Cortisol in alcoholics with a disordered aggression control

0306-4530/5’2 SS.OO+O.OO M-P==plc

Psyckmnvrmdocrinology, Vol. 17, No. 1. pp. 4%54,1992 F’rintedin GreatBritain

CORTISOL IN ALCOHOLICS WITH A DISORDERED AGGRESSION CONTROL L. BUYDENS-BRANCHEY and M. H. BRANCHEY Departmentof Psychiatry, MountSinai School of Medicine, City College of New York, New York City, and Veterans Administration Medical Center, Bronx, New York, U.S.A. (Received 31 July 1989; injnalform

2 November 1989)

SUMMARY Considerable evidence exists that the limbic system and the hypothalamus play an important role in the BPA axis disturbances found in depressive disorders. Evidence also exists that the limbic system plays a role in the modulation of aggressive behavior. Yet the I-IPA function of individuals with a disordered regulation of aggression has received little scrutiny. Because aggressive behavior has been observed to be extensively correlated with heavy alcohol use, we explored the BPA function of alcoholics who had had a life-long history of violence. Basal 0700h cords01 was measured in 4 consecutive wk following cessation of drinking in 19 alcoholics with a history of depression, and 17 alcoholics with a history of violent behavior, eight of whom had been incarcerated because of the severity of their violent acts. When compared with alcoholics with no problem in mood or aggression regulation, significant cords01 increases were found in the group of patients who had been incarcerated for violent acts and not in any other group. This increase persisted for 4 wk after cessation of drinking. A variety of variables, including several measures of alcohol consumption, amounts of benzodiazepines used for detoxification, and liver function tests, failed to show significant associations with cortisol. Data are interpreted as indicating that individuals displaying severe forms of violence could have a dysregulated BPA function revealed by exposure to excessive amounts of alcohol.

INTRODUCTION

SINCETHEDISCOVERY that the adrenal glands were involved in responses to stress, investigators have studied the hypothalamo-pituitary-adrenal (IPA) function of a large number of psychiatric patients in an attempt to correlate abnormalities in this function with emotional factors. Disturbances in TWA function have been described in depressed patients more frequently than in any other category of psychiatric patients. Among the alterations observed in these patients arc an increased secretion of cortisol, disturbances in feedback mechanisms as manifested by nonsuppression of cortisol release to dexamethasone, an exaggerated adrenocortical response to ACTH, and alterations in the circadian rhythms of AC’lX and cortisol release. The method utilized most frequently to test the HPA function is the dexamethasone suppression test (DST). The test is not considered to be specific for depression, but abnormal responses seem to Addresscorrespondenceand reprint requests to: Dr. Lame Buydens-Branchey, Brooklyn V. A. Medical Center, 800 Poly Place, Brooklyn NY 11209, USA. 45

46

L. BWDENS-BRANCHEY& M. H. BRANCHEK

increase with increased severity of depressive symptomatology, the highest rate of abnormal responses being associated with severe suicidality (Arana ef al., 1985). It has been suggested that the persistent HPA axis hyperactivity found in depression could derive in part from the persistent release of increased amounts of corticotropin releasing hormone, secondary to alterations in hypothalamic, limbic, or higher brain structures (Carroll, 1982; Stokes & Sikes, 1988; Truffinet & Lecrubier, 1988). It also has been suggested that the limbic system and the hypothalamus could be the sites of primary pathology in affective disorders (Carroll, 1982). Considerable evidence exists that, in addition to its role in the modulation of mood, the limbic system plays a role in the modulation of aggressive behavior (Wilson, 1960; Monroe, 1986; Post, 1986), but the wealth of data on HPA axis dysregulation in depressive illness has not been matched by data on HPA function of patients with problems in aggmssion control. Individuals with problems in the control of aggression are difficult to study, because they do not necessarily come to medical attention for these problems. However, a large percentage of these patients abuse alcohol and other drugs, and they do come to medical attention when they experience physical or psychological problems resulting from alcoholism or drug abuse. According to the U. S. Department of Health and Human Services, nearly half of all homicides were found to have been perpetrated by individuals who had been drinking. Alterations in HPA function have been described in alcoholics during chronic drinking as well as during withdrawal (Mendelson et al., 1971; Stokes, 1973). In these patients, HPA disturbances have been studied in the context of depressive episodes, but their associations with problems in aggression control have not been explored. Because, in the past, we had observed biological abnormalities in a population of violent alcohol abusers (Branchey et al., 1984; 1988; Buydens-Branchey et al., 1989), we wondered whether indices of a dysregulated HPA axis function could be found in these patients as well. SUBJECTS AND METHODS The subjects were selected from a population of male alcoholic patients admitted to a detoxification unit for up to 1 wk before their transfer to a rehabilitation unit, where they stayed for 3-4 wk. Treatment on the detoxification unit was conservative, and psychotropic medication was used only if necessary and for no longer than 3-4 days. Patients in need of medication were given benzodiazepines, usually chlordiazepoxide, but sometimes diazepam. Patients were excluded if they had major physical illness (including heart, lung and GI tract diseases, endocrine diseases and cancers), major abnormalities of liver function tests (albumin level of I 3 g/dl, SGOT or 8CiPT of > 150 U/l, total bilirubin level of > 2 mg/dl, alkaline phosphatase of > 250 U/l, or evidence of cirrhosis on liver biopsy), a history of use of narcotics or other psychotropic drugs during the year preceding admission, or evidence of an organic mental disorder or a history of symptoms of schizophrenia or mania, as assessed with DSM-III criteria (American Psychiatric Association, 1980). Patients with a history of primary depression (preceding the onset of alcoholism or evident during episodes of prolonged sobriety and unrelated to drinking or withdrawal episodes) also were excluded. This last criterion was used in an attempt to select patients whose depression was secondary to alcohol abuse. These patients represent 90% of alcoholics experiencing depressive episodes. Among individuals who satisfied DSM-III diagnostic criteria for alcohol dependence, who were 25-60 yr of age, and who did not meet our exclusion criteria, three different groups of patients were selected after having given informed consent. They were individuals who had never had problems in mood or aggression control, individuals with a past history of depression but no history of aggression, and individuals with a past history of aggression but no history of depression. This last group was further subdivided into patients who were never incarcerated for violent acts and patients who had been incarcerated for such acts.

CORTWLINAGGRESSIVEALCOHOLICS

47

Patient selection was done after the administration of the Structured Clinical Interview for DSM-III (SCID) (Spitzer & Williams, 1984) and the Schedule for Affective Disorders and Schizophrenia (SADS) (Spitzer & Endicott, 1978). In addition, information about past history of aggressive behavior was obtained with a questionnaire adapted from the Brown-Goodwin Assessment for Life History of Aggression (Brown et al., 1981). This questionnaire was aimed at eliciting information about the following behavioral categories: problems with discipline in the Army, problems with discipline at work, assaults on people, property damage, incarceration for assaultive behavior, incarceration for other crimes, and crimes which did not result in incarceration. In each category, non-occurrence was scored 0, one event was scored as 1; two events or “a few” were scored as 2; three events, “several” or “frequent” were scored as 3; and four or more events, “many”, or “numerous”, were scored as 4. Patients with a score of 28 were considered to have a past history of aggression. Presence of aggressive tendencies was also assessed with the Buss-Durkee Inventory, which consists of six hostility subscales, a suspicion subscale and a guilt subscale (Buss & Durkee, 1957). The test is self-administered; in order to obtain more reliable data, the inventory was modified in such a way that it could be administered by an investigator. Drinking history data were obtained. They included the age at the beginning of regular and excessive drinking, the number of years of regular drinking, the age at which alcohol abuse was perceived to be a problem, and the amount of ethanol consumed during the month preceding admission. Levels of benzodiazepines given on the detoxification unit were recorded on a daily basis. Diazepam doses were converted into chlordiazepoxide equivalents. Blood was collected at 07OOh after an overnight fast during the Iirst, second, third and fourth weeks following cessation of drinking. Plasma was separated and kept frozen at -80’ C until assayed. Plasma total cortisol was measured in duplicate samples by radioimmunoassay. The inter- and intra-assay coefficients of variation were 5.7% and 7.58, respectively. This study was approved by the Bronx Veterans Administration Medical Center Institutional and Human Subjects Committees. The Code of Ethics of the World Medical Association was adhered to. RESULTS

Disorders in mood and aggression regulation The population studied consisted of 43 patients: 19 patients had no history of depression or

aggression; 7 patients had a history of depression but no history of aggression; and 17 patients had a history of aggression but no history of depression. Eight of the seventeen individuals who had displayed violent tendencies had been incarcerated for having committed violent acts. These acts included murder, assaults on others, including officers in the army or police officers, with or without the use of a deadly weapon, and repeated wife battering. Scores obtained with the questionnaire on life history of aggression were 3.16k1.54 for patients with no history of depression or aggression, 3.33k1.63 for patients with a history of depression but not aggression, and 11.88 f 3.35 for patients with a history of aggression but not depression. The effect on cortisol of time elapsed since last drink was first assessed in the entire population by an analysis of variance for repeated measures. This analysis revealed a significant effect of time (p
48

L. BWDENCBRANCHEY & M. H. BRANCHM

1 WEEK

2 WEEKS

I

No depresslon

m

Depression

4 WEEKS

3 WEEKS

LENGTH OF SOBRIETY FIG. 1: Sequential 0700h cortisol in patient groups, Repeated-measures

ANCOVArevealed a significant effect

of time (p-zO.05) but no dtfference between groups.

1 WEEK

2 WEEKS

3 WEEKS

I

No aggression

BBI

Aggression

4 WEEKS

LENGTH OF SOBRIETY RG. 2: &quenw 07OOh cortisol in patient groups. Repeated-measures oftime(P<0.02)andaslBnificantdiffirencebetween~Ps(P
ANCOW revealed a -cant

e&x3

49

CORTIWLINAGGRESMIALCOHOLICS

Figure 2 shows the sequential 07OOh cortisol values for group I patients and for patients with a life-long history of aggression (group III). ANCOVA for repeated measures showed a significant effect of time (I, c 0.02) and a significant difference between groups (p
3 30 ri

0

T

I WEEK

m

No aggression Aggression -

no jail

Jail for violent acts

2 WEEKS

3 WEEKS

4 WEEKS

LENGTH OF SOBRIETY FIG. 3: Sequential 07OOh cortisol ln patient groups. Repeated-measures ANCOVA rcvcakd a s&nUicantefrcct oftime (pcO.001). a @nitkant difference between groups (pcO.001). and a sign&ant group xttme MeracUon (p
50

L. BUYDENS-BRANCHEY & M. H. BRANCHEY

Buss- Durkee Inventory and cortisol The data on aggression reported above were derived from patients’ past histories of violence and imprisonment. The question of whether there was a relationship between the presence of hostile tendencies at the time our studies were conducted and HPA function was addressed. The behavior of our patients on the inpatient unit gave us li.ttle information about their hostile tendencies. They were not psychotic and had no opportunity to express their anger. The Buss-Durkee, administered as part of our test battery, did give us information about the patients’ angry feelings. Correlations were calculated between cortisol values and the sum of the six Buss-Durkee hostility subscales, the guilt subscale, and the suspicion subscale. Significant correlations were found between cortisol and the sum of the hostility subscales during the second (r=.33; pcO.Ol), third (r=.33; p
between cortisol and the guilt and suspicion

subscales were not significant.

Drinking history variables The Table shows the age at the beginning of excessive drinking, the number of years of regular drinking, and the amount of pure alcohol consumed daily (in grams) during the month preceding admission for the various patient subgroups. The same table also gives the patients’ age at the time of admission. Groups II, III, IBa and IBb were each compared to group I by Student’s t-tests. Significant differences between groups were found only in amount of alcohol consumed prior to admission: Groups II and IIlb were significantly different from group I @ c 0.05 for each group). DRINKING HISTORY VARIABLES Age at start of d-inking

Years of regular drinking

Daily amount of pure alcohol (grams)

45.72 12.1

25.2 f 8.4

20.5 f 13.4

225.4 f 89.7

7

40.4 f 8.0

22.8 It:6.3

17.6 f 8.8

368.0f 177.1*

17

42.2 + 9.8

22.4+ 5.1

19.8 f 10.9

411.7f 341.7*

9

43.3 f 7.3

21.5k2.7

21.7f8.3

374.9 f 292.1

8

41.0 f 12.3

23.3 -+6.9

17.6 f 13.3

453.1 f 407.1*

N

GROUP I No depression or aggression history

19

Age

GROUP II

History of depression, no history of aggression GROUP III

depression history history of aggression

No

GROUP IIIa

No depression history, history of aggression without incarceration GROUP IIIb No depression history history of aggression resulting in incarceration

values are

means&SD.

Student’s t-tests were used to compare groups II, III. an asterisk were significant lp
comparisons marked with

IIIa and IIIb to greup I. ‘Ibe

CORIEOL IN AGGRESS~ ALCOHOLICS

51

Liver function tests There was no difference among the patient groups in incidence of minor medical complica-

tions or liver function test results. Some patients had slight-to-moderate elevations of SGGT and SGPT, but there was no association between these elevations and cortisol levels. Benzodiazepines

Correlations were calculated between cortisol levels and total amounta of benzodiazpines taken up to the time of the blood collections. This was done for the first and all subsequent weeks. These correlations were not significant, whether calculated for the entire population or for patient subgroups. DISCWSION

When measured in the entire sample, cortisol was the highest during the week that followed cessation of drinking and decreased over the following 3 wk. Others have observed an increase in corticosteroids in alcoholics during chronic drinking (Mendelson et al., 1971; Stokes, 1973) followed by a decrease during abstinence (Stokes, 1973). Factors underlying these changes have not yet been elucidated. In the present study, the decrease in cortisol over tune was more abrupt in those patients incarcerated for violent acts, as indicated by the significant group x time interaction in the comparison of this group of patients with the nonviolent individuals. The lowest cortisol values were found in individuals with no history of depression or aggression. Compared to this group, patients with a history of secondary depression had a slight but nonsignificant increase in cortisol which persisted throughout the 4 wk of the study. The majority of investigators who have explored the HPA function of depressed alcoholics have concluded that there was no relationship between increases in basal cortisol levels or abnormal DST responses and depression (Abou-Saleh et al., 1984; Ravi et al., 1984; Wiienbring et al., 1984; Del Port0 et al., 1985; Johnson & Perry, 1986; Nelson et al., 1986; h-win et al., 1988). Others have disagreed and have reported either a cortisol elevation or an increased frequency of DST nonsuppressors among depressed alcoholics (Targum et al., 1982; Dackis et al., 1986; Zem et al., 1986). Sources of variability between studies could include diiercnces in the time elapsed between the last drink and the cortisol measures, as well as differences in the relative proportion of primary and secondary depressives (whether primary or secondary depressives were selected was usually not specified). Secondary depressions are more frequent in alcoholics and may represent biologically different disorders that do not share all the biological characteristics of primary depressions and thus might not present the high frequency of HPA axis disturbances found in primary depressions. Patients with a history of aggression had cortisol levels significantly higher than those of patients with no problem in mood or aggression control. When the group of violent patients was subdivided, only those whose violent acts had been frequent enough and severe enough to result in incarceration differed significantly from the patients with no history of depression or aggression. Values for patients who had a history of aggression but had never been incarcerated wem elevated but were not significantly different from those of individuals with no disturbance in aggression control. Thus, the severity of aggression appears to be related to the magnitude of disturbances in the I-PA axis. An overexposure to alcohol seems to stimulate cortisol secretion, but not in the same way in all alcoholics. In the present study, the HPA axis disturbances appeared to be more pronounced in patients who displayed severe forms of violence. The cortisol levels of these patients were

52

L. BLNDENS-BRANCHEY & M. H. BRANCHM

higher shortly after they stopped drinking and decreased more abruptly thereafter. These patients could be hypersensitive to the effects of alcohol. Whether the elevated values still found after 4 wk of sobriety represent a delayed return to levels that would be similar to those of nonaggressive patients or a return to higher baseline levels is not known; the HPA axis properties of violent individuals were not studied before they started drinking or during prolonged periods of abstinence. The abnormal DST status found by others in some alcoholics who were not depressed could have been partly due to the presence of some patients with violent tendencies. The increased cortisol found in individuals jailed for violent acts appears not to have been caused by the larger amounts of alcohol they had ingested. They drank more than patients with no problem in depression or aggression regulation, but the same was true of patients with a history of depression, and yet no significant increase in cortisol was found in this last subgroup. In addition, the amounts of alcohol ingested were used as covariates in all the statistical analyses. An absence of correlation between the quantity of alcohol consumed and disturbances of the IIPA axis has been reported by others (Ravi et al., 1984; Willenbring et al., 1984; Zem et al., 1986). A number of studies of the effects of alterations in liver function on responses to the DST have not shown the existence of any relationship between liver dysfunction and DST nonsuppression (Del Port0 er al., 1985; Johnson & Perry, 1986; Nelson et al., 1986; Zem et al., 1986). These studies were conducted in alcoholics whose liver pathology were not severe. We made similar observations. The patients involved in the present study did not have severe liver dysfunction. Some of them had slight-to-moderate elevations of SGOT and SGPT, but we observed no relationship between these elevations and changes in cortisol levels. We did not find significant associations between cortisol levels and amounts of benzodiazepines taken by p.atients during the first days of their stay in the hospital. This finding is in agreement with the observations of others who found that benzodiazepines do not affect steroid metabolism significantly and do not interfere with the DST (Mendlewicz et al., 1982; Breier et al., 1986). In this study basal measures of plasma cortisol were used as indices of HPA axis function. Basal measures, unlike those that follow challenges, provide information about the resting activity of the HPA system. The disadvantage of isolated measures is their unpredictability due to the episodic release of IIPA axis hormones. In the present study, this was partly circumvented by the measurement of cortisol in groups of patients and not in isolated individuals. Stokes and Sikes (1988) indicated that there has been a renewed interest in measurements of basal levels of total plasma cortisol and that data from their own laboratory suggests that “8:O0 AM cortisol has the same specificity, sensitivity and predictive value for depression as does the DST, when used in the same patient population.” In our study, as in the majority of studies published to date, total cortisol was measured. Approximately 75% of total plasma cortisol is bound to globulin whose binding sites are saturated when the hormone concentration exceeds 20 pg/dl (Muldoon & West@& 1967). The remaining 25% is divided into a pool loosely bound to albumin (15%) and a pool that circulates freely (10%). It is only the free cortisol (and possibly the loosely bound pool) that is considered to be biologically active and that crosses the blood brain barrier. The long-term administration of glucocorticoids has been found to produce a loss of hippocampal neurons (Sapolsky et al., 1985). and it has been suggested that the hippocampal damage contributes to the abnormal mood as well as to cognitive and behavioral disturbances in depressive illness. The existence of such a mechanism could also be operative in violent alcoholics in light of our finding in these patients of cortisol levels higher than 20 pg/dl (presumably accompanied by

CORTWL IN AGGRESSWE ALCOHOLICS

53

marked increases in the free and loosely bound forms of the hormone). Of note is our previous finding, in patients incarcerated for violent acts, of deficits in the P3 component of eventrelated potentials, which are believed to originate in the hippocampus and amygdala (Branchey ef al., 1988). In summary, this study offers preliminary evidence that alcoholics who have a tendency to exhibit severe forms of violence have disturbances of the HPA axis, as manifested by an increase in basal cortisol under the influence of alcohol. Whether these disturbances antedate alcohol abuse, are a consequence of hippocampal damage subsequent to exposure to the high cortisol levels occurring during exposure to excessive amounts of alcohol, or result from a combination of these and possibly other factors is not presently known. Acknowledgements: This study was supported by the Veterans Administration

and by Grant AA6510 from

the National Institute on Alcohol Abuse and Alcoholism. REFERENCES Abou-Saleh MT, Merry J. Coppen A (1984) Dexamethasone suppression test in alcoholism. Acta Psychiutr Scund69: 112-116. American Psychiatric Association (1980) Diagnostic and StatisticalManual of Mental Disorders, 3rd edition. APA Press, Washington DC. Arana GW, Baldessarini RJ, Grnsteen M (1985) The dexamethasone suppression test for diagnosis and prognosis in psychiatry. Arch Gen Psychiatry 42: 1193- 1204. Branchey L. Branchey M, Shaw S, Lieber CS (1984) Depression, suicide and aggression in alcoholics and their relationship to plasma amino acids. PsychiatryRes 12: 219-226. Branchey MI-I, Buydens-Branchey L, Lieber CS (1988) P3 in alcoholics with disordered regulation of aggression. PsychiatryRes 25: 49-58. Breier A, Chamey DS, Meninger CR (1986) Intravenous diaxepam fails to change growth hormone and cortisol secretion in humans. Psychiatry Res 18:293-299. Brown GL, Ballenger JC, Minichiello MD, Goodwin FK (1981) Human aggression and its relationship to CSF 5-l-BAA, MBPG, and HVA. In: Sandler M (Ed) Psychopharmacology of Aggression. Raven Press, New York, pp 131-147. Buss AH, Durkee A (1957) An inventory for assessing different kinds of hostility. J Consult Clin Psycho1 21: 343-349. Buydens-Branchey L, Branchey M, Noumair D, Lieber CS (1989) Age of alcoholism onset: II. Relationship to susceptibility to serotonin precursor availability. Arch Gen Psychiatry 46: 231-236. Carroll BJ (1982) ‘Ibe dexamethasone suppression test for melancholia. Br J Psychiatry 140:292-304. Da&is CA, Stuckey RF, Gold MS, Pottash LC (1986) Dexamethasone suppression test testing of depressed alcoholics. Alcohol Clin Exp Res 10:59-60. Del Port0 JA, Monteiro MG, Laranjeira RR, Jorge MR, Masur J (1985) Reversal of abnormal dexamethasone suppression test in alcoholics abstinent for four weeks. Biol Psychiatry 20: 1156-1160. Irwin MR. Risch SC, Brown SA, Howard TL, Smith TL, Knapp S, Schuckit MA (1988) Urinary free cords01 in depressed alcoholic patients. Biol Psychiatry 24: 713-716. Johnson B, Perry JC (1986) The relationship between depression and the dexamethasone suppression test following alcohol withdrawal in a psychiatric population. J Clin Psychophawnacol6: 343-349. Mendelson JH, Ogata M, Mello NK (1971) Adrenal function and alcoholism. I. Serum cortisol. Psychosom Med 33: 145-157. Mendlewicz J. Charles G, Fran&son JM (1982) The dexamethasone suppression test in affective disorder: relationship to clinical and genetic subgroups. Br J Psychiatry 141: 464-470. Monroe RR (1986) Episodic behavioral disorders ad limbic ictus. In: Doane BK, Livingstone KE (Eds) The Limbic System: Functional Organization and Clinical Disorders. Raven Press, New York, pp 251-266. Muldoon TG, Westphal U (1967) Steroid-protein interactions. XI. Isolation and characterization of corticosteroid-binding globulin from human plasma. J Biol Chem 242: 5636-5643.

54

L. BUYDENS-BRANCHEY & M. H. BRANCHEY

Nelson WH, Sullivan P, Khan A, Tamragouri RN (1986) The effect of age on dexamethasone suppression test results in alcoholic patients. Am J Psychiatry 143: 237-239. Post RM (1986) Dose limbic system dysfunction play a role in affective illness? In: Doane BK, Livingstone KE (Eds) The Limbic System: Funcdonal Organization and Clinical Disorders. Raven Press, New York, pp 229-249. Ravi SD, Dorus W, Park YN, Collins MC, Reid RW, Borge GF (1984) The dexamethasone suppression test and depressive symptoms in early and late withdrawal from alcohol. Am J Psychiatry 141: 1445-1448. Sapolsky RM, I&y LC, McEwen BS (1985) Prolonged glucocorticoid exposure reduces hippocampal neuron number: implication for aging. J Neurosci 5: 1222- 1227. Spitzer RL, Endicott J (1978) Schedule for Affective Disorders and Schizophrenia. New York State Psychiatric Institute, New York. Spitzer RL, Williams JBW (1984) Structured Clinical Interview for DSM-III (SCID). New York State Psychiatric Institute, New York. Stokes PE (1973) Adrenocortical activation in alcoholics during chronic drinking. Ann NY Acad Sci 215: 77-82. Stokes PE, Sikes CR (1988) The hypothalamic-pituitary-adrenocortical axis in major depression. Neural Clin 6: 1-19. Targum SD, Wheadon DE, Chastek CT, McCabe WJ, Advani MT (1982) Dysregulation of hypothalamic-pituitary-adrenal axis function in depressed alcoholic patients. J Affect Disord 4: 347-353. Truffinet P, Lecrubier Y (1988) La s&r&ion de cords01 chez le dkprimd. Encephale 14: 1-17. Willenbring ML, Morley JE, Niewoehner CB, Heilman RO, Carlson CH, Shafer RB (1984) Adrenocortical hyperactivity in newly admitted alcoholics: prevalence, course and associated variables. Psychoneuroendocrinology 9: 415-422. Wilson SAK (1960) Neurology. Edward Arnold, London. Zem MA, Halbreich U, Bacon K, Galanter M, Kang BJ, Gasparini F (1986) Relationship between serum cortisol, liver function, and depression in detoxified alcoholics. Alcohol Clin Eq Res 10:320-322.