Abstract
In vivo testing is a crucial part of product development. Most in vivo toxicology studies are conducted in a heavily regulated environment, in accordance with sound scientific principles, and in an ethical manner. Statistical methods are an important component of these scientific principles, and many regulatory authorities provide guidelines or recommendations for statistical analyses and reporting. We propose a rigorous and comprehensive statistical approach to assess the impact of smoke exposure during in vivo comparative studies using statistical equivalence with variable equivalence limits based on historical data. The analysis enables comparison of the effect of a new/test aerosol to its reference/control while leveraging information about reference item variation across multiple studies. This approach is helpful in assessing the relevance of observed effects while incorporating sources of variability estimated using data from the current and previous experiments. The proposed method was used to analyze an OECD (Organization for Economic Co-operation and Development) 90-day inhalation study to determine the effects of exposure to smoke emitted by the new 1R6F reference cigarette as compared to the older 3R4F reference cigarette. Data from previous OECD 90-day in vivo studies using 3R4F cigarette smoke were used to improve our assessment of the effects observed in the current study.