Estimates of appropriate number of rats: interaction with housing environment.
Mering, S., Kaliste-Korhonen, E. and Nevalainen, T.
Laboratory Animals, 35(1), 80-90 (2001).
An extensive list of physiological parameters from previous experiments was re-analysed in order to evaluate the effects of enrichment, cage type and group size on the within-group variation and hence on the number of animals needed in studies using Wistar rats. The independent factors studied in these experiments included the provision of aspen gnawing blocks for enrichment, solid bottom cages (SBCs) and grid ¯oor cages (GFCs) and animal number per cage (varied from 1±4). SOLO power analysis was used to calculate the smal lest number of animals (n)needed to detect an arbitrarily chosen 20% effect size, when signi®cance was set at Pˆ0.05 and statist ical power at 0.90. N ratios (nlarger=nsmaller) were calculated for the effect of enrichment, cage type and group size to compare the `treatment group’ with the `control group’. The n values of adrenal gland, interscapular brown adipose tissue (BAT ) and epididymal adipose tissue (EAT )weights seemed to varymost, whereas ®nal body weight (FBW) and growth seemed to be the least variabl e ones. According to one-sample t-test, the N ratios of most physiological parameters differed signi®cant ly from zero (except the ones of FBW) indicating that n values in `treatment’ and `control’ groups were unequal. The results indicate that some of the physiological parameters are susceptible to variability attribut able to environmental modi®cat ions in general whereas some are not. Furthermore, they suggest that the variation of different parameters may vary from one experiment to another and between different environments thus hindering the estimat ions of appropriate number of animals.