Skip to Menu Skip to Search Contact Us Skip to Content

This paper describes several case studies in which new clinical trial methods were employed during the early phase clinical development of therapeutic compounds.

The topics covered include the implementation of first-in-human multi-center, combined protocols with patient arms for early proof of concept. Further, the usefulness of biomarkers in these early proof of concept trials will be discussed, followed by the use of effective modeling- and simulation tools. The paper will end with some lessons learned based on a case study.

New Early Phase Clinical Trial Drivers

The first question is: what drives these new trends in Early Phase? The primary reason for the trends is the low productivity of the pharmaceutical industry’s pipeline, combined with the greater regulatory and clinical requirements to increase safety. Another important driver is the accelerating patent expirations that make push pharmaceutical companies to increase the speed of drug development in order to maximize the compounds patented time on the market as an approved product. There is also an increasing role for payers, with a heightened political focus and more attention for patient involvement. 

All of these factors combined result in a traditional R&D business model that is no longer sustainable and must evolve in order to increase productivity. Moreover, if a drug is to make it to the market, the developing company must show true value for the patient. The time of me-too products has gone. Therefore, there is an absolute need to partner with different stakeholders for innovation and success with new development strategies.  

Combined Protocol

One of the first new methods that was implemented during this new evolution, was the first-in-human, multi-center combined protocols (1). These study designs helped to introduce the concept of performing one single project in early development that could lead to a faster go-no-go decision.

In this combined protocol design, it is possible to merge the classic single ascending dose in healthy volunteers, with food interaction, multiple ascending dose, and lastly early proof of concept in patients. This approach can be further optimized by including clinical trial sites across multiple geographies at different pivotal points in the trial. Western European sites can typically start up without delay and recruit the healthy volunteers quickly for the first part of the combined protocol. However, in our Western European units, we traditionally see a slower recruitment of patients, 2 to 4 patients per month, at which point it becomes advantageous to enrol centres in Central and Eastern Europe. These additional centers have relatively close proximity to the start up center, high quality work,  appropriate clinical experience and most importantly high recruitment rates of 10 to 20 patients per month.

The early combined start-up in Western Europe makes it possible to integrate new pharmacodynamic models and biomarkers in the first arms of the study. This type of study allows the pharmaceutical development company to make a go-no-go decision on the drug candidate in three to twelve months, faster than by the traditional route. 

Advantages & Disadvantages

The main advantages and disadvantages of using this type of designs are summarized in the table below.

With one single project included in one or more protocols, the investigator team is able to gain more experience in a short time with this particular therapy and with new PK/PD models. PK/PD models change regularly with new molecules but within one project the same method can be used more often. By having more reproducible data, it will become easier to do the dose selection for Phase II trials. 

Second, one can improve the safety risks by exposing fewer subjects to the new drug, depending on the desired endpoints in this phase. Moreover, all safety data will be collected in one database, thus preserving a robust clinical safety assessment. 

Third, by combining different studies in one protocol, it is easier to adjust many parameters during these adaptive studies. Flexible dosing is the most well-known and most frequently used adjustment, but one can also adapt treatment settings, eligibility criteria, sample size or even the design and randomisation during the execution of the trial. It is even possible to add other dose groups and different treatment regimens by relatively easy adjustment of the protocol. As a worst case scenario, it may be necessary to submit an amendment for which has an approval timeframe of  14 days after submission.

Fourth, the scientific input from the CRO will increases and becomes more valuable, because there is only one investigator team they build up experience with the molecule much faster, bringing new ideas exploring efficacy and eventually concerns about safety. 

By doing all this, the single largest advantage is reducing time to market, with easily 1/3 of a year more sales. Costs will also decrease through optimisation of the logistics processes such as reducing the number of submissions to authorities to one and training only one group of investigators.  

Of course there are also disadvantages. Preparing such protocols is a much more complicated task than simple studies. As monitoring as complex in this early phase, is a very intensive and continuous process, this is best done by the same monitor. It can become complicated to organise the logistics of this even with the relatively short distance between Western and Central & Eastern Europe. Another disadvantage is that some part of the staff and certainly the volunteers use different local languages. Even more important are the cultural differences that play a role in harmonising all types of procedures to implemented be used in a short time.

Supplies can cause a problem, not so much for the timely supply of the investigative drug but more for the core medication and laboratory material, so it is necessary to install a fast distribution system to quickly solve the issues.

Another disadvantage is the variability between local safety labs due to different methods used and also to the differences in reporting the results. And lastly, even if there is a relative proximity, sample transport can still be a problem when you have an adaptive design and want to use online PK information to make a fast decision on how to proceed.

  

Biomarkers in Proof of Concept

The second trend is the implementation of biomarkers in early proof of concept studies. Two examples will help to clarify the practical challenges.

The first case is a study in neuropathic pain where SGS used the Quantitative Sensory Testing method (QST) with thermodes that stimulate the skin from low to high temperatures to define the Heath Pain Perception Threshold, HPPT. The drug was a TRPV-1 receptor antagonist.  The results show a clear and statistically significant increase in the post-dose pain threshold of 3.9°C for the active drug, versus 0.6°C for the placebo compared to the pre-dose status. These results allow one to feel quite comfortable that this new drug can be effective. However, the absolute temperature reached for the new drug was 50.5°C which is the maximum allowed threshold to avoid burn lesions. This means that the effect is probably underestimated. Moreover the QST method tests mostly for small fibre neuropathies (A delta, and C-fibers, not inflammatory component). However, in neuropathic pain conditions there are different mechanisms involved where the actual neuropathic changes are just a part of the pain mechanism. So by doing only one test the real disease model is not fully reflected. 

The second example is from a drug for Alzheimer’s disease. In this study, continuous cerebrospinal fluid (CSF) sampling was performed over 36 hours in 36 subjects to test a number of possibly relevant biomarkers (2). There is a dose dependent and sustained reduction of the biomarker by the experimental drug under study, providing clear indication for a possible therapeutic effect. However, there are some initial increases in biomarkers relative to baseline and a very large upward drift induced by placebo (fig 1). It has been shown that CSF sampling frequency and/or volume contributes to the variability in the biomarker levels under study. So it is very important to have a clear insight in the baseline values and in the placebo effects in healthy volunteers before actually studying the effect of a drug in patients as was shown in this experiment (5).

 

There are a number of conditions where a biomarker can be useful in the early phase of development. First and foremost, the marker must be relevant for the disease model. As shown in the first example, it is not often certain that one is studying a clinically relevant effect nor that the best biomarker is chosen at this early stage of development. 

The use of a biomarker at this stage must also be feasible for staff and for patients. For example, in the second case presented, if the proper trained anaesthesiologist or neurosurgeon is not accessible for putting in the catheter and collecting the CSF, it will not be acceptable for the patient and a low patient burden is an absolute must. Staff experience is also key in this regard, but it must be taken into account that the selected biomarkers will change from study to study, so the necessary experience must be built up quickly before the start.

So the question is how does one make biomarkers useful at this early stage of clinical development? First of all, it is best to use a set of biomarkers from which one or more can be selected later in development. Thorough training is required for the staff for each biomarker study, because even manipulating the same biomarker can evolve over time. Therefore, there is a need to run pilot studies before the actual drug study in order to obtain preliminary information on the biomarkers used. Fluctuations in baseline values and placebo effects, like that of putting a CSF catheter or the sampling frequency, must be studied and understood. These preliminary studies are also part of the training. The bottom line is: start these preparations early in the study setup, even before finalising and submitting the protocol.

Modeling and Simulation

The next tool to add to these new development techniques in early phase, to accelerate the time to decision, is effective modeling and simulation (M&S). This method is intended to predict the future, extrapolating information from existing data.

Often the interest does not lie in the type of patients or the settings under study, but to use the data to look into the future and into different settings. The tool helps to clarify complex systems such as non-linear effects, time dependency, multiple sources and visits, basic noise. It makes it possible to integrate information across time, dose-levels, and studies and even across drugs. By applying these techniques optimisation of the design of further studies is possible. All this information can also be used as a knowledge repository to gain more insight in the disease, fill in the gaps in data and quantify uncertainty.

The case of a monoclonal antibody that binds to interleukin 6 (Il-6) and is used for rheumatoid arthritis, illustrates the added value of M&S (3).The first model that was developed was a Target Mediated Drug Disposition Model. (TMDD) predicting the target-time profile as a function of dose and route of administration for the effect on IL-6 and CRP. So you learn what dose to use, what timing and route of administering of the study drug and when sampling for these biomarkers has to be done. A second model was a pharmacotherapeutic model, describing the casual link between these biomarkers and a clinical endpoint, Disease Activity Scale 28 (DAS28), helping to optimize therapy.

But what is needed to guarantee a successful use of M&S?  First, starting preparation of the model on time before collecting any data and not when you have the data. Second is focusing on the next phase of development in order to support the decisions to take regarding POC, Go-no-go, dose selection. Therefore it is necessary to correctly understand all the issues that have to be addressed by M&S before being really involved. Also the buy-in from all team members is needed including the senior management that in the end makes the final decisions. Of course the quality of the data must be very high, which is most often the case for in-house data, but sometimes it’s necessary to include public data as well, for which this guarantee cannot be given. Also the biomarkers used to develop the model have to be identified and validated; otherwise one cannot obtain relevant outcomes. Therefore it is necessary to use pharmacometricians with high levels of experience in technical aspects of M&S, as well as in overall drug development and excellent communication skills. These key personnel need uninterrupted time blocks and strong collaboration with an inter-disciplinary team of clinicians, pharmacologists, statisticians, pharmacokineticists and others.

Lessons Learned

The lessons learned from all of these examples are based on the experience and feedback of the personnel actually involved in the execution of combined protocols. Part of this experience comes from the following case (4). The study is a combined first administration to human study with single and multiple-ascending dose arms and a POC arm for a hepatitis C drug (HCV). The single-ascending dose part was done in 18 healthy volunteers, the multiple ascending part in another 18 healthy volunteers, with twice a day administration for 5 days. For the POC arm 18 HCV patients were randomized to 2 doses in a 2:1 ratio with placebo. All patients were followed for 12 days and apart from clinical safety evaluations, plasma samples were collected for pharmacokinetic data and in the POC arm viral load was measured in the HCV patients as a biomarker. The results of the patient part demonstrate a fast decreasing viral load after 2 days with 2 different doses compared to placebo. It is important to stress that there was only one approval cycle for health authorities and medical ethical committees of both centers with a submission to approval delay of 16 days. Submission to first patient enrolled (FPI) was 18 days, submission to first dosing 35 days and first dosing to last patient out L(LPO) 6 months. The study was done in the SGS unit in Belgium and in the unit of a partner in Moldova.

Other experiences have been added to this one, leading to a number of lessons. The first lesson learned is that in- and exclusion criteria must be discussed in much more detail than typically done for a single center study and discussed with each center involved, before finalizing the protocol. The standard of care for patients must be looked after in detail as the access to this care and local culture cause important differences even if all centres follow the same guidelines. This can be illustrated with a case where inhaled corticosteroids were used in asthma patients: same guidelines were in place, but were interpreted differently due to different use in different countries. It can make a huge difference in recruitment and in the interpretation of some tests.

Further, shifts in starting timelines happen often in early phase for many reasons such as toxicological data that are not ready available on time, and so on. This has a negative effect on recruitment but the effect can be completely different between two centers, due to cultural reasons, local habits, and mode of transportation certainly in countries with a central referral system. All this must be taken into account in the set-up of the trial.

Another lesson is that, when using local safety labs, what is often done for practical reasons and speed, transfer of safety laboratory data are still best done in the traditional way, by paper transfer, as most labs do not have the right software tools. Use of local labs for screening of biomarkers for eligibility is logical, because it increases speed of enrolment. Quality can be checked by auditing, but one of the issues, that was encountered, was the lack of continuity in lab supplies for the additional number of tests for study candidates, so this must also be checked.
 
In all preparations it must never be forgotten that even in experienced centers, that what is obvious for one center is absolutely not obvious for the other. In general communication is key in preparation, and never take “we have no further questions”, as a guarantee that everything is clear. Starting up a new study means that a new site qualification must be done each time since things will change over time at every clinical trial center. Details are needed about recent staff turnover as these can influence the fast recruitment. Also financial health can temporarily change causing problems with suppliers.

Every time study related processes, the use of source documents and pharmacy processes have to be checked thoroughly. Centralization of clinical supplies and labs for outcome results is of course the preferred way of working, but it is very important to use providers that are familiar with your procedures so that they can be proactive in their contacts with the centres. All investigators in experienced phase I units, had extensive training in GCP but nevertheless there can be important differences in knowledge, so additional training is really needed. The most important lesson from all this, is that you need additional and sufficient time before you start these combined protocols for evaluation, organization, training and for a very clear communication plan.

Conclusion

There is an urgent need to use innovative techniques in early development. Multi-center combined protocols including early involvement of patients and implementation of biomarkers can increase productivity in early drug development and adding modeling and simulation to these tools can further stimulate efficiency. However if we want to do all this properly, we must absolutely shift from a “rush-to-do” mentality to “reflection and thought”.  So part of the reduction in time we can gain with these techniques, must be spent on more thorough upfront preparation.

Author

Robert Lins, MD, PhD
SGS Life Science Services - Clinical Research
Senior Clinical Advisor

References

  1. Underwood D.  Flexible Phase I Study Designs : Expediting Early Clinical Drug Development. Drug Information Journal 2012; 46:13-18;

  2. Forman MS, Tseng J, Palcza J, et al. The Novel BACE Inhibitor MK-8931
    Dramatically Lowers CSF Ab Peptides in Healthy Subjects:  Results from a Rising Single Dose Study. American Academy of Neurology, New Orleans, LA, USA, 2012.

  3. Levi M, Charoin JE, Frey N, Delor I, Jacqmin P.  A mechanistic target mediated drug disposition (TMDD) model is required to correctly estimate the bioavailability of a subcutaneous formulation of Tocilizumab (TCZ), a monoclonal antibody with non-linear kinetics. American Conference on Pharmacometrics, American Conference on Pharmacometrics, Mashantucket, CT, USA, 2009, 

  4. Detishin V. Haazen W2, Robison H3, et al. Virological Response, Safety, and pharmacokinetic profile following Single- and multiple-dose administration of ACH-1625 Protease Inhibitor to Healthy Volunteers and HCV Genotype-1 Patients. European Association for the Study of the Liver, EASL, Vienna, Austria, 2010

  5. Li J, Llano D, Ellis T, et al. Effect of human cerebrospinal fluid sampling frequency on amyloid-ß levels. Alzheimer’s & Dementia 2011:1-9