The Use of Trend Tests to Determine a No-Observable-Effect Level in Animal Safety Studies

Abstract: Proper statistical analysis of data from subchronic and chronic stud-ies is a key aspect of their data analysis and interpretation. We discuss the motivation and illustrate the use of trend tests for continuous quantitative data from such studies. Because responses are often more severe (increase or de-crease) with increasing dose, trend tests are the most appropriate statistical procedures to apply. Several parametric and nonparametric trend tests are defined and applied to a small set of male liver weight data from a 14-day rat study. Shortcomings of more standard statistical procedures applied to data from such studies are discussed as well. We recommend the use of trend tests as the customary method for analyzing data from these studies. Key Words: Statistical methods-Data analysis-Subchronic study-Chronic study-Safety testing. Statistical analyses of data from animal safety studies have commonly been part of the routine data analysis and interpretation in such studies for many years. However, despite the advances in statistical science and particularly with respec