Mixture-based gatekeeping procedures in adaptive clinical trials.

Clinical trials with data-driven decision rules often pursue multiple clinical objectives such as the evaluation of several endpoints or several doses of an experimental treatment. These complex analysis strategies give rise to "multivariate" multiplicity problems with several components or sources of multiplicity. A general framework for defining gatekeeping procedures in clinical trials with adaptive multistage designs is proposed in this paper. The mixture method is applied to build a gatekeeping procedure at each stage and inferences at each decision point (interim or final analysis) are performed using the combination function approach. An advantage of utilizing the mixture method is that it enables powerful gatekeeping procedures applicable to a broad class of settings with complex logical relationships among the hypotheses of interest. Further, the combination function approach supports flexible data-driven decisions such as a decision to increase the sample size or remove a treatment arm. The paper concludes with a clinical trial example that illustrates the methodology by applying it to develop an adaptive two-stage design with a mixture-based gatekeeping procedure.

An enhanced mixture method for constructing gatekeeping procedures in clinical trials.

It is increasingly common to encounter complex multiplicity problems with several multiplicity components in confirmatory Phase III clinical trials. These components are often based on several endpoints (primary and secondary endpoints) and several dose-control comparisons. When constructing a multiplicity adjustment in these settings, it is important to control the Type I error rate over all multiplicity components. An important class of multiple testing procedures, known as gatekeeping procedures, was derived using the mixture method that enables clinical trial sponsors to set up efficient multiplicity adjustments that account for clinically relevant logical relationships among the hypotheses of interest. An enhanced version of this mixture method is introduced in this paper to construct more powerful gatekeeping procedures for a specific type of logical relationships that rely on transitive serial restrictions. Restrictions of this kind are very common in Phase III clinical trials and the proposed method is applicable to a broad class of multiplicity problems. Several examples are provided to illustrate the new method and results of simulation trials are presented to compare the performance of gatekeeping procedures derived using this method and other available methods.

Mixture-based gatekeeping procedures for multiplicity problems with multiple sequences of hypotheses.

Complex multiplicity problems arise in drug development programs with several sets of clinical objectives. This article considers a common setting with two sources of multiplicity induced by the analysis of multiple dose levels based on ordered endpoints. This results in multiplicity problems with multiple sequences of null hypotheses of no effect. Type I error rate inflation in problems of this type is typically addressed by using gatekeeping procedures that account for the hierarchical structure of the trial objectives. A general method for building gatekeeping procedures, known as the mixture method, tends to be conservative in problems with several sequences of hypotheses. This article defines a modified mixture method and shows that this method provides a power advantage over the standard mixture method. In addition, it is demonstrated that in special cases the modified mixture method allows for a stepwise testing algorithm, which facilities the implementation of gatekeeping procedures and general decision making. The new methodology is illustrated using two clinical trial examples.

Statistical challenges in a regulatory review of cardiovascular and CNS clinical trials.

There are several challenging statistical problems identified in the regulatory review of large cardiovascular (CV) clinical outcome trials and central nervous system (CNS) trials. The problems can be common or distinct due to disease characteristics and the differences in trial design elements such as endpoints, trial duration, and trial size. In schizophrenia trials, heavy missing data is a big problem. In Alzheimer trials, the endpoints for assessing symptoms and the endpoints for assessing disease progression are essentially the same; it is difficult to construct a good trial design to evaluate a test drug for its ability to slow the disease progression. In CV trials, reliance on a composite endpoint with low event rate makes the trial size so large that it is infeasible to study multiple doses necessary to find the right dose for study patients. These are just a few typical problems. In the past decade, adaptive designs were increasingly used in these disease areas and some challenges occur with respect to that use. Based on our review experiences, group sequential designs (GSDs) have borne many successful stories in CV trials and are also increasingly used for developing treatments targeting CNS diseases. There is also a growing trend of using more advanced unblinded adaptive designs for producing efficacy evidence. Many statistical challenges with these kinds of adaptive designs have been identified through our experiences with the review of regulatory applications and are shared in this article.

Superchain procedures in clinical trials with multiple objectives.

This paper introduces a new class of multiple testing procedures for addressing multiplicity problems arising in clinical trials with multiple objectives grouped into families. The families may correspond to equally important sets of objectives (co-primary endpoints) or ordered sets of objectives (primary and secondary endpoints). The procedures, termed superchain procedures, serve as an extension of several classes of other multiple testing procedures, including chain procedures and parallel gatekeeping procedures. Superchain procedures exhibit several desirable features, including flexible decision rules that can be tailored to a broad class of win criteria in confirmatory clinical trials. Additionally, superchain procedures enable the trial's sponsor to efficiently incorporate available distributional information to improve overall power. Finally, superchain procedures rely on straightforward testing algorithms, which facilitates their development. We illustrate the new methodology by using clinical trial examples with two and three families of objectives.

Multistage and mixture parallel gatekeeping procedures in clinical trials.

Gatekeeping procedures have been developed to solve multiplicity problems arising in clinical trials with hierarchical objectives where the null hypotheses that address these objectives are grouped into ordered families. A general method for constructing multistage parallel gatekeeping procedures was proposed by Dmitrienko et al. (2008). The objective of this paper is to study two related classes of parallel gatekeeping procedures. Restricting to two-family hypothesis testing problems, we first use the mixture method developed in Dmitrienko and Tamhane (2011) to define a class of parallel gatekeeping procedures derived using the closure principle that can be more powerful than multistage gatekeeping procedures. Second, we show that power of multistage gatekeeping procedures can also be improved by using α-exhaustive tests for the component procedures. Extensions of these results for multiple families are stated. Illustrative examples from clinical trials are given.

Method of balanced adjustment in testing co-primary endpoints.

In a clinical trial, if there are three or more co-primary endpoints, the type II error could increase depending on the correlation among the endpoints and their treatment effect sizes. To keep the type II error under control one may have to consider larger sample sizes. However, in cases where treatment effect size of at least one of the endpoints is likely to be small, the required sample size estimates can exceed reasonable bounds. Patel (1991) proposed an approach that adjusts the significance level for testing each primary endpoint based on the idea of restricting the null space. In Chuang-Stein et al. (2007), the upward adjustment to the significance levels is based on controlling an average type I error rate. In the scenario that statistical significance of each individual hypothesis is not required, we introduce a compromise testing approach in which the significance level for a co-primary endpoint is adjusted upward only if the treatment shows high significance in one (or more than one) of the remaining co-primary endpoints. The adjustment depends on the correlation among the endpoints: larger adjustment is needed for cases of smaller correlation. The method is applicable for the scenario where the null space is restricted. Our testing approach controls maximum joint false positive rate over the restricted null space.

The US Food and Drug Administration's perspective on the new antidepressant vortioxetine.

OBJECTIVE: This article summarizes the US Food and Drug Administration's (FDA's) review of the New Drug Application for vortioxetine, especially the clinical efficacy and safety data. It emphasizes the issues that were important to the FDA's approval decision, particularly the difference in the effective dose in domestic and foreign studies, and notes several new labeling features, specifically, description of time course of treatment response and detailed sexual dysfunction evaluation. DATA SOURCES: The data sources were the original raw data sets for all clinical trials included in the development program for vortioxetine, as well as the sponsor's original analyses of these data. Data were available from 51 human trials involving vortioxetine, and included a total of 7,666 healthy volunteers and patients with a diagnosis of major depressive disorder (MDD) or generalized anxiety disorder who were exposed to at least 1 dose of vortioxetine for a total of 2,743 patient-years. RESULTS: Vortioxetine was effective in treating MDD in the United States at a dose of 20 mg/d. The recommended starting dose is 10 mg once daily without regard to food, with increase to 20 mg/d if the 10 mg/d dose is tolerated. For patients who do not tolerate 20 mg/d, 10 mg/d can be used and 5-mg/d dose can be considered. Vortioxetine can be discontinued abruptly, but it is recommended that doses of 15 mg/d or 20 mg/d be reduced to 10 mg/d for 1 week prior to full discontinuation to avoid potential withdrawal symptoms. Although the non-US maintenance study showed that maintenance doses of 5 to 10 mg/d were effective, a clinical judgment needs to be made to decide the maintenance dose in the United States. The applicant has agreed to conduct a US maintenance dose-response study covering the US-approved dose range. Vortioxetine's adverse event profile is similar to that of other selective serotonin reuptake inhibitors (SSRIs). Nausea is the most common adverse event and is dose dependent. No dose adjustment is needed based on age, gender, or the presence of renal or mild to moderate hepatic impairment. The maximum recommended dose is 10 mg/d in known cytochrome P450 2D6 poor metabolizers. CONCLUSIONS: Vortioxetine is a new treatment for MDD, and its adverse event profile is similar to that of other SSRIs.

Sex-related differences in QTc effects potential of drugs.

The objective of this study was to examine the drug-induced sex differences in corrected QT (QTc) interval by re-analyzing the data collected in thorough QT studies submitted to the US Food and Drug Administration (FDA). We examined 64 studies available in the FDA database by performing a time-matched, baseline adjusted ANCOVA on the QTc response stratified by sex. We used several summaries to capture the differences between males and females in drug response QTc effects. They included sample means, upper confidence intervals, and areas under the curves. At baseline, females tend to have a higher QTc response than males. After treatment, various summaries suggest that females tend to have a higher QTc effect than males. However, the magnitude of the difference is small and is often not statistically significant. Several limitations can be raised about these available data: 1) available QT studies were not designed to examine the sex differences in QTc effects, 2) the findings were undermined by large variations seen in QT data, and 3) our summary statistics are descriptive in nature and are not for inferential purposes. Nonetheless, the results suggest that females tend to have a higher QTc effect than males, although the difference tends to be small. Further research is needed to formally address the question.