You are on page 1of 8

Glucocorticoid Feedback Sensitivity and Adrenocortical Responsiveness in Posttraumatic Stress Disorder

Evan D. Kanter, Charles W. Wilkinson, Allen D. Radant, Eric C. Petrie, Dorcas J. Dobie, Miles E. McFall, Elaine R. Peskind, and Murray A. Raskind
Background: Decreased basal cortisol levels have been reported in individuals with posttraumatic stress disorder (PTSD). There is evidence for enhanced negative feedback sensitivity of the hypothalamic-pituitary-adrenal (HPA) axis in PTSD, which could account for this, but other possible mechanisms have not been ruled out. We examined the HPA axis employing a metyrapone-cortisol infusion protocol designed to study negative feedback sensitivity. Methods: Vietnam combat trauma exposed subjects met DSM-IV criteria for PTSD. Exclusion criteria included substance abuse and most medications. Endogenous feedback inhibition was removed by blocking cortisol synthesis with oral metyrapone and reintroduced by intravenous infusion of cortisol. In a placebo condition, subjects received oral placebo and normal saline infusion. Serial blood samples drawn over 4 hours were assayed for adrenocorticotrophic hormone (ACTH), cortisol, and 11deoxycortisol. Selected samples were assayed for cortisol binding globulin (CBG) and dehydroepiandrosterone (DHEA). Results: Basal plasma cortisol was significantly decreased in PTSD subjects (n 13) compared with control subjects (n 16). No significant difference in the ACTH response to cortisol infusion following metyrapone was observed; however 11-deoxycortisol was significantly decreased in PTSD subjects. In addition, CBG was significantly increased in PTSD subjects, and DHEA was significantly decreased in both PTSD and combat-exposed control subjects. Conclusions: These observations suggest decreased adrenocortical responsiveness may be an additional or alternative mechanism accounting for low cortisol in PTSD. Biol Psychiatry 2001;50:238 245 2001 Society of Biological Psychiatry Key Words: PTSD, adrenal cortex, ACTH, cortisol, metyrapone, DHEA

Introduction
osttraumatic stress disorder (PTSD) is associated with abnormalities in neuroendocrine systems (Yehuda 1998), including alterations in the hypothalamic-pituitaryadrenal (HPA) axis. Given the long-established role of the HPA axis in the acute response to stress, it is not surprising that HPA axis abnormalities appear to be present in PTSD. The nature and direction of these abnormalities are counterintuitive, however. The increase in plasma cortisol seen in response to acute stress is well documented (Chrousos and Gold 1992; Selye 1956), and increased basal cortisol levels are consistently found in patients with major depression (Sachar 1975). In contrast, most studies have shown decreased basal cortisol levels in PTSD (for review, see Yehuda 1997). Intensity of combat exposure is correlated with PTSD symptom severity (Buydens-Branchey et al 1990) and negatively correlated with plasma cortisol concentration (Boscarino 1996), suggesting a pathophysiologic relationship between HPA axis function and PTSD symptoms. One possible mechanism that could account for low cortisol concentrations in PTSD is increased sensitivity of the HPA axis to feedback inhibition by cortisol (Yehuda 1997). Although patients with major depression often fail to suppress cortisol concentrations following dexamethasone (DEX), patients with PTSD hypersuppress cortisol concentrations in response to DEX (Goenjian et al 1996; Heim et al 1998; Stein et al 1997; Yehuda et al 1993). Demonstration of a decrease in the number of glucocorticoid receptors on peripheral lymphocytes in response to DEX in combat veterans with PTSD, but not in control subjects (Yehuda et al 1995), provides additional evidence of an enhanced sensitivity to cortisol in PTSD. This is again opposite of the pattern seen in major depression, in which decreased sensitivity of lymphocyte glucocorticoid receptors is observed (Gormley et al 1985; Whalley et al 1986).
0006-3223/01/$20.00 PII S0006-3223(01)01158-1

From the Mental Illness Research, Education, and Clinical Center (EDK, DJD, MEM, ERP, MAR) and Geriatric Research, Education, and Clinical Center (CWW), Veterans Affairs Puget Sound Health Care System; and the Department of Psychiatry and Behavioral Sciences, University of Washington School of Medicine (EDK, CWW, ADR, ECP, DJD, MEM, ERP, MAR), Seattle, Washington. Address reprint requests to Dr. Evan Kanter, VA Puget Sound Health Care System, Mental Illness Research, Education, and Clinical Center (116 MIRECC), 1660 South Columbian Way, Seattle, WA 98108. Received September 29, 2000; revised March 19, 2001; accepted March 22, 2001.

2001 Society of Biological Psychiatry

HPA Axis in PTSD

BIOL PSYCHIATRY 2001;50:238 245

239

Yehuda et al (1996) have used the metyrapone stimulation test as a means to examine sensitivity to removal of glucocorticoid negative feedback of the HPA axis in PTSD subjects. The 11 -hydroxylase inhibitor metyrapone blocks the final step in cortisol biosynthesis, decreasing plasma cortisol concentrations and increasing concentrations of the precursor compound 11-deoxycortisol, which does not have feedback activity. Plasma adrenocorticotrophic hormone (ACTH) increases as a result of decreased feedback inhibition at the hypothalamus and pituitary by cortisol. In Yehudas study, combat veterans with PTSD exhibited a greater increase in ACTH release from the pituitary and a greater accumulation of the cortisol precursor 11-deoxycortisol in response to metyrapone than did normal control subjects, suggesting increased feedback inhibition by cortisol in PTSD. The evidence to date, including the DEX suppression test and the metyrapone stimulation test, supports a model of enhanced glucocorticoid negative feedback in PTSD. Other possible mechanisms might contribute to the low plasma cortisol seen in PTSD, however, and have not been ruled out. These include decreased sensitivity of the pituitary or the adrenal as well as decreased concentrations of corticosteroid binding globulin (CBG). We sought to examine negative feedback sensitivity in PTSD directly using the endogenous ligand cortisol. We used a protocol of cortisol infusion following pretreatment with metyrapone, originally developed to examine the effect of human aging on feedback inhibition of the HPA axis by cortisol (Wilkinson et al 1997, 2001). Following suppression of endogenous cortisol by oral metyrapone, cortisol was infused intravenously. Serial measurements were made of plasma ACTH, cortisol, and 11-deoxycortisol. The feedback effect of the exogenous cortisol could be determined by measuring the decrease in circulating ACTH concentrations. We hypothesized that circulating ACTH would decrease more rapidly in response to cortisol infusion in PTSD subjects than in normal control subjects, a finding that would be consistent with enhanced negative feedback sensitivity to cortisol in PTSD. Considering other hypotheses to account for hypocortisolemia in PTSD, we assayed samples for CBG. We also measured another corticosteroid, dehydroepiandrosterone (DHEA) as an additional probe of adrenal function.

System, Seattle Division. Subjects were in good general health as determined by history, physical exam, and laboratory studies. Complete blood count, electrolytes, glucose, liver, kidney, and thyroid function tests, and urinalysis were all within normal limits. Urine toxicology screens were negative. Subjects had no substance abuse for at least 3 months and were free of all medications known to affect the HPA axis for at least 30 days, including antidepressants and other psychotropic medications. Diagnosis of PTSD was established using the Clinician Administered PTSD Scale (CAPS; Blake et al 1995). The Structured Clinical Interview for DSM-IV (SCID; First et al 1995) was used to assess other Axis I disorders. PTSD subjects also completed the Revised Combat Exposure Scale (RCS; Gallops et al 1981), Mississippi Scale for Combat-Related PTSDRevised (MISS; Keane et al 1988), PTSD Checklist (PCL; Weathers et al 1993), and Hamilton Depression Rating Scale (HAM-D; Hamilton 1960). An age-comparable group of control subjects without PTSD (mean age 50.2 0.8) were community volunteers recruited by newspaper advertisement, advertisement to VA employees, and word of mouth. Nine of these were veterans who had histories of combat exposure. Control subjects completed a nonpatient version of the SCID, the PCL, and the HAM-D. Combat-exposed control subjects also completed the RCS. The PTSD subjects had a mean score on the CAPS (maximum 136) of 84.5 4.0 and the MISS (maximum 195) of 121.4 3.4, reflecting severe PTSD symptoms. Mean scores on the 14-point RCS were 11.5 0.7 for PTSD subjects and 8.4 1.1 for combat-exposed control subjects. Mean scores on the PCL (range 17 85) were 69.6 2.1 for PTSD subjects, 20.7 1.7 for combat-exposed control subjects, and 20.4 2.0 for noncombat control subjects. Mean scores on the HAM-D were 18.5 1.8 for PTSD subjects, 1.0 0.4 for combat-exposed control subjects, and 1.7 0.9 for noncombat control subjects. Of the 13 PTSD subjects, 11 also met SCID criteria for major depression or dysthymia. No other Axis I disorders were diagnosed in the PTSD subjects, and none were diagnosed in the control subjects. Six of 13 PTSD subjects were unemployed at the time of the study; two of 16 control subjects were unemployed. Body mass index was 27.6 0.9 for PTSD subjects and 28.1 1.4 for control subjects.

Procedures
Each subject underwent two study conditions in random order at least 4 days apart. In the experimental condition, subjects were administered 750 mg of metyrapone (Metopirone, Ciba-Geigy) orally at 0600 and again at 0900. Subjects were instructed to take nothing by mouth after midnight the night before the study. Except for two PTSD subjects who completed the study as inpatients, all subjects drove to the hospital. They were instructed to awaken at 0600, take the first metyrapone dose, and then call the voicemail of the study nurse to confirm that the medication was taken at the appropriate time and that they were on their way to the hospital. At 0830, an intravenous catheter was placed in an antecubital vein of each arm and kept patent with a slow infusion of normal saline. One intravenous catheter was used for blood sampling and the other for infusion. At 0855 and 0900 baseline

Methods and Materials


Subjects
The study was approved by the Human Subjects Review Committee of the University of Washington, and written informed consent was obtained from all subjects. Thirteen Vietnam combat veterans (mean age 50.5 0.6) were recruited from inpatient and outpatient PTSD programs at the VA Puget Sound Health Care

240

BIOL PSYCHIATRY 2001;50:238 245

E.D. Kanter et al

blood samples were drawn. Beginning at 0900, cortisol (hydrocortisone sodium succinate, Pharmacia, Peapack, NJ) was infused at 0.03 mg/kg/hour in 300 mL normal saline over 150 min. Blood sampling occurred every 15 min for 240 min following baseline. In the placebo condition, subjects received placebo tablets and a normal saline infusion matched for volume.

Assays
We measured ACTH by a double antibody sandwich radioimmunometric assay (Nichols Institute Diagnostics, San Juan Capistrano, CA) following a modified version of the commercial protocol. In this assay, cross-reactivity of shorter forms of ACTH and melanotropins is negligible. The detection limit is 2 pg/mL. Intra- and interassay coefficients of variation are 6.3% and 12.8%, respectively. Cortisol was measured by radioimmunoassay (RIA) in unextracted plasma. Samples were diluted with phosphate buffer and heated for 20 min at 80C to denature binding globulins. Cortisol antiserum was obtained from ICN Biomedicals (Costa Mesa, CA). The detection limit for cortisol is 0.5 g/dL. Intra- and interassay coefficients of variation are 4.6% and 10.2%, respectively. Cross-reactivity of the antiserum with 11-deoxycortisol is 10%. Assay values for cortisol were corrected by subtracting 10% of the 11-deoxycortisol values for each sample. 11-deoxycortisol was measured by RIA in unextracted plasma to which steroid binding globulin inhibitor was added. Antiserum, standards, and 125I-labeled 11-deoxycortisol were obtained from ICN Biomedicals. Cross-reactivity of the antiserum with cortisol is 1.7%. Diluted aliquots of 5 L of plasma were used in assaying samples from metyrapone-treated subjects and 20 L samples from control subjects. The sensitivity of the assay is 0.5 ng/mL. Intra- and interassay coefficients of variation are 5.2% and 14.4%, respectively. We measured CBG by radial immunodiffusion with Cortisol Binding Globulin-NL Kits from The Binding Site (San Diego, CA). Duplicate 5- L plasma samples and standards were added to individual wells in agarose gel plates containing a monospecific antibody (sheep antihuman corticosteroid binding globulin). Plates were incubated under moist conditions at room temperature for 96 hours, and the diameter of the precipitin rings was read with an electronic reader. Minimum sensitivity of the assay is 0.1 mol/L. Intra- and interassay coefficients of variation are 2.9% and 4.4%, respectively. Free cortisol was calculated from assayed values of total plasma cortisol and CBG using the method of Coolens et al (1987). Plasma concentrations of DHEA were measured with the DSL-8900 DHEA Radioimmunoassay Kit from Diagnostic Systems Laboratories (Webster, TX) following the manufacturers protocol. Cross-reactivity of the antiserum with DHEA-sulfate is 0.02%, and cross-reactivities with other steroids are all less than 1%. Sensitivity of the assay is 0.03 ng/mL. All samples in this study were measured in a single assay, and the intraassay coefficient of variation was 2.8%.

analysis of variance (ANOVA) for repeated measures. Differences in hormonal baseline values between groups were compared with two-tailed, unpaired t tests. Determination of the time at which ACTH was significantly inhibited by cortisol feedback was made by calculating the difference in ACTH concentration between the postmetyrapone baseline and each subsequent time point and analyzing the differences at each time by single sample t tests. We compared DHEA concentrations in the three treatment subgroups using one-way ANOVA followed by Fishers protected least significant difference (PLSD) post hoc test for paired comparisons. Area under the curve (AUC) for plasma free cortisol was calculated with the trapezoidal method and group differences in AUC were analyzed with two-tailed, unpaired t tests.

Results
Because the ACTH, cortisol, and 11-deoxycortisol responses and CBG concentrations of the combat-exposed and noncombat control subjects did not significantly differ in either placebo or drug condition (all t 1.0, p .3), the two groups were combined for analysis of these variables. All results are presented as means standard error of the mean (SEM).

Placebo Condition Measurements


In the placebo condition, plasma cortisol (Figure 1B) was significantly lower in PTSD subjects compared with control subjects throughout the protocol [F(1, 27) 12.6, p .005]. During placebo infusion ACTH (Figure 1A) and 11-deoxycortisol (Figure 1C) concentrations did not change over time and did not differ between PTSD and control groups. Concentrations of 11-deoxycortisol were minimal in both groups.

Pituitary and Adrenal Responses to Metyrapone


Metyrapone treatment substantially reduced initial plasma cortisol (Figure 1B) concentrations and increased initial ACTH (Figure 1A) and 11-deoxycortisol (Figure 1C) concentrations. Metyrapone treatment induced significantly greater decreases in cortisol concentrations (t 2.3, p .05) and significantly greater increases in 11-deoxycortisol concentrations (t 2.4, p .05) in control subjects compared with PTSD subjects. The magnitude of the increase in ACTH concentrations did not differ significantly between PTSD and control groups (t 1.7, p .10).

Statistical Analysis
Hormonal responses of PTSD and control groups during the course of experimental sessions were analyzed by two-way

Plasma ACTH, Cortisol, and 11-Deoxycortisol Responses to Exogenous Cortisol Infusion


There was no significant difference in ACTH concentration (Figure 1A) between PTSD subjects and control

HPA Axis in PTSD

BIOL PSYCHIATRY 2001;50:238 245

241

Figure 2. Plasma cortisol binding globulin (CBG) concentrations in subjects with posttraumatic stress disorder (n 13) and control subjects (n 16). The first sample in the placebo condition and the first and last samples in the metyraponecortisol (met-cort) condition were assayed for CBG. Vertical lines indicate SEM. Concentrations of CBG were significantly increased in PTSD subjects compared with control subjects [F(1, 27) 8.6, p .01].

Figure 1. Hypothalamic-pituitary-adrenal axis responses to metyrapone-cortisol protocol in subjects with posttraumatic stress disorder (PTSD; filled circles, n 13) and control subjects (open circles, n 16). Graphs in right-hand column represent responses to infusion of cortisol at 0.03 mg/kg/hour following oral administration of 1500 mg metyrapone in divided doses. Graphs in left-hand column represent placebo responses. Vertical lines indicate SEM. (A) Adrenocorticotrophic hormone (ACTH): Differences between PTSD subjects and control subjects in the effect of exogenous cortisol on the ACTH response were not significant [F(1, 27) 3.0, p .10]. (B) Cortisol: Drug condition represents accumulation of cortisol from exogenous infusion. Placebo condition demonstrates significantly decreased baseline plasma cortisol in PTSD subjects compared with control subjects [F(1, 27) 12.6, p .005]. (C) 11-deoxycortisol: Plasma concentrations of 11-deoxycortisol were significantly lower in PTSD subjects than in control subjects [F(1, 27) 10.7, p .005].

difference in plasma cortisol between PTSD and control subjects over the course of the experiment [F(1, 27) 0.95, p .34]. The maximum levels of plasma cortisol reached as a result of the infusion did not differ significantly between PTSD and control groups (t 1.7, p .10). Plasma levels of 11-deoxycortisol (Figure 1C) were significantly lower in PTSD subjects compared with control subjects throughout the protocol [F(1, 27) 10.7, p .005].

Measurement of Plasma CBG and Calculated Free Cortisol


Plasma CBG concentrations (Figure 2) were measured in the first sample of the placebo condition and the first and last samples from the metyrapone-cortisol condition. We found that CBG was significantly increased in PTSD subjects compared with control subjects [F(1, 27) 8.6,

subjects [F(1, 27) 3.0, p .10] and no significant group by time interaction. The ACTH concentration never declined significantly in either the PTSD or control group by single sample t tests of the difference scores at each time point. During the 150 min of exogenous cortisol infusion, plasma cortisol levels (Figure 1B) steadily increased to approximately twice the baseline level. After the infusion was stopped, cortisol levels decreased over 90 min, nearly returning to baseline levels. There was no significant

Figure 3. Calculated plasma free cortisol concentrations in subjects with posttraumatic stress disorder (PTSD; filled circles, n 13) and control subjects (open circles, n 16). Vertical lines indicate SEM. In the placebo condition (left) free cortisol concentrations were significantly decreased in PTSD subjects compared with control subjects [F(1,27) 20.4, p .0001]. In the metyrapone-cortisol condition (right) the difference in free cortisol concentrations between PTSD subjects and control subjects did not reach significance [F(1, 27) 4.0, p .055].

242

BIOL PSYCHIATRY 2001;50:238 245

E.D. Kanter et al

Figure 4. Plasma dehydroepiandrosterone (DHEA) concentrations in PTSD subjects (n 13), combat control subjects (n 9), and noncombat control subjects (n 7). The first two samples in both placebo and metyrapone-cortisol (met-cort) conditions were assayed for DHEA. Vertical lines indicate SEM. Concentrations of DHEA were significantly decreased in both PTSD subjects and combat control subjects compared with noncombat control subjects in both placebo and drug conditions (all p .05). Concentrations of DHEA did not significantly differ between PTSD subjects and combat control subjects.

p .01]. In the placebo condition, calculated free cortisol (Figure 3) was significantly lower in PTSD subjects compared with control subjects throughout the protocol [F(1, 27) 20.4, p .0001]. In the metyrapone-cortisol condition, there was a trend for free cortisol to be decreased in the PTSD group that did not reach the level of significance [F(1, 27) 4.0, p .055]. There was no significant difference between the groups in an area under the curve analysis (t 1.4, p .16).

Measurement of Plasma DHEA


Plasma DHEA concentrations (Figure 4) were measured in the first and second samples of each condition (preinfusion baseline). Unlike other parameters, DHEA concentrations differed significantly between the combat and noncombat control groups in both placebo and metyrapone-cortisol conditions (placebo t 2.2, p .05; drug t 3.4, p .005). The two control groups were treated separately for purposes of analyzing DHEA data. Differences in DHEA concentrations among groups were significant by ANOVA [F(2, 24) 10.4, p .001]. Post hoc comparisons using Fishers PLSD showed that in both placebo and metyrapone-cortisol conditions DHEA concentrations were not significantly different between PTSD subjects and combat control subjects; DHEA concentrations were significantly decreased in both PTSD subjects and combat control subjects compared with noncombat control subjects (all p .05).

Discussion
There were three main findings in the study. First, baseline plasma cortisol levels were significantly lower in PTSD subjects than control subjects, consistent with previous

studies. Second, no evidence of increased sensitivity of glucocorticoid negative feedback was observed in this protocol. Third, a variety of data suggested that adrenocortical responsiveness was decreased in subjects with PTSD. Additional findings were increased concentrations of CBG in subjects with PTSD and decreased concentrations of DHEA in subjects with combat exposure. One possible explanation for decreased plasma cortisol concentrations would be decreased concentrations of CBG. Because 95% of plasma cortisol is protein bound (Dunn et al 1981), a relative deficiency of CBG in the PTSD subjects might result in a lower concentration of bound cortisol, which is inactive. Total cortisol would be decreased, but the concentration of physiologically active free cortisol might be the same. In our study, CBG was actually increased in the PTSD subjects compared with control subjects. Consequently, differences between PTSD subjects and control subjects appear greater for free cortisol than for total cortisol. The finding of increased CBG in the hypocortisolemic PTSD subjects is consistent with observations in the rat that removal of corticosterone by adrenalectomy increases the production and secretion of CBG (Feldman et al 1979; Levin et al 1987). No significant difference was seen by ANOVA between PTSD subjects and control subjects in the effect of exogenous cortisol on circulating ACTH. These data therefore do not provide support for enhanced glucocorticoid negative feedback of the HPA axis. It is possible that a significant difference in ACTH suppression would be seen using a higher cortisol infusion dose. Given the a priori hypothesis of enhanced negative feedback in PTSD patients, the 0.03 mg/kg/hour dose was chosen because it would produce only a modest suppression of ACTH and might thereby accentuate the difference between PTSD and control subjects. A 0.06 mg/kg/hour dose has produced a more robust suppression of ACTH in normal subjects in previous studies in our laboratory (Wilkinson et al 1997, 2001). Repeating the experiment using the higher dose might produce more conclusive results. Another possible explanation for the lack of the expected ACTH response to cortisol is that there may have been less of a feedback signal in the PTSD group due to the observed decrease in free cortisol. Yehuda (1997) has proposed that trauma sensitizes the HPA axis so that there is enhanced glucocorticoid feedback inhibition, thus explaining the reported low cortisol concentrations (Yehuda 1997), hypersuppression of cortisol following DEX (Goenjian et al 1996; Heim et al 1998; Stein et al 1997; Yehuda et al 1993), and increased ACTH response to metyrapone-induced removal of cortisol feedback (Yehuda et al 1996). It is possible, however, that decreased adrenal secretion of cortisol could also contribute to this pattern of findings. For example, Wittert et al

HPA Axis in PTSD

BIOL PSYCHIATRY 2001;50:238 245

243

(1996) found that intense physical training leads to decreased adrenal responsiveness to ACTH, consistent with an association between physical stress and decreased adrenal function. Furthermore, an analogy with the HPA axis characteristics of major depression suggests the possibility of adrenal hyporesponsiveness. Both decreased glucocorticoid negative feedback sensitivity (Young et al 1991) and increased adrenal sensitivity to ACTH (Jaeckle et al 1987) have been demonstrated in major depression. If a pattern of HPA axis abnormalities in PTSD is emerging that is opposite that seen in depression, it seems reasonable to hypothesize that decreased adrenal sensitivity would be observed in PTSD. If adrenal hyporesponsiveness were present, one might have expected to see increased ACTH levels in response to low cortisol in PTSD subjects. The absence of an increase in ACTH might be explained if increased negative feedback sensitivity were also present. The finding of significantly lower levels of plasma 11-deoxycortisol in PTSD subjects compared with control subjects lends further support to the possibility of decreased adrenocortical responsiveness as an additional or alternative mechanism to explain observations of low cortisol in PTSD. After metyrapone treatment, ACTH was not significantly different between PTSD subjects and control subjects, but the magnitude of the changes in both cortisol and 11-deoxycortisol was significantly decreased in the PTSD group, again providing evidence of decreased adrenal responsiveness in PTSD subjects. The observed difference in adrenal sensitivity could represent a differential response to ACTH, perhaps at the level of the ACTH receptor, or it could be a function of some other factor that acts synergistically with or in opposition to ACTH. An ACTH stimulation test would be a means to test directly this hypothesis. Although there has been no report of an ACTH challenge in the PTSD literature, there have been reports of CRF challenges in PTSD. Smith et al (1989) reported a blunted ACTH response to CRF in subjects with PTSD compared with control subjects, with normal levels of cortisol. This would be consistent with a model of enhanced negative feedback by cortisol. On the other hand, Heim et al (1998) reported different results in a population of women with chronic pelvic pain, many of whom had PTSD. In this study, ACTH levels were normal and cortisol levels were decreased compared with control subjects, a pattern that would be consistent with adrenal hyporesponsiveness. Plasma DHEA concentrations were decreased in PTSD subjects and in combat-exposed control subjects compared with control subjects without combat exposure. These differences were amplified when adrenal stimulation was increased by treatment with metyrapone, in a pattern similar to that observed for 11-deoxycortisol. Although the

data show an association between decreased DHEA and combat exposure, rather than PTSD, the findings are still consistent with adrenal hyporesponsiveness. A precursor to many other steroids, including testosterone and estrogen, DHEA is secreted in large concentrations from the adrenal cortex. It possesses weak androgenic activity; its intrinsic physiologic function is unknown (Kroboth et al 1999). Administration of DHEA has been reported to have positive effects on mood and memory (Wolkowitz et al 1997). The finding of decreased DHEA in PTSD subjects is intriguing because affective disturbance and difficulty with memory for nontraumatic events are symptoms commonly reported in the disorder. A potential concern about studies that use DEX to evaluate glucocorticoid feedback sensitivity is that DEX differs from endogenous glucocorticoids in significant ways. First, DEX and the human endogenous glucocorticoid cortisol have different affinities for corticosteroid receptor subtypes (DeKloet et al 1991). Second, the two ligands differ in the level of the HPA axis at which the feedback inhibition is recognized (Kovacs and Makara 1988; Miller et al 1992). Because DEX does not penetrate well into the brain and exerts feedback effects primarily at the pituitary, using it to test suprapituitary components of the HPA axis may be problematic (De Kloet et al 1998). A potentially significant difference between this study and that of Yehuda (1996) is the dose of metyrapone used. In Yehudas study, 2500 mg of metyrapone was given in a single dose. At this dose some of the subjects may have experienced nausea. Because nausea is one of the most powerful known stimulators of the HPA axis (Eversmann et al 1978), this could have been a confounding factor. In our study, metyrapone was given in two doses of 750 mg. With administration of metyrapone in this manner, no subjects in our study experienced any gastrointestinal side effects. As a result of the lower metyrapone dose, ACTH did not increase in response to metyrapone to the same degree as it did in the Yehuda study. Approximately 50% suppression of endogenous cortisol was achieved in normal subjects. This degree of suppression was sufficient to enable demonstration of differences in glucocorticoid feedback sensitivity between young and old subjects in previous studies (Wilkinson et al 1997, 2001). An additional consequence of the study design is that the full effect of the first metyrapone dose and the effect of the second dose on ACTH concentrations are masked by the cortisol infusion. The majority of the PTSD subjects in this study also met criteria for major depression or dysthymia. This is to be expected because signs and symptoms sufficient to meet criteria for a depressive disorder are frequently comorbid with PTSD (Brady et al 2000). In the National Comorbidity Survey, lifetime prevalence of major depression in

244

BIOL PSYCHIATRY 2001;50:238 245

E.D. Kanter et al

individuals with PTSD was 48% (Kessler et al 1995). Highly symptomatic, treatment-seeking patients with PTSD such as those in our sample may have even greater rates of comorbid depression. Chronicity of PTSD further increases comorbidity (Breslau et al 1991). The comorbid depression in this study may be unavoidable and does not appear to have substantially affected the results. If it had, one would expect to observe increased cortisol levels in the depressed PTSD subjects. The veterans in this study suffer from a severe form of chronic PTSD resulting from prolonged combat trauma exposure that occurred more than 30 years in the past. It is not clear to what extent these findings can be extrapolated to other groups with PTSD. This population is also notable for having a lifetime prevalence of alcohol or other drug dependence greater than 70% (Kulka et al 1990). Although no subjects had substance abuse within 3 months of the study, past substance use is an uncontrolled variable of unknown consequence. Because HPA axis activity is circadian, sleep is another uncontrolled variable with possible relevance. Sleep disorders are common in PTSD and the PTSD subjects would be expected to have had less sleep than control subjects, on average, the night before the study. Finally, employment status is uncontrolled; PTSD subjects were more likely to be unemployed than were control subjects. The data in this report provide further support for the observation of decreased cortisol concentrations in chronic PTSD and suggest that the mechanisms underlying this observation are complex. Our study does not rule out the possibility that increased sensitivity of the HPA axis to glucocorticoid feedback inhibition is one of the mechanisms, but such increased feedback sensitivity was not observed under the conditions of this protocol. These results do suggest that decreased responsiveness of the adrenal cortex contributes to the decreased cortisol concentrations observed in PTSD.
This work was supported by the Department of Veterans Affairs and a Thomas H. Holmes Research Scholar Award (EDK). We gratefully acknowledge the excellent technical assistance of Jill Perander, Hollie Holmes, Elizabeth Colasurdo, Robert Beckham III, Shannon Boldt, and Carl Sikkema.

References
Blake DD, Weathers FW, Nagy LM, Kaloupek DG, Gusman FD, Charney DS, et al (1995): The development of a clinicianadministered PTSD scale. J Trauma Stress 8:7590. Boscarino JA (1996): Posttraumatic stress disorder, exposure to combat, and lower plasma cortisol among Vietnam veterans: Findings and clinical implications. J Consult Clin Psychol 64:191201. Brady KT, Killeen TK, Brewerton T, Lucerini S (2000): Comor-

bidity of psychiatric disorders and posttraumatic stress disorder. J Clin Psychiatry 61(suppl 7):2232. Breslau N, Davis GC, Andreski P, Peterson E (1991): Traumatic events and posttraumatic stress disorder in an urban population of young adults. Arch Gen Psychiatry 48:216 222. Buydens-Branchey L, Noumair D, Branchey M (1990): Duration and intensity of combat exposure and posttraumatic stress disorder in Vietnam veterans. J Nerv Ment Dis 178:582587. Chrousos GP, Gold PW (1992): The concepts of stress and stress system disorders: Overview of physical and behavioral homeostasis. JAMA 267:1244 1252. Coolens J-L, Van Baelen H, Heyns W (1987): Clinical use of unbound cortisol as calculated from total cortisol and corticosteroid-binding globulin. J Steroid Biochem 26:197202. De Kloet ER (1991): Brain corticosteroid receptor balance and homeostatic control. Front Neuroendocrinol 12:95164. De Kloet ER, Vreugdenhil E, Oitzl MS, Joels M (1998): Brain corticosteroid receptor balance in health and disease. Endocr Rev 19:269 301. Dunn JF, Nisula BC, Rodbard D (1981): Transport of steroid hormones: Binding of endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma J Clin Endocrinol Metab 53:58 68. Eversmann T, Gottsman M, Ehlich E, Unlbrect G, Von Werder K, Scriba P (1978): Increased secretion of growth hormone, prolactin, antidiuretic hormone and cortisol induced by the stress of motion sickness. Aviat Space Environ Med 49:53 57. Feldman D, Mondon CE, Horner JA, Weiser JN (1979): Glucocorticoid and estrogen regulation of corticosteroid-binding globulin production by rat liver. Am J Physiol 237:E493 499. First, MB, Spitzer RL, Gibbon M, Williams JBW (1995): Structured Clinical Interview for DSM-IV Axis I Disorders. New York: New York State Psychiatric Institute, Biometrics Research Department. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56 62. Heim C, Ehlert U, Hanker JP, Hellhammer DH (1998): Abuserelated posttraumatic stress disorder and alterations of the hypothalamic-pituitary-adrenal axis in women with chronic pelvic pain. Psychosom Med 60:309 318. Gallops M, Laufer RS, Yager T (1981) Revised combat scale. In: Laufer RS, Yager T, Frey-Wouters E, Donnellan J, editors. Legacies of Vietnam: Comparative Adjustments of Veterans and their Peers, Vol 3. Washington, DC: U.S. Government Printing Office, 125. Goenjian AK, Yehuda R, Pynoos RS, Steinberg AM, Tashjian M, Yang RK, et al (1996): Basal cortisol, dexamethasone suppression of cortisol, and MHPG in adolescents after the 1988 earthquake in Armenia. Am J Psychiatry 153:929 934. Gormley GJ, Lowy MT, Reder AT, Hospelhorn VD, Antel JP, Meltzer HY (1985): Glucocorticoid receptors in depression: Relationship to the dexamethasone suppression test. Am J Psychiatry 142:1278 1284. Jaeckle RS, Kathol RG, Lopez JF, Meller WH, Krummel SJ (1987): Enhanced adrenal sensitivity to exogenous cosyntropin (ACTH124) stimulation in major depression. Arch Gen Psychiatry 44:233240. Keane TM, Caddell JM, Taylor KL (1988): Mississippi scale for

HPA Axis in PTSD

BIOL PSYCHIATRY 2001;50:238 245

245

combat related posttraumatic stress disorder: Three studies in reliability and validity. J Consulting Clin Psychology 56:85 90. Kessler RC, Sonnega A, Bromet EJ, Hughes M, Nelson CB (1995): Posttraumatic stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry 52:1048 1060. Kovacs KJ, Makara GB (1988): Corticosterone and dexamethasone act at different brain sites to inhibit adrenalectomyinduced adrenocorticotropin hypersecretion. Brain Res 474: 205210. Kroboth PD, Salek FS, Pittenger AL, Fabian TJ, Frye RF (1999): DHEA and DHEA-S: A review. J Clin Pharmacol 39:327 348. Kulka RA, Schlenger WE, Fairbank JA, Hough R (1990): Trauma and the Vietnam War Generation: Report of the Findings from the National Vietnam Veterans Readjustment Study. New York: Brunner/Mazel. Levin N:Akana SF, Jacobson L, Kuhn RW, Siiteri PK, Dallman MF (1987): Plasma adrenocorticotropin is more sensitive to transcortin production or thymus weight to inhibition by corticosterone in rats. Endocrinology 121:1104 1110. Miller AH, Spencer RL, Pulera M, Kang S, McEwen BS, Stein M (1992): Adrenal steroid receptor activation in rat brain and pituitary following dexamethasone: Implications for the dexamethasone suppression test. Biol Psychiatry 32:850 869. Sachar EJ (1975): Twenty-four-hour cortisol secretory patterns in depressed and manic patients. Prog Brain Res 42:8191. Selye H (1956): The Stress of Life. New York: McGraw-Hill. Smith MA, Davidson J, Ritchie JC, Kudler H, Lipper S, Chappell P, et al (1989): The corticotropin-releasing hormone test in patients with posttraumatic stress disorder. Biol Psychiatry 26:349 355. Stein MB, Yehuda R, Koverola C, Hanna C (1997): Enhanced dexamethasone suppression of plasma cortisol in adult women traumatized by childhood sexual abuse. Biol Psychiatry 42:680 686. Weathers FW, Litz BT, Herman DS, Huska JA, Keane TM (1993, October): The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility. Paper presented at the 9th annual meeting of the International Society for Traumatic Stress Studies, San Antonio, TX.

Whalley LJ, Borthwick N, Copolov D, Dick H, Christie JE, Fink G (1986): Glucocorticoid receptors and depression. BMJ 292:859 861. Wilkinson CW, Peskind ER, Raskind MA (1997): Decreased hypothalamic-pituitary-adrenal axis sensitivity to cortisol feedback inhibition in human aging. Neuroendocrinology 65:79 90. Wilkinson CW, Petrie EC, Murray SR, Colasurdo EA, Raskind MA, Peskind ER (2001): Human glucocorticoid feedback inhibition is reduced in older individuals: Evening study. J Clin Endocrinol Metab 86:545550. Wittert GA, Livesey JH, Espiner EA, Donald RA (1996): Adaptation of the hypothalamopituitary adrenal axis to chronic exercise stress in humans. Med Sci Sports Exerc 28:10151019. Wolkowitz OM, Reus VI, Roberts E, Manfredi F, Chan T, Raum WJ, et al (1997): Dehydroepiandrosterone (DHEA) treatment of depression. Biol Psychiatry 41:311318. Yehuda R (1997): Sensitization of the hypothalamic-pituitaryadrenal axis in posttraumatic stress disorder. Ann N Y Acad Sci 821:5775. Yehuda R (1998): Neuroendocrinology of trauma and posttraumatic stress disorder. In: Yehuda R, editor. Psychological Trauma. Washington, DC: American Psychiatric Press, 97 131. Yehuda R, Boisoneau D, Lowy MT, Giller EL (1995): Doseresponse changes in plasma cortisol and lymphocyte glucocorticoid receptors following dexamethasone administration in combat veterans with and without posttraumatic stress disorder. Arch Gen Psychiatry 52:583593. Yehuda R, Levengood RA, Schmeidler J, Wilson S, Guo LS, Gerber D (1996): Increased pituitary activation following metyrapone administration in posttraumatic stress disorder. Psychoneuroendocrinology 21:116. Yehuda R, Southwick SM, Krystal JM, Charney DS, Mason JW (1993): Enhanced suppression of cortisol following dexamethasone administration in combat veterans with posttraumatic stress disorder and major depressive disorder. Am J Psychiatry 150:83 86. Young EA, Haskett RF, Murphy-Weinberg V, Watson SJ, Akil H (1991): Loss of glucocorticoid fast feedback in depression. Arch Gen Psychiatry 48:693 699.