Abstract
Spinal cord injury research in experimental animals aims to define mechanisms of tissue damage and identify interventions that can be translated into effective clinical therapies. Highly reliable models of injury and outcome measurement are essential to achieve these aims and avoid problems with reproducibility. Functional scoring is a critical component of outcome assessment and is currently commonly focused on open field locomotion (the "BBB score"). Here we analyze variability of observed locomotor outcome after a highly regulated spinal cord contusion in a large group of rats that had not received any therapeutic intervention. Our data indicate that, despite tight regulation of the injury severity, there is considerable variability in open-field score of individual rats at 21 days after injury, when the group as a whole reaches a functional plateau. The bootstrapped reference interval (that defines boundaries that contain 95% scores in the population without regard for data distributional character) for the score at 21 days was calculated to range from 2.3 to 15.9 on the 22-point scale. Further analysis indicated that the mean day 21 score of random groups of 10 individuals drawn by bootstrap sampling from the whole study population varies between 9.5 and 13.5. Wide variability between individuals implies that detection of small magnitude group-level treatment effects will likely be unreliable, especially if using small experimental group sizes. To minimize this problem in intervention studies, consideration should be given to assessing treatment effects by comparing proportions of animals in comparator groups that attain pre-specified criterion scores.