- Email: [email protected]
HPA-axis abnormalities in psychiatrically well controls
265
Psychiatry Research. 20, 265-273 Elsevier
HPA-Axis
Abnormalities
in Psychiatrically
Well Controls
William
Coryell
and Mark Zimmerman
Received
March 24, 1986; revised version received June 4. 1986; accepted July IO, 1986.
Abstract. The considerable divergence in the literature describing dexamethasone suppression test (DST) nonsuppression rates among psychiatrically well controls led the authors to explore screening procedures as a possible source of variance. Using what they judged to be a typical screening procedure, the authors eliminated 69 of 128 subjects responding to an advertisement seeking psychiatrically and physically healthy subjects. Structured interviews for Axis I and II disorders then revealed that 36% of the remaining 59 subjects nevertheless had historical evidence of psychiatric illness; the 38 who passed all phases of screening underwent three monthly DSTs, which demonstrated the stability over time of both normal and abnormal results. Moreover, a family study revealed that relatives of nonsuppressors had a significantly higher morbid risk of affective disorder (20.5% vs. 8.8%), particularly mania and hypomania.
Key Words. Dexamethasone,
family study, controls,
affective disorder.
In a recent review of studies which described dexamethasone suppression test (DST) results among psychiatrically well controls (submitted for publication), over one half of 53 studies reported nonsuppression rates of 0. Yet, published rates extended as high as 35% nonsuppression (Stokes et al., 1984). Other authors have reported rates which, though only half that reported by Stokes et al. (1984), were still substantially higher than those of the majority; thus, Amsterdam et al. (1982) reported a rate of 15.1% and Hallstrom et al. (1983), a rate of 14.6%. Overall, I I (20.8%) of these studies reported rates exceeding 10%. While Stokes et al. (1985) have suggested the importance of age distribution, few others have explored reasons for these interstudy differences. We designed the following study to do so. Specifically, we wished to address the following hypotheses. First, we suspected that the screening interviews typically used for controls lack appropriate sensitivity to biologically significant affective disorder and that a more thorough and systematic psychiatric screening would isolate this subgroup. This possibility appeared particularly important since recruitment procedures might preferentially attract subjects who have high rates of affective disorder and are therefore driven by curiosity to participate in such studies. Second, we predicted that familial illness, when assessed by relatively sensitive procedures, would discriminate psychiatrically well controls with abnormal DST results from those who suppressed normally. This prediction was counterintuitive in that the DST, when given serially to depressed patients undergoing antidepressant therapy, appears to be a state marker; test results often normalize as depressive
William Coryell, M.D.. is Prol’cssor. and Mark Zimmermnn, B.A.. is Kesearch Assistant. Ilniversity 01 Iowa College of Medicine, Department ol Psychiatry, 500 Newton Rd., lows City. IA 52242, USA. (Keprinl requests to Dr. W. Coryell.) 0165-17Xl/X7/$03.50
@ 19X7 Flscvicr Science I’uhlishcrs
H.V.
266
symptoms improve and, if they do not, the likelihood of relapse is high (Carroll, 1985). Coppen et al. (1983), however, described bipolar patients well controlled on lithium and found that nonsuppression was frequent (33/98 or 33.7% vs. 9/79 or I I .4% for controls, x2 = 12.0, I#‘= 1, p < 0.001) and that DST results were not related to ongoing measures of affective morbidity. This led us to suspect that DST results in some cases might reflect an underlying predisposition to affective disorder that might be revealed in familial psychopathology, if not in the proband’s own psychiatric history. Buchsbaum et al. (1976), in fact, proposed this biochemical high-risk paradigm a decade ago, although they chose platelet monoamine oxidase activity as a probe rather than hypothalamic-pituitary-adrenal (HPA) axis function. Finally, we predicted that multiple dexamethasone suppression tests would increase the sensitivity of abnormal results in patients with major depression without compromising the specificity with which it distinguished patients with major depression from controls. Thus, we wished to extend earlier findings showing that multiple testing procedures enhance the ability to distinguish between primary and secondary depression (Coryell et al., 1983) and between bipolar and unipolar depression (Coryell and Schlesser, 1983). Specifically, we expected this increase in sensitivity to identify “false negative”subjects, (i.e., subjects whose psychiatric illness is missed on routine screening and revealed with a more thorough assessment). Methods Potential subjects responded to an advertisement tion at the University of Iowa which read:
placed in a daily in-house hospital publica-
Wanted: Physically healthy persons who have never had a problem with depression or any other psychiatric difficulties are needed to participate in a study of a blood test for depression. Compensation $85. Of 12X subjects who responded, 46% responded negatively to all questions listed in Table I Positive responses to questions IO and I I were systematically followed by questions eliciting Research Diagnostic Criteria (RDC) symptoms for alcoholism, major depression, minor depression, and chronic intermittent depression. If answers indicated that the subject had ever met criteria for these disorders, he or she was excluded at that point. Any positive answers to question I2 resulted in exclusion whether or not the subject met criteria for a specific disorder. One of us (M.Z.) interviewed each of these 59 subjects with the Schedule for Affective Disorders and Schizophrenia-Lifetime (SADS-L) version (Endicott and Spitxer, 1978) and the Structured Interview for DSM-//I Personality Disorders (SIDP), a semistructured interview with demonstrated reliability (Stangl et al., 1985) designed to elicit clinical information necessary for the diagnosis of all I>SM-//I Axis II disorders. Each subject was also given two self-rating scales for depression: the Carroll Depression Scale (CDS) (Carroll et al., 1981). a self-report scale approximation of the Hamilton Rating Scale (Hamilton, 1967). and the Inventory to Diagnose Depression (IDD), an instrument based on I)SM-///criteria for major depression (Zimmerman et al., 19X6). Finally, subjects were given two 0.5 mg tablets ol dexamethasone with instructions to ingest these at I I p.m. and to return the following day between 2 and 5 p.m. for blood sampling. M.Z. applied Research Diagnostic Criteria (RDC) (Spitter et al., 197X) as well as DSM-III Axis II diagnoses (American Psychiatric Association, 1980) without knowledge of DST results. Those subjects without histories of either RDC or DSM-//I Axis II disorders were invited to repeat the DST after 4 and 8 weeks. Subjects who participated in the structured interviews and the first DST were paid $55; those who went on to receive repeated DSTs received an additional $15 for each DST.
267 Table 1. Screenina auestions Have you lost weight recently? (How much?)
N
2.
Do you ever self-induce vomiting?
N
3.
Do you take birth control pills?
N
4.
Are you pregnant, or did you give birth in the past month?
N
5.
Are you receiving any kind of hormone therapy?
N
6.
Are you taking any medications for a cold, allergies, asthma, Y
N
Have you been physically ill in the last month?
Y
N
8.
Do you have diabetes?
Y
N
9.
Have you ever had a seizure or epilepsy?
Y
N
Y
N
Y
N
Y
N
high blood pressure, nerves, etc?
IF YES: Are you taking medication for this?
Y
N
10.
Was there a period in your life when you drank too much?
11.
Have you ever had a period that lasted at least a week when you felt down,
12.
Have you ever consulted a counselor, psychologist, or psychiatrist
blue, sad, or depressed? about a problem you were having?
Sub_jects who underwent multip!e DST testing were asked after the last test to describe any physical illness they might have experienced during this 8-week period. After being assured that their answers would not affect compensation, subjects were also asked whether they had indeed taken the dexamethasone as instructed for each test. Each subject who passed both initial and intensive interview screening procedures participated in the family study. One of two research assistants contacted all living first degree relatives > age 17 regardless of location and administered the Diagnostic Interview Schedule (DIS) (Robins et al., 1981) in person, if possible, or by telephone if not. Each of these raters had undergone training in the use of the DIS at a week-long course specifically designed for raters in the National Institute of Mental Health (NIMH) Environmental Catchment Area study (Regier et al., 1984). Furthermore, these raters were blind to proband neuroendocrine results and to proband diagnoses as well since a family study of psychiatric probands was ongoing at the same time. Each proband provided information according to the Family History-Research Diagnostic Criteria (FH-RDC) (Endicott et al., 1978) describing each first degree relative. A “best informant” identified by the proband was similarly interviewed. In addition, we sought hospital records for each relative who gave a history of psychiatric hospitalization. A psychiatrist (W.C.) then reviewed all interview and record material to render a final diagnosis. Of 228 first degree relatives, 43 (18.9%)) were not interviewed. Twenty-two of these relatives were dead; nine could not be located; and three could not be interviewed (one was deaf and two were severely demented). Only nine or 4. IV0 of those located refused to participate. Of the I85 successful interviews, 53 (28.6%) were in person and I32 (71.4yo) were by phone. The following data include diagnoses both for interviewed relatives and for those not interviewed (because they refused, were dead, couldn’t be located); thus, family history and family study approaches are combined. lJsing ages of onset for probands with histories of major depression, mania, and hypomania assessed by the NlMH Collaborative Study for Depression -Clinical Branch (Katz et al., 1978), we applied the Stromgren (1935) method for age correction. We used regression analyses to determine correlations between continuous variables and we used Fisher’s exact test for categorical comparisons with four cells since most of these included expected values of less than five. Morbid risk figures were compared according to the method of Breborowicz and Trzebiatowska-Trzeciak (1976). Cortisol was measured by radioimmunoassay using a specific cortisol antiserum obtained from Damon Diagnostics, Needham Heights, MA. There was no cross-reactivity with dexa-
268 methasone, progesterone, testosterone, or I I -deoxycortisol. The intra-and interassay coefficients of variation were 12.2% and I I. IYe, respectively, at concentrations near the cutoff to distinguish suppressors from nonsuppressors. We chose a threshold of 5 ,ug/dl because it provided the optimum separation between depressed patients and a previous sample of control subjects (Zimmerman et al., 1985).
Results Of the 59 subjects who passed initial phone screening, 21 (35.6%) were excluded after the SADS and SIDP interviews. Of these, 12 (18.0% of 59) had a lifetime history of major depression (none were in a current episode); eight had a diagnosis at a definite level, five had received treatment, one had been hospitalized for depression, one had a history of hypomania, and none had a history of mania. Two other subjects had a lifetime diagnosis of minor depression, one of hypomania, one of phobic disorder, one of alcoholism, and one of drug abuse. Another subject met criteria for histrionic personality disorder and two others were excluded because they had sought treatment for “emotional difficulties,” though they did not meet criteria for any Axis 1 or II disorder. The mean (k SD) baseline postdexamethasone cortisol for those 38 subjects who passed screening by structured interviews was 3.1 f 2.7 ng/ ml, and five (13.2%) had values > 5 ng/ ml. The 2 1 subjects who did not pass screening by structured interview had similar values; their mean (+ SD) postdexamethasone cortisol was 2.8 + 1.8, and 14.3% were nonsuppressors. One of these nonsuppressors had no psychiatric diagnosis but had sought counseling; both of the other two had had episodes of major depression which, in one case, had led to treatment. The 59 subjects with baseline dexamethasone suppression tests had mean (k SD) and median ages of 36.2 k 11.2 years and 33 years, respectively, and 55.9% were female. There was little correlation between age and log-transformed postdexamethasone cortisol (r = -0.11, p = 0.4 1). Specifically, the mean (+ SD) postdexamethasone cortisol in the subjects below the younger median age was 3.0 f 2.6, while the older subjects had an identical mean cortisol of 3.0 f 2.1 ng/ ml. None of the eight subjects over age 50 were nonsuppressors. Moreover, neither weight nor the ratio between weight and height correlated with postdexamethasone cortisol; r values were -0.08 and 0.06, respectively. Postdexamethasone cortisols bore no relationship to CDS scores, either in the entire group of 59 subjects (r = 0.07) or at the second or third DST for the 38 never-ill subjects (r = 0.08 and r = 0.06, respectively). According to IDD responses, none of the 59 subjects had major depression at entry to the study, although five had a lifetime diagnosis of major depression; only one of these was a nonsuppressor at baseline. Of 114 (38 X 3) tests possible for subjects who passed all screening phases, DST tt2 was missing for one subject and test #3 was unavailable for another two subjects. Five of the 38 subjects who received multiple tests were nonsuppressors at baseline; all of these were nonsuppressors on at least one of the two subsequent tests (Table 2) while only one of the remaining 33 subjects had a subsequent abnormal value 01 = 0.0000 1, Fisher’s exact test). Thus, with two exceptions, suppressors remained suppressors and nonsuppressors remained nonsuppressors.
269 Two subjects admitted to taking less than the specified amount of dexamethasone on one or more tests, but neither had any postdexamethasone cortisol value indicating nonsuppression. None of the six who had had physical illnesses during the study and only one of the eight who had taken any medication were nonsuppressors at any time.
Table 2. Serial postdexamethasone values amona never-ill controls DST 1
DST 2
cortisol DST 3
14.3
5.0
-
10.6
6.7
8.0
8.1
1.8
10.3
7.0
5.4
7.4
6.5
12.0
11.9
4.6
4.4
2.2
3.2
2.2
2.9
2.8
1.3
1.7
2.7
2.1
2.6
2.6
2.5
3.5
2.6
1.7
2.1
2.5
4.6
2.2
2.4
2.2
2.0
2.4
2.1
3.0
2.4
2.0
2.4
2.3
1.7
3.1
2.3
2.1
1.8
2.2
2.6
2.4
2.1
1.5
2.6
2.1
1.7
3.9
2.1
1.8
2.8
2.1
2.8
2.1
2.0
1.9
1.8
2.0
1.8
2.4
1.9
1.7
1.6
1.9
1.8
2.0
1.9
1.7
1.9
1.8
1.8
1.7
1.8
1.7
1.6
1.8
1.7
2.2
1.7
2.0 -
1.6
1.7 1.6
18.6
2.0 -
1.4
1.6
1.7
1.6
2.0
1.9
1.6
1.7
2.1
1.5
1.8
1.8
1.5
3.2
1.6
DST = dexamethasone suppressiontes!.
270 This relationship between postdexamethasone cortisols measured at widely spaced intervals applied also to subjects with values that were consistently below those defining nonsuppression (Table 3). Thus, postdexamethasone cortisols appear to be a stable trait among psychiatrically well individuals, even within the range generally considered nonpathological. Table 3. Correlations between baseline and subsequent dexamethasone cortisols (log-transformed)
All
subjects
Persistent
suppressors
post-
DST 1 vs. DST 2
DST 1 vs. DST 3
DST 2 vs.
n
r
n
r
n
r
38
0.472
37
0.863
36
0.433
32
0.38’
32
0.43’
31
0.03
DST 3
DST = dexamethasone suppresston test l.p
These results were compared with data from another study (Coryell et al.. 1983) in which depressed outpatients were given DSTs at baseline and at intervals during desipramine therapy (Table 4). Sensitivity to primary depression was, in fact, markedly increased by multiple testing, while specificity-in this case the ability to identify controlsPwas barely compromised by multiple testing of controls. Table 4. Effects of multiple testing on DST sensitivity
Primary
depression’
Controls
(n = 36)
(n = 21)
and specificity
% Nonsuppression on first test
% Nonsuppression on any test
14.3
47.6
13.2
15.8
Specificity
0.87
P
NS
0.84 < 0.01
1 See Coryell et al. (1983).
Relatives of the six controls who were nonsuppressors at any time had higher morbid risks for major depression, mania, and hypomania than did the relatives of controls with consistently normal DST results, although none of these comparisons were statistically significant (Table 5). Group differences were significant when the manic spectrum disorders or when major depression. mania. and hypomania were combined. If those sub.jects with a family history of bipolar spectrum disorder were excluded, the proportion who were nonsuppressors on any test dropped from 15.8vc to 12.1%. With the exclusion of probands with a family history of any affective disorder (major depression. mania, or hypomania). the proportion fell to 8.3% roughly half that 01 the full sample.
271
Table 5. Morbid risk (MR) for affective disorder among relatives of never-ill controls by dexamethasone suppression test response SuppressorsIllness in relatives
all tests (n = 32)
Nonsuppressorsany test (n = 6)
z9P
Major depression # affected # at risk MR %
11
6
147
43
7.5
14.0
z = 1.31
Manic # affected # at risk MR %
1
2
141
41
0.7
4.9
z = 1.85, p < 0.10
Hypomania’ # affected
2
2
# at risk MR %
125 1.6
28 7.1
z = 1.66, p < 0.10
Mania or hypomania It affected # at risk MR %
3
4
142
42
2.1
9.5
z = 2.2,
p < 0.05 Mania, hypomania,
or major
depression # affected # at risk MR %
13
9
147
44
8.8
20.5
z=
2.11,
p < 0.05 1. Only interviewed relatives.
Discussion The initial screening procedure tested here appears to be quite inadequate, although we suspect it is typical of most studies attempting to isolate psychiatrically never-ill controls for biological research. While this lack of sensitivity to lifetime psychiatric illness seems to be of little consequence if the biological measure of interest is a simple outpatient DST, it may be of critical importance to other endeavors, particularly those concerned with alleged trait markers. In our hands, the most commonly used structured interviews for assessing lifetime psychiatric diagnosis, the SADS-L and the DIS, generally take < 45 minutes to complete in nonpatients. We believe this would be time well spent in such studies. These data give more weight to previous assertions (Coryell et al., 1983) that a positive DST result is more meaningful than a negative one when the DST is given under appropriate circumstances; that is, when it is used as a umfirmator~v trst and the true prevalence of endogenous depression in the sample being tested is not inappropriately low (Baldessarini et al., 1983).
272 We were surprised both at the high rate of nonsuppression in such a carefully screened sample and at the stability of postdexamethasone results, which extended even into the nonpathological end of the spectrum. Thus, some substantial proportion of apparently well individuals may exhibit this HPA-axis “abnormality” as a stable characteristic. If so, we are unable to explain why so many studies of normal controls have sampled none of them while other studies have included a relatively large fraction. Nonsuppressors were not distinguished in this study by weight, age, lifetime psychiatric diagnosis, admitted compliance with the procedure, or level of depressive symptoms as measured by the CDS in cross-section. A family history of manic spectrum disorders did distinguish nonsuppressors from suppressors, although only with borderline statistical significance. Attempts at replication should be undertaken since the identification of a familial marker for bipolar spectrum disorder would be of unquestionable importance both to genetic research and, if the test proved to have predictive power (nonsuppressors subsequently developed affective disorder), to preventive medicine. Attempts at replication would do well to focus more carefully on hypomania since this disorder is often missed (Depue et al., 1981) or diagnosed unreliably (Andreasen et al., 1981). The use of methods developed specifically for the detection of bipolar spectrum disorders (Depue et al., 198 I) might, when applied to relatives, more clearly distinguish psychiatrically well probands with and without HPA-axis abnormalities. In any case, replication of our findings should be followed by longitudinal studies to determine whether these HPA-axis abnormalities have predictive significance. References American
Psychiatric Association. DSM-III: Diagnostic. and Statistical Manual of’ Mental 3rd ed. APA, Washington, DC (1980). Amsterdam, J.D., Winokur, A., Caroff, SM., and Conn, J. The dexamethasone suppres-
Disorders.
affective disorder and healthy control subjects. American
sion test in outpatients
with primary
Journal of’ Psychiatry,
139, 287 ( 1982).
Andreasen, McDonald-Scott,
N., Grove, W.M., Shapiro, P. Reliability
R.W., Keller, M.B., of lifetime diagnosis. Archives
Hirschfeld,
R.M.A.,
of General Ps_vchiatry,
and
38,400
(1981). Baldessarini, R.J., Finkelstein, S., and Arana, G.W. The predictive power of diagnostic tests and the effect of prevalence of illness. Archives oJ’ General Psychiatry. 40,569 (1983). Breborowicz, G., and Trzebiatowska-Trzeciak, 0. A method for testing differences in morbidity risk for affective psychoses. Acta Psychiatrica Scandinavica, 54, 353 (1976). Buchsbaum, M.S., Coursey, R.D., and Murphy, D.L. The biochemical high-risk paradigm: Behavioral and familial correlates of low platelet monoamine oxidase activity. Science. 194, 339 ( 1976). Carroll, B.J. Dexamethasone suppression test: A review of contemporary confusion. Jour-
nal of’ Clinical Psychiatry.
46, I3 ( 1985).
Carroll, B.J., Feinberg, M., Smouse, P.E., Rawson, S.G., and Greden, J.F. The Carroll Rating Scale for Depression: I. Development, reliability and validation. British Journal of Psychiatry. 138, I94 ( I98 I ). Coppen, A., Abou-Saleh, M., Milln, P., Metcalfe, M., Harwood, J., and Bailey, J. Dexamethasone suppression test in depression and other psychiatric illness. British Journal of
Psychiatry,
142, 498 (I 983).
Coryell, W., and Schlesser, M.A. Dexamethasone depression: Stability across hospitalizations. Psychiatry
suppression
test response
Research. 8, 179 (1983).
in major
273 Coryell, W., Smith, R., Cook, B., Moucharafieh, S., Dunner, F., and House, D. Serial dexamethasone suppression test results during antidepressant therapy: Relationship to diagnosis and clinical change. fsychia(ry Research, 10, I65 (1983). Depue, R.A., Slater, J.F., Wolfstetter-Kausch, H., Klein, D., Goplerud, E., and Fart-, D. A behavioral paradigm for identifying persons at risk for bipolar depressive disorders: A conceptual framework and five validation studies. Journal of Abnormal Psychology, 90,381 (1981). Endicott, J., Andreasen, N., and Spitzer, R.L. Family History Research Diagnostic Criteria. 3rd ed. New York State Psychiatric Institute, New York (1978). Endicott, J., and Spitzer, R.L. A diagnostic interview: The Schedule for Affective Disorders and Schizophrenia. Archives of General Psychiatry, 35, 837 (1978). Hallstrom, T., Samuelsson, S., Balldin, J., Walinder, J., Bengtsson, C., Nystrom, E., Andersch, B., Lindstedt, G., and Lundberg, P. Abnormal dexamethasone suppression test in normal females. British Journal of Psychiatry, 142,489 (1983). Hamilton, M. Development of a rating scale for primary depressive illness. British Journal qf’Social and Clinical Psychology, 6, 278 (1967). Katz, M.M., Secunda, S.K., Hirschfeld, R.M.A., and Koslow, S.H. NIM H Clinical Branch Research Collaborative Program on the Psychobiology of Depression. Archives of General Psychiatry, 36, 765 (1979). Pfohl, B., Stangl, D., and Zimmerman, M. The implications of DSM-III personality disorders for patients with major depression. Journal of Affective Disorders, 7,309 (1984). Regier, D.A., Myers, J.K., Kramer, M., Robins, L.N., Blazer, D.G., Hough, R.L., Eaton, W.W., and Locke, B.Z. ‘The NIMH Epidemiologic Catchment Program. Archives of General Psychiatry, 41,934 (I 984). Robins, L.N., Helzer, J.E., Croughan, J., Williams, J.B.W., and Spitzer, R.L. National Institute of Mental Health Diagnostic Interview Schedule. Archives of General Psychiatry, 38, 381 (1981). Spitzer, R.L., Endicott, J., and Robins, E. Research Diagnostic Criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773 (1978). Stangl, D., Pfohl, B., Zimmerman, M., Bowers, W., and Corenthal, C. A structured interview for the DSM-III personality disorders. Archives of General Psychiatry, 42,591 (1985). Stokes, P.E., Stall, P.M., Koslow, S.H., Maas, J.W., Davis, J.M., Swarm, A.C., and Robins, E. Pretreatment DST and the hypothalamic-pituitary-adrenocortical function in depressed patients and comparison groups. Archives of General Psychiatry, 41, 257 (1984). Stokes, P.E., Stall, P.M., Koslow, S.H., Maas, J.W., Davis, J.M., Swarm, A.C., and Robins, E. Limited utility of the I mg dexamethasone suppression test as a measure of hypercortisolism. Archives of General Psychiatry, 42,202 (1985). Stromgren, E. Zum ersatz des Weinbergachen Abyekerzten Verfahrens. Zeitschrtft Gesamte fiir Neurologie und fsychiatrie. 153,784 (1935). Zimmerman, M., Coryell, W., Corenthal, C., and Wilson, S. A self-report scale to diagnose major depressive disorder. Archives of General Psychiatry, 43, 1075 (1986). Zimmerman, M., Stangl, D., and Coryell, W. The research diagnostic criteria for endogenous depression and the dexamethasone suppression test: A discriminant function analysis. Psychiatry Research, 14, I97 ( 1985).
Psychiatry Research. 20, 265-273 Elsevier
HPA-Axis
Abnormalities
in Psychiatrically
Well Controls
William
Coryell
and Mark Zimmerman
Received
March 24, 1986; revised version received June 4. 1986; accepted July IO, 1986.
Abstract. The considerable divergence in the literature describing dexamethasone suppression test (DST) nonsuppression rates among psychiatrically well controls led the authors to explore screening procedures as a possible source of variance. Using what they judged to be a typical screening procedure, the authors eliminated 69 of 128 subjects responding to an advertisement seeking psychiatrically and physically healthy subjects. Structured interviews for Axis I and II disorders then revealed that 36% of the remaining 59 subjects nevertheless had historical evidence of psychiatric illness; the 38 who passed all phases of screening underwent three monthly DSTs, which demonstrated the stability over time of both normal and abnormal results. Moreover, a family study revealed that relatives of nonsuppressors had a significantly higher morbid risk of affective disorder (20.5% vs. 8.8%), particularly mania and hypomania.
Key Words. Dexamethasone,
family study, controls,
affective disorder.
In a recent review of studies which described dexamethasone suppression test (DST) results among psychiatrically well controls (submitted for publication), over one half of 53 studies reported nonsuppression rates of 0. Yet, published rates extended as high as 35% nonsuppression (Stokes et al., 1984). Other authors have reported rates which, though only half that reported by Stokes et al. (1984), were still substantially higher than those of the majority; thus, Amsterdam et al. (1982) reported a rate of 15.1% and Hallstrom et al. (1983), a rate of 14.6%. Overall, I I (20.8%) of these studies reported rates exceeding 10%. While Stokes et al. (1985) have suggested the importance of age distribution, few others have explored reasons for these interstudy differences. We designed the following study to do so. Specifically, we wished to address the following hypotheses. First, we suspected that the screening interviews typically used for controls lack appropriate sensitivity to biologically significant affective disorder and that a more thorough and systematic psychiatric screening would isolate this subgroup. This possibility appeared particularly important since recruitment procedures might preferentially attract subjects who have high rates of affective disorder and are therefore driven by curiosity to participate in such studies. Second, we predicted that familial illness, when assessed by relatively sensitive procedures, would discriminate psychiatrically well controls with abnormal DST results from those who suppressed normally. This prediction was counterintuitive in that the DST, when given serially to depressed patients undergoing antidepressant therapy, appears to be a state marker; test results often normalize as depressive
William Coryell, M.D.. is Prol’cssor. and Mark Zimmermnn, B.A.. is Kesearch Assistant. Ilniversity 01 Iowa College of Medicine, Department ol Psychiatry, 500 Newton Rd., lows City. IA 52242, USA. (Keprinl requests to Dr. W. Coryell.) 0165-17Xl/X7/$03.50
@ 19X7 Flscvicr Science I’uhlishcrs
H.V.
266
symptoms improve and, if they do not, the likelihood of relapse is high (Carroll, 1985). Coppen et al. (1983), however, described bipolar patients well controlled on lithium and found that nonsuppression was frequent (33/98 or 33.7% vs. 9/79 or I I .4% for controls, x2 = 12.0, I#‘= 1, p < 0.001) and that DST results were not related to ongoing measures of affective morbidity. This led us to suspect that DST results in some cases might reflect an underlying predisposition to affective disorder that might be revealed in familial psychopathology, if not in the proband’s own psychiatric history. Buchsbaum et al. (1976), in fact, proposed this biochemical high-risk paradigm a decade ago, although they chose platelet monoamine oxidase activity as a probe rather than hypothalamic-pituitary-adrenal (HPA) axis function. Finally, we predicted that multiple dexamethasone suppression tests would increase the sensitivity of abnormal results in patients with major depression without compromising the specificity with which it distinguished patients with major depression from controls. Thus, we wished to extend earlier findings showing that multiple testing procedures enhance the ability to distinguish between primary and secondary depression (Coryell et al., 1983) and between bipolar and unipolar depression (Coryell and Schlesser, 1983). Specifically, we expected this increase in sensitivity to identify “false negative”subjects, (i.e., subjects whose psychiatric illness is missed on routine screening and revealed with a more thorough assessment). Methods Potential subjects responded to an advertisement tion at the University of Iowa which read:
placed in a daily in-house hospital publica-
Wanted: Physically healthy persons who have never had a problem with depression or any other psychiatric difficulties are needed to participate in a study of a blood test for depression. Compensation $85. Of 12X subjects who responded, 46% responded negatively to all questions listed in Table I Positive responses to questions IO and I I were systematically followed by questions eliciting Research Diagnostic Criteria (RDC) symptoms for alcoholism, major depression, minor depression, and chronic intermittent depression. If answers indicated that the subject had ever met criteria for these disorders, he or she was excluded at that point. Any positive answers to question I2 resulted in exclusion whether or not the subject met criteria for a specific disorder. One of us (M.Z.) interviewed each of these 59 subjects with the Schedule for Affective Disorders and Schizophrenia-Lifetime (SADS-L) version (Endicott and Spitxer, 1978) and the Structured Interview for DSM-//I Personality Disorders (SIDP), a semistructured interview with demonstrated reliability (Stangl et al., 1985) designed to elicit clinical information necessary for the diagnosis of all I>SM-//I Axis II disorders. Each subject was also given two self-rating scales for depression: the Carroll Depression Scale (CDS) (Carroll et al., 1981). a self-report scale approximation of the Hamilton Rating Scale (Hamilton, 1967). and the Inventory to Diagnose Depression (IDD), an instrument based on I)SM-///criteria for major depression (Zimmerman et al., 19X6). Finally, subjects were given two 0.5 mg tablets ol dexamethasone with instructions to ingest these at I I p.m. and to return the following day between 2 and 5 p.m. for blood sampling. M.Z. applied Research Diagnostic Criteria (RDC) (Spitter et al., 197X) as well as DSM-III Axis II diagnoses (American Psychiatric Association, 1980) without knowledge of DST results. Those subjects without histories of either RDC or DSM-//I Axis II disorders were invited to repeat the DST after 4 and 8 weeks. Subjects who participated in the structured interviews and the first DST were paid $55; those who went on to receive repeated DSTs received an additional $15 for each DST.
267 Table 1. Screenina auestions Have you lost weight recently? (How much?)
N
2.
Do you ever self-induce vomiting?
N
3.
Do you take birth control pills?
N
4.
Are you pregnant, or did you give birth in the past month?
N
5.
Are you receiving any kind of hormone therapy?
N
6.
Are you taking any medications for a cold, allergies, asthma, Y
N
Have you been physically ill in the last month?
Y
N
8.
Do you have diabetes?
Y
N
9.
Have you ever had a seizure or epilepsy?
Y
N
Y
N
Y
N
Y
N
high blood pressure, nerves, etc?
IF YES: Are you taking medication for this?
Y
N
10.
Was there a period in your life when you drank too much?
11.
Have you ever had a period that lasted at least a week when you felt down,
12.
Have you ever consulted a counselor, psychologist, or psychiatrist
blue, sad, or depressed? about a problem you were having?
Sub_jects who underwent multip!e DST testing were asked after the last test to describe any physical illness they might have experienced during this 8-week period. After being assured that their answers would not affect compensation, subjects were also asked whether they had indeed taken the dexamethasone as instructed for each test. Each subject who passed both initial and intensive interview screening procedures participated in the family study. One of two research assistants contacted all living first degree relatives > age 17 regardless of location and administered the Diagnostic Interview Schedule (DIS) (Robins et al., 1981) in person, if possible, or by telephone if not. Each of these raters had undergone training in the use of the DIS at a week-long course specifically designed for raters in the National Institute of Mental Health (NIMH) Environmental Catchment Area study (Regier et al., 1984). Furthermore, these raters were blind to proband neuroendocrine results and to proband diagnoses as well since a family study of psychiatric probands was ongoing at the same time. Each proband provided information according to the Family History-Research Diagnostic Criteria (FH-RDC) (Endicott et al., 1978) describing each first degree relative. A “best informant” identified by the proband was similarly interviewed. In addition, we sought hospital records for each relative who gave a history of psychiatric hospitalization. A psychiatrist (W.C.) then reviewed all interview and record material to render a final diagnosis. Of 228 first degree relatives, 43 (18.9%)) were not interviewed. Twenty-two of these relatives were dead; nine could not be located; and three could not be interviewed (one was deaf and two were severely demented). Only nine or 4. IV0 of those located refused to participate. Of the I85 successful interviews, 53 (28.6%) were in person and I32 (71.4yo) were by phone. The following data include diagnoses both for interviewed relatives and for those not interviewed (because they refused, were dead, couldn’t be located); thus, family history and family study approaches are combined. lJsing ages of onset for probands with histories of major depression, mania, and hypomania assessed by the NlMH Collaborative Study for Depression -Clinical Branch (Katz et al., 1978), we applied the Stromgren (1935) method for age correction. We used regression analyses to determine correlations between continuous variables and we used Fisher’s exact test for categorical comparisons with four cells since most of these included expected values of less than five. Morbid risk figures were compared according to the method of Breborowicz and Trzebiatowska-Trzeciak (1976). Cortisol was measured by radioimmunoassay using a specific cortisol antiserum obtained from Damon Diagnostics, Needham Heights, MA. There was no cross-reactivity with dexa-
268 methasone, progesterone, testosterone, or I I -deoxycortisol. The intra-and interassay coefficients of variation were 12.2% and I I. IYe, respectively, at concentrations near the cutoff to distinguish suppressors from nonsuppressors. We chose a threshold of 5 ,ug/dl because it provided the optimum separation between depressed patients and a previous sample of control subjects (Zimmerman et al., 1985).
Results Of the 59 subjects who passed initial phone screening, 21 (35.6%) were excluded after the SADS and SIDP interviews. Of these, 12 (18.0% of 59) had a lifetime history of major depression (none were in a current episode); eight had a diagnosis at a definite level, five had received treatment, one had been hospitalized for depression, one had a history of hypomania, and none had a history of mania. Two other subjects had a lifetime diagnosis of minor depression, one of hypomania, one of phobic disorder, one of alcoholism, and one of drug abuse. Another subject met criteria for histrionic personality disorder and two others were excluded because they had sought treatment for “emotional difficulties,” though they did not meet criteria for any Axis 1 or II disorder. The mean (k SD) baseline postdexamethasone cortisol for those 38 subjects who passed screening by structured interviews was 3.1 f 2.7 ng/ ml, and five (13.2%) had values > 5 ng/ ml. The 2 1 subjects who did not pass screening by structured interview had similar values; their mean (+ SD) postdexamethasone cortisol was 2.8 + 1.8, and 14.3% were nonsuppressors. One of these nonsuppressors had no psychiatric diagnosis but had sought counseling; both of the other two had had episodes of major depression which, in one case, had led to treatment. The 59 subjects with baseline dexamethasone suppression tests had mean (k SD) and median ages of 36.2 k 11.2 years and 33 years, respectively, and 55.9% were female. There was little correlation between age and log-transformed postdexamethasone cortisol (r = -0.11, p = 0.4 1). Specifically, the mean (+ SD) postdexamethasone cortisol in the subjects below the younger median age was 3.0 f 2.6, while the older subjects had an identical mean cortisol of 3.0 f 2.1 ng/ ml. None of the eight subjects over age 50 were nonsuppressors. Moreover, neither weight nor the ratio between weight and height correlated with postdexamethasone cortisol; r values were -0.08 and 0.06, respectively. Postdexamethasone cortisols bore no relationship to CDS scores, either in the entire group of 59 subjects (r = 0.07) or at the second or third DST for the 38 never-ill subjects (r = 0.08 and r = 0.06, respectively). According to IDD responses, none of the 59 subjects had major depression at entry to the study, although five had a lifetime diagnosis of major depression; only one of these was a nonsuppressor at baseline. Of 114 (38 X 3) tests possible for subjects who passed all screening phases, DST tt2 was missing for one subject and test #3 was unavailable for another two subjects. Five of the 38 subjects who received multiple tests were nonsuppressors at baseline; all of these were nonsuppressors on at least one of the two subsequent tests (Table 2) while only one of the remaining 33 subjects had a subsequent abnormal value 01 = 0.0000 1, Fisher’s exact test). Thus, with two exceptions, suppressors remained suppressors and nonsuppressors remained nonsuppressors.
269 Two subjects admitted to taking less than the specified amount of dexamethasone on one or more tests, but neither had any postdexamethasone cortisol value indicating nonsuppression. None of the six who had had physical illnesses during the study and only one of the eight who had taken any medication were nonsuppressors at any time.
Table 2. Serial postdexamethasone values amona never-ill controls DST 1
DST 2
cortisol DST 3
14.3
5.0
-
10.6
6.7
8.0
8.1
1.8
10.3
7.0
5.4
7.4
6.5
12.0
11.9
4.6
4.4
2.2
3.2
2.2
2.9
2.8
1.3
1.7
2.7
2.1
2.6
2.6
2.5
3.5
2.6
1.7
2.1
2.5
4.6
2.2
2.4
2.2
2.0
2.4
2.1
3.0
2.4
2.0
2.4
2.3
1.7
3.1
2.3
2.1
1.8
2.2
2.6
2.4
2.1
1.5
2.6
2.1
1.7
3.9
2.1
1.8
2.8
2.1
2.8
2.1
2.0
1.9
1.8
2.0
1.8
2.4
1.9
1.7
1.6
1.9
1.8
2.0
1.9
1.7
1.9
1.8
1.8
1.7
1.8
1.7
1.6
1.8
1.7
2.2
1.7
2.0 -
1.6
1.7 1.6
18.6
2.0 -
1.4
1.6
1.7
1.6
2.0
1.9
1.6
1.7
2.1
1.5
1.8
1.8
1.5
3.2
1.6
DST = dexamethasone suppressiontes!.
270 This relationship between postdexamethasone cortisols measured at widely spaced intervals applied also to subjects with values that were consistently below those defining nonsuppression (Table 3). Thus, postdexamethasone cortisols appear to be a stable trait among psychiatrically well individuals, even within the range generally considered nonpathological. Table 3. Correlations between baseline and subsequent dexamethasone cortisols (log-transformed)
All
subjects
Persistent
suppressors
post-
DST 1 vs. DST 2
DST 1 vs. DST 3
DST 2 vs.
n
r
n
r
n
r
38
0.472
37
0.863
36
0.433
32
0.38’
32
0.43’
31
0.03
DST 3
DST = dexamethasone suppresston test l.p
These results were compared with data from another study (Coryell et al.. 1983) in which depressed outpatients were given DSTs at baseline and at intervals during desipramine therapy (Table 4). Sensitivity to primary depression was, in fact, markedly increased by multiple testing, while specificity-in this case the ability to identify controlsPwas barely compromised by multiple testing of controls. Table 4. Effects of multiple testing on DST sensitivity
Primary
depression’
Controls
(n = 36)
(n = 21)
and specificity
% Nonsuppression on first test
% Nonsuppression on any test
14.3
47.6
13.2
15.8
Specificity
0.87
P
NS
0.84 < 0.01
1 See Coryell et al. (1983).
Relatives of the six controls who were nonsuppressors at any time had higher morbid risks for major depression, mania, and hypomania than did the relatives of controls with consistently normal DST results, although none of these comparisons were statistically significant (Table 5). Group differences were significant when the manic spectrum disorders or when major depression. mania. and hypomania were combined. If those sub.jects with a family history of bipolar spectrum disorder were excluded, the proportion who were nonsuppressors on any test dropped from 15.8vc to 12.1%. With the exclusion of probands with a family history of any affective disorder (major depression. mania, or hypomania). the proportion fell to 8.3% roughly half that 01 the full sample.
271
Table 5. Morbid risk (MR) for affective disorder among relatives of never-ill controls by dexamethasone suppression test response SuppressorsIllness in relatives
all tests (n = 32)
Nonsuppressorsany test (n = 6)
z9P
Major depression # affected # at risk MR %
11
6
147
43
7.5
14.0
z = 1.31
Manic # affected # at risk MR %
1
2
141
41
0.7
4.9
z = 1.85, p < 0.10
Hypomania’ # affected
2
2
# at risk MR %
125 1.6
28 7.1
z = 1.66, p < 0.10
Mania or hypomania It affected # at risk MR %
3
4
142
42
2.1
9.5
z = 2.2,
p < 0.05 Mania, hypomania,
or major
depression # affected # at risk MR %
13
9
147
44
8.8
20.5
z=
2.11,
p < 0.05 1. Only interviewed relatives.
Discussion The initial screening procedure tested here appears to be quite inadequate, although we suspect it is typical of most studies attempting to isolate psychiatrically never-ill controls for biological research. While this lack of sensitivity to lifetime psychiatric illness seems to be of little consequence if the biological measure of interest is a simple outpatient DST, it may be of critical importance to other endeavors, particularly those concerned with alleged trait markers. In our hands, the most commonly used structured interviews for assessing lifetime psychiatric diagnosis, the SADS-L and the DIS, generally take < 45 minutes to complete in nonpatients. We believe this would be time well spent in such studies. These data give more weight to previous assertions (Coryell et al., 1983) that a positive DST result is more meaningful than a negative one when the DST is given under appropriate circumstances; that is, when it is used as a umfirmator~v trst and the true prevalence of endogenous depression in the sample being tested is not inappropriately low (Baldessarini et al., 1983).
272 We were surprised both at the high rate of nonsuppression in such a carefully screened sample and at the stability of postdexamethasone results, which extended even into the nonpathological end of the spectrum. Thus, some substantial proportion of apparently well individuals may exhibit this HPA-axis “abnormality” as a stable characteristic. If so, we are unable to explain why so many studies of normal controls have sampled none of them while other studies have included a relatively large fraction. Nonsuppressors were not distinguished in this study by weight, age, lifetime psychiatric diagnosis, admitted compliance with the procedure, or level of depressive symptoms as measured by the CDS in cross-section. A family history of manic spectrum disorders did distinguish nonsuppressors from suppressors, although only with borderline statistical significance. Attempts at replication should be undertaken since the identification of a familial marker for bipolar spectrum disorder would be of unquestionable importance both to genetic research and, if the test proved to have predictive power (nonsuppressors subsequently developed affective disorder), to preventive medicine. Attempts at replication would do well to focus more carefully on hypomania since this disorder is often missed (Depue et al., 1981) or diagnosed unreliably (Andreasen et al., 1981). The use of methods developed specifically for the detection of bipolar spectrum disorders (Depue et al., 198 I) might, when applied to relatives, more clearly distinguish psychiatrically well probands with and without HPA-axis abnormalities. In any case, replication of our findings should be followed by longitudinal studies to determine whether these HPA-axis abnormalities have predictive significance. References American
Psychiatric Association. DSM-III: Diagnostic. and Statistical Manual of’ Mental 3rd ed. APA, Washington, DC (1980). Amsterdam, J.D., Winokur, A., Caroff, SM., and Conn, J. The dexamethasone suppres-
Disorders.
affective disorder and healthy control subjects. American
sion test in outpatients
with primary
Journal of’ Psychiatry,
139, 287 ( 1982).
Andreasen, McDonald-Scott,
N., Grove, W.M., Shapiro, P. Reliability
R.W., Keller, M.B., of lifetime diagnosis. Archives
Hirschfeld,
R.M.A.,
of General Ps_vchiatry,
and
38,400
(1981). Baldessarini, R.J., Finkelstein, S., and Arana, G.W. The predictive power of diagnostic tests and the effect of prevalence of illness. Archives oJ’ General Psychiatry. 40,569 (1983). Breborowicz, G., and Trzebiatowska-Trzeciak, 0. A method for testing differences in morbidity risk for affective psychoses. Acta Psychiatrica Scandinavica, 54, 353 (1976). Buchsbaum, M.S., Coursey, R.D., and Murphy, D.L. The biochemical high-risk paradigm: Behavioral and familial correlates of low platelet monoamine oxidase activity. Science. 194, 339 ( 1976). Carroll, B.J. Dexamethasone suppression test: A review of contemporary confusion. Jour-
nal of’ Clinical Psychiatry.
46, I3 ( 1985).
Carroll, B.J., Feinberg, M., Smouse, P.E., Rawson, S.G., and Greden, J.F. The Carroll Rating Scale for Depression: I. Development, reliability and validation. British Journal of Psychiatry. 138, I94 ( I98 I ). Coppen, A., Abou-Saleh, M., Milln, P., Metcalfe, M., Harwood, J., and Bailey, J. Dexamethasone suppression test in depression and other psychiatric illness. British Journal of
Psychiatry,
142, 498 (I 983).
Coryell, W., and Schlesser, M.A. Dexamethasone depression: Stability across hospitalizations. Psychiatry
suppression
test response
Research. 8, 179 (1983).
in major
273 Coryell, W., Smith, R., Cook, B., Moucharafieh, S., Dunner, F., and House, D. Serial dexamethasone suppression test results during antidepressant therapy: Relationship to diagnosis and clinical change. fsychia(ry Research, 10, I65 (1983). Depue, R.A., Slater, J.F., Wolfstetter-Kausch, H., Klein, D., Goplerud, E., and Fart-, D. A behavioral paradigm for identifying persons at risk for bipolar depressive disorders: A conceptual framework and five validation studies. Journal of Abnormal Psychology, 90,381 (1981). Endicott, J., Andreasen, N., and Spitzer, R.L. Family History Research Diagnostic Criteria. 3rd ed. New York State Psychiatric Institute, New York (1978). Endicott, J., and Spitzer, R.L. A diagnostic interview: The Schedule for Affective Disorders and Schizophrenia. Archives of General Psychiatry, 35, 837 (1978). Hallstrom, T., Samuelsson, S., Balldin, J., Walinder, J., Bengtsson, C., Nystrom, E., Andersch, B., Lindstedt, G., and Lundberg, P. Abnormal dexamethasone suppression test in normal females. British Journal of Psychiatry, 142,489 (1983). Hamilton, M. Development of a rating scale for primary depressive illness. British Journal qf’Social and Clinical Psychology, 6, 278 (1967). Katz, M.M., Secunda, S.K., Hirschfeld, R.M.A., and Koslow, S.H. NIM H Clinical Branch Research Collaborative Program on the Psychobiology of Depression. Archives of General Psychiatry, 36, 765 (1979). Pfohl, B., Stangl, D., and Zimmerman, M. The implications of DSM-III personality disorders for patients with major depression. Journal of Affective Disorders, 7,309 (1984). Regier, D.A., Myers, J.K., Kramer, M., Robins, L.N., Blazer, D.G., Hough, R.L., Eaton, W.W., and Locke, B.Z. ‘The NIMH Epidemiologic Catchment Program. Archives of General Psychiatry, 41,934 (I 984). Robins, L.N., Helzer, J.E., Croughan, J., Williams, J.B.W., and Spitzer, R.L. National Institute of Mental Health Diagnostic Interview Schedule. Archives of General Psychiatry, 38, 381 (1981). Spitzer, R.L., Endicott, J., and Robins, E. Research Diagnostic Criteria: Rationale and reliability. Archives of General Psychiatry, 35, 773 (1978). Stangl, D., Pfohl, B., Zimmerman, M., Bowers, W., and Corenthal, C. A structured interview for the DSM-III personality disorders. Archives of General Psychiatry, 42,591 (1985). Stokes, P.E., Stall, P.M., Koslow, S.H., Maas, J.W., Davis, J.M., Swarm, A.C., and Robins, E. Pretreatment DST and the hypothalamic-pituitary-adrenocortical function in depressed patients and comparison groups. Archives of General Psychiatry, 41, 257 (1984). Stokes, P.E., Stall, P.M., Koslow, S.H., Maas, J.W., Davis, J.M., Swarm, A.C., and Robins, E. Limited utility of the I mg dexamethasone suppression test as a measure of hypercortisolism. Archives of General Psychiatry, 42,202 (1985). Stromgren, E. Zum ersatz des Weinbergachen Abyekerzten Verfahrens. Zeitschrtft Gesamte fiir Neurologie und fsychiatrie. 153,784 (1935). Zimmerman, M., Coryell, W., Corenthal, C., and Wilson, S. A self-report scale to diagnose major depressive disorder. Archives of General Psychiatry, 43, 1075 (1986). Zimmerman, M., Stangl, D., and Coryell, W. The research diagnostic criteria for endogenous depression and the dexamethasone suppression test: A discriminant function analysis. Psychiatry Research, 14, I97 ( 1985).