Relationships Among Plasma Dehydroepiandrosterone Sulfate and Cortisol Levels, Symptoms of Dissociation, and Objective Performance in Humans Exposed to Acute Stress

Charles A. Morgan III, MD, MA; Steve Southwick, MD; Gary Hazlett, PsyD; Ann Rasmusson, MD; Gary Hoyt, PhD; Zoran Zimolo, MD; Dennis Charney, MD

Context: Recently, a growing body of research has pro Results: The DHEA-S-cortisol ratios during stress were vided evidence that dehydroepiandrosterone sulfate significantly higher in subjects who reported fewer symp (DHEA-S) is involved in an organism's response to stress toms of dissociation and exhibited superior military per and that it may provide beneficial behavioral and neu formance. rotrophic effects. Conclusions: These data provide prospective, Objective: To investigate plasma DHEA-S and cortisol empiri cal evidence that the DHEA-S level is increased levels, psychological symptoms of dissociation, and mili by acute tary performance. stress in healthy humans and that the DHEA-S-cortisol ratio may index the degree to which an individual is buff Design: Prospective study. ered against the negative effects of stress.

Setting and Participants: Twenty-five healthy sub jects enrolled in military survival school. Arch Gen Psychiatry. 2004;61:819-825

vented by DHEA-S.[7,8], In addition, DHEA-S (DHEA) is an endoge prevents corticosterone-induced perfor nous hormone secreted mance decrements. [ 9, , 1Taken0 ] together, by the adrenal cortex in these findings suggest that DHEA-S may response to adrenocor- play a significant role in modulating vul tiocotropin-releasing hormone (ACTH).[1] nerability of the organism to negative con Dehydroepiandrosterone and its sulfated sequences of stress. derivative, DHEA-S, were originally In humans, levels of DHEA-S peak From the National Centerfor thought to be produced in situ in brain around ages 20 to 25 years and then de Post-TraumaticStress Disorder, tissue and were referred to as neuro cline to values of 20% to 30% at approxi Veterans Affairs New England steroids. At[2] present, however, it is be mately 70 to 80 years of age.[11-14] Levels of Healthcare System. West lieved that the only source of brain DHEA-S are also Haven, reduced in a number of Conn (Drs Morgan, DHEA-S is from the periphery.[ 3 ] A grow medical Southwick, Rasmusson, illnesses such as end-stage renal and ing body of research has provided evi [ 15 ] 6] Zimolo); the Department of disease, liver disease,[1type 2 diabetes dence that [ 17] Psychiatry, Yale University DHEA-S is involved in an or mellitus, coronary artery disease,[18] and School of Medicine, New ganism's response to stress and that it may rheumatoid arthritis.[19]Levels of DHEA-S Haven, Conn (Drs Morgan, provide beneficial behavioral and neuro have also been noted to be reduced in in Southwich, Rasmusson, Hoyt, trophic effects. dividuals with chronic fatigue syn and Zimolo); the Psychological In mice, DHEA-S release has been drome,[20] depression, [21-23 ] anxiety,[ 24 ]an Applications Directorate,John shown to be triggered by stress-induced orexia nervosa,[25 and] schizophrenia.[2 6 ]In F. Kennedy Special Warfare release of ACTH and also exhibits memory posttraumatic Training stress disorder (PTSD), lev Center and School, enhancing, antidepressant, anxiolytic, and Fort Bragg, els of DHEA or DHEA-S have been vari NC (Dr Hazlett); antiaggression properties.[3 -6 Neuronal] the National Institutes and able. Levels have been noted to be in of glial survival and differentiation Health, Bethesda, Md have also creased in Israeli soldiers and Kosovo been (Dr Charney); and Fleet shown to be enhanced by DHEA-S refugees with PTSD, but low in Kosovo Aviation Special Operations in dissociated cultures of mouse embryo refugees with PTSD who also had comor [ 7, 8] TrainingGroup PacificNavy brain and intact rats. Hippocampal neu bid depression. Recently, increased re Site 2, Naval Air Station, North rotoxicity induced by corticosterone, lease of DHEA was found in response to Island, Coronado, Calif oxidative stressors, and the glutamate ago ACTH administration in women with (DrHoyt). nist N-methyl-D-aspartate (NMDA) is pre- PTSD but a negative relationship be-

ARCH GEN PSYCHIATRY/VOL 61. AUG 2004 WWW.ARCHGENPSYCHIATRY.COM 819 tween the degree of ACTH-induced DHEA release and tivities. immediately following the collection of salivary samples, PTSD symptomatology (A.R., C.A.M., S.S., and D.C., un baseline plasma samples were collected by one of the current published data, 2004). investigators and Vladimir Coric, MD. Baseline salivary samples were again collected at 7:45 AM. Subjects then completed a modi Consistent with preclinical data regarding the bal fied self-report version of the Clinician-Administered Disso ance of DHEA-s to glucocorticoid levels, the DHEA-S ciative States Scale (CADSS) to rate their symptoms of disso cortisol ratio in humans has been significantly associ [ 3 6] [ 27 29] ciation during the classroom phase. ated with the degree of functional impairment - or the assesses the frequency and intensity of state [ 30] The CADSS response to clinical intervention. Taken together, these symptoms of dissociation. The items of the instrument are de preclinical and clinical studies provide indirect evi signed to assess how perceptually in touch (or out of touch) dence that the DHEA-S level and DHEA-S-cortisol ratio an individual is vis-a-vis his or her environment during spe may play a role in modulating the impact of stress on a cific conditions (nonstressed and stressed). Although some of variety of processes in humans. the items on the scale ask about one's sense of physical self (eg, The present study was designed to evaluate levels "Doyou feel as if you are looking at things outside of your body?" of DHEA-S and cortisol, the DHEA-S-cortisol ratio, and and "Do you feel as if you are watching the situation as an ob other items ask about cognitive or per the relationship of these hormone indexes with stress server or spectator?"), ceptual distortions (eg, "Do colors seem to be diminished in induced symptoms of dissociation and objective perfor intensity?"-"Do sounds almost disappear or become much stron in survival mance in military personnel participating ger than you would have expected?" "Do you space out or in school training. Previous investigations have demon some way lose track of what is going on?" and "Do you see things strated that survival school represents a valid, reliable as if you were in a tunnel, or looking through a wide-angle pho model for the study of acute, uncontrollable stress in hu tographic lens?"). The self-report scale contains 19 items, each mans.[ 3 1-34 ]The results of these studies demonstrate that of which is rated by subjects on a Likert scale of 0 (not at all) the stress experienced by subjects during survival school to 4 (extremely). A total score of 76 is possible. activates biological threat-response systems and elicits psychological symptoms of dissociation on a scale of mag STRESS SAMPLES nitude comparable to the degree of arousal and disso At the conclusion of the didactic phase of the training, sol ciation noted in humans responding to real-world, threat diers participated in an experiential phase of survival training. experiences. Clinical studies have provided evidence to-life During this phase, they were confined in a mock prisoner of that peritraumatic symptoms of dissociation represent a [ 3 5 ] war camp (POWC). In the POWC, each subject experienced risk factor for the development of PTSD and that such various types of psychological stress. Broadly speaking, these symptoms are positively associated with the stress included interrogations and problem-solving dilemmas de [ 33] induced release of glucocorticoids. On the basis of this signed to test the trainees' ability to use the information they and the preclinical and clinical literature suggesting that learned during the didactic phase. As noted in previous pub [ 313] an increased DHEA-S-cortisol ratio may buffer against lications, -3 these interrogations result in robust increases in the negative impact of stress,[ 78,, 30 ]we hypothesized that cortisol and catecholamine levels, heart rate, and subjective dis individuals with a higher DHEA-S-cortisol ratio during tress and significant reductions in testosterone level. During subjects also underwent uniform food and stress exposure would be protected from the negative im the POWC stage, sleep deprivation. Before interrogation stress, all subjects had pact of survival school stress as evidenced by fewer symp been deprived of food for approximately 8 hours and were physi military performance. toms of dissociation and superior cally inactive. During the 30-minute exposure to interroga tion stress, subjects remained standing and relatively immo bile; they did not engage in exercise or physical exertion. Immediately after interrogation, subjects were moved to a sec Twenty-six consecutively recruited active duty military per ond room identical in appearance to the first, where their blood sonnel were the subjects of this study. One female subject was and saliva samples were collected by the research team (C.A.M. removed from the data set, resulting in a study population of and G.H.) between 4:30 and 5 PM. Subjects provided the saliva 25. As designated by their military operational specialty, 8 sub samples before the venipuncture. After this, subjects contin jects were naval aviators, and 17 were nonaviating marines. The ued to undergo uniform sleep and food deprivation until their mean age was 25 years (SD, 4.4 years). Six subjects (24%) were release from the POWC. All participants were monitored by married, and 19 (76%) were single. The average number of years the survival school medical staff during the POWC stage and in the service was 4.6 (SD, 4.1). received water on a uniform schedule. The methods used in this study have been reported in de tail elsewhere. [31-14]In brief, before enrollment in this investiga MILITARY PERFORMANCE tion, each participant completed in-processing into the sur vival training course. Recruitment of subjects was conducted Survival school instructors performed an objective appraisal of by the principal investigator (C.A.M.). All subjects gave writ observable military-relevant performance of each participant ten, informed consent. As per survival training course require during the POWC phase of survival school. These perfor ments, all subjects provided documentation of physical exami mance assessment scores are part of the survival school pro nation and medical and psychiatric clearance before enrollment. gram and are not available to the public. The overall rating score, All subjects were free of illicit substances as documented by however, is designed to reflect how well a participant in train results of urine toxicologic screening. ing is able to demonstrate specific behaviors and problem solving abilities while experiencing acute stress. The perfor BASELINE ASSESSMENT mance ratings are scored on a scale that ranges from 0 (no skills demonstrated) to a maximum score of 4 (excellent demonstra Five days before stress exposure, baseline saliva samples were tion of skills). Because these performance scores were gener obtained at 4 PM on the second day of didactic (classroom) ac- ated independent of the research team and the measures col-

ARCH GEN PSYCHIATRY/VOL 61, AUG 2004 WWW.ARCHGENPSYCHIATRY.COM 820 lected by the research team, they represent a double-blind Exploratory post hoc analyses were planned for individual opportunity to assess the relationship between operationally CADSS items in the event that a significant relationship relevant military performance and the psychobiological was mea observed between DHEA-S-cortisol ratios and CADSS sures that were of interest to the research team.[33] total score. Finally, separate stepwise linear regression analyses were used to examine whether or how well the RECOVERY SAMPLES independent vari ables of age, rank, time in service, or the DHEA-S-salivary cor tisol ratio would explain the variance in stress-induced symp Twenty-four hours after the conclusion of POWC stress, re toms of dissociation (the CADSS total score) and in objectively covery plasma and saliva samples were collected in all sub assessed military performance. jects. Because of programmatic constraints within the Navy sur vival school program, the collection of recovery samples occurred at 7:45 AM and not 4 PM as in our previous investigations of the Army survival school program. In addition to providing blood results and saliva samples, subjects once again completed the CADSS. Subjects were asked to complete the CADSS using the stress The variables of they experienced during the interrogation phase of the POWC age, rank, and time in the service did not as their reference point. contribute to variance in the hormone or the psychologi Plasma samples were spun down in a refrigerated centri cal data, and thus were removed from the analyses. fuge, pipetted into microtubules, and frozen within 40 min utes of venipuncture. Samples were stored at -70°C from the PSYCHOLOGICAL MEASURES time of initial collection until analyses were performed. Sali vary samples were frozen and shipped with plasma samples to The mean CADSS score at baseline was 1.0 our laboratory within 24 hours of collection. (SD, 1.6), whereas the poststress CADSS score was 17.4 (SD, 13.0). This increase was statistically significant (F SALIVARY CORTISOL ANALYSIS METHODS 1,24=40.8 [P<.001]). In the published literature this would be con sidered a moderate level of dissociation.[ 37 ] Saliva was collected in Salivette tubes (Sarstedt Inc, Newton, NC), centrifuged, and pipetted into two 1.5-mL plastic vials. MILITARY The samples were shipped on dry ice to our laboratory and stored PERFORMANCE SCORES at -70 degreesC until assayed. Salivary cortisol levels were analyzed by means ofradioimmunoassay (IncStar Corp, Stillwater, Minn). The mean, objectively assessed military performance The intra-assay and interassay coefficients of variation were 4.2% rating for subjects was 2.3 (SD, 0.7), with a range of 1.0 and 6.1%, respectively. to 3.8.

PLASMA DHEA-S ANALYSIS METHODS HORMONE VALUES Frozen plasma samples were used and processed in batch by Compared with baseline, there was a means of DHEA-S commercial radioimmunoassay kits (Diag significant in crease in DHEA-S nostic Systems Laboratories, Inc, Webster, Tex). Determina level in response to stress (baseline, 27.81 pg/dL [SD, tions were performed on duplicate 10-mL plasma samples ac 11.06 pg/dL] [SI units (calculated with cording to the manufacturers recommendations, using supplied a conversion factor of 0.02714), 0.755mu mol/L {SD, 0.30 rabbit polyclonal anti-DHEA-S antibody-coated tubes. Plasma mu mol/L)]; stress, 60.12 pg/dL [SD, 26.15 pg/dL] [1.63 DHEA-S concentrations were measured with a sensitivity (de pmol/L (SD, 0.71 mumol/L)]; F1, 21,=76.1 [P<.001]), and tectability) of approximately 2 pg/dL and intra-assay and in it remained significantly increased compared with base terassay coefficients of variation of 9% and 15%, respectively. line at the recovery time point (27.81 pg/dL [SD, 11.06 mug/dL] [0.755mu mol/L (SD, 0.30 mumol/L ] vs 37.32 DATA ANALYSIS mug/dL [SD, 16.98 mug/dL] [1.01 mu mol/L (SD, 0.46 mumol/L ]; F[ 1, 8 =22.4 [P<.001]). The DHEA-S values Separate repeated-measures analyses of variance (ANOVAs) us at recovery were significantly reduced ing the time factors baseline-stress and baseline-recovery were compared with the stress time point (F 6 2 performed to detect whether exposure to stress significantly af 1,1 7= 8. [P<.001]). Plasma cortisol level was fected levels of plasma DHEA-S and plasma and salivary cor also significantly increased by exposure to survival school tisol. To control for diurnal variation, the cortisol samples col stress and remained significantly elevated compared with lected at the 4 PM point at baseline were compared with the baseline at the recovery time point (baseline, 8.6 pg/dL cortisol samples collected at the stress time point. Similarly, [237.3 nmol/L [SD, 3.8 mug/dL (104.8 nmol/L}]; the ANOVAs examining stress, whether salivary cortisol levels had re 31.1 mug/dL [858.0 nmol/L] [SD, 5.8 pg/dL (160.0 nmol/ turned to the reference range after the conclusion of the train L)]; recovery, 19.9 mug/dL [549.0 nmol/L] [SD, 7.0 ing used the 7:45 AM salivary cortisol pg/dL samples, because these (193.1 nmol/L)]) (F[1,2= 168.2 were closely associated 2 [P<.001] and F1, 23=55.8 with the time of saliva assessment on [P<.001], the day of recovery (7:45 AM). baseline vs stress and baseline vs recovery, respectively). Similarly, Pearson correlation analyses were used to evaluate the re compared with baseline levels, salivary cortisol lationships among the assessed hormone levels (DHEA-S and level was significantly increased by ex cortisol) and the relationships among the hormones and the posure to survival school stress (baseline, 0.14 mug/dL [3.9 independent variables of age and weight. nmol/L] [SD, 0.04 pg/dL (1.1 nmol/L)] ; stress, 0.87 mug/dL Spearman rank correlation analyses were used to com [24.0 nmol/L] [SD, 0.45 mug/dL (12.4 nmol/L)]) (F1, = 62.5 pare the DHEA-S--salivary 24 cortisol ratios from the baseline, stress, [P<.001]). However, compared with baseline morning and recovery time points with psychological symptoms of dis values, morning salivary cortisol levels collected at re sociation (CADSS total scores) and military performance scores. covery (recovery value, 0.49mu g/dL [13.5 nmol/L] [SD,

ARCH GEN PSYCHIATRY/VOL 61, AUG 2004 www.ARCHGENPSYCHIATRY.COM 821 CADSS items 2 (feeling unreal as if in a dream; r=-0.44; [P<.041), 6 (feeling disconnected from one's body; r=-0.54 [P=.009]), 10 (colors appearing to be dimin ished in intensity; r=-0. 6 [P= .0041), 11 (feeling as if one is viewing the world in a tunnel or looking through a wide angle lens; r= -0.6 [P= .003]), 15 (feelings of being spaced out or of losing track of what is going on; r=-0.44 [P=.04]), 16 (sounds disappearing or becoming much stronger than expected; r=-0.51 [P=.01]), 17 (things hav ing a special sense of clarity; r=-0.49 [P=.02]), and 18 (feeling as if one is looking at world as if in a fog, people appearing far away or unclear; r=-0.72 [P<.001]). With regard to the relationship between objective Z-Score DHEA-S-Cortisol Ratio During Stress military performance and the indexes of DHEA-S and cor tisol levels and dissociation, as shown in Figure 2, there Figure 1. Correlation between dehydroepiandrosterone sulfate was a significant positive correlation between the DHEA (DHEA-S)-cortisol ratio and stress-induced symptoms of dissociation in 22 healthy subjects enrolled in military survival school (R= 0.29; Spearman S-salivary cortisol ratio during stress and the military per p=-0.63 (P= .0021). formance scores (r= 0.61 [P=.008]) and a significant nega tive correlation between stress-induced levels of salivary cortisol and military performance (r=-0.51 [P<.01]). In addition, and similar to the findings of our previous in vestigations, there was a significant negative relation ship between stress-induced symptoms of dissociation (CADSS scores) and military performance (r=-0.51 [P<.01]). No significant relationships were observed be tween baseline hormone values, recovery hormone val ues, DHEA-S-cortisol ratios at baseline or at recovery, and the outcome measures of dissociative symptoms and military performance scores in response to stress.

REGRESSION ANALYSES

Stepwise linear regression analysis using poststress CADSS Figure 2. Correlation between dehydroepiandrosterone sulfate dissociation scores as the dependent variable and the (DHEA-S--cortisol ratio during stress and military performance in 18 healthy subjects enrolled in military survival school (R2=0.37; Spearman p=5.2 DHEA-S-salivary cortisol ratio and plasma levels of DHEA-S [P=.03]). and cortisol during stress as the independent variables showed that the model was significant (F1 ,19= 7 [P= .02]). 2 0.20 pg/dL {5.5 nmol/L}]) tended to be significantly higher The adjusted multivariate coefficient of determination (R ) only at the recovery time point (F 1,24=3.7 [P=.06]). for the model was 0.24 for the predictor DHEA-S salivary cortisol ratio during stress. The model did not im CORRELATION ANALYSES prove when the variables plasma levels of DHEA-S and cor tisol during stress were added. The standardized beta No significant correlations were observed between plasma coefficient value for DHEA-S-salivary cortisol ratio dur levels of DHEA-S and cortisol or between plasma levels ing stress was -0.53, with a t value of -2.6 (P=.02). of DHEA-S and salivary levels of cortisol at any of the Stepwise linear regression analysis using military per assessment time points (baseline and stress or recovery). formance scores as the dependent variable and the DHEA Similarly, Spearman rank correlation analyses did not re S-salivary cortisol ratio and plasma levels of DHEA-S and veal a significant correlation between plasma cortisol level cortisol during stress as the independent variables showed and poststress CADSS scores (r=0.16; P=.41). How that the model was significant (F1 ,16=5.6 [P=.03]). The 2) ever, there was a significant positive relationship be adjusted multivariate coefficient of determination (R for tween salivary cortisol level during stress exposure and the model was 0.23 for the predictor DHEA-S-salivary the poststress CADSS scores (r= 0.4; P<.05) and a trend cortisol ratio during stress. The model did not improve for a significant negative relationship between stress when the variable plasma level of DHEA-S or cortisol dur induced levels of DHEA-S and poststress CADSS scores ing stress was added. The standardized betacoefficient value (r=-0.4; P=.08). Finally, as shown in Figure 1, there for DHEA-S-salivary cortisol ratio during stress was 0.52, was a significant negative correlation between the DHEA with a t value of 2.4 (P=.03). S-salivary cortisol ratio during stress and the poststress CADSS scores (r=-0.63; P=.002). Analysis of individual CADSS items indicated that there were significant negative relationships between The principal finding of this study is that individuals with the DHEA-S-salivary cortisol ratio during stress and a higher DHEA-S-salivary cortisol ratio during stress ex-

ARCH GEN PSYCHIATRY/VOL 61, AUG 2004 WWW.ARCHGENPSCHIATRY.COM 822 perienced fewer symptoms of dissociation and exhib nuclear uptake of activated glucocorticoid receptors in ited superior military performance. Because indexes of the neurons of the hippocampus.[ 46 ] Dehydroepiandros dissociation and military performance presumably in terone also protects against excitatory amino acid- and dex central processes, the data herein provide support oxidative stress-induced damage, restores cortisol that in healthy humans, the ratio or balance between cir induced decrements in long-term potentiation, regu culating levels of DHEA-S to unbound cortisol may help lates programmed cell death, and promotes neurogen buffer against centrally mediated, negative [ 7 effects of stress. esis in the hippocampus. ,8,47-50]Thus it is possible that One implication of the present findings is that a low the military personnel exhibiting higher DHEA-S DHEA-S-cortisol ratio may be associated with vulner cortisol ratios during extreme stress in this study came ability to stress-induced symptoms of dissociation. In the to survival training with brain structures and functional future it may be fruitful to conduct clinical trials de capacities that protected them from dissociation during signed to prospectively evaluate whether augmentation the interrogation stress. of DHEA-S levels in humans, before the time of their ex Preclinical and clinical data suggest that the re posure to stress, will confer a protective effect, as evi duced levels of dissociation in subjects with an increased denced by diminished peritraumatic dissociation and im DHEA-S-salivary cortisol ratio during stress proved cognitive performance. may be due, in part, to the action of DHEA-S at NMDA receptors and/or Military performance was significantly and posi at the gamma-aminobutyric acid-benzodiazepine receptor com tively associated with DHEA-S-cortisol ratios. Because plex, perhaps at the level of the hippocampus or frontal the rating scales used by the military have not been made cortex. Dehydroepiandrosterone sulfate serves as a available to nega the public, the degree to which these scales tive modulator of the y-aminobutyric acid-benzodiaz relate to more traditional [ 40, 512,] cognitive or psychological epine receptor complex, and there is evidence of al measures has not yet been fully established. However, tered benzodiazepine receptor modulation and sensitivity in 3 separate investigations, psychological symptoms of in stress-related disorders characterized by symptoms of dissociation-as measured 53 by the CADSS-have been dissociation, such as PTSD. Furthermore, DHEA-S also shown to be associated with poor military performance. positively modulates NMDA receptors, which have been Because the military performance scores reflect the ca implicated in dissociative phenomena in humans as as pacity of soldiers' [ 37 ] cognitive and decision-making abili sessed by the CADSS. Thus, it is reasonable to speculate ties during stress, the present finding of a positive rela that the reduced levels of dissociation in subjects with a tionship between military performance scores and the high DHEA-S--cortisol ratio may be due, in part, to the effect DHEA-S-cortisol ratio adds weight to the idea that in of DHEA on NMDA receptors and/or the effect of DHEA creased levels of DHEA-S are associated with an anti at the y-aminobutyric acid-benzodiazepine receptor com stress effect on human cognition. These findings are plex, perhaps at the level of the hippocampus. consistent with those of previous clinical studies link There are several limitations to this study. First, all ing superior cognitive performance with higher con subjects in survival school training experienced food centrations of DHEA.[ 3 8 ] deprivation before stress exposure. Because dieting in Although no preclinical studies deal specifically with clinically obese individuals has been reported to result symptoms of dissociation, [ 54 much is known about the re in an increase in DHEA-S level, ]it is theoretically pos lationship between levels of cortisol, DHEA-S, and glu sible that the increase in DHEA-S levels noted in the tamate and stress-induced neurotoxicity. However, it is present subjects may have been, in part, influenced by not known at present whether stress-induced dissocia this factor. However, because food deprivation was kept tion in humans is related to stress-induced neurotoxic uniform across subjects, this factor cannot account for ity, as reported in the preclinical literature. the significant relationships between the DHEA-S-cor A host of preclinical investigations have shown that tisol ratios and the symptoms of dissociation and mili glucocorticoids can be neurotoxic and that DHEA-S ex tary performance. erts a potent antiglucocorticoid effect peripherally and All subjects also experienced sleep deprivation, rais centrally. For[7] example, prolonged exposure to high lev ing the possibility that alterations in the diurnal varia els of circulating corticosterone in rats increases the age tion in DHEA-S and cortisol levels may have contrib related rate at which hippocampal pyramidal neurons are uted to the findings of this study. Although possible, this lost.[ 3 9 ]Glucocorticoids have also been shown to poten is unlikely to explain the present findings for 2 reasons. tiate neurodegeneration induced by anoxia and gluta First, the findings of previous studies in military sub mate analogues. The neurotoxic effects of glucocorti jects exposed to high stress have shown that the diurnal coids can be attenuated or blocked by in vivo and in vitro variation of hormones is extremely small relative to the administration of DHEA-S. [ 78,] very large alterations of these hormones induced by ex Preclinical investigations [ 31 ] suggest several possible posure to acute stress. Second, sleep deprivation was mechanisms by which DHEA-S or DHEA may protect kept uniform across all subjects. against dissociation or improve cognition in humans. For Another limitation is the relatively small number of example, antiglucocorticoid effects of DHEA have been subjects in this study. It is possible that the large amount demonstrated in many tissues, including brain.[40-45]Within of the variance explained in the regression analyses may the brain, region-specific metabolism of DHEA may ul be due to the relative homogeneity of these military per timately control the nature of DHEA effects on cogni sonnel, and the present data may not generalize to the tion and behavior.[ 46 ] For instance, 7 alpha-hydroxylated me general population. However, the data may be relevant tabolites of DHEA have been shown to interfere with the for civilian groups (eg, firefighters, police officers, and

ARCH GEN PSYCHIATRY/VOL 61, AUG 2004 WWW.ARCHGENPSYCHIATRY.COM 823 16. Floreani A,Titta M, Plebani M,Faggian D, Chiaramonte M, Naccarato R. Sex hor emergency personnel) who are at increased risk for stress a mone changes in post-menop usal women with primary biliary cirrhosis (PBC) exposure and stress-related illness. and with cryptogenic chronic liver disease. Clin Exp Obstet Gynecol. 1991:18: 229-234. Submitted for publicationJuly 22, 2003; final revision re 17. Yamauchi A, Takei I, Kasuga A, Kitamura Y, Ohashi N, Nakano S, Takayama S, T. Depression of dehydroepiandrosterone February.13, 2004; accepted February 18, 2004. Nakamoto S, Katsukawa F, Saruta ceived men: comparison between non-insulin dependent diabe Insti in Japanese diabetic This study was supported by US Army Research tes mellitus and impaired glucose tolerance. Eur J Endocrino. 1996;135:101 tute for Environmental Medicine, Natick, Mass; the Robert 104. Mitchell Centerfor Repatriated POW Studies, Pensacola, 18. Herrington DM, Gordon GB,Achuff SC, Trejo JF, Weisman HJ, Kwitervich P0, Fla; and the National Centerfor Post-TraumaticStress Dis Pearson TA. Plasma dehydroepiandrosterone and dehydroepiandrosterone sul fate inpatients undergoing diagnostic coronary angiography. JAm Coll Cardiol. order, West Haven, Conn. 1990;16:862-870. We thank Jeremy Cordova, MS; CPO John Burkhart, 19. Masi AT, Sex hormones and rheumatoid arthritis: cause or effect relationships USN; and Vladimir Coric, MD, whose assistancegreatly fa in a complex pathophysiology? Clin Exp Rheumatol. 1995;13:227-240. cilitated the completion of this project. 20. Scott LV, Salahuddin F,Cooney J, Svec F, Dinan TG. Differences in adrenal ste The views expressed in this article reflect those of the roid profile in chronic fatigue syndrome, in depression and in health. J Affect Oisord. 1999;54:129-137. do not represent those of the US government authors and 21. Goodyer 1M,Herbert J, Altham PME, Pearson J, Secher SM, Shiers HM. Adrenal or the funding agencies. secretion during major depression in 8-16 year olds, 1:altered diurnal rhythms Correspondence: Charles A. Morgan III, MD, MA, in salivary cortisol and dehydroepiandrosterone (DHEA) at presentation. Psy c/o 116A, National Centerfor Post-TraumaticStress Disor chol Med. 1996;26:245-256. F,Chan T, Raum WJ, Ormiston S, der, Veterans Affairs New England HealthcareSystem, 950 22. Wolkowitz OM, Reus VI, Roberts E, Manfredi R, CanickJ, Brizendine L,Weingartner H. dehydroepiandrosterone (DHEA) Ave, West Haven, CT 06516 (charles.a.morgan Johnson Campbell treatment of depression. BiolPsychiatry.1997;41:311-318. @yale.edu). 23. Micheal A, Jenaway A, Paykel ES,Herbert J. Altered salivary dehydroepiandros terone levels in major depression in adults. Biol Psychiatry. 2000;48:989-995. 24. Fava M, Rosenbaum JF, MacLaughlin RA, Tesar GE,Pollack MH, Cohen LS, Hirsch M. Dehydroepiandrosterone-sulfate/cortisol ratio in panic disorder. Psy chiatry Res. 1989;28:345-350. 25. Zumoff B, Walsh BT, Katz JL, Levin J. Rosenfeld RS, Kream J, Weiner H. Sub 1. Nieschlag E, Loriaux DL, Ruder HJ, Zucker IR, Kirschner MA, Lipsett MB. The normal plasma dehydroepiandrosterone to cortisol ratio in anorexia nervosa: secretion of dehydroepiandrosterone and dehydroepiandrosterone sulphate in a second hormonal parameter of ontogenic regression. J Clin Endocrinol Metab. man. J Endocrinol. 1973;57:123-134. 1983,56:668-672. 2. Majewska MD, Demirgoren S, Spivak CE,London ED. The neurosteroid dehy gonadal steroid hormones in young schizo droepiandrosterone sulfate is an allosteric antagonist of the GABAA receptor. Brain 26. Oades RD,Schepker R. Serum Psychoneuroendocrinology. 1994;19:373-385. Res. 1990;526:143-146. phrenic patients. HP, Hansson LO,Theorell T. Elevated blood levels of dehydroepi 3. Compagnone NA, Mellon SH. Neurosteroids: biosynthesis and function of these 27. Soendergaard vary with symptom load in posttraumatic stress disor novel neuromodulators. Front Neuroendocrinol. 2000;21:1-56. androsterone sulphate from a longitudinal study of refugees in Sweden. Psychother Psy 4. Melchior CL, Ritzmann RF. Dehydroepiandrosterone is an anxiolytic in mice on der: findings the plus maze. Pharmacol Biochem Behav. 1994;47:437-441. chosom. 2002;71:298-303. IM, Herbert J, Altham PME. Adrenal steroid secretion and major de 5. Robel P, Baulieu EE.Dehydroepiandrosterone (DHEA) is a neuroactive neuro 28. Goodyer III: influence of cortisol/DHEA ratio at presentation steroid. Ann N YAcad Sci. 1995;774:82-110. pression in 8-16 year olds, subsequent rates of disappointing life events and persistent major depres 6. Roberts E. Dehydroepiandrosterone (DHEA) and its sulfate (DHEAS) as neural on facilitators: effects on brain tissue in culture and on memory in young and old sion. Psychol Med. 1998;28:253-255. A longitudinal study of hormonal reactions ac mice: a cyclic GMP hypothesis of action of DHEA and DHEAS in nervous system 29. Soendergaard HP, Theorell T. recently resettled refugees. Psychother Psychosom. andother tissues. In: Kalimi M, Regelson W, eds. The Biologic Role of Dehy companying life events in droepiandrosterone (dHEA). Berlin, Germany: W de Gruyter; 1990:13-42. 2003;72:49-58. Antoni MH, Kumar M, Ironson G, McCabe P, Fernandez JB, Fletcher 7. Kimonides VG, Khalibi NH, Svendsen CN,Sofroniew MV, Herbert J. Dehydro 30. Cruess DG, stress management buffers de epiandrosterone (DHEA) and DHEA-sulfate (DHEAS) protect hippocampal neu MA, Schneiderman N. Cognitive-behavioral in the cortisol/ rons against excitatory amino acid-induced neurotoxicity. Proc Natl Acad Sci creases indehydroepiandrosterone sulfate (DHEA-S) and increases and perceived stress among HIV USA. 1998;95:1852-1857. DHEA-S ratio and reduces mood disturbance 1999:24:537-549. 8. Kimonides VG, Spillantini MG, Sofroniew MV, Fawcett JW, Herbert J. Dehydro seropositive men. Psychoneuroendocrinology. J, Hazlett G, Fox P, Southwick SM, Charney DS, epiandrosterone antagonizes the neurotoxic effects of corticosterone and trans 31. Morgan CAIII, Wang S, Mason experiencing military survival train location of stress-activated protein kinase 3 inhippocampal primary cultures. Greenfield G. Hormone profiles inhumans Neuroscience. 1999;89:429-436. ing. Bio/ Psychiatry. 2000;47:891-901. SM, Rasmusson A, Hauger R, Charney DS. 9. Fleshner M, Pugh CR, Tremblay D, Rudy JW. DHEA-S selectively impairs con 32. Morgan CA III, Wang S, Southwick exposed to military survival training. BiolPsy textual-fear conditioning: support for the antiglucocorticoid hypothesis. Behav Plasma neuropeptide-Y in humans Neurosci. 1997;111:512-517. chiatry. 2000;47:902-909. Hazlett G,Rassmusson A, Anderson G, Charney DS. Re 10. Frye CA, Lacey EH.The neurosteroids DHEA and DHEAS may influence cogni 33. Morgan CAIll, Wang S, neuropeptide Yand human perlor tive performance by altering affective state. Physiol Behav. 1999;66:85-92. lationships among cortisol, catecholamines, Med. 2001;63:412-442. 11. Labrie F,Belanger A,Cusan L, Gomez J-L, Candas B.Marked decline inserum mance during uncontrollable stress. Psychosom Richardson G,Schnurr P,Southwick SM. Symp concentrations of adrenal C19 sex steroid precursors and conjugated androgen 34. Morgan CA Ill, Hazlett G.Wang S, stress: a pro metabolites during aging. J Clin Endocrinol Metab. 1997;82:2396-2402. toms ofdissociation in humans experiencing acute uncontrollable Am J Psychiatry. 2001;158:1239-1247. 12. Orentreich N, Brind JL, Rizer RL, Vogelman JH. Age changes and sex differ spective investigation. KL, Best SR, Wentworth KA. Lon ences inserum dehydroepiandrosterone sulfate concentrations throughout adult 35. Marmar CR, Weiss DS, Metzler TJ, Delucchi of continuing distress following critical inci hood. J Clin Endocrinol Metab. 1984;59:551-555. gitudinal course and predictors in emergency services personnel. J Nerv Ment Dis. 1999;187: 13. Orentreich N, Brind JL, Bogelman JH, Andres R, Baldwin H. Long-term longitu dent exposure dinal measurements of plasma dehydroepiandrosterone sulfate in normal men. 15-22. Putnam FW, Southwick SM, Marmar C,Charney DS, Ma J Clin Endocrinol Metab. 1992;75:1002-1004. 36. Bremner JD, Krystal JH, of dissociative states with the Clinician-Administered Dis 14. Sulcova J, Hill M, Starka L. Age and sex related differences in serum levels of zure CM. Measurement (CADSS). J Trauma Stress. 1998;11:125-136. unconjugated dehydroepiandrosterone and its sulphate innormal subjects. JEn sociative States Scale Oren DA, Berman RM, Hu XS, Cappiello A, Krystal JH. docrinol. 1997;154:57-62. 37. Anand A, Charney DS, of the neuropsychiatric effects of ketamine with lamotrigine: sup 15. Zumoff B, Walter L, Rosenfeld RS, Strain JJ, Degen K, Strain GW, Levin J, Fuku Attenuation effects of N-methly-D-aspartate receptor antago shima D. Subnormal plasma adrenal androgen levels in men with uremia. JClin port for hyperglutamatergic 2000;57:270-276. Endocrinol Metab. 1980;51:801-805. nists. Arch Gen Psychiatry.

ARCH GEN PSYCHIATRY/VOL 61, AUG 2004 WWW.ARCHGENPSYCHIATRY.COM 824 38. Wolf OT, Kudielka BM, Helihammer DH, Helihammer J, Kirschbaum C.Oppos 47. Bastianetto S, Ramassamy C,PoirierJ, Quirion R. Dehydroepiandrosterone (DHEA) ing effects of DHEA replacement in elderly subjects on declarative memory and protects hippocampal cells from oxidative stress-induced damage. MolBrain Res. attention afterexposure to alaboratory stressor. Psychoneuroendocrinology. 1998; 1999;66:35-41. 23:817-629. 48. Kaminska M, Harris J, Gilsbers K, Dubrovsky B. Dehydroepiandrosterone sul 39. Hortnagl H, Berger ML, Havelec L, Hornykeiwicz 0. Role of glucocorticoids in fate (OHEAS) counteracts decremental effects of corticosterone on dentate gy the cholinergic degeneration in rat hippocampus induced by ethylcholine aziridi rus LTP: implications for depression. Brain Res Bull. 2000;52:229-234. nium (AF64A). JNeurosci.1993;13:2939-2945. 49. Karishma KK,Herbert J. Dehydroepiandrosterone (DHEA) stimulates neurogen 40. Demirgoren S, Majewska MD, Spivak CE,London ED. Receptor binding and elec esis in the hippocampus of the rat, promotes survival of newly formed neurons trophysiological effects of dehydroepiandrosterone sulfate, an antagonist of the and prevents corlicosterone-induced suppression. Eur J neurosci. 2002;16:445 GABA A receptor. Neuroscience. 1991;45:127-135. 453. 41. Kishimoto Y, Hoshi M. Dehydroepiandrosterone sulphate in rat brain: incorpo 50. Zhang L. Li B, Ma W, Barker JL, Chang YH, Zhao W, Rubinow DR. Dehydroepi ration from blood and metabolism in vivo. JNeurochem. 1972;19:2207-2215. androsterone (DHEA) and its sulfated derivative (DHEAS) regulate apoptosis dur 42. Blauer KL, Poth M, Rogers W, Bernton E.DHEA antagonizes the suppressive ef ing neurogenesis by triggering the Akt signaling pathway in opposing ways. Brain fects of dexamethasone on lymphocyte proliferation. Endocrinology. 1991;129: Res Mol Brain Res. 2002;98:58-66. 3174-3179. 51. Majewska MD, Schwartz RD. Pregnenolone-sulfate: an endogenous antagonist 43. Browne ES, Wright BE,Porter JR, Svec F. Dehydroepiandrosterone: antigluco of the gamma-aminobutyricacid receptorcomplex in brain? Brain Res. 1987;404:355 corticoid action in mice. Am JMedSci. 1992;303:366-371. 360. 44. Daynes RA, Dudley DJ, Araneo BA. Regulation of murine lymphokine produc 52. Majewska MD, Mienville J-M, Vicini S. Neurosteroid pregnenolone sulfate an tion in vivo, It: dehydroepiand rosterone is a natural enhancer of interleukin 2 syn tagonizes electrophysiological responses to GABA in neurons. NeurosciLett.1988; thesis by helper T cells. EurJ Immunol. 1990;20:793-802. 90:279-284. 45. Morlin R, Starka L. Neurosteroid 7-hydroxylation products in the brain. lntRev 53. Gavish M, Laor N, Bidder M, Fisher D, Fonia 0, Muller U, Reiss A, Wolmer L, Neurobiol.2001;46:79-95. Karp L, Weizman R. Altered platelet peripheral type benzodiazepine receptor in 46. Rose KA, Stapleton G,Dott K,Kieny MP, Best R,Schwarz M, Russell DW, Bjorkheim posttraumatic stress disorder. Neuropsychopharmacology 1996;14:181-186. I, Seckl J, Lathe R. Cyp7b, a novel brain cytochrome P450, catalyzes the syn 54. Jakubowicz DJ. Beer NA, Beer RM, Nestler JE. Disparate effects of weight re thesis of neurosteroids 7aw-hydroxy dehydroepiandrosterone and 7a-hydroxy preg duction by diet on serum dehydroepiandrosterone-sulfate levels in obese men nenolone. Proc Natl Acad SciUSA. 1997;94:4925-4930. and women. J Clin Endocrinol Metab. 1995,80:3373-3376.

ARCH GEN PSYCHIATRYIVOL 61, AUG 2004 WW.ARCHGENPSYCHIATRY.COM 825