The hippocampus is a brain area involved in memory that animal studies have shown to be sensitive to the effects of stress. Meanwhile, studies in adults with post-traumatic stress syndrome (PTSD) related to either early childhood abuse or adult traumas like combat have found smaller hippocampal volume. Two recent meta-analyses of the literature on PTSD and hippocampal volume (Kitayama and colleagues 2005, which I co-authored, and one by Michael Smith, also in 2005) found smaller volume for both the left and the right hippocampi in both men and women with chronic PTSD. But does stress cause the hippocampi to shrink, or do smaller hippocampi simply make people more vulnerable to stress? A number of factors have been implicated as possible mechanisms for stress-related reductions in hippocampal volume, including increases in cortisol, increased levels of glutamate (a neurotransmitter) and decreased levels of brain-derived neurotrophic factor (BDNF). But so far no one has been able to determine definitively whether stress shrinks the hippocampus or simply preys more heavily on people that already have small hippocampi. Asking the question sooner One way to answer this question is to study children who have recently undergone stress and see if the stress seems to affect their hippocampi. The few studies published to date, however, have not even found smaller hippocampal volume in children with PTSD. A study published last March in Pediatrics, however, examined the hippocampi of 15 children (six boys and nine girls) age 8 to 14 with PTSD related to various kinds of physical, emotional and sexual abuse and neglect and got results a bit more meaningful. The study was conducted by a team of University of New Orleans and Stanford Unversity researchers led by Stanford psychiatrist Allan Reiss. Diagnosis of PTSD was established with the Clinician Administered PTSD Scale for Children and Adolescents (CAPS-CA), a common diagnostic tool. The researchers then took two important measurements: They measured salivary cortisol, elevations of which are associated with PTSD, and took magnetic resonance imaging (MRI) scans of the children's brains. They did this testing twice -- once soon after diagnosis, and once about 12 to 18 months later. They found a significant negative correlation between cortisol levels and other PTSD symptom severity at baseline and changes in right hippocampal volume over time. In other words, patients with the highest baseline cortisol and most severe PTSD symptoms had the greatest decreases in hippocampal volume over time. Catching the change This finding is significant because it addresses the puzzle about why studies had shown that adults with histories of early childhood abuse had smaller hippocampal volume, whereat the few studies performed in children (including an earlier study by Reiss and his collaborators) did not show any difference from controls. Studies in animals have shown that early stress may not necessarily cause an immediate effect on the hippocampus, but rather a delayed response leading to changes in hippocampal structure during development. Meanwhile, a study of twins led by Mark Gilbertson suggests a genetic contribution to smaller hippocampal volume in patients with PTSD; in twin pairs in which one twin had PTSD and the other did not, both would have small but similarly small hippocampi, suggesting that the smaller hippocampal volume was a risk factor for PTSD rather than a result of stress. Take all this together -- the Reiss paper, the animal literature showing hippocampal damage as a result of stress, and the twin study showing small hippocampi as risk factor for PTSD rather than result -- and it seems likely that there is a combined genetic and environmental contribution to smaller hippocampal volume in PTSD. A small hippocampus makes you more vulnerable to stress's ill effects, including damage to the hippocampus. Particularly intriguing is the correlation the Reiss study found between high cortisol and progressive reductions in hippocampal volume over time. As the authors note, one possible explanation for this finding is that high cortisol has a toxic effect on hippocampal neurons. Although there is debate on this topic, several studies in animals have reported this effect. Another possibility is that elevated cortisol is a marker of severity of PTSD status rather than a cause, and individuals with the most severe PTSD are also those with the highest baseline cortisol and the greatest propensity to develop hippocampal volume reductions over time. It isn't clear how these two possibilities could be disentangled in research studies involving clinical populations. A letter to the editor of Pediatrics, the journal in which Reiss's piece appeared, called "Stressing about posttraumatic stress disorder," various methodological issues were raised, such as accuracy of measurement of the hippocampus, small sample size, use of multiple statistical tests and a prior longitudinal study in children with PTSD that did not show changes in hippocampal volume over time. However, as the authors pointed out in their reply, MRI measurement of hippocampal volume can be highly accurate, and small sample sizes are not a reason to discount the results of an otherwise solid study. In fact, if statistically significant results can be found with a smaller sample size, that may in fact increase the strength of the findings. Overall, this study provides important information regarding the effects of stress on the hippocampus, strengthens evidence that small hippocampal size may be both risk factor and result of PTSD -- and points to the value of more future studies that follow patients with PTSD longitudinally. J. Douglas Bremner, M.D., is professor of psychiatry and radiology and director of the Clinical Neuroscience Research Unit at Emory University Medical School in Atlanta, Georgia.