A meta-analysis of clinical pathology results from South African bred Sprague Dawley rats

Abstract

Rats are laboratory animals that play a pivotal role in pre-clinical toxicity testing. There are two major classes of rats available for this exercise, inbred and outbred stocks. Inbred stocks are well known for their genetic uniformity, which normally results in less variation phenotypically when responding to pharmacological effect as opposed to outbred stocks during pre-clinical toxicity trials. Despite these animals being derived and classified, they are not resistant to genetic drift, which can influence the design, results and reproducibility of the said animal model. While genetic drift may not result in major changes in the genetic characteristics of the animals, it can also have more insidious changes in the clinical pathology of the animals. Since South Africa represents a small breeding colony, we undertook a retrospective study for this study to ascertain if there is a risk of genetic drift existing in the Sprague Dawley rats. A meta-analysis of the mean parameter values obtained from four previously conducted independent studies was undertaken. Statistical heterogeneity was interpreted as the variability of pathological parameters across the four studies considered. The 𝐼�����2 statistic was used to describe the percentage of variation across the studies that is due to heterogeneity rather than chance To determine whether statistical heterogeneity is greater than it would be by chance, a chi-squared test of the hypothesis of heterogeneity of effects using Cochrane’s Q statistic was conducted. Random effects models were a natural choice based on the assumption that toxicity levels (via measured parameters) were not identical but follow some distribution. The model represents a lack of knowledge about why real, or apparent, genetic drift effects differ across the studies by considering the differences as if they were random. From the meta-analysis the 90% confidence intervals were determined and compared to the reference intervals available from Charles River laboratories, a major rodent breeder in the world. There were variations for parameter across the three main categories of white blood cells, red blood cells and serum chemistry. While some variables did fall within the published reference intervals, there were many deviations. For instance, absolute monocyte counts were higher than the normal ranges for both genders while mean corpuscular haemoglobin concentration and haematocrit had lower values than normal ranges. While the animals selected for this analysis are known to be healthy, and also confirmed on pathology, these post mortem findings supports our concerns of genetic drift being present and interfering with clinical pathology values. This has implications on future toxicology studies conducted on inbred SD rats and the interpretation of results, as the Charles River published reference intervals are unlikely to be relevant to South Africa. It is thus recommended that the results determined by meta-analysis should be validated or a completely new reference interval be established in these SD rats as per recommended statistical methodology for determining clinical pathology reference intervals.