IMPAIRED SERUM CORTISOL STRESS RESPONSE IS A PREDICTOR OF EARLY RELAPSE

K. Junghanns1,*, J. Backhaus1, U. Tietz1, W. Lange2, J. Bernzen3, T. Wetterling4, L. Rink5 and M. Driessen2

1 Department of Psychiatry and Psychotherapy, University of Luebeck,
2 Gilead IV, Psychiatric Hospital, Bielefeld,
3 Clinic Holstein, Luebeck,
4 Department of Psychiatry and Psychotherapy, University of Frankfurt/Main and
5 Institute of Immunology, University of Aachen, Germany

Received 1 October 2002; in revised form 14 November 2002; accepted 18 November 2002


    ABSTRACT
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Aims: to investigate a possible association of cortisol stress response during early abstention with relapse. Methods: Thirty-six alcohol-dependent males, half of them with a comorbid anxiety disorder, and 15 healthy controls were exposed to a standardized psychosocial stress test. Thirty-one of the patients were assessed for relapse 6 weeks after discharge. Results: The relapsers showed almost no cortisol responses in the stress test. Comorbid anxiety disorder influenced neither stress response nor relapse. Conclusions: During early abstention from alcohol, reduced stress-responsivity of the hypothalamo–pituitary–adrenocortical axis seems to be connected to early relapse.


    INTRODUCTION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Chronic alcohol consumption is associated with several abnormalities in the regulation of the hypothalamo– pituitary–adrenocortical (HPA) axis. Thus, drinking alcoholics show a blunted response of adrenocorticotropic hormone (ACTH) to several acute intervening stressors (Berman et al., 1990Go; Wand and Dobs, 1991Go). During withdrawal, elevation of cortisol and ACTH levels could be shown (Bannan et al., 1984Go; Iranmanesh et al., 1989Go; von Bardeleben et al., 1989Go; Adinoff et al., 1991Go; Heinz et al., 1995Go; Hundt et al., 2001Go), especially in alcoholics with a comorbid depression (Burov et al., 1986Go).

Two to 6 weeks after cessation of drinking, baseline plasma, salivary and urinary cortisol values return to normal, but the cortisol response to stress remains blunted (Errico et al., 1993Go; Bernardy et al., 1996Go; Lovallo et al., 2000Go), with a gradual, but incomplete, normalization of HPA parameters within 12 weeks (Ehrenreich et al., 1997Go).

As yet, no study has investigated the relationship between the HPA axis stress response and relapse. Along with a gradual normalization of ACTH and cortisol responses, the risk of relapse also decreases with sustained abstention (Whitworth et al., 1996Go; Project Research Group MATCH, 1997Go). On the assumption that these two phenomena are related, we hypothesized that alcoholics with a particularly blunted HPA axis response during early abstention have an increased risk of relapse, since their stress responsivity is especially disturbed. As comorbid anxiety disorder was shown to be associated with an increased risk of relapse (Driessen et al., 2001Go), and as the effect of comorbid anxiety disorder on HPA axis responsivity was unclear, early abstinent alcohol-dependent patients with and without comorbid anxiety disorder were included in the study.


    PATIENTS AND METHODS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
The study started after approval through the local ethics committee. All subjects gave their written informed consent to participate. All alcoholics fulfilled the DSM-IV criteria for alcohol dependence (American Psychiatric Association, 1994Go). They were recruited from our in-patient motivation enhancement therapy, which consists of a 3-week psychotherapeutic programme starting after detoxification. They had to have no major organic comorbidity and had to be unmedicated for at least 1 week. They did not have any clinical symptoms of alcohol withdrawal at this time. The psychiatric diagnoses were established with the help of a computerized version of the Composite International Diagnostic Interview (DIA-X; Wittchen and Pfister, 1997Go). In addition, a complete medical and psychiatric history, including data on the duration of dependence and family history of alcoholism, was taken and a complete physical examination was performed. None of the participating subjects fulfilled the criteria of an antisocial personality disorder. All participants were interviewed for nicotine and illicit drug use. There was no indication of the latter. As depression per se influences the HPA axis response, alcohol-dependent patients with a comorbid major depressive episode were excluded.

Altogether, 36 male alcohol-dependent patients were recruited after cessation of drinking for at least 2 weeks (Table 1Go): 18 of them had no additional psychiatric disorder (group A), whereas the other 18 had a current comorbid anxiety disorder as the only comorbidity (group A + A). Of the latter, 13 had a panic disorder with or without agoraphobia, four had a comorbid social phobia and two received both comorbid diagnoses. Two patients of the non-comorbid group and three patients with a comorbid anxiety disorder dropped out for follow-up, as they entered a long-term treatment within the follow-up period. They were included in the evaluation of stress response, but were not considered for the calculations on early relapse. The healthy controls (group C) were recruited by advertisements. They underwent the same diagnostic procedures and were unmedicated. The three study groups were matched with regard to age.


View this table:
[in this window]
[in a new window]
 
Table 1. Descriptive characteristics (means ± SD) for the alcoholics without any comorbidity (A), alcoholics with comorbid anxiety disorder (A + A) and healthy controls (C)
 
Immediately before the psychosocial stress test, subjects completed the state anxiety questionnaire (STAI-X1) of the State Trait Anxiety Inventory (Spielberger et al., 1970Go). The trait anxiety questionnaire (STAI-X2) was rated at the time of inclusion into the study. For studying the cortisol stress response, the standardized Trier Social Stimulation Test (TSST) was applied (Kirschbaum et al., 1992Go). During this test of about 15 min, an arithmetic task and a virtual 5-min video presentation for a job has to be prepared and presented before an evaluating emotionally neutral audience.

The experiments always started with the insertion of an intravenous cannula into an antecubital vein at 12.00. The baseline values for serum cortisol, pulse, systolic and diastolic blood pressure were assessed at 13.00 (T1), after which the TSST was performed. The response parameters were investigated again immediately after the stress test (T2), and then after 20 (T3), 40 (T4) and 60 min (T5). The collected blood was transferred to serum tubes, centrifuged, and serum was isolated and frozen at -20°C until the assay was performed. Cortisol levels were measured by specific radio-immunoassay techniques (Cortisol Assay Kit, cat. # 1841; Coulter-Immunotech Diagnostics). The lower limit of detection was 10 nmol. The intra- and interassay coefficients of variation were 3.9 and 6.6%, respectively.

Six weeks after discharge from the hospital, the patients were interviewed face to face about relapse, the alcohol level was controlled via a breathalyser and blood samples for the assessment of {gamma}-glutamyl transferase, mean corpuscular volume of the erythrocytes and carbohydrate-deficient transferrin were taken. The interviewer was blind with regard to the results of the TSST. Relapse was defined as consumption of at least one alcoholic beverage since discharge from the hospital. Confirmation of the interview data through a third person was collected. No patient denied a relapse when faced with the laboratory or third person reports indicating relapse. The day of first drink was calculated with the time-line follow-back strategy.

The data on serum cortisol were analysed applying analyses of variance (ANOVAs) for area under the curve (AUC) for the five time points. AUC was calculated according to the trapezium rule (Matthews et al., 1990Go). With respect to blood pressure, ANOVAs with repeated measurements for three-group analyses were performed. Significant group differences were evaluated with post hoc Bonferroni tests. T-tests for independent samples and {chi}2 tests were applied where appropriate. For the calculations, the statistical program SPSS version 10 was used with a two-sided alpha of 0.05.


    RESULTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Patients with and without comorbid anxiety disorder vs controls
The three groups (A, A + A, C) did not differ with respect to age and body mass index. The groups A and A + A showed no significant difference concerning duration of alcohol dependence, duration of abstention before participation in the study, and a positive family history of alcoholism. Group A + A expectedly scored higher in the state trait anxiety inventory (STAI), whereas groups A and C did not differ. Groups A and A + A did not differ with respect to relapse rate (Table 1Go).

There was a significant time effect for pulse, diastolic and systolic blood pressure (ANOVA: F = 2.68, P < 0.05 for pulse; F = 4.9, P < 0.01 for diastolic and F = 11.8, P < 0.01 for systolic pressure), but no interaction effect (time x group) for these parameters during testing (Fig. 1Go).



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 1. Blood pressure and stress test. There were significant time but no time x group effects with respect to diastolic (bottom) and systolic (top) blood pressure before and after TSST (means ± SEM). **P < 0.01.

 
While the baseline (T1) serum cortisol value showed no difference between groups (ANOVA: F = 2.13, P > 0.05), there was a significant group difference with respect to the cortisol stress response (ANOVA for AUC: F = 4.4, P < 0.05), with both groups of alcoholics showing a lower response than the healthy controls (see Fig. 2Go).



View larger version (41K):
[in this window]
[in a new window]
 
Fig. 2. Serum cortisol stress response. Area under the curve (AUC) (means ± SEM) is given for cortisol differentiated by the groups’ alcohol dependence with (A + A) and without (A) comorbid anxiety disorder, and healthy controls (C). ANOVA with Bonferroni post hoc tests: *P < 0.05; ns, not significant.

 
Relapsers vs abstainers vs controls
Abstainers, relapsers and controls did not differ with respect to the baseline cortisol value (means ± SEM: 213.9 ± 32.5, 226.9 ± 40.9 and 290.2 ± 36.6 nmol/l, respectively; ANOVA: F = 1.31, P > 0.05).

Concerning the cortisol stress response, the relapsers had a significantly lower AUC of serum cortisol than the healthy controls, while the abstainers did not differ significantly from the controls (relapse: 16 447.7 ± 2869.0; controls: 26 253.5 ± 2566.1; abstainers: 18 115.0 ± 2280.1; ANOVA: F = 2.052, P < 0.05; post hoc test: relapsers vs controls: P < 0.05; abstainers vs controls and relapsers vs abstainers: P > 0.05).

This difference between relapsers and abstainers in comparison with the healthy controls could be further substantiated by the results of the immediate cortisol response as calculated by the difference between the cortisol baseline value (T1) and the value at T2 (delta-cortisol) (Fig. 3Go). While the relapsers showed almost no response, the abstainers again did not differ significantly from the healthy controls (ANOVA: F = 3.21, P = 0.05; post hoc: relapsers vs controls: P < 0.05; relapsers vs abstainers and abstainers vs controls: P > 0.05). Delta-cortisol values above 30 nmol/l were produced only by 25% of the relapsers, as compared with 46.7% of the abstainers and 66.7% of the controls ({chi}2: 6.76, P < 0.05). In accordance with our hypothesis, there was a positive correlation between the days of abstention after discharge and delta-cortisol (r = 0.39, P < 0.05).



View larger version (25K):
[in this window]
[in a new window]
 
Fig. 3. Immediate cortisol stress response. Delta cortisol (i.e. T2 – T1) values are given for alcoholics grouped according to relapse at week 6 of follow-up (means ± SEM). ANOVA with Bonferroni post hoc tests: *P < 0.05; ns, not significant.

 
A comorbid anxiety disorder was not associated with the risk of relapse. Relapsers and abstainers both differed from controls, but not from each other with respect to trait anxiety scores (ANOVA: F = 8.73, P < 0.001; post hoc: relapsers vs controls: P < 0.01; abstainers vs controls: P < 0.01; abstainers vs relapsers: P > 0.05). For state anxiety there were no significant group differences (ANOVA: F = 1.97, P > 0.05). Furthermore, there were no significant differences between relapsers and abstainers with regard to duration of dependence (14.0 ± 6.6 vs 10.4 ± 7.5 years, t = 1.35, df = 29, P > 0.05), duration of abstinence before TSST (3.5 ± 0.5 vs 4.6 ± 0.5 weeks, t = 2.01, df = 29, P > 0.05), or a positive family history of alcoholism among first-degree relatives (54.5 vs 36.8%, {chi}2 = 0.889). Nicotine use neither correlated with basic cortisol level nor with delta-cortisol or AUC values (P > 0.05).


    DISCUSSION
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
While the attenuated cortisol stress response during early abstention shown in other studies (Errico et al., 1993Go; Bernardy et al., 1996Go; Ehrenreich et al., 1997Go; Lovallo et al., 2000Go) could be replicated in the present study, our novel finding is that of a particularly low HPA axis stress response among the early relapsers. This difference cannot be accounted for by duration of a dependent drinking pattern, duration of abstinence or state anxiety. Neither did we find an association with a comorbid anxiety disorder. The contrasting higher risk of relapse for comorbid anxiety found after a longer follow-up period (Driessen et al., 2001Go) might indicate that, with a normalizing HPA axis response, different factors become more relevant for relapse. Our results with respect to comorbid anxiety must be regarded with certain methodological reservations, however. Despite the fact that, in the applied standardized diagnostic procedure, the diagnosis is only given if the probands explicitly denied a substance induction of the symptoms, another diagnostic testing after a longer follow-up period would have been helpful to settle this difficult differential-diagnostic issue.

In contrast to the blunted HPA axis response, all groups showed an equal and significant stress reaction to the TSST with respect to blood pressure and pulse rate. In the light of other results (Ehrenreich et al., 1997Go), this can tentatively be interpreted as a normal noradrenergic stress responsiveness during early abstention.

There is an analogy of the study results on a psychosocial stress response in alcoholics with some studies on the HPA axis response to alcoholic beverages. Thus, Schuckit et al.(1987Go, 1988)Go found an attenuated ACTH and cortisol response to alcohol in non-alcoholic children of alcoholics and, in two other studies, a blunted response of ACTH and cortisol in drinking alcoholics could be demonstrated (Berman et al., 1990Go; Wand and Dobs, 1991Go). This could mean that a diminished cortisol response during early abstention is not a result of a disturbed HPA axis function due to heavy alcohol consumption, but, rather, a marker for a stronger predisposition to a dependent drinking pattern with a higher risk of relapse as a consequence. This is supported by a recently published animal study demonstrating that the increased ethanol self-administration of early abstinent male Wistar rats can be attenuated by central injection of corticotropin-releasing factor (Valdez et al., 2002Go). As our data seem to point in the same direction, it appears important to explore further a connection between HPA axis (dys-)regulation and relapse. For further studies, a larger sample size with assessment of more neurobiological parameters of the HPA axis and inclusion of female probands are recommended to increase the impact of the results.

While the present data on this small sample do not support an association of the attenuated response with the duration of alcohol dependence or with a positive family history of alcoholism, factors which usually contribute to the severity of alcoholism, it remains to be elucidated in further studies with a larger sample size, whether such an impaired HPA axis response is connected to the drinking pattern before abstention.


    ACKNOWLEDGEMENTS
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
This study was supported by a grant from the University of Luebeck.


    FOOTNOTES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
* Author to whom correspondence should be addressed at: Department of Psychiatry and Psychotherapy, University of Luebeck, Ratzeburger Allee 160, 23538 Luebeck, Germany. Back


    REFERENCES
 TOP
 FOOTNOTES
 ABSTRACT
 INTRODUCTION
 PATIENTS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 REFERENCES
 
Adinoff, B., Risher-Flowers, D., De Jong, J., Ravitz, B., Bone, G. H., Nutt, D. J., Roehrich, L., Martin, P. R. and Linnoila, M. (1991) Disturbances of hypothalamic–pituitary–adrenal axis functioning during ethanol withdrawal in six men. American Journal of Psychiatry 148, 1023–1025.[Abstract]

American Psychiatric Association (1994) Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association, Washington, DC.

Bannan, L. T., Potter, J. F., Beevers, D. G., Saunders, J. B., Walters, J. R. and Ingram, M. C. (1984) Effect of alcohol withdrawal on blood pressure, plasma renin activity, aldosterone, cortisol and dopamine beta-hydroxylase. Clinical Science 66, 659–663.[ISI][Medline]

Berman, J. D., Cook, D. M., Buchman, M. and Keith, L. D. (1990) Diminished adrenocorticotropin response to insulin-induced hypoglycemia in nondepressed, actively drinking male alcoholics. Journal of Clinical Endocrinology and Metabolism 71, 712–717.[Abstract]

Bernardy, N. C., King, A. C., Parsons, O. A. and Lovallo, W. R. (1996) Altered cortisol response in sober alcoholics: an examination of contributing factors. Alcohol 13, 493–498.[CrossRef][ISI][Medline]

Burov, Y. V., Treskov, V. G., Vedernikova, N. N. and Shevelyova, O. S. (1986) Types of alcohol withdrawal syndrome and dexamethasone suppression test. Drug and Alcohol Dependence 17, 81–88.[CrossRef][ISI][Medline]

Driessen, M., Meier, S., Hill, A., Wetterling, T., Lange, W. and Junghanns, K. (2001) The course of anxiety, depression and drinking behaviours after completed detoxification in alcoholics with and without comorbid anxiety and depressive disorders. Alcohol and Alcoholism 36, 249–255.[Abstract/Free Full Text]

Ehrenreich, H., Schuck, J., Stender, N., Pilz, J., Gefeller, O., Schilling, L., Poser, W. and Kaw, S. (1997) Endocrine and hemodynamic effects of stress versus systemic CRF in alcoholics during early and medium term abstinence. Alcoholism: Clinical and Experimental Research 21, 1285–1293.[ISI][Medline]

Errico, A. L., Parsons, O. A., King, A. C. and Lovallo, W. R. (1993) Attenuated cortisol response to biobehavioral stressors in sober alcoholics. Journal of Studies on Alcohol 54, 393–398.[ISI][Medline]

Heinz, A., Rommelspacher, H., Graf, K. J., Kurten, I., Otto, M. and Baumgartner, A. (1995) Hypothalamic-pituitary-gonadal axis, prolactin, and cortisol in alcoholics during withdrawal and after three weeks of abstinence: comparison with healthy control subjects. Psychiatry Research 56, 81–95.[CrossRef][ISI][Medline]

Hundt, W., Zimmermann, U., Pottig, M., Spring, K. and Holsboer, F. (2001) The combined dexamethasone-suppression/CRH-stimulation test in alcoholics during and after acute withdrawal. Alcoholism: Clinical and Experimental Research 25, 687–691.[CrossRef][ISI][Medline]

Iranmanesh, A., Veldhuis, J. D., Johnson, M. L. and Lizarralde, G. (1989)24-hour pulsatile and circadian patterns of cortisol secretion in alcoholic men. Journal of Andrology 10, 54–63.[Abstract/Free Full Text]

Kirschbaum, C., Wüst, S. and Hellhammer, D. (1992) Consistent sex differences in cortisol responses to psychological stress. Psychosomatic Medicine 54, 648–657.[Abstract]

Lovallo, W. R., Dickensheets, S. L., Myers, D. A., Thomas, T. L. and Nixon, S. J. (2000) Blunted stress cortisol response in abstinent alcoholic and polysubstance-abusing men. Alcoholism: Clinical and Experimental Research 24, 651–658.[ISI][Medline]

Matthews, J. N. S., Altman, D. G., Campbell, M. J. and Royston, P. (1990) Analysis of serial measurements in medical research. British Medical Journal 300, 230–235.[ISI][Medline]

Project Research Group MATCH (1997) Matching alcoholism treatments to client heterogeneity: treatment main effects and matching effects on drinking during treatment. Journal of Studies on Alcohol 58, 7–29.[ISI][Medline]

Schuckit, M. A., Gold, E. and Risch, C. (1987) Plasma cortisol levels following ethanol in sons of alcoholics and controls. Archives of General Psychiatry 44, 942–945.[Abstract]

Schuckit, M. A., Risch, S. C. and Gold, E. O. (1988) Alcohol consumption, ACTH level, and family history of alcoholism. American Journal of Psychiatry 145, 1391–1395.[Abstract]

Spielberger, C. D., Gorsuch, R. L. and Lushene, R. E. (1970) STAI, Manual for the State-Trait-Anxiety-Inventory. Consulting Psychologist Press, Palo Alto, CA.

Valdez, G. R., Roberts, A. J., Chan, K., Davis, H., Brennan, M., Zorrilla, E. P. and Koob, G. F. (2002) Increased ethanol self-administration and anxiety-like behavior during acute ethanol withdrawal and protracted abstinence: regulation by corticotropin-releasing factor. Alcoholism: Clinical and Experimental Research 26, 1494–1501.[ISI][Medline]

von Bardeleben, U., Heuser, I. and Holsboer, F. (1989) Human CRH stimulation response during acute withdrawal and after medium-term abstention from alcohol abuse. Psychoneuroendocrinology 14, 441–449.[CrossRef][ISI][Medline]

Wand, G. S. and Dobs, A. S. (1991) Alterations in the hypothalamic-pituitary-adrenal axis in actively drinking alcoholics. Journal of Clinical Endocrinology and Metabolism 72, 1290–1295.[Abstract]

Whitworth, A. B., Fischer, F., Lesch, O. M., Nimmerrichter, A., Oberbauer, H., Platz, T., Potgieter, A., Walter, H. and Fleischhacker, W. W. (1996) Comparison of acamprosate and placebo in long-term treatments of alcohol dependence. Lancet 347, 1438–1442.[ISI][Medline]

Wittchen, H.-U. and Pfister, H. (eds.) (1997) DIA-X, a computerized version of the CIDI. Swets and Zeitlinger, Frankfurt.