During the initial period of abstinence, many cocaine-dependent individuals experience anhedonia, craving, fatigue, insomnia, and dysphoria (1–4). These symptoms are thought to be associated with cocaine-induced alterations in monoaminergic transmission (5–7). Increased levels of depressive symptoms have also been associated with increased attrition from treatment programs (8) and increased cocaine use at follow-up (9, 10). The association between these symptoms and the euphoric effects of cocaine has not been studied. Examining this association may be particularly helpful in understanding the experience of the individual who reuses cocaine after a period of abstinence. Determining how such symptoms are related to the subjective response that cocaine produces on reuse could help identify risk factors associated with relapse.

Participants included 17 non-treatment-seeking, cocaine-dependent men recruited from the community through advertisements in local newspapers. The mean age of the group was 38.2 years old (range=25 to 47), and the mean education level was 13.0 years (range=9 to 17). The group consisted of 12 African Americans, three Caucasians, one Latino, and one American Indian. All participants met DSM-III-R criteria for cocaine dependence, reported using at least 0.5 g of cocaine or crack per week (mean=5.1 g/week, SD=2.6) for the 6 months before the study, and had positive results on a urine benzoylecgonine test within 2 weeks of entering the study. All participants’ normal manner of cocaine administration was smoking. Past use of alcohol or other drugs except opiates was acceptable for participation as long as individuals did not meet DSM-III-R criteria for any other substance dependence.

Participants did not have histories of stroke, epilepsy, or axis I psychotic or mood disorders that lasted longer than 2 weeks. Pregnant women were excluded. Results of laboratory testing, including CBC, electrolytes, ECG, and EEG, were within predetermined limits. An HIV antibody test was nonreactive. Participants gave written informed consent after being apprised of the study risks. Subjects were reimbursed for participation. Urine samples for drug toxicology screens were collected every several days or whenever participants left the inpatient unit in order to ensure abstinence from self-administered cocaine or other nonstudy drugs.

Individuals participated in a 14-day inpatient study to determine the safety of a cocaine challenge during the administration of potential medication treatments for cocaine dependence. On admission, participants were administered the Addiction Severity Index (11) and the Brief Symptom Inventory (12) to assess general psychiatric distress during the 30 days before study entry. Subjects were asked to provide daily ratings of their current levels of craving for cocaine; self-reported depressive symptoms, as indexed by the Beck Depression Inventory (13); and general psychiatric distress, as indexed by the Brief Symptom Inventory.

Four days after admission, participants received a 20-mg dose of cocaine given intravenously at 1 mg/second in order to screen out subjects who tolerated cocaine poorly in the laboratory. On the following day, participants were administered 40 mg of cocaine intravenously at 1 mg/second. Before cocaine administration and at 3, 6, 10, 20, 30, and 40 minutes after cocaine administration, participants provided verbal ratings on a 0–10-point scale for self-reported high produced by cocaine (0=no effect, 10=maximal effect). After each administration of cocaine, participants were monitored with beat-to-beat blood pressure, heart rate recordings, and continuous ECG monitoring.

Subjective high following cocaine administration was quantified by calculating area under the curve. Pearson correlations were employed to determine the relationship among depressive symptoms indexed by the Beck Depression Inventory, craving for cocaine, and area under the curve for high.

Self-rated high did not significantly correlate with age, education, amount of cocaine used over the last 6 months, years of cocaine use, days since last use of cocaine, craving for cocaine, or general psychiatric symptoms on the day of cocaine administration or during the 30 days before admission (p>0.14). The correlation between area under the curve for high and Beck Depression Inventory scores on the day of infusion was significant (r=0.56, df=15, p=0.02). Because craving for cocaine and Beck depression scores on the day of cocaine administration correlated significantly (r=0.72, df=15, p=0.003), a partial correlation controlling for craving was employed. This did not affect the significant correlation between Beck depression scores and area under the curve for high (r=0.64, df=12, p=0.01).

Subsequent analyses revealed that on the day of cocaine administration, seven of the 17 participants reported clinically significant levels of depressive symptoms (i.e., Beck depression score of 11 or more) (14). To identify the symptoms that differentiated between the groups, a gamma statistic was employed (15). Gamma provides a method for examining the relationship among ordinal variables, such as Beck Depression Inventory responses. Because gamma was calculated for each of the 20 items on the Beck Depression Inventory, a family-wise correction set alpha at 0.01 (16).

Participants with higher Beck depression scores were more likely to respond positively to the following items: feelings of failure (gamma=–0.97, df=15, p<0.001), anhedonia (gamma=–0.96, df=15, p<0.001), feelings of guilt (gamma=–0.92, df=15, p<0.001), self-hatred (gamma=–1.00, df=15, p<0.001), self-blame (gamma=–1.00, df=15, p<0.001), indecisiveness (gamma=–1.00, df=15, p<0.001), motivation (gamma=–1.00, df=15, p=0.003), and fatigue (gamma=–1.00, df=15, p=0.003). Participants with higher Beck depression scores were not more likely to report sadness, discouragement, tearfulness, suicidality, or other symptoms of syndromal depression (p>0.05).

The results from this study revealed that increased levels of self-reported depressive symptoms during abstinence were positively associated with subjective ratings of high after the administration of 40 mg of cocaine on the same day. These results require consideration of a novel element in why depressive symptoms were a valid predictor for treatment outcome in some studies (9, 10). The greater high that cocaine administration produces in individuals experiencing more depressive symptoms during abstinence could result in an increased likelihood of relapse, not because of a greater drive to self-medicate depression, but because the drug effect itself is stronger in individuals suffering greater depressive symptoms. Therefore, individuals with more severe depressive symptoms may be especially vulnerable to relapse because they experience a greater high after taking cocaine. An association between intensity of drug-induced high and liability to relapse has not been demonstrated, however.

It is also noteworthy that other studies did not find an association between depression and relapse (17, 18). The inconsistency may be a function of the variability of the content and the psychometric properties of the questionnaires employed to measure depressive symptoms. Our results show that only a subset of the depressive symptoms indexed by the Beck Depression Inventory is associated with the subjective high induced by cocaine. This subset of symptoms is not assessed by the Brief Symptom Inventory, which includes a depression scale. Consequently, the Brief Symptom Inventory depression scale was not associated with area under the curve for high. These findings underscore the need to consider the content and psychometric properties of the depression scales used in future studies.

The items differentiating between individuals with high and low scores on the Beck Depression Inventory are not those that are typically seen in syndromal depression. Because these symptoms appear to be associated with the high induced by cocaine, it may be that these symptoms are important predictors of cocaine addiction and relapse.

A limitation of this study is that only 17 participants completed the protocol; thus, replication is necessary to determine the stability of the findings. Furthermore, it is unknown if the symptoms experienced during abstinence in a hospital are the same symptoms that an individual would experience outside the hospital. Environmental factors have been shown to influence mood and behavior during abstinence (19, 20).

Future research efforts could aim at more precisely identifying the symptoms experienced during cocaine abstinence in a nonhospital setting and examining the association between these symptoms and relapse rates. If certain symptoms do, indeed, predict response to cocaine use and treatment outcome, then treating these symptoms may be especially important in minimizing the probability of relapse. A better understanding of addicts’ experiences during abstinence is crucial to the treatment of cocaine addiction.

Received Sept. 18, 1998; revision received Feb. 23, 1999; accepted April 9, 1999. From the Department of Psychiatry and Biobehavioral Sciences, UCLA School of Medicine. Address reprint requests to Dr. Newton, Department of Psychiatry, UCLA Neuropsychiatric Institute, Rm. C8-846, 760 Westwood Blvd., Los Angeles, CA 90024; tnewton@ucla.edu (e-mail). Supported by grant DA-50038 from the National Institute on Drug Abuse (Dr. Ling).