A case study of virtual control group verification using historical control data at a single facility

In general toxicity studies in rodents, groups are initially assigned based on body weight, highlighting the importance of matching the body weights of test article groups when creating Virtual Control Groups (VCGs). In this study, we aimed to generate VCGs by linking toxicity parameters to initial body weights and to verify the significance of body-weight matching. The Concurrent Control Groups (CCGs) from nine 4-week rat studies at a single facility served as Historical Control Data (HCD). VCGs were generated by random sampling of initial body weights in HCD, and a ± 5 g range from the mean initial body weight of the test article group was used as the body-weight-matching criteria. VCGs were used to assess statistical differences in aspartate aminotransferase (AST), alanine aminotransferase (ALT), and liver weight at the end of one specific study. Of the 10 VCGs generated, two had deviations of more than 5 g from the mean initial body weight value. Statistically significant differences in AST and ALT levels observed in CCGs tended to become non-significant in VCGs, regardless of the mean initial body weight of VCGs. Meanwhile, VCGs outside the ± 5 g range revealed new statistically significant differences in liver weight. Liver weight was weakly correlated with initial body weight, suggesting that deviations in initial body weight influenced the statistical significance of liver-weight differences at the end of the study. The importance of matching body weight in VCGs was confirmed, particularly for toxicity parameters that correlate with initial body weight values.