BIOL PSYCHIATRY 1990;28:657-664
657
The Dexamethasone Suppression Test and Thyrotropin-Releasing Hormone Stimulation Test in Posttraumatic Stress Disorder* Thomas R. Kosten, Victor Wahby, Earl Giller, Jr, and John Mason
Male veterans with posttraumatic stress disorder (PTSD) (n = 11), including 6 with concurrent major depressive disorder (MDD), were compared to veterans with MDD alone (n = 18) and to 28 controls in their response to the dexamethasone suppression test (DST) and thyrotropin-releasing hormone (TRH) stimdation tests. We found higher levels of 4 eM serum cortisol and lower peak thyroid-stimulating hormone (TSH) response to TRtI in the MDD patients than in either the PTSD patients or controls, in spite of equivalent levels of depression for MDD and PISD. DST suppression (cortisol < 5 mg/dl) occurred h~ 90% of control, 90% of PTSD, and 78% of MDD subjecls, whereas TRH blunting (dTSHm~ < 7 I~Ulml) occurred in 28% of control, 27% of PTSD, and 67% of MDD subjects. Rather than blunting, four PTSD patients (36%) and only 10% of the control and MDD subjects had high TSH responses (13-24 I~U/ml), which may be linked to high noradrenergic activity, since subclinical hypothroidism seemed unlikely.
Introduction Posttraumatic stress disorder (PTSD) is an anxiety disorder with symptoms that can substanti~ly overlap those of a major depressive disorder (MDD) (Rundell et -al. 1986). This overlap is further suggested by high rates of affective disorders in the relatives of patients with PTSD and by response to antidepressant medications (Burnstein 1984; Hogben and Cornfield 1981; Frank et al. 1988). In previous work with these patients we have reported relatively high levels of depressive symptoms, and almost half have met criteria for concurrent MDD (Frank et al. 1988; Mason et al. 1986; Kosten et al. 1987). Although these patients frequently report depressive symptoms, they have demonstrated neuroendocrine alterations that substantially differ from patients wi.th MDD (Mason et al. 1986; Kosten et al. 1987). In particular, a pilot study of PTSD inpatients showed low
From the Departments of Psychiatry, West Haven Veterans Administration Hospital. Yale University School of Medicine, and Research Office, North Chicago Veterans Administration Hospital, and University of Connecticut Health Center, School of Medicine. *Presented in part at the New Research session of the American Psychiatric Association annual meeting, Montreal, Canada, May 1988. Address reprint requests to Thomas R. Kosten, M.D., Yale University, Substance Abuse Treatment Unit, 27 Sylvan Avenue, New Haven. CT 06519. Received October 14, 1988; revised January 13, 1989.
l~blished 1990 by Elsevier Science Publishing Co., Inc.
0006-3223/901500.00
658
BIOL PSYCHIATRY 1990;28:657-664
T.R. Kosten et al.
24-hr urinary cortisol levels, rather than the elevations typically seen in MDD (Mason et al. 1986). This lower cortisol excretion has suggested that the response to dexamethasone administration (DST) in PTSD wo~ld be quite different than in MDD, and the nonsuppression rates in PTSD would be quite low. Three recent studies have examined the DST in PTSD and all found low rates of nonsuppression in patients with PTSD, although one study suggested that DST nonsuppression was higher in PTSD patients with concurrent MDD (Halbreich et al. 1988; Kauffman et al. 1987; Kudler et al. 1987). Thus, DST response in PTSD is controversial and needs further study. Since blunting of the TSH response to thyrotropin-releasing hormone (TRH) has been suggested as a marker for MDD, it would also be of interest in PTSD (Schuckit 1985; Banki et al. 1984). We therefore examined both of these tests, as well as 24-hr urinary free cortisol excretion, in a group of hospitalized male veterans with PTSD and compared their response to those of veterans with MDD and to controls. Challenge testing was repeated in as many patients as possible towards the end of hospitalization in a sick-well design to further examine state-related changes in test results. Methods The all-male sample included 11 patients with PTSD, 18 with MDD, and 28 controls. The 29 patients (42 _+ 11 years) and 28 controls (37 _+ 12 years) were comparable in age and other demographics. Patient diagnoses were made by DSM-III criteria for PTSD and by Research Diagnostic Criteria (RDC) for MDD (Endicott and Spitzer 1978; Spitzer et al. 1978; American Psychiatric Association 1980). All of the MDD patients met endogenous criteria. Symptom levels were assessed using the Hamilton Depression Rating Scale (HDRS). Current substance abusers, patients with major medical illnesses, and those over age 65 were excluded. Six PTSD and six MDD patients were taking medications at the time of initial testing. One PTSD and two MDD patients were receiving tricyclic antidepressants, and four PTSD and four MDD patients were taking neuroleptics. Neuroleptics were given for control of acutely agitated and psychotic behavior at admission. The mean neuroleptic dose in chlorpromazine equivalents was 309 mg (SD = 170) for the medicated PTSD patients and 340 mg (SD = 200) for the medicated MDD patients. Repeat DST and TRH testing was also done 4-6 weeks later (before hospital discharge) in 7 PTSD and 11 MDD patients and in 13 controls. The standard l-mg overnight DST and a 9 AM TRH challenge were performed on different days within the first week of admission to the hospital for the patients and were separated by several days for the outpatient controls. A 24-hr urinary free cortisol was also obtained during the first week on the day prior to the DST. The DST included serum cortisol levels at 4 PM the day following an 11 PM dose of 1 mg dexamethasone (Carroll et al. 1981). The cortisol radioimmunoassay (Clinical Assays, Inc.) has an interassay coefficient of variance (CV) of 4% and a sensitivity of 0.7 Ixg/dl. The TRH challenge used a 500-1~g intravenous bolus of TRH (Hoechst-Roussel) followed by six blood samples over the next 2 hr measuring TSH (Wahby et al. 1988; Schuckit 1985; Banki et al. 1984). Baseline TSH as well as total and free thyroxine were assessed before starting the TRH infusion. Maximal TSH response to TRH, dTSHmax(peak minus baseline), was used for further analyses. Serum thyroxine (1"4) and TSH were measured using radioimmunoassays (Clinical Assays, Inc.) with an interassay CV of 2.2% for T4, 3.2% for free I"4, and 4.3% for TSH. Sensitivity for TSH was 2.1 pdU/ml.
DST and TRH Stimulation Test in PTSD
BIOL PSYCHIATRY 1990:28:657-664
659
Results The results for the initial TRH challenge and the DST are shown in Figure 1 for the three groups. As shown in the top part of Figure 1, the peak TSH response on the TRH challenge was significantly lower in the MDD group (7.1 +_ 0.8 I~IU/ml) than in the PTSD (10.7 +_ 1.8 IdU/ml) and control (C) groups (9.4 _+ 0.8 tdU/ml) (F = 5.2, df = 1,55, p < 0.03, for contrast between MDD versus PTSD and C). Baseline total I"4, free I"4, and TSH levels were equivalent across the three groups, as shown in Table 1. On the DST (middle part of Figure 1), the 4 PM serum cortisol was significantly higher in the MDD group (4.1 _+ 1.8 ttg/dl) than in the PTSD (1.8 _+ 1.0 I~g/dl) and control groups (1.9 _+ 0.6 I~g/dl) (F = 5.3, df = 1,55, p < 0.03, for contrast between MDD versus PTSD and C). Because nine patients were taking medications at the time of challenge testing, we conducted further analyses to adjust for possible medication effects. When analyses were repeated excluding medicated patients, the cortisol levels following the DST (F = 5.1, df = 1,46, p < 0.03) and the TSH levels following the TRH challenge (F = 3.8, df = 1,46, p < 0.06) remained significantly different among the three groups. Although these biological tests significantly distinguished the PTSD from the MDD group, depressive symptom levels were equivalent between the two groups with mean Hamilton depression scores of 25 _+ 3 for the PTSD and of 24 _+ 3 for the MDD group, as shown in the bottom part of Figure 1. Controls reported virtually no depressive symptoms. The 24-hr urinary cortisol levels were significantly higher in the MDD (70 +_ 8 mg) than in the PTSD (50 -+ 5 mg) and control (5.5 _+ 5 mg) groups (F = 4.1, df = 1,55, p < 0.05, for contrast between MDD versus PTSD + C). Our findings were also analyzed dichotomously in terms of published criteria for DST nonsuppression and for blunted TSH response (Carroll et al. 1981; Banki et al. 1984). On the DST we defined nonsuppression as a cortisol level above 5 p,g/dl. As Figure 2 illustrates, a higher rate of nonsuppression was found in the MDD (22%) than in the PTSD (10%) or control (10%) groups, but this difference was not significant. For the TRH challenge we used a response of less than 7 tdU/ml from baseline as indicative of bluntin~ As shown in Figure 2, a higher rate of blunting was found in the MDD (67%) patients than in the PTSD (27%) or control (28%) subjects (X2 = 6.0, df = l, p < 0.05). Four PTSD patients (36%) also had relatively high TSH responses, ranging from 13 to 24 p,U/ml, whereas only three controls (11%) and two MDD patients (11%) had TSH responses greater than 13 ttU/ml (X2 = 4.4, df = 1, p < 0.05). These high TSH responses were not associated with low levels of free T4, however, and the correlations between baseline TSH and I'4 were not significant for either the total sample (r = 0.08, df = 56) or the PTSD patients alone (r = 0.15, df = 10). None of the six PTSD patients with concurrent MDD were DST nonsuppressors, but three of the six (50%) had blunted TStt responses to TRH. Interestingly, two of the four augmented TSH responses also occurred among the PTSD patients with MDD. Detailed individual data are given in Table 2 for the PTSD patients, including DST and TRH testing results, symptom levels on the BPRS and HDRS, medications, and concurrent diagnoses. Two PTSD patients had endogenous depression, no patient had a concurrent anxiety disorder other than IYrSD, and none were current alcoholics or substance abusers. The DST and TRH tests were repeated in a subsample of 7 PTSD and 11 MDD patients and 13 controls. For the patients, repeat testing was done just before discharge in a sickwell design. For the DST, three nonsuppressors in the MDD group returned to a normal DST response, and the mean 4 PM serum cortisol declined by 2.8 - 4.5 I~g/dl (t = 2.1,
660
T.R. Kosten et al.
BIOL P S Y C H I A T R Y
1990;28:657-664
TSH DELTA
15 12 . E
eee ,ee4 ,eeoe, e 4 ,ee eee ,.ee4 e~,e ~41, e 4 e;.e.~e. ,co e • ,!1,
,eel ee, e ~e,oq eee pee,q eee peel eee
'i?it ' ee
DST 4 PM SERUM CORT~30L
m
"1o
o~
0 30
Ld n,,
HAMILTON DEPRESSION SCALE SUM
20
o
¢.3 (,q
10 ~\\\\\\\\ KX\~'%X\XN
PTSD
(n=11)
MDD (E) (n-18)
CONTROl. (n=28)
Figure 1. TSH levels in response to TRH challenge, 4 PM serum cortisol levels during dexamethasone suppression testing, and Hamilton Depression Scale rating among controls and patients with PTSD or major depressive disorder (MDD). PTSD = posttraumatic stress disorder; MDD (E) = major depressive disorder (endogenous).
BIOL PSYCHIATRY 1990;28:657-664
DST and TRH Stimulation Test in PTSD
661
Table 1. Baseline Thyroid Indices in Patients With Major Depressive Disorder (MDD), Pos~aumatic Stress Disorder (PTSD), and Controls (Mean _ SD)
Thyroid ind~-'x
Depressives (n = 18)
Subject Group PTSD (n = 11)
Controls (n = 28)
Total thyroxine (T4I (ttg/dl) Free 1"4 (ng/dl) TSH (l~IU/ml)
8.7 ~ !.4 1.5 - 0.3 3.3 _ 0.7
8.6 = 2.0 1.5 ± 0.4 3.4 _ 0.9
7.8 _ 1.3 1.5 _ 0.3 3.0 ± 1.0
3O
80 Ill=
r~ O o3 of) lad iv" 13. n
DST n,, ,,, a
l ; ~ = TRH TEST ,,
60
....
,
20
IK\\\ ~'~\~
012. z
03
i
%,%,~,%
o3 I Z 0 z
40
,
T
I---
•
/ ~
% ~_N.N ,,\\,
I''" I
20
I" z :3
%.'%,~.N
0
.... MDD (E)
PTSD
0
CONTROL
Figure 2. Percentage of DST nonsuppressors and of blunted TSH response to TRH among controls and patients with PTSD or major depressive disorder (MDD). PTSD = postwaumatic stress disorder; MDD (g) = major depressive disorder (endogenous),
Table 2. Data Summary for Individual Posttraumatic Stress Disorder (PTSD) Patients Patient no.
DST 4 PM TRH response BPRS sum HDRS sum
1
2
3
4
5
6
?
8
9
10
11
2.2 8.6 28 16
3.7 4.7 22 33
1.0 5.7 22 31
1,4 3.0 19 37
1.5 9.4 37 5
1.2 7.8
0.5 14.0 28 29
0.8 13.0 23 23
0.2 18.1 27 23
0.2 9.3 24 30
7.6 24.0 23 17
0 0
AD
Medication°
0
N
0
0
N
N
N
0
0
Concurrent DXb
A
ED
ED
MDD
PS A
MDD
PS
MDD SA
.M~D SA
0
~N, neurolepfic; AD, tricyclicantidepressant; 0, no other medication. ~dDD, major depressive disorder; El), major depressive disorder, et~dogenoussubtype; PS, paranoid schizophrenia;SA, substance able; A, alcoholism;0, no other disorder.
662
BIOL PSYCHIATRY 1990:28:657-664
T.R. Kosten et al.
p > 0.06). The PTSD group showed no significant change ( - 0 . 3 _+ 1.1 Ixg/dl). At discharge the post-DST cortisol levels for neither the MDD nor PTSD groups differed from controls (2.4 Ixg/dl). The peak change in TSH did not change for the PTSD patients ( - 0.4 -4- 3 ixIU/ml), but significantly increased for the MDD pai~en:s (2.3 - 1.6 IzIU/ml) (t = 5.1, p < 0.0003). Although two PTSD patients continued to have blunted TSH responses, none of the MDD patients' responses remained blunted. At discharge the TSH response did not differ from controls (7.8 ixIU/ml). On the Hamilton Depression Scale, the MDD patients generally improved, with a drop in score from 22.7 _ 9.6 to 13.4 +__ 9.7 (t = 3.6, p < 0.006), whereas the PTSD patients dropped from 24.8 + 10.4 to 17 +_ 7.4 (~ = 2.5, p < 0.04). Discussion Recent work by our group has shown that inpatient veterans with PTSD have relatively low urinary cortisol levels and high adrenergic activity coupled with downregulated alphaadrenergic receptors (Mason zt al. 1986; Kosten et al. 1987; Perry et al. 1987). In this study we again found no evidence for increased pituitary adrenal cortical activity in PTSD, as DST suppression rates were similar to controls. Dexamethasone suppression has generally been found in PTSD patients, although a recent study suggested that PTSD patients with concurrent MDD may be nonsuppressors on the DST (I-!albreich et at. 1988; Kauffman et al. 1987; Kudler et al. 1987). We found only one nonsuppressor among our PTSD patients, and the ~st suppresser on the DST, in spite of substantial depressive symptoms. None of the PTSD patients with concurrent MDD were nonsuppressors, including the two with endogenous features, in support of the Halbreich et at. study (1988). Thus, in contrast to patients with MDD, patients witii PTSD appear to have relatively low levels of adrenal cortical activity, even when they met criteria for concurrent MDD. Although our studies have focused on inpatients who may have other clinical features differentiating them from outpatient samples of PTSD patients, Halbreich et al. (1988) have described equivalent results with PTSD outpatients with severe depression. The "normal" response that the PTSD patients had to TRH challenge, rather than the blunted response seen in patients with MDD, provides further support for the diagnostic and perhaps pathophysiological distinction between these two disorders, in spite of substantial symptomatic overlap (Rundell et ai. 1986; Kosten et al. 1987). These results are based on a small sample, however, and differ from the one O~|ler study using the TRH challenge in PTSD. In that study (Kauffman et al. 1987), four of eight PTSD patients had a blunted response, in contrast to our evidence for possibly augmented TRH responses. Although the pathophysiology for TSH blunting in MDD is not clear, a proposed mechanism has been altered pituitary receptor sensitivity and either hypersecretion of dopamine or hyposecretion of norepinephrine (Terry 1984). Some studies have suggested that norepinephrine may act through alpha-adrenergic recepto~ to increase TRH release leading to increased levels of TSH (Nilsson et al. 1974; Yoshimura et al. 1977). Thus, patients with MDD may hyposecrete norepinephrine and thereby have blunted TSH responses to TRH challenge. Applying this reasoning to PTSD patients, the normal response observed in the PTSD patients may indirectly relate to their high adrenerg';c activity (Kosten et al. 1987). "]['herelatively elevated levels of adrenergic activity in PTSD would stimulate TRH and subsequently TSH release. This might counterbalance any increa~¢d release of dopamine associated with the depressive symptoms in these PTSD patients. This speculation would suggest that some PTSD patients with minimal depressive
DST and TRH Stimulation Test in PTSD
BIOLPSYCHIATRY 1990;28:657-664
663
symptoms may have augmented rather than blunted TSH responses to TRH. In fact, four of our PTSD patients showed relatively large TSH responses, ranging from 13 to 24 ttU/ml, although two of them also met criteria for MDD. Further examination of this hypothesis would require careful adjustment for concurrent tfiiodothyronine (T3) and T4 levels during TRH challenge, as thyroxine levels are powerful feedback regulators of TSH levels. The low levels of cortisol in PTSD patients would also fit with this formulation, since high cortisol may also blunt TSH response to TRH (Patel et al. 1972). The alternative explanation for the exaggerated TSH response is subclinical hypothyroidism, but the baseline free T4 levels in these four patients were normal (range = ! .331.62 ng/dl). Furthermore, with hypothyroidism a strong inverse correlation would be expected between 1"4 and TSH, and no significant correlation was found for either the total sample or the PTSD patients alone. Although several treatment studies, including a recent double-blind trial, support the efficacy of antidepressant treatment of PTSD patients, the mechanism of action for these medications in PTSD and MDD are probably not similar (Burnstein 1984; Hogben and Cornfield 1981; Frank et al. 1988). The current study is consistent with several previous ones in suggesting that the biology of PTSD is distinctive and in several ways opposite to that of MDD (Mason et ai. 1986; Kosten et al. 1987; Halbreich et al. 1988; Kauffman et al. 1987). Future work should explore these biological distinctions as keys to understanding traumatic disorders as well as leads for new pharmacological treatment strategies, Supportwas providedby the Veteran~AdministrationResearchFunds,P,e~.,~h ScientistAwa_~_s_.fromA D . ~ A to JM #K05-MH-00346 and to TRK #K02-DA00112.
References American Psychiatric Association (1980): Diagnostic and Statistical Manual, 3rd ed. Washington, DC: American Psychiatric Press. Banki CM, Arato M, Papp Z (1984): Thyroid stimulation test in healthy subjects and psychiatric patients. Acta Psychiatr Scand 70:295-303. Bumstein A (1984): Treatment of post traumatic stress disorder with imipramine. Psychosomatics 25:681-687. Carroll BJ, Feinberg M, Greden JF, Tarika J, Albala AA, Haskett RF, et al (1981): A specific laboratory test for the diagnosis of melancholia. Arch Gen Psychiatry 38:1-15. Endicott J, Spitzer RL (1978): A diagr,Jst;~c interview: The schedule for affective disorders and schizophrenia. Arch Gen Psychiatry 35:837-844. Frank JB, Kosten TR, Gi!ler EL, Dan E (1988): A randomized clinical trial of phenelzine and imipramine for post traumatic stress disorder. Am J Psychiatry 145:1289-1291. Halbreich U, Olympia J, Glogowski J, Carson S, Axelrod S, Yeh CM (1988): The importance of past psychological trauma and pathophysiological process as determinants of current biological abnormalities. Arch Gen Psychiatry 45:293-294. Hogben G, Cornfield R (1981): Treatment of traumatic war neurosis with phenelzine. Arch Gen Psychiatry 38:440-445. Kauffman CD, Reist C, Djenderedjian A, Nelson JN, Haier R.! (1987): Biological markers of affective disorders and posttraumatic stress disorder: A pilot study with desipramine. J Clin Psychiatry 48:366-367. Kosten TR, Mason JW, Giller EL, Ostroff R, Harkness L (1987): Sustained urinary norepinephrine and epinephrine elevations in post traumatic stress disorder. Psychoneuroendocrinology 12:1320.
664
BIOL PSYCHIATRY 1990;28:657-664
T.R. Kosten et al.
Kudler H, Davidson J, Meador K, Lipper S, Ely T (1987): DST and post traumatic stress disorder. Am J Psychiatry 144:1068-1071. Mason JW, Giller EL, Kosten TR, Ostroff R, Podd L (1986): Urinary free cortisGl levels in post traumatic stress disorder. J Nerv Ment Dis 174:145-149. Nilsson KO, Thorell Jl, Mikflet A (1974): Effect of thyrotropih releasing hormone on the release of thyrotropin and other pituitary hormones in man under basal conditions and following adrenergic blocking agents. Acto Endocriaol 76:24--48. Patel YC, Alford FP, Burger HG (1972): 24 hour plasma thyrotropin profile. Clin Sci 43:71-77. Perry BD, Giller EL, Southwick SM (1987): Altered platelet alpha-2-adreaergic receptor binding sites in post-traumatic stress disorder. Am J Pcychiatry 144:1511-1512. Rundell JR, Ursanl RJ, Silberman EK (1986): Posttraumatic stress disorder's relationship of depressive illness. Am J Psychiatry 143:267-268. Schuckit MA (1985): Trait and state markers of a predisposition to psychopathology. In Michels R, Cavenar JO, Brodie HKH, et al (eds), Psychiatry. Philadelphia: Lippincott. Spitzer RL, Endicott J, Robins E (1978): Research diagnostic criteria: Rationale and reliability. Arch Gen Psychiatry 35:773-789. Terry LC (1984): Catecholamine regulation of growth hormone and thyrotropin in mood disorder. In Brown GM, et al (eds), Neuroendocrinology and Psychiatric Disorder. New York: Raven, pp 237-254. Wahby VS, Ibrahim GA, Giller EL, et al (1988): Thyrotropin response to thyrotropin releasing hormone in RDC schizodepressed men. J Affect Disord 15:81-85. Yoshimura M, Hachiya T, Ochi Y, et al (1977): Suppression of elevated s e n ~ TSH levels in hypothyroidism by fusaric acid. J Clin Endocrinol Metab 45:95-98.
657
The Dexamethasone Suppression Test and Thyrotropin-Releasing Hormone Stimulation Test in Posttraumatic Stress Disorder* Thomas R. Kosten, Victor Wahby, Earl Giller, Jr, and John Mason
Male veterans with posttraumatic stress disorder (PTSD) (n = 11), including 6 with concurrent major depressive disorder (MDD), were compared to veterans with MDD alone (n = 18) and to 28 controls in their response to the dexamethasone suppression test (DST) and thyrotropin-releasing hormone (TRH) stimdation tests. We found higher levels of 4 eM serum cortisol and lower peak thyroid-stimulating hormone (TSH) response to TRtI in the MDD patients than in either the PTSD patients or controls, in spite of equivalent levels of depression for MDD and PISD. DST suppression (cortisol < 5 mg/dl) occurred h~ 90% of control, 90% of PTSD, and 78% of MDD subjecls, whereas TRH blunting (dTSHm~ < 7 I~Ulml) occurred in 28% of control, 27% of PTSD, and 67% of MDD subjects. Rather than blunting, four PTSD patients (36%) and only 10% of the control and MDD subjects had high TSH responses (13-24 I~U/ml), which may be linked to high noradrenergic activity, since subclinical hypothroidism seemed unlikely.
Introduction Posttraumatic stress disorder (PTSD) is an anxiety disorder with symptoms that can substanti~ly overlap those of a major depressive disorder (MDD) (Rundell et -al. 1986). This overlap is further suggested by high rates of affective disorders in the relatives of patients with PTSD and by response to antidepressant medications (Burnstein 1984; Hogben and Cornfield 1981; Frank et al. 1988). In previous work with these patients we have reported relatively high levels of depressive symptoms, and almost half have met criteria for concurrent MDD (Frank et al. 1988; Mason et al. 1986; Kosten et al. 1987). Although these patients frequently report depressive symptoms, they have demonstrated neuroendocrine alterations that substantially differ from patients wi.th MDD (Mason et al. 1986; Kosten et al. 1987). In particular, a pilot study of PTSD inpatients showed low
From the Departments of Psychiatry, West Haven Veterans Administration Hospital. Yale University School of Medicine, and Research Office, North Chicago Veterans Administration Hospital, and University of Connecticut Health Center, School of Medicine. *Presented in part at the New Research session of the American Psychiatric Association annual meeting, Montreal, Canada, May 1988. Address reprint requests to Thomas R. Kosten, M.D., Yale University, Substance Abuse Treatment Unit, 27 Sylvan Avenue, New Haven. CT 06519. Received October 14, 1988; revised January 13, 1989.
l~blished 1990 by Elsevier Science Publishing Co., Inc.
0006-3223/901500.00
658
BIOL PSYCHIATRY 1990;28:657-664
T.R. Kosten et al.
24-hr urinary cortisol levels, rather than the elevations typically seen in MDD (Mason et al. 1986). This lower cortisol excretion has suggested that the response to dexamethasone administration (DST) in PTSD wo~ld be quite different than in MDD, and the nonsuppression rates in PTSD would be quite low. Three recent studies have examined the DST in PTSD and all found low rates of nonsuppression in patients with PTSD, although one study suggested that DST nonsuppression was higher in PTSD patients with concurrent MDD (Halbreich et al. 1988; Kauffman et al. 1987; Kudler et al. 1987). Thus, DST response in PTSD is controversial and needs further study. Since blunting of the TSH response to thyrotropin-releasing hormone (TRH) has been suggested as a marker for MDD, it would also be of interest in PTSD (Schuckit 1985; Banki et al. 1984). We therefore examined both of these tests, as well as 24-hr urinary free cortisol excretion, in a group of hospitalized male veterans with PTSD and compared their response to those of veterans with MDD and to controls. Challenge testing was repeated in as many patients as possible towards the end of hospitalization in a sick-well design to further examine state-related changes in test results. Methods The all-male sample included 11 patients with PTSD, 18 with MDD, and 28 controls. The 29 patients (42 _+ 11 years) and 28 controls (37 _+ 12 years) were comparable in age and other demographics. Patient diagnoses were made by DSM-III criteria for PTSD and by Research Diagnostic Criteria (RDC) for MDD (Endicott and Spitzer 1978; Spitzer et al. 1978; American Psychiatric Association 1980). All of the MDD patients met endogenous criteria. Symptom levels were assessed using the Hamilton Depression Rating Scale (HDRS). Current substance abusers, patients with major medical illnesses, and those over age 65 were excluded. Six PTSD and six MDD patients were taking medications at the time of initial testing. One PTSD and two MDD patients were receiving tricyclic antidepressants, and four PTSD and four MDD patients were taking neuroleptics. Neuroleptics were given for control of acutely agitated and psychotic behavior at admission. The mean neuroleptic dose in chlorpromazine equivalents was 309 mg (SD = 170) for the medicated PTSD patients and 340 mg (SD = 200) for the medicated MDD patients. Repeat DST and TRH testing was also done 4-6 weeks later (before hospital discharge) in 7 PTSD and 11 MDD patients and in 13 controls. The standard l-mg overnight DST and a 9 AM TRH challenge were performed on different days within the first week of admission to the hospital for the patients and were separated by several days for the outpatient controls. A 24-hr urinary free cortisol was also obtained during the first week on the day prior to the DST. The DST included serum cortisol levels at 4 PM the day following an 11 PM dose of 1 mg dexamethasone (Carroll et al. 1981). The cortisol radioimmunoassay (Clinical Assays, Inc.) has an interassay coefficient of variance (CV) of 4% and a sensitivity of 0.7 Ixg/dl. The TRH challenge used a 500-1~g intravenous bolus of TRH (Hoechst-Roussel) followed by six blood samples over the next 2 hr measuring TSH (Wahby et al. 1988; Schuckit 1985; Banki et al. 1984). Baseline TSH as well as total and free thyroxine were assessed before starting the TRH infusion. Maximal TSH response to TRH, dTSHmax(peak minus baseline), was used for further analyses. Serum thyroxine (1"4) and TSH were measured using radioimmunoassays (Clinical Assays, Inc.) with an interassay CV of 2.2% for T4, 3.2% for free I"4, and 4.3% for TSH. Sensitivity for TSH was 2.1 pdU/ml.
DST and TRH Stimulation Test in PTSD
BIOL PSYCHIATRY 1990:28:657-664
659
Results The results for the initial TRH challenge and the DST are shown in Figure 1 for the three groups. As shown in the top part of Figure 1, the peak TSH response on the TRH challenge was significantly lower in the MDD group (7.1 +_ 0.8 I~IU/ml) than in the PTSD (10.7 +_ 1.8 IdU/ml) and control (C) groups (9.4 _+ 0.8 tdU/ml) (F = 5.2, df = 1,55, p < 0.03, for contrast between MDD versus PTSD and C). Baseline total I"4, free I"4, and TSH levels were equivalent across the three groups, as shown in Table 1. On the DST (middle part of Figure 1), the 4 PM serum cortisol was significantly higher in the MDD group (4.1 _+ 1.8 ttg/dl) than in the PTSD (1.8 _+ 1.0 I~g/dl) and control groups (1.9 _+ 0.6 I~g/dl) (F = 5.3, df = 1,55, p < 0.03, for contrast between MDD versus PTSD and C). Because nine patients were taking medications at the time of challenge testing, we conducted further analyses to adjust for possible medication effects. When analyses were repeated excluding medicated patients, the cortisol levels following the DST (F = 5.1, df = 1,46, p < 0.03) and the TSH levels following the TRH challenge (F = 3.8, df = 1,46, p < 0.06) remained significantly different among the three groups. Although these biological tests significantly distinguished the PTSD from the MDD group, depressive symptom levels were equivalent between the two groups with mean Hamilton depression scores of 25 _+ 3 for the PTSD and of 24 _+ 3 for the MDD group, as shown in the bottom part of Figure 1. Controls reported virtually no depressive symptoms. The 24-hr urinary cortisol levels were significantly higher in the MDD (70 +_ 8 mg) than in the PTSD (50 -+ 5 mg) and control (5.5 _+ 5 mg) groups (F = 4.1, df = 1,55, p < 0.05, for contrast between MDD versus PTSD + C). Our findings were also analyzed dichotomously in terms of published criteria for DST nonsuppression and for blunted TSH response (Carroll et al. 1981; Banki et al. 1984). On the DST we defined nonsuppression as a cortisol level above 5 p,g/dl. As Figure 2 illustrates, a higher rate of nonsuppression was found in the MDD (22%) than in the PTSD (10%) or control (10%) groups, but this difference was not significant. For the TRH challenge we used a response of less than 7 tdU/ml from baseline as indicative of bluntin~ As shown in Figure 2, a higher rate of blunting was found in the MDD (67%) patients than in the PTSD (27%) or control (28%) subjects (X2 = 6.0, df = l, p < 0.05). Four PTSD patients (36%) also had relatively high TSH responses, ranging from 13 to 24 p,U/ml, whereas only three controls (11%) and two MDD patients (11%) had TSH responses greater than 13 ttU/ml (X2 = 4.4, df = 1, p < 0.05). These high TSH responses were not associated with low levels of free T4, however, and the correlations between baseline TSH and I'4 were not significant for either the total sample (r = 0.08, df = 56) or the PTSD patients alone (r = 0.15, df = 10). None of the six PTSD patients with concurrent MDD were DST nonsuppressors, but three of the six (50%) had blunted TStt responses to TRH. Interestingly, two of the four augmented TSH responses also occurred among the PTSD patients with MDD. Detailed individual data are given in Table 2 for the PTSD patients, including DST and TRH testing results, symptom levels on the BPRS and HDRS, medications, and concurrent diagnoses. Two PTSD patients had endogenous depression, no patient had a concurrent anxiety disorder other than IYrSD, and none were current alcoholics or substance abusers. The DST and TRH tests were repeated in a subsample of 7 PTSD and 11 MDD patients and 13 controls. For the patients, repeat testing was done just before discharge in a sickwell design. For the DST, three nonsuppressors in the MDD group returned to a normal DST response, and the mean 4 PM serum cortisol declined by 2.8 - 4.5 I~g/dl (t = 2.1,
660
T.R. Kosten et al.
BIOL P S Y C H I A T R Y
1990;28:657-664
TSH DELTA
15 12 . E
eee ,ee4 ,eeoe, e 4 ,ee eee ,.ee4 e~,e ~41, e 4 e;.e.~e. ,co e • ,!1,
,eel ee, e ~e,oq eee pee,q eee peel eee
'i?it ' ee
DST 4 PM SERUM CORT~30L
m
"1o
o~
0 30
Ld n,,
HAMILTON DEPRESSION SCALE SUM
20
o
¢.3 (,q
10 ~\\\\\\\\ KX\~'%X\XN
PTSD
(n=11)
MDD (E) (n-18)
CONTROl. (n=28)
Figure 1. TSH levels in response to TRH challenge, 4 PM serum cortisol levels during dexamethasone suppression testing, and Hamilton Depression Scale rating among controls and patients with PTSD or major depressive disorder (MDD). PTSD = posttraumatic stress disorder; MDD (E) = major depressive disorder (endogenous).
BIOL PSYCHIATRY 1990;28:657-664
DST and TRH Stimulation Test in PTSD
661
Table 1. Baseline Thyroid Indices in Patients With Major Depressive Disorder (MDD), Pos~aumatic Stress Disorder (PTSD), and Controls (Mean _ SD)
Thyroid ind~-'x
Depressives (n = 18)
Subject Group PTSD (n = 11)
Controls (n = 28)
Total thyroxine (T4I (ttg/dl) Free 1"4 (ng/dl) TSH (l~IU/ml)
8.7 ~ !.4 1.5 - 0.3 3.3 _ 0.7
8.6 = 2.0 1.5 ± 0.4 3.4 _ 0.9
7.8 _ 1.3 1.5 _ 0.3 3.0 ± 1.0
3O
80 Ill=
r~ O o3 of) lad iv" 13. n
DST n,, ,,, a
l ; ~ = TRH TEST ,,
60
....
,
20
IK\\\ ~'~\~
012. z
03
i
%,%,~,%
o3 I Z 0 z
40
,
T
I---
•
/ ~
% ~_N.N ,,\\,
I''" I
20
I" z :3
%.'%,~.N
0
.... MDD (E)
PTSD
0
CONTROL
Figure 2. Percentage of DST nonsuppressors and of blunted TSH response to TRH among controls and patients with PTSD or major depressive disorder (MDD). PTSD = postwaumatic stress disorder; MDD (g) = major depressive disorder (endogenous),
Table 2. Data Summary for Individual Posttraumatic Stress Disorder (PTSD) Patients Patient no.
DST 4 PM TRH response BPRS sum HDRS sum
1
2
3
4
5
6
?
8
9
10
11
2.2 8.6 28 16
3.7 4.7 22 33
1.0 5.7 22 31
1,4 3.0 19 37
1.5 9.4 37 5
1.2 7.8
0.5 14.0 28 29
0.8 13.0 23 23
0.2 18.1 27 23
0.2 9.3 24 30
7.6 24.0 23 17
0 0
AD
Medication°
0
N
0
0
N
N
N
0
0
Concurrent DXb
A
ED
ED
MDD
PS A
MDD
PS
MDD SA
.M~D SA
0
~N, neurolepfic; AD, tricyclicantidepressant; 0, no other medication. ~dDD, major depressive disorder; El), major depressive disorder, et~dogenoussubtype; PS, paranoid schizophrenia;SA, substance able; A, alcoholism;0, no other disorder.
662
BIOL PSYCHIATRY 1990:28:657-664
T.R. Kosten et al.
p > 0.06). The PTSD group showed no significant change ( - 0 . 3 _+ 1.1 Ixg/dl). At discharge the post-DST cortisol levels for neither the MDD nor PTSD groups differed from controls (2.4 Ixg/dl). The peak change in TSH did not change for the PTSD patients ( - 0.4 -4- 3 ixIU/ml), but significantly increased for the MDD pai~en:s (2.3 - 1.6 IzIU/ml) (t = 5.1, p < 0.0003). Although two PTSD patients continued to have blunted TSH responses, none of the MDD patients' responses remained blunted. At discharge the TSH response did not differ from controls (7.8 ixIU/ml). On the Hamilton Depression Scale, the MDD patients generally improved, with a drop in score from 22.7 _ 9.6 to 13.4 +__ 9.7 (t = 3.6, p < 0.006), whereas the PTSD patients dropped from 24.8 + 10.4 to 17 +_ 7.4 (~ = 2.5, p < 0.04). Discussion Recent work by our group has shown that inpatient veterans with PTSD have relatively low urinary cortisol levels and high adrenergic activity coupled with downregulated alphaadrenergic receptors (Mason zt al. 1986; Kosten et al. 1987; Perry et al. 1987). In this study we again found no evidence for increased pituitary adrenal cortical activity in PTSD, as DST suppression rates were similar to controls. Dexamethasone suppression has generally been found in PTSD patients, although a recent study suggested that PTSD patients with concurrent MDD may be nonsuppressors on the DST (I-!albreich et at. 1988; Kauffman et al. 1987; Kudler et al. 1987). We found only one nonsuppressor among our PTSD patients, and the ~st suppresser on the DST, in spite of substantial depressive symptoms. None of the PTSD patients with concurrent MDD were nonsuppressors, including the two with endogenous features, in support of the Halbreich et at. study (1988). Thus, in contrast to patients with MDD, patients witii PTSD appear to have relatively low levels of adrenal cortical activity, even when they met criteria for concurrent MDD. Although our studies have focused on inpatients who may have other clinical features differentiating them from outpatient samples of PTSD patients, Halbreich et al. (1988) have described equivalent results with PTSD outpatients with severe depression. The "normal" response that the PTSD patients had to TRH challenge, rather than the blunted response seen in patients with MDD, provides further support for the diagnostic and perhaps pathophysiological distinction between these two disorders, in spite of substantial symptomatic overlap (Rundell et ai. 1986; Kosten et al. 1987). These results are based on a small sample, however, and differ from the one O~|ler study using the TRH challenge in PTSD. In that study (Kauffman et al. 1987), four of eight PTSD patients had a blunted response, in contrast to our evidence for possibly augmented TRH responses. Although the pathophysiology for TSH blunting in MDD is not clear, a proposed mechanism has been altered pituitary receptor sensitivity and either hypersecretion of dopamine or hyposecretion of norepinephrine (Terry 1984). Some studies have suggested that norepinephrine may act through alpha-adrenergic recepto~ to increase TRH release leading to increased levels of TSH (Nilsson et al. 1974; Yoshimura et al. 1977). Thus, patients with MDD may hyposecrete norepinephrine and thereby have blunted TSH responses to TRH challenge. Applying this reasoning to PTSD patients, the normal response observed in the PTSD patients may indirectly relate to their high adrenerg';c activity (Kosten et al. 1987). "]['herelatively elevated levels of adrenergic activity in PTSD would stimulate TRH and subsequently TSH release. This might counterbalance any increa~¢d release of dopamine associated with the depressive symptoms in these PTSD patients. This speculation would suggest that some PTSD patients with minimal depressive
DST and TRH Stimulation Test in PTSD
BIOLPSYCHIATRY 1990;28:657-664
663
symptoms may have augmented rather than blunted TSH responses to TRH. In fact, four of our PTSD patients showed relatively large TSH responses, ranging from 13 to 24 ttU/ml, although two of them also met criteria for MDD. Further examination of this hypothesis would require careful adjustment for concurrent tfiiodothyronine (T3) and T4 levels during TRH challenge, as thyroxine levels are powerful feedback regulators of TSH levels. The low levels of cortisol in PTSD patients would also fit with this formulation, since high cortisol may also blunt TSH response to TRH (Patel et al. 1972). The alternative explanation for the exaggerated TSH response is subclinical hypothyroidism, but the baseline free T4 levels in these four patients were normal (range = ! .331.62 ng/dl). Furthermore, with hypothyroidism a strong inverse correlation would be expected between 1"4 and TSH, and no significant correlation was found for either the total sample or the PTSD patients alone. Although several treatment studies, including a recent double-blind trial, support the efficacy of antidepressant treatment of PTSD patients, the mechanism of action for these medications in PTSD and MDD are probably not similar (Burnstein 1984; Hogben and Cornfield 1981; Frank et al. 1988). The current study is consistent with several previous ones in suggesting that the biology of PTSD is distinctive and in several ways opposite to that of MDD (Mason et ai. 1986; Kosten et al. 1987; Halbreich et al. 1988; Kauffman et al. 1987). Future work should explore these biological distinctions as keys to understanding traumatic disorders as well as leads for new pharmacological treatment strategies, Supportwas providedby the Veteran~AdministrationResearchFunds,P,e~.,~h ScientistAwa_~_s_.fromA D . ~ A to JM #K05-MH-00346 and to TRK #K02-DA00112.
References American Psychiatric Association (1980): Diagnostic and Statistical Manual, 3rd ed. Washington, DC: American Psychiatric Press. Banki CM, Arato M, Papp Z (1984): Thyroid stimulation test in healthy subjects and psychiatric patients. Acta Psychiatr Scand 70:295-303. Bumstein A (1984): Treatment of post traumatic stress disorder with imipramine. Psychosomatics 25:681-687. Carroll BJ, Feinberg M, Greden JF, Tarika J, Albala AA, Haskett RF, et al (1981): A specific laboratory test for the diagnosis of melancholia. Arch Gen Psychiatry 38:1-15. Endicott J, Spitzer RL (1978): A diagr,Jst;~c interview: The schedule for affective disorders and schizophrenia. Arch Gen Psychiatry 35:837-844. Frank JB, Kosten TR, Gi!ler EL, Dan E (1988): A randomized clinical trial of phenelzine and imipramine for post traumatic stress disorder. Am J Psychiatry 145:1289-1291. Halbreich U, Olympia J, Glogowski J, Carson S, Axelrod S, Yeh CM (1988): The importance of past psychological trauma and pathophysiological process as determinants of current biological abnormalities. Arch Gen Psychiatry 45:293-294. Hogben G, Cornfield R (1981): Treatment of traumatic war neurosis with phenelzine. Arch Gen Psychiatry 38:440-445. Kauffman CD, Reist C, Djenderedjian A, Nelson JN, Haier R.! (1987): Biological markers of affective disorders and posttraumatic stress disorder: A pilot study with desipramine. J Clin Psychiatry 48:366-367. Kosten TR, Mason JW, Giller EL, Ostroff R, Harkness L (1987): Sustained urinary norepinephrine and epinephrine elevations in post traumatic stress disorder. Psychoneuroendocrinology 12:1320.
664
BIOL PSYCHIATRY 1990;28:657-664
T.R. Kosten et al.
Kudler H, Davidson J, Meador K, Lipper S, Ely T (1987): DST and post traumatic stress disorder. Am J Psychiatry 144:1068-1071. Mason JW, Giller EL, Kosten TR, Ostroff R, Podd L (1986): Urinary free cortisGl levels in post traumatic stress disorder. J Nerv Ment Dis 174:145-149. Nilsson KO, Thorell Jl, Mikflet A (1974): Effect of thyrotropih releasing hormone on the release of thyrotropin and other pituitary hormones in man under basal conditions and following adrenergic blocking agents. Acto Endocriaol 76:24--48. Patel YC, Alford FP, Burger HG (1972): 24 hour plasma thyrotropin profile. Clin Sci 43:71-77. Perry BD, Giller EL, Southwick SM (1987): Altered platelet alpha-2-adreaergic receptor binding sites in post-traumatic stress disorder. Am J Pcychiatry 144:1511-1512. Rundell JR, Ursanl RJ, Silberman EK (1986): Posttraumatic stress disorder's relationship of depressive illness. Am J Psychiatry 143:267-268. Schuckit MA (1985): Trait and state markers of a predisposition to psychopathology. In Michels R, Cavenar JO, Brodie HKH, et al (eds), Psychiatry. Philadelphia: Lippincott. Spitzer RL, Endicott J, Robins E (1978): Research diagnostic criteria: Rationale and reliability. Arch Gen Psychiatry 35:773-789. Terry LC (1984): Catecholamine regulation of growth hormone and thyrotropin in mood disorder. In Brown GM, et al (eds), Neuroendocrinology and Psychiatric Disorder. New York: Raven, pp 237-254. Wahby VS, Ibrahim GA, Giller EL, et al (1988): Thyrotropin response to thyrotropin releasing hormone in RDC schizodepressed men. J Affect Disord 15:81-85. Yoshimura M, Hachiya T, Ochi Y, et al (1977): Suppression of elevated s e n ~ TSH levels in hypothyroidism by fusaric acid. J Clin Endocrinol Metab 45:95-98.
