adverse events, lab results, etc)?
ii) how much time needs to be built in for sample analysis and data transfer?
iii) how much time is required by the pharmacokineticist, statistician and other study team members to analyse and report the data to the team?
If appropriate measures are in place, it can in most cases be possible to complete ii) and iii) within a week of the last data point being recorded. Dose escalation decisions are usually made jointly between the clinical development team and the investigator, therefore during review meetings care must be taken so that any information that might reveal the treatment blind is not shared with the investigator or other study site personnel.
From all the above considerations, it can be seen that planning is key. Once everything is in place, the study can proceed efficiently through the various cohorts to completion. The resulting time saving compared to doing several separate protocols can have high impact in terms of reducing cost and overall development timelines.
This paper has provided an overview of some of the ways in which Phase I packages can be tailored to the needs of the specific compound and conducted more efficiently using combination, flexible protocols. Considerations that need to be made at the planning stage have been outlined and the practicalities discussed so that best use can be made of emerging PK and safety data to enable informed decisions regarding study progression.
Apart from combination studies, use of novel statistical designs can be used to improve Phase I or other clinical pharmacology studies in other ways. For example, the use of incomplete block designs (where each subject receives a subset of the treatments under study) can enable more treatments to be tested within a cross-over study than might otherwise be thought practical. This may arise in situations where different formulations of the same compound are to be evaluated. Factorial designs can also be used to help explore several factors simultaneously in one study (e.g. for ascertaining the optimal dose levels for a combination product). For compounds where the expected variability of the pharmacokinetic parameters is high, resulting in a large sample size, the use of interim analysis or sequential designs can offer the opportunity to complete the study more quickly with fewer subjects, if the primary objective has been reached or the study looks as if it will fail to achieve it.
In conclusion, the use of combination studies, flexible protocols and novel statistical designs can significantly improve Phase I and clinical pharmacology packages, with benefits both in terms of timelines and costs.
1. ‘Guidelines for Phase I Clinical Trials’ (2007) Association of the British Pharmaceutical Industry at www.abpi.org.uk/publications