Simon Kirby

Pharmacological reversal of a pain phenotype in iPSC-derived sensory neurons and patients with inherited erythromelalgia.

A selective Nav1.7 sodium channel blocker reduced hyperexcitability of iPSC-derived sensory neurons and alleviated pain in a subpopulation of patients with an inherited pain disorder. A gain in pain control Subtype-specific blockade of sodium channel Nav1.7, which is important for firing of peripheral pain-signaling neurons, is a major focus of pain research. In a new study, Cao et al. created iPSC-derived sensory neurons from patients with inherited erythromelalgia (IEM), a painful disorder in which gain-of-function Nav1.7 mutations produce hyperexcitability and hyperresponsiveness to warmth in peripheral sensory neurons. The investigators show that a new selective Nav1.7 sodium channel blocker normalized the phenotype of iPSC-derived sensory neurons carrying IEM mutations and blocked pain perception in human subjects with IEM. These results provide proof of principle that selective Nav1.7 blockade may be useful in pain alleviation. In common with other chronic pain conditions, there is an unmet clinical need in the treatment of inherited erythromelalgia (IEM). The SCN9A gene encoding the sodium channel Nav1.7 expressed in the peripheral nervous system plays a critical role in IEM. A gain-of-function mutation in this sodium channel leads to aberrant sensory neuronal activity and extreme pain, particularly in response to heat. Five patients with IEM were treated with a new potent and selective compound that blocked the Nav1.7 sodium channel resulting in a decrease in heat-induced pain in most of the patients. We derived induced pluripotent stem cell (iPSC) lines from four of five subjects and produced sensory neurons that emulated the clinical phenotype of hyperexcitability and aberrant responses to heat stimuli. When we compared the severity of the clinical phenotype with the hyperexcitability of the iPSC-derived sensory neurons, we saw a trend toward a correlation for individual mutations. The in vitro IEM phenotype was sensitive to Nav1.7 blockers, including the clinical test agent. Given the importance of peripherally expressed sodium channels in many pain conditions, our approach may have broader utility for a wide range of pain and sensory conditions.

A Quantitative Approach for Making Go/No-Go Decisions in Drug Development

There are many decision points along the product development continuum. Formal clinical milestones, such as the end of phase 1, phase 2a (proof of mechanism or proof of concept), and phase 2b provide useful decision points to critically evaluate the accumulating data. At each milestone, sound decisions begin with asking the right questions and choosing the appropriate design as well as criteria to make go/no-go decisions. It is also important that knowledge about the new investigational product, gained either directly from completed trials or indirectly from similar products for the same disorder, be systematically incorporated into the evaluation process. In this article, we look at metrics that go beyond type I and type II error rates associated with the traditional hypothesis test approach. We draw on the analogy between diagnostic tests and hypothesis tests to highlight the need for confirmation and the value of formally updating our prior belief about a compounds effect with new data. Furthermore, we show how incorporating probability distributions that characterize current evidence about the true treatment effect could help us make decisions that specifically address the need at each clinical milestone. We illustrate the above with examples.

Candidate biomarkers for the diagnosis and prognosis of drug‐induced liver injury: An international collaborative effort

Current blood biomarkers are suboptimal in detecting drug‐induced liver injury (DILI) and predicting its outcome. We sought to characterize the natural variabilty and performance characteristics of 14 promising DILI biomarker candidates. Serum or plasma from multiple cohorts of healthy volunteers (n = 192 and n = 81), subjects who safely took potentially hepatotoxic drugs without adverse effects (n = 55 and n = 92) and DILI patients (n = 98, n = 28, and n = 143) were assayed for microRNA‐122 (miR‐122), glutamate dehydrogenase (GLDH), total cytokeratin 18 (K18), caspase cleaved K18, glutathione S‐transferase α, alpha‐fetoprotein, arginase‐1, osteopontin (OPN), sorbitol dehydrogenase, fatty acid binding protein, cadherin‐5, macrophage colony‐stimulating factor receptor (MCSFR), paraoxonase 1 (normalized to prothrombin protein), and leukocyte cell‐derived chemotaxin‐2. Most candidate biomarkers were significantly altered in DILI cases compared with healthy volunteers. GLDH correlated more closely with gold standard alanine aminotransferase than miR‐122, and there was a surprisingly wide inter‐ and intra‐individual variability of miR‐122 levels among healthy volunteers. Serum K18, OPN, and MCSFR levels were most strongly associated with liver‐related death or transplantation within 6 months of DILI onset. Prediction of prognosis among DILI patients using the Model for End‐Stage Liver Disease was improved by incorporation of K18 and MCSFR levels. Conclusion: GLDH appears to be more useful than miR‐122 in identifying DILI patients, and K18, OPN, and MCSFR are promising candidates for prediction of prognosis during an acute DILI event. Serial assessment of these biomarkers in large prospective studies will help further delineate their role in DILI diagnosis and management.

The shrinking or disappearing observed treatment effect.

It is frequently noted that an initial clinical trial finding was not reproduced in a later trial. This is often met with some surprise. Yet, there is a relatively straightforward reason partially responsible for this observation. In this article, we examine this reason by first reviewing some findings in a recent publication in the Journal of the American Medical Association. To help explain the non-negligible chance of failing to reproduce a previous positive finding, we compare a series of trials to successive diagnostic tests used for identifying a condition. To help explain the suspicion that the treatment effect, when observed in a subsequent trial, seems to have decreased in magnitude, we draw a conceptual analogy between phases II-III development stages and interim analyses of a trial with a group sequential design. Both analogies remind us that what we observed in an early trial could be a false positive or a random high. We discuss statistical sources for these occurrences and discuss why it is important for statisticians to take these into consideration when designing and interpreting trial results.

Adaptive modelling of dose–response relationships using smoothing splines

We consider the use of smoothing splines for the adaptive modelling of dose-response relationships. A smoothing spline is a nonparametric estimator of a function that is a compromise between the fit to the data and the degree of smoothness and thus provides a flexible way of modelling dose-response data. In conjunction with decision rules for which doses to continue with after an interim analysis, it can be used to give an adaptive way of modelling the relationship between dose and response. We fit smoothing splines using the generalized cross-validation criterion for deciding on the degree of smoothness and we use estimated bootstrap percentiles of the predicted values for each dose to decide upon which doses to continue with after an interim analysis. We compare this approach with a corresponding adaptive analysis of variance approach based upon new simulations of the scenarios previously used by the PhRMA Working Group on Adaptive Dose-Ranging Studies. The results obtained for the adaptive modelling of dose-response data using smoothing splines are mostly comparable with those previously obtained by the PhRMA Working Group for the Bayesian Normal Dynamic Linear model (GADA) procedure. These methods may be useful for carrying out adaptations, detecting dose-response relationships and identifying clinically relevant doses.

An Example of an Unblinded, Third-Party Interim Analysis for Sample Size Re-Estimation

An interim analysis for sample size re-estimation is often carried out in a clinical trial when there is concern about the uncertainty of the estimate of variability used for the prestudy sample size estimation. The use of an unblinded third party to do the interim analysis instead of conventional blinded methods may be possible and may lead to the desired power for the study being more likely to be achieved. An example of a clinical trial where the unblinded, third-party approach was used is described. For the interim results obtained, the decision not to increase the sample size would also have been reached had the project team used either of two commonly adopted blinded sample size re-estimation methods. However, simulation results from some plausible alternative scenarios based on the example show that an unblinded, third-party interim analysis for sample size re-estimation can provide more accurate results than the two blinded methods.

Designing Phase 4 Trials

Phase 4 trials are conducted for a variety of reasons. These include investigating a marketed drug in pediatric patients, comparing a drug head-to-head with another drug, investigating the effect of a drug at a lower/higher dose or with different administration schedules, studying a drug in combination with other drugs, or testing a drug for other indications. In this chapter, we cover the design of Phase 4 trials from the perspective of obtaining a prior distribution for the treatment effect from past trials. We first focus on comparing different drugs in a network meta-analysis and proceed to consider a trial comparing a drug against a comparator using information obtained from the network meta-analysis. We also discuss how prior distributions for treatment effects may be obtained by other means such as PK/PD modeling.

A Frequentist Decision-Making Framework

In this chapter, we review the Frequentist approach to hypothesis testing and in particular the two-action decision problem. The Frequentist approach to statistical inference involves examining probability via its long-run frequency interpretation. Procedures for assessing evidence and making decisions under the Frequentist approach are calibrated by how they would perform when they are used repeatedly. In the context of drug development, the two actions correspond to progressing or not progressing a drug for further development. These two actions are fundamental to decision-making in drug development.

Choosing Metrics Appropriate for Different Stages of Drug Development

In this chapter, we give an overview of metrics that are useful to evaluate designs for determining if a new drug is efficacious at each of the three premarketing clinical development stages. The three stages are to determine if the new drug exhibits a positive proof of concept, to explore a possible dose-response relationship, and to confirm a hypothesized drug effect. The focus on efficacy is due to the generally well-defined endpoints to decide the beneficial effect of a new drug. Nevertheless, the approach is equally applicable to safety endpoints if there are specific safety endpoints that can be used to anchor design considerations and decision rules. Deliberations of the metrics for each stage will be further elaborated in Chaps. 7– 9 respectively.

Discounting Prior Results to Account for Selection Bias

In this chapter, we consider possible selection bias that can occur due to the fact that a Phase 2 trial is usually selected for use in future trial planning only when it has produced a positive result. We first illustrate the phenomenon of selection bias and relate it to the general concept of regression to the mean. We take a simplified situation where a single Phase 2 trial comparing a single dose versus a control produced a promising treatment effect. We review several approaches that have been proposed to discount the observed treatment effect in the single Phase 2 trial when results from the trial are used to plan a Phase 3 trial. While we have hinted at the existence of selection bias in earlier chapters, Chap. 12 is the first time we consider this issue thoroughly. We offer some recommendations on how to watch out for selection bias when planning for late stage trials.