Fluoxetine in Depressed AUD Adolescents: A 1-Year Follow-up Evaluation

Jack R. Cornelius, M.D., M.P.H., Oscar G. Bukstein, M.D., M.P.H., Ihsan M. Salloum, M.D.,M.P.H., Thomas M. Kelly, Ph.D., D. Scott Wood, Ph.D., and Duncan B. Clark, M.D., Ph.D.

ABSTRACT

The authors conducted the first naturalistic 1-year follow-up evaluation of 10 adolescentswith comorbid major depressive disorder and an alcohol use disorder (AUD) who had previ-ously participated in an acute phase study of open-label fluoxetine plus psychotherapy (Cor-nelius et al. 2001). The goal of this follow-up evaluation was to determine whether thedecreases in drinking and in depressive symptoms that were noted during the acute phasestudy persisted at the follow-up evaluation. At the 1-year follow-up evaluation, the groupcontinued to demonstrate significantly fewer depressive symptoms (according to the 24-itemHamilton Rating Scale for Depression) and a lower frequency of drinking (drinking days inthe last 30 days) than they had demonstrated at the baseline of the acute phase study. Surpris-ingly, all of the subjects had chosen to discontinue their antidepressant medication by thesecond month of their naturalistic follow-up period. Three subjects had experienced a re-lapse of their major depression during the follow-up period, and three others demonstrated apersistence of their original depressive episode throughout the follow-up period. Also, thenumber of drinks per drinking day continued to be high (about five per day), which was notsignificantly different from the baseline level. Thus, the long-term therapeutic effects of anacute phase trial of fluoxetine plus psychotherapy were limited. The high rate of recurrenceor persistence of major depression in our sample and in a previous sample of nonalcoholicadolescents with major depression (Emslie et al. 1998) and the significant levels of drinkingof our comorbid adolescents suggest that longer term treatment may be needed for at leastsome adolescents with major depressive disorder and alcohol use disorder.

33

JOURNAL OF CHILD AND ADOLESCENT PSYCHOPHARMACOLOGYVolume 14, Number 1, 2004© Mary Ann Liebert, Inc.Pp. 33–38

INTRODUCTION

COMORBIDITY INVOLVING major depression andsubstance use disorder is common among

adolescents, but treatment studies involvingthis comorbid population are scarce (Buksteinet al. 1992). Recently, the first long-term (1year) naturalistic follow-up study of a selec-tive serotonin reuptake inhibitor (SSRI) med-

ication was conducted involving a sample ofadolescents with major depression withoutaccompanying substance abuse (Emslie et al.1998). The results of that study demonstratedefficacy for fluoxetine versus placebo for de-creasing the depressive symptoms of adoles-cents with major depression who did notdisplay comorbid substance abuse. However,it is unclear to what extent those results apply

Pittsburgh Adolescent Alcohol Research Center, Department of Psychiatry, University of Pittsburgh School of Med-icine, Pittsburgh, Pennsylvania.

D. Scott Wood, Ph.D. served as the statistical consultant.

to adolescents with comorbid major depres-sion and a substance use disorder, becauseadolescents with substance use disorderswere excluded from participation in thatstudy. Riggs et al. (1997) conducted a 7-weekopen-label trial of fluoxetine in a study sam-ple involving adolescents with major depres-sion and a variety of comorbid substance usedisorders. The results of that open-labelstudy suggested within-group efficacy forfluoxetine for decreasing the depressivesymptoms of those patients, although nomeasures of substance use were utilizedbecause the subjects were assessed in a con-trolled environment. Also, no long-term fol-low-up evaluations were conducted in thatstudy, so no preliminary conclusions could bedrawn concerning the long-term efficacy offluoxetine among this population of adoles-cents with comorbid conditions.

Recently, our own research team conductedan open-label trial of fluoxetine among adoles-cents with comorbid major depression and analcohol use disorder (alcohol abuse or alcoholdependence) (AUD) (Cornelius et al. 2001). Theacute phase (which took 12 weeks) results fromthat study demonstrated a significant within-group reduction of both depressive symptomsand alcohol use among the population of ado-lescents with comorbid conditions. To date,however, there have been no reports of anydouble-blind, placebo-controlled studies of flu-oxetine or any other SSRI medication amongcomorbid adolescents. Our own research grouprecently conducted a double-blind, placebo-controlled study involving adults with comor-bid major depression and alcohol dependence(Cornelius et al. 1997). The results of that studydemonstrated efficacy for fluoxetine for de-creasing both the depressive symptoms anddrinking among adults with comorbid majordepression and alcohol dependence. A 1-yearnaturalistic follow-up study was conducted in-volving the subjects who had participated inthat study. The results of that naturalistic 1-year follow-up study of fluoxetine in comorbidadults suggested continuing efficacy for fluox-etine for treating both the depressive symp-toms and the drinking of that adult comorbidpopulation (Cornelius et al. 2000). To date, how-ever, there have been no reports of any long-

term (1 year or longer) evaluations of the effi-cacy of fluoxetine or any other SSRI medica-tion among adolescents with comorbid majordepression and an alcohol use disorder. Con-sequently, at present, it is unclear whether flu-oxetine or any other SSRI medication is oflong-term benefit for decreasing the depressivesymptoms or the alcohol use of that comorbidpopulation of adolescents.

In this report, we present data from the first1-year naturalistic follow-up evaluation offluoxetine treatment of adolescents with co-morbid major depression and an alcohol usedisorder who had previously completed a con-trolled, acute phase, open-label trial of fluoxe-tine (Cornelius et al. 2001).

METHODS

This study was conducted at the PittsburghAdolescent Alcohol Research Center (which isfunded by the National Institute on AlcoholAbuse and Alcoholism), which is part of theUniversity of Pittsburgh. All patients were re-cruited from radio and newspaper advertise-ments. Before entry into the protocol, thestudy was explained, and written informedconsent was obtained after all procedures hadbeen fully detailed. The study was approvedby the University of Pittsburgh InstitutionalReview Board.

Thirteen patients participated in the 12-week acute phase of the study. Inclusion crite-ria for the acute phase included the following:(1) current Diagnostic and Statistical Manualof Mental Disorders, fourth edition (DSM-IV;American Psychiatric Association 1994), diag-noses of major depression and either alcoholdependence or alcohol abuse at baseline, fol-lowing any needed detoxification and a subse-quent 1-week washout period; (2) consumingat least 10 drinks in the month prior to enter-ing the study; and (3) a score of greater than 12at baseline on the 24-item Hamilton RatingScale for Depression (HAM-D-24). No onewho participated in the acute phase study wasexcluded from participation in the follow-upstudy.

During the acute phase of the study, sub-jects were treated using an open-label study

34 CORNELIUS ET AL.

design. All subjects were initially given one10-mg capsule of fluoxetine each morning forthe first 2 weeks, which was then increased to20 mg of fluoxetine each morning thereafter, ifno medication side effects were noted at the10-mg dose. All patients in the study toleratedthe 10-mg initial dose without any side effects,so all subjects were increased to the 20-mgdose of fluoxetine each morning. Verificationof compliance with medication was assessedby pill counts. All patients also receivedweekly supportive psychotherapy sessionsduring the first 4 weeks of the study and bi-weekly sessions thereafter, accompanied bymeetings with the attending psychiatrist(J.R.C.). All ratings for the study were con-ducted by experienced raters at the PittsburghAdolescent Alcohol Research Center. Addi-tional details about the acute phase studyhave been published elsewhere (Cornelius etal. 2001).

Ten of the 13 patients who had participatedin the acute phase study signed informed con-sent for the 1-year follow-up study. All 10 ofthe subjects who had given signed informedconsent for the 1-year follow-up study subse-quently participated in the 1-year follow-upevaluation. The other 3 subjects who had par-ticipated in the acute phase study chose not toparticipate in the follow-up study. All 3 ofthese subjects stated that they had decided notto participate in the follow-up assessment be-cause they were doing well and did not wantany further contact with treatment providersor research staff. Following completion of the12-week acute phase medication trial, the sub-jects were referred back to their local doctorsfor long-term naturalistic treatment by physi-cians and therapists unrelated to the study.These physicians were generally the ones whohad provided treatment to the subjects prior totheir participation in the acute phase study.The subjects did not receive protocol medica-tions during this time but instead receivedevaluation and treatment as deemed necessaryby their local treating physician. The 1-yearfollow-up evaluation was conducted during asingle follow-up interview 1 year from thebaseline of the acute phase study, which was 9months following the completion of the acutephase study.

The outcome measures used during the 1-year follow-up evaluation were the same asthose used during the acute phase of the study.Observer-rated depressive symptoms were as-sessed with the HAM-D-24 (Hamilton 1960).Drinking behavior was evaluated using thetimeline follow-back method (Sobell et al.1988). In addition, standardized diagnostic in-struments were used to assess the current di-agnoses of major depressive disorder andalcohol use disorder at the 1-year follow-upevaluation. Depressive diagnoses were evalu-ated using the Schedule for Affective Dis-orders and Schizophrenia for School-AgeChildren–Present and Lifetime Version (K-SADS-PL; Kaufman et al. 1997), and alcoholand drug diagnoses were determined usingthe Structured Clinical Interview for DSM-IV(SCID; Martin et al. 2000). Faculty from ouralcohol research center (Martin et al. 2000)have validated these diagnostic instrumentswith adolescent substance abuse populations.These same two diagnostic instruments (K-SADS-PL and SCID) were also administered atbaseline, so changes in diagnoses could becompared between the beginning and end ofthe study. On the K-SADS-PL a prior period ofeuthymia lasting at least 3 months was re-quired to define a recurrence of major depres-sion. Statistical analysis was performed usingpaired t tests. All tests were two-tailed.

RESULTS

All 10 of the subjects who had signed up toparticipate in the 1-year follow-up study at theend of the acute phase study actually com-pleted the 1-year follow-up assessment 9months later. Of these subjects, 2 were malesand 8 were females; 9 were Caucasian and 1was African American. They ranged in agefrom 15 to 20 years at the 1-year follow-upevaluation. Nine of the subjects had a diagno-sis of alcohol dependence, and 1 subject had adiagnosis of alcohol abuse. The mean durationof the index major depressive disorder episodeat the baseline of the acute phase study was11.7 months, with a range of 3–36 months. Themean duration of the index alcohol use disor-der at the baseline of the acute phase study

FLUOXETINE IN DEPRESSED ADOLESCENTS 35

was 15.6 months, with a range of 3–38 months.In four cases the major depressive disorder di-agnosis came before the alcohol use disorderdiagnosis; in four cases the alcohol use disor-der diagnosis came before the major depres-sive disorder diagnosis; and in 2 cases theonset of the major depressive disorder diagno-sis and the onset of the alcohol use disorder di-agnosis occurred in the same month.

Surprisingly, none of the 10 subjects hadcontinued taking antidepressant medicationmore than 2 months into the follow-up period.When asked why they had chosen to discon-tinue medication treatment, the subjects con-sistently stated that they felt substantially lessdepressed and that their drinking had gener-ally decreased, so they no longer felt the needto be treated with an antidepressant medica-tion. In addition, the subjects often stated thattheir family members and their treating physi-cians had not stressed continued treatmentwith medication, possibly because to date thelong-term effectiveness of this treatment hasnot been established. Also, 9 of the 10 subjectshad been in their first episode of major depres-sion upon entering the acute phase treatmentstudy, so they had not experienced a recur-rence of depression, and thus they may havebeen less concerned about a recurrence of de-pression upon discontinuing their medication.However, 6 of the 10 subjects displayed a cur-rent major depressive episode at the time ofthe 1-year follow-up evaluation, whereas only4 subjects did not. Three of those 6 subjectscontinued to demonstrate signs of their origi-nal major depression episode throughout thefollow-up period, as shown on the K-SADS-PL, whereas the other 3 subjects experienced arecurrence of their major depression duringthe follow-up period. These recurrences oc-curred at 2, 3, and 4 months after their SSRImedication had been discontinued. One ofthese three recurrences occurred in the onlysubject who had displayed recurrent major de-pression upon entry into the acute phasestudy. In contrast, only 4 subjects continued todemonstrate a current alcohol use disorder onthe SCID, whereas the other 6 subjects nolonger continued to demonstrate a current al-cohol use disorder, despite persistent low lev-els of drinking.

At the 1-year follow-up evaluation, the sub-jects continued to demonstrate less depressivesymptomatology than they had displayed atthe baseline of the acute phase study (HAM-D-24 score = 6.8 ± 5.8 vs. 25.1 ± 6.9; t = 10.022, df =9, p < 0.001). However, no significant change inlevel of depressive symptoms was noted be-tween the end of the 12-week acute phasetreatment study and the level of depressivesymptoms at the 1-year follow-up evaluation(6.8 ± 5.8 vs. 5.6 ± 4.9, ns). Thus, no significantchange in level of depressive symptoms wasnoted in the 10 subjects as a group during thecourse of the follow-up period; however, someincrease in depressive symptoms was noted inthe 3 subjects who had demonstrated a recur-rence of their major depression during the fol-low-up period.

The frequency of drinking was measuredusing the number of drinking days during thelast 30 days. At the 1-year follow-up evalua-tion, the subjects continued to demonstratefewer days of drinking during the last 30 daysthan they had demonstrated at the baseline ofthe acute phase study (6.3 days vs. 10.7 days;t = 2.28, df = 9, p = 0.048). However, no signifi-cant change in number of drinking days wasnoted between the end of the 12-week acutephase treatment study and the number notedat the 1-year follow-up evaluation (6.3 days vs.5.1 days, ns).

The quantity of drinking was measuredusing the number of drinks per drinking day.At the 1-year follow-up evaluation, the sub-jects continued to drink the same number ofdrinks per drinking day that had been noted atthe baseline of the acute treatment study (4.9drinks/drinking day vs. 4.6 drinks/drinkingday, ns).

DISCUSSION

This study provides data from what we be-lieve is the first 1-year naturalistic follow-upstudy evaluating the potential long-term effi-cacy of an acute phase trial of an SSRI pluspsychotherapy in the treatment of adoles-cents with comorbid major depression and analcohol use disorder. The results of this studyappear to demonstrate some residual benefit

36 CORNELIUS ET AL.

from the acute phase therapy with fluoxetineand psychotherapy, as shown by levels of de-pressive symptoms and frequency of drink-ing that continued to be significantly lowerthan they had been at baseline of the acutephase study. However, the subjects exhibiteda high rate of recurrent and persistent majordepression during the follow-up period. Inaddition, they continued to demonstrateheavy drinking when they did drink, with amean drinks per drinking day of about five,which indicates that most subjects drank todrunkenness when they did drink. Thus, thelong-term therapeutic effects of an acutephase trial of fluoxetine plus psychotherapywere limited.

The results of this study also demonstratethat this comorbid population of adolescentsis very likely to discontinue taking antide-pressant medication shortly after the worst ofthe depressive symptoms have abated. Thisfinding is very different from the pattern ofmedication taking that we have noticed in ourprevious work with comorbid adults. Specifi-cally, our previous 1-year follow-up studyinvolving adults with comorbid major depres-sion and alcohol dependence showed thatmost of the subjects continued to remain onantidepressant medication at the 1-year fol-low-up evaluation (Cornelius et al. 2000),whereas none of the comorbid adolescent sub-jects in our current study continued to taketheir medications beyond 2 months into thefollow-up period. It should be noted, how-ever, that the overwhelming majority of thecomorbid adults in our previous study werein a recurrent episode of major depressionwhen they were recruited into the treatmentstudy, whereas the adolescents in our currentstudy were virtually all (90%) in their first epi-sode of major depression when they firstentered the acute phase treatment study. Con-sequently, the possibility of a recurrence of de-pression may have been a more convincingrationale for continuing medication for theadults than for the adolescents. The low rateof subjects still taking antidepressant medica-tion at 1-year follow-up may also in part re-flect the reluctance to treat adolescents withpharmacological methods, as has been shownby Clark et al. (2002), which may in turn re-

flect the lack of studies evaluating or support-ing this treatment strategy.

Recurrent or persistent depression was a fre-quent problem among the subjects in our ado-lescent sample, which is consistent withEmslie et al.’s (1998) previous findings. Specif-ically, Emslie et al. (1998) reported that 40% oftheir adolescents with major depression (withno drug or alcohol use disorder) suffered a re-currence of major depression within 12months, which they concluded was a higherrate of recurrence of depression than is gener-ally noted among adults. Thus, it appears thatrecurrence or persistence of depression is com-mon among adolescents treated for major de-pression, regardless of the presence or absenceof comorbid alcohol use disorders upon initialpresentation for treatment. Indeed, the rate ofcurrent major depression noted among ourpopulation of adolescents at 1-year follow-up(60%) may even have been higher than thatnoted by Emslie et al.’s non-alcohol-abusingadolescents (40%). This finding suggests thatlonger term treatment for depression may bewarranted for at least some adolescents withcomorbid major depression and an alcohol usedisorder, despite their apparent reluctance tocontinue antidepressant medication beyondthe most severe phase of their original depres-sive episode.

Our current study is limited by its modestsample size and the naturalistic nature of itsstudy design. The findings of our currentstudy suggest that large, controlled, long-termstudies are warranted to assess the efficacy ofSSRI antidepressants in adolescents with co-morbid major depression and an alcohol usedisorder.

ACKNOWLEDGMENT

This research was supported in part bygrants from the National Institute on AlcoholAbuse and Alcoholism (R01 AA013370, R01AA13397, R01 AA11929, K02 AA00291, K08AA00280, K24 AA00301, P50 AA08746), fromthe National Institute on Drug Abuse (R01DA14635, P50 DA05605), and from the Veter-ans Administration (MIRECC to VISM 4, Starsand Stripes Network).

FLUOXETINE IN DEPRESSED ADOLESCENTS 37

REFERENCES

American Psychiatric Association: Diagnostic andStatistical Manual of Mental Disorders, 4th ed.(DSM-IV). Washington (DC), American Psychi-atric Association, 1994.

Bukstein OG, Glancy LJ, Kaminer Y: Patterns ofaffective comorbidity in a clinical population ofdually diagnosed adolescent substance abusers.J Am Acad Child Adolesc Psychiatry 31:1041–1045, 1992.

Clark DB, Bukstein O, Cornelius J: Alcohol use dis-orders in adolescents: Epidemiology, diagnosis,psychosocial interventions, and pharmacologicaltreatment. Paediatr Drugs 4:493–502, 2002.

Cornelius JR, Bukstein OG, Birmaher B, SalloumIM, Lynch K, Pollock NK, Gershon S, Clark D:Fluoxetine in adolescents with major depressionand an alcohol use disorder: An open-label trial.Addict Behav 26:735–739, 2001.

Cornelius JR, Salloum IM, Ehler JG, Jarrett PJ, Cor-nelius MD, Perel JM, Thase ME, Black A: Fluoxe-tine in depressed alcoholics: A double-blind,placebo-controlled trial. Arch Gen Psychiatry54:700–705, 1997.

Cornelius JR, Salloum IM, Haskett RF, Daley DC,Cornelius MD, Thase ME, Perel JM: Fluoxetineversus placebo in depressed alcoholics: A 1-yearfollow-up study. Addict Behav 25:307–310, 2000.

Emslie GJ, Rush AJ, Weinberg WA, Kowatch RA,Carmody T, Mayes TL: Fluoxetine in child andadolescent depression: Acute and maintenancetreatment. Depress Anxiety 7:32–39, 1998.

Hamilton M: A rating scale for depression. J NeurolNeurosurg Psychiatry 23:56–62, 1960.

Kaufman J, Birmaher B, Brent D, Rao U, Flynn C,Moreci P, Williamson D, Ryan N: Schedule for Af-fective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version(K-SADS-PL): Initial reliability and validity data.J Am Acad Child Adolesc Psychiatry 36:980–988,1997.

Martin CS, Pollock NK, Bukstein OG, Lynch KG:Inter-rater reliability of the SCID alcohol andsubstance use disorders section among adoles-cents. Drug Alcohol Depend 59:173–176, 2000.

Riggs PD, Mikulich SK, Coffman LM, Crowley TJ:Fluoxetine in drug-dependent delinquents withmajor depression: An open trial. J Child AdolescPsychopharmacol 7:87–95, 1997.

Sobell LC, Sobel MB, Leo GI, Cancilla A: Reliabilityof a timeline method: Assessing normal drinkers’reports of recent drinking and a comparativeevaluation across several populations. Br J Ad-dict 83:393–402, 1988.

Address reprint requests to:Jack R. Cornelius, M.D., M.P.H.

Western Psychiatric Institute and Clinic3811 O’Hara Street

Pittsburgh, PA 15213

E-mail: jcornel@pitt.edu

38 CORNELIUS ET AL.