- Email: [email protected]
Serum thyroid hormone abnormalities in psychiatric disease
Serum Thyroid Hormone Abnormalities Richard P. Levy, Jonathan
B. Jensen,
in Psychiatric
Disease
Victor G. Laus, David P. Agle, and llona M. Engel
Psychiatric illness is a cause of “euthyroid sick syndrome” (ESS), defined as abnormal concentrations of circulating iodothyronines in euthyroid subjects with nonthyroidal illness (NTI). We describe a prospective study of 150 consecutive psychiatric admissions studied by endocrine and psychologic techniques. Based on 150 admission blood samples, we found a 7% incidence of ESS and with serial samples (74 patients) the incidence was 27%. demonstrating that ESS can develop after hospital admission. Of the 20 patients with ESS, 11 had elevation of both serum total T4 concentrations (T4) and free thyroxine index (FTI) while their serum total T3 concentrations (T3) remained normal; 5 had elevation of FTI without elevation of T4 or T3; and 4 had low T4 and low FTI and normal TSH. In 2 of the 4 patients in the last category, the T3 was also low. The free T3 index (FT311 was normal in all but 1 patient who had low FT3l and FTI, low T4 and T3, and normal TSH. The serum thyroid hormone abnormalities were transient in the ESS patients during the 10 day period with 2 exceptions; 1 patient had persistently elevated T4 and FTI with normal T3 and FT3l values while another patient had persistently depressed T4 and FTI without abnormality of FT3l or TSH. Multivariate statistical analysis demonstrated a difference (P -C .06) in the psychologic attributes of somatic and autonomic symptoms in ESS patients compared to controls. We conclude that ESS is as common amongst psychiatric admissions as in general hospital patients previously studied and that blood thyroid function tests should be interpreted cautiously in all hospitalized patients.
I
T HAS BEEN demonstrated that a variety of acute and chronic disorders can alter serum concentrations of free thyroid hormones without causing overt hyperthyroidism or hypothyroidism. The term euthyroid sick syndrome (ESS) has been used to describe patients who are clinically euthyroid, ill with nonthyroidal illness (NTI), who have abnormal thyroid function tests; i.e., a low serum T3 concentration (T3) and an elevated reverse T3 (rT3),‘.2 a low free thyroxine index (FTI) not accompanied by an elevated concentration of TSH,’ or alternatively a high FTI not accompanied by an elevated blood T3.3,4 In a preliminary study of the incidence of the syndrome in a university hospital we noted ESS in several patients in whom no obvious cause other than psychiatric illness was found. A series of serum samples from 98 consecutive new psychiatric admissions (Series I) was then studied and 9 patients with ESS were detected using the criteria and laboratory techniques described later. McLarty’ also noted abnormal T3 and serum total T4 concentrations (T4)
From the Department of Psychiatry and Thyroid Center, University Hospitals of Cleveland and Case Western Reserve University. Cleveland. Ohio. Dr. Jensen is now with the Division of Chiid and Adolescent Psychiatry. University of Minnesota Hospitals. Dr. Levy is now Director of the Division of Medicine, and Professor of Internal Medicine, Ajiliated Hospitals at Canton. Northeastern Ohio Universities College of Medicine. Presented in part at the 4th Great Lakes Regional Endocrine Society Meeting, Rochester, New York, 1980. Received for publication December 8, 1980. Supported in part bv teaching grant #ITO-IMH-15022-01 PI from the National Institute of Mental Health. Address reprint requests to Dr. Richard P. Levy, Division of Medicine, Timken Mercy Medical Center, 1320 Timken Mercy Drive NW, Canton, Ohio 44708. ii 1981 by Grune & Stratton, Inc. 0026-0495/81/301 I-0004$0l.00/0
1060
without thyroid disease in a large psychiatric inpatient population and Cohen and Swigar reported findings consistent with ESS in 9% of newly admitted psychiatric patients.4 To identify the psychic factors correlated with ESS, a retrospective but blind analysis of Series I comparing the hospital charts of the nine ESS patients and concurrent control cases was conducted by the psychiatrists (JBJ and DPA), evaluating the presence of general motor activity, catabolic stress, life change, and anxiety in each patient. Though no parameters demonstrated statistical significance, a trend was noted implicating an association between the ESS and inability to handle anxiety effectively. Accordingly, a second sample of patients (Series 11) admitted to a psychiatric hospital was defined using a more sophisticated evaluation of anxiety. Serial thyroid function tests were done. The data to be reported from Series II demonstrated that ESS often begins during psychiatric hospitalization and that there is a statistically significant relationship between ESS and certain manifestations of anxiety. MATERIALS
AND
METHODS
Patients All patients admitted on six days of the week to the adult psychiatry service at the University Hospitals of Cleveland from December, 1977, through May, 1978, were included in the study. (Saturday was excluded since admission blood samples could not be obtained within 24 hr.) The hospital population serves a wide range of socioeconomic groups and draws from the greater metropolitan area of Cleveland, Ohio. The patients included all diagnostic categories and all degrees of emotional disturbance common to psychiatric inpatient units. The endocrinologist (RPL) determined which patients had ESS. The names of patients with ESS and controls matched for age, sex, and race were sent to two other investigators, a psychiatrist (JBJ) and a psychologist (IME). After patient consent
Metabolism, Vol. 30. No.
11
(November). 198 1
1061
THYROID HORMONES AND PSYCHIATRIC DISEASE
was obtained, the psychiatrist and psychologist, ignorant of whether the patient was ESS or control, conducted the psychological measurement within 72 hr of admission. After the interviews were scored, each patient’s chart was reviewed for medications, concurrent medical illness, and history of thyroid disease. Patients were excluded when medications or illness that are known to affect the thyroid function tests were found and patients with known thyroid disease were analyzed separately. Within 3 days of the psychologic examination the endocrinologist interviewed and examined each patient and the medical record for evidence of thyroid disease and concurrent medical illness.
Psychological
Measurements
The psychiatrist and psychologist administered the Buss Anxiety Rating Scale, as modified by Schalling,‘.8 and the patient completed a State-Trait Anxiety Inventory (STAI)9 which was scored mechanically
Laboratory
MeaSurements
A second blood sample was obtained immediately after the STAI was completed, and 3 days and 7 days afterwards. All 4 serum samples were frozen promptly, stored at -20°C and then ran simultaneously in the same analytic run. Tt. owing tests were done: serum total T4 and T3 concentration\ were measured by specific radioimmunoassays developed in our laboratory by modifying Larsen’s techniques. ‘O.” The “resin” T3 uptake test (RT3U) was done by a talc uptake method.” In this method the radioactivity in the supernatant rather than in the talc is measured. It is calculated as the ratio of the counts per minute (cpm) of the normal serum pool over those of the sample cpm, normal pool cpm. sample
= RT3U
thus yielding values below normal when TBG is increased. The circulating free T4 and T3 were estimated from their respective indices; i.e., the free T4 index (FTI) was the product of total T4 value and the RT3U. This follows the principles published by Bermuder. Surks and Oppenheimer.’ The 95% confidence limits (“normal” values) of these tests are: total T4 = 5--12 ug/dl, total T3 = 80-200 ng/dl. RT3U = 0.80--l. 18,FTI = 5-I 2, Free T3 index (FT31) = 80-200, TSH =
Patients who had been separated into the ESS group and a control group were compared on I2 anxiety measurements and 4 covariates by a multivariate analysis of covariance. The anxiety measures included the 10 combined scores from both raters from the Buss Anxiety Rating Scale, and the State/Trait Anxiety scores from the Spielberger STAI. The four covariates were dummy variables indicating the presence or absence of four classes of drugs. The drugs were classified as major tranquilizers (chlorpromazine and haloperidol), minor tranquilizers (diazepam and chlordiazepoxide). antidepressants (imipramine and amitriptyline) and sedatives (barbiturates). The interaction effects were pooled with the error term, because they were not statistically significant. The statistical package used for the analysis of variance was “Multivariance.“”
RESULTS Endocrine
In Series I, 34 males and 64 females, there were 4 men and 5 women with ESS, an incidence of 9%. Six of the ESS patients had high T4, high FTI and normal T3 and FT31; 3 had low T4, low FTI and normal T3, FT3I and TSH. Two patients were frankly hypothyroid (low T4 and FTI and high TSH). None were hyperthyroid. Of the 175 consecutive admissions comprising Series II, an initial admission blood sample was obtained within 24 hr on 55 males and 95 females, or 85% of the total population admitted. The age range was 14.-78-yr-old. Timely blood samples could not be obtained from the remainder because admission samples were not drawn within 24 hr or because the patient had been transferred to the psychiatric unit from another part of the hospital. Based on the admission blood sample, 4 men and 6 women of the 150 subjects were found to have an abnormal FTI without other evidence of thyroid disease, a 7% incidence of ESS. This value is not statistically significantly different from the 9% incidence found in the preliminary study of psychiatric inpatients at this hospital (Series I). It is also identical to the 7%’ incidence of ESS in the general University Hospitals population.‘J Seventy-four patients without previously known thyroid disease consented to the full experimental protocol, including all of those found to have ESS on admission, and IO more were found to develop ESS during hospitalization, an incidence of 27% in the 74 patients. Ten other patients were found to have thyroid disease which included 4 cases of nontoxic diffuse goiter, 4 cases of hypothyroidism, and 1 case each of spontaneously resolving hyperthyroidism and lithium induced goiter. By exclusion. no patient had concurrent confounding nonthyroidal, medical illness. Of the 20 patients with ESS (Table 1). 1 I had elevation of both T4 and FTI while their T3 remained normal; 5 had elevation of FTI without elevation of T4 or T3; and 4 had low T4 and low FTI and normal TSH. In 2 of the 4 patients in the last category. the T3 was also low. The FT31 was normal in all but I patient who had low FT31 and FTI. low T4 and T3. and normal TSH. We analyzed the variance of’ the data in Table I ;\s the ratio of between subjects vs. within subjects. The ratio ranged from 2.6-6.3 for the various laboratory tests done, indicating the discriminating power of the tests. Correlation (r) of the day-to-day values for each of the 3 thyroid function tests ranges from 0.75 0.95. T4 and T3 values are also found to bc well correlated with each other on all 4 study days with r between 0.50 and
T4.
12.6
12.6
M
F
M
F
M
F
M
M
F
M
M
ID
11
12
13
14
15
16
17
18
19
20
155
135
90
145
12.0
11.5
5.3
10.5
.29
5.3-
.72-
1.30
5.0-
13.2
1.00 1.23 1.06 ,030
12.3 4.9 Il.1 4.8 10.6 .85
141 85 161 156
1.30
*Samples unavailable.
IOC175
11.9
2.9
128
6.9
126
131
165
75
145
105
135
165
145
13.8
4.8-
.28
7.8
.84
9.7
3.8
13.3
3.8
10.0
5.0
15.2
11.2
10.5
9.5
10.5
11.4
13.9
11.9
Day6
11.9
5.7-
.26
6.2
.80
10.1
4.4
14.9
4.6
9.3
5.4
14.1
11.0
11.8
13.4
11.4
11.6
13.1
11.4
10.7
13.5
7.8
12.9
10.5
2.4
8.7
FTI
184
65-
4.2
121
5.2
109
116
120
60
115
loo
150
105
90
100
115
125
110
140
125
126
80
140
105
70
90
ng/dl
T3.
TSH = < lOuU/ml.
1.30
.77-
,020
1.07
,028
1.07
1.16
1.12
1.21
0.93
1.08
0.93
0.98
,.I2
1.41
1.08
1.02
0.94
0.96
1.10
1.09
12.4 9.7
1.00
0.85
1.13
1.20
1.05
RT3U
7.8
15.2
9.3
2.0
8.3
uQ/dl
T4.
ng/dl, FT31 = 80-200,
172
9%
3.2
126
7.2
131
161
165
92
?44
112
139
153
181
122
140
130 100
170
95
132
133
148
SO
165
115
160
120
145
80
105
173
160 135
69
114
FT31
65
100
nQ/dl
T3.
5.1-
.27
8.8
.96
11.1
5.9
11.1
6.0
12.2
4.8
11.5
13.2
20.0
14.6
12.7
13.0
10.8
11.6
12.2
14.5
10.3
12.2
11.6
2.8
11.2
FTI
Day3
1.18, FTI = 5- 12, T3 = 80-200
14.6
184
4.5-
175
.73-
.29
82-
,020
8.7
4.0
1.01
0.99
126
7.1
137
1.22
4.5
95 1.07
11.2 1.03
0.93
14.2
139
143
1.22 1.25
16.0
1.08
1.03
0.63
0.88
1.11
1.02
1.00
0.78
12.0
115 168
90-
3.5
122
7.9
130
131
135
70
95
11.8
158
95% confidence limits: T4 = 5- 12 ug/dl, RTBU = 0.80-
12.0
8.6
.29
1.04
.71
,022
,030
.91
5.7
10.8
10.1
1.19
1.19
6.8
1.21
0.97
1.06
1.03
1.08
6.4
Range
150
11.5
10.5
1.21
1.05
0.92
165
12.0
0.96
12.6
170
165
13.0
13.0
108
1.03
13.2
168
125
13.7
0.86
195
12.8
0.86
Il.0
142
12.7
1.16
10.3
98
120
175
13.1
0.99
15.7
120 14.2
95
10.7
189 1.08
1.22
2.3
147 79
RT3U 1.14
T4. UQ/dl 9.8
FT31
173
135
12.6
160
13.3
65
2.8
12.7
0.89
4.8
*SE
T3.
DQ/dl
100
1
14.0
FTI
Day
1.03
I.18
1.22
1.47
RTBU
10.4
9.5
5.6
10.8
5.0
11.2
13.0
10.0
Mean
Controls In - 441
15.9
9.5
14.9
F
10.8
M
7
M
13.2
F
6
9
12.3
6
14.2
F
F
11.3
M
3
5
2.3
F
2
4
9.5
UQ/dl
F
sex
1
No.
gs
Table 1, Thyroid Function Test Results in Series II
194
7%
4.1
129
4.8
115
135
134
72
107
140
103
101
141
125
128
103
134
138
136
80
119
119
84
94
FT31
1.3
1.8
1.5
"Uhl
TSH,
12.0
4.8-
.29
7.6
.69
9.9
5.0
4.3 .
9.2
.
11.0
13.8
11.5
8.5
10.4
9.0
14.4
11.3
10.4
8.6
6.8
13.5
11.6
.
s.e
UQ/dl
T4.
0.92
1.33
.70-
,021
1.o*
,039
1.06
1.34
1.14 .
0.92
0.91 .
0.92
1.03
1.41
1.03
0.91
0.66
11.2
5.5-
.25
8.0
.60
10.2
6.7
4.9 .
8.5
10.0 *
12.7
11.8
12.0
10.7
8.2
12.4
10.4
11.9
9.1
1.06 1.14
6.9
12.8
13.3
11.1 .
FTI
Day10
1.01
0.95
1.15
1.26 .
RTJU
176
80-
3.7
116
4.9
Ill
123
SO .
120
145 .
125
110
100
135
105
107
145
115
86
80
105
120
nQ/df 95 .
73.
.
165
81-
3.3
123
5.7
118
165
91 .
110
132 .
133
113
141
139
96
92
133
132
90
82
100
138
120
FT31
THYROID
HORMONES
AND
PSYCHIATRIC
1063
DISEASE
0.66, thus showing that these two tests were not completely independent variables. RT3U values however, did not correlate with the other two tests. The T4 and FTI values were statistically significantly different (p c .025) when controls were compared with ESS patients on each study day, but neither T3 nor RT3U values differed significantly. These relationships between the thyroid function tests were the expected ones. The serum thyroid hormone abnormalities were transient in the ESS patients during the 10 day period with 2 exceptions: 1 patient had persistently elevated T4 and FTI with normal T3 and FT3I values while another patient had persistently depressed T4 and FTI without abnormality of FT31 or TSH. As noted previously, 10 patients had ESS on admission. Of the remainder with normal admission thyroid function tests, 6 had abnormal results in the second blood sample and in the remaining 4 the abnormal result was evident by the third blood sample. In 10 patients only I of the 4 blood samples was abnormal, in 6 patients 2 samples were abnormal, in 3 patients 3 samples were abnormal, and all 4 samples were abnormal in I patient Psychological Because the psychological data were incomplete in 9 patients and I patient was found to be taking phenytoin, only 64 patients remained for the multivariate analysis of psychological factors and presence of ESS. Of the 64 patients 18 were ESS and 46 served as controls. The results of the STAl were dropped after they were found not to distinguish ESS from control patients. The multivariate analysis showed that the mean vectors of ESS and control groups were different at the p : .06 level (Table 2). The greatest contributor Table 2.
Relationship
Between
Psychologic Variables
and
Euthyroid Sick Syndrome Univariate Vanable Somatic
Global
-0.69
5.36
0.02
-0.62
4.39
0.04
restlessness
(0)
10.50
2.87
0.09
1.56
0.21
-0.24
0.64
0.42
io.22
0.56
0.45
-0.18
0.37
0.54 0.64
anxety
(R1
0.37
(01 (0)
Disquietude
(R1
(RI symptoms
(0)
-0.13
0.22
restlessness
(R)
-0.12
0.16
0.68
-0.10
0.13
0.71
Autonomvz Muscular
Dlstractablllty
(R)
Multwariate
analysis.
Multivariate
F =
*ASick, verse
P.c
(Rj
Distractabllity
Worry
F
Ml = 1 and 56)
symptoms
symptoms
Autonomic Muscular
ASack, well*
well:
control
R-Reported. O-Observed.
n = 64,df
=
10a
to this difference was somatic symptoms* as reported by the patient. The other important contributor was autonomic symptoms? as reported by the patient. These two items discriminated at the p < .02 and p d .04 level respectively. As seen in Table 2, all other factors had minor discriminating ability. After the means were adjusted for the effects of the covariates. the ESS group reported significantly less somatic symptoms and autonomic symptoms compared to the control group. DISCUSSION
That a variety of NT1 alters the concentrations of circulating iodothyronines is clear; the underlying mechanisms for these changes and their significance is much less clear. While a decrease in production of T3 from T4 has been well demonstrated, recent studies leave unclear whether or not there are changes in the concentration of the circulating free T4. Chopra and coworkers found that different methods of estimating serum free thyroxine which agree well in the absence of NT1 gave conflicting results in patients with a variety of nonthyroidal illnesses.” These investigators found that in NTI the FTI correlated better with a direct RIA for free T4 than with the classical equilibrium dialysis method. In any case our work, as well as that of Cohen and Swigar, adds psychiatric illness to the already long list of NTl’s that can confound the interpretation of thyroid function. We have also demonstrated that. similar to other illnesses, ESS can occur within days in patients with psychiatric illness. Cohen and Swigar’s work seems to have deficiencies: the FTI method they used is an older one utilizing a dextran column for separating bound and free T4 and the outdated PBI method for estimating total T4, making their findings hard to compare with those of Chopra, et al., cited above, and with ours. They did not measure T3. They report a remarkably high number of psychiatric patients with “secondary hypothyroidism,” 8 out of the 12 with thyroid disease. Furthermore, in none of the 8 patients was the diagnosis confirmed by appropriate laboratory investigation. In our patients the abnormalities tended to be only moderately deviant from normal, tended more to be elevation of T4 and FTI rather than depression and were usually transient. Thus the physician will almost always be able to avoid misdiagnosis in these patients by repeating the abnormal thyroid function tests. Cohen and Swigar reported a 9% incidence of abnor-
nd 47, P < 0.06
1.90
Adjusted
expressed
estimate in standard
of mean deviation
differences umts.
of euthyroid
sick
*Pains, physical
aches,
TPerspiration, ing,
diaturbing
bodily
\ensatwnh.
blthout
apparent
illness.
palpitations.
flushing, frequent
breathing disturbance. cxcessivr micturition. nnusw. vertigo.
swallow-
1064
LEVY ET AL.
ma1 thyroid function tests based on admission blood samples only;4 in contrast, when we evaluated thyroid function tests throughout 10 days of psychiatric hospitalization, we found abnormalities in 27% of the patients. A recent abstract suggests that in certain psychiatric illnesses a transient hyperthyroid-like illness can occur.16 Those workers reported that 19% of acute admissions to a Veterans Administration Center had high FTI and FT31. These findings from a different population are different from our results and those of Cohen and Swigar. Examination of the patients and their charts by both an endocrinologist and a psychiatrist failed to reveal consistent evidence for starvation or any of the other previously documented causes of the syndrome in those with ESS. The data were examined using multivariate analysis with covariate adjustment for drug effects, and a statistically significant association between the IO psychologic variables and ESS emerged. Drug effects are important considerations in anxiety measurements.” indeed, anxiety measures have been used to n-1 .sure the efficacy of psychoactive medications. Of particular interest was that, when considered separately, two variables, decreased reported somatic symptoms and decreased reported autonomic symptoms, discriminated ESS from control patients. Thus, the present study suggests that there are different anxiety components in those patients who have ESS compared to those who do not. This intrigu-
ing relationship is the subject of further investigation currently in progress. The 95% confidence interval was used for the relevant blood tests to distinguish ESS from control psychiatric patients. Our studies of patients in a general hospital,‘4 our preliminary survey of 98 psychiatric patients (Series I) using the same diagnostic criteria, and the 150 patients in Series II showed the overall prevalence of ESS estimated from a single admission blood sample to be 7.3% without statistically significant differences between the groups. Serial sampling in the psychiatric patients during the 4 study days revealed a 27% incidence (we are not aware of any comparable serial study). These results do not permit statistically significant conclusions about the true incidence of ESS in psychiatric populations. Nevertheless, since the 95% confidence limits are often used clinically to distinguish “abnormal” from normal, many hospitalized psychiatric patients are at risk of being-misdiagnosed as having thyroid disease. Our findings reemphasize the importance of history and physical examination and clinical evaluation. In the hospitalized patient thyroid function tests taken alone should be interpreted cautiously. ACKNOWLEDGMENT The authors wish to thank Dr. Henry Cluck of Case Western Reserve University for his invaluable assistance in doing the statistical analysis, and Amy McGowan for her technical assistance.
REFERENCES 1, Chopra IJ, Solomon DH, Hepner GW. et al: Misleadingly low free thyroxine index and usefulness of reverse triiodothyronine measurement in nonthyroidal illnesses. Ann Intern Med 90:905%
IO. Larsen PR. Dockalova J, Sipula D, et al: immunoassay of thyroxine in unextracted human serum. J Clin Endocrinol Metab 37:177-182, 1973
912. 1979 2. Bermudez F, Surks MI, Oppenheimer JH: High incidence of decreased serum triiodothyronine concentration in patients with nonthyroidal disease. J Clin Endocrinol Metab 41:27-40. 1975 3. Birkhauser M, Burer T, Busset R, et al: Diagnosis of hyperthyroidism when serum thyroxine alone is raised. Lancet 2:53-56, 1977 4. Cohen KL and Swigar ME: Thyroid function screening in psychiatric patients. JAMA 242:2.54-257, 1979 5. McLarty DC, Ratcliffe WA, Ratclime JG, et al: A study of thyroid function in psychiatric in-patients. Br J Psychiatry 133:211218, 1978 6. Buss AH, Wiener M, Durkee A, et al: The measurement of anxiety in clinical situations. J Consult Psycho1 19: 125-l 29, 1955 7. Schalling D, Cronholm B, Asberg M, et al: Ratings of psychic and somatic anxiety indicants. Acta Psychiatr Stand 49:353-368, 1973 8. Schalling D, Cronholm 8. Asberg M: Components of state and trait anxiety as related to personality and arousal, in Levi L (ed.): Emotions-Their parameters and measurements. New York, Raven Press, 1975, pp 603-6 I7 9. Spielberger CD, Gorsuch RL, Lushene RE: STAI manual for the state-trait anxiety inventory. Palo Alto, California, Consulting Psychologist Press, Inc., 1970, pp 20-2 I
I I. Larsen PR: Direct immunoassay human serum. J Clin Invest 51:1939-1942. 12. Parslow M: T3/Talc say Manual. Los Angeles: 1974, pp 18-19
of triiodothyronine 1972
in
Uptake (assessment) RadioimmunoasNichols Institute for Endocrinology.
13. Finn JD: Multivariance; univariate and multivariate analysis ofvariance, covariance, and regression. A Fortran IV Program. Ann Arbor. Michigan: National Educational Resources, 1972 14. Levy RP, Laus VG, Hamlin CR: Hyperthyroxinemia and hypothyroxinemia in hospitalized patients. (Abstract) Endocrine Society Annual Meeting, 1976 15. Chopra IJ, Van Herle AJ, Chua Teco CN, et al: Serum free thyroxine in thyroidal and nonthyroidal illnesses: a comparison of measurements by radioimmunoassay, equilibrium dialysis. and free thyroxine index. J Clin Endocrinol Metab 51:135-143, 1980 16. Morley JE and Shafer RB: Stress induced hyperthyroxinemia: a new syndrome? (Abstract) Clinical Research Proceedings 28:761 A, 1980 17. deBonis M: Content analysis of 27 anxiety inventories and rating scales, in Pichot P (ed): Psychological Measurements in Psychopharmacology. Mod Probl Pharmaco Psychiat (Karger, Base1 1974). 7:221-237
B. Jensen,
in Psychiatric
Disease
Victor G. Laus, David P. Agle, and llona M. Engel
Psychiatric illness is a cause of “euthyroid sick syndrome” (ESS), defined as abnormal concentrations of circulating iodothyronines in euthyroid subjects with nonthyroidal illness (NTI). We describe a prospective study of 150 consecutive psychiatric admissions studied by endocrine and psychologic techniques. Based on 150 admission blood samples, we found a 7% incidence of ESS and with serial samples (74 patients) the incidence was 27%. demonstrating that ESS can develop after hospital admission. Of the 20 patients with ESS, 11 had elevation of both serum total T4 concentrations (T4) and free thyroxine index (FTI) while their serum total T3 concentrations (T3) remained normal; 5 had elevation of FTI without elevation of T4 or T3; and 4 had low T4 and low FTI and normal TSH. In 2 of the 4 patients in the last category, the T3 was also low. The free T3 index (FT311 was normal in all but 1 patient who had low FT3l and FTI, low T4 and T3, and normal TSH. The serum thyroid hormone abnormalities were transient in the ESS patients during the 10 day period with 2 exceptions; 1 patient had persistently elevated T4 and FTI with normal T3 and FT3l values while another patient had persistently depressed T4 and FTI without abnormality of FT3l or TSH. Multivariate statistical analysis demonstrated a difference (P -C .06) in the psychologic attributes of somatic and autonomic symptoms in ESS patients compared to controls. We conclude that ESS is as common amongst psychiatric admissions as in general hospital patients previously studied and that blood thyroid function tests should be interpreted cautiously in all hospitalized patients.
I
T HAS BEEN demonstrated that a variety of acute and chronic disorders can alter serum concentrations of free thyroid hormones without causing overt hyperthyroidism or hypothyroidism. The term euthyroid sick syndrome (ESS) has been used to describe patients who are clinically euthyroid, ill with nonthyroidal illness (NTI), who have abnormal thyroid function tests; i.e., a low serum T3 concentration (T3) and an elevated reverse T3 (rT3),‘.2 a low free thyroxine index (FTI) not accompanied by an elevated concentration of TSH,’ or alternatively a high FTI not accompanied by an elevated blood T3.3,4 In a preliminary study of the incidence of the syndrome in a university hospital we noted ESS in several patients in whom no obvious cause other than psychiatric illness was found. A series of serum samples from 98 consecutive new psychiatric admissions (Series I) was then studied and 9 patients with ESS were detected using the criteria and laboratory techniques described later. McLarty’ also noted abnormal T3 and serum total T4 concentrations (T4)
From the Department of Psychiatry and Thyroid Center, University Hospitals of Cleveland and Case Western Reserve University. Cleveland. Ohio. Dr. Jensen is now with the Division of Chiid and Adolescent Psychiatry. University of Minnesota Hospitals. Dr. Levy is now Director of the Division of Medicine, and Professor of Internal Medicine, Ajiliated Hospitals at Canton. Northeastern Ohio Universities College of Medicine. Presented in part at the 4th Great Lakes Regional Endocrine Society Meeting, Rochester, New York, 1980. Received for publication December 8, 1980. Supported in part bv teaching grant #ITO-IMH-15022-01 PI from the National Institute of Mental Health. Address reprint requests to Dr. Richard P. Levy, Division of Medicine, Timken Mercy Medical Center, 1320 Timken Mercy Drive NW, Canton, Ohio 44708. ii 1981 by Grune & Stratton, Inc. 0026-0495/81/301 I-0004$0l.00/0
1060
without thyroid disease in a large psychiatric inpatient population and Cohen and Swigar reported findings consistent with ESS in 9% of newly admitted psychiatric patients.4 To identify the psychic factors correlated with ESS, a retrospective but blind analysis of Series I comparing the hospital charts of the nine ESS patients and concurrent control cases was conducted by the psychiatrists (JBJ and DPA), evaluating the presence of general motor activity, catabolic stress, life change, and anxiety in each patient. Though no parameters demonstrated statistical significance, a trend was noted implicating an association between the ESS and inability to handle anxiety effectively. Accordingly, a second sample of patients (Series 11) admitted to a psychiatric hospital was defined using a more sophisticated evaluation of anxiety. Serial thyroid function tests were done. The data to be reported from Series II demonstrated that ESS often begins during psychiatric hospitalization and that there is a statistically significant relationship between ESS and certain manifestations of anxiety. MATERIALS
AND
METHODS
Patients All patients admitted on six days of the week to the adult psychiatry service at the University Hospitals of Cleveland from December, 1977, through May, 1978, were included in the study. (Saturday was excluded since admission blood samples could not be obtained within 24 hr.) The hospital population serves a wide range of socioeconomic groups and draws from the greater metropolitan area of Cleveland, Ohio. The patients included all diagnostic categories and all degrees of emotional disturbance common to psychiatric inpatient units. The endocrinologist (RPL) determined which patients had ESS. The names of patients with ESS and controls matched for age, sex, and race were sent to two other investigators, a psychiatrist (JBJ) and a psychologist (IME). After patient consent
Metabolism, Vol. 30. No.
11
(November). 198 1
1061
THYROID HORMONES AND PSYCHIATRIC DISEASE
was obtained, the psychiatrist and psychologist, ignorant of whether the patient was ESS or control, conducted the psychological measurement within 72 hr of admission. After the interviews were scored, each patient’s chart was reviewed for medications, concurrent medical illness, and history of thyroid disease. Patients were excluded when medications or illness that are known to affect the thyroid function tests were found and patients with known thyroid disease were analyzed separately. Within 3 days of the psychologic examination the endocrinologist interviewed and examined each patient and the medical record for evidence of thyroid disease and concurrent medical illness.
Psychological
Measurements
The psychiatrist and psychologist administered the Buss Anxiety Rating Scale, as modified by Schalling,‘.8 and the patient completed a State-Trait Anxiety Inventory (STAI)9 which was scored mechanically
Laboratory
MeaSurements
A second blood sample was obtained immediately after the STAI was completed, and 3 days and 7 days afterwards. All 4 serum samples were frozen promptly, stored at -20°C and then ran simultaneously in the same analytic run. Tt. owing tests were done: serum total T4 and T3 concentration\ were measured by specific radioimmunoassays developed in our laboratory by modifying Larsen’s techniques. ‘O.” The “resin” T3 uptake test (RT3U) was done by a talc uptake method.” In this method the radioactivity in the supernatant rather than in the talc is measured. It is calculated as the ratio of the counts per minute (cpm) of the normal serum pool over those of the sample cpm, normal pool cpm. sample
= RT3U
thus yielding values below normal when TBG is increased. The circulating free T4 and T3 were estimated from their respective indices; i.e., the free T4 index (FTI) was the product of total T4 value and the RT3U. This follows the principles published by Bermuder. Surks and Oppenheimer.’ The 95% confidence limits (“normal” values) of these tests are: total T4 = 5--12 ug/dl, total T3 = 80-200 ng/dl. RT3U = 0.80--l. 18,FTI = 5-I 2, Free T3 index (FT31) = 80-200, TSH =
Patients who had been separated into the ESS group and a control group were compared on I2 anxiety measurements and 4 covariates by a multivariate analysis of covariance. The anxiety measures included the 10 combined scores from both raters from the Buss Anxiety Rating Scale, and the State/Trait Anxiety scores from the Spielberger STAI. The four covariates were dummy variables indicating the presence or absence of four classes of drugs. The drugs were classified as major tranquilizers (chlorpromazine and haloperidol), minor tranquilizers (diazepam and chlordiazepoxide). antidepressants (imipramine and amitriptyline) and sedatives (barbiturates). The interaction effects were pooled with the error term, because they were not statistically significant. The statistical package used for the analysis of variance was “Multivariance.“”
RESULTS Endocrine
In Series I, 34 males and 64 females, there were 4 men and 5 women with ESS, an incidence of 9%. Six of the ESS patients had high T4, high FTI and normal T3 and FT31; 3 had low T4, low FTI and normal T3, FT3I and TSH. Two patients were frankly hypothyroid (low T4 and FTI and high TSH). None were hyperthyroid. Of the 175 consecutive admissions comprising Series II, an initial admission blood sample was obtained within 24 hr on 55 males and 95 females, or 85% of the total population admitted. The age range was 14.-78-yr-old. Timely blood samples could not be obtained from the remainder because admission samples were not drawn within 24 hr or because the patient had been transferred to the psychiatric unit from another part of the hospital. Based on the admission blood sample, 4 men and 6 women of the 150 subjects were found to have an abnormal FTI without other evidence of thyroid disease, a 7% incidence of ESS. This value is not statistically significantly different from the 9% incidence found in the preliminary study of psychiatric inpatients at this hospital (Series I). It is also identical to the 7%’ incidence of ESS in the general University Hospitals population.‘J Seventy-four patients without previously known thyroid disease consented to the full experimental protocol, including all of those found to have ESS on admission, and IO more were found to develop ESS during hospitalization, an incidence of 27% in the 74 patients. Ten other patients were found to have thyroid disease which included 4 cases of nontoxic diffuse goiter, 4 cases of hypothyroidism, and 1 case each of spontaneously resolving hyperthyroidism and lithium induced goiter. By exclusion. no patient had concurrent confounding nonthyroidal, medical illness. Of the 20 patients with ESS (Table 1). 1 I had elevation of both T4 and FTI while their T3 remained normal; 5 had elevation of FTI without elevation of T4 or T3; and 4 had low T4 and low FTI and normal TSH. In 2 of the 4 patients in the last category. the T3 was also low. The FT31 was normal in all but I patient who had low FT31 and FTI. low T4 and T3. and normal TSH. We analyzed the variance of’ the data in Table I ;\s the ratio of between subjects vs. within subjects. The ratio ranged from 2.6-6.3 for the various laboratory tests done, indicating the discriminating power of the tests. Correlation (r) of the day-to-day values for each of the 3 thyroid function tests ranges from 0.75 0.95. T4 and T3 values are also found to bc well correlated with each other on all 4 study days with r between 0.50 and
T4.
12.6
12.6
M
F
M
F
M
F
M
M
F
M
M
ID
11
12
13
14
15
16
17
18
19
20
155
135
90
145
12.0
11.5
5.3
10.5
.29
5.3-
.72-
1.30
5.0-
13.2
1.00 1.23 1.06 ,030
12.3 4.9 Il.1 4.8 10.6 .85
141 85 161 156
1.30
*Samples unavailable.
IOC175
11.9
2.9
128
6.9
126
131
165
75
145
105
135
165
145
13.8
4.8-
.28
7.8
.84
9.7
3.8
13.3
3.8
10.0
5.0
15.2
11.2
10.5
9.5
10.5
11.4
13.9
11.9
Day6
11.9
5.7-
.26
6.2
.80
10.1
4.4
14.9
4.6
9.3
5.4
14.1
11.0
11.8
13.4
11.4
11.6
13.1
11.4
10.7
13.5
7.8
12.9
10.5
2.4
8.7
FTI
184
65-
4.2
121
5.2
109
116
120
60
115
loo
150
105
90
100
115
125
110
140
125
126
80
140
105
70
90
ng/dl
T3.
TSH = < lOuU/ml.
1.30
.77-
,020
1.07
,028
1.07
1.16
1.12
1.21
0.93
1.08
0.93
0.98
,.I2
1.41
1.08
1.02
0.94
0.96
1.10
1.09
12.4 9.7
1.00
0.85
1.13
1.20
1.05
RT3U
7.8
15.2
9.3
2.0
8.3
uQ/dl
T4.
ng/dl, FT31 = 80-200,
172
9%
3.2
126
7.2
131
161
165
92
?44
112
139
153
181
122
140
130 100
170
95
132
133
148
SO
165
115
160
120
145
80
105
173
160 135
69
114
FT31
65
100
nQ/dl
T3.
5.1-
.27
8.8
.96
11.1
5.9
11.1
6.0
12.2
4.8
11.5
13.2
20.0
14.6
12.7
13.0
10.8
11.6
12.2
14.5
10.3
12.2
11.6
2.8
11.2
FTI
Day3
1.18, FTI = 5- 12, T3 = 80-200
14.6
184
4.5-
175
.73-
.29
82-
,020
8.7
4.0
1.01
0.99
126
7.1
137
1.22
4.5
95 1.07
11.2 1.03
0.93
14.2
139
143
1.22 1.25
16.0
1.08
1.03
0.63
0.88
1.11
1.02
1.00
0.78
12.0
115 168
90-
3.5
122
7.9
130
131
135
70
95
11.8
158
95% confidence limits: T4 = 5- 12 ug/dl, RTBU = 0.80-
12.0
8.6
.29
1.04
.71
,022
,030
.91
5.7
10.8
10.1
1.19
1.19
6.8
1.21
0.97
1.06
1.03
1.08
6.4
Range
150
11.5
10.5
1.21
1.05
0.92
165
12.0
0.96
12.6
170
165
13.0
13.0
108
1.03
13.2
168
125
13.7
0.86
195
12.8
0.86
Il.0
142
12.7
1.16
10.3
98
120
175
13.1
0.99
15.7
120 14.2
95
10.7
189 1.08
1.22
2.3
147 79
RT3U 1.14
T4. UQ/dl 9.8
FT31
173
135
12.6
160
13.3
65
2.8
12.7
0.89
4.8
*SE
T3.
DQ/dl
100
1
14.0
FTI
Day
1.03
I.18
1.22
1.47
RTBU
10.4
9.5
5.6
10.8
5.0
11.2
13.0
10.0
Mean
Controls In - 441
15.9
9.5
14.9
F
10.8
M
7
M
13.2
F
6
9
12.3
6
14.2
F
F
11.3
M
3
5
2.3
F
2
4
9.5
UQ/dl
F
sex
1
No.
gs
Table 1, Thyroid Function Test Results in Series II
194
7%
4.1
129
4.8
115
135
134
72
107
140
103
101
141
125
128
103
134
138
136
80
119
119
84
94
FT31
1.3
1.8
1.5
"Uhl
TSH,
12.0
4.8-
.29
7.6
.69
9.9
5.0
4.3 .
9.2
.
11.0
13.8
11.5
8.5
10.4
9.0
14.4
11.3
10.4
8.6
6.8
13.5
11.6
.
s.e
UQ/dl
T4.
0.92
1.33
.70-
,021
1.o*
,039
1.06
1.34
1.14 .
0.92
0.91 .
0.92
1.03
1.41
1.03
0.91
0.66
11.2
5.5-
.25
8.0
.60
10.2
6.7
4.9 .
8.5
10.0 *
12.7
11.8
12.0
10.7
8.2
12.4
10.4
11.9
9.1
1.06 1.14
6.9
12.8
13.3
11.1 .
FTI
Day10
1.01
0.95
1.15
1.26 .
RTJU
176
80-
3.7
116
4.9
Ill
123
SO .
120
145 .
125
110
100
135
105
107
145
115
86
80
105
120
nQ/df 95 .
73.
.
165
81-
3.3
123
5.7
118
165
91 .
110
132 .
133
113
141
139
96
92
133
132
90
82
100
138
120
FT31
THYROID
HORMONES
AND
PSYCHIATRIC
1063
DISEASE
0.66, thus showing that these two tests were not completely independent variables. RT3U values however, did not correlate with the other two tests. The T4 and FTI values were statistically significantly different (p c .025) when controls were compared with ESS patients on each study day, but neither T3 nor RT3U values differed significantly. These relationships between the thyroid function tests were the expected ones. The serum thyroid hormone abnormalities were transient in the ESS patients during the 10 day period with 2 exceptions: 1 patient had persistently elevated T4 and FTI with normal T3 and FT3I values while another patient had persistently depressed T4 and FTI without abnormality of FT31 or TSH. As noted previously, 10 patients had ESS on admission. Of the remainder with normal admission thyroid function tests, 6 had abnormal results in the second blood sample and in the remaining 4 the abnormal result was evident by the third blood sample. In 10 patients only I of the 4 blood samples was abnormal, in 6 patients 2 samples were abnormal, in 3 patients 3 samples were abnormal, and all 4 samples were abnormal in I patient Psychological Because the psychological data were incomplete in 9 patients and I patient was found to be taking phenytoin, only 64 patients remained for the multivariate analysis of psychological factors and presence of ESS. Of the 64 patients 18 were ESS and 46 served as controls. The results of the STAl were dropped after they were found not to distinguish ESS from control patients. The multivariate analysis showed that the mean vectors of ESS and control groups were different at the p : .06 level (Table 2). The greatest contributor Table 2.
Relationship
Between
Psychologic Variables
and
Euthyroid Sick Syndrome Univariate Vanable Somatic
Global
-0.69
5.36
0.02
-0.62
4.39
0.04
restlessness
(0)
10.50
2.87
0.09
1.56
0.21
-0.24
0.64
0.42
io.22
0.56
0.45
-0.18
0.37
0.54 0.64
anxety
(R1
0.37
(01 (0)
Disquietude
(R1
(RI symptoms
(0)
-0.13
0.22
restlessness
(R)
-0.12
0.16
0.68
-0.10
0.13
0.71
Autonomvz Muscular
Dlstractablllty
(R)
Multwariate
analysis.
Multivariate
F =
*ASick, verse
P.c
(Rj
Distractabllity
Worry
F
Ml = 1 and 56)
symptoms
symptoms
Autonomic Muscular
ASack, well*
well:
control
R-Reported. O-Observed.
n = 64,df
=
10a
to this difference was somatic symptoms* as reported by the patient. The other important contributor was autonomic symptoms? as reported by the patient. These two items discriminated at the p < .02 and p d .04 level respectively. As seen in Table 2, all other factors had minor discriminating ability. After the means were adjusted for the effects of the covariates. the ESS group reported significantly less somatic symptoms and autonomic symptoms compared to the control group. DISCUSSION
That a variety of NT1 alters the concentrations of circulating iodothyronines is clear; the underlying mechanisms for these changes and their significance is much less clear. While a decrease in production of T3 from T4 has been well demonstrated, recent studies leave unclear whether or not there are changes in the concentration of the circulating free T4. Chopra and coworkers found that different methods of estimating serum free thyroxine which agree well in the absence of NT1 gave conflicting results in patients with a variety of nonthyroidal illnesses.” These investigators found that in NTI the FTI correlated better with a direct RIA for free T4 than with the classical equilibrium dialysis method. In any case our work, as well as that of Cohen and Swigar, adds psychiatric illness to the already long list of NTl’s that can confound the interpretation of thyroid function. We have also demonstrated that. similar to other illnesses, ESS can occur within days in patients with psychiatric illness. Cohen and Swigar’s work seems to have deficiencies: the FTI method they used is an older one utilizing a dextran column for separating bound and free T4 and the outdated PBI method for estimating total T4, making their findings hard to compare with those of Chopra, et al., cited above, and with ours. They did not measure T3. They report a remarkably high number of psychiatric patients with “secondary hypothyroidism,” 8 out of the 12 with thyroid disease. Furthermore, in none of the 8 patients was the diagnosis confirmed by appropriate laboratory investigation. In our patients the abnormalities tended to be only moderately deviant from normal, tended more to be elevation of T4 and FTI rather than depression and were usually transient. Thus the physician will almost always be able to avoid misdiagnosis in these patients by repeating the abnormal thyroid function tests. Cohen and Swigar reported a 9% incidence of abnor-
nd 47, P < 0.06
1.90
Adjusted
expressed
estimate in standard
of mean deviation
differences umts.
of euthyroid
sick
*Pains, physical
aches,
TPerspiration, ing,
diaturbing
bodily
\ensatwnh.
blthout
apparent
illness.
palpitations.
flushing, frequent
breathing disturbance. cxcessivr micturition. nnusw. vertigo.
swallow-
1064
LEVY ET AL.
ma1 thyroid function tests based on admission blood samples only;4 in contrast, when we evaluated thyroid function tests throughout 10 days of psychiatric hospitalization, we found abnormalities in 27% of the patients. A recent abstract suggests that in certain psychiatric illnesses a transient hyperthyroid-like illness can occur.16 Those workers reported that 19% of acute admissions to a Veterans Administration Center had high FTI and FT31. These findings from a different population are different from our results and those of Cohen and Swigar. Examination of the patients and their charts by both an endocrinologist and a psychiatrist failed to reveal consistent evidence for starvation or any of the other previously documented causes of the syndrome in those with ESS. The data were examined using multivariate analysis with covariate adjustment for drug effects, and a statistically significant association between the IO psychologic variables and ESS emerged. Drug effects are important considerations in anxiety measurements.” indeed, anxiety measures have been used to n-1 .sure the efficacy of psychoactive medications. Of particular interest was that, when considered separately, two variables, decreased reported somatic symptoms and decreased reported autonomic symptoms, discriminated ESS from control patients. Thus, the present study suggests that there are different anxiety components in those patients who have ESS compared to those who do not. This intrigu-
ing relationship is the subject of further investigation currently in progress. The 95% confidence interval was used for the relevant blood tests to distinguish ESS from control psychiatric patients. Our studies of patients in a general hospital,‘4 our preliminary survey of 98 psychiatric patients (Series I) using the same diagnostic criteria, and the 150 patients in Series II showed the overall prevalence of ESS estimated from a single admission blood sample to be 7.3% without statistically significant differences between the groups. Serial sampling in the psychiatric patients during the 4 study days revealed a 27% incidence (we are not aware of any comparable serial study). These results do not permit statistically significant conclusions about the true incidence of ESS in psychiatric populations. Nevertheless, since the 95% confidence limits are often used clinically to distinguish “abnormal” from normal, many hospitalized psychiatric patients are at risk of being-misdiagnosed as having thyroid disease. Our findings reemphasize the importance of history and physical examination and clinical evaluation. In the hospitalized patient thyroid function tests taken alone should be interpreted cautiously. ACKNOWLEDGMENT The authors wish to thank Dr. Henry Cluck of Case Western Reserve University for his invaluable assistance in doing the statistical analysis, and Amy McGowan for her technical assistance.
REFERENCES 1, Chopra IJ, Solomon DH, Hepner GW. et al: Misleadingly low free thyroxine index and usefulness of reverse triiodothyronine measurement in nonthyroidal illnesses. Ann Intern Med 90:905%
IO. Larsen PR. Dockalova J, Sipula D, et al: immunoassay of thyroxine in unextracted human serum. J Clin Endocrinol Metab 37:177-182, 1973
912. 1979 2. Bermudez F, Surks MI, Oppenheimer JH: High incidence of decreased serum triiodothyronine concentration in patients with nonthyroidal disease. J Clin Endocrinol Metab 41:27-40. 1975 3. Birkhauser M, Burer T, Busset R, et al: Diagnosis of hyperthyroidism when serum thyroxine alone is raised. Lancet 2:53-56, 1977 4. Cohen KL and Swigar ME: Thyroid function screening in psychiatric patients. JAMA 242:2.54-257, 1979 5. McLarty DC, Ratcliffe WA, Ratclime JG, et al: A study of thyroid function in psychiatric in-patients. Br J Psychiatry 133:211218, 1978 6. Buss AH, Wiener M, Durkee A, et al: The measurement of anxiety in clinical situations. J Consult Psycho1 19: 125-l 29, 1955 7. Schalling D, Cronholm B, Asberg M, et al: Ratings of psychic and somatic anxiety indicants. Acta Psychiatr Stand 49:353-368, 1973 8. Schalling D, Cronholm 8. Asberg M: Components of state and trait anxiety as related to personality and arousal, in Levi L (ed.): Emotions-Their parameters and measurements. New York, Raven Press, 1975, pp 603-6 I7 9. Spielberger CD, Gorsuch RL, Lushene RE: STAI manual for the state-trait anxiety inventory. Palo Alto, California, Consulting Psychologist Press, Inc., 1970, pp 20-2 I
I I. Larsen PR: Direct immunoassay human serum. J Clin Invest 51:1939-1942. 12. Parslow M: T3/Talc say Manual. Los Angeles: 1974, pp 18-19
of triiodothyronine 1972
in
Uptake (assessment) RadioimmunoasNichols Institute for Endocrinology.
13. Finn JD: Multivariance; univariate and multivariate analysis ofvariance, covariance, and regression. A Fortran IV Program. Ann Arbor. Michigan: National Educational Resources, 1972 14. Levy RP, Laus VG, Hamlin CR: Hyperthyroxinemia and hypothyroxinemia in hospitalized patients. (Abstract) Endocrine Society Annual Meeting, 1976 15. Chopra IJ, Van Herle AJ, Chua Teco CN, et al: Serum free thyroxine in thyroidal and nonthyroidal illnesses: a comparison of measurements by radioimmunoassay, equilibrium dialysis. and free thyroxine index. J Clin Endocrinol Metab 51:135-143, 1980 16. Morley JE and Shafer RB: Stress induced hyperthyroxinemia: a new syndrome? (Abstract) Clinical Research Proceedings 28:761 A, 1980 17. deBonis M: Content analysis of 27 anxiety inventories and rating scales, in Pichot P (ed): Psychological Measurements in Psychopharmacology. Mod Probl Pharmaco Psychiat (Karger, Base1 1974). 7:221-237