| 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. |
|||||||||
HIGHLIGHTS AND BOOK REVIEWS |
Senior Director, Early Development and Clinical Pharmacology, Wyeth Research
Manager, Pilot Plant R & D, Cadila Pharmaceuticals Ltd, Ahemdabad, India
Address for reprints: Correspondence regarding the Highlights and Book Reviews section may be directed to Peter L. Bonate, PhD, Genzyme Corporation, 4545 Horizon Hill Blvd, San Antonio, TX 78229; e-mail: peter.bonate{at}genzyme.com.
RESEARCH HIGHLIGHTS
Andrew MA, Easterling TR, Carr DB, et al. Amoxicillin pharmacokinetics in pregnant women: modeling and simulation of dosage strategies. Clin Pharmacol Ther. 2007;81:547-556.
Amoxicillin has been recommended by the American College of Obstetricians and Gynecologists (ACOG) as 1 treatment for postexposure anthrax prophylaxis if the bacteria are penicillin sensitive. Furthermore, the Centers for Disease Control and Prevention (CDC) recommends "if test results show that the anthrax bacteria can be killed by penicillin antibiotics, the use of amoxicillin is recommended to prevent the development of anthrax disease in people who have been exposed to anthrax, when other antibiotics are not as safe to use such as with children and pregnant women" (the bold text is verbatim from its Web page). The recommended dose was 500 mg orally 3 times daily for 60 days (the recommended dose for nonpregnant women), despite there being no published studies on the pharmacokinetics of amoxicillin in pregnant women. Andrew and colleagues published the results of a phase I trial in 17 healthy pregnant women who were administered single-dose amoxicillin 500 mg after 
6 hours of fasting on 3 separate occasions: second trimester, 18 to 22 weeks' gestation; third trimester, 30 to 34 weeks' gestation; and 3 months ±2 weeks postpartum, with the latter serving as a nonpregnant control visit. Serial plasma samples were collected for 12 hours, as was urine in 4-hour intervals. Samples were analyzed for amoxicillin using high-performance liquid chromatography (HPLC) with mass spectral detection. Amoxicillin pharmacokinetics were characterized noncompartmentally for renal excretion and compartmentally for plasma parameters using a 1-compartment model with nonlinear absorption followed by summarization using the standard 2-stage approach.
To examine the variability in concentrations at each pregnancy stage and at varying doses, Monte Carlo simulation was used based on the population estimates obtained following the 2-stage approach. The following regimens were examined: 500 mg every 4, 6, and 8 hours; 750 mg every 6 hours; and 1000 mg every 6 hours. A total of 500 individuals were simulated for each pregnancy state; from these, concentration-time profiles after 4 doses were simulated. The peak and trough concentration for each subject was obtained and summarized using descriptive statistics.
Changes in amoxicillin pharmacokinetics were observed across pregnancy states. Apparent oral clearance and renal clearance were higher while pregnant. The increase in renal clearance appeared to be due to an increase in filtration and secretory clearance or decreased reabsorption in the kidneys. Half-life was shorter during pregnancy. Volume of distribution was not reported, so any changes to that parameter are unknown. The results from these changes suggest that amoxicillin exposure will be less while pregnant and that maintenance of trough concentrations will be difficult.
Monte Carlo simulations indicated that a major revision in dosage recommendations should be made in pregnant women to meet the anthrax minimum inhibitory concentrations (MICs) required at pseudo-steady state. Increased clearance and decreased half-life resulted in lower area under the curve and maximal concentrations such that more frequent dosing is needed to offset these pregnancy-induced changes. However, the 4-hour dosing interval that may be needed to achieve MICs would be difficult to maintain from a compliance perspective and raises the possibility that amoxicillin may not be an appropriate therapy for postexposure anthrax prophylaxis. However, the same situation may exist for other postexposure anthrax therapies, such as ciprofloxacin, but studies such as this one for amoxicillin are lacking, and so their use is more caveat emptor. It will be interesting to see whether the CDC or ACOG changes its recommendations based on this report.
Barrett JS, Gupta M, Mondick JT. Model-based drug development applied to oncology. Expert Opin Drug Discov. 2007;2:185-209.
Kummar S, Gutierrez M, Doroshow JH, Murgo A. Drug development in oncology: classic cytotoxics and molecularly targeted agents. Br J Clin Pharmacol. 2006;62:15-26.
Jeff Barrett and colleagues have published a very informative review on the application of modeling and simulation (M&S) applied to the development of oncolytics. The review discusses how M&S can be used at the various stages of drug development to help guide decision making. Two case studies are also discussed: the docetaxel story started by Rene Bruno et al in their classic paper published in the Journal of Pharmacokinetics and Biopharmaceutics in 1997 (v. 24, pp. 152-172) and actinomycin D, a lesser known drug first introduced into the clinic in 1954, in which M&S was used to help guide development of a pediatric protocol conducted through the Children's Oncology Group. Model-based drug development (MBDD) uses mathematical models of pharmacokinetic, pharmacodynamic disease progression and trial designs to model the link between drug administration and outcome. This paradigm differs from traditional paradigms that base dosing on empirical results, and this is where the article from Kummar and colleagues comes in.
Shivaani Kummar and colleagues published an article less than a year ago on drug development in oncology in the British Journal of Clinical Pharmacology. The other authors of this article include Anthony Murgo, head of Early Clinical Trial Development at the National Cancer Institute (NCI), and James Doroshow, senior editor for Pharmacology in Clinical Cancer Research and director of the Division of Cancer Treatment and Diagnosis at the NCI. These authors carry a lot of weight in the field of drug development of oncolytics. This article presents a good survey on the differences in the development of cytotoxic drugs, which are usually dosed to the maximally tolerated dose (MTD), and cytostatics, which may have their maximal effect at doses less than the MTD. The article discusses phase I designs, phase II designs, use of biomarkers, the importance of drug interactions, and the future use of pharmacogenomics in the development of cancer drugs. What was lacking in this article was any mention whatsoever of MBDD or M&S of any kind.
These 2 articles highlight the chasm that lies between pharmacologists/pharmacometricians and clinicians. The article by Kummar et al was written by clinicians; the article by Barrett and colleagues was written by clinical pharmacologists/modelers. The sad fact of the matter is that few companies are embracing M&S in the development of oncolytics. It would be great to see an article such as Barrett's in the Journal of Clinical Oncology. Those who practice M&S need to do better at introducing these concepts to physicians and presenting their results at clinical meetings, instead of the usual meetings such as those for the American College of Clinical Pharmacy, American Society of Clinical Pharmacology and Therapeutics, or American Association of Pharmaceutical Scientists. More presentations need to be made at the meetings of the American Society of Clinical Oncology, American Association for Cancer Research, or European Organization for Research and Treatment of Cancer on the use of this technology to leverage the information collected during drug development. Only then will MBDD become a true reality in oncology.
Bauer RJ, Guzy S, Ng C. A survey of population analysis methods and software for complex pharmacokinetic and pharmacodynamic models with examples. AAPS J. 2007; 9:Article 7.
Although NONMEM is the de facto leader in software for nonlinear mixed-effects modeling, there are a number of other programs available, including PDx-MCPEM, S-Adapt, Monolix, and WinBugs. Robert Bauer, Serge Guzy, and Chee Ng present a review of the various programs, their estimation algorithms, and a comparison of their parameter estimates when applied to 4 different data sets. Their review of the estimation algorithms is the most up-to-date review on this topic in some time and shows how the programs differ in how they estimate the parameters in a model. The review also presents a comparison of the features of the software such as type of interface and ease of use. The examples they use to compare the programs range from a simple 2-compartment model to a complex pharmacokinetic-pharmacodynamic model with nonlinear elimination in the pharmacokinetic model coupled to an indirect response model.
There were too many conclusions in this study to list them all. Some of the major ones include that, among the 2-stage hierarchical models, NONMEM first-order approximation performed well with the well-behaved 2-compartment model when the variance components were <10%. Accuracy was improved if the data and model were log-transformed prior to analysis so that a homoscedastic error could be assumed. NONMEM first-order conditional estimation (FOCE) gave accurate population values regardless of the magnitude of the intersubject variability but took much longer to run than the first-order approximation method. If the data are very sparse, the FOCE can lead to inaccurate values and may require a more accurate estimation method such as the Laplacian option. The latter method, however, is not very robust, and one should use the method only after obtaining final estimates from FOCE. The authors conclude that given today's computer speeds, NONMEM's first-order approximation algorithm should never be used for a final analysis and that FOCE be used when the data are reasonably rich in nature, can be expressed as an analytical form, and extensive covariate modeling and hypothesis testing needs to be done.
They also conclude that the expectation-maximization (EM) algorithms (S-Adapt, MC-PEM, and Monolix) have greater stability in analyzing complex pharmacokinetic-pharmacodynamic models and have reduced bias compared to likelihood approximation methods. In their more complex examples with 8 or more estimable parameters, the EM algorithms were more stable and faster than NONMEM's FOCE. WinBugs typically took twice as long as S-Adapt and 4 to 5 times longer than NONMEM FOCE for the first 2 examples. But for the more complex models, it was as fast as the EM algorithms and took no longer than NONMEM's FOCE. In example 4, which used differential equations, WinBugs took many hours compared to 30 minutes for the EM algorithms. Because the 3-stage hierarchical model, as used in WinBugs, is more computationally expensive than the likelihood approximation methods, it is not recommended to use WinBugs with poor initial values. The authors speculate that in the future, modelers will first work out the model by performing 2-stage hierarchical analyses and then performing a 3-stage hierarchical Bayesian analysis. This article provides a good introduction to the different methods available to modelers to tackle problems, and it provides a glimpse into some of the more esoteric programs being used today (which may be the disruptive technologies of the future).
Ellenbogen JM, Hu PT, Payne JD, Titone D, Walker M. Human relational memory requires time and sleep. Proc Natl Acad Sci USA. 2007;104:7723-7728.
This study had nothing to do with clinical pharmacology but was fascinating just the same. Relational memory is the ability to generalize across different stores of information and allows one to make innovative memory decisions in new situations. One such example is transitive inference, so that if one learns the individual premises A > B and B > C (which are called premise pairs), then one can infer that A > C (which is called the inference pair), despite never learning that directly. In practice, if you learned that driving south on Interstate-95 will take you from Boston to New York and that you can also drive south on I-95 from New York to Washington, DC, you could infer that driving south on I-95 will take you from Boston to Washington, DC, despite never learning this specifically. However, despite learning the premise pairs, sometimes the connection to the inferential pair is never made.
Jeffrey Ellenbogen and colleagues studied how time and sleep affect the ability to make these inferential connections. A group of 56 mentally healthy participants between 18 and 30 years old who abstained from caffeinated beverages for 24 hours prior to and during the study performed an initial training session where they were asked to learn a series of premise pairs: A > B, B > C, C > D, D > E, and E > F, which consisted of a series of novel visual patterns. They were then immediately tested for their ability to have learned the premise pair. Following an offline period of 20 minutes, 12 hours, or 24 hours, participants were randomized to groups and tested for their ability to form novel transitive, inferential relations (eg, B > E).
There were no differences in the groups' ability in the time taken to learn the premise pairs or in the groups' immediate test recall. Furthermore, even after an offline period, the ability to retain the premise pairing was retained across groups. However, after an offline period, the groups showed striking differences in their ability to draw transitive inferences. The group with a 20-minute offline gap had a worse performance than the 12- and 24-hour groups (75% correct). Chance alone would predict 50% correct. Hence, the longer offline period resulted in a better ability to make inferential judgments.
As a second test, subjects in the 12-hour group were further randomized to be trained in the evening and, following a night of sleep, performed their inferential test (the sleep group) or to be trained in the morning and tested later that day in the evening (the wake group). Both groups showed identical training and immediate recall testing results, as well as retained knowledge of the premise pairs during the inferential testing session. Although the overall ability to form transitive inferences was not statistically higher in the sleep group (79%) compared to the wake group (72%), when examined on the basis of a 1-degree (eg, C > A) or 2-degree (eg, D > A) separation, a qualitative difference emerged. The sleep group was able to form significantly more 2-degree transitive inferences (93%) than the wake group (70%). Interestingly, in a series of poststudy interviews, subjects in the sleep group did not show any greater confidence in their inferential ability than subjects in the wake group.
This study shows that it is better to have offline periods after learning so that the necessary transitive formations can occur and that sleep improves the ability to form higher order transitive relationships. Interestingly, the formation of these relationships did not appear to be consciously apparent to the subject. We have all seen examples in our own lives where taking some time off from a problem helps us to make a breakthrough and hence the old phrase "sleep on it." This study scientifically confirms it.
Reynolds KK, Valdes R Jr, Hartung BR, Linder MW. Individualizing warfarin therapy. Future Med. 2007;4:11-31.
Warfarin has been marketed as an anticoagulant for more than 50 years. Despite this, dosing of warfarin is still largely empirical, with dosing adjustments based on having an international normalized ratio (INR), a measure of clotting time, within some therapeutic range. It is widely known that warfarin exhibits large between-subject variability and that adverse events, particularly bleeding-related ones, are common during the first year of therapy. In the past few years, there have been many reports on genetic associations with warfarin therapy dose requirements. Kristen Reynolds and colleagues have written an excellent review article on these studies and how genomic information could be used to help guide warfarin dosing.
Warfarin is a mixture of S- and R-enantiomers, with the S-warfarin being 3 to 5 times more active than R-warfarin. S-warfarin is almost exclusively metabolized by CYP2C9 to inactive metabolites, whereas the R-enantiomer is metabolized by CYP3A4, 1A2, and 2C19. Many studies have shown that age, sex, ethnicity, weight, height, body surface area, heart disease, liver disease, drug interactions, dietary vitamin K intake, CYP2C9 genotype, and vitamin K epoxide reductase complex subunit 1 (VKORC1) all affect warfarin dosing. Of these, CYP2C9 accounts for 5% to 20% of the variability in dosing requirements. When this is combined with the VKORC1-1639 G>A genotype, 35% to 40% of the variability in dosing requirements can be explained. With the addition of other patient characteristics, such as age, sex, and weight, these numbers can further be increased to approximately 50% to 60%.
As Reynolds et al state, "The direct application of clinical pharmacogenetics to the creation of individualized dosing strategies has not yet been achieved to an appreciable extent." How true. Most physicians would argue that even with genomic information, they would not know how to use this information; also, why should an expensive genomic test be done when they can use a cheaper, simple test such as INR to empirically dose warfarin? Arguments such as these are holding up less and less. Many studies, which are summarized in the article, have shown that warfarin-associated adverse drug reactions (ADRs) are common and that patients with atypical genotypes are at increased risk for ADRs. A recent study has shown that integrating genetic testing into warfarin therapy could avoid 85 000 serious bleeding events and 17 000 strokes annually, resulting in a savings to the US health care system of $1.1 billion.
Reynolds et al also discuss scenarios about how routine genomic testing and genotype-guided maintenance dosing could be implemented. This is a masterful summary of recent events and should be a must-read for those interested in individualized therapy.
Bansal S, DeStefano A. Key elements of bioanalytical method validation for small molecules. AAPS J. 2007;9:Article 11.
Surendra Bansal and Anthony DeStefano present a very readable and useful summary of the key elements in bioanalytical method validation, with an emphasis on the chromatographic analysis of small molecules based on the Food and Drug Administration (FDA) guidance and the 2006 Bioanalytical Methods Validation white paper. Useful summary tables on parameter requirements are presented—for example, spacing of quality control (QC) samples and acceptance criteria for the lower limit of quantification standards, as well as parameter definitions, are presented. The topics covered are comprehensible and not verbose. This would be a very useful document to send to analytical sites that might not be familiar with the FDA guidance on method validation. For example, a study might require analysis of some special peptide that cannot be routinely analyzed and can only be done by an academic investigator. This document would be a useful reference for the investigator to have prior to sample analysis so that both the sponsor and the investigator can agree on how to best validate the method.
Kallen A. A simple absorption model for dose-escalating studies. J Pharmacokinet Pharmacodyn. 2007;34:251-261.
Anders Kallen proposes a very simple modification to the first-order rate of absorption that can account for first-order, zero-order, and some type of mixed-order absorption. Recall that in a first-order absorption model, where loss of drug occurs from the dosing compartment, the rate of loss is proportional to the amount in the compartment (A) or, mathematically, dA/dt = -K*A, where K is the rate of loss from the compartment per time. For a zero-order loss from the compartment, dA/dt = -K, where -K has units of the amount per time. Kallen proposes that the rate equation be modified to dA/dt = -K*A
; when equals 0, this reduces to zero order, and when equals 1, this reduces to first-order loss. Intermediate values of could result in some type of mixed-order loss from the dosing compartment. This is an interesting model because, although not discussed, an expanded version of this could allow to vary, thereby allowing the model to convert from, say, a zero-order input to a first-order input. Kallen illustrates the application of the model to data obtained from a dose-escalating, first-time-in-man study.
Liu M, Wei L, Zhang J. Review of guidelines and literature for handling missing data in longitudinal clinical trials with a case study. Pharm Stat. 2006;5:7-18.
Missing data are found in every study for a variety of reasons—patients do not show up because they have moved, samples break in the centrifuge, or the nurse forgets to fill in a field on a case report form. More and more, pharmacokinetic data are being linked to pharmacodynamic data; as such, the issue of handling missing data is falling into the realm of the pharmacokinetic modeler. Liu, Wei, and Zhang have presented a review on missing data theory and methods that is not too complex for the clinician or modeler to understand. The authors discuss missing data models such as selection models and pattern mixture models, missing data terminology such as missing at random and nonignorability, parameter estimation using likelihood or imputation in the face of missing data, and guidelines in designing clinical trials. They conclude with a case study of missing data in a diabetes trial. The example was a bit unusual in the sense that they were concerned with 2 primary endpoints and had to consider the case for a joint missing data mechanism, but other than this, the case was interesting. This was a very nice and simple introduction to the missing data problem and possible solutions.
Ludbrook J. Writing intelligible prose for biomedical journals. Clin Exp Pharmacol Physiol. 2007;34:508-514.
Writing scientific papers is not a topic usually taught in graduate school. Very few scientists get any more than ad hoc or on-the-job training prior to graduating. As such, most scientists write in a complex technical manner that is becoming increasingly difficult to comprehend. John Ludbrook recently published an essay on writing intelligible prose in biomedical journals. He recommends that scientists need a firm grasp of English grammar, usage, and style, which should occur during schooling, followed by constant practice, especially in the style and format for biomedical journals. In other words, get a good foundation and then practice, practice, practice. The problem, as he sees it, is that good reading skills are needed by almost all vocations, but the writing of prose is only needed in a few, including technical fields. Furthermore, most journal editors do not have the time or staff to thoroughly edit a poorly written manuscript. Most reviewers rarely do more than a cursory editorial job during their review of papers submitted for publication. If it is too late for formal schooling, then he recommends reading books and essays on scientific English. He notes that some journals, such as Chest, have periodic columns on medical writing tips and that some universities have short courses on medical writing. He also recommends the use of freelance science editors in editing manuscripts—they are well worth the money, in his opinion. He offers some simple tips, strategies, and words that can be used to increase clarity in writing for biomedical journals. What I liked about this article was that it brings to light a neglected aspect in science, and being aware of a problem is the first step toward fixing it.
Yengi LG, Leung L, Kao J. The evolving role of drug metabolism in drug discovery and development. Pharm Res. 2007;24:842-858.
Testa B, Kramer S. The biochemistry of drug metabolism—an introduction: Part 1. Principles and overview. Chem Biodiversity. 2006;3:1053-1101.
Testa B, Kramer S. The biochemistry of drug metabolism—an introduction: Part 2. Redox reactions and their enzymes. Chem Biodiversity. 2007;4:257-405.
In 1990, 40% of drug attrition was due to problems related to pharmacokinetics. That fact is well known within our field. To understand the importance that drug metabolism has made to drug development is to recognize that a decade later, that number was down to 10%. Lilian Yengi, Louis Leung, and John Kao have written an excellent review on what has happened to drug metabolism over the past 10 years, as well as including some commentary on the history of drug metabolism. The review is well structured and highly informative with a wealth of very useful references.
The articles by Bernard Testa and Stefanie Kramer are not articles in the traditional sense but a novel application of technology and publishing. These authors are lecturers on drug metabolism to graduate-level students. Their lectures are done using Microsoft PowerPoint. What they have done in these articles is to reprint their slides, 1 to a page, and then underneath present an extensive caption to accompany the slide. The articles conclude with an extensive reference list. The first article is an introduction, and the second article is on redox reactions. They plan to publish 5 more articles in the same format, followed by collecting the entire series and publishing it as a book. For someone wishing to learn drug metabolism, so far these articles are worth looking at.
Zandvliet AS, Huitema ADR, Copalu W, et al. CYP2C9 and CYP2C19 polymorphic forms are related to increased indisulam exposure of higher risk of severe hematologic malignancy. Clin Cancer Res. 2007;13:2970-2976.
Indisulam is a sulfonamide anticancer agent that is well tolerated but only has limited single-agent activity and is currently being studied for the treatment of solid tumors. Phase I studies have shown that the dose-limiting toxicity with indisulam was neutropenia and thrombocytopenia. Pharmacokinetically, indisulam is metabolized by CYP2C9 and CYP2C19 to form a hydroxylate metabolite that is immediately conjugated to form an O-glucuronide and O-sulfate metabolite. Population analysis showed that clearance could best be described by 2 parallel pathways, one saturable and the other non-saturable. Wide between-subject variability (BSV) was observed, 45% for Vmax.
Anthe S. Zandvliet and colleagues present the results of an analysis in 46 Caucasian and 21 Japanese patients who were genotyped for the *2, *3, *4, and *6 polymorphisms of CYP2C9 and the *2, *3, *4, *5, and *6 polymorphisms for CYP2C19. Indisulam pharmacokinetics were extensively characterized in these patients as part of the phase I program. Using a previously reported population pharmacokinetic model as the starting point, the elimination model was expanded to include the genotype as a means to further explain the BSV in indisulam pharmacokinetics. Letting P be the population parameter in wild-type patients and assuming that homozygous patients will have twice the impact of heterozygous mutations, their model for Vmax, Km, and linear clearance was
 |
Using this model, a 27% decrease in Vmax was found between wild-type and CYP2C9*2 polymorphs and a 38% decrease in linear clearance between wild-type and CYP2C19*3 polymorphs. Patients with these mutations had a higher AUC than wild-type patients. Due to the small number of patients, a quantitative analysis of CYP polymorphisms and grade of neutropenia could not be made. However, at higher dose levels, a relationship between total dose and number of mutations and grade of neutropenia was observed. The pharmacokinetic model was coupled to a cell transit model previously developed for absolute neutrophil count (ANC). Using Monte Carlo simulation, they showed that the relative risk of having dose-limiting neutropenia was 40% higher in patients having a single polymorphism. Homozygous mutations increased to a 2-fold risk. There did not appear to be a difference in the risk of neutropenia between Caucasians and Asians after controlling for genotype. They then used Monte Carlo simulation to determine the degree of dose reduction needed to equalize the degree of relative risk in patients with high-risk mutations. They found that each CYP2C19 mutation required a dose reduction of 100 mg/m2 and that the reduction for a CYP2C9*3 mutation was 50 mg/m2.
This study beautifully illustrates the role of modeling and simulation in helping to guide dosing regimens in oncology and further illustrates the increasing importance that pharmacogenetics may play in personalized medicine.
BOOK REVIEWS
Dubovsky SL, Dubovsky AN. Psychotropic Drug Prescriber's Survival Guide: Ethical Mental Treatment in the Age of Big Pharma. New York: W. W. Norton; 2007; 187 pp., $27.50.
As someone who has worked for pharmaceutical companies all his life, this reviewer obviously has a different perspective on the drug development process than the authors, who are academicians. As such, I think this book provides a completely unbalanced perspective of drug development and the business side of pharmaceuticals. The book is essentially a 187-page rant on the evils of big pharma and its role in medical research as it relates to psychotropic drugs. The basic premise is that big pharma companies publish research that is mostly biased, flawed, or an outright lie designed to make their drug look "better" than a competitor's product. No one is spared in this book; the authors criticize pretty much all aspects of industry-sponsored drug research, the motives of industry, the role of the FDA in the process, and medical journals for publishing the studies.
This book was ostensibly written because "clinicians need to be aware of the financial motivations of drug companies, and interpret industry findings accordingly." Only the most naive physician would be unaware that the pharmaceutical industry is a business and the job of a sales representative is sales. The book references heavily Marcia Angell's 2005 book on the pharmaceutical industry, and readers familiar with that book, its allegations, and tone will know exactly what they can expect from this book. Authors of such one-sided portrayals do little to help clinicians or the public understand the complexity of drug development or the risk-benefit relationship that must be evaluated for almost all therapeutic interventions, including pharmacologic treatment.
Moyé LA. Statistical Reasoning in Medicine: The Intuitive P-Value Primer. 2nd ed. New York: Springer; 2006; 301 pp., $54.95.
Some of us involved in clinical research have likened statistical knowledge to driving a car. Can we learn to be good drivers without completely understanding the principles of internal combustion or rack- and-pinion steering? To continue the analogy, with formal education in both mathematics and medicine, as well as experience as a physician, researcher, epidemiologist, and statistician, Lemuel Moyé, MD, PhD, has written this text with the goal of making us more aware of the traffic laws governing many of the common tools involving the interpretation of P values used in clinical trials without deriving the equations and using much mathematics beyond simply stating the equations of interest. His goal has been achieved, and although the text does not provide extensive operational direction, most readers will be very happy with the result and better able to make choices in research and statistical design and understand how choices may influence the interpretation of the results.
The book is written in a wonderfully descriptive manner. It is intended to be read cover to cover, although the chapters on subgroup analyses and regression analyses may be consulted several times by new researchers. The book begins with a brief review of the origins and early history of statistical methodology. The background information is very helpful to providing a context for why certain statistical tests have the properties that they do. Several chapters begin with a brief vignette that provides insight into the topic that will be covered, and there are helpful diagrams included that help clarify the principles presented. Within each topic, the author provides helpful context for the implications of different choices made in the design of the study and statistical testing procedures. Several important case studies are presented with references for further reading as well as identification of any potential biases that the author might have in discussing the cases. The book is well organized with a helpful table of contents and index.
Readers may find it helpful to bookmark the pages where equations are presented in text so that they are more easily available for reference in later chapters. The author includes equations as footnotes in some cases, but the reader may have to flip through the book looking for the explanation of a term. There are appendices that provide more discussion on standard normal probabilities and sample size calculations. Although the version that I reviewed was from the first printing, and corrections may already have been made, there are several unfortunate typos, including some referring to incorrect references and nonexistent appendices.
The book is intended for a nonmathematically inclined audience and, although it is well referenced, researchers will need additional texts to serve as user guides to actually implement various statistical tests. However, the book will be helpful for investigators and practitioners who make extensive use of literature. Nonstatistical members of clinical development teams and teachers helping students with science fair projects would all find this book interesting and helpful.
Abdel-Magid AF, Caron S. Fundamentals of Early Clinical Drug Development: From Synthesis Design to Formulation. New York: John Wiley; 2006; 323 pp.
This book is "deceptive" from its size and the large amount of information contained within regarding the development of drug compound. The authors have compiled relevant information from across the different fraternities concerned with pharmaceutical development. For instance, the book provides an opportunity for the medicinal drug chemist to understand the difficulties an organic chemist faces in the development of a particular route of synthesis and to appreciate the difficulties and criticalities an engineer faces during scale-up. An engineer can further understand the importance of an otherwise simple process of crystallization to develop a particular form/habit/size of an active product ingredient as a critical factor to formulation for chemists and scientists.
Another impressive thing about the book is its practical relevance and industrial applications. One author has provided his own industrial experience and challenges faced when solving problems. From the very beginning, enough illustrative examples are provided to inspire "experienced" and "not so experienced" scientists/chemists/engineers to think beyond "the box" while tackling problems in their areas. Practical relevance is also highlighted by including chapters such as "Outsourcing—The Challenge of Science, Speed, and Quality," which accounts for economic feasibilities.
Although the book covers a wide area, it fails to provide detailed insight, which of course can be attributed to its size and its actual purpose of being a "reference." This drawback is compensated to quite an extent by a long list of references at the end of each chapter, which can encourage readers to quench their thirst for more detailed information. Looking at the spectrum of areas covered by this book, I feel the book will impress management groups involved in product development and launch apart from chemists, scientists, and chemical engineers associated with drug development, from lab synthesis to formulation. Last, the content and the title of the book do not seem synonymous. The title of the book at first glance gives an impression that it discusses fundamentals of clinical drug development that have more relevance to medicinal and formulation chemists and may not appeal to synthetic chemists, process development engineers, and product development managers. Skimming the chapters inside can highlight its usefulness. A more appropriate title, such as Fundamentals of Drug Development: Synthesis Design, Scale-Up, and Formulation, would have made more sense.
BOOKS AVAILABLE FOR REVIEW
The following books are available for review. If you would like to write a review for the journal, please contact Peter Bonate at peter.bonate{at}genzyme.com.
Arya DP, ed. Aminoglycoside Antibiotics: From Chemical Biology to Drug Discovery. New York: John Wiley; 2007.
Bacchieri A, Della Cioppa G. Fundamentals of Clinical Research: Bridging Medicine, Statistics, and Operations. New York: Springer; 2007.
Ekins S, ed. Computational Toxicology: Risk Assessment for Pharmaceutical and Environmental Chemicals. New York: John Wiley; 2007.
Hasko G, Cronstein BN, Szabo C, eds. Adenosine Receptors: Therapeutic Aspects for Inflammatory and Immune Diseases. Boca Raton, FL: CRC Press; 2006.
Hernandez MA, Rathinavelu A. Basic Pharmacology: Understanding Drug Actions and Reactions. Boca Raton, FL: CRC Press; 2006.
Hofbauer KG, Anker SD, Inui A, Nicholson JR, eds. Pharmacotherapy of Cachexia. Boca Raton, FL: CRC Press; 2006.
Kleinbaum DG, Sullivan KM, Barker ND. A Pocket Guide to Epidemiology. New York: Springer; 2007.
Langel U, ed. Handbook of Cell-Penetrating Peptides. 2nd ed. Boca Raton, FL: CRC Press; 2006.
Singh M. Vaccine Adjuvants and Delivery Systems. New York; John Wiley; 2007.
Strom BL, Kimmel SE, eds. Textbook of Pharmacoepidemiology. New York: John Wiley; 2006.
Vandenbroeck K. Cytokine Gene Polymorphisms in Multifactorial Conditions. Boca Raton, FL: CRC Press; 2006.
LETTER FROM THE SECTION EDITOR
If you are interested in contributing a summary of an article you think is particularly notable and should be featured in the Research Highlights of the journal, please contact Peter Bonate at peter.bonate{at}genzyme.com or contact the Journal of Clinical Pharmacology through the managing editor at m.spraycar{at}verizon.net.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati 
Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
