Reliability of the Mouse Grimace Scale in C57BL/6JRj Mice

SIMPLE SUMMARY: The protection and welfare of laboratory animals is ethically and legally required. Any possible impairment of well-being caused by husbandry or experiment must be reduced to a minimum. A prerequisite for ameliorating animal welfare is that we recognize the current well-being state of the animals by using measuring tools, for instance, the grimace scale. Facial expressions of mice, the most frequently used laboratory animals, can be analyzed using the Mouse Grimace Scale, which consists of five action units, i.e., orbital tightening, nose and cheek bulge, ear position, and whisker change. Since it is important that the animal welfare measuring tools are reliable and objective, we investigated the reliability of the Mouse Grimace Scale. Our results indicated good agreement between all observers applying the Mouse Grimace Scale. However, when investigating the action units individually, the best agreement was achieved for orbital tightening and the poorest agreement for nose as well as cheek bulge, depending on the observers’ experience levels. Against this background, we critically discuss factors that potentially influence the reliability of the Mouse Grimace Scale to improve its application. ABSTRACT: To maintain and foster the welfare of laboratory mice, tools that reliably measure the current state of the animals are applied in clinical assessment. One of these is the Mouse Grimace Scale (MGS), a coding system for facial expression analysis. Since there are concerns about the objectivity of the MGS, we further investigated its reliability. Four observers (two experienced and two inexperienced in use of the MGS) scored 188 images of 33 female and 31 male C57BL/6JRj mice. Images were generated prior to, 150 min, and two days after ketamine/xylazine anesthesia. The intraclass correlations coefficient (ICC = 0.851) indicated good agreement on total MGS scores between all observers when all three time points were included in the analysis. However, interrater reliability was higher in the early post-anesthetic period (ICC = 0.799) than at baseline (ICC = 0.556) and on day 2 after anesthesia (ICC = 0.329). The best agreement was achieved for orbital tightening, and the poorest agreement for nose and cheek bulge, depending on the observers’ experience levels. In general, experienced observers produced scores of higher consistency when compared to inexperienced. Against this background, we critically discuss factors that potentially influence the reliability of MGS scoring.