Reproducibility of ethanol elimination rates in long-sleep and short-sleep mice

Ethanol metabolism was measured in long-sleep (LS) and short-sleep (SS) mice on two occasions separated by 1 week to test for repeatability. Mice were injected intraperitoneally with either 1.5 or 4.0 g/kg ethanol and the linear decline of blood ethanol level was measured. The parameters measured in each animal were linear ethanol elimination rate (EER), peak blood ethanol level, volume of distribution and Widmark ratio (r). Reproducibility was assessed using two statistical methods, paired t tests and Pearson correlations. Paired t tests indicated good reproducibility since the two replicate determinations did not differ significantly from each other. The other widely used indicator of reproducibility, the correlation coefficient (Pearson r) between the two measurements, was nonsignificant in almost every case, indicating poor reproducibility. This occurs because the range of values of EER is fairly narrow; thus, an individual is likely to fall anywhere within that narrow range from one day to the next, and the rank ordering of the individuals may not be maintained. Although parameters such as EER and volume of distribution appear to be reproducible for populations, they may have little or no utility as covariates in genetic analyses of individual differences in response to ethanol.