Fluoxetine

INTRODUCTION

Addiction can be viewed as a dysregulation of brain reward systems resulting in compulsive use and loss of control over drug taking (1). Therefore, pharmacological treatments for drug dependence have targeted decreasing the rewarding effects of the drug or decreasing response to triggers, such as stress or environmental cues. Studies in rodents have linked corticosteroids with animal models of addiction in each of these areas [reviewed in Ref. (2)]. Cocaine administration elevates peripheral corticosteroid levels (3). This may contribute to the addictive properties of cocaine (2). Elevating plasma cortisol levels, either by exogenous administration or stress, facilitates the acquisition of stimulant administration in preclinical rodent models (4). Surgical or chemical adrenalectomy decreases self-administration (5) and reinstatement (2,6,7) in laboratory animals. However, studies of rat reinstatement models suggest that drug-, cue-, and stress-induced reinstatement of drug seeking have different biochemical and anatomical pathways [reviewed in Refs. (8,9)]. The hypothalamic-pituitary-adrenal (HPA) axis may play a more critical role in stress-induced reinstatement (8-10) than exposure to drug or cues.

Cues conditioned to cocaine administration activate the adrenocortical axis in both preclinical (11) and human laboratory research (12). Preclinical researchers hypothesize that increases in cortisol may contribute to craving and drug-seeking behavior. The relationship of the adrenocortical response after cocaine cues to subjective craving effects has yet to be studied. Cocaine cues reliably induce not only craving responses but also increases in other subjective measures such as euphoria and anxiety in human laboratory studies. It is not clear to which positive or negative subjective responses glucocorticoids contribute, if any. DeVries et al. (11) hypothesized in their preclinical study that the cue-induced increase in cortisol "could further predispose an organism to the reinforcing effects of the drug or enhance the susceptibility to drug-taking behavior." Alternatively, DeVries noted the possibility that conditioned effects may be related to the anxiogenic properties of cocaine. Deroche and colleagues (6) argued in favor of the former possibility (i.e., increasing reinforcing effects). In our earlier clinical studies of cue-induced craving, we found that haloperidol blocked all subjective responses to cocaine cues (anxiety, craving, and euphoria) (12), whereas nicotinic agents selectively modulated craving (nicotine pretreatment increased and mecamylamine pretreatment decreased) without affecting euphoria or physiological responses (13,14). Because glucocorticoids have been shown to have both reinforcing and anxiogenic effects, blocking cue-induced increases in cortisol could conceivably block all of the subjective responses like haloperidol. Alternately, as hypothesized by DeVries et al. (11), they may contribute selectively to either cue-induced reinforcement, anxiety, or other effects.

Corticosteroid-depleting agents, such as metyrapone, widely used by neuroendocrine researchers, produce addisonian-like side effects when administered chronically to patients. An alternative approach is the use of serotonergic medications, such as the antidepressant fluoxetine, that are well tolerated in widespread clinical use. Even though these medications acutely activate the adrenocortical axis, they have the opposite effect after chronic administration, producing a hyporesponsiveness (15,16). Chronic administration of serotonergic and other classes of antidepressants upregulates glucocorticoid type II receptors in the hippocampus. These receptors are believed to mediate feedback inhibition of the glucocorticoid axis (17). In the study described below, we evaluated glucocorticoid responses to cocaine cues in cocaine addicts enrolled in a clinical fluoxetine trial. This enabled us to evaluate the relationship between cue-induced activation of the adrenocortical axis and craving, high, and anxiety. If cortisol contributes to the reinforcing effects of cocaine cues, as hypothesized by DeVries et al. (11), Deroche et al. (6), and Goeders (2), then blocking the increase in cortisol should block craving or euphoric responses. Alternatively, if cortisol is related more to anxiogenic effects, then blocking its increase should attenuate cue-induced increases in anxiety.

MATERIALS AND METHODS

Clinical Trial General Design

This study recruited from a larger (82 subject), 11-week, double-blind, randomly assigned pharmacotherapy trial of fluoxetine for cocaine dependence. The clinical trial was undertaken primarily to replicate earlier encouraging double-blind study results by Batki and colleagues (18). The design was similar (e.g., recruiting subjects from the same population). Subjects were recruited by newspaper ad. They took 40 mg of fluoxetine [consistent with a clinically approved dosage for the treatment of depression (19)] or placebo for 11 weeks. Candidates for the study taking psychotropic medications were excluded. Potential subjects underwent physical examination and routine laboratory screening. All were diagnosed as cocaine-dependent with the Structured Clinical Interview for DSM-IV (SCID-IV) and reported cocaine as the primary illicit drug used. Patients with bipolar affective disorder or psychotic disorders were excluded from the treatment and cue reactivity studies. Patients dependent on heroin were also excluded. Neither individuals with additional drug dependencies, except for heroin, nor patients with depression or anxiety disorders were excluded. To do so would have selected for a population of subjects not representative of those in treatment (virtually all of whom are users of other drugs and many with depression or anxiety disorders). To exclude such subjects would decrease generalizability of pharmacotherapy to the majority of cocaine addicts. Use of other illicit drugs was monitored with random urine toxicology testing and by subject report twice a week. Urine was also collected weekly for quantitative levels of benzoylecgonine, a metabolic product of cocaine and objective quantitative indication of recent use. All subjects participated in weekly cognitive-behavioral relapse prevention groups.

Subject Description

Twenty-nine subjects were recruited from participants in the fluoxetine clinical trial. Upon recruitment for the fluoxetine treatment study, subjects were approached for participation in the cue reactivity study. Subjects willing to participate were tested for cue reactivity. Noncue-reactive subjects were excluded as in earlier studies (12). Including noncue-reactive patients would diminish the statistical power to detect a medication's ability to block cue reactivity.

Cue reactivity was assessed by showing a 30-second segment of a videotape containing images of crack cocaine preparation rituals. Potential subjects who experienced no increase in "desire to use" after viewing the video were excluded. The tape screening excluded approximately 30% of patients who were not responsive (denied any craving). Other studies have reported a similar proportion of patients denying cocaine craving in response to cues (20). Exclusion of nonresponders yielded 29 cue-reactive patients to begin the study. Subjects gave informed consent. They were paid for participating. The study was approved by the University of California, San Francisco, Committee on Human Research.

Study Design

Cue reactivity procedures were similar to our earlier studies (12,13). The major difference from earlier studies is that a parallel group design was used to evaluate chronic medication effects as opposed to the earlier single dose, crossover study design. Subjects participated in two cue reactivity sessions, one during the baseline period of the clinical trial when all subjects were administered placebo and the second after approximately 5 weeks of fluoxetine or placebo treatment. The 5-week time period was selected because poor treatment retention (common in this patient population) precludes intervals much longer than 1 month. This interval is consistent with the time course of therapeutic efficacy in our past cocaine clinical pharmacotherapy trial (the majority of clinical improvement occurred by 3 weeks) (18).

The patients' reactivity to neutral and cocaine cues was evaluated on both pretreatment and medication study days. Ideally, we would have preferred to evaluate neutral cue reactivity on a different day from cocaine cues as in our past study (12). However, with the current study, this would have incurred the cost of an additional day of hospitalization. In our past outpatient studies, we administered neutral cues on the same day as cocaine cues (13,14). We only observed modest effects of neutral cues that did not affect our ability to detect significant pharmacological effects on responses to cocaine cues (13). Procedures were done at the same time each study day to minimize diurnal variation. Subjects were hospitalized by 4:00 PM the afternoon prior to the study session to decrease the variability from sleep and circadian rhythm and from external stressors. Given the short duration of pharmacokinetic effects associated with cocaine (21), admitting patients the day before should be sufficient to preclude direct effects from recent abuse.

At approximately 8:30 AM of the study day, a heparin-treated catheter (heplock) was inserted into a prominent arm vein. At 9:00 AM the neutral cues procedure began and at 10:00 AM the cocaine cues procedure began.

Neutral Cues

A 5-minute neutral cue videotape showed outdoor scenes of acorns and seashells. Following the tape, subjects handled seashells for approximately 15 minutes and made a design with them to incorporate an equal amount of tactile and motor stimuli as with the cocaine cue procedure.

Cocaine Cues

The 5-minute cocaine cue videotape contained auditory and visual cues specific to crack cocaine with actors simulating preparing, purchasing, and smoking crack cocaine. The composition of the film was similar to one provided by Childress and associates except that our film had cues specific to the San Francisco Bay Area, such as regional dialect and paraphernalia. In the subsequent period lasting approximately 15 minutes, patients handled cocaine paraphernalia including a crack pipe and benzocaine-derived white 5 to 15 mm in diameter crystals comparable in size to that typically used and difficult to distinguish from cocaine. The psychological measures and plasma cortisol sampling were made before and after neutral cues and before and after cocaine cues (approximately 25 minutes after the start of the cue procedure). We chose to focus on these time points rather than include later time points in the analysis, because the time immediately following the cocaine cues was the mean peak plasma cortisol concentration time point and the cortisol level rapidly returned to baseline afterward in some subjects.

Measures

We used the Within Session Rating Scale (22) to assess craving for cocaine before and after cue exposure: patients estimated the intensity of two craving measures ("desire to use" and "likely to use") and various other drug-related states on a 100-ram visual analogue scale: above appropriate numbers on the scale were the following modifiers: "not at all," "mildly," "moderately," and "extremely." Anxiety was calculated from the mean of the "nervous," "anxious," and "fearful" visual analogue scales for each subject. Reported cocaine outcome variables included days of cocaine use per week and cost of cocaine used per week (both derived from the Quantitative Cocaine Inventory (QCI) (23). Our objective outcome measure, urine concentration of benzoylecgonine (the major metabolite of cocaine), was determined by gas chromatography (24). Plasma cortisol levels were measured by protein-binding globulin assay (PBG) (25).

Data Analysis

In the primary analysis, we compared the change in each measure from the period immediately before cue exposure to the period immediately after cue exposure. These changes were analyzed within a session, between the two sessions (baseline and poststudy medication), and between the two groups during the final postmedication session or for the change between sessions. Instead of using a single analysis of variance (ANOVA) model, we chose a single degree of freedom test focused on our hypotheses. These changes were analyzed with a two-sample Wilcoxon rank sum W test using exact p values for between-groups analyses and a Wilcoxon matched-pairs Z test for within-subjects changes, because this is a more conservative statistical approach in view of the limited sample size and the observed range in distribution of some of the dependent variables. As a secondary analysis, we estimated the correlation between changes in craving and plasma cortisol with clinical outcome measures using Kendall's tau (SPSS, Inc., Chicago IL).

RESULTS

Twenty-two of 29 randomized subjects completed both cue reactivity sessions. Twelve of these subjects received active fluoxetine and 10 received placebo. Demographics are in Table 1. Besides nicotine and caffeine, these primarily cocaine-dependent subjects rarely used drugs from other categories except sometimes alcohol and marijuana. No one appeared intoxicated or in withdrawal at the time of testing. Chronic fluoxetine was well tolerated with no major adverse effects.

Subjective measures increased significantly from the precocaine cues presentation to postcocaine cues at the pretreatment session (Table 2). As expected, there was no significant difference between the placebo and fluoxetine groups in the placebo run-in session. Response measures did not increase significantly in the pretreatment session after neutral cues presentation and the "likely to use" subjective measure significantly decreased (Z = -2.0; p < 0.05).

Pretreatment session mean (median) [+ or -] SD cortisol plasma levels ([micro]g/ dL) showed a circadian drop from 11.0 (10.0) [+ or -] 4.5 at 9:00 AM just prior to the neutral cues to 10.1 (9.1) [+ or -] 3.9 at 9:30 AM following the neutral cues to 9.2 (9.3) [+ or -] 3.3 at 10:00 AM just prior to cocaine cues. After cocaine cues, cortisol level significantly increased to 10.3 (10.1) [+ or -] 3.8 (Z = -2.1; p < 0.04) (Table 1).

During the second session following approximately 5 weeks of fluoxetine or placebo treatment, the intensity of subjective measures increased significantly after cocaine cues (Table 3). No significant differences were found in subjective measures when comparing the placebo group with the fluoxetine group (Table 3). However, comparing the cue reactivity session before medication to the second session after medication (within group change between sessions), the fluoxetine-treated group showed a significant increase in the "likely to use" measure alter cocaine cue presentation (Z = - 2.0; p < 0.05). The placebo group showed a nonsignificant decrease. This change in reactivity from pretreatment to treatment cue sessions (change score) was also statistically significantly different between groups (W = 80; p < 0.03) (Table 4, Fig. 1). This difference appears to be due in part to a statistically significantly (p < 0.01) lower precues rating of "likely to use" in the fluoxetine-treated group during the treatment period session (mean [+ or -] SD, 24 [+ or -] 21 in the active treatment session compared to 65 [+ or -] 29 in the placebo run-in session). Other subjective responses ("desire to use," "high," and "anxiety") were not significantly different between groups.

[FIGURE 1 OMITTED]

Adrenocortical axis responsiveness to cocaine cues as measured by cortisol was significantly different between the fluoxetine and placebo group at the second session (W = 145.5; p < 0.05) (Fig. 2). Plasma cortisol levels were similar after fluoxetine or placebo treatment prior to cocaine cues. Cocaine cues significantly elevated cortisol in the placebo group (Z = -2.3; p < 0.03), consistent with the significant increase induced by cocaine cues in the larger group of subjects evaluated at baseline prior to medication randomization. However, no significant effect of cocaine cues on cortisol was observed after fluoxetine treatment. The rise in cortisol levels and subjective responses were not related.

[FIGURE 2 OMITTED]

Urine benzoylecgonine levels at the time of the second session (but not at the end of the study) were statistically significantly greater in the placebo group (W = 168; p = 0.0001). Urine benzoylecgonine level was unrelated to cortisol response or subjective response. Neither the subjective craving and hormonal responses to cocaine cues at either the first or second session nor the change in reactivity between sessions were correlated significantly with treatment outcome measures (reported days of use in the past week, cost of cocaine use in the past week, and benzoylecgonine levels) at the time of the second session or at the end of the study for the group as a whole.

DISCUSSION

The increased subjective measures of craving and high and the rise in cortisol levels after the presentation of cocaine cues in the first session suggest that the cues were effective in eliciting craving and other expected subjective responses. The rise in cortisol level replicates our earlier work (12) and is consistent with preclinical cocaine-conditioning studies (11). The significant effect on cortisol was comparable to that reported in a psychological stress imagery paradigm (26), but somewhat less than that produced by cocaine itself (27). This response pattern cannot be attributed to nonspecific effects, because the presentation of the neutral cues produced only minimal changes or lower ratings while the cortisol level continued its circadian decrease.

Fluoxetine was associated with significantly less cortisol reactivity to cocaine cues than that of the placebo group after 5 weeks of treatment. To our knowledge, this is the first clinical demonstration of dampened adrenocortical reactivity to cocaine cues induced by chronic antidepressant administration. Chronic administration of serotonergic and other classes of antidepressant agents appear to upregulate glucocorticoid type II receptors in the hippocampus, increasing feedback inhibition of the glucocorticoid axis (17). Increasing feedback inhibition could attenuate the mild activation of the adrenocortical axis produced by cocaine cues.

Most subjective euphoric, craving, and anxiety responses to cocaine cues were not significantly changed by fluoxetine treatment. One measure of cocaine craving on the Within Sessions Rating Scale, "likely to use," was significantly increased across sessions by chronic fluoxetine, whereas other measures of craving and anxiety were not significantly changed. This is quite possibly a statistical artifact. However, there are several possible explanations of why the "likely to use" measure could be modulated in the absence of a change in other subjective response measures, particularly "desire to use." Many patients who deny what they identify as craving are nevertheless aware that drug cues trigger drug-seeking behavior in them (making them "likely to use"). Clinically, after engaging in treatment, some patients are more comfortable acknowledging the risk of using cocaine (reflected in the "likely to use" measure) than acknowledging desire, which is associated with considerable stigma. Although our subjects also showed a robust response in "desire to use," it is still likely that these two craving ratings measure different aspects of craving. "Likely to use" could reflect the state of powerlessness emphasized by 12-step treatment approaches (28) and quantified by the well-validated "locus of control" measure (29). An increased rating of "likely to use" could also indicate an increased awareness of impulse to use. Miller and Gold (30) found that cocaine-dependent persons reported that craving was not a major reason for relapse; they attributed it to impulsive action with no known cause. Animal and human studies have found fluoxetine to decrease impulsiveness (31-34). Therefore, a change in feeling "likely to use" in response to cocaine cues in the absence of a change in "desire to use" might still be clinically relevant.

The change in our cue-induced "likely to use" response may be due to a serotonin effect unrelated to HPA axis activation. Because fluoxetine increases synaptic serotonin levels, increased cue-induced craving is consistent with a report by Satel and colleagues that an acute serotonin depletion diet diminishes cocaine cue-induced craving (35). However, decrease in serotonin turnover and neuronal firing from chronic fluoxetine administration (36) may more closely resemble serotonin depletion, which in rats decreases cocaine-seeking behavior during extinction but enhances responding for cocaine after a priming dose (37). Fluoxetine also attenuates cocaine seeking in rodents during extinction (38,39). The decreased preexposure rating of "likely to use" in our study during fluoxetine treatment along with the relative increase in rise of "likely to use" in response to cues is consistent with these preclinical findings. However, chronic fluoxetine also attenuates cocaine-seeking behavior in rats exposed to the self-administration environment or cues (light/tone stimuli) but not to priming (38,39). Perhaps, in humans an intense exposure to cocaine cues more closely resembles priming in rats (especially because humans report some "high" or "euphoria" in response to cues) than exposure to cues. Therefore, degree of exposure to environmental cues may strongly influence outcome to treatment with this class of drug.

The finding of decreased cortisol responsivity to cocaine cues in the absence of a decreased effect on subjective craving does not support the relationship hypothesized by preclinical researchers between corticosteroids, craving, and drug-seeking behavior (40). With respect to stress-induced reinstatement, Shalev and colleagues (41) have suggested that, although corticosteroids play a permissive role, their effects are not associated with stress-induced increases in corticosteroid concentrations. Perhaps corticosteroids play a permissive role not only for stress-induced reinstatement but for cue-induced craving also. We have reported that cocaine cues activate peripheral and central measures of dopaminergic function (12). Fluoxetine increases extracellular dopamine in the prefrontal cortex (42), and chronic administration of this drug may modulate incentive motivational responses to cocaine cues. It remains to be determined how these measures are clinically relevant to craving or drug-seeking behavior. The ability to selectively enhance or antagonize different aspects of cue response could increase our understanding of this process and lead to more effective treatments for cocaine dependence.

ACKNOWLEDGEMENTS

Statistical consultation was provided by Kevin Delucchi, Ph.D., and William Hargreaves, Ph.D. Reese T. Jones, M.D., provided helpful editorial comments. Thomas Everhart, Ph.D., analyzed the urine for benzoylecgonine concentrations. Kaye Welch provided editorial assistance. This work was supported by NIDA Grant No. P50-DA09253, San Francisco Treatment Research Center; NIDA Grant No. T32-DA07250 (NIDA-sponsored fellowship) and NIH Grant No. 5-MOI-RR00083 from the Division of Research Resources (for the General Clinical Research Center, San Francisco General Hospital).

REFERENCES

(1.) Koob GF, Le Moal M. Drug addiction, dysregulation of reward, and allostasis. Neuropsychopharmacology 2001; 24:97-129.

(2.) Goeders NE. Stress and cocaine addiction. J Pharmacol Exp Ther 2002; 301:785-789.

(3.) Dackis CA, Gold MS. Biological aspects of cocaine addiction. In: Volkow ND, ed. Cocaine in the Brain. Houston, TX: Ritgers University Press, 1988.

(4.) Goeders NE, Guerin GF. Non-contingent electric footshock facilitates the acquisition of intravenous cocaine self-administration in rats. Psychopharmacology 1994; 114:63 70.

(5.) Goeders NE, Guerin GF. Effects of surgical and pharmacological adrenalectomy on the initiation and maintenance of intravenous cocaine self-administration in rats. Brain Res 1996; 722:145-152.

(6.) Deroche V, Marinelli M, Le Moal M, Piazza PV. Glucocorticoids and behavioral effects of psychostimulants. II. Cocaine intravenous self-administration and reinstatement depend on glucocorticoid levels. J Pharmacol Exp Ther 1997; 281:1401-1407.

(7.) Goeders NE, Clampitt DM. Potential role for the hypothalamopituitary-adrenal axis in the conditioned reinforcer-induced reinstatement of extinguished cocaine seeking in rats. Psychopharmacology (Berl) 2002; 161:222-232.

(8.) Shaham Y, Shalev U, Lu L, De Wit H, Stewart J. The reinstatement model of drug relapse: history, methodology and major findings. Psychopharmacology (Berl) 2003; 168: 3-20.

(9.) Shalev U, Grimm JW, Shaham Y. Neurobiology of relapse to heroin and cocaine seeking: a review. Pharmacol Rev 2002: 54:1-42.

(10.) Erb S, Shaham Y, Stewart J. The role of corticotropin-releasing factor and corticosterone in stress- and cocaine-induced relapse to cocaine seeking in rats. J Neurosci 1998; 18:5529-5536.

(11.) DeVries AC, Taymans SE, Sundstrom JM, Pert A. Conditioned release of corticosterone by contextual stimuli associated with cocaine is mediated by corticotropin-releasing factor. Brain Res 1998; 786:39-46.

(12.) Berger SP, Hall S, Mickalian JD, Reid MS, Crawford CA, Delucchi K, Carr K, Hall S. Haloperidol antagonism of cue-elicited cocaine craving. Lancet 1996; 347:504-508.

(13.) Reid MS, Mickalian JD, Delucchi KL, Hall SM, Berger SP. An acute dose of nicotine enhances cue-induced cocaine craving. Drug Alcohol Depend 1998; 49:95-104.

(14.) Reid MS, Mickalian JD, Delucchi KL, Berger SP. A nicotine antagonist, mecamylamine, reduces cue-induced cocaine craving in cocaine-dependent subjects. Neuropsychopharmacology 1999; 20:297-307.

(15.) Lesch KP, Hob A, Schulte HM, Osterheider M, Muller T. Long-term fluoxetine treatment decreases 5-HTIA receptor responsivity in obsessive-compulsive disorder. Psychopharmacology (Berl) 1991; 105:415-420.

(16.) Jensen JB, Jessop DS, Harbuz MS, Mork A, Sanchez C, Mikkelsen JD. Acute and long-term treatments with the selective serotonin reuptake inhibitor citalopram modulate the HPA axis activity at different levels in male rats. J Neuroendocrinol 1999; 11:465-471.

(17.) Peiffer A, Veilleux S, Barden N. Antidepressant and other centrally acting drugs regulate glucocorticoid receptor messenger RNA levels in rat brain. Psychoneuroendocrinology 1991; 16:505-515.

(18.) Batki SL, Washburn AM, Delucchi K, Jones RT. A controlled trial of fluoxetine in crack cocaine dependence. Drug Alcohol Depend 1996; 41:137-142.

(19.) Physician's Desk Reference. Montvale, NJ: Medical Economics Company, Inc., 1998.

(20.) Robbins SJ, Ehrman RN, Childress AR, O'Brien CP. Relationships among physiological and self-report responses produced by cocaine-related cues. Addict Behav 1997; 22:157-167.

(21.) Hart CL, Jatlow P, Sevarino KA, McCance-Katz EF. Comparison of intravenous cocaethylene and cocaine in humans. Psychopharmacology (Berl) 2000; 149:153-162.

(22.) Childress AR, McLellan AT, O'Brien CP. Abstinent opiate abusers exhibit conditioned craving, conditioned withdrawal and reductions in both through extinction. Br J Addict 1986; 81:655-660.

(23.) Gawin FH, Kleber HD. Cocaine abuse treatment. Open pilot trial with desipramine and lithium carbonate. Arch Gen Psychiatry 1984; 41:903-909.

(24.) Jacob P III, Elias-Baker BA, Jones RT, Benowitz NL. Determination of benzoylecgonine and cocaine in biologic fluids by automated gas chromatography. J Chromatogr 1987; 417:277-286.

(25.) Murphy BE. Some studies of the protein-binding of steroids and their application to the routine micro and ultramicro measurement of various steroids in body fluids by competitive protein-binding radioassay. J Clin Endocrinol Metab 1967; 27:973-990.

(26.) Sinha R, Catapano D, O'Malley S. Stress-induced craving and stress response in cocaine dependent individuals. Psychopharmacology (Berl) 1999; 142:343-351.

(27.) Ward AS, Collins ED, Haney M, Foltin RW, Fischman MW. Ketoconazole attenuates the cortisol response but not the subjective effects of smoked cocaine in humans. Behav Pharmacol 1998; 9:577-586.

(28.) Gilbert FS. Development of a "steps questionnaire." J Stud Alcohol 1991; 52:353-360.

(29.) Vielva I, Iraurgi I. Cognitive and behavioural factors as predictors of abstinence following treatment for alcohol dependence. Addiction 2001; 96:297-303.

(30.) Miller NS, Gold MS. Dissociation of "conscious desire" (craving) from and relapse in alcohol and cocaine dependence. Ann Clin Psychiatry 1994; 6:99-106.

(31.) Bizot J, Le Bihan C, Puech AJ, Hamon M, Thiebot M. Serotonin and tolerance to delay of reward in rats. Psychopharmacology (Berl) 1999; 146:400-412.

(32.) Fairbanks LA, Melega WP, Jorgensen MJ, Kaplan JR, McGuire MD. Social impulsivity inversely associated with CSF 5-HIAA and fluoxetine exposure in vervet monkeys. Psychopharmacology 2001; 24:370-378.

(33.) Alcami Pertejo M, Peral Guerra M, Gilaberte I. Open study of fluoxetine in children with autism. Actas Esp Psiquiatr 2000; 28:353-356.

(34.) Silva H, Jerez S, Paredes A, Salvo J, Renteria P, Ramirez A, Montes C. Fluoxetine in the treatment of borderline personality disorder. Actas Luso Esp Neurol Psiquiatr Cine Afines 1997; 25:391-395.

(35.) Satel SL, Krystal JH, Delgado PL, Kosten TR, Charney DS. Tryptophan depletion and attenuation of cue-induced craving for cocaine. Am J Psychiatry 1995; 152:778-783.

(36.) Wong DT, Bymaster FP, Engleman EA. Prozac (fluoxetine, Lilly 110140), the first selective serotonin uptake inhibitor and an antidepressant drug: twenty years since its first publication. Life Sci 1995; 57:411-441.

(37.) Tran-Nguyen LT, Bellew JG, Grote KA, Neisewander JL. Serotonin depletion attenuates cocaine seeking but enhances sucrose seeking and the effects of cocaine priming on reinstatement of cocaine seeking in rats. Psychopharmacology (Berl) 2001; 157:340-348.

(38.) Burmeister JJ, Lungren EM, Neisewander JL. Effects of fluoxetine and d-fenfluramine on cocaine-seeking behavior in rats. Psychopharmacology (Berl) 2003; 168:146-154.

(39.) Baker DA, Tran-Nguyen TL, Fuchs RA, Neisewander JL. Influence of individual differences and chronic fluoxetine treatment on cocaine-seeking behavior in rats. Psychopharmacology (Berl) 2001; 155:18-26.

(40.) Goeders NE, Guerin GF. Role for corticosterone in intravenous cocaine self-administration in rats. Neuroendocrinology 1996; 64:337-348.

(41.) Shalev U, Marinelli M, Baumann MH, Piazza PV, Shaham Y. The role of corticosterone in food deprivation-induced reinstatement of cocaine seeking in the rat. Psychopharmacology (Berl) 2003; 168:170-176.

(42.) Pozzi L, Invernizzi R, Garavaglia C, Samanin R. Fluoxetine increases extracellular dopamine in the prefrontal cortex by a mechanism not dependent on serotonin: a comparison with citalopram. J Neurochem 1999; 73:1051-1057.

Debra S. Harris, M.D., (1), * Steven L. Batki, M.D., (2) and S. Paul Berger, M.D. (3)

* Correspondence: Debra S. Harris, M.D., University of Cincinnati and Cincinnati VA Medical Center, 3200 Vine St., Cincinnati, OH, 45220, USA; E-mail: debra.harris4@med.va.gov.

(1) Department of Psychiatry, University of Cincinnati and Cincinnati VA Medical Center, Cincinnati, Ohio, USA

(2) Department of Psychiatry, SUNY--Upstate Medical University and Syracuse VA Medical Center, Syracuse, New York, USA

(3) Department of Psychiatry, Portland Veterans Affairs Medical Center/ Oregon Health Sciences University, Portland, Oregon, USA

COPYRIGHT 2004 Taylor & Francis Ltd.
COPYRIGHT 2005 Gale Group