Commentary on Improving the Efficiency of Clinical Pharmacology Studies.

I read the commentary by Polasek and Schuck1 on improving the efficiency of clinical pharmacology studies with a great deal of interest. The authors correctly point out that drug development is expensive and becoming more so with every passing day. The authors also point out that, far from being a necessary evil on the way to Phase 3, Phase 1 studies provide critical information that is needed for future development. These studies should be performed in a thoughtful manner and are every bit as important as Phase 2 and 3 studies. Although it is desirable to be efficient in Phase 1, any attempt at “short-cutting” Phase 1 is likely to result in problems down the line. In that sense, although Phase 1 represents a small portion of the overall cost of new drug development, performing the right studies at the right time in Phase 1 is likely to result in reduced costs to the sponsor, and should lead to more favorable development timelines. One topic that the authors should have mentioned in their introduction is that while Phase 1 studies have increased in cost over the years, the actual number of studies needed in Phase 1 have decreased dramatically. For example, because of advancements in our ability to predict drug interactions (through the use of human liver microsomes, recombinant cytochrome P450 enzymes, and human hepatocytes), many fewer clinical drug interaction studies are performed today, as compared to even 20 years ago. Recent efficiencies in regulatory requirements related to the thorough QT study requirement will decrease the need for a dedicated QT study for many programs, since the drug's effect on the QT interval in many cases may be determined during the first-in-human study.2 These are very important advancements and have done much to streamline clinical development during Phase 1 and beyond. It is always important to look at what we in the pharmaceutical industry are doing in whatever phase of development we are in and look for more efficient development pathways. The authors, who are consultants to the pharmaceutical industry, have many years of experience designing and analyzing Phase 1 trials, as well as advising clients on what specific studies are needed. Based on their experience, they have listed a number of steps that they believe the industry should take to improve the efficiency of Phase 1 (and perhaps decrease costs). I will take each of these recommendations in turn and provide comments on each one. I agree with the authors when they state that many Phase 1 protocols are long and complicated, but, in contrast to the authors, I think that much of this complexity is needed. Protocol complexity is driven by a number of factors, including regulatory requirements, the ever-increasing emphasis on acute safety and tolerability, and the desire to squeeze as much value as possible from each study (by including several parts to each protocol). The authors have 2 specific recommendations. The first is to reduce the number of exclusion/inclusion criteria, on the grounds that, by definition, healthy volunteers have no significant medical history. The problem with this line of thinking is that, while a given volunteer might not have a significant medical history, they may indeed have undiagnosed medical problems, which only become apparent at the screening stage. For example, first-degree heart block is asymptomatic, usually requires no treatment, and is often detected during Phase 1 screening of prospective volunteers. Is such a subject a “healthy” volunteer? I would argue no, because whereas first-degree heart block is usually a benign condition, participation in a clinical trial for a new drug for which the safety profile is completely unknown, and the individual will not benefit from the treatment represents an unfavorable risk–benefit to the subject. Therefore, such subjects must be screened out, and this, in turn, requires that such dysrhythmias be specifically listed in the protocol. The second recommendation that the authors give is to include the list of study assessments in tabular form only. This is likely to be met with considerable pushback from clinical contract research organizations (CROs), as well as regulators, neither of whom want to have to interpret how a given study procedure is to be performed. I know from experience that even when a given procedure is written down explicitly, it can be interpreted in various ways; including study procedures only in tabular form would exacerbate the problem. Removing placebo subjects from first-in-human studies would serve neither the volunteers nor the pharmaceutical industry well. Although the current paradigm of 6 active–2 placebo is not perfect, it gives the study team a reasonable estimate of the acute tolerability of the new drug, which is about all that can be expected from a first-in-human study enrolling 36–60 subjects. Inclusion of placebo subjects allows at least an initial assessment of safety; if 3 of 6 actives have nausea and vomiting and none of the placebos do, is this not useful information? I would argue that it is. I have been in many situations where the inclusion of placebo subjects allowed the correct attribution of adverse events during study conduct. Safety Review Committees do indeed review the adverse event profiles for both active and placebo subjects when determining the subsequent dose, so a lack of these data is likely to be met with considerable pushback. I suspect that regulatory authorities, as well as Institutional Review Boards (IRBs), would not look favorably on the lack of placebo subjects in first-in-human trials. Blinding is another matter. Most of the first-in-human studies that I have been involved with have been single-blind; the investigator and subjects are blinded, but the sponsor is not. This seems to be a good compromise in that it allows the sponsor to react quickly to any perceived safety signal, while preserving the blind for the investigator, who can assess the adverse event without bias. I agree with the authors that double-blinding is seldom necessary. This is a sound recommendation and, based on my experience, is being done on a fairly regular basis now. Most sponsors include, in addition to the standard single-ascending-dose and multiple-dose assessments, an initial assessment of the effect of food for studies involving oral dosage forms, and many include an initial assessment of potential drug interactions, based on in vitro data. An assessment of the QT prolongation potential is also increasingly included in these studies. As the authors point out, inclusion of these additional objectives can streamline clinical development and allow go/no-go decisions more quickly than the standard sequential approach. This is a sound recommendation theoretically, but in practice it may be difficult. The reasoning behind the traditional recruitment of young healthy men in Phase 1 trials is that (1) they are not, in general, taking any chronic medications; and (2) they are not, for the most part, suffering from any chronic diseases. Recruiting older volunteers is more difficult than one might think, because this population is typically taking at least 1 chronic medication (eg, antihypertensives) and may have some other chronic diseases, which can complicate the safety analysis of Phase 1 trials. Because of this, the screening of older people for such trials is likely to result in a high screen failure rate, which is not very efficient. Generally, in these trials, women taking effective contraception are eligible to participate; however, in most cases, at this stage of development, reproductive toxicology studies have not been performed, so there is a regulatory limit on what can be done here. Increasing minority recruitment in clinical trials is a desirable but complex multifaceted issue that deserves considerable study in its own right. I do agree with the authors that the weight inclusion criteria could be relaxed in most Phase 1 trials. This would speed recruitment and allow an initial assessment of the effect of weight on the pharmacokinetics of the study drug. I agree with the authors on all these recommendations. Many of these recommendations are in fact routinely performed presently by most pharmaceutical companies. It should be noted that such approaches (eg, physiologically based pharmacokinetic models for the predictions of drug interactions) have varying rates of acceptance within the regulatory agencies. In the case of the US Food and Drug Administration, there can be disagreement between the Office of Clinical Pharmacology and the individual review division on what is acceptable, as the review divisions may not have the expertise or experience with these newer approaches. This can be specific to the therapeutic area and is probably a major reason why there is not total adoption of these methods by sponsors. This recommendation is certainly desirable from the standpoint of the pharmaceutical industry but may not be perceived that way by other stakeholders. If there are too many former industry scientists and physicians on IRBs, many may question whether the reviews that these IRBs produce can be objective and unbiased. IRBs serve as an essential protection for patients and volunteers in clinical research; this role would be diminished if the perception were that the pharmaceutical industry is “packing” them with sympathetic members. In my own experience, there are many good IRBs out there, and they produce thoughtful and insightful reviews of protocols. It is also important to remember that, whereas some individual IRB members may not have direct experience in the pharmaceutical industry, these scientists see a lot of study protocols during the course of their work and thus have a reasonable view of contemporary drug development. What the authors call “micromanaging” seems to me to be good study oversight, which is required by all regulatory authorities and represents responsible stewardship of the well-being of patients and volunteers. The fact is that not all adverse events take time to evolve; some happen quickly. It is essential for sponsors to show that they are monitoring their studies and reviewing adverse events as they happen so that signals potentially pointing to more serious events can be evaluated in a timely manner. Using modern online tools to monitor safety data frequently is demonstrating appropriate oversight by sponsors, and such tools pay for themselves when serious events do happen. I am not quite sure what to make of this recommendation. For example, I am not sure why “Obsessively curating data judged as not clinically significant (NCS) by investigators, such as parameters slightly outside reference ranges…” is judged to be undesirable. Sponsors are required to report all adverse events and laboratory data, whether they are judged clinically significant or not. Getting this right is an exacting, tedious process that requires many people, but I am not sure that I would characterize the work as “menial,” as the authors imply. Rather, it is essential in order that CROs maintain compliance with the regulations. Considering that receiving 1 Form 483 might spell the end of a CRO's business, it is not surprising that a lot of time and effort is put into quality assurance processes by them. As scientists, we often do not have a clear picture of what it takes to maintain regulatory compliance. I suggest that we leave that to the regulatory experts in CROs who understand those regulations much better than we do and who, I am sure, are implementing the regulations in the most efficient way possible. Here, the authors are recommending that Phase 1 studies not include small cohorts of patients with the disease in order to get an initial impression of efficacy, on the grounds that the study will be too underpowered to be interpretable. As the authors point out, it can be very informative if a good biomarker exists, and perhaps less so if such a biomarker is not available. I do not think blanket recommendations can be made here. If one has an informative biomarker or some way to assess efficacy and/or target engagement of the compound in a meaningful way in patients, then it makes sense to do that as early as possible. If that situation does not exist, then I agree that a patient cohort is not ideal. In my time in both large and small pharma, this was never a routine decision; considerable thought is put into the question of when to dose patients, so in essence, sponsors are already doing this. I agree with the authors that this is needed; however, it is easier said than done. The lack of clinical pharmacology training in medical and nursing schools has been documented elsewhere.3 Like most things in drug development, this is a complex, multifaceted problem that defies a quick fix. The authors have put forth a number of recommendations for improving the drug development process, particularly in Phase 1. Although I do not agree with some of them and believe we are doing quite a few of them already, it is a useful exercise to examine the process as a whole from time to time, to look for areas where we can improve. In that respect, the article by Polasek and Schuck should be read and discussed by the clinical pharmacology community. Michael Fossler is an employee of Cytel and provides consultation services to the pharmaceutical industry.