Experimental Design for Sampling Multiple Offspring per Litter to Evaluate Reproductive Developmental Effects Following In Utero Through Lactational Chemical Exposure in Sprague

• Activity Number: 363
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Zhenxu J. Ma*+ and Paul I. Feder and Don R. Bergfelt and Ralph L. Cooper and David P. Houchens
• Companies: Battelle Memorial Institute and Battelle Memorial Institute and Environmental Protection Agency and Environmental Protection Agency and Battelle Memorial Institute
• Keywords: sampling design ; endocrine disruptor ; power analysis ; mixture of normal and Poisson ; mixture of normal and binomial ; interlitter and intralitter variances

Abstract

Interaction of certain chemicals with the endocrine system has been reported to result in reproductive developmental effects in humans and wildlife. Data for the present study were made available by the US Environmental Protection Agency to examine the added power obtained by increasing the number of F1 male or female offspring selected per litter and ability to detect statistically significant effects on hormone-dependent endpoints. Reproductive developmental endpoints were recorded as continuous, count and categorical measurements following in utero through lactational exposure to VIN, DBP and two chemical mixtures. Power analyses were conducted for each endpoint to evaluate the effects of sampling 1, 3 or 5 pups per litter involving 20 or 10 litters. Hierarchical distributions and Monte Carlo simulations were conducted, which incorporated between-pup and between-litter variabilities. Power analyses demonstrated that, regardless of litter size, there was a greater increase of power going from 1 to 3 pups per litter compared to 3 to 5 pups per litter. Moreover, by increasing number of pups sampled per litter, comparable results could be obtained with a reduced number of litters.