University of Texas Health Science Center, START Center, 7703 Floyd Curl Drive, Mail Stop 7792, San Antonio, TX 78229, USA
Received 14 October 2002; in revised form 13 December 2002; in revised form 19 March 2003; accepted 21 March 2003
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In addition to the role of ondansetron by alcoholic subtype, changes in anxiety may contribute to improved drinking. Anxiety is common among treatment-seeking alcoholics (Ross et al., 1988) and its presence may be associated with relapse (Kushner et al., 1990
, 2000a
,b
). It would therefore be reasonable to predict that decreased anxiety among treatment-seeking alcoholics would facilitate a reduction in drinking behaviour. As an adjunct to our previous finding that dose of ondansetron, age of onset and pre-treatment drinking predict post-treatment drinking behaviour, we tested the hypothesis that reduced anxiety is associated with improvement in drinking outcomes above and beyond subtype differences associated with ondansetron response.
![]() |
SUBJECTS AND METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Design
Study design was a 2 (early onset alcoholism versus late onset alcoholism) x 4 (ondansetron doses of 0, 1, 4 and 16 µg/kg) factorial design. Groups were balanced with respect to age, gender and past 90 days drinking level. Abstinence was not a study entry criterion; however, participants reported a desire to stop drinking and to participate in psychosocial treatment. All subjects received weekly standardized CBT based on The Cognitive Behavioral Coping Skills Therapy Manual (Kadden et al., 1994), and Treating Alcohol Dependence: A Coping Skills Training Guide (Monti et al., 1989). During the course of the 12-week study, subjects received CBT and either ondansetron or placebo.
Experimental questionnaires
We examined subject responses to two of the administered questionnaires as follows.
(1) A scheduled assessment of drinking (time-line follow-back). Drinking behaviour was measured as: (1) drinks per day the average number of drinks for the week; (2) drinks per drinking day the average number of drinks the individual consumed when he drank; (3) per cent days abstinent the percentage of days out of the week the individual did not drink at all. Data were collected at baseline (prior to double-blinding) and the last treatment visit attended (end-state). It should be noted that using end-state drinking as the outcome measure differs from our previous analyses, in which treatment outcome was defined as average drinking during treatment. For the purposes of this analysis, end-state drinking was selected as the best representation of treatment outcome, because we wanted to determine the effects of anxiety on participants final drinking level after undergoing treatment. In addition, we sought to include a more diverse sample of treatment-seeking alcoholics, participants who completed treatment and those who prematurely terminated treatment. Therefore, instead of being required to complete the full 12 weeks of treatment, all participants who were enrolled for at least 3 weeks were included. We chose a cut-off of 3 weeks so that all participants would have received at least 2 weeks of their assigned treatment (psychotherapy plus ondansetron 1, 4 or 16 µg/kg twice a day, or placebo). On their first week all participants received placebo, while the medication was dispensed on subsequent weeks according to randomization.
(2) The Profile of Mood States tensionanxiety scale (POMS).
The POMS (McNair et al., 1971) is a 65-item adjective checklist rated by the subject on a 5-point Likert scale. It consists of six factors: tensionanxiety, depressiondejection, anger hostility, vigor, fatigue and confusion. For the present study we used only the tensionanxiety scale, which is a 9-item scale with scores ranging from 0 to 36. Subjects completed the scale at pre-treatment and at their last visit (end-state). Change in anxiety during the course of treatment was calculated by subtracting the two scores.
Statistical analyses
Using a type III least squares regression weighted by completion status, we first reconstructed the model used in our previously published results from this clinical trial. We replicated the finding that pre-treatment drinking and the interaction between ondansetron dose and age of onset predicted drinking behaviour. Next, we looked at adding baseline anxiety and change in anxiety to determine whether either affected the model. Change in anxiety from baseline to last visit contributed significantly to the model; baseline anxiety alone did not. We therefore added change in anxiety level to the model. Then we recalculated the regression equation. The resulting model was compared with our original model to determine whether the addition of the variable (change in anxiety) enhanced our ability to predict end-state drinking. This was accomplished by comparing the adjusted R2 value. Finally, we considered how anxiety changed during treatment by looking at descriptive statistics and comparing pre-treatment anxiety with end-state anxiety using a paired t-test.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
At baseline the average number of drinks per day was 3.7 ± 3.8, that of drinks per drinking day was 5.5 ± 4.3 and per cent days abstinent was 0.43 ± 0.36. Both drinks per day and drinks per drinking day ranged from 0 to 27, and per cent days abstinent ranged from 0 to 1. At end-state the average number of drinks per day was 2.0 ± 2.7, drinks per drinking day was 3.3 ± 3.6 and per cent days abstinent was 0.59 ± 0.39. Both drinks per day and drinks per drinking day ranged from 0 to 17, and per cent days abstinent ranged from 0 to 1. Drinking decreased significantly from baseline to the last visit (all P values were <0.001).
Using regression equations we examined the relationship between anxiety and drinking. Table 1 shows the results obtained in our replication of the original model. As before, we found that pre-treatment drinking and the interaction between ondansetron dose and age of onset predicted drinking outcome (drinks per day, drinks per drinking day and per cent days abstinent). Next we examined the effects of adding baseline anxiety and change in anxiety level to the above-mentioned equation. Change in anxiety made a significant contribution to the equation [for drinks per day: F(1,239) = 10.6, P < 0.001; for drinks per drinking day: F(1,239) = 12.71, P < 0.001; for per cent days abstinent: F(1,239) = 6.02, P < 0.05], whereas baseline anxiety did not (all P values >0.5). Therefore, we added change in anxiety to the model. We also examined its interaction with age of onset and ondansetron dose. We found that the two-way and three-way interactions were not significant predictors of end-state drinking (all P values >0.15). The variables that accounted for a significant amount of the variance in end-state drinking were changes in anxiety, pre-treatment drinking, and the interaction between ondansetron dose and age of onset (all P values <0.05). Beta weights are presented in Table 2
. Table 3
shows the R2adj, for the two models. For all three drinking measures, adding changes in anxiety to the original prediction equation significantly increased the adjusted R2 (all P values <0.05).
|
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Next, as hypothesized, we found that anxiety was a significant predictor of drinking outcome. Changes in anxiety predicted end-state drinking above and beyond age of onset and ondansetron dose. Also, changes in anxiety uniformly predicted drinking behaviour for both early and late onset alcoholics, as well as across ondansetron dose (including placebo). Therefore, change in anxiety was an independent factor affecting drinking outcome. The correlation between the factors was positive, with those reporting greater decreases in anxiety tending to be drinking less. One possible explanation for these results is that reductions in anxiety resulted in decreased drinking behaviour. This explanation is consistent with the tension reduction hypothesis and its corollaries, which state that some people use alcohol to cope with stress (Cappell and Greeley, 1987; Sher, 1987
). In support of the contention that individuals use alcohol to decrease stress and anxiety, alcohol use tends to reduce panic attacks (Kushner et al., 1996
), simple phobic responses (Lindman et al., 1980
) and performance anxiety (Abrams et al., 2001
). In fact, just the expectation of reduced anxiety predicts alcohol use (Kushner et al., 1994
; Cooper et al., 1995
). In addition, medication trials have found that alcoholics with comorbid anxiety decrease anxiety as well as drinking behaviour in response to an anxiolytic. For example, 65% of patients with alcoholism and panic disorder showed improvement in global functioning when treated with the combination of the anxiolytic fluoxetine and disulfiram. Also, those with comorbid alcoholism and anxiety drank less when receiving the anxiolytic busiprone, compared with placebo. In contrast, a third study found no difference between busiprone and placebo (for a thorough review see Swift, 1999
). Recently, Randall et al. (2001)
have examined the effects of psychotherapy on individuals diagnosed with comorbid alcoholism and anxiety. In that study, participants were treated with either standard therapy for alcoholism or standard therapy plus exposure therapy for social phobia. Surprisingly, those receiving the combined therapy, compared with the standard therapy alone, drank more. In explaining these unexpected findings, the above authors suggested that methodological complications may have influenced the results. For example, individuals may have been drinking during the exposure exercises, thus artificially inflating alcohol use. Therefore, although it seems plausible to suggest that individuals use alcohol to cope with anxiety, treatment targeting the underlying anxiety amongst alcoholics has not consistently resulted in the expected reduction in drinking behaviour.
An alternative explanation for the relationship between alcohol use and anxiety is that alcohol use may actually increase anxiety. This hypothesis stems from the finding that alcohol withdrawal can result in anxiety, irritability, and sleep and mood disturbances, even if the withdrawal symptoms are minimal and occur between drinking episodes (Hall, 1984; Roelofs and Dikkenberg, 1987
). Another possibility is that the relationship between drinking behaviour and anxiety was due to the influence of a third factor. One possibility is that the relationship was due to the CBT that all participants received. The psychotherapy may have enhanced coping ability, social support (Bandura, 1969
) and confidence levels (Marlatt and Gordon, 1985
), all of which may have reduced both anxiety and drinking behaviour. Alternatively, a situational variable, such as coming to the clinic, may have resulted in decreases in both. This is because the participants may have felt less stress about their drinking problem once they had taken the first step towards obtaining help. At the same time, drinking levels may have decreased because the participants focused on how much they were drinking, and were therefore motivated to decrease their consumption.
One limitation of the present study is that we could not distinguish between state and trait anxiety levels, because our only measure of anxiety was the POMS. The POMS, a measure of state anxiety, is sensitive to mood fluctuations. A measure of trait anxiety would be less influenced by transient changes and situational influences. Another limitation is that the majority of our participants reported low pre-treatment anxiety. Interestingly, we nonetheless found a relationship between changes in anxiety and drinking outcome. Thus our results suggest that this relationship may exist for non-pathologically anxious alcoholics.
Future studies should measure both state and trait anxiety to enhance our understanding of the relationship between anxiety and drinking. In addition, these findings could be extended by addressing whether results generalize to a patient population with a wider range of anxiety levels (e.g. a dual diagnosis population with comorbid anxiety). As ondansetron may have anxiolytic effects (Freeman et al., 1997; Roychoudhury and Kulkarni, 1997
), future studies should address whether such effects affect drinking in individuals with comorbid anxiety and alcoholism.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
FOOTNOTES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders, 3rd edn, revised. American Psychiatric Association, Washington, DC.
Bandura, A. (1969) Principles of Behavior Modification. Holt, Rinehart, & Winston, New York.
Cappell, H. and Greeley, J. (1987) Alcohol and tension reduction: An update on research and theory. In Psychological Theories of Drinking and Alcoholism, Blane, H. T. and Leonard, K. E., eds, pp. 1453. Guilford Press, New York.
Cooper, M., Frone, M., Russell, M. and Mudar, P. (1995) Drinking to regulate positive and negative emotions: A motivational model of alcohol use. Journal of Personality and Social Psychology 69, 9901005.[CrossRef][ISI][Medline]
Freeman, A. M., III, Westphal, J. R., Norris, G. T., Roggero, B. A., Webb, P. B., Freeman, K. L., Rush, J. A., Hearne, E. M., III and Evoniuk, G. (1997) Efficacy of ondansetron in the treatment of generalized anxiety disorder. Depression and Anxiety 5, 140141.[CrossRef][Medline]
Hall, S. M. (1984) The abstinence phobias: Links between substance abuse and anxiety. International Journal of the Addictions 19, 613631.[ISI][Medline]
Johnson, B. A., Cloninger, C. R., Roache, J. D., Bordnick, P. S. and Ruiz, P. (2000a) Age of onset as a discriminator between alcoholic subtypes in a treatment-seeking outpatient population. American Journal on Addictions 9, 1727.[CrossRef][ISI][Medline]
Johnson, B. A., Roache, J. D., Javors, M. A., DiClemente, C. C., Cloninger, C. R., Prihoda, T. J., Bordnick, P. S., Ait-Daoud, N. and Hensler, J. (2000b) Ondansetron for reduction of drinking among biologically predisposed alcoholic patients: A randomized controlled trial. Journal of the American Medical Association 284, 963971.
Kushner, M. G., Sher, K. J. and Beitman, B. D. (1990) The relation between alcohol problems and the anxiety disorders. American Journal of Psychiatry 147, 685695.[Abstract]
Kushner, M. G., Sher, K. J., Wood, M. D. and Wood, K. W. (1994) Anxiety and drinking behavior: Moderating effects of tension reduction alcohol outcome expectancies. Alcoholism: Clinical and Experimental Research 18, 852860.[ISI][Medline]
Kushner, M. G., Mackenzie, T. B., Fiszdom, J., Valentiner, D. P., Foa, E. and Wangensteen, D. (1996) The effects of alcohol consumption on laboratory induced panic and state anxiety. Archives of General Psychiatry 53, 264270.[Abstract]
Kushner, M. G., Abrams, K., Thuras, P. and Hanson, K. L. (2000a) Individual differences predictive of drinking to manage anxiety among non-problem drinkers with panic disorder. Alcoholism: Clinical and Experimental Research 24, 448458.[ISI][Medline]
Kushner, M. G., Thuras, P., Abrams, K., Brekke, M. and Stritar, L. (2000b) Anxiety mediates the association between anxiety sensitivity and coping-related drinking motives in alcoholism treatment patients. Addictive Behaviors 26, 869885.[CrossRef][ISI]
Lindman, R., Alexanderson, G. and Lindfors, B. (1980) Anxiety and alcohol: limitation of tension reduction theory in nonalcoholics. Report from the Department of Psychology at ABO Akademi, Monograph Supplement 1. Turku, Finland: ABO Academi University Press.
Lynskey, M. T. (1998) The comorbidity of alcohol dependence and affective disorders: Treatment implications. Drug and Alcohol Dependence 52, 201209.[CrossRef][ISI][Medline]
Marlatt, G. A. and Gordon, J. R. (1985) Relapse Prevention. Guildford Press, New York.
McNair, D., Lorr, M. and Droppleman, L. (1971) Profile of Mood States. Educational and Industrial Testing Service, San Diego, CA.
Randall, C. L., Thomas, S. and Thevos, A. K. (2001) Concurrent alcoholism and social anxiety disorder: A first step toward developing effective treatments. Alcoholism: Clinical and Experimental Research 25, 210220.[CrossRef][ISI][Medline]
Roelofs, S. M. and Dikkenberg, G. M. (1987) Hyperventilation and anxiety: Alcohol withdrawal symptoms decreasing with prolonged abstinence. Alcohol 4, 215220.[CrossRef][ISI][Medline]
Ross, H. E., Glaser, F. B. and Germanson, T. (1988) The prevalence of psychiatric disorders in patients with alcohol and other drug problems. Archives of General Psychiatry 45, 10231031.[Abstract]
Roychoudhury, M. and Kulkarni, S. K. (1997) Antianxiety profile of ondansetron, a selective 5-HT3 antagonist, in a novel animal model. Methods and Findings in Experimental and Clinical Pharmacology 19, 107111.[ISI][Medline]
Sher, K. J. (1987) Stress response dampening. In Psychological Theories of Drinking and Alcoholism, Blane, H. T. and Leonard, K. E., eds, pp. 227271. Guilford Press, New York.
Swift, R. M. (1999) Drug therapy: drug therapy for alcohol dependence. New England Journal of Medicine 340, 14821490.