Brief Communication

Depressive Symptoms and Alcohol Consumption Among Nonalcoholic Depression Patients Treated With Desipramine

Benjamin I Goldstein, MD, PhD¹, Ayal Schaffer, MD, FRCPC², Anthony Levitt, MD, FRCPC³, Ari Zaretsky, MD, FRCPC², Russell T Joffe, MD, FRCPC⁴, Virginia Wesson, MD¹, R Michael Bagby, PhD³

Objective: There are few data addressing the effect of alcohol consumption on response to antidepressants among nonalcoholics with depression. Similarly, the effect of antidepressant treatment on alcohol consumption in this group is not yet understood. This study focuses on changes in depressive symptoms and alcohol consumption in response to treatment with desipramine. Method: Twenty-seven nonalcoholic outpatients with major depression (as determined by the Schedule for Affective Disorders and Schizophrenia-Lifetime Version) completed measures of depression (that is, the 17-item Hamilton Depression Rating Scale and the Beck Depression Inventory) and alcohol consumption at intake and after 5 weeks of open treatment with desipramine. Subjects were characterized as minimal or mild-to-moderate drinkers. Results: There was no significant difference between the groups with respect to effectiveness of antidepressant treatment. Analysis for repeated measures demonstrated that alcohol consumption with desipramine was significantly lower after treatment than at intake (F = 4.8, df 23:2, P < 0.01). Further, carbohydrate consumption was also significantly lower after treatment than at intake (F = 4.4, df 23:2, P < 0.05). Conclusion: Desipramine treatment appeared to result in decreases in alcohol consumption in nonalcoholic patients with depression. Further research is needed to elucidate the effect of alcohol consumption on the course and outcome of major depressive illness among nonalcoholics as well as the effect of antidepressant medication on alcohol consumption in this population. (Can J Psychiatry 2004;49:859–862) Click here for author affiliations.  Clinical Implications Desipramine may attenuate alcohol consumption among nonalcoholics with depression. Previous studies have suggested that nonalcoholic patients with depression would benefit by minimizing their consumption of alcohol. Changes in alcohol and carbohydrate consumption during antidepressant treatment may be independent of changes in depressive symptoms. Limitations The study sample was relatively small. The length of this trial may not have been sufficient to detect the effect of alcohol on response to desipramine. Alcohol consumption was determined by self-report and may be an underestimate of actual alcohol consumption. Key Words: depression, alcohol consumption, desipramine, carbohydrate consumption Résumé : Symptômes dépressifs et consommation d’alcool chez des patients non alcooliques souffrant de dépression et traités à la désipramine

The cooccurrence of alcohol use disorders has been reported to adversely affect the course, treatment, and prognosis of major depressive disorder (1,2). What is less clearly understood is the significance of mild-to-moderate alcohol consumption on treatment response among patients with major depression. To our knowledge, Worthington and colleagues reported the only findings regarding nonabusive alcohol consumption in depression patients treated with fluoxetine (3). They found that the level of alcohol consumption at baseline predicted significantly poorer response to treatment and a nonsignificant trend toward lowered alcohol consumption in the treatment group. There are, however, no previous studies of tricyclic antidepressants (TCAs) in this respect. 

This study prospectively examines the effect of pretreatment alcohol consumption on response to treatment with desipramine, as well as the effect of desipramine treatment on alcohol consumption in nonalcoholic subjects with depression. We hypothesized that subjects who abstained or drank minimally would show greater depressive symptom reduction than would those with moderate alcohol consumption. 

Method

The study sample comprised 27 consecutive outpatients (21 women and 6 men) referred to the Depression Clinic at the Clarke site of the Centre for Addiction and Mental Health, University of Toronto. At baseline, after they provided written informed consent, all patients were administered the Schedule for Affective Disorders and Schizophrenia Lifetime Version (SADS-LV, 4), the 17-item Hamilton Depression Rating Scale (HDRS, 5), the Beck Depression Inventory (BDI, 6), and the Food Pattern Questionnaire (FPQ, 7). 

The SADS-LV and the HDRS interviews were conducted by a psychiatric nurse trained in the administration of these instruments and blind to the patients’ responses on the self-report measures (that is, the BDI and the FPQ). According to the SADS-LV, all patients in the study fulfilled research diagnostic criteria (RDC, 8) for unipolar, nonpsychotic major depression. None of the patients had a concurrent medical illness that was unstable or that would preclude antidepressant treatment; all were medication-free for a minimum of 2 weeks prior to the study. All patients then received 5 weeks of treatment with open-label desipramine at a target daily dosage of 2.5 mg per kg of body weight (or less if this was not tolerated). Participants then completed posttreatment HDRS, BDI, and FPQ. 

The FPQ requires that participants indicate the consumption frequency of various foods and beverages, as follows: never or almost never, 1 to 3 times monthly, 1 to 3 times weekly, 4 to 6 times weekly, or daily. The questionnaire was scored in Likert-type fashion, with raw scores of 0 to 4 for each of the categories, respectively. The FPQ categories of interest for this investigation include carbohydrates (22 items), non-alcoholic beverages (7 items), and alcoholic beverages (4 items). 

We determined drinking status according to frequency of alcohol consumption. Minimal alcohol consumption was defined as a score of less than 2 on the alcohol category of the FPQ (that is, the sum of the scores on beer, wine, liquor, and liqueur items, with a maximum possible score of 16 and a maximum reported score of 7). Mild-to-moderate alcohol consumption was defined as a score of 2 or greater on the alcohol category of the FPQ. 

We performed all statistical analyses, using the Statistical Package for the Social Sciences, version 11.0 (9). Minimal drinkers were compared with mild to moderate drinkers with respect to demographic variables. We performed chi-square analysis to compare the percentage of female patients in each group, and we used t tests to compare mean age at baseline and mean age at onset of depression. 

We determined the difference in the change in alcohol consumption for the minimal vs mild-to-moderate alcohol consumption groups, using a 2 × 2 (that is, group by time) repeated-measures analysis of variance (ANOVA). Similarly, we used ANOVA for repeated measures to evaluate whether change over time in depressive symptoms was different in the minimal vs mild-to-moderate alcohol consumption groups, with both change in carbohydrate consumption and change in nonalcoholic beverage consumption included as covariates. 

Results

We recruited 27 subjects (6 men and 21 women). Demographic data are as follows: 83% of subjects in the mild-to-moderate alcohol consumption group were women, and 67% of subjects in the minimal alcohol consumption group were women (c ² = 0.91, P = ns). The mean age of subjects in the mild-to-moderate alcohol consumption group was 36.7 years, SD 8.1, compared with 38.8 years, SD 10.5, among subjects in the minimal alcohol consumption group (t = 0.57, P = ns). Mean age at onset of depression was 24.1 years, SD 12.0, among subjects in the mild-to-moderate alcohol consumption group, compared with 27.1 years, SD 13.7, among subjects in the minimal alcohol consumption group (t = 0.53, P = ns).

Mean pre- and posttreatment scores for the study outcome measures are presented in Table 1. The notable findings are as follows: There was a significant main effect of group on alcohol consumption, as expected. There was also a significant effect of time on alcohol consumption, that is a reduction of alcohol consumption with treatment. However, there was no significant group-by-time interaction and no significant main effect of time, group, or group-by-time interaction on nonalcoholic beverage consumption. There was a significant reduction of carbohydrate consumption over time; however, there was no significant effect of group and no significant group-by-time interaction. When we used an analysis of covariance (ANCOVA) to control for depressive symptoms, the findings for alcohol (F = 8.2, df 23:2, P < 0.01) and carbohydrate (F = 4.4, df 23:2, P < 0.05) consumption remained significant.

Table 1  Clinical characteristics of alcohol consumption groups
	Alcohol consumption		Analysis of variance
	Mild-to-moderate mean (SD)	Minimal mean (SD)	Group Fa	Time Fa	Group ´ time Fa
Beck Depression Inventory			0.04	38.6***	0.04
Pretreatment	31.3 (7.8)	31.6 (5.2)
Posttreatment	18.4 (12.7)	19.6 (7.7)
Hamilton Depression Rating Scale			0.55	25.8***	0.05
Pretreatment	18.9 (3.3)	20.1 (3.9)
Posttreatment	11.5 (6.4)	13.2 (19.6)
Carbohydrate consumptionb			2.9	4.4*	0.04
Pretreatment	21.7 (6.8)	26.1 (7.9)
Posttreatment	18.8 (7.0)	23.8 (8.5)
Nonalcoholic beverage consumptionc			0.09	0.08	0.08
Pretreatment	10.4 (3.4)	10.7 (4.4)
Posttreatment	10.4 (3.1)	11.0 (3.7)
Alcoholic beverage consumptiond			14.3***	5.6*	4.8*
Pretreatment	3.7 (1.6)	0.4 (0.5)
Posttreatment	2.2 (1.9)	0.3 (0.5)
*P < 0.05; **P < 0.01; ***P < 0.001 adf = 23:2 bSum of the scores on 22 items (each item scored on 0 to 4 Likert scale: 0 = never or almost never, 1 = 1 to 3 times monthly, 2 = 1 to 3 times weekly, 3 = 4 to 6 times weekly, 4 = daily; cSum of the scores on 7 items; dSum of the scores on 4 items

Using Pearson’s correlation coefficient, we found no significant correlation between change in alcohol consumption and change in depressive symptoms among the sample as a whole (BDI r = 0.11, P = ns; HDRS r = 0.02, P = ns). 

There was a trend toward significant correlation between change in carbohydrate consumption and change in depressive symptoms among the sample as a whole (BDI r = 0.37, P = 0.06; HDRS r = 0.36, P = 0.07). The correlation between change in carbohydrate consumption and change in alcohol consumption was not significant (r = 0.03, P = ns). 

Discussion

To our knowledge, the present study is the first to report a decrease in the alcohol consumption of nonalcoholics with depression treated with a TCA. In a group of nonalcoholic patients with depression, scores on an alcohol measure decreased by approximately 40% after 5 weeks of treatment with desipramine. By comparison, SSRIs have been reported to reduce alcohol consumption among patients with alcoholism by 15% to 20% (10). However, direct comparison is limited by the nature of the differences in both the clinical characteristics of these groups (that is, persons with and without alcoholism) and the alcohol measures used. Nonetheless, this finding remains notable in that patients were not specifically instructed to decrease their alcohol consumption during the treatment period. 

The absence of any significant change in nonalcoholic beverage consumption suggests that the decrease in alcohol consumption is not owing to an effect of desipramine on beverage consumption in general. There was, however, a decrease in carbohydrate food intake. Though there was no significant correlation between the change in alcohol consumption and the change in carbohydrate consumption, it remains possible that desipramine impacts a common element of carbohydrate consumption in general, whether in the form of food or alcohol. The decrease in alcohol consumption may also be due to the effect of time or such nonspecific factors as contact with research assistants associated with clinical trial involvement.

The data did not support the primary hypothesis that the effect of desipramine on depressive symptoms would differ, based on the frequency of alcohol consumption. One possible reason for this is that the amount of alcohol consumed was not sufficient to render an effect on depressive symptoms. The FPQ does not provide a quantitative volumetric assessment of alcohol consumption; thus among individuals with a similar frequency of alcohol consumption, the amount consumed might have varied. Nonetheless, the FPQ remains a useful instrument for assessing alcohol consumption because it allows for the categorization of alcohol consumption among individuals who would not be distinguished by using the SADS-LV or standard alcoholism screening instruments. 

Methodological limitations of the present study include the small sample size and the exclusive use of self-report in determining alcohol consumption. A longer follow-up duration might also have revealed unrecognized between-group differences in depressive symptoms. 

The primary clinical implication of this study is that nonalcoholic patients with depression may derive the additional benefit of less frequent alcohol consumption as a result of treatment with desipramine. Future studies with large samples incorporating quantification of the amount of alcohol consumed are needed to confirm these findings.

References

1. Grant BF, Harford TC. Comorbidity between DSM-IV alcohol use disorders and major depression: results of a national survey. Drug Alcohol Depend 1995;39:197–206.

2. Leibenluft E, Madden PA, Dick SE, Rosenthal NE. Primary depressives with secondary alcoholism compared with alcoholics and depressives. Compr Psychiatry 1993;34:83–6.

3. Worthington J, Fava M, Agustin C, Alpert A, Nierenberg AA, Pava JA, and others. Consumption of alcohol, nicotine, and caffeine among depressed outpatients. Psychosomatics 1996;37:518–22.

4. Spitzer RL, Endicott J. Schedule for affective disorders and schizophrenia lifetime version. 3rd ed. New York (NY): New York State Psychiatric Institute; 1979.

5. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psych 1960;23:56–62.

6. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71.

7. The DCCT research group. The diabetes control and complications trial (DCCT): design and methodologic considerations for the feasibility phase. Diabetes 1986;35:530–45.

8. Spitzer R, Endicott J, Robins LN. Research diagnostic criteria: rationale and reliability. Arch Gen Psychiatry 1978;35:773–82.

9. SPSS Inc. SPSS Base 10.0 Applications Guide. Chicago (IL): SPSS Inc; 1999.

10. Naranjo CA, Knoke D. The role of selective serotonin reuptake inhibitors in reducing alcohol consumption. J Clin Psychiatry 2001;62(Suppl 20):S18–S25.

Author(s)

Manuscript received October 2003, revised, and accepted November 2003.

1Resident, Department of Psychiatry, University of Toronto, Toronto, Ontario.

2Assistant Professor, Department of Psychiatry, University of Toronto, Toronto, Ontario.

3Professor, Department of Psychiatry, University of Toronto, Toronto, Ontario.

4Professor, Department of Psychiatry and Dean, UMDNJ-New Jersey Medical School, Newark, New Jersey.

Address for correspondence: Dr AJ Levitt, Department of Psychiatry, Sunnybrook and Women’s College Health Sciences Centre, Room FG 46, 2075 Bayview Avenue, Toronto, ON M4N 3M5

e-mail: anthony.levitt@sw.ca