INTRODUCTION
Treatment of Depression in Drug-Dependent Patients: Effects on Mood and Drug Use Edward V. Nunes and Frederic M. Quitkin
Symptoms of depression and anxiety are common in patients with substance use disorders (Meyer 1986; Schuckit 1986). In the general population, mood and anxiety disorders convey increased risk for substance use disorders (Regier et al. 1990). Further, mooddisordered substance abusers have poor prognoses (Rounsaville et al. 1982b, 1986b, 1987; Weissman et al. 1976; Kosten et al. 1986; LaPorte et al. 1981; Loosen et al. 1990; Carroll et al. 1993). Thus, evaluation and appropriate management of affective disorders should be a useful treatment adjunct with the potential of improving outcome of substance abuse. Nevertheless, controversy continues to surround this clinical problem, and approaches vary widely among clinicians. Further, the problem of the depressed substance abuser raises important theoretical questions about the etiology and pathogenesis of substance abuse disorders. Mood syndromes observed in substancedependent patients often resolve soon after abstinence or the initiation of specific treatment such as methadone (Weddington et al. 1990; Rounsaville et al. 1986a; Schuckit 1986; Willis and Osbourne 1978; DeLeon et al. 1973), suggesting a "substance-induced" (American Psychiatric Association 1994) syndrome (i.e., toxicity or withdrawal) or transient adjustment reactions. However, in 10 percent or more of these patients in various clinical samples, depression persists (Nakamura et al. 1983; Johnson and Perry 1986; Rounsaville et al. 1986a; Croughan et al. 1981). These persons may have a mood disorder that is independent of substance abuse. Because both substance abuse and mood disorders are common in the general population, it can be expected that some individuals will have both disorders by chance alone. Another possibility is that a subgroup has mood disorders that contribute to the etiology of substance abuse. In fact, a selfmedication hypothesis has been advanced (Khantzian 1985; Quitkin et al. 1972) suggesting that some individuals use drugs because they provide temporary relief from symptoms of depression or anxiety. Depression or anxiety may be the sole etiology or one of several causal factors, including other genetic and environmental vulnerabilities, in substance abuse. Also, depression or anxiety may alter the course of substance abuse. For example, depression is a frequent internal cue triggering drug craving (Marlatt and Gordon 1985; Daley and Marlatt 1992). Thus, an initially independent depressive disorder, through classical conditioning (Childress et al. 1994), may become linked to relapse and perpetuate substance abuse. The evidence that the self-medication hypothesis plays a contributory role in some substance abuse stems mainly from clinical observations (Khantzian 1985; Marlatt and Gordon 1985) and epidemiologic data (Regier et al. 1990). However, the strongest test of the hypothesis would be one that directly addresses its clinical utility, namely, whether treatment of depression or anxiety alters the course and outcome of substance abuse. Specifically, if depression contributes to the etiology of substance abuse, then antidepressant treatment should improve substance abuse outcome. Relatively few studies of this type have been undertaken, and most have methodologic problems. This chapter reviews this literature as well as the authors' recent studies, drawing tentative conclusions and developing suggestions for future research. The present approach to evaluating the literature is summarized in table 1 (Nunes et al., in press). Studies evaluated are those in which patients with substance use disorders who also display evidence of depression receive antidepressant medication treatment. If there is no medication-placebo difference in both mood and substance use outcome (right column, table 1), and particularly if the placebo-mood response is high, a transient substance-induced mood syndrome or adjustment reaction is suggested. If mood improves on medication compared to placebo, but substance use does not (middle column, table 1), it suggests a true mood disorder that is independent of substance abuse. Finally, if both mood and substance use improve on medication (left column, table 1), it suggests that the depression contributes to the etiology of substance abuse, as in self-medication. Antidepressant Treatment in Alcoholism
In the 1960's and 1970's, nine placebo-controlled studies of antidepressant medications (mainly tricyclics) (TCAs) in alcoholic TABLE 1. Predicted outcome of antidepressant treatment across
mood disorder/substance abuse relationships.
Mood improves (vs. placebo) Drug use improves (vs. placebo)
Depression contributes to substance abuse: "self-medication" Yes
Depression and substance abuse are independent Yes
Yes
No
Depression is substance-induced No (high placebo resonse) No
SOURCE: Adapted from Nunes et al. 1994.
patients were reported. These studies have been reviewed extensively by Ciraulo and Jaffe (1981) and by Liskow and Goodwin (1987). No study demonstrated superiority of TCAs except for some short-lived effects, probably attributable to amelioration of withdrawal symptoms. However, both research pairs concluded that the studies were seriously flawed and that further study of antidepressant medication treatment of depressed alcoholics was needed. Both the doses of TCAs and the trial lengths (mostly 3 weeks or less) were inadequate. Outcome measures were narrowly focused on either depression or drinking behavior, but not both. Methods of diagnosing affective disorder were either unspecified or based on cross-sectional scales, which are not adequate measures of primary affective disease in the setting of alcoholism (Keeler et al. 1979). One promising pilot study did demonstrate successful open treatment with imipramine of a small series of alcohol and sedative abusers with panic disorder (Quitkin et al. 1972). This study had the advantage of a carefully diagnosed, homogeneous sample, but required replication in larger controlled trials. In a first replication attempt, Nunes and colleagues (1993) conducted an open-label trial of imipramine followed by double-blind, placebocontrolled discontinuation for responders. Subjects were outpatients who currently met criteria for Diagnostic and Statistical Manual of Mental Disorders, 3d ed. revised (DSM-III-R) (American Psychiatric Association 1987) alcohol abuse or dependence and also had DSM- III-R major depression or dysthymia. Experienced research psychiatrists interviewed and diagnosed the patients. Depressive syndromes were either chronologically primary, antedating the onset of alcohol abuse on a lifetime basis, had persisted during past abstinent periods, or were chronic. Eighty-five patients met inclusion criteria and entered the trial, and 60 completed the minimum adequate trial of 6 weeks of imipramine. The mean dose of imipramine was 263+77 mg/day and the mean blood level was 368+264 nanograms per milliliter (ng/mL). In addition to weekly visits with a treating psychiatrist, each patient received one weekly session of alcoholism counseling, and all patients were encouraged to attend Alcoholics Anonymous. Of the 60 completers, 27 (45 percent) were rated as "responders" to open imipramine after the initial 12 weeks with a substantial improvement in both mood Clinical Global Impression (CGI) change score of 2, "much improved," or 1, "very much improved," and drinking behavior. A rating of substantial improvement in drinking required either abstinence (18 cases, 30 percent) or a substantial reduction in quantity consumed and an absence of functional impairment. Another three patients responded after increases in imipramine dosage, and five more responded after brief courses of disulfiram, so that a total of 35 (58 percent) were ultimately called responders. Twenty-three of the responders entered and completed the 6-month, double-blind discontinuation phase in which they were randomized either to remain on imipramine or taper off imipramine onto placebo. The principal endpoint was relapse during the 6-month followup period, defined as loss of either mood response or drinking response or both. The relapse rate was lower on imipramine (31 percent, 4/13) than on placebo (70 percent, 7/10), a difference that approached statistical significance (Fisher's exact p = 0.09, two-tailed). Most relapses involved near-simultaneous return of both depression and drinking. This study differed from previous research by providing a medication trial of adequate dosage and duration, and by selecting depression via syndromal criteria rather than cross-sectional symptoms. The results suggest that antidepressant medication treatment is useful in depressed outpatient alcoholics, both in treating depression and in inducing remission of drinking and preventing relapse. The findings provide preliminarily support for the hypothesis that depression plays some role in the etiology of drinking in a selected subgroup. Replication is clearly needed in larger controlled trials, along with further work on developing criteria for selecting medicationresponsive depressed alcoholics. The majority of patients selected for this trial had depression that was chronologically primary, because the investigators felt that this history would characterize self-medicators. Interestingly, Mason and Kocsis (1991) recently completed a methologically sound, placebo-controlled trial of desipramine in alcoholics who all had depression that was chronologically secondary, but had persisted during inpatient detoxification. Their results also suggested that desipramine was useful in treating both depression and drinking (Mason and Kocsis 1991). This suggests that the primarysecondary distinction may be of limited utility as a selection criterion, since both primary and secondary depressions appear to respond to medication. Persistence of depression after inpatient detoxification may be more useful, although in practice it is not always possible to arrange hospitalization. different individuals with the same set of mental and drug use disorders are more alike than different individuals who have only one disorder. Assume, for example, that there are many types of depression and many possible etiologies of depression. From this assumption, it is clear that a study of people with depression is a study of a very heterogeneous group of people. Assume, also, that nicotine affects only people with a certain type of depression in a certain way, and that these are the people who become addicted to nicotine. In this scenario, those individuals who have a homogeneous type of depression coupled with nicotine addiction may be more like each other than a group of heterogeneously depressed people who are not addicted to nicotine. This scenario seems quite possible in light of Glassman's work (this volume) on the relationship between depression and nicotine. Research on the treatment of individuals with comorbid mental and addictive disorders holds promise for a greater understanding of the relationship among these disorders and the potential for better treatments. To date, however, research in this area has been limited. A meeting was held on September 27 and 28, 1994, to highlight some of the ongoing research in this area and to stimulate further research. Not only was research on the treatment of comorbid mental and addictive disorders addressed, but the additional problem of human immunodeficiency virus (HIV) in the context of these comorbid disorders was a topic of focus. The findings presented at the meeting could not be viewed as a definitive statement on this complex subject. On the contrary, only a limited number of combinations of comorbid mental and addictive disorders have been researched, and no one type could be fully addressed within the confines of any one meeting. The meeting was chaired by Lisa Simon Onken, Ph.D., Jack Blaine, M.D., Sander Genser, M.D., M.P.H., and Arthur MacNeill Horton, Jr., Ed.D. of the National Institute on Drug Abuse. Presentations were given by David Barlow, Ph.D., Robert Brooner, Ph.D., Kate Carey, Ph.D., Linda Cottler, Ph.D., Francine Cournos, M.D., John Docherty, M.D., Alexander Glassman, M.D., Bridget Grant, Ph.D., Edward V. Nunes, M.D., Kim Mueser, Ph.D., Bruce J. Rounsaville, M.D., Paul Satz, Ph.D., Andrew Shaner, M.D., and George Woody, M.D. The presentations given by these scientists underscore the promise that research on comorbid mental and addictive disorders holds for future treatment advances and resultant public health benefits. Just as the meeting could not fully address the full range of comorbid mental and addictive disorders and associated HIV issues, neither can this monograph. However, the chapters that follow, written by many of the participants of the meeting, are examples of exciting research being done in this important area, and they help to define the need for further research. It is the hope of all the editors of this monograph that the readers will be inspired by the contributions that follow.
REFERENCES
Regier, D.; Farmer, M.E.; Rae, D.S.; Locke, B.Z.; Keith, S.J.; Judd, L.L.; and Goodwin, F.K. Comorbidity of mental disorders with alcohol and other drug abuse. Results from the Epidemiologic Catchment Area (ECA) study. JAMA 264:2511-2518, 1990.
AUTHORS
Lisa Simon Onken, Ph.D. Jack D. Blaine, M.D. Sander Genser, M.D., M.P.H. Arthur MacNeill Horton, Jr., Ed.D. Division of Clinical Research National Institute on Drug Abuse 5600 Fishers Lane Rockville, MD 20857 Depression, Substance Use, and Sexual Orientation as Cofactors in HIV-1 Infected Men: Cross-Cultural Comparisons Paul Satz, Hector F. Myers, Mario Maj, Fawzy Fawzy, David L. Forney, Eric G. Bing, Mark A. Richardson, and Robert Janssen
INTRODUCTION
The co-occurrence of major depression in medical populations has been the subject of much controversy in the past decade. Although some investigators have suggested that reports of an increased prevalence of depression may, in part, be due to misclassification based on physician reliance on self-report methods (Perez-Stable et al. 1990) or on failure to adjust for symptoms induced by physical illness (Plumb and Holland 1977), most studies have suggested that the prevalence of depression is high, although often undetected by the primary care physician (Perez-Stable et al. 1990; Schulberg et al. 1985). Schulberg and colleagues (1985) state that it is unclear whether this oversight reflects true limitations in the physician's diagnostic acumen, his/her lack of concern for social implications, or whether it is an artifact of existing classification procedures. Regardless of the reasons for its underreporting, the detection and treatment of depression is crucial, especially for medically ill patients, because depressive disorders may adversely affect survival, length of hospital stay, compliance with therapy, ability to care for oneself, and quality of life (Schulberg et al. 1985). These concerns are particularly relevant to human immunodeficiency virus type 1 (HIV-1), where the co-occurrence of major depression has received only limited recent attention. Based on initial reports using chart reviews (Perry and Tross 1984), it has been suggested that over 83 percent of hospitalized acquired immunodeficiency syndrome (AIDS) patients have significant disturbances in mood. Unfortunately, anecdotal studies such as this fail to use structured diagnostic interviews to distinguish transitory dysphoria in response to the clinical condition and its treatment from syndromal depression. The latter disorder is more serious and merits direct clinical intervention because it may be predictive of more accelerated course and early mortality. In addition, such reporters do not investigate other cofactors that could account for the co-occurrence of depression in HIV-1 disease. The purpose of this chapter, therefore, is to examine the evidence of the co-occurrence of major depression in persons infected with HIV1, with special attention to the potential role that cofactors such as substance use and sexual orientation (i.e., being gay or bisexual) might play in accounting for the association. The chapter is organized into two parts. Part I presents a brief summary of the literature on syndromal depression in HIV-1 that was part of a larger review on the assessment of mood disorder in medical populations (Satz et al., in press). Part II presents a reanalysis of data from two large recent cohort studies of HIV-1 in populations in the United States and abroad, the World Health Organization (WHO) Multicentre Study of HIV-1 (Maj et al. 1994a, 1994b). The latter provides a more direct test of the relationship of syndromal depression in HIV-1, with special focus on substance use and sexual orientation as important cofactors.