| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Sign In to gain access to subscriptions and/or personal tools.
|
|
|
|||||||
Sign In to gain access to subscriptions and/or personal tools. |
|||||||||
LETTER TO THE EDITOR |
Centre for Applied Pharmacokinetic Research, University of Manchester, Manchester, UK
Address for correspondence: Malcolm Rowland, Centre for Applied Pharmacokinetic Research, University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, UK; e-mail: mrow190539{at}aol.com.
Bertino et al1 recently published an interesting position paper on behalf of the American College of Clinical Pharmacology (ACCP) on the place of microdosing as a tool to accelerate drug candidate selection and development. One study discussed in considerable depth was the results of the CREAM (Consortium for Resourcing and Evaluating AMS [Accelerator Mass Spectrometry] Microdosing) trial,2 whose primary objective was to evaluate the ability of microdosing to predict the pharmacokinetics at therapeutic doses of 5 drugs: warfarin, ZK253, midazolam, diazepam, and erythromycin. The reader is encouraged to read the associated publication2 explaining why each drug was chosen. Bertino et al1 challenge several conclusions made in this widely discussed publication. First, that diazepam exhibited linear kinetics, because the mean half-life was 25% higher and the volume of distribution 30% lower for the microdose than the therapeutic dose. Second, that assessment of the oral bioavailability of ZK253 could be made when measurement of the unchanged drug after oral microdosing was below the detection level of the AMS assay. As the chair of the scientific committee that oversaw this trial, I feel that I must respond to these challenges, as well as comment more broadly on other issues raised.
The central issue in the potential use of microdosing is not one of robustly demonstrating strict pharmacokinetic linearity but rather that prediction at therapeutic doses is sufficiently close to that subsequently observed to permit the correct decision to be made whether to take a compound forward, at least with respect to its pharmacokinetics. Here, the commonly held view, widely adopted when using allometric scaling of animal to human pharmacokinetics, is that prediction within 2-fold of the mean observed in a phase I study is acceptable, which was certainly so for the diazepam microdose data. Bertino et al1 argue that one may wish to apply narrower limits when predicting phase I doses for narrow therapeutic index compounds. Although laudable, it is perhaps unrealistic given the uncertainty in the human safety profile, activity of metabolites, and other factors.
The challenge about the oral bioavailability assessment of ZK253 is misdirected. The issue with this experimental drug was its poorly understood, very high interspecies variability in absorption among test animals, leading to a total lack of confidence in what to expect in humans. In fact, it was subsequently found to be <1%, too low to warrant further development. The important lesson here is the benefit of administering the microdose intravenously, as well as orally, to define systemic exposure associated with a known body burden. So, although Bertino et al1 are correct in stating that, in the absence of measurable values after the oral microdose, an accurate estimate of oral bioavailability cannot be made, a maximal value could be assigned by reference to the intravenous data, assuming that the actual values after oral dosing were just below the assay limit. Indeed, the answer (0.16%) would have allowed the company to stop the development program much earlier than otherwise.
I will now move on to some more general issues raised in the position paper. One is sampling times. Bertino et al1 state that the CREAM trial is somewhat unrealistic in that for all 5 drugs studied, their pharmacokinetics at therapeutic doses were known, allowing blood sampling to be close to optimal, whereas at the preclinical stage, such information is lacking. Although strictly true, as with any first-dose-in-human study, the investigator uses all prior animal and in vitro data, perhaps suitably incorporated into a physiologically based model,3 to predict the anticipated human pharmacokinetics and hence appropriate blood sampling times. Furthermore, additional samples are usually included extending over long times and spread logarithmically, recognizing the polyexponential decline profile of many drugs. Hence, events after a microdose should be reasonably well characterized.
Another concern raised is that the subject sample size in currently published (and unpublished) reports is too small to provide robust results. This raises 2 issues. First, the one of qualifying the microdose approach, which generally involves a crossover design of micro and therapeutic doses, is concerned primarily with intrasubject variability, often relatively small. Moreover, a substantial body of human therapeutic data usually already exists to serve as reference. For example, although not shown in the CREAM trial publication, the individual dose-normalized microdose profiles for both diazepam and midazolam lay within the 95% prediction interval of the concentration-time profiles obtained from the population analysis of a substantial body of therapeutic data. The second issue is the drug development situation, where in phase 0, there would only be microdose data and intersubject variability, which is the major issue. Although increasing statistical power by increasing the sample size is clearly desirable, it is doubtful that this request will be often heeded, owing to the wish to limit cost and resources at this very early stage. Nonetheless, any prior information, such as that the compound is primarily eliminated by an enzyme with a very high intersubject variability, such as CYP3A4, should be taken into consideration. That said, even with relatively small numbers of subjects per study, one could use methods such as fuzzy simulation4 to generate an expected envelope of the concentration-time profile in a larger population with a specified level of certainty.
Yet another expressed concern was whether microdosing can (accurately) predict the pharmacokinetics of drugs that phenotypically express polymorphism in enzymes or transporters. The reason for this concern is unclear. Apart from events such as induction or mechanism-based inhibition, pharmacokinetics reflects rather than perturbs the state of an individual, and as such there is no fundamental reason to believe that microdosing would be any better or worse than a therapeutic dose in reflecting polymorphisms.
The above comments aside, other important questions remain. Among these are the following: can microdosing reasonably predict pharmacokinetics at therapeutic doses for drugs for which reliable predictive in vitro methods are currently lacking or uncertain, including those drugs that are primarily eliminated by phase II conjugative mechanisms, by cytosolic oxidative enzymes, or are absorbed or eliminated predominantly by transporters? Will coupling microdose data with in vitro data that characterize likely saturable processes (eg, metabolism, transport) materially improve the prediction of pharmacokinetics at therapeutic doses? These are among the questions currently being addressed in a European Union supported project (www.eumapp.com). Ultimately, the aim is to know what characteristics of a compound or situations lend themselves to the successful application of microdosing and those that do not or are more problematic, so that we may channel our best efforts in the right direction in the use of microdosing. This will only come about by asking the right questions, designing the right validation studies, and appropriately analyzing the resultant data.
Finally, although there is certainly a place for microdosing in aiding drug development, it is but one approach in the toolbox to predict human pharmacokinetics. Moreover, in the world of business decision making, pharmacokinetic data will always be placed within the broader frame of a pharmacodynamic biomarker profile, as well as safety and commercial considerations.
 |
ACKNOWLEDGEMENTS |
|---|
 TOP ACKNOWLEDGEMENTS REFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPACKNOWLEDGEMENTSREFERENCES |
|---|
1. Bertino JS, Greenberg HE, Reed MD. American College of Clinical Pharmacology position statement on the use of microdosing in the drug development process. J Clin Pharmacol. 2007;47: 418-422.
2. Lappin G, Kuhnz H, Jochemsen R, et al. Use of microdosing to predict pharmacokinetics at the therapeutic dose: experience with five drugs. Clin Pharmacol Ther. 2006;80: 203-217.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
3. Jones HM, Parrott N, Jorga K, Lave T. A novel strategy for physiologically based predictions of human pharmacokinetics. Clin Pharmacokinet. 2006;45: 511-542.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
4. Gueorguieva II, Nestorov IA, Rowland M. Fuzzy modeling and simulation of pharmacokinetic models: case study of whole body physiologically based model of diazepam. J Pharmacokinet Pharmacodyn. 2004;31: 185-213.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  C. J. Horvath and M. N. Milton The TeGenero Incident and the Duff Report Conclusions: A Series of Unfortunate Events or an Avoidable Event? Toxicol Pathol, April 1, 2009; 37(3): 372 - 383. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
