Interpreting the toxicologic significance of alterations in anogenital distance: potential for confounding effects of progeny body weights

Anogenital distance (AGD) is an endpoint that was recently added to the U.S . EPA testing guidelines for reproductive toxicity studies. This endpoint i s sensitive to hormonal effects of test chemicals. It is possible that appa rent alterations in AGD might occur after treatment with agents that affect overall pup body size. In such cases, hormonal activity might be associate d incorrectly with the test treatment. The analyses in this report evaluate d statistical correlations between pup body weight and AGD in control litte rs. AGDs were measured on postnatal day 1 in 1501 pups derived from 113 unt reated female Sprague-Dawley rats in two independent two-generation reprodu ctive toxicity studies. Significant correlations were detected between AGD and body weight and between AGD and the cube root of body weight. In males, AGD increased 0.26 mm for each 1 g increase in body weight. In females, AG D increased 0.13 mm per I g increase in body weight. Although there were es sentially no differences between the regression models developed to predict AGD in either males or females using body weight as a covariate and those based on the cube root of body weight, such similarities in predictivity mi ght not occur in larger animals with broader weight ranges than those encou ntered in this analysis. Normalization of AGD by dividing by body weight si gnificantly overcompensated for differences in body size. Normalizing with the cube root of body weight resulted in an AGD/cube root of body weight ra tio that was constant across the range of body weights observed in this stu dy. In conclusion, as a preferred method to account for body size effects o n AGD, analysis of covariance is recommended. If a normalization is done di rectly, the ratio of AGD to the cube root of body weight is the more approp riate metric. (C) 1999 Elsevier Science Inc. All rights reserved.