Pharmacokinetics — Uses and abuses

Summary It is apparent from studying recent articles on pharmacokinetics that a number of misunder-standings exist, both about the design of experiments and the analysis of results. The purpose of this paper is to outline many of the common pitfalls associated with the design of experiments and also the limitations upon the analysis of results. The paper describes mathematical, laboratory and clinical aspects which must be examined in designing a protocol for pharmacokinetic experiments. Simulated data is presented to demonstrate the dangers of using standard computer programs for parameter estimation. Even when convergence is obtained the answers may be dependent on the method employed. A mathematical model is of little use unless a reasonable amount of good, accurate data is obtained.     

Essential properties of drug-targeting delivery systems

How, if at all, can drug delivery help to create ideal drugs? After four decades of trying, an effective site-specific drug-delivery system has not yet been developed. This review draws attention to the pharmacokinetic conditions that must be met to achieve a successful performance by site-selective drug-carrier delivery systems. In a drug-carrier approach, a drug is attached to a macromolecular carrier via a chemically labile linker. The carrier transports the drug to its site of action and releases it at the target site. For this simple approach to work, several fundamental conditions (nonspecific interactions, target site access, drug release and drug suitability) must be satisfied. The importance of these essential requirements, not always recognized in the development of drug-delivery systems, is discussed and illustrated by recent examples selected from the literature.     

On some “disadvantages” of the population approach

In a seminal article on population pharmacokinetic modeling, researchers demonstrated how means and variances of pharmacokinetic parameters for a patient population could be inferred from sparse data collected under conditions of routine patient care. But they also identified 4 potential concerns about their methodology: unobserved confounding variables may bias the inferences; conditions under which data are collected may lead to inaccuracies of reporting or recording; correlations among important predictor variables may reduce statistical efficiency; and costs cannot be controlled by principles of study design. Experiences are reviewed that relate to these potential disadvantages. A method is presented for diagnosing the possible presence of confounding. A model is constructed and applied that captures the influences of data inaccuracies. An example of selecting from among correlated covariates is summarized. Finally, a methodology for optimal study design is reviewed and applied to an example.     

Pharmacokinetics/pharmacodynamics and the stages of drug development: Role of modeling and simulation

Pharmacokinetic (PK) and pharmacodynamic (PD) modeling and simulation (M&S) are well-recognized powerful tools that enable effective implementation of the learn-and-confirm paradigm in drug development. The impact of PK/PD M&S on decision making and drug development risk management is dependent on the question being asked and on the availability and quality of data accessible at a particular stage of drug development. For instance, M&S methodologies can be used to capture uncertainty and use the expected variability in PK/PD data generated in preclinical species for projection of the plausible range of clinical dose; clinical trial simulation can be used to forecast the probability of achieving a target response in patients based on information obtained in early phases of development. Framing the right question and capturing the key assumptions are critical components of the "learn-and-confirm" paradigm in the drug development process and are essential to delivering high-value PK/PD M&S results. Selected works of PK/PD modeling and simulation from preclinical to phase III are presented as case examples in this article.     

Possible synergistic prostate cancer suppression by anatomically discrete pomegranate fractions

We investigated whether dissimilar biochemical fractions originating in anatomically discrete sections of the pomegranate (Punica granatum) fruit might act synergistically against proliferation, metastatic potential, and phosholipase A2 (PLA2) expression of human prostate cancer cells in vitro . Proliferation of DU 145 human prostate cancer cells was measured following treatment with a range of therapeutically active doses of fermented pomegranate juice polyphenols (W) and sub-therapeutic doses of either pomegranate pericarp (peel) polyphenols (P) or pomegranate seed oil (Oil). Invasion across Matrigel by PC-3 human prostate cancer cells was measured following treatment with combinations of W, P and Oil such that the total gross weight of pomegranate extract was held constant. Expression of PLA2, associated with invasive potential, was measured in the PC-3 cells after treatment with the same dosage combinations as per invasion. Supra-additive, complementary and synergistic effects were proven in all models by the Kruskal-Wallis non-parametric H test at p < 0.001 for the proliferation tests, p < 0.01 for invasion, and p < 0.05 for PLA2 expression. Proliferation effects were additionally evaluated with CompuSyn software median effect analysis and showed a concentration index CI < 1, confirming synergy. The results suggest vertical as well as the usual horizontal strategies for discovering pharmacological actives in plants.     

Pseudo cluster randomization: a treatment allocation method to minimize contamination and selection bias

In some clinical trials, treatment allocation on a patient level is not feasible, and whole groups or clusters of patients are allocated to the same treatment. If, for example, a clinical trial is investigating the efficacy of various patient coaching methods and randomization is done on a patient level, then patients who are receiving different methods may come into contact with each other and influence each other. This would create contamination of the treatment effects. Such bias might be prevented by randomization on the coaches level. The patients of a coach constitute a cluster and all the subjects in that cluster receive the same treatment. Disadvantages of this approach may be reduced statistical efficiency and recruitment bias, as the treatment that a subject will receive is known in advance. Pseudo cluster randomization avoids this, because in pseudo cluster randomization, not everybody in a certain cluster receives the same treatment, just the majority. There are two groups of clusters: in one group the majority of subjects receive treatment A, while a limited number receive treatment B. In the other group of clusters the proportions are reversed. The statistical properties of this method are described. When contamination is present, the method appears to be more efficient than randomization on a patient level or on a cluster level.     

A comparison of intent-to-treat and per-protocol results in antibiotic non-inferiority trials

While the intent-to-treat (ITT) analysis is widely accepted for superiority trials, there remains debate about its role in non-inferiority trials. It is often said that the ITT tends to be anti-conservative in the demonstration of non-inferiority. This concern has led to some reliance on per-protocol (PP) analyses that exclude patients on the basis of post-baseline events, despite the inherent bias of such analyses. We compare ITT and PP results from antibiotic trials presented to the public at the FDA's Anti-infective Drug Advisory Committee from 1999 to 2003. While the number of available trials is too small to produce clear conclusions, these data did not support the assumption that the ITT would lead to smaller treatment difference than the PP, in the setting of antibiotic trials. Possible explanations are discussed.     

Factorial design provides evidence to guide practice of anaesthesia

Many scientific articles are written merely to get something published, neglecting the clinician who would like the medical literature to guide their practice. Evidence-based medicine is expected to help in clinical decision-making. Systematic reviews of the literature followed by a meta-analysis of randomized, controlled trials (RCT) have claimed to represent the highest strength of evidence. However, the results published in meta-analyses have not always been confirmed in subsequent large RCTs. An analysis of 12 large RCTs and 19 meta-analyses addressing the same questions found that the outcomes of these large RCTs were not predicted accurately 35% of the time by previously published meta-analyses. Therefore, meta-analyses of several small RCTs do not obviate the need for large, multicentre RCTs, which can still be considered as a gold standard for the development of clinical guidelines or practice plans. Moreover, large RCTs using a factorial design can be highly efficient because they can answer several clinical questions at the same time and offer the only systematic approach to investigate an interaction of combinations in multimodal approaches.     

统计学的三型理论及其在生物医学科研中的应用

目的从学术角度指出误用和滥用统计学问题产生的症结,并创造性地提出能够从根本上解决前述问题的“统计学的三型理论”,即问题的“表现型”、“原型”和“标准型”。方法从理论与实践两个方面阐明前述提及的“三型理论”是合理应用统计学的要领,而且,通过实例说明“三型”之间是互相有联系的。用“三型理论”去解释和分析生物医学科研中的实验设计和统计分析问题,可以帮助人们透过事物的现象看清其本质。结果通过研究发现：有些问题的“三型”之间是彼此相同的,有些问题的原型就是标准型,然而,还有些问题的“三型”是彼此不同的。特别是在某些多因素实验研究中,研究者犯了“对照不全”的错误,导致与其“原型”对应的“标准型”根本不存在,这是一个需要引入“拆分”的思想和方法方可解决的问题。结论一旦弄清了每个具体问题的“三型”,合理进行实验设计与正确选择统计分析方法处理特定的实验资料就是水到渠成的事了。“统计学的三型理论”,可以帮助人们在运用统计学的过程中不犯统计学错误或至少会大大降低误用统计学的机会,可在很大程度上提高生物医学科研的质量、水平和速度;可以提高统计学教材的质量和教学效果,为生物医学统计学的学科发展指明了前进的方向。