Tetrahydrocortisol/Tetrahydrocortisone Ratio (H4F/H4E) as an Indicator of Depressive Feelings HAROLD PERSKY, P H D
Tetrahydrocortisol (H4F) and tetrahydrocortisone (H4E) were determined for a group of 15 depressed patients and 25 normal controls following the administration of cortisol-4-C14. The ratio, H4F /H4E, was significantly greater for the depressed patients than for the normal controls (P < 0.001), although the sum of the two metabolites, H4F + H4E, was not significantly different. Three of four objective test measures of depressive feelings likewise discriminated between the two groups. All four depression measures correlated significantly with the H4F /H4E ratio at the 0.01 level or better. It is suggested that the data of the present report and those of Zumoff et al. (J Clin Endocrinol Metab 39:1120,1974) may be explained by the paradigm: depressive feelings -» reduced thyroid function -» elevated H4F/H4E ratio.
Some time ago, Basu, Persky, and Zuckerman (2) reported urinary levels of H4F Zumoff and his coworkers (1) recently and H4E following administration of reported that the urinary ratio of cortisol-4-C14 for a group of depressed patetrahydrocortisol /tetrahydrocortisone tients and a group of normal subjects. Re(H4F/H4E) following administration of calculation of these data indicated that the cortisol-4-C14 was significantly greater for depressed patients had significantly a group of patients with "miscellaneous" higher H4F/H4E ratios than the control diseases than for a group of normal con- subjects and that this ratio was trols. This increase in ratio occurred de- significantly correlated with objective spite the fact that the sum of the two measures of depressive feelings. This remetabolites, H4F + H4E, as well as the cor- port is therefore being submitted to sugtisol production rate, PRF, were almost gest an alternative, but not necessarily inidentical for the two groups. Since this consistent, explanation of the two sets of increased ratio was obtained in a group of findings. unselected patients, these investigators termed it a "nonspecific illness effect." On the basis of a similar, elevated ratio obSUBJECTS AND METHODS tained in a group of hypothyroid patients by these investigators, they suggested that The subjects of this study were 25 healthy men who their group of miscellaneous patients ranged in age from 21 to 44 years and 15 hospitalized, depressed men, ages 21-68. The control subjects were might also be suffering from mild all employees of a research laboratory whose posihypothyroidism. INTRODUCTION
From the Department of Psychiatry, University of Pennsylvania and the Philadelphia General Hospital, Philadelphia, Pennsylvania 19104. Received for publication June 9, 1975; Revision Received August 18, 1975.
tions varied from junior technician to senior investigator. The patients were classified as depressive reaction and were assessed for the purpose of this study during their first week of hospitalization. Both controls and patients were tested in identical fashion and at the same time of day. All subjects were tested individually.
Psychosomatic Medicine Vol. 38, No. 1 (January-February 1976) Copyright ° 1976 by the American Psychosomatic Society, Inc. Published by American Elsevier Publishing Company, Inc.
13
HAROLD PERSKY At 9:00 AM ± 15 min, the subject was given a selfrating checklist and questionnaire. Immediately after completion, approximately 30 min later, he voided, a blood sample was drawn, and 0.2 /iCi of cortisol-4-C14 in 1 ml of 5 % glucose was injected through the needle in situ. Urine was collected thereafter for 24 hr, completion of collection being assessed by creatinine determination. Plasma cortisol (F) was determined by double chromatographic separation on silica gel and subsequent fluorescent identification, according to Braunsberg and James (3). Following /3-glucuronidase hydrolysis of a urine aliquot and extraction, H4F, H4E, and allo-H4F were separated by column chromatography according to Flood et al. (4). Each moiety was determined by a modified PorterSilber reaction (5). Cortisol production rate (PRF) was determined from the specific activities of H4F and H4E. An average of these two estimates (PRF) was used for the subsequent report. The psychological battery administered to all subjects included the Manifest Anxiety Scale (MMPIMAS), Depression Scale (MMPI-D), and Lie Scale (MMPI-L) from the Minnesota Multiphasic Personality Inventory (6), the Buss-Durkee Hostility Inventory (B-D) (7), and Zuckerman's Multiple Affect Adjective Check List (MAACL) (8). These objective test instruments contain a variety of subscales, some of which are described more fully later in the text.
and H4E, and ratio of H4F /H4E for the 15 depressed patients and 25 control subjects. The H4F/H4E ratio for the controls is somewhat lower than the value reported by Zumoff et al. (1). However, the depressed patients have a significantly higher ratio than the controls (0.66 vs. 0.38), a difference significant beyond the 0.001 level. Figure 1 shows the distribution of actual values of the H4FIR4E ratios for the controls and depressed patients. The distribution of the sum of the two metabolites, H4F + H4E, is shown in Fig. 2. The mean values for the two groups are almost identical. Similarly, Fig. 3 indicates that the cortisol production rates of the two groups are not significantly different, although the depressed patients' rates tend to be somewhat higher. While considerable psychological data are available for both the normals and the patients, the depression measures are of particular interest in connection with the argument to be advanced. All of the depression measures used are of an objective nature rather than clinical judgments. Their correlations with clinical ratings RESULTS have been shown to be significant at better than the 0.01 level in a previous study (9). Table 1 gives the plasma level of cortisol Table 2 summarizes the mean values of the (F), urinary H4F and H4E outputs per day, MAACL-D, MMPI-D, MMPI-D 1, and B-D cortisol production rate (PRF), sum of H4F Factor I scores for the normal subjects and TABLE 1.
H«F (mg/d) H«E (mg/d) PRF (mg/d) H4F+HUE (mg/d) H4F/H4E 3
14
Comparison of Depressed Patients with Normal Subjects: Adrenocortical Measures Normal subjects'1
Depressed patients*
13.5 ±7.0 1.45+0.52 3.89±1.28 14.5 ±5.3 5.35±1.66 0.38±0.10
23.8 ±13.2 2.16± 0.98 3.29± 1.15 17.3 ± 6.3 5.45± 2.01 0.66± 0.20
3.22 3.01 1.50 1.52 0.17 5.97
Mean ± SD; bsignificant beyond the 1% level; csiginficant beyond the 0.1% level.
Psychosomatic Medicine Vol. 38, No. 1 (January-February 1976)
H4FIH4E A N D DEPRESSION
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 1. Distribution of H4FH.iE ratios for normal subjects and depressed patients.
ences between the two groups. The B-D Factor I Scale is one of two factors into which the items of the Hostility Inventory consistently cluster. It has been termed a negative hostility factor (as contrasted to Factor 11,an aggression factor) and has also been considered to be an inward-directed hostility factor (11). In the latter sense, it can be considered to be a measure of depressive feelings. All of these depression measures, with the exception of the MAACL-D, may be considered to be trait measures of depression while the latter is a state measure, i.e., the MAACL-D assesses current feelings of depression while the other measures assess both current and more persistent aspects of depressive mood.
Although the MAACL-D mean score was greater for the depressed patients than for the controls, the mean difference was short of the conventional level of significance. However, the three trait measures of depression significantly differentiated between the two groups: MMPI-D (0.05 < P < 0.10), MMPI-Dl (P < 0.001), and B-D Factor I (P < 0.001). The
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 2. Distribution of H4F + H4E sums for normal subjects and depressed patients.
the depressed patients. The MMPI-Dl scale is one of five scales into which the Depression Scale of the MMPI may be subdivided and is presumed to assess subjective feelings of depression (10). The other four scales did not yield any significant differ-
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 3. Distribution of cortisol production rates for normal subjects and depressed patients.
Psychosomatic Medicine Vol. 38, No. 1 (January-February1976)
15
HAROLD PERSKY TABLE 2. Measure MAACL-D MMPI-D MMPI-Di B-D Factor I
Comparison of Depressed Patients with Normal Subjects: Depression Measures Normal subjects'1
Depressed patients3
12.5±6.0 20.4±4.4 6.2 + 3.3 -0.45±0.67
19.4±5.3 31.1 ±6.8 16.2±5.3 0.74±0.87
1.40 1.86 4.25 4.45
'Mean ± SD; bsignificant beyond the 10% level;c significant beyond the 0.1% level.
other four MMPI-D subscales tended to be higher for the depressed patients but short of conventional levels of significance. Since several of the depression measures differentiated between patients and controls, correlation coefficients (r) were calculated for each of these measures and the H4F IH*E ratio. Significant correlations were obtained in each instance: MAACL-D (r = 0.36,P
Tetrahydrocortisol (H4F) and tetrahydrocortisone (H4E) were determined for a group of 15 depressed patients and 25 normal controls following the administration of cortisol-4-C14. The ratio, H4F /H4E, was significantly greater for the depressed patients than for the normal controls (P < 0.001), although the sum of the two metabolites, H4F + H4E, was not significantly different. Three of four objective test measures of depressive feelings likewise discriminated between the two groups. All four depression measures correlated significantly with the H4F /H4E ratio at the 0.01 level or better. It is suggested that the data of the present report and those of Zumoff et al. (J Clin Endocrinol Metab 39:1120,1974) may be explained by the paradigm: depressive feelings -» reduced thyroid function -» elevated H4F/H4E ratio.
Some time ago, Basu, Persky, and Zuckerman (2) reported urinary levels of H4F Zumoff and his coworkers (1) recently and H4E following administration of reported that the urinary ratio of cortisol-4-C14 for a group of depressed patetrahydrocortisol /tetrahydrocortisone tients and a group of normal subjects. Re(H4F/H4E) following administration of calculation of these data indicated that the cortisol-4-C14 was significantly greater for depressed patients had significantly a group of patients with "miscellaneous" higher H4F/H4E ratios than the control diseases than for a group of normal con- subjects and that this ratio was trols. This increase in ratio occurred de- significantly correlated with objective spite the fact that the sum of the two measures of depressive feelings. This remetabolites, H4F + H4E, as well as the cor- port is therefore being submitted to sugtisol production rate, PRF, were almost gest an alternative, but not necessarily inidentical for the two groups. Since this consistent, explanation of the two sets of increased ratio was obtained in a group of findings. unselected patients, these investigators termed it a "nonspecific illness effect." On the basis of a similar, elevated ratio obSUBJECTS AND METHODS tained in a group of hypothyroid patients by these investigators, they suggested that The subjects of this study were 25 healthy men who their group of miscellaneous patients ranged in age from 21 to 44 years and 15 hospitalized, depressed men, ages 21-68. The control subjects were might also be suffering from mild all employees of a research laboratory whose posihypothyroidism. INTRODUCTION
From the Department of Psychiatry, University of Pennsylvania and the Philadelphia General Hospital, Philadelphia, Pennsylvania 19104. Received for publication June 9, 1975; Revision Received August 18, 1975.
tions varied from junior technician to senior investigator. The patients were classified as depressive reaction and were assessed for the purpose of this study during their first week of hospitalization. Both controls and patients were tested in identical fashion and at the same time of day. All subjects were tested individually.
Psychosomatic Medicine Vol. 38, No. 1 (January-February 1976) Copyright ° 1976 by the American Psychosomatic Society, Inc. Published by American Elsevier Publishing Company, Inc.
13
HAROLD PERSKY At 9:00 AM ± 15 min, the subject was given a selfrating checklist and questionnaire. Immediately after completion, approximately 30 min later, he voided, a blood sample was drawn, and 0.2 /iCi of cortisol-4-C14 in 1 ml of 5 % glucose was injected through the needle in situ. Urine was collected thereafter for 24 hr, completion of collection being assessed by creatinine determination. Plasma cortisol (F) was determined by double chromatographic separation on silica gel and subsequent fluorescent identification, according to Braunsberg and James (3). Following /3-glucuronidase hydrolysis of a urine aliquot and extraction, H4F, H4E, and allo-H4F were separated by column chromatography according to Flood et al. (4). Each moiety was determined by a modified PorterSilber reaction (5). Cortisol production rate (PRF) was determined from the specific activities of H4F and H4E. An average of these two estimates (PRF) was used for the subsequent report. The psychological battery administered to all subjects included the Manifest Anxiety Scale (MMPIMAS), Depression Scale (MMPI-D), and Lie Scale (MMPI-L) from the Minnesota Multiphasic Personality Inventory (6), the Buss-Durkee Hostility Inventory (B-D) (7), and Zuckerman's Multiple Affect Adjective Check List (MAACL) (8). These objective test instruments contain a variety of subscales, some of which are described more fully later in the text.
and H4E, and ratio of H4F /H4E for the 15 depressed patients and 25 control subjects. The H4F/H4E ratio for the controls is somewhat lower than the value reported by Zumoff et al. (1). However, the depressed patients have a significantly higher ratio than the controls (0.66 vs. 0.38), a difference significant beyond the 0.001 level. Figure 1 shows the distribution of actual values of the H4FIR4E ratios for the controls and depressed patients. The distribution of the sum of the two metabolites, H4F + H4E, is shown in Fig. 2. The mean values for the two groups are almost identical. Similarly, Fig. 3 indicates that the cortisol production rates of the two groups are not significantly different, although the depressed patients' rates tend to be somewhat higher. While considerable psychological data are available for both the normals and the patients, the depression measures are of particular interest in connection with the argument to be advanced. All of the depression measures used are of an objective nature rather than clinical judgments. Their correlations with clinical ratings RESULTS have been shown to be significant at better than the 0.01 level in a previous study (9). Table 1 gives the plasma level of cortisol Table 2 summarizes the mean values of the (F), urinary H4F and H4E outputs per day, MAACL-D, MMPI-D, MMPI-D 1, and B-D cortisol production rate (PRF), sum of H4F Factor I scores for the normal subjects and TABLE 1.
H«F (mg/d) H«E (mg/d) PRF (mg/d) H4F+HUE (mg/d) H4F/H4E 3
14
Comparison of Depressed Patients with Normal Subjects: Adrenocortical Measures Normal subjects'1
Depressed patients*
13.5 ±7.0 1.45+0.52 3.89±1.28 14.5 ±5.3 5.35±1.66 0.38±0.10
23.8 ±13.2 2.16± 0.98 3.29± 1.15 17.3 ± 6.3 5.45± 2.01 0.66± 0.20
3.22 3.01 1.50 1.52 0.17 5.97
Mean ± SD; bsignificant beyond the 1% level; csiginficant beyond the 0.1% level.
Psychosomatic Medicine Vol. 38, No. 1 (January-February 1976)
H4FIH4E A N D DEPRESSION
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 1. Distribution of H4FH.iE ratios for normal subjects and depressed patients.
ences between the two groups. The B-D Factor I Scale is one of two factors into which the items of the Hostility Inventory consistently cluster. It has been termed a negative hostility factor (as contrasted to Factor 11,an aggression factor) and has also been considered to be an inward-directed hostility factor (11). In the latter sense, it can be considered to be a measure of depressive feelings. All of these depression measures, with the exception of the MAACL-D, may be considered to be trait measures of depression while the latter is a state measure, i.e., the MAACL-D assesses current feelings of depression while the other measures assess both current and more persistent aspects of depressive mood.
Although the MAACL-D mean score was greater for the depressed patients than for the controls, the mean difference was short of the conventional level of significance. However, the three trait measures of depression significantly differentiated between the two groups: MMPI-D (0.05 < P < 0.10), MMPI-Dl (P < 0.001), and B-D Factor I (P < 0.001). The
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 2. Distribution of H4F + H4E sums for normal subjects and depressed patients.
the depressed patients. The MMPI-Dl scale is one of five scales into which the Depression Scale of the MMPI may be subdivided and is presumed to assess subjective feelings of depression (10). The other four scales did not yield any significant differ-
NORMAL SUBJECTS
DEPRESSED PATIENTS
Fig. 3. Distribution of cortisol production rates for normal subjects and depressed patients.
Psychosomatic Medicine Vol. 38, No. 1 (January-February1976)
15
HAROLD PERSKY TABLE 2. Measure MAACL-D MMPI-D MMPI-Di B-D Factor I
Comparison of Depressed Patients with Normal Subjects: Depression Measures Normal subjects'1
Depressed patients3
12.5±6.0 20.4±4.4 6.2 + 3.3 -0.45±0.67
19.4±5.3 31.1 ±6.8 16.2±5.3 0.74±0.87
1.40 1.86 4.25 4.45
'Mean ± SD; bsignificant beyond the 10% level;c significant beyond the 0.1% level.
other four MMPI-D subscales tended to be higher for the depressed patients but short of conventional levels of significance. Since several of the depression measures differentiated between patients and controls, correlation coefficients (r) were calculated for each of these measures and the H4F IH*E ratio. Significant correlations were obtained in each instance: MAACL-D (r = 0.36,P
