Industry Comments Seek Changes on Endpoints Guidance
Novartis says that a recent draft guidance on “Multiple Endpoints in Clinical Trials” needs to make the distinction between secondary and exploratory endpoints more clear. In just-posted comments, the company says the guidance provides a definition of primary, secondary and exploratory endpoints, where all secondary endpoints require multiplicity adjustments, along with primary endpoints, but not the exploratory endpoints. It suggests that the document could elaborate more on the distinction by taking into account “clinical importance and potential regulatory actions (e.g. additional labeling claims vs. descriptive statements in a package insert).”
The draft guidance was issued in January, and it discusses the problems posed by multiple endpoints in the analysis and interpretation of drug/biologic study results and how these problems can be managed in clinical trials. Most drug clinical trials contain multiple endpoints to assess the effects of the drug and to document a product’s ability to favorably affect one or more disease characteristics, the document notes. “As the number of endpoints analyzed in a single trial increases, the likelihood of making false conclusions about a drug’s effects with respect to one or more of those endpoints becomes a concern if there is not appropriate adjustment for multiplicity,” it says. The guidance describes various strategies for grouping and ordering endpoints for analysis and applying statistical methods for managing multiplicity within a study in order to control the chance of making erroneous conclusions about a drug’s effects.
Comments from AstraZeneca signal disappointment that a harmonized global position on handling multiplicity has not been reached. “It is strongly encouraged that this should be further explored between regulators around the world in order to produce a harmonized position via [International Council for Harmonization],” it says.
AstraZeneca’s comments take issue with FDA’s statement that: “Both the issues and methods that apply to multiple endpoints also apply to other sources of multiplicity, including multiple doses, time points, or study population subgroups.” The company says that this position is not correct. “It is agreed that most of the methods are the same for other sources of multiplicity but the issues associated with the methods are different,” it explains. “For example, defining a correlation structure between different dose groups has more logic than defining correlation structure for different endpoints... To focus the document only on multiple endpoints misses an opportunity to provide detailed guidance on other sources of multiplicity such as multiple dose groups.”
Comments from GlaxoSmithKline note that the guidance is “specifically focused on studies when null hypothesis significance tests will be used to determine efficacy and/or safety. However there are hypothesis generating/ exploratory studies where alternative approaches are used, e.g. multivariate inferential methods via joint modeling of endpoints, or the Bayesian approach to Type I error control, etc.” The company asks the agency to provide “further discussion about the issue of multiplicity in studies not intended to demonstrate effectiveness and support drug approval and some of the methods that are more applicable for these studies (i.e. multivariate inferential methods, Bayesian methods, etc.).”