Abstract

This report presents a statistical analysis of an investigation into the subchronic toxicity of patulin in rats. The method used is analysis of variance with explanatory factors 'concentration of patulin' and 'sex'. The factor 'concentration' was split into (orthogonal) polynomials. The resulting regression model can be used to estimate concentrations that have a toxic effect of a given magnitude, with confidence limits. The lower limit of one such estimate can be used for regulatory purposes, as an alternative tot the NOEC-value. The effects of patulin on the creatinin clearance and feed conversion are described. Finally, recommendations are given with respect to the design of toxicological experiments in general.