Individuals with posttraumatic stress disorder (PTSD) have disturbances of both of the major stress-response systems in the body: the hypothalamic-pituitary-adrenal (HPA) axis and the sympathoadrenomedullary system. Both of these stress-response systems have well-described effects on immune function, suggesting that regulation of the immune system may be disturbed in individuals with PTSD.

Most, but not all, studies of adults with longstanding PTSD have found low basal levels of urinary or plasma cortisol and greater baseline sympathoadrenomedullary activity (1). In addition, PTSD has been associated with greater HPA axis and sympathoadrenomedullary reactivity to stress (1, 2). There are at least two ways that these disturbances of stress-response systems in individuals with PTSD could exacerbate cell-mediated immune system reactivity. First, low basal levels of glucocorticoids could predispose subjects with PTSD to enhanced immune activation. Chronic elevations of glucocorticoid levels globally suppress immune system reactivity (3, 4), and conversely, a number of chronic inflammatory disorders have been associated with lower basal HPA axis activity (5). Second, enhanced sympathoadrenomedullary and HPA axis reactivity to stress could also enhance cell-mediated immune function since there is preclinical evidence that acute stress and acute administration of catecholamine and glucocorticoid hormones can enhance cell-mediated immunity (4).

This study was designed to determine whether delayed-type hypersensitivity, an integrated, in vivo skin test measure of cell-mediated immunity, is enhanced in subjects with chronic PTSD and whether the delayed-type hypersensitivity response is related to basal cortisol levels. Because the reaction develops over several days in vivo, it provides an opportunity to integrate the influences of basal activity of both the HPA axis and the autonomic nervous system, as well as the influence of periodic activation of these systems by environmental stressors.

Subjects with PTSD were recruited from a group of individuals who were evaluated for participation in a psychotherapy treatment study for women with PTSD due to childhood physical and/or sexual abuse. Diagnosis of PTSD was determined with the Structured Clinical Interview for DSM-IV (SCID) and the Clinician-Administered PTSD Scale (6). Nine subjects with PTSD had comorbid major depression, seven had comorbid panic disorder, and nine had comorbid social phobia. The subjects with PTSD had a mean score of 74 (SD=16) on the Clinician-Administered PTSD Scale, indicating moderate to severe symptom severity.

All of 15 comparison subjects were free of current or past psychiatric disorders, as determined by the SCID, and reported no history of physical or sexual abuse or other major trauma. The PTSD and normal comparison groups were matched for age (mean=34 years, SD=8, versus mean=35 years, SD=9, respectively), race, time of day of skin testing, and menstrual cycle phase or menopausal status.

Exclusion criteria included regular use of alcohol or tobacco, current medical illness and history of psoriasis, atopic dermatitis, or any autoimmune disease. Three subjects with PTSD were taking psychotropic medications at the time of the study; two were using venlafaxine, and one was taking fluoxetine. All subjects gave written informed consent before participation.

Three recall antigens were given to the subjects intradermally (0.1 ml) at three sites on the flexor surface of one forearm. Candida and trichophyton (both at 1000 pnu/ml, Bayer Corp, Elkhart, Ind.) and tetanus toxoid (1 lfu/ml, Connaught Laboratories, Stillwater, Pa.) were used because there is a high probability that all subjects had previous exposure to these antigens. Forty-eight hours later, the size of the erythema response to each skin test was scored as the mean of its diameters in two perpendicular directions, and the three scores were averaged to produce a single delayed-type hypersensitivity measure.

Blood samples for serum cortisol measurements were drawn immediately before placement of the skin test. A circadian cortisol profile was determined on the day before the skin testing by collection of six salivary cortisol samples every 3 hours, from 8:00 a.m. to 11:00 p.m.

Two-tailed t tests were used to compare single time point variables between groups of subjects. Repeated-measures analysis of variance was used to examine salivary cortisol measures. Data are presented as means and standard deviations.

In relation to the comparison subjects, the subjects with PTSD had significantly higher mean delayed-type hypersensitivity reactions: mean=5.9 mm (SD=4.4) versus mean=2.9 mm (SD=3.0) (t=2.1, df=29, p<0.04). Responses to specific antigens were as follows, respectively: candida: mean=6.8 mm (SD=4.6) versus mean=4.4 mm (SD=6.6), tetanus: mean=5.2 mm (SD=5.9) versus mean=1.6 mm (SD=3.4), and trichophyton: mean=5.7 mm (SD=6.2) versus mean=2.7 mm (SD=5.7). Circadian measures of salivary-free cortisol levels did not differ between PTSD patients and comparison subjects (main effect: F=0.0, df=1, 27, p=0.97; interaction: F=0.4, df=5, 135, p=0.86). Plasma cortisol levels tended to be lower in the subjects with PTSD than in the comparison subjects: mean=217 nmol/liter (SD=60), versus mean=293 nmol/liter (SD=171) (t=1.7, df=29, p=0.11).

The power to detect the effect of comorbid diagnoses on delayed-type hypersensitivity and cortisol measures was limited by the group size. Delayed-type hypersensitivity reactions tended to be smaller in size by almost half in the PTSD subjects with major depression (N=9) than in those without comorbid major depression (N=7): mean=4.2 mm (SD=3.7) versus mean=8.0 mm (SD=4.5) (t=1.9, df=14, p<0.08). There was no evidence of a difference in mean salivary cortisol or plasma cortisol levels in the patients with PTSD with major depression and in those without major depression: saliva: mean=14.3 nmol/liter (SD=6.5) versus mean=19.7 nmol/liter (SD=5.6) (t=0.9, df=13, p=0.38), plasma: mean=210 nmol/liter (SD=62) versus mean=226 nmol/liter (SD=60) (t=0.5, df=14, p=0.60).

When the three subjects with PTSD who were taking antidepressants were compared to the remaining subjects with PTSD (N=13), the subjects taking antidepressants appeared to have a smaller mean delayed-type hypersensitivity size (mean=4.2 mm, SD=2.3, versus mean=6.3 mm, SD=4.7, respectively), indicating that the enhanced delayed-type hypersensitivity response in the entire PTSD group was not related to antidepressant use.

This study demonstrated enhanced delayed-type hypersensitivity reactions in women with PTSD related to childhood sexual and physical abuse. This finding is consistent with one previous report of enhanced delayed-type hypersensitivity in a group of male Australian Vietnam war veterans with long-standing PTSD compared to veterans who were not exposed to combat (7). Enhanced delayed-type hypersensitivity reactions in subjects with PTSD is also compatible with reports of higher plasma levels of pro-inflammatory cytokines (8, 9) and activated T lymphocytes (10) in individuals with PTSD.

Enhanced delayed-type hypersensitivity in subjects with long-standing PTSD contrasts with reports of smaller delayed-type hypersensitivity in subjects with melancholic depression (11), a syndrome characterized by hypercortisolemia. Although enhanced delayed-type hypersensitivity responses in the PTSD group were not clearly associated with lower salivary or plasma cortisol levels, there tended to be smaller plasma cortisol levels in the PTSD group. It is possible that a single midday plasma cortisol measurement and 1 day of circadian salivary cortisol measurements were not sensitive enough to detect basal hypocortisolism in the PTSD group. On the other hand, hypocortisolism is a less consistent finding in PTSD studies that include premenopausal women (1).

In summary, women with long-standing PTSD demonstrated enhanced delayed-type hypersensitivity, a physiological in vivo measure of cell-mediated immunity. Greater inflammatory responses may contribute to a variety of medical illnesses that have been associated with PTSD (12).

Received Nov. 5, 2001; revised Feb. 25, 2003; accepted March 17, 2003. From the Department of Psychiatry, Weill Medical College, Cornell University; New York University Child Study Center, New York; and the Department of Molecular Virology, Immunology, and Medical Genetics, College of Medicine, and the Department of Oral Biology, College of Dentistry, Ohio State University, Columbus, Ohio. Address reprint requests to Dr. Altemus, Department of Psychiatry, Weill Medical College, 1300 York Ave., Box 244, New York, NY 10021; maltemus@med.cornell.edu (e-mail). Funded in part by a grant from the Dewitt Wallace Reader’s Digest Fund to Dr. Altemus, an NIMH grant (MH-57883) to Dr. Cloitre, and a research grant from the John D. and Catherine T. MacArthur Foundation to Dr. Dhabhar. The authors thank Jessica Zonana for help with manuscript preparation, Ruirong Yang for help with cortisol assays, Karen Heffernan, Lisa Cohen, and Karestan Koenan for help with clinical characterization of PTSD subjects; and Bruce McEwen for advice and support.