Comparison of Induced and Independent Major Depressive Disorders in 2,945 Alcoholics

SCHUCKIT, MAJOR Am J DEPRESSIVE Psychiatry TIPP, BERGMAN, 154:7, DISORDERS July 1997 ET AL. IN ALCOHOLICS Comparison of Induced and Independent Depressive Disorders in 2,945 Alcoholics Marc A. Schuckit, M.D., Jayson E. Tipp, M.A., Michael Bergman, B.A., Wendy Reich, Ph.D., Victor M. Hesselbrock, Ph.D., and Tom L. Smith, Ph.D. Objective: Depressive episodes among alcohol-dependent men and women are heterogeneous in causation and clinical course. This study tested three hypotheses regarding the rates and clinical characteristics of two potential subtypes of these affective states: those that appear to be substance-induced mood disorders and those that are independent major depressive episodes. Method: Semistructured, detailed interviews were administered to 2,945 alcohol-dependent subjects as part of the Collaborative Study on the Genetics of Alcoholism. With the use of a time line method for determining the type of mood disorder among probands, relatives, and comparison subjects, individuals with histories of the two types of mood disorders were compared. Results: depressive episodes with an onset before the development of alcohol dependence or during a subsequent long abstinence period (i.e., independent depressions) were observed in 15.2% of the alcoholics, while 26.4% reported at least one substance-induced depressive episode. According to a logistic regression analysis, the subjects with independent (as compared to substance-induced) major depressive episodes were more likely to be married, Caucasian, and female, to have had experience with fewer drugs and less treatment for alcoholism, to have attempted suicide, and, on the basis of personal interviews with family members, to have a close relative with a major mood disorder. Conclusions: These results support the contention that it is possible to differentiate between what appear to be substance-induced and independent depressive episodes in alcoholics. Such differentiation might be important for establishing prognosis and optimal treatment. (Am J Psychiatry 1997; 154:948 957) Received Sept. 27, 1996; revision received Feb. 5, 1997; accepted Feb. 10, 1997. From the Department of Psychiatry, University of California, San Diego, and the VA Medical Center, San Diego. Address reprint requests to Dr. Schuckit, Department of Psychiatry (116A), VA Medical Center, 3350 La Jolla Village Dr., San Diego, CA 92161-2002. The Collaborative Study on the Genetics of Alcoholism, of which this study is a part, is supported by National Institute on Alcohol Abuse and Alcoholism grants AA-08401, AA-08402, and AA-08403. This research was also supported by NIAAA grant AA-05526 and by the VA Research Service. A s many as 80% of alcoholic men and women complain of depressive symptoms, and at least onethird meet the criteria for a major depressive episode (1 4). However, these mood disturbances might represent a heterogeneous group of syndromes with different prognoses and treatment needs (5 10). One relevant issue relates to what DSM-IV labels as substance-induced mood disorders with depressive features. The problem arises because intoxication with alcohol can cause temporary, but at times severe, affective symptoms, even in subjects with no history of clinically relevant depression (11 16). The identification of these intoxication- or withdrawal-related mood disorders is important because, contrary to what is expected with independent major depressive episodes (17), the substance-related depressive syndromes are likely to improve markedly within days or weeks of abstinence (18 23) or perhaps a bit longer (24 26). This relatively rapid temporal resolution is supported both by observations of the course of depressive symptoms in subjects participating in heavy-drinking experiments (13 15) and by studies of alcohol-dependent individuals who abstain (16, 18 22). There is not unanimous agreement, however, about the meaning of depressive syndromes among substancedependent persons. Some researchers propose that it is the overall magnitude of the psychiatric impairment that is important, not the specific type of psychiatric syndrome (27). Others question the overall relevance of subgroups, believing that high levels of psychiatric symptoms are inherent in alcohol-dependent individuals, even when sober (28). Yet other research challenges the belief that affective disturbances of any type are related to the clinical outcome, and there are disagreements about the optimal way to identify and characterize potentially substanceinduced disorders (26, 29 31). 948 Am J Psychiatry 154:7, July 1997

SCHUCKIT, TIPP, BERGMAN, ET AL. Parts of the disagreement might relate to differences among the groups studied, others might reflect the philosophies of the investigators, but some probably are the result of the definitions of the syndromes used. Until recently, distinctions among affective states in alcoholics have focused mostly on a primary versus secondary distinction based solely on the chronology of development of the syndromes (4). In this approach, the psychiatric disorder with onset at the earliest age is labeled as primary, and subsequent conditions are noted as secondary, tertiary, and so on. But this might be too restrictive, because independent depressive episodes, with their need for longer-term treatments, might develop during periods of abstinence but still be labeled as secondary. In that context, recent evidence has indicated that 70% or more of alcohol-dependent persons report at least one period of abstinence lasting 3 or more months in the course of their alcohol dependence, and there is an average of two such occurrences in which the mean duration exceeds 12 months (32). Reflecting methodological questions, there are areas of disagreement in the literature regarding the characteristics most closely associated with depression of any type among alcohol-dependent persons. Some studies report lower levels of impulsivity and violent behavior among depressed alcoholic individuals (33), while others report greater experience with drugs and higher levels of alcohol intake along with more alcohol-related problems (3, 33 35). These issues are important because a more intense involvement with substances of abuse might characterize substance-induced depression, perhaps relating to the pharmacological actions of the drugs involved. At least five investigations (3, 26, 33, 36, 37) have attempted to clarify some of these questions by looking at more carefully defined subgroups of depressed alcoholic subjects on the basis of the primary/secondary approach. The majority of these studies indicated that secondary affective disorders were seen more frequently than primary, or first-appearing, major depressive episodes (3, 26, 37). Several studies, but not all, indicated the possibility of more intense withdrawal syndromes, higher scores on the Michigan Alcohol Screening Test (MAST), and a greater likelihood of attending Alcoholics Anonymous (AA) among subjects with secondary mood disorders (3, 36, 37). While the studies revealed many similarities in the pattern of depressive symptoms (33, 37, 38), alcoholics with primary depression might have had longer-lasting or more numerous depressive episodes, as well as a history of treatment aimed specifically at a depressive disorder (36, 37). Finally, in some studies the patients with primary or potentially independent depression were also more likely to have reported depressive episodes, but not alcoholism, in close relatives (37, 39). If corroborated, this familial pattern of independent depressive disorders would potentially validate the diagnostic classification. Most studies, however, have described a limited number of patients from demographically narrow groups of alcoholics without sufficient numbers of both genders. While the emphasis on primary versus secondary depressive syndromes has added potential precision to the descriptions, the ability to draw more general conclusions is threatened by differences between studies in how depressive episodes were defined. Few studies broadened the definition of an independent mood disorder to indicate episodes that occurred during longerterm periods of abstinence. The present study took advantage of the ongoing Collaborative Study on the Genetics of Alcoholism. The analyses tested the hypothesis that the more narrowly defined substance-induced disorders would still be seen more often than independent depressive episodes. It was also hypothesized that, consistent with the possibility that some affective conditions might reflect pharmacological drug effects, persons with substance-induced depression would demonstrate more severe alcohol and drug histories, even after the impact of antisocial problems overall is controlled for. Third, it was proposed that among subjects with independent mood disorders there would be a higher rate of similar conditions in their first-degree relatives, a finding that might in part validate the independent versus substance-induced distinction. METHOD In the six-center Collaborative Study on the Genetics of Alcoholism project (5, 32, 40 42), after a full explanation of the procedures, subjects gave written informed consent to enter the study. Individuals were selected if they met the DSM-III-R criteria for alcohol dependence and the Feighner criteria (43) for alcoholism, regardless of additional psychiatric diagnoses (41). Evaluations were carried out by trained and closely supervised interviewers using the Semi-Structured Assessment for the Genetics of Alcoholism (42), a face-to-face instrument that reviews patterns of substance use and problems and establishes 17 DSM- III-R axis I diagnoses as well as antisocial personality disorder. In the semistructured interview, the onset of a substance use disorder is defined as the first age at which three or more of the potential nine DSM-III-R criterion items for dependence have occurred. Then, all episodes of abstinence lasting 3 or more months following the onset of alcohol dependence are noted (32). A 3-month period was selected to be consistent with the approach used in prior research and to strike a balance between periods of abstinence that might be long enough to be remembered but not so brief as to be too numerous to document (32). The interview also determines the age at onset of major depressive episodes. Independent major depression was defined as an episode that occurred either before the onset of alcohol dependence or during a period of 3 or more months of abstinence. However, major depressive syndromes that occurred during a period of active alcohol dependence were considered to have been substance-induced. Depressive episodes related to a recent loss (i.e., bereavement) or an illness/injury were defined as organic/other induced and are not discussed here. Full Semi-Structured Assessment for the Genetics of Alcoholism interviews were carried out with all of the alcohol-dependent subjects, with their first-degree relatives and extended families, and with comparison subjects and their relatives. Thus, data on rates of disorders and their relation to alcoholism in relatives were generated from personal semistructured interviews in which the same procedures as those for probands were used. Regarding comparison subjects, each center selected normal controls through a variety of methods including random mailing of questionnaires to university students, selection from driver s license records, and selection of general medical patients. For the analyses reported below, alcohol-dependent comparison subjects were included. Am J Psychiatry 154:7, July 1997 949

MAJOR DEPRESSIVE DISORDERS IN ALCOHOLICS TABLE 1. Demographic Characteristics of 2,945 Alcohol-Dependent Subjects Grouped According to History of Variable Group 1: Independent (N=449, 15.2%) Group 2: Substance- Induced (N=777, 26.4%) Group 3: No (N=1,719, 58.4%) Analysis Significant Post Hoc Group Comparisons Effect Size a Mean SD Mean SD Mean SD F (df=2, 2942) Age (years) 38.1 11.28 37.0 10.18 39.1 13.15 8.43*** 0.01 Education (years) 12.8 2.13 12.4 2.09 12.7 2.36 8.03*** 0.01 N % N % N % χ 2 (df=2) Female gender 233 51.9 229 29.5 483 28.1 95.82*** 1 versus 2***; 0.18 Race/ethnicity 37.44*** 1 versus 2***; 0.11 Caucasian 386 86.0 555 71.4 1,276 74.2 34.78*** 1 versus 2***; 0.11 African American 37 8.2 149 19.2 279 16.2 26.19*** 1 versus 2***; 0.09 Hispanic 17 3.8 48 6.2 117 6.8 5.60 Other 9 2.0 25 3.2 47 2.7 1.57 Marital status 95.46*** 1 versus 2***; 2 versus 0.18 3***; Married 202 45.0 234 30.1 784 45.6 55.70*** 1 versus 2***; 0.14 Separated 35 7.8 85 10.9 100 5.8 20.39*** 0.08 Divorced 92 20.5 196 25.2 246 14.3 44.93*** ; 0.12 Widowed 4 0.9 13 1.7 22 1.3 1.39 Never married 116 25.8 249 32.0 567 33.0 8.49* 1 versus 2**; 0.05 Employed full-time 234 52.1 335 43.1 894 52.0 18.00*** 1 versus 2**; 0.08 Original proband 113 25.2 390 50.2 483 28.1 133.74*** 1 versus 2***; 0.21 a Magnitude of effect as described by Cohen (44): ω 2 for F (0.01=small, 0.06=medium, 0.15=large); w for χ 2 (0.10=small, 0.30=medium, 0.50= large). *p<0.05. **p<0.01. ***p<0.001. Using this approach, we identified a total of 2,945 alcohol-dependent men and women. Individuals were placed in group 1 if they had ever met the criteria for an independent major depressive episode as defined above. From the remainder, group 2 included subjects who had had a major depression only in the context of alcoholism. Finally, group 3 was composed of alcohol-dependent individuals who had never met criteria for any major depression. Quantitative measures were compared among groups with analysis of variance (ANOVA) and with Tukey s honestly significant difference post hoc test for multiple group comparisons when the overall F statistic was significant. Because of the large study group size, estimates of the magnitude of effect (ω 2 for ANOVA and w for chisquare) as described by Cohen (44) are also reported. Finally, a stepwise logistic regression was used to evaluate the potential relevance of the multiple characteristics that differentiated groups 1 and 2 in the univariate analyses. RESULTS Semi-Structured Assessment for the Genetics of Alcoholism interviews from master file 49 of the collaborative study provided data on 3,104 alcohol-dependent individuals; data on 2,945 of these were relevant to our three groups after exclusion of 159 individuals because of incomplete data or medical disorders. The 2,945 remaining subjects, 945 (32.1%) of whom were women, included 986 initial probands (33.5%), 1,802 relatives of probands (61.2%), and 157 alcohol-dependent individuals (5.3%) originally selected from the comparison population. Data relating to the high rate of substance-induced conditions are presented in table 1. While 58.4% of the alcohol-dependent subjects had never experienced a major depressive episode (group 3), 15.2% of the total had had an independent major depression (group 1), including 200 subjects for whom the independent depression occurred only during a period of abstinence. About 26% reported information consistent with only a substance-induced depression (group 2). Thus, consistent with the first hypothesis, the rate of substanceinduced disorders remained high despite the expanded concept of independent depression. There were a number of differences in demographic characteristics across groups, although the magnitude of effect was usually small (e.g., age and education). Thus, the study group as a whole was about 38 years old and had about 12 years of education. Reflecting similar or slightly larger differences in effect size, group 1 subjects differed from groups 2 and 3 in having a higher proportion of women and Caucasians, while a higher proportion of group 2 subjects were probands, but fewer of them were married or were employed fulltime at interview. The potential impact of several of these demographic differences on the remaining group dissimilarities is discussed later in the Results section. To address the second hypothesis that persons with substance-induced depression would demonstrate more severe alcohol and drug histories table 2 and table 3 present the histories of substance use and related prob- 950 Am J Psychiatry 154:7, July 1997

SCHUCKIT, TIPP, BERGMAN, ET AL. TABLE 2. Histories of Substance Use and Problems Among Alcohol-Dependent Subjects Grouped According to History of Variable Group 1: Independent (N=449) Group 2: Substance- Induced (N=777) Group 3: No (N=1,719) Analysis Significant Post Hoc Group Comparisons Mean SD Mean SD Mean SD F df Age at onset of regular drinking (years) 17.5 5.61 16.9 4.72 17.7 5.10 7.86*** 2, 2941 0.01 Age at onset of alcohol dependence 23.9 8.07 23.8 7.69 25.1 9.32 7.43*** 2, 2931 2 versus 3** 0.01 (years) 1 versus 3* Maximum number of drinks 26.1 20.70 32.7 23.78 25.9 19.29 29.66*** 2, 2942 1 versus 2*** 0.02 per day (ever) Number of drinks per day 6.0 7.00 10.1 10.27 6.5 8.23 17.79*** 2, 997 1 versus 2*** 0.01 (last 6 months) b Number of drinking days per 3.0 2.19 3.8 2.30 3.6 2.17 5.55** 2, 997 1 versus 2** 0.01 week (last 6 months) b 1 versus 3* Total time abstinent (months) 52.7 63.94 31.5 42.00 39.7 56.34 14.98*** 2, 1850 1 versus 2*** 0.01 2 versus 3* Number of abstinent periods 1.4 1.28 1.2 1.18 1.1 1.22 10.43*** 2, 2942 1 versus 2* 0.01 Effect Size a N % N % N % χ 2 df Alcoholism treatment history Any treatment 226 50.3 559 72.0 798 46.4 139.82*** 2 1 versus 2*** 0.22 Inpatient treatment 164 36.5 476 61.3 621 36.1 145.22*** 2 1 versus 2*** 0.22 Alcoholics Anonymous 209 46.5 500 64.3 698 40.6 116.80*** 2 1 versus 2*** 0.19 1 versus 3* Drug use history Number of different drugs used Mean SD Mean SD Mean SD F df 1.6 1.63 2.3 1.76 1.5 1.54 c c 1 versus 2*** 0.05 Number of times drug was used Marijuana 360.9 436.01 533.3 448.46 389.1 431.63 29.35*** 2, 2393 1 versus 2*** 0.02 Amphetamine 455.8 1194.75 444.5 1207.15 308.8 889.00 2.87 2, 1309 Cocaine 362.2 1115.03 547.9 1266.65 385.0 1143.18 3.73* 2, 1665 2 versus 3* 0.01 Opiates 216.2 700.20 458.1 1517.71 321.1 1280.37 2.07 2, 927 Sedative/hypnotics 258.3 986.63 190.6 513.90 167.6 748.87 1.10 2, 1105 Pattern of drug use Marijuana 21 or more times Amphetamines 11 or more times N % N % N % χ 2 df 368 82.0 679 87.4 1,349 78.5 28.15*** 2 1 versus 2** 0.01 218 48.6 426 54.8 669 38.9 58.66*** 2 1 versus 2* 0.14 Cocaine 11 or more times 228 50.8 546 70.3 894 52.0 80.08*** 2 1 versus 2*** 0.19 Opiates 11 or more times 148 33.0 352 45.3 430 25.0 103.08*** 2 1 versus 2*** 0.19 Sedative/hypnotics 11 or more times 192 42.8 388 49.9 528 30.7 90.19*** 2 1 versus 2*** 0.17 a Magnitude of effect as described by Cohen (44): ω 2 for F (0.01=small, 0.06=medium, 0.15=large); w for χ 2 (0.10=small, 0.30=medium, 0.50= large). b Among individuals who drank alcohol in the last 6 months. c χ 2 =72.37, df=2, p<0.001. *p<0.05. **p<0.01. ***p<0.001. Am J Psychiatry 154:7, July 1997 951

MAJOR DEPRESSIVE DISORDERS IN ALCOHOLICS TABLE 3. Diagnostic Pattern for Alcohol-Dependent Subjects and First-Degree Relatives Grouped According to History of Among Subjects Group 1: Independent (N=449) Group 2: Substance- Induced (N=777) Group 3: No (N=1,719) Variable N % N % N % Analysis: χ 2 (df=2) Significant Post Hoc Group Comparisons Subject s primary (first appearing) 742.61*** 1 versus 2***; 2 versus 0.50 diagnosis 3***; Alcohol dependence 169 37.6 422 54.3 1,219 70.9 189.18*** 1 versus 2***; 2 versus 0.25 3***; Antisocial personality disorder 101 22.5 216 27.8 253 14.7 62.01*** 1 versus 2*; ; 0.15 Drug dependence 28 6.2 93 12.0 172 10.0 10.44** 1 versus 2***; 1 versus 3** 0.06 depression 106 23.6 0 0.0 0 0.0 611.26*** 1 versus 2***; 0.46 Mania or dysthymia 7 1.6 8 1.0 3 0.2 3.05 anxiety disorder 38 8.4 37 4.8 62 3.6 18.89*** 1 versus 2**; 0.08 Subject s substance diagnoses/ treatment Drug dependence (ever) 230 51.2 528 67.9 712 41.4 151.07*** 1 versus 2***; 2 versus 0.23 3***; Marijuana 153 34.1 382 49.1 464 27.0 116.06*** 1 versus 2***; 2 versus 0.20 3***; 1 versus 3** Amphetamines 131 29.2 172 22.1 198 11.5 51.10*** ; 0.13 Cocaine 108 24.0 368 47.4 437 25.4 132.98*** 1 versus 2***; 0.21 Opiates 43 9.6 126 16.2 108 6.3 61.06*** 1 versus 2***; 2 versus 3***; 1 versus 3* 0.14 Sedative/hypnotics 62 13.8 126 16.2 83 4.8 100.66*** ; 0.19 Any treatment for drug abuse 103 22.9 302 38.9 342 19.9 103.39*** 1 versus 2***; 0.19 Narcotics Anonymous/Cocaine 78 17.4 226 29.1 256 14.9 70.94*** 1 versus 2***; Anonymous 0.16 Number of first-degree relatives 461 1,587 1,911 Relatives diagnoses Independent major depression 104 22.6 297 18.7 289 15.1 17.13*** ; 0.07 Independent mania 16 3.5 14 0.9 19 1.0 20.70*** 1 versus 2***; 0.07 Alcohol dependence 161 34.9 576 36.3 629 32.9 4.42 Drug dependence 102 22.1 313 19.7 342 17.9 5.33 Antisocial personality disorder 38 8.2 144 9.1 140 7.3 3.42 a Magnitude of effect as described by Cohen (44): ω 2 for F (0.01=small, 0.06=medium, 0.15=large); w for χ 2 (0.10=small, 0.30=medium, 0.50=large). *p<0.05. **p<0.01. ***p<0.001. Effect Size a lems across the groups. As hypothesized, subjects with substance-induced depression (group 2) had some evidence of more intense alcohol and drug involvement (table 2). This included a higher maximum number of drinks in 24 hours (a drink was defined as the amount of alcohol in 12 ounces of beer, 4 ounces of wine, or 1 to 1.5 ounces of 80-proof beverage), a higher average number of drinks per day, more days per week on which drinking occurred in the 6 months before interview, a lower average number of months of abstinence since the onset of alcohol dependence, and involvement with more categories of drugs. However, the effect sizes for these differences were relatively small. It is also possible that the periods of abstinence reflect the manner of subject selection, since the longer the abstinence, the greater the period of time a subject is eligible to have a diagnosis of an independent depressive disorder. At larger effect sizes, more group 2 subjects received treatment for alcoholism and participated in AA, and they were more likely to have reported using cocaine, amphetamines, opiates, and sedative/hypnotics. The information in table 3 complements the conclusions supported by table 2. Group 2 subjects were the most likely to meet the criteria for antisocial personality disorder and drug dependence. Reflecting the construction of the three groups, the subjects with independent depressive disorders were the only ones for whom major depression was the primary or first-appearing illness. Not directly related to the manner of assignment to groups, subjects in group 1 were also most likely to have had a first-appearing major anxiety disorder. Table 3 also offers information relevant to the third hypothesis regarding the familial pattern of illness. To avoid the problems incurred if each subject is considered in relation to all other family members, the family history data are limited to interviewed first-degree relatives of probands. A significantly higher proportion of the firstdegree relatives of persons in group 1 had independent major depression or mania, but the groups did not differ regarding the rates of alcohol and drug dependence and antisocial personality disorder in relatives. The clinical characteristics of the most severe depressive 952 Am J Psychiatry 154:7, July 1997

SCHUCKIT, TIPP, BERGMAN, ET AL. TABLE 4. Clinical Characteristics of the Most Severe Depressive Episode Among 1,226 Alcohol-Dependent Subjects With Variable Group With Independent (N=449) Group With Substance- Induced (N=777) Analysis Mean SD Mean SD F df Age at onset of major depression (years) 28.3 12.09 30.4 9.79 11.36** 1, 1211 0.01 Total number of DSM-III-R depressive symptoms b 7.2 1.24 7.4 1.26 3.57 1, 1224 Number of days hospitalized 41.6 56.34 35.3 65.51 0.36 1, 148 Effect Size a N % N % χ 2 df DSM-III-R depressive symptoms Diminished pleasure 433 96.4 726 93.4 4.96* 1 0.06 Weight gain/loss 341 76.0 643 82.8 8.32** 1 0.08 Sleep disturbance 421 93.8 693 89.2 7.17** 1 0.08 Psychomotor agitation/retardation 268 59.7 482 62.0 0.66 1 Fatigue 382 85.1 666 85.7 0.09 1 Feelings of worthlessness 369 82.2 676 87.0 5.25* 1 0.08 Problems concentrating 387 86.2 691 88.9 2.01 1 Thoughts of death 210 46.8 401 51.6 2.66 1 Other depressive symptoms Hallucinations/delusions 24 5.4 63 8.1 3.30 1 Decreased functioning 347 77.3 610 78.5 0.25 1 Attempted suicide 136 30.3 193 24.8 4.31* 1 0.06 Treatment history Sought help 209 46.6 307 39.5 5.78* 1 0.07 Medication prescribed 116 25.8 163 21.0 3.82 1 Hospitalized 54 12.0 97 12.5 0.06 1 Received ECT 8 1.8 7 0.9 1.83 1 a Magnitude of effect as described by Cohen (44): ω 2 for F (0.01=small, 0.06=medium, 0.15=large); w for χ 2 (0.10=small, 0.30=medium, 0.50=large). b Of nine. *p<0.05. **p<0.01. episodes for the subjects in groups 1 and 2 are shown in table 4. At a small effect size, members of groups 1 and 2 were slightly different in age at onset of their first major depressive episode (despite the way that groups were defined) but similar in most aspects of the history of treatment for mood disorders and in most symptoms of depression. For those symptoms that did differ, there was no consistent pattern regarding whether group 1 or group 2 subjects had higher rates of problems. The univariate analyses in tables 1 3 do not indicate which of the potential independent variables significantly differentiating groups 1 and 2 remain robust when considered in a multivariate context. Therefore, a logistic regression analysis was carried out on data of the group 1 and group 2 subjects, with group 1 membership as the dependent variable. For these analyses, in order to include all subjects, not just probands, the family history variable was the proportion of subjects with at least one relative having one of the five disorders listed at the bottom of table 3. The specific items used in the logistic regression included all those listed in table 5 plus those that did not enter the equation, including treatment for alcoholism, participation in AA, a history of drug dependence, and seeking treatment for depression. Because of their relationship to the criteria used for group 1 membership, neither a primary diagnosis of alcoholism nor the number of months of abstinence was used in the analysis. The first data column of table 5 indicates that a history of at least one independent major depressive episode was related to the demographic characteristics of female gender, Caucasian race, having ever been married, and not being one of the original probands. Group 1 membership was also predicted by use of fewer drugs, a family history of independent depression or of independent mania, and a personal history of a suicide attempt. Because of the large differences between groups 1 and 2 in the proportion of women, the second and third data columns of table 5 present the results separately for the two genders. The same items generally had results in the same direction for men and women separately, although marital status and family history were predictors of independent major depression for men, while a suicide attempt was a predictor for women. Two additional items entered this equation: a lower maximum number of drinks and a history of antisocial personality disorder, both of which predicted independent depression for women. Regarding the second hypothesis, there is evidence in both genders of a relationship of more alcohol or drug involvement for group 2, and regarding the third hypothesis, a family history of affective illness was most evident for women in group 1. Table 5 also presents results separately for proband and nonproband alcohol-dependent subjects. This step is important, because for probands the analysis explored the predictors evaluated in subjects who were not biologically Am J Psychiatry 154:7, July 1997 953

MAJOR DEPRESSIVE DISORDERS IN ALCOHOLICS TABLE 5. Logistic Regression Analyses of Characteristics Significantly Associated With Independent Among 1,226 Alcohol- Dependent Subjects With Independent or Substance-Induced Depressive Episodes Variable Total Group (N=1,226) χ 2 (df=1) Odds Ratio Male Subjects (N=764) χ 2 (df=1) Odds Ratio Female Subjects (N=462) χ 2 (df=1) Odds Ratio Probands (N=503) χ 2 (df=1) Odds Ratio Nonproband Subjects (N=723) χ 2 (df=1) Female gender 27.23*** 2.00 31.42*** 2.48 Caucasian race 19.64*** 2.10 6.79** 1.76 13.98*** 2.76 17.02*** 2.46 Married 10.93*** 1.56 16.47*** 2.05 13.09*** 1.81 Proband 31.10*** 0.45 9.64** 0.58 17.49*** 0.36 Maximum number of drinks per day 4.17* 0.98 Number of different drugs used 18.32*** 0.85 5.98* 0.89 11.97*** 0.79 23.01*** 0.79 Primary diagnosis of antisocial personality disorder 5.94* 2.15 Any first-degree relative with independent mania 8.25** 2.01 6.30* 2.27 17.59*** 5.30 Any first-degree relative with independent depression 4.22* 1.31 4.60* 1.45 5.11* 1.44 Ever attempted suicide 7.81** 1.51 6.06* 1.72 10.96*** 2.12 *p<0.05. **p<0.01. ***p<0.001. Odds Ratio related to one another. Here most of the demographic variables and substance use patterns were found only to predict group 1 membership among relatives of probands, but regarding the third hypothesis, some aspects of a family history of independent affective disorder appear relevant for both probands and nonprobands. Finally, it is important to note that group 2 membership was based on any substance-induced depressive episode. While 76.6% (N=595) of the group 2 subjects had experienced at least one alcohol-induced depressive episode, for the remainder, other drugs were involved. Therefore, the logistic regression analyses were repeated for the group with alcohol-induced affective syndromes. The results closely resembled those shown for the total study group in table 5. DISCUSSION The results of this study support a number of conclusions. Consistent with prior reports (1, 2, 5, 6, 30), there was a high rate of affective disturbances in these alcoholic individuals. In the present study group, 41.6% of the alcohol-dependent men and women had depressive symptoms serious enough to qualify for a diagnosis of a major depressive episode at some time during their alcoholic course. The data support the first hypothesis, in that more than 60% of the alcoholic subjects with histories of depression reported a substance-induced depressive episode (3, 26, 36, 37). This predominance of substanceinduced conditions occurred despite our broadening the definition of independent affective disorders to include depressive episodes that only developed during periods of abstinence. While not all investigators agree (30, 45), some other studies have demonstrated how important the distinction between primary and secondary or between induced and independent depressive disorders can be. Most evaluations of substance-induced or secondary depression indicate that while the mood disorders persist in the context of continued heavy drinking, they are likely to improve markedly within days or weeks of abstinence (13 16, 18 21, 46). On the other hand, affective disturbances that are primary and independent might tend to remain for more extended periods of time (17), even in alcoholic individuals (47, 48). Thus, the distinction between substance-induced and independent syndromes gives useful data regarding the potential need for antidepressant medications, and it offers information about the prognosis for the affective disturbance (49). Consistent with some prior studies of primary versus secondary depressive disorders (6, 33, 38), the present data indicate few differences with large effect sizes in the clinical symptoms in the two types of depressive episodes. Thus, clinicians might not be able to use the pattern of depressive symptoms alone to identify those patients most likely to experience long-lasting depressive episodes following abstinence. Some investigators have reported a greater number of depressive episodes and a larger number of days of treatment for subjects with primary depression (36, 37), but our data did not corroborate these findings. The disparity might reflect the broader independent versus substance-induced approach used here. The data do, however, show that the 449 alcohol-dependent persons with independent depressive episodes were more likely to report a history of at least one suicide attempt (33), although this effect appears to have been produced by female subjects and probands. Several demographic characteristics might help identify subjects with independent, and potentially more long-lasting, depressive episodes. As might be expected for major depressive disorders (17), in this study independent major depression was more common among women and, consistent with the report of Roy et al. (3), 954 Am J Psychiatry 154:7, July 1997

SCHUCKIT, TIPP, BERGMAN, ET AL. among men with a stable marital history. These characteristics were not solely a reflection of drug use patterns or antisocial personality disorder, as the findings remained significant in a logistic regression analysis. Consistent with the second hypothesis, group 1 subjects showed some evidence of less intense alcohol and drug intake and problem histories (3, 36, 37). A lower number of categories of drugs used remained a robust item in all logistic regression analyses in table 5, but a lower maximum number of drinks per day related only to women. These findings are, in general, consistent with the lower score on the MAST, the lower proportion who had ever attended AA, and the lesser drug involvement among subjects with independent depression in several reports in the literature (3, 34, 36, 37). If accurate, the higher level of drug involvement or of alcohol intake among group 2 subjects at least in part supports the probably substance-induced nature of their affective disturbances. However, in the absence of a clinically obvious animal model of substance-induced depression, and reflecting concerns about protection of human subjects, this relationship remains inferential. Regarding the third hypothesis, the data in table 3 and table 5 confirm prior reports of a higher prevalence of independent depressive disorders in first-degree relatives of subjects with independent major depression (20, 26, 33, 39, 50 52). The present findings occurred in the context of personal interviews with these relatives, evaluations that incorporated the same time linebased distinction between substance-induced and independent depressions used for the probands. The increased risk of independent affective episodes in relatives of subjects with independent depression is consistent with the validity of the distinction between group 1 and group 2 individuals. The possibility that independent major depression might not be related to familial alcoholism but might reflect a family history of major affective disorder is consistent with our own 8.2-year follow-up of 450 sons of alcoholic men and control subjects (53). In that study sons of alcoholics had a greater risk of alcoholism but showed no heightened rate of independent major depressive disorders findings consistent with some other prospective studies of children of alcoholic parents (54 57). Also supporting this lack of a close tie between vulnerability to affective disorder and vulnerability to alcoholism are two follow-up studies of subjects with a major mood disorder in childhood who showed no subsequent higher rate of alcoholism (58, 59). However, it is likely that other types of familial relationships between affective and alcoholic disorders might also be important, including evidence that some genes might predict a general vulnerability to psychiatric syndromes overall (60) and that in some families alcoholism and affective disorders might be more closely tied together in a syndrome labeled as depressive spectrum disease (61, 62). While they were not the direct focus of this study, there are some treatment implications of the distinction between substance-induced and independent depression. Several reviews have indicated that antidepressant medications have little role in treating the usual alcoholic patient (63, 64), while they have noted deficiencies in the investigations reviewed (63, 65). Also, a recent 12-week study of antidepressants in 69 actively drinking alcohol-dependent outpatients (48) reported that imipramine had no effect on drinking outcome but was associated with greater improvement in depression for subjects with primary depression. However, a second trial involved 6 months of treatment with desipramine for primary alcoholics, comparing 12 depressed subjects taking the active drug with 10 taking placebo, in which a decrease in depression ratings and a lower rate of full relapse was reported for the former group (35). These results indicate that further studies of antidepressants in depressed alcoholics are warranted, even if the mood disturbance appears to be secondary. As for the present study, despite the large group of subjects and the intense efforts made to gather detailed information, the results must be viewed in the context of the method used. All of the data are historical, and while some treatment records of individuals who had received inpatient care were reviewed, the study would have been strengthened by interviews with additional informants. Second, no follow-up interviews have yet been carried out with this group, so no prospective data are available. The third caveat is the necessity of relying on clinical histories in distinguishing between substance-induced and secondary and independent, or primary, depressive disorders. Thus, the ability of the substance use to induce the depression could not be directly evaluated, with the result that the group 2 condition might have been more properly called substancecoincident. Finally, the average age of the subjects reported here was about 38 years, and it is possible that different results might be observed in older or younger groups. Regarding the latter, data are being gathered from adolescent children of the subjects in the Collaborative Study on the Genetics of Alcoholism, and future analyses will evaluate the generalizability of those findings to the younger group. ACKNOWLEDGMENTS The Collaborative Study on the Genetics of Alcoholism (H. Begleiter, State University of New York, Health Sciences Center at Brooklyn, Principal Investigator; T. Reich, Washington University, St. Louis, Co-Principal Investigator) includes six different centers where data collection takes place. The six sites and principal investigators and co-investigators are Indiana University (J. Nurnberger, Jr., P.M. Conneally); University of Iowa (R. Crowe, S. Kuperman); University of California, San Diego, and Scripps Institute (M. Schuckit, F. Bloom); University of Connecticut (V. Hesselbrock); State University of New York, Health Sciences Center at Brooklyn (H. Begleiter, B. Porjesz); and Washington University (T. Reich, C.R. Cloninger). REFERENCES 1. Regier DA, Farmer ME, Rao DS, Locke BZ, Keith SJ, Judd LL, Goodwin FK: Comorbidity of mental disorders with alcohol and other drug use. JAMA 1990; 264:2511 2518 2. Kessler RC, Nelson CB, McGonagle KA, Eclund MJ, Frank RG, Am J Psychiatry 154:7, July 1997 955

MAJOR DEPRESSIVE DISORDERS IN ALCOHOLICS Leof PJ: The epidemiology of co-occurring addictive and mental disorders. Am J Orthopsychiatry 1996; 66:17 31 3. Roy A, DeJone J, Lamparski D, George T, Linnoila M: among alcoholics: relationship to clinical and cerebrospinal fluid variables. Arch Gen Psychiatry 1991; 48:428 432 4. Schuckit MA: The clinical implications of primary diagnostic groups among alcoholics. Arch Gen Psychiatry 1985; 42:1043 1049 5. Schuckit MA, Tipp JE, Bucholz KK, Nurnberger JI, Hesselbrock V, Crowe RR: The lifetime rates of three major mood disorders and four major anxiety disorders in alcoholics and controls. Addiction (in press) 6. Hesselbrock MN, Hesselbrock VM, Tennen H, Meyer RE, Workman KL: Methodological considerations in the assessment of depression in alcoholics. J Consult Clin Psychol 1983; 51: 399 405 7. Winokur G, Cook B, Liskow B, Fowler R: Alcoholism in manic depressive (bipolar) patients. J Stud Alcohol 1993; 54:574 576 8. Morrison JR: The family histories of manic-depressive patients with and without alcoholism. J Nerv Ment Dis 1975; 160:227 229 9. Dunner DL, Hensel BM, Fieve RR: Bipolar illness: factors in drinking behavior. Am J Psychiatry 1979; 136:583 585 10. Winokur G, Clayton P, Reich T: Manic-Depressive Illness. St Louis, CV Mosby, 1969 11. Schuckit MA: Drug and Alcohol Abuse: A Clinical Guide to Diagnosis and Treatment, 4th ed. New York, Plenum Medical, 1995 12. Birnbaum I, Taylor T, Parker E: Alcohol and sober mood state in female social drinkers. Alcohol Clin Exp Res 1983; 7:362 368 13. Isbell H, Fraser HF, Wikler A, Belleville RE, Eisenman AJ: An experimental study of the etiology of rum fits and delirium tremens. J Stud Alcohol 1955; 16:1 33 14. Gibson S, Becker J: Changes in alcoholics self-reported depression. Q J Stud Alcohol 1973; 34:829 836 15. Tamerin JS, Weiner S, Mendelson JH: Alcoholics expectancies and recall of experiences during intoxication. Am J Psychiatry 1970; 126:1697 1704 16. Mayfield DG: Psychopharmacology of alcohol: affective change with intoxication, drinking behavior, and affective state. J Nerv Ment Dis 1968; 146:314 321 17. Akiskal HS: Mood disorders: clinical features, in Comprehensive Textbook of Psychiatry, 6th ed, vol 1. Edited by Kaplan HI, Sadock BJ. Baltimore, Williams & Wilkins, 1995, pp 1123 1152 18. Brown SA, Schuckit MA: Changes in depression among abstinent alcoholics. J Stud Alcohol 1988; 49:412 417 19. Davidson KM: Diagnosis of depression in alcohol dependence: changes in prevalence with drinking status. Br J Psychiatry 1995; 166:199 204 20. Willenbring ML: Measurement of depression in alcoholics. J Stud Alcohol 1986; 47:367 372 21. Pettinati HM, Sugerman A, Maurer HS: Four-year MMPI changes in abstinent and drinking alcoholics. Alcohol Clin Exp Res 1982; 6:487 494 22. Penick EC, Powell BJ, Liskow BI, Jackson JO, Nickel EJ: The stability of coexisting psychiatric syndromes in alcoholic men after one year. J Stud Alcohol 1988; 49:395 405 23. Dackis CA, Gold MS, Pottash ALC: Evaluating depression in alcoholics. Psychiatry Res 1986; 17:105 109 24. Coryell W, Keller M, Endicott J, Andreasen N, Clayton P, Hirschfeld R: Bipolar II illness: course and outcome over a fiveyear period. Psychol Med 1989; 19:129 141 25. Pottenger M, McKernon J, Patrie LE, Weismann MM, Ruben HL, Newberry P: The frequency of persistence of depressive symptoms in the alcohol abuser. J Nerv Ment Dis 1978; 166: 562 570 26. O Sullivan K, Whillans P, Daly M, Carroll B, Clare A, Cooney J: A comparison of alcoholics with and without coexisting affective disorder. Br J Psychiatry 1983; 143:133 138 27. Rounsaville BJ, Kosten TR, Weissman MM, Kleber HD: Prognostic significance of psychopathology in treated opiate addicts. Arch Gen Psychiatry 1986; 43:739 745 28. Gerard DL, Saenger G, Wile R: The abstinent alcoholic. Arch Gen Psychiatry 1962; 6:99 111 29. Powell BJ, Penick EC, Nickel EJ, Liskow BI, Riesenmy KD, Campion SL, Brown EF: Outcomes of co-morbid alcoholic men: a 1-year follow-up. Alcohol Clin Exp Res 1992; 16:131 138 30. Penick EC, Powell BJ, Nickel EJ, Bingham SF, Riesenmy KR, Read MR, Campbell J: Co-morbidity of lifetime psychiatric disorder among male alcoholic patients. Alcohol Clin Exp Res 1994; 18:1289 1293 31. Hesselbrock VM, Hesselbrock MN, Workman-Daniels KL: Effect of major depression and antisocial personality on alcoholism: course and motivational patterns. J Stud Alcohol 1986; 47: 207 212 32. Schuckit MA, Tipp JE, Smith TL, Bucholz KK: Periods of abstinence following the onset of alcohol dependence in 2,230 men and women. J Stud Alcohol (in press) 33. Cornelius JR, Salloum IM, Mezzich J, Cornelius MD, Fabrega HF Jr, Ehler JG, Ulrich RF, Thase ME, Mann JJ: Disproportionate suicidality in patients with comorbid major depression and alcoholism. Am J Psychiatry 1995; 152:358 364 34. Schuckit M: Alcoholic patients with secondary depression. Am J Psychiatry 1983; 140:711 714 35. Mason BJ, Kocsis JH, Ritvo EC, Cutler RB: A double-blind placebo-controlled trial of desipramine in primary alcoholics stratified on the presence or absence of major depression. JAMA 1996; 275:1 7 36. Powell BJ, Read MR, Penick EC, Miller NS, Bingham SF: Primary and secondary depression in alcoholic men: an important distinction? J Clin Psychiatry 1987; 48:98 101 37. Hasegawa K, Mukasa H, Nakazawa Y, Kodama H, Nakamura K: Primary and secondary depression in alcoholism clinical features and family history. Drug Alcohol Depend 1991; 27: 275 281 38. Weissman MM, Pottenger M, Kleber H, Ruber HL, Williams D, Thompson WD: Symptom patterns in primary and secondary depression. Arch Gen Psychiatry 1977; 34:854 862 39. Hesselbrock V, Tennen H, Stabenau J, Hesselbrock M: Affective disorder in alcoholism. Int J Addict 1983; 18:435 444 40. Schuckit MA, Hesselbrock VM, Tipp J, Nurnberger JI, Anthenelli RM, Crowe RR: The prevalence of major anxiety disorders in relatives of alcohol-dependent men and women. J Stud Alcohol 1995; 56:309 317 41. Schuckit MA, Tipp JE, Anthenelli RM, Bucholz KK, Hesselbrock VM, Nurnberger JI Jr: Anorexia nervosa and bulimia nervosa in alcohol-dependent men and women and their relatives. Am J Psychiatry 1996; 153:74 82 42. Bucholz KK, Cadoret R, Cloninger CR, Dinwiddie SH, Hesselbrock VM, Nurnberger JI Jr, Reich T, Schmidt I, Schuckit MA: A new, semi-structured psychiatric interview for use in genetic linkage studies: a report on the reliability of the SSAGA. J Stud Alcohol 1994; 55:149 158 43. Feighner JP, Robins E, Guze SB, Woodruff RA Jr, Winokur G, Munoz R: Diagnostic criteria for use in psychiatric research. Arch Gen Psychiatry 1972; 26:57 63 44. Cohen J: A power primer. Psychol Bull 1992; 112:115 159 45. Woodruff RA, Guze SB, Clayton PJ, Carr D: Alcoholism and depression. Arch Gen Psychiatry 1973; 28:97 100 46. Brown SA, Inaba RK, Gillin JC, Schuckit MA, Stewart MA, Irwin MR: Alcoholism and affective disorder: clinical course of depressive symptoms. Am J Psychiatry 1995; 152:45 52 47. Merry J, Reynolds CM, Bailey J, Coppen A: Prophylactic treatment of alcoholism by lithium carbonate: a controlled study. Lancet 1976; 2:481 482 48. McGrath PJ, Nunes EV, Stewart JW, Goldman D, Agosti V, Ocepek-Welikson K, Quitkin FM: Imipramine treatment of alcoholics with primary depression. Arch Gen Psychiatry 1996; 53:232 240 49. Mueller TI, Lavori PW, Keller MB, Swartz A, Warshaw M, Hasin D, Coryell W, Endicott J, Rice J, Akiskal H: Prognostic effect of the variable course of alcoholism on the 10-year course of depression. Am J Psychiatry 1994; 151:701 706 50. Merikangas KR, Leckman JF, Prusoff BA, Pauls DL, Weissman MM: Familial transmission of depression and alcoholism. Arch Gen Psychiatry 1985; 42:367 372 51. Winokur G, Coryell W, Akiskal HS, Maser JD, Keller MB, Endicott J, Mueller T: Alcoholism in manic-depressive (bipolar) ill- 956 Am J Psychiatry 154:7, July 1997

SCHUCKIT, TIPP, BERGMAN, ET AL. ness: familial illness, course of illness, and the primary-secondary distinction. Am J Psychiatry 1995; 152:365 372; correction, 152:1106 52. Aneshensel CS, Huba GJ:, alcohol use, and smoking over one year: a four-wave longitudinal causal model. J Abnorm Psychol 1983; 92:134 150 53. Schuckit MA, Smith TL: An 8-year follow-up of 450 sons of alcoholic and control subjects. Arch Gen Psychiatry 1996; 53: 202 210 54. Hill SY, Hruska DR: Childhood psychopathology in families with multigenerational alcoholism. J Am Acad Child Adolesc Psychiatry 1992; 31:1024 1030 55. Merikangas KR, Prusoff BA, Weissman MM: Parental concordance for affective disorders. J Affect Disord 1988; 15:279 290 56. Reich W, Earls F, Frankel O, Shayka JJ: Psychopathology in children of alcoholics. J Am Acad Child Adolesc Psychiatry 1993; 32:995 1002 57. Knop J, Goodwin DW, Jensen P, Penick E, Pollock V, Gabrielli W, Teasdale TW, Mednick SA: A 30-year follow-up study of the sons of alcoholic men. Acta Psychiatr Scand 1993; 370:48 53 58. Harrington R, Fudge H, Rutter M, Pickles A, Hill J: Adult outcomes of childhood and adolescent depression. Arch Gen Psychiatry 1990; 47:465 473 59. Rao U, Ryan ND, Birmaher B, Dahl RE, Williamson DE, Kaufman J, Rao R, Nelson B: Unipolar depression in adolescents: clinical outcome in adulthood. J Am Acad Child Adolesc Psychiatry 1995; 34:566 578 60. Kendler KS, Walters EE, Neale MC, Kessler RC, Heath AC, Eaves LJ: The structure of the genetic and environmental risk factors for six major psychiatric disorders in women. Arch Gen Psychiatry 1995; 52:374 383 61. Winokur G: The development and validity of familial subtypes in primary unipolar depression. Pharmacopsychiatry 1982; 15: 142 146 62. Cadoret RJ, Winokur G, Langbehn D, Troughton E, Yates WR, Stewart MA: spectrum disease, I: the role of gene-environment interaction. Am J Psychiatry 1996; 153:892 899 63. Liskow BI, Goodwin DW: Pharmacological treatment of alcohol intoxication, withdrawal and dependence: a critical review. J Stud Alcohol 1987; 48:356 370 64. Litten RZ, Allen JP: Pharmacotherapies for alcoholism: promising agents and clinical issues. Alcohol Clin Exp Res 1991; 15: 620 633 65. Ciraulo DA, Alderson LM, Chapron DJ, Jaffe JH, Subbarao B, Kramer PA: Imipramine disposition in alcoholics. J Clin Psychopharmacol 1982; 2:2 7 Am J Psychiatry 154:7, July 1997 957