An animal model of posttraumatic stress disorder: the use of cut-off behavioral criteria.

Considerable heterogeneity exists in the response of human subjects exposed to extreme traumatizing events. Posttraumatic stress disorder (PTSD) is diagnosed in 20-30% of those exposed. Clinical studies of this population employ stringent inclusion/exclusion criteria, yet animal studies have routinely included the entire exposed population as the study population. We examined the effect of grouping stressed rats according to the magnitude of their response on the statistical analysis of behavioral models. Exposure to a predator stimulus was used as the stress paradigm. Response magnitude was assessed in two consecutive behavioral tests measuring anxiety- and stress-related behaviors and was used to divide the animals into groups. The two extremes were studied, that is, those clearly "maladapted" and those clearly "well-adapted," using arbitrarily selected severity-measures, the "cut-off behavioral criteria" (CBC). Data for the partially affected middle group were discarded for reasons of clarity. The hypothalamic-pituitary-adrenal axis and heart rate variability were analyzed for the entire exposed population and then reexamined according to the CBCs. When the CBCs were applied, we found PTSD-like symptoms in only 22.0% of exposed rats. Compared to controls and to well-adapted exposed rats, the behaviorally maladapted rats displayed disordered physiological measures. They had significantly higher plasma corticosterone and ACTH levels, increased sympathetic activity, diminished vagal tone, and increased sympathovagal balance. These differences surfaced only when data were analyzed according to the CBCs. Animals respond to stress heterogeneously, resembling humans. Overlooking heterogeneity in responses obscures the results of biobehavioral data analysis. We submit that animals exposed to trauma should be divided into groups according to the magnitude of their response and be studied accordingly.