Standards in behavioral teratology testing: test variability and sensitivity.

Abstract

Regulatory guidelines have produced a need to develop behavioral screening techniques to accompany teratology and reproduction testing. In the repeated use of a provisional test battery, we have found that conclusions about the behavioral teratogenic potential of test compounds are most likely to be revealed using tests having intermediate levels of variability. The results obtained from examining animals exposed developmentally to brominated vegetable oil (BVO as 2.0, 1.0, 0.5, or 0.25 percent of the diet) and comparisons of the tests' coefficients of variation, offer an empirical example of this concept. Preweaning tests of locomotion and reflex development demonstrated numerous instances of developmental delay in the BVO-treated subjects, but postweaning tests revealed few abnormalities. Behavioral testing revealed functional deficits from BVO administration at doses lower than those which have adverse effects on reproduction. Examination of the tests' variability by using coefficients of variation as a comparative index, disclosed that the postweaning test variability was almost twice that of the preweaning tests. Thus, the lack of effects of BVO on most of the postweaning tests should not be conclusively interpreted as indicative of recovery of function, because this pattern is equally likely to have resulted from the lower sensitivity of these tests. An acceptable standard for future behavioral teratology screening requires a close examination of test variability, as it appears to be an important element in the sensitivity and, hence, the interpretation of such procedures.