Hypersensitivity of the hypothalamic-pituitary-adrenal axis to naloxone in post-traumatic stress disorder

Hypersensitivity of the hypothalamic-pituitary-adrenal axis to naloxone in post-traumatic stress disorder

BIOL PSYCHIATRY 1993;33:585-593 585 Hypersensitivity of the Hypothalamic-PituitaryAdrenal Axis to Naloxone in Post-Traumatic Stress Disorder Gregory...

743KB Sizes 13 Downloads 72 Views

BIOL PSYCHIATRY 1993;33:585-593

585

Hypersensitivity of the Hypothalamic-PituitaryAdrenal Axis to Naloxone in Post-Traumatic Stress Disorder Gregory I. Hockings, Jeffrey E. Grice, Warren K. Ward, Margaret M. Waiters, Graerne R. Jensen, and Richard V. Jackson

Naloxone, which increases endogenous corticotropin-releasing hormone (CRH) release by blocking an inhibitory opioidergic tone on the hypothalamic-pituitary-adrenal (HPA) axis, was administered in a dose-response protocol to seven healthy volunteers and 13 patients with treated posttraumatic stress disorder (PTSD). Six of the PTSD patients showed an it,creased hormonal response to the lowest naloxone dose (6 l~g/kg) compared to both the control subjects and the other PTSD patients. This difference persisted on detailed subgroup analysis, although it was less marked at the highest naloxone dose (125 ttg/kg). The responses of the other seven PTSD patients were indistinguishable from those of the control group. The greater responses of the six PTSD patients could not be explained on the basis of associated psychiatric illnesses or psychotropic drug therapy, and did not correlate with standard psychological testing or severity of PTSD. The results of this preliminary study therefore suggest that a hypersensitivity of the HPA axis to endogenous CRH stimulation may occur in PTSD.

Key Words: Adrenocorticotropin, conisol, naloxone, hypothalamus, post-traumatic stress disorder, corticotropin-releasing hormone

Introduction Posttraumatic stress disorder (PTSD) may affect individuals following exposure to extreme stress such as combat, physical assault, or natural ~isaster (Helzer et al 1987). Abnormal activity of the autonomic nervous system (Kolb 1987; Paige et al 1990) and of the hypothalamic-pituitaryadrenal (HPA) axis (Smith et al 1989) have been suggested From the Neuroendocrine Research Unit, University of Queensland Department of Medicine (GIH, JEG, MMW, RVJ) and the Department of Psychiatry (WKW, GRJ), Greenslopes Hospital, Brisbane, Australia. Address reprint requests to: Dr. Richard V. Jackson, Neuroendocrine Research Unit, University Dep',u'tment of Medicine, Graenslopes Hospital, Newdegate Street, Greenslopes, Brisbane, Queensland, 4120, Australia. Received June 29, 1991; revised January 2, 1993. © 1993 Society of Biological Psychiatry

as the basis of some of the characteristic features of this disorder. However, studies of neuroendocrine function in PTSD, which have measured basal plasma cortisol concentrations, cortisol suppression by dexamethasone, and 24.hr urinary free cortisol and catecholamine levels have reported inconsistent and sometimes conflicting findings (Mason et al 1986; Kosten et al 1987; Kudler et al 1987; Halbreich et al 1989; Dinan et al 1990; Kosten et al 1990; Pitman and Orr 1990; Yehuda et al 1990). In this pilot study, we have administered the opioid antagonist naloxone, using a dose-response protocol, to 13 Vietnam veterans with severe, combat-related PTSD. In high doses, naloxone increases the secretion of endogenous corticotropin-releaGing hormone (CRH) from the hy0006-3223/93/$06.00

586

alOL PSYCHIATRY 1993:33:585-593

pothalamus by blocking an endogenous inhibitory opioidergic tone on CRH-secreting parvocellular neurones of the paraventricular nucleus (Grossman and Besser 1982; Jackson et al 1990). This results in increased release of adrenocorticotrophic hormone (ACTH) and cortisol in normal humans (AI-Damluji et al 1990; Hockings et al 1991). Because naloxone stimulates the entire HPA axis, the eL fects of its administration may provide additional information on neuroendocrine function in PTSD.

Methods

Subjects Thirteen men who were Vietnam veterans aged 40-48 years participated in the study; 11 were inpatients, and two were outpatients. Exclusion criteria included a history of drug or alcohol abuse within the previous 2 months (all the veterans had a past history of alcohol abuse), major active medical illness, or organic brain syndromes. All were receiving drug treatment for PTSD, and this was not withdrawn in view of the severity of their condition. The control group for the naloxone studies comprised seven healthy men aged 20-46 years. All were nonsmokers and medication-free. Both control subjects and veterans underwent a history, physical examination, and screening tests, which included biochemical profile, full blood count, urinalysis, electrocardiogram, thyroid function tests, testosterone, follicle-stimulating hormone (FSH), luteinizing hor,'nonv (LH), and prolactin levels. All the veterans were diagnosed by the same psychiatrist (GP,J) as having combat-related PTSD according to DSM.LU.R criteria (American Psychiatric Association 1987). The severity of PTSD was scored by a structured questionnaire, the Frederick Reaction Index Scale (copyright 1988 Dr. Calvin J. Frederick, VA Medical Center, West Los Angeles, CA, USA); this scores the severity of 20 symptoms of PTSD to give a maximum possible result of 80. The presence of associated psychiatric diagnoses was determined by the application of the Research Diagnostic Criteria (RDC) (Spitzer et al 1978). In addition, each veteran was assessed by the following psychological tests: Hamilton Depression Rating Scale, Beck Depression Inventory, and State-Trait Anxiety Inventory (STAI). A Melancholia Score, derived from the Hamilton Depression Scale results, was also obtained for each patient.

Study Design Naloxone stimulation tests were performed in the afternoon, when basal secretory activity for ACTH and cortisol is low. Subjects were fasted after a light meal taken not later than 10 AM. At about 1 PM a forearm venous cannula was inserted for both drawing of blood samples and in-

G.i. Hockings et al

fusion of isotonic saline, and two 10-ml basal blood samples for measurement of ACTH and cortisol were taken 30 and 45 min later. Ten milliliters of isotonic saline (as placebo) was then administered intravenously at time 0 rain, and an additional four blood samples were drawn at 5-rain intervals. This sequence was repeated at 30-min intervals following administration of intravenous naloxone (Narcan; The Boots Company, North Rocks, N.S.W., Australia) in doses of 6 p~g/kg body weight at + 30 min, 25 Ixg/kg at +60 rain, and 125 ttg/kg at +90 min. After completion of the final sequence, two further samples were taken at 30-rain intervals. Blood pressure and heart rate were monitored, and the subjects were blinded to the sequence of injections. The study protocol was approved by the University of Queensland Ethics Committee and by the Greenslopes Hospital Ethics Committee. Each subject gave written, informed consent for their participation in the studies.

Hormone Assays Plasma ACTH was measured in unextracted plasma by radioimmunoassay (Nicholson et al 1984). The limit of detection of the assay was !. 1 pmol/L. Interassay and intraassay coefficients of variation at 7.7 pmollL were 7.8% and 3.7% respectively. Plasma cortisol was measured by direct-phase high pressure liquid chromatography (HPLC) using a previously described method (Grice et al 1991). Recovery of cortisol was 95%-100%. The limit of detection was 25 nmol/L. lnterassay and intraassay coefficients of variation at 165 nmoi/L were 6% and 4.5%. respectively.

Statistical Analysis Results are expressed as mean -- SEM. Mean ACTH and cortisoi levels were analysed by between-within multivariate analysis of variance (MANOVA) with repeated measures, and mean peak hormone levels, mean peak changes (deltas), and areas under the hormone-time curves (AUC) were assessed by one-way between groups analysis of variance (ANOVA). The PTSD subgroups were compared regarding associated psychiatric conditions, drugs and psychological tests by the Mann-Whitney U test. Correlations between nongrouping variables (e.g., psychological test scores) and hormone peaks, deltas, and areas under the hormone-time curves were made by Pearson product-moment correlation. Mean ages were compared by independent t-tests. Subgroup analysis was validated by Bartlett's test for homogeneity of variance (Snedecor and Cochran 1980), and the existence of two separate PTSD subgroups was confirmed by cluster analysis using k-means (Hartigan 1975). Analyses were performed on the software package CSS (Statsoft, Tulsa, OK).

Naloxone Hypersensitivity in PTSD

BIOL PSYCHIATRY

587

1993;33:585-593

ResuRs Psychiatric Assessment and Treatment All patients fulfilled both RDC and DSM-HI-R criteria for the diagnosis of PTSD. In addition, the severity of PTSD was rated as very severe in 11 cases, and severe in the other 2 cases, according to both the Frederick Reaction Index Scale and clinical assessment (GRJ). The mean scores for the various psychometric tests were as follows: Hamilton Depression 20.1 -+ 5.9; Beck Depression 27.5 ± 3.4; Melancholia 12.5 +- 8.6; STAI state 60.1 _ 13.7; STAI trait 58.5 _+_ 9.9. The individual RDC diagnosis, Frederick and Hamilton depression scores, and types of psychotropic drugs each patient was receiving at the time of his naloxone stimulation test are shown in Table 1.

Plasma ACTH and Plasma Cortisol When the mean plasma ACTH and cortisol levels in the PTSD group at each sampling point were compared with the corresponding values for the control group, no significant differences were found. Similarly, there were no significant differences between these two groups for ACTH or cortisol regarding mean peak hormone levels, mean peak hormone changes from baseline, or mean areas under the hormone-time curvcs (p > 0.1). However, six of the PTSD patients each had a greater ACTH response to the lowest dose of naloxone (6 Ixg/kg) than any of the other PTSD patients or any of the control subjects (Figure 1). Subgroup analysis was therefore performed, with these six

PTSD patients being designated the hyperresponsive PTSD subgroup (PTSD-H), and the remaining seven PTSD patients the normally responsive PTSD subgroup (PTSD-N). For the integrated area under the ACTH-time curve in response to the 6 Izg/kg naloxone injection, significant differences in variance (Bartlett's test) existed between the control group and the PTSD subjects when analyzed as either a single overall PTSD group (p < 0.01) or as the two PTSD subgroups defined above (p < 0,005). On cluster analysis, the data fit a bimodal dis~bution for the PTSD group alone (p < 0.05, ANOVA) and for all subjects combined (i.e., controls plus PTSD) (p < 0.005, ANOVA), corresponding to the PTSD-H and PTSD-N subgroups as previously defined. There was a significant interaction between subgroup and time on a 2-way ANOVA for both ACTH (F = 1.93, df - 2,23, p < 0.001) and cortisol (F = !.74, df = 2,23, p < 0.005). This is in marked contrast to the corresponding results when the PTSD subjects were analyzed as a single group rather than two subgroups; no significant interactions occurred for ACTH (F = 0.50, df = 1,23, p > 0.9) or cortisol (F = 1.34, df = 1,23, p > 0.1). The remainder of the statistical analysis compares the two P'rSD subgroups with each other and with the control group. Before the first (placebo) injection, the mean basal hormone levels (defined as the mean of the levels at - 15 rain and 0 min) for the control, overall PTSD, PTSD-N, and PTSD-H groups, respectively were 3.2 -+ 2.1, 3.2 -+ 1.8, 3.0 _+ 2.3 and 3.4 +- 1.2 pmol/L (ACTH)

Table !. Results of Clinical and Psychological Assessments, Responses to Low-Dose Naloxone, and Psychotropic Drug Therapy in PTSD Patients Patient no.

Frederick Score °

! 2 3 4 5 6 7 8 9 10 lI 12 13

69 47 65 73 77 64 80 65 69 69 52 76 64

Associated conditions (RDC)b asp none MD; mel anx mda MD MD; mel MD; reel MD; mel cyc; ocd MD; mel lxla MD

Hamilton depression

Naloxone response ~

17 6 23 13 19 20 29 22 25 19 26 20 22

N N H H H H N H

N N N N H

Psychotropic drug classes" benz tea tca; ptz; lith; benz . tea; benz tea; lith; benz tea; benz tea; benz tea; benz tca; benz benz tea; maoi; benz tea; ptz; benz tea; lith; benz

aFrederick index score indicates the severity of PTSD symptoms (see text), with a higher score indicating increased severity. bAssocialed conditions satisfying the Research Diagnostic Criteria (see text) included antisocial personality (asp), major depression (MD), melancholia (mcl), generalized anxiety (anx), minor depression wRh anxiety (mda), cyclothymic personality (cyc), obsessive.compulsive disorder (ocd) and panic disorder with agarophobia (pda). ~N = normal response, H = hyperresponse to 6 I~g/kg naloxono (see text). ~C1~¢~ ~,f psychotropic drugs included Iricyclic antidepressants (tea), phenothiazinas (ptz), monoamine oxidase inhibitors (maoi), lithium carbonate (lith), and benzodiazcph~es (benz).

588

IlIOLPSYCHIATRY

G.I. Hockings et al

1993;33:585-593

Ii

¢

30C

20. 0

~ U

~

100-

"|

-10-

0 0

I> "ID

3

A &

t

o

A

-20

o3

Figure 1. Scatter plot of individual ACTH responses to 6 ttg/kg naloxone in 7 healthy control subjectsand 13 PTSD patients. AUC = integrated area under the ACTH-time curve; A = maximal change in plasma ACTH from basal after naloxone administration (basal taken as plasma ACTH level immediately before the naloxone injection). Brackets indicate an out-of-range point (! 18.6 pmol.min/L).

-2 0 Controls

,, PTSD-N

¢ PTSD-H

and 180 --- 55, 158 -+ 63, 172 -+ 77, and 142 -+ 40 nmol/L (cortisol). These results indicate low and nonstressed pituitary-adrenal axis activity. There were no significant differences in these mean basal hormone levels or in levels at either of the two basal sampling points between the groups (p > 0.1). After the first (placebo) injection, no changes in mean hormone levels occurred within the PTSD-H or PTSD-N subgroups (p > 0.1), whereas mean ACTH and conisol fell within the control group (p < 0.05). There were no significant differences after placebo administration in either hormone level between any two of the groups at any time point (p < 0. I). The mean hormone levels at each sampling point, and significant differences among the three subject groups, are shown in Figure 2. Significant (p < 0.05) differences in ACTH occurred between the PTSD.H and PTSD-N groups, and between the PTSD-H and control groups, after each of the three naloxone injections (although only the response to the first naloxone injection was used to define the subgroups). Significant differences in cortisol occurred between the PTSD-H ~nd control groups after the two larger naloxone injections. No differences between the FI'SD-N and control groups occurred at any time point for either hormone (p > 0.1). The mean integrated areas under the hormone-time curves, the mean peak hormone levels, and mean peak changes from basal after each injection of naloxone are shown in Figures 3 and 4. Significant (p < 0,05) differences were present between the PTSD-H and control groups (for ACTH and cortisol) and between the PTSD-H and PTSD-N groups (for ACTH). Again, no differences between the PTSD-N and control groups occurred for either hormone (/7 > 0.1).

There were no significant correlations between basal plasma ACTH or cortisol and any of the hormonal responses to naloxone (p > 0.1). Additionally, the number of time points at which hormone levels of the PTSD-H subgroup were significantly different from either the PTSDM subgroup or the controls did not change when the ANOVA was repeated using the basal plasma cortisol level as a covariate. There were also no significant correlations between the basal hormone levels and the presence of major depression or the Hamilton Depression score (p > 0.1). However, there was a significant negative correlation between mean basal plasma cortisol and severity of PTSD (r = -0.583, n = 13, p < 0 . 0 5 ) . No subject (control or PTSD) experienced any symptoms, adverse effects, or overt stress response following administration of naloxone or placebo. No significant changes in blood pressure or heart rate occurred in any subject during any of the tests (p > 0.1).

Effects of Associated Psychiatric Conditions, Psychotropic Drugs, and Age on Plasma ACTH and Cortisol Responses to Naloxone The PTSD-H and PTSD-N groups did not differ significantly (p > 0. I) with regard to presence of associated major depression or melancholia. Additionally, no differences in mean peak hormone levels, mean peak changes from basal, or mean integrated areas under the hormonetime curves (hormone response parameters) were found on data reanalysis with the PTSD patients subgrouped on the basis of the presence or absence of each of these two conditions (p > 0.1). Other RDC diagnoses did not occur with sufficient frequency to permit meaningful evaluation (Table 1).

Naloxone Hypersensitivity in PTSD

BIOLPSYCHIATRy

589

1993;33:585-593

15

::: .

, PTSD-H

'

dl \"

10

ft.

0

"

I

I

I

i

"

"

i l l

"

I

.

.

.

.

.

i

.

.

.

.

.

I

.

.

.

.

.

450

l

.

.

.

.

.

J

.

.

.

.

.

I

a

l

l

.

.

.

.

.

*

.

.

.

.

.

*

.

.

.

jo

4OO 35O

-~ 300

250

Figure 2. Mean plasma immunoreactive ACTH (top panel) and cortisol (bottom panel) responses to intravenous injections of placebo (10 ml normal saline, 0 rain) and naloxone (6 I.~g/kg, + 30 rain; 25 Itg/kg, +60 rain; 125 Itg/kg, + 90 rain); these injections are indicated by P, N I, N2, N3, respectively, Items b, c, d indicate points at whichp < 0.05 between groups: b--p < 0,05 for both PTSD-H versus controls and PTSD-H versus PTSD-N;c--p < 0.05 for PTSDH versus controls only; d---p < 0.05 for PTSD-H versus PTSD-N only.

"~o 200 o

150

i

100 5O

-30

0



.

.

.

.

.

P

NI

N2

!

t

t



0

.

.

.

.

.

*

30

.

.

.

.

[

|

N3

r .

60

.

.

.

.

*

.

.

.

90

.

.



120

150

180

TIME (rains)

There were also no differences between the PTSD-H and PTSD-N groups with regard to scores on each of the psychological tests (p > 0.1). No significant correlations occurred between the result of any of the psychological tests and any of the hormone response parameters. There were no significant differences (p > 0.1) between the PTSD-H and FI"SD-N subgroups with regard to classes of psychotropic drugs taken by the individual subjects (Table 1). Also, no effects of drug therapy on hon, one responte parameters were detectable when each class of drug was looked at separately (p > 0. !). Although there was a significant difference in the mean age of the control group and the overall PTSD group (27.0 - 9.8 years cf, 43.6 -.+ 1.8 years, respectively;p < 0.01), there was no such difference in the ages of the two PTSD subgroups (43.6 - 2.4 years for PTSD-N cf. 43.7 ± 0.8

years for PTSD-H; p > 0.1). No significant correlation between age and any of the ACTH or cortisol parameters was found (p > 0.1).

Discussion In this study, the function of the HPA axis was assessed by the hormonal responses to naloxone administration. Animal studies indicate that naloxone acts on the HPA axis solely via hypothalamic CRH release, as preadministration of CRH antiserum abolished naloxone-induced ACTH secretion in rats (Nikolarakis et al 1987) and pituitary stalk transsection abolished naloxone-induced cortisol release in pigs although a full response to exogenous CRH was retained (Estienne et al 1988). This mechanism of action is also supported by the finding that the combined

590

BIOLPSYCHIATRY 1993;33:585-593

225]

2ooi c

175

G.I. Hockings et al

900 BOO 700 :~

co.v=

K'~ PTSD-N ~ PTSD-H

t

"

Z o

i4oo 300 200 100 0

:"

-2 J

,eoo ,500

i

-100 .340

i3oo

75 ©

65

'g

55

i2so ,i220

35

t

1 _5 J

*lr T

AI

A2

A3

A4

il

: 180

Figure 3. Mean integrated areas under the plasma ACTH-time (top panel) and plasma cortisol-time curves following naloxone administration. A1, A2, A4 indicatethe areas afterthe first(6 ~g/kg), second (25 i~g/kg),and third 025 Ixg/kg) injections of naloxone;A3 representsthe total area following the first two injections, and A5 the total area after all three injection. The vertic',dbars indicateSEM. The symbols show degrees of statistical significancebetweenthe indicatedgroup and the PTSD-H group; **p < 0.01; *p < 0.05; t (trend), p < 0.10.

.140

• 100 .60

i

,20

~

,-20

A5

administration of naloxone and exogenous arginine vasopressin (AVP) to normal humans results in a synergistic ACTH response (Hockings et al 1992), similar to that previously reported from combined ovine CRH and AVP administration (DeBold et al 1984). Six of the 13 PTSD patients in this study clearly showed greater individual ACTH responses to low-dose naloxone than the other PTSD patients and the control subjects. Significant differences were consistently present between this hyperresponsive PTSD subgroup and the control group for both ACTH and cortisol, and between the two FrSD subgroups for ACTH. These differences were greatest at the lowest dose of naloxone (6 ~g/kg) and least at the largest dose (125 p,g/kg). This may be because the negative feedback from the greater cortisol responses to the first two naloxone injections in the hyperresponders probably reduced the differences between them and the other subjects in regard to the responses to the final, largest dose of naloxone These results point to a hypersensitivity of the HPA axis to endogenous CRH stimulation in some PTSD patients. In normal humans, naloxone doses less than 1 mg do not alter ACTH or cortisol levels (Tolls et al 1982; Grossman et al 1986). Intermediate doses, such as 25 p,g/kg,

cause nonsignificant rises (Grossman et al 1986), whereas doses of 100-125 p,g/kg have consistently resulted in significant increases (Jackson et al 1990; AI-Damluji et al 1990; Hockings et al 1991). Some studies have used higher doses of 15-20 mg or more, which probably produce a maximal or near maximal response (Volavka et al 1979; Grossman et al 1982; Grossman and ClementJones 1983). Because the present study included two doses at the low end of the dose-response curve, and the tests were performed in the afternoon when basal ACTH and cortisol levels are low, its design was particularly suited to the detection of any increase in the hormone responses to naloxone stimulation. The magnitude of the responses in the normal subjects was similar to those in previously published studies using the same doses (cited above), although greater than in some studies that have given larger doses (e.g., Extein et al 1982; Grossman and Besser 1982). This probably reflects methodological differences, including morning rather than afternoon testing in the latter reports. Our findings in the present study are in contrast to the report of a blunted ACTH response to ovine CRH in six of eight PTSD patients (Smith et al 1989). This may be caused by methodological differences, because a dose-

Naloxone Hypersensitivityin PTSD

18

]

[ - 7 Controls

let

PTSD-N

/

&

121

ntOL PSYCHIATRY

1993;33:585-593

!

§

t

10

!

t

8



t

i

: :. "

~"

2

..l

Figure 4. Mean peak plasma immunoreactive ACTH (top panel) and cortisol (bottom panel) levels and changes following naloxone administration. Items !, 2, and 3 indicate the responsesto the first (6 itg/kg), second (25 ILg/kg),and third (125 i~g/kg) injections of naloxone, respectively. The vertical bars indicate SEM. The symbols show degrees of statistical significancebetweenthe indicated group and the PTSD-H group; **p < 0.01; *p < 0.05; t (trend), p < 0.10.

0

~ i

~

40O

t

300 2OO

591

t

t

1°I 1

2

PEAKS

3

1

2

3

PEAK CHANGES

response protocol was not used in the CRH study, and these patients had recently ceased psychotropic drug therapy; such withdrawal can cause abnormal activity of the HPA axis (Meador-Woodruff and Greden 1988). Also, on subgroup analysis only those subjects with associated major depression showed a statistically significant decrease in the ACTH response to ovine CRH, suggesting that the results of this study may be caused by the effects of depression rather than PTSD. Alternatively, our findings may be compatible with those of Smith et al if a suprapituitary abnormality of HPA axis function occurs in PTSD, as administration of exogenous CRH (unlike naloxone) tests only the pituitary-adrenal component of the axis. The results of the present study must be regarded as preliminary, in view of the small number of subjects stud-

ied, the choice of healthy volunteers as the control group, and the possibility that other factors such as the patients' associated psychiatric conditions may have influenced the results. No statistical correlations were found between the hormonal responses and either comorbid psychiatric disease or severity of FI'SD. However, seven of the PTSD patients tested (including four of the six hyperresponders) also had major depression, and it is well established that HPA axis function is often abnormal in depression tGoid et al 1986; Stokes and Sikes 1988). Naloxone testing has not demonstrated any difference in cortisol responses between depressed patients and normal controls (Judd et al 1981; Extein et al 1982; Zis and Garland 1991). However, in the absence of a control group of depressed patients without PTSD, we cannot exclude the possibility that our

592

BIOLPSYCHIATRY

G.I. Hockings et al

1993;33:585-593

results reflect the effects of depression, rather than PTSD, in some of our patients. All the PTSD patients in this study were receiving psychotropic drugs, which can influence HPA axis function (Meador-Woodmff and Greden 1988). We considered drug withdrawal impractical, both because of the clinical implications (all the patients had severe PTSD) and because psychotropic drug withdrawal itself can cause hyperactivity of the HPA axis for a variable period of time (MeadorWoodruff and Greden 1988). The effect of this drug therapy on the subjects' hormone responses is uncertain, although there is evidence suggesting that this does not provide an adequate explanation for the naloxone hypersensitivity of the PTSD-H subgroup. First, the group of drug-free normal volunteers showed no significant differences in their results from the PTSD-N subgroup, despite the members of this latter group being on multiple psychotropic drugs at the time of their naloxone tests. Second, the two PTSD subgroups were reasonably well matched for the various psychotropic drug classes except for alprazolam and lithium (Table 1), both of which were being taken by a greater proportion of the subgroup showing the hypersensitive responses, although in neither case was the difference statistically significant. Finally, urinary-free cortisol levels in PTSD patients are not altered by psychotropic drug therapy (Mason et al 1986; Yehuda et al 1990), and plasma cortisol concentrations and lymphocyte

glucocorticoid receptor numbers are similar in medicated and drug-free PTSD patients (Yehuda et al 1991). These findings suggest that HPA axis function in PTSD predominantly reflects the effects of the disease itself rather than medication status. However, further studies are needed to determine if hypersensitivity to naloxone also occurs in unmedicated PTSD patients. In conclusion, the results of this study, although clearly preliminary, suggest that there is an increased ACTH response to naloxone, especially at low doses, in some patients with severe, combat-related, treated PTSD. This finding points to an abnormality of HPA axis function, which may be at the suprapituitary level. Larger, more definitive studies employing additional control groups (e.g., depression without PTSD, combat exposure without PTSD) are necessary to confirm these results. This research was supported by a grant from the Central Health and Medical Research Committee of the Department of Veterans' Affairs of the Commonwealth of Australia. Dr. Hockings was supported by the Edwin S. Tooth Research Scholarship of The Universityof Queensland. We wish to thank Dr. Brace Lawfordfor permitting us to include one of his patients in this study, Cannel Hill and Mark Hastings for performing the psychological testing, and Myma Bamfield for her highly competent secretarial assistance. We appreciate the cooperation of the medical, nursing, and secretarial staff of Wards 14 and 41, Greenslopes Hospital. We also thank the patients with PTSD who freely gave of their time to participate in this study.

References AI-Damluji S, l~oulou~, P, White A, Besser M (1990): The role of alpha-2-adret: ,, ~'ceptors in the control of ACTH secretion; interaction with the opioid system. Neuroendocrinology 51:76-81. American Psychiatric Association (1987): Diagnostic and Statistical Manual of Mental Disorders, 3rd ed rev. Washington, DC: American Psychiatric Press. DeBold CR, Sheldon WR, DeChemey GS, et al (1984): Arginine vasopressin potentiates adrenocorticotropin release induced by ovine corticotropin-releasing factor. J Clin Invest 73:533538. Dinan '1"(3, Barry S, Yatham LN, Mobayed M, Brown I (1990): A pilot study of a neuroendocrine test battery in posttraumatic stress disorder. Biol Psychiatry 28:665-672. Estienne MJ, Kesner JS, Barb CR, Kraeling MJ, Rampacek GB (1988): On the site of action of naloxone-stimulated cortisol secretion in gilts. Life Sci 43:161--166. Extein I, Pottash ALC, Gold MS (1982): A possible opioid receptor dysfunction in some depressive disorders. Ann NY Acad $ci 398:113-119. Gold PW, Loriaux DL, Roy A, et al (1986): Responses to corticotropin-releasing hormone in the hypercortisolism of depression and Cushing's disease: Pathophysiologic and diagnostic implications. N Engl J Med 314:1329-1335. Grice JE, Jackson J, Penfold PJ, Jackson RV (1991): Adrenocorlicotmpin hyperresonsiveness in myotonic dystrophy fol-

lowing oral fenfluramine administration. J Neuroendocrinol 3:69-73. Gmssman A, Besser GM (1982): Opiates control ACTH through a noradrenergic mechanism. Clin Endocrinol (Oxf) 17:287290. Grossman A, Clement-Jones V (1983): Opiate receptors: Enkenphalins and endorphins. Clin Endocrinol Metab 12:3156. Gmssman A, Galliard RD, McCartney P, Rees LH, Besser GM (1982): Opiate modulation of the pituitary-adrenal axis: Effects of stress and circadian rhythm. Clin Endocrinol 17:279286.

Gmssman A, Moult PJA, Cunnah D, Besser M (1986): Different opioid receptors are involved in the modulation of ACTH and gnnadotropin release in man. Neuroendocrinology 42:357360. Halbreich U, Olympia J,Carson S, et al (1989): Hypothalamopituitary-edrenal activity in endogenously depressed posttraumatic stress disorder patients. Psychoneuroendocrinology 14:365-370. Hartigan JA (1975): Clustering Al&orithms. New York: Wiley. Helzer JE, Robins LN, McEvoy L (1987): Post-traumatic stress disorder in the general population. Findings of the epidemiologic catchment area survey. N Engl J Med 317:1630--1634. Hockings GI, Grice JE, Waiters MM, Jackson RV (1991): L-

Naloxone Hypersensitivity in PTSD

type calcium channels and CRH-mediated ACTH and cortisol release in humans. Clin Exp Pharmacol Physiol 18:67-71. Hockings GI, Grice JE, Crosbie GV, Waiters MM, Jackson RV (1992): Naloxone and AVP administration to normal humans results in a synergistic ACTH response which is blunted by nifedipine [abstract]. Proceedings of the Ninth International Congress on Endocrinology, Aug 30--Sept51992, Nice, France, p. 84. Jackson RV, Grice JE, Jackson AJ, Hockings GI (1990): Naloxone-induced ACTH release in man is inhibited by clonidine. Clin Exp Pharmacol Physiol 17:179-184. Judd LL, Janowsky DS, Zettner A, Huey LY, Takahashi KI (1981): Effects of naloxone-HCi on cortisol levels in patients with affective disorder and normal controls. Psychiatry Res 4:277-283. Kolb LC (1987): A neuropsychological hypothesis explaining posttraumatic stress disorders. Am J Psychiatry 144:989-995. Kosten TR, Mason JW, Giller EL, Ostroff RB, Harkness L (1987): Sustained urinary norepinephrine and epinephrine elevation in post-traumatic stress disorder. Psychoneuroendo. crinology 12:13-20. Kosten TR, Wahby K, Giller E, Mason J (1990): The dexamethasone suppression test and thyrotropin-releasing hormone stimulation test in post-traumatic stress disorder. Bioi Psychiatry 28:657-664. Kudler H, Davidson J, Meador K, Lipper S, Ely T (1987): The DST and post-traumatic stress disorder. Am J Psychiatry 144:1068-1071. Mason JW, Giller EL, Kosten TR, Ostroff RB, Podd L (1986): Urinary free-cortisol levels in posttraumatic stress disorder patients. J Nerv blent Dis 174:145-149. Meatier-Woodruff JH, Greden JF (1988): Effects of psychotropic medications on hypothalamic-pituitary-adrenal regulation. Endocrinol Metab Clin North Am 17:225-234. Nicholson WE, Davis DR, Sherrel BJ, Orth DN(1984): Rapid radioimmunoassay for corticotrophin in unextracted human plasma. Clin Chem 30:259-265.

BIOLPSYCHIATRY 1993;33:585-593

593

Nikolarakis K, Pfeiffer A, Stalla GK, Herz A (1987): The role of CRH in the release of ACTH by opiate agonists and antagonist in rats. Brain Res 421:373-376. Paige SR, Reid GM, Allen MG, Newton JEO (1990): Psychophysiological correlates of posttraumatic stress disorder in Vietnam veterans. Biol Psychiatry 27:419-430. Pitman RK, Orr SP (1990): Twenty-four hour urinary cortisol and catecholamine excretion in combat-related post-traumatic stress disorder. Biol Psychiatry 27:245-247. Smith MA, Dav~dson J, Ritchie JC, et al (1989): The corticotrnpin-releasing hormone test in patients with posuraumatic stress disorder. Biol Psychiatry 26:349-355. Snedecor GW, Cochran WG (1980): Statistical Methods, 7th ed. Ames Iowa. Iowa State University Press. Spitzer RL, Endicott J, Robins E (1978): Research diagnostic criteria: rationale and reliability. Arch Gen Psychiatry 35:773782. Stokes PE, Sikes CR (1988): The hypothalamic-pituitary-adrenocortical axis in major depression. Endocrinol Metab Clin North Am 17:1-19. Tolls G, Jukier L, Wiesen M, Krieger DT (1982): Effect of naloxone on pituitary hypersecretory syndromes. J Clin Endocrinol M etab 54:780-784. Volavka J, Cho D, Mallya A, Bauman J (1979): Naloxone increases ACTH and cortisol levels in man. N Engl J Med 300; 1056-1057. Yehuda R, Southwick SM, Nussbaum G, Wahby V, Giller EL, Mason JW (1990): Low urinary cortisol excretion in patients with posttraumatic stress disorder. J Nerv Merit Dis 178:366369. Yehuda R, Lowy MT, Southwick SM, Shaffer D, Giller EL, (1991):Lymphocyte glucocorticoid receptor number in posttraumatic stress disorder. Am J Psychiatry 148:499-504. Zis AP, Garland JE (1991): Opioid peptides and depression: The neuroendocrine approach. Balliere's Clin Endocrinol Metab 5:97-117.