Background: Growing evidence for clinically significant differences between the sexes has attracted the attention of researchers. However, failures to report a test animal sex and balance the sex ratios of study samples remain widespread in preclinical investigations. We analyzed the sex-reporting rate and sex distributions of test animals in published oncology studies.
Methods: We selected five oncology journals included in the Scientific Citation Index (SCI) based on impact factors. We identified preclinical investigations with in vivo mouse experiments published in 2015 for inclusion in our study sample. We classified each article by whether or not it reported test subject sex, and by which sex was included. We also recorded whether there were justifications for using one particular sex in single-sex studies (e.g., anatomical reasons) and whether sex-based analyses were conducted for both-sex studies.
Results: We surveyed a total of 382 articles. Half (50.3%) failed to report test animal sex. Among articles that did report sex, 91.7% were single-sex studies, of which 69.4% did not provide any justifications for using the sex included in the study. Relatively few studies 15.7 studies included animals of both sexes, and only 2.3 studies conducted sex-based analyses. These findings are consistent with those of previous research that used other methods to collect data from the literature such as text mining, but our analysis of the provision of justifications for using one sex versus the other is a novel feature.
Conclusions: Many researchers overlook test subject sex as a factor, but test animal sex should be reported in all preclinical investigations to enhance the reproducibility of research and avoid faulty conclusions drawn from one-sided studies.
Keywords: animal experimentation; bias; data accuracy; data curation; research design; research subjects.
Copyright © 2018 The Korean Society of Critical Care Medicine.