Filer.case.edu

The Treatment for Adolescents
With Depression Study (TADS)
Long-term Effectiveness and Safety Outcomes
Context: The Treatment for Adolescents With Depres-
cal Global Impressions–Improvement score of much or very sion Study evaluates the effectiveness of fluoxetine hy- drochloride therapy, cognitive behavior therapy (CBT),and their combination in adolescents with major depres- Results: Intention-to-treat analyses on the Children’s De-
pression Rating Scale–Revised identified a significanttime ϫ treatment interaction (P Ͻ .001). Rates of re- Objective: To report effectiveness outcomes across 36
sponse were 73% for combination therapy, 62% for fluox- etine therapy, and 48% for CBT at week 12; 85% for com-bination therapy, 69% for fluoxetine therapy, and 65% Design and Setting: Randomized, controlled trial con-
for CBT at week 18; and 86% for combination therapy, ducted in 13 academic and community sites in the United 81% for fluoxetine therapy, and 81% for CBT at week 36.
States. Cognitive behavior and combination therapies were Suicidal ideation decreased with treatment, but less so not masked, whereas administration of placebo and fluox- with fluoxetine therapy than with combination therapy etine was double-blind through 12 weeks, after which or CBT. Suicidal events were more common in patients treatments were unblinded. Patients assigned to pla- receiving fluoxetine therapy (14.7%) than combination cebo were treated openly after week 12, and the placebo group is not included in these analyses by design.
Conclusions: In adolescents with moderate to severe de-
Participants: Three hundred twenty-seven patients aged
pression, treatment with fluoxetine alone or in combi- 12 to 17 years with a primary DSM-IV diagnosis of ma- nation with CBT accelerates the response. Adding CBT to medication enhances the safety of medication. Tak-ing benefits and harms into account, combined treat- Interventions: All treatments were administered per
ment appears superior to either monotherapy as a treat- ment for major depression in adolescents.
Main Outcome Measures: The primary dependent mea-
Trial Registration: clinicaltrials.gov Identifier:
sures rated blind to treatment status by an independent evaluator were the Children’s Depression Rating Scale–Revised total score and the response rate, defined as a Clini- Arch Gen Psychiatry. 2007;64(10):1132-1144 MAJORDEPRESSIVEDISOR- weeks)effectivenessofthefollowing3ac-
tive treatments for adolescents with MDD: cant morbidity and family burden1,2 as well etine therapy plus CBT) and, in the short cide.3,4 Hence, improvements in the treat- significant public health value. To this end, described the study aims, rationale, and de- sign5; the sample characteristics6; and the Author Affiliations: The TADS
Team authors and their
trolled trial that is intended to evaluate the (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
nitive behavior therapy alone was less effective than com- responders at the end of stage 1 were treated openly by the TADS bination therapy or fluoxetine therapy alone and was not team; placebo responders were offered telephone follow-up and, significantly more effective than placebo.
if they relapsed, their choice of open fluoxetine therapy, CBT, Almost 30% of the TADS participants exhibited clini- or combination therapy. Results past 12 weeks for placebo- cally significant suicidal ideation at baseline, with 2% re- treated patients will be reported separately.
Medication management in fluoxetine therapy and combi- porting high-intensity suicidal ideation. Suicidal ide- nation therapy included monitoring patient status and medi- ation decreased significantly during 12 weeks in all cation effects during 20- to 30-minute visits across 36 weeks treatment groups and showed the greatest improve- of treatment and general encouragement about the effective- ment with the combination of fluoxetine therapy and CBT.
ness of pharmacotherapy for MDD. Dosages of fluoxetine hy- Consistent with recent findings showing a slightly el- drochloride began at 10 mg/d, and were increased if necessary evated risk of treatment-emergent suicidality in pa- to 40 mg/d at week 6. At the week 12 visit, the dosage was in- tients receiving antidepressants,8 harm-related events were creased to a maximum of 60 mg/d in partial responders. In stages more common in patients receiving fluoxetine.7 In a sec- 2 and 3, the dosage of fluoxetine was held constant unless ad- ondary analysis that used the Columbia coding scheme verse events required a reduction, and the visit schedule was for suicidal events,9 24 suicidal events occurred during switched to maintenance visits every 6 weeks after week 18.
the 12-week short-term treatment period, including 5 of Cognitive behavior therapy in the TADS included fifteen 1-hoursessions during the first 12 weeks of treatment. Partial respond- 107 patients (4.7%) receiving combination therapy, 10 ers were then given 6 additional weeks of weekly CBT; full re- of 109 (9.2%) receiving fluoxetine therapy, 5 of 111 (4.5%) sponders were given biweekly CBT during the same interval receiving CBT, and 3 of 112 (2.7%) receiving placebo.
(stage 2). After week 18, CBT was provided every 6 weeks on Statistically, only fluoxetine therapy was associated with a maintenance visit schedule (stage 3). Combination therapy more suicidal events than placebo (P=.04; odds ratio [OR], consisted of all the components of the medication manage- 3.7; 95% confidence interval, 1.00-13.7).
ment and CBT protocols plus limited interaction between the Before study inception, we hypothesized that combi- CBT therapist and the pharmacotherapist.
nation therapy would show a larger, more rapid short- The TADS used the following 2 primary outcome mea- term treatment benefit than fluoxetine therapy or CBT.
sures assessed at baseline and weeks 6, 12, 18, 24, 30, and 36 We additionally hypothesized that the advantage for com- by an independent evaluator who was blind to the treatmentcondition: (1) the scalar Children’s Depression Rating Scale– bination therapy relative to fluoxetine therapy and CBT Revised (CDRS-R) total score19 and (2) responder status (de- would be evident across 9 months of randomized treat- fined as much or very much improved) on the CGI-I.20 Data ment.5 In a test of this hypothesis, we now report effec- from the adolescent self-report, the Reynolds Adolescent De- tiveness outcomes for combination therapy, fluoxetine pression Scale,21 was included because of the prominent place therapy, and CBT across the full 36 weeks of random- accorded adolescent self-report in the CBT literature.22 Psy- ized treatment. For ethical and feasibility reasons, pa- chometric properties and intercorrelations for all measures were tients assigned to placebo were treated openly after week 12, and the placebo group is not included in these analy- To ascertain suicidal ideation, we used a “flag” score of 31 or ses by design. Reflecting the paramount public health greater on the Suicidal Ideation Questionnaire–Junior High School importance of suicidality,10 we also report rates of clini- Version (SIQ-Jr),23 which denotes suicidal ideation of sufficientseverity to warrant prompt clinical evaluation. To identify pro- cally significant suicidal ideation and treatment- spectively assessed treatment-emergent suicidal events, we used the Columbia University classification scheme of the US Foodand Drug Administration analyses of antidepressant-associated suicidal events.8 In the Columbia coding scheme, classificationas a suicidal event requires a suicide attempt, preparatory ac-tion toward suicidal behavior, or suicidal ideation. Self-harm with- The TADS methods have been extensively documented in pre- out suicidal ideation or intent, such as cutting or other forms of vious publications,5-7,11-18 and only those aspects of the study self-mutilation, are not included in this definition. All suicidal that are directly relevant to the weeks 0 to 36 analyses will be events met the TADS adverse event reporting threshold of func- presented herein. All participants and at least 1 parent for each tional impairment or seeking medical attention. Coding of treat- participant provided informed consent/assent. The institu- ment-emergent suicidal events was completed with the evalua- tional review board at each site approved and monitored the tor blinded to treatment assignment and course using the protocol. Safety monitoring was performed quarterly by the Columbia Suicidality Classification Group.
NIMH Data Safety and Monitoring Board.
The primary effectiveness and safety analyses were con- Participants and all study staff remained masked in the “pills- ducted using an intention-to-treat (ITT) principle in which the only” conditions (fluoxetine therapy and placebo) until the end analysis included all patients randomized to treatment regard- of stage 1 (week 12). Patients and treatment providers in the less of adherence to study treatment or procedures. To mini- combination and CBT conditions were aware of treatment as- mize confounding, we also conducted supplementary ob- signment. Cognitive behavior therapy, fluoxetine therapy, and served cases (OCs) analyses that included only those data combination therapy responders or partial responders (mini- elements for which the patient was still in his or her assigned mally improved or better on a clinician-assigned Clinical Global treatment arm at the time of the assessment. A teenager was Impressions–Improvement [CGI-I] score) entered stage 2, a considered an observed case if any of the following did not oc- 6-week maintenance/consolidation phase, followed by stage 3, cur before the assessment or event date: investigator-initiated an 18-week maintenance phase. For ethical reasons, to main- provision of out-of-protocol treatment administered under the tain the sample, and to protect the ecological validity of the fluox- adjunctive services and attrition prevention provisions of the etine therapy condition during long-term treatment, placebo protocol (termed premature termination), study dropout, or dis- and fluoxetine treatments were unmasked at the end of stage continuation of assigned treatment based on treatment nonre- 1. Placebo-treated patients who were deemed partial or non- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Statistical analyses on the primary scalar outcome mea- are non-Hispanic white; 11.3% are African American; and sures were conducted using random coefficients regression (RR) 9.8% are Hispanic. With a range of mild to severe de- models.24 Specifically, the influence of treatment on outcome pression (CDRS-R total raw score, 45-98), the mean (SD) was modeled as a function of fixed effects for treatment (with CDRS-R raw score at entry was 59.8 (10.3). Translated time defined as the natural logarithm of days since random- to a normed T score (standardized to a mean of 50 and ization) and clinical site (and their 2- and 3-way interaction SD of 10) of 75.3 (6.5), on average the TADS sample shows terms), as well as the random effects for patient andpatientϫtime interactions. The final model included linear and moderate to moderately severe MDD. This level of de- quadratic time effects and their significant interactions. Site was pression is consistent with mean (SD) Clinical Global Im- retained but its interactions were omitted from the final model pressions Severity Scale20 and Clinical Global Assess- owing to statistical nonsignificance. Generalized estimating equa- ment Scale scores of 4.7 (0.8) and 49.8 (7.4), respectively.
tions (GEEs) for binary outcomes were used to compare the This was the first episode of depression for 86.5% of the probability of treatment response over time in the 3 treatment TADS patients, with a mean (SD) duration of 75.1 (86.7) arms. The model included treatment, time, treatment ϫtime, weeks. On the CDRS-R suicide item, 28.1% of patients and site. The time effect was linear and, thus, it was not nec- were defined as having at least minimal suicidal ide- essary to include a quadratic term in the GEE model. For this ation (CDRS-R item 13 score of Ն2), with 1.2% endors- analysis, the original 13 sites were collapsed into 10 sites to im- ing severe depression (CDRS-R item 13 score of Ն6). In- prove the stability of the model (low enrolling sites were com-bined into a single site). The RR and GEE models are tolerant cluding dysthymia, more than half of the sample (52.0%) to data missing at random in the dependent variable; accord- was comorbid for at least 1 other psychiatric disorder.
ingly, no imputation methods were used.
Forty-one of 327 patients (12.5%) met DSM-IV criteria General linear models and tests for differences in propor- for attention-deficit/hyperactivity disorder and, of these, tions (␹2 tests) were performed to evaluate differences across 9 of 41 (22.0%) were taking a psychostimulant at study treatment groups at baseline. The rate of clinically significant entry. The modal family income was $50 000 to $74 000, suicidality on the SIQ-Jr and of treatment-emergent suicidal with a range of less than $5000 to more than $200 000; events in each treatment arm was compared using ␹2 and Fisher 41.7% lived in a single-parent home; and 26.7% had been exact tests. For treatment-emergent suicidal events, Wald ORs suspended or expelled from school. No statistically sig- and 95% confidence intervals were calculated to provide an in- nificant differences among the 3 treatment groups on any dicator of the risk of active treatments relative to each other.
For hypotheses stipulated in the statistical plan for the 2 pri- of the specified baseline characteristics were noted.
mary outcomes, the nominal significance level was set a prioriat a 2-tailed type I error rate of .05 for the omnibus tests de- PATIENT DISPOSITION
signed to compare all 3 treatment arms. If the treatment or atreatmentϫtime (linear or quadratic) interaction term was sig- As shown in Figure 1, 2804 patients were screened by
nificant, then pairwise comparisons were conducted using a telephone (gate A). Of these, 1088 signed consent for evalu- closed test procedure with an ␣ of .05 for each test. In the event ation of inclusion and exclusion criteria (gate B) and 439 of a nonsignificant omnibus result, a sequential rejective ap- completed the baseline assessment and were subse- proach25 was planned (but not needed) in all but the safety analy- quently randomized to treatment (gate C). This report con- ses for which the samples sizes were deemed too small to war- cerns only the 327 patients randomized to 1 of the 3 ac- To evaluate the magnitude of the influence of combination tive treatment conditions: combination therapy (n=107), therapy and fluoxetine therapy relative to CBT, we calculated fluoxetine therapy (n=109), and CBT (n=111). Of these, the effect size and the number needed to treat (NNT) for the 270 (82.6%) continued in their assigned randomized treat- primary scalar and binary outcomes, respectively. Effect sizes ment across the 12-week short-term treatment period. By (Hedges g) were calculated as ME−MC/SDpooled, where ME rep- the end of stage 1, 23 of 327 patients (7.0%) prematurely resents the adjusted mean of experimental treatment, MC rep- terminated, 34 of 327 (10.4%) exited owing to with- resents the adjusted mean of the comparison treatment, and drawal of consent or loss to follow-up, and 28 of 327 SDpooled represents pooling of the standard deviations from within (8.6%) were referred to open treatment owing to nonre- both groups.26 The NNT was calculated according to methods sponse at the end of stage 1. Two hundred forty-three of 327 patients (74.3%) remained in the study at week 36.
Analyses were conducted using SAS statistical software (ver- sion 8.2; SAS Institute, Cary, North Carolina), with PROC Of the 327 patients, 178 (54.4%) remained in the treat- MIXED used for RR and PROC GENMOD used for GEE and ment condition to which they initially had been random- ized (n=68 for combination therapy, n=55 for fluoxetinetherapy, and n=55 for CBT). Eighty-four of 327 (25.7%)exited the study because of loss to follow-up or with- drawal of consent (n=21 for combination therapy [19.6%],n=32 for fluoxetine therapy [29.4%], and n=31 for CBT PATIENT CHARACTERISTICS
[27.9%]). Ninety-six of 327 patients (29.4%) discontin-ued their randomized treatment before week 36 owing to The full TADS sample (N = 439) has been extensively de- premature termination or nonresponse at the end of stage scribed and compared with clinical and epidemiological 1 (n=25 for combination therapy [23.4%], n=39 for fluox- etine therapy [35.8%], and n=32 for CBT [28.8%]), with The clinical and demographic characteristics of the 327 only combination therapy vs fluoxetine therapy proving adolescents assigned to an active treatment arm (fluox- statistically significant (P=.046; OR, 1.8; 95% confidence etine therapy, CBT, or combination therapy) follow. The interval, 1.0-3.3). Among the 149 patients who discon- mean (SD) age was 14.6 (1.5) years; 45.0% are male; 74.0% tinued randomized treatment across the 36 weeks, the pri- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
treatment owing to nonresponse (as per protocol) Figure 1. Flow diagram for the Treatment for Adolescents With Depression Study. Reasons for discontinuation before randomization have been previously
reported.7 All 112 patients assigned to the placebo condition discontinued randomized treatment at the end of stage 1 as per protocol and are not included in the
current randomization through week 36 analysis. Observed cases (OCs) are cases in a randomized treatment arm at the assessment point. CBT indicates cognitive
behavior therapy; DC, discontinuation of treatment owing to premature termination, nonresponse at the end of stage 1, or study exit; Exit, exit from the study
owing to withdrawal of consent or loss to follow-up.
mary reason for termination was clinician-initiated dis- 6, 12, 18, 24, 30, and 36 identified early superiority for continuation of treatment (n=96 [64.4%]) or departure combination therapy and fluoxetine therapy relative to CBT from the study without prior termination of treatment at 6 and 12 weeks, whereas combination therapy and fluox- etine therapy did not separate, thereby replicating previ-ously published results.7 Combination therapy and fluox- BENEFITS
etine therapy retained superiority relative to CBT at week18, and combination therapy remained superior to CBT Table 1 presents the ITT and OC adjusted means and
at week 24; all 3 treatments converged at weeks 30 and standard deviations for the CDRS-R and Reynolds Ado- 36. The OC analyses matched the ITT results, with planned lescent Depression Scale, as well as the adjusted CGI-I contrasts showing that fluoxetine therapy retained supe- response probabilities broken out for each treatment group riority over CBT to week 24 and combination therapy re- by assessment point. Table 2 presents planned con-
tained superiority over CBT to week 30.
trasts by treatment week for each treatment arm. Figure 2
With a positive response defined as a CGI-I of 1 (very graphs the ITT and OC trajectories on the CDRS-R for much improved) or 2 (much improved) as the depen- the 3 active treatments across 36 weeks.
dent variable, GEE analyses indicated that, for the ITT cases, Intention-to-treat RR analyses on CDRS-R total score the effect of treatment was statistically significant (␹22=8.8; across time identified a statistically significant linear time P = . 0 1 ) , a s w a s t i m e ( ␹25 = 72.0; P Ͻ .001) and effect (F1,296=46.2; P Ͻ.001), timeϫtreatment interac- treatmentϫtime (␹210=19.6; P=.03); the effect of site was tion (F2,296 = 15.4; P Ͻ .001), quadratic time effect nonsignificant (␹29=10.7; P=.30). Adjusted response rates (F1,287=12.8; PϽ.001), and quadratic timeϫtreatment in- (adjusted for site) at 12 weeks were 73% for combination teraction (F2,287=13.0; PϽ.001). The fixed effect of site was therapy, 62% for fluoxetine therapy, and 48% for CBT.
significant (F12,311=2.0; P=.02). Planned contrasts at weeks These GEE estimated responses rates are similar to the pre- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Table 1. Adjusted Mean CDRS-R and RADS Total Score and CGI-I Response Probabilitiesa
Mean ± SD Score
Cases/Treatment
Baseline
Abbreviations: CBT, cognitive behavior therapy; CDRS-R, Children’s Depression Rating Scale–Revised; CGI-I, Clinical Global Impressions–Improvement; FLX, fluoxetine hydrochloride therapy; ITT, intention-to-treat; NA, not applicable; OCs, observed cases; RADS, Reynolds Adolescent Depression Scale.
a For CDRS-R, the total score was rated by blinded clinical evaluator; for CGI-I, estimated response probability was from ratings by a blinded clinical evaluator; and for RADS, the total score was rated by the patient. All means are adjusted for fixed and random effects as specified in the analytic models.
viously published week 12 response rates (71% for com- bination therapy maintained superiority relative to CBT bination therapy, 61% for fluoxetine therapy, and 43% for at weeks 24 and 30. The OC results matched the ITT re- CBT), in which the last-observation-carried-forward sults, with planned contrasts indicating that combina- method was used to impute missing data in a logistic re- tion therapy retained superiority relative to CBT to week gression analysis.7 Adjusted response rates at week 18 were 36, whereas CBT and fluoxetine therapy were not sta- 85% for combination therapy, 69% for fluoxetine therapy, tistically different from week 18 onward.
and 65% for CBT for the ITT cases. At week 18, combi- The magnitude and, hence, clinical significance of the nation therapy maintained statistical superiority relative influence of combination therapy and fluoxetine therapy to fluoxetine therapy and CBT, which did not differ from relative to CBT was evaluated by calculating effect sizes each other. By week 24, all 3 treatments converged and (Hedges g) using the CDRS-R and Reynolds Adolescent remained so to 36 weeks. Adjusted response rates at week Depression Scale adjusted scores and by calculating the 36 were 86% for combination therapy, 81% for fluox- NNT using adjusted response rates. An effect size of 0.2 etine therapy, and 81% for CBT. Observed cases analyses is small, 0.5 is moderate, and 0.8 is large; corresponding indicated that combination therapy maintained superior- values for the NNT are 10, 5, and 2. As shown in Table 3,
ity over fluoxetine therapy and CBT at week 18, whereas the magnitude of the treatment effect for combination fluoxetine therapy did not differ from CBT after week 12.
therapy and fluoxetine therapy relative to CBT de- Intention-to-treat RR analyses on longitudinal Rey- creased across time for the ITT and OC samples, with nolds Adolescent Depression Scale scores identified a sta- combination therapy but not fluoxetine therapy main- tistically significant linear time effect (F1,276=14.9; P=.001), taining a numerically significant, if at times clinically mod- time ϫtreatment interaction (F2,276= 9.9; PϽ.001), qua- est, advantage over CBT at all time points. The advan- dratic time effect (F1,265= 10.4; P = .001), and quadratic tage of combination therapy over fluoxetine therapy time ϫtreatment interaction (F2,265=6.7; P=.002). The relative to CBT is most evident on adolescent self-report fixed effect of site was significant (F12,314=3.0; P=.003).
As shown in Table 2, planned contrasts at weeks 6, 12,18, 24, 30, and 36 identified early superiority for com- SUICIDAL IDEATION AND EVENTS
bination therapy and fluoxetine therapy relative to CBTat 6 and 12 weeks, thereby replicating previously pub- With respect to clinically significant suicidal ideation, 97 lished results.7 Combination therapy and fluoxetine of 320 patients with an SIQ-Jr (30.3%) met the SIQ-Jr therapy did not differ. Fluoxetine therapy retained sta- suicidality flag criterion at baseline, including 42 of 106 tistical superiority relative to CBT through week 18. Com- (39.6%) for combination therapy, 28 of 107 (26.2%) for (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Table 2. Planned Contrasts for the CDRS-R and RADS Total Scores and CGI-I Response Probabilities
Contrast
F
df
P Value
F
df
P Value
Contrast
␹2 Test
P Value
␹2 Test
P Value
Abbreviations: CBT, cognitive behavior therapy; CDRS-R, Children’s Depression Rating Scale–Revised; CGI-I, Clinical Global Impressions–Improvement; CGI-I TR, CGI-I predicted probabilities of treatment response; FLX, fluoxetine hydrochloride therapy; ITT, intention to treat; NA, not applicable; OCs, observedcases; OR, odds ratio; RADS, Reynolds Adolescent Depression Scale.
(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Figure 2. Adjusted mean Children’s Depression Rating Scale–Revised (CDRS-R) total scores. A, intention-to-treat (ITT) analysis. B, and observed case (OC) (B)
analyses. CBT indicates cognitive behavior therapy.
Table 3. Magnitude of the Effect of Combination Therapy and Fluoxetine Relative to CBT
ITT Population
OC Population
Effect Size, Hedges g
Effect Size, Hedges g
Contrast
CGI-I, NNT
CGI-I, NNT
Abbreviations: CBT, cognitive behavior therapy; CDRS-R, Children’s Depression Rating Scale–Revised; CGI-I, Clinical Global Impressions–Improvement; FLX, fluoxetine hydrochloride therapy; ITT, intention-to-treat; NNT, number needed to treat; OCs, observed cases; RADS, Reynolds Adolescent Depression Scale.
fluoxetine therapy, and 27 of 107 (25.2%) for CBT. Pair- of 63 for combination therapy, 8 of 55 (14.5%) for fluox- wise comparisons indicated that combination therapy had etine therapy, and 2 of 53 (3.8%) for CBT. Fluoxetine an excess of suicidal ideation at baseline relative to fluox- therapy showed higher rates of suicidal risk compared etine therapy (P = .04) and CBT (P = .02). At week 12, 31 with CBT (P = .04) and combination therapy (P = .01), of 278 patients who completed the questionnaire met the whereas combination therapy and CBT did not differ.
SIQ-Jr flag criterion (11.2%), including 8 of 90 (8.9%) Among the observed cases, fluoxetine therapy showed for combination therapy, 18 of 97 (18.6%) for fluox- elevated rates compared with combination therapy etine therapy, and 5 of 91 (5.5%) for CBT. For the ob- (P=.002) and a trend toward elevated rates compared with served cases, 24 of 257 (9.3%) met the SIQ-Jr flag crite- CBT (P=.09), whereas combination therapy and CBT did rion at week 12, including 5 of 84 (6.0%) for combination therapy, 14 of 89 (15.7%) for fluoxetine therapy, and 5 Table 4 presents rates and Table 5 presents planned
of 84 (6.0%) for CBT. Thus, by week 12, patients treated contrasts and ORs for treatment-emergent suicidal events with fluoxetine continued to show more clinically sig- by treatment group for the ITT and OC samples. During nificant suicidal ideation than those treated with CBT stage 1, 6.7% of ITT and 6.1% of OC patients had a sui- (PϽ.01) or, as a trend, with combination therapy (P=.06).
cidal event. From 0 to 36 weeks of treatment, 9.8% of Among the observed cases, fluoxetine therapy was sig- ITT and 8.0% of OC patients experienced a suicidal event.
nificantly different from CBT (P = .04) and combination Of these, 22 of 32 ITT events (69%) and 20 of 26 OC therapy (P = .04), which did not differ. At week 36, 15 of events (77%) occurred during the first 12 weeks of treat- 228 patients (6.6%) who completed an SIQ-Jr met the ment. Across 12 weeks of treatment, suicidal events were SIQ-Jr flag criterion, including 2 of 79 (2.5%) for com- more common in patients treated with fluoxetine alone bination therapy, 10 of 73 (13.7%) for fluoxetine therapy, in the ITT sample (11.0%, compared with 4.7% for com- and 3 of 76 (3.9%) for CBT. For the observed cases, 10 bination therapy and 4.5% for CBT) and in the OC sample of 171 (5.8%) met the SIQ-Jr flag criterion, including 0 (9.2%, compared with 4.7% for combination therapy and (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
4.5% for CBT). Likewise, across 36 weeks of treatment,suicidal events were more common in patients treated Table 4. Patients With Suicidal Eventsa
with fluoxetine alone in the ITT sample (14.7%, com-pared with 8.4% for combination therapy and 6.3% for No. (%) of Patients
CBT) and in the OC sample (11.0%, compared with 7.5% for combination therapy and 5.4% for CBT). Although Treatment
Patients
Population
Population
only the 0- to 36-week fluoxetine therapy vs CBT con- trast proved statistically significant (OR, 2.6; P=.04), ORs at 12 and 36 weeks indicate approximately twice the risk of a treatment-emergent suicidal event in patients treated with fluoxetine therapy than in patients treated with CBT or combination therapy. There were no completed sui- Abbreviations: CBT, cognitive behavior therapy; FLX, fluoxetine Relative to our previous report at 12 weeks,7 data on the hydrochloride therapy; ITT, intention-to-treat; OCs, observed cases.
benefits and harms during 36 weeks of randomized treat- Suicidal events include suicide attempt, preparatory action toward suicidal behavior, or suicidal ideation as adjudicated by the Columbia ment provide important new information. First, clini- Suicidality Classification Group. Youths with multiple suicide-related events cally meaningful improvement occurred in all 3 active are counted only once and the most severe code is represented.
treatment conditions, with convergence on most endpoints by week 36. Second, treatment with fluoxetinealone or in combination with CBT produced more rapid applicable to those obviously ill depressed adolescents improvement in MDD symptoms than did CBT alone.
who few would disagree ought to receive evaluation and Third, although rarely statistically superior, combina- treatment as a part of routine clinical practice.28 tion therapy proved numerically superior to fluoxetinetherapy on most end points, with the advantage for com- COMBINATION THERAPY
bination therapy most apparent on the OC analyses andon adolescent self-report. Fourth, CBT catches up to fluox- Overall, combination therapy in the TADS proved ro- etine therapy at the midpoint of treatment and to com- bustly superior to CBT and modestly and inconsistently bination therapy toward the end of treatment. Fifth, clini- better than fluoxetine therapy. Secondary analyses of out- cally significant suicidal ideation persists in a minority comes at 12 weeks show that the combination of CBT of patients and is significantly more common in pa- and fluoxetine but not fluoxetine alone proved superior tients treated with fluoxetine alone than with combina- to CBT and placebo with respect to probability of remis- tion therapy or CBT. Sixth, despite state-of-the-art sion,29 function and quality of life,30 and multiple mea- treatment and notable improvement in depression, sures of acceptability, tolerability, and safety, including treatment-emergent suicidal events occurred in 10% of suicidality.9 Of the 16 possible week 12 end points ex- TADS patients during 9 months of treatment, with amined so far, combination therapy proved superior to most of these occurring early in treatment. Patients CBT and placebo on greater than 90% of the week 12 end treated with fluoxetine alone were twice as likely as pa- points and to fluoxetine therapy half the time, whereas tients treated with combination therapy or CBT to ex- fluoxetine therapy was superior to CBT and placebo in perience a suicidal event, indicating that CBT may just under half.31 Although the reasons for the advan- protect against treatment-emergent suicidal events in tage of combination therapy relative to fluoxetine therapy patients taking fluoxetine. After taking benefit and risk are unclear, it is clear that the TADS was underpowered into account, we conclude that the combination of to detect the identified 10% difference between combi- fluoxetine and CBT appears superior to either mono- nation therapy and fluoxetine therapy at week 12 or 18 therapy as a long-term treatment strategy for MDD in on the primary categorical outcome. This difference (NNT = 10), although at the margin of clinical detec-tion, is nonetheless of considerable public health rel- GENERALIZABILITY
Given the difference in the sampling frame and CBT Although the TADS sample spans the range from mild intervention strategies, it is difficult to directly compare to severe depression, most patients (97%) fell in the mod- the TADS findings with those of the few other trials in erate to severe range of illness as characterized by a mean depressed youth that included a combined treatment con- CDRS-R score 2.5 SD above the mean (98th percentile), dition. One methodologically flawed and severely un- high rates of comorbidity, and a strikingly prolonged me- derpowered study by Melvin and colleagues32 showed no dian (42 weeks) and mean (75 weeks) current episode advantage for combined treatment over medication or duration. Given a mix of younger and older teenagers of CBT, but it did show an advantage for CBT over sertra- both sexes, minority representation roughly proportion- line hydrochloride. Another study comparing CBT com- ate to US population values, and wide variability in so- bined with medication vs medication alone in a mildly cioeconomic circumstances,6 the results should be broadly ill population showed a weak and inconsistent CBT (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Table 5. ORs and Treatment Contrasts for Suicidal Events
ITT Population
OC Population
Planned Contrast
␹2 Test
P Valuea
OR (95% CI)
␹2 Test
P Valuea
OR (95% CI)
Abbreviations: CBT, cognitive behavior therapy; CI, confidence interval; FLX, fluoxetine hydrochloride therapy; ITT, intention-to-treat; OCs, observed cases; OR, a Unless otherwise indicated, P values were calculated using 2 ϫ 2 ␹2 test results.
b Fisher exact test was used to calculate the P value owing to 25% of the cells having expected counts less than 5.
effect.33 A report from the randomized controlled trial most to 24 and to combination therapy by weeks 30 to 36. Thus, like the TADS—the British Adolescent Depression Anti- although some have speculated that CBT was poorly con- depressant and Psychotherapy Trial—is slated for pub- structed or implemented,38 the long-term outcomes pre- lication in late 2007. By intention, the Adolescent De- sented herein make it more likely that the severity of ill- pression Antidepressant and Psychotherapy Trial and the ness simply delayed the onset of benefit for CBT relative TADS both include the Health of the Nation Outcome to combination therapy and fluoxetine therapy, a finding Scales for Children and Adolescents. At 12 weeks, the that is widely acknowledged in adult studies of CBT rela- TADS showed a robust advantage for combined treat- tive to medication.40 Given that a recent meta-analysis of ment on this measure34; combined data set analyses, which the CBT literature for depressed youth showed a rela- were built into the structure of both trials, should am- tively small effect size (0.34),41 efforts to understand the plify and extend the results of the individual studies. Fu- mechanisms by which CBT alters the course of MDD and ture analyses of the TADS also should shed light on who in turn to revise CBT protocols so as to increase their effect is likely to respond to combined treatment or to 1 of the 2 monotherapies and whether adherence to treatment me-diates these differences. These future analyses are ex- FLUOXETINE THERAPY
pected to generate hypotheses that will inform future stud-ies in this area.
In contrast to the unenthusiastic view of medications heldby some,42,43 we and others7,28,44,45 concluded from the WHAT IS THE ROLE OF CBT?
short-term treatment data that the benefits of includingfluoxetine (we make no claims about other antidepres- In short-term studies of CBT, one-third of depressed youth sants) in the short-term treatment of adolescent MDD were do not respond.35,36 In the TADS CBT condition, which readily apparent and clinically meaningful. In second- blended the best available elements of cognitive, behav- ary analyses at 12 weeks, we have now shown that fluox- ioral, and family treatments,11,15,16 more than half of the etine alone or in combination with CBT accelerates re- patients did not respond at 12 weeks, which represents a sponse relative to CBT and to placebo.46 Thus, when substantially larger proportion of patients with an inad- considered in light of 2 other positive randomized con- equate response compared with other CBT trials and well trolled trials for fluoxetine therapy in pediatric MDD,47,48 below the hypothesized 60% response rate. In the CBT lit- short- and long-term data from the TADS unequivo- erature, it is conventional to argue that treatment should cally confirm that fluoxetine is an effective treatment for continue longer than 12 weeks to maximize the probabil- adolescents with moderate to severe MDD.
ity of a response and to minimize the possibility of a re-lapse.37 Given that the TADS patients also were more se- SUICIDALITY
verely ill than most patients in published CBT trials,6 12weeks of CBT may not have been long enough to allow Although the absolute magnitude of the medication- CBT to separate from placebo in the short term.5,38 In sup- attributable risk is low (number needed to harm, 50), treat- port of this proposition, which replicates and extends pre- ment-emergent suicidal events are an important public vious results in the CBT literature where 16 to 20 weeks health concern in children and adolescents treated with of treatment is the norm,39 the response rate for CBT im- antidepressants.8,10 Findings from the TADS across 9 proved from 48% to 65% at week 18, whereas the re- months of treatment confirm and extend results at 12 sponse rate for fluoxetine therapy moved from 62% to 69%.
weeks7 showing that patients treated with fluoxetine alone Although causal interpretations are limited by the ab- are more likely than patients treated with combination sence of a placebo control beyond week 12, the data show therapy or CBT to show clinically significant suicidal ide- that CBT caught up with fluoxetine therapy by weeks 18 ation (on the SIQ-Jr) and treatment-emergent suicidal (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
events (on the adverse event report). As an effectiveness taneously remitting.34 Thus, it is unlikely that improve- trial for which suicidal events were not a primary end point, ment within 36 weeks primarily reflected spontaneous the TADS was not structured properly to segment risk or remission. It is more likely that a reduction in relapse to establish the mechanism by which fluoxetine therapy coupled with gradual improvement in MDD symptoms may enhance risk or that CBT may exert a protective effect.
with continued treatment was responsible for the ob- The increased risk of a suicidal event with fluoxetine served overall improvement rate of approximately 85%.
therapy cannot simply be due to improvement because (1) In comparative treatment trials that include medica- combination therapy, which began with statistically higher tion and psychotherapy, double-blind administration of suicidal ideation, improved as much as or more than fluox- the medication therapy but not the psychosocial treat- etine therapy while patients receiving combination therapy ment conditions is possible. This is a design choice, not had a rate of suicidal events that approximated CBT alone a design flaw,43,54 which constrains the question to an ef- and (2) suicidality did not increase in the CBT group dur- fectiveness (primarily information for clinical decision ing the 12- to 24-week interval, when a large portion of makers) rather than efficacy (dismantling treatment ele- the CBT benefit occurred. Recalling that medication and ments) aim.5,7,31 In the absence of a CBT-placebo group CBT were administered in a coordinated fashion,12 the mean and a fluoxetine therapy–nonsupportive psychotherapy dose of fluoxetine hydrochloride in combination therapy comparison group, it is not possible to conclusively claim (28 mg) was lower than that for fluoxetine hydrochlo- the superiority of combination therapy relative to CBT ride alone (32 mg) at 12 weeks.7 Although there is no evi- or combination therapy relative to fluoxetine therapy, re- dence from the TADS to suggest that fluoxetine induces spectively. Similarly, in the absence of an active psycho- mania or “behavioral activation,” patients treated with com- therapy comparison group, it is not possible to claim that bination therapy show fewer psychiatric and nonpsychi- the benefits of TADS CBT are unique to CBT compared atric adverse events than patients treated with fluoxetine with a different type of psychosocial treatment. How- alone.9 Thus, it is at least possible if not likely that the added ever, 3 lines of evidence suggest that the advantage shown risk associated with fluoxetine monotherapy was dose re- by combined treatment is not artifactual. First, low ex- lated. More likely, CBT alone or in combination with fluox- pectation of benefit predicted a poorer outcome at 12 etine mitigates suicidality, perhaps by minimizing the prob- weeks; however, expectancy did not differentially mod- ability that ideation will lead to an attempt, decreasing erate outcomes,34 suggesting that expectancy alone can- the likelihood or improving the management of stressful not account for the advantage of combined treatment. Sec- psychosocial events, decreasing family conflict and ond, if expectancy effects accounted for all the advantage enhancing family problem solving ability, or providing skills of combination therapy over fluoxetine therapy, one might to manage negative affects, agitation, irritability, or dis- expect that unblinding the fluoxetine condition at 12 weeks should reduce the superiority of combinationtherapy over fluoxetine therapy at 18 weeks. Instead, the LIMITATIONS
numerical superiority of combination therapy over fluox-etine therapy increases once unblinding takes place, and We discuss design challenges and controversies in de- combination therapy remains superior to CBT longer than tail elsewhere5,31 and restrict our comments herein to 3 fluoxetine therapy does. Third, combined treatment is areas of concern: (1) the absence of a placebo group af- robust to differences in patients, therapists, and settings ter week 12, (2) expectancy effects, and (3) the influ- across all the outcomes examined so far,31,34 suggesting that rater bias is not playing a substantive role in the effect We readily acknowledge that it is impossible to con- clude that patients would not have reached equivalent Finally, patients assigned to combined treatment ex- week 36 outcomes simply because of the passage of time perienced somewhat greater contact time than did pa- without a placebo group or, better, an untreated control tients assigned to fluoxetine therapy or CBT alone. As group, both of which were considered unfeasible for ethi- pointed out earlier, the TADS did not include the proper cal and practical reasons.50 Thirty percent to 70% of youth control conditions to disentangle time and attention from with MDD recover during the first year of illness, al- the other active components of treatment. Although analy- though 30% subsequently relapse.51 Naturalistic fol- ses of the influence of treatment adherence on outcome low-up of randomized controlled trials of psycho- may help decompose the extra benefit associated with therapy52 and placebo-substitution trials of fluoxetine combined treatment, the TADS was never intended to ask therapy37 also show a high rate of relapse, even against a questions involving the mechanism of treatment ben- background of substantial improvement. In the only trial efit. Rather, it was designed to ask the simple question of maintenance psychotherapy in depressed teens, CBT of whether combined treatment was advantageous rela- booster sessions did not reduce the rate of recurrence but tive to CBT and fluoxetine therapy because they would did appear to accelerate recovery among participants who be delivered in clinical practice: the answer to this ques- were still depressed at the end of the short-term phase.53 The mean episode duration was longer than 1 year, andmore than half the TADS sample had received previoustreatment within the current episode.6 Secondary analy- ses at week 12 also showed that shorter episode dura-tion predicted a better outcome at 12 weeks, which is the The 2 central findings from the TADS are that (1) fluox- opposite of what would be predicted if episodes were spon- etine alone or in combination with CBT accelerates im- (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
provement of depression relative to CBT alone and (2) Kolker, and Karyn Riedal (Carolinas Medical Center, adding CBT to fluoxetine therapy minimizes persistent Charlotte, North Carolina); Norah Feeny, Robert Find- suicidal ideation and treatment-emergent suicidal events.
ling, Sheridan Stull, and Susan Baab (Case Western Re- This leads us to make 4 key recommendations for health serve University, Cleveland, Ohio); Elizabeth B. Weller, care decision makers at all levels of the health care sys- Michele Robins, Ronald A. Weller, and Naushad Jessani tem. First, identification and provision of evidence- (The Children’s Hospital of Philadelphia, Philadelphia, based treatment to adolescents with moderate to severe Pennsylvania); Bruce Waslick (Baystate Health/Tufts Uni- depression would likely have a positive public health ben- versity, Springfield, Massachusetts), Michael Sweeney, and efit and should be encouraged. Second, because accel- Randi Dublin (Columbia University, New York, New erating symptom reduction by using medication is an im- York); John Walkup, Golda Ginsburg, Elizabeth Kas- portant clinical outcome in psychiatry, as it is in other telic, and Hyung Koo (The Johns Hopkins University, Bal- areas of medicine, use of fluoxetine should be made widely timore, Maryland); Christopher Kratochvil, Diane May, available, not discouraged. Third, given equivalent if de- Randy LaGrone, and Brigette Vaughan (University of Ne- layed results for CBT monotherapy and increased pro- braska, Omaha); Anne Marie Albano (Columbia Univer- tection from suicidality, CBT or other evidence-based psy- sity), Glenn S. Hirsch and Elizabeth Podniesinki (New chosocial treatments55 should be made readily available York University, New York); Mark Reinecke, Bennett Lev- as part of comprehensive treatment for depressed ado- enthal, Gregory Rogers, and Rachel Jacobs (University lescents. Because most adolescents with depression re- of Chicago, Chicago, Illinois, and Northwestern Univer- ceive no or unproved psychosocial treatments, which sity, Evanston, Illinois); Sanjeev Pathak, Jennifer Wells, oddly enough are sometimes sanctioned by expert guide- Sarah Arszman, and Arman Danielyan (Cincinnati Chil- lines,56 conforming to this recommendation will require dren’s Hospital Medical Center, Cincinnati, Ohio); Paul a significant shift in current practice.57 Until this oc- Rohde, Anne Simons, James Grimm, and Stephenie Frank curs, fluoxetine monotherapy, delivered in the context (University of Oregon, Eugene); Graham Emslie, Beth of regular clinical management and careful clinical moni- Kennard, Carroll Hughes, and Taryn L. Mayes (The Uni- toring, will remain an important stop-gap measure in pa- versity of Texas Southwestern Medical Center, Dallas); tients for whom the earliest possible response is deemed David Rosenberg, Nili Benazon, Michael Butkus, and Marla clinically meaningful. Fourth, although not yet at the level Bartoi (Wayne State University, Detroit, Michigan); and of evidence necessary to mandate a standard of care, we Kelly Posner, for the Columbia University Suicidality Clas- believe that the literature on evidence-based treatment sification Group. James Rochon (Duke Clinical Re- for adolescent MDD has reached sufficient maturity that search Institute, Durham) is statistical consultant.
it should fall within the realm of informed consent.50 Financial Disclosure: Dr Findling has received research
Future reports from the TADS will address second- support, acted as a consultant, and/or served on a speak- ary outcomes; moderators and mediators, including ad- er’s bureau for Abbott Laboratories, AstraZeneca Pharma- herence to treatment; remission, relapse, and recovery ceuticals, Bristol-Myers Squibb, Celltech-Medeva, Eli Lilly during randomized treatment and during a year-long natu- and Company, Forest Laboratories, GlaxoSmithKline, ralistic follow-up period; the nature and use of ancillary Johnson & Johnson, New River, Novartis, Otsuka, Pfizer, treatments; and short- and long-term cost-effectiveness, Sanofi-Aventis, Shire Pharmaceuticals, Solvay, and Wy- among other topics of interest to decision makers at all eth Research. Dr Posner has received funding from the US levels of the health care system. Although TADS manu- Food and Drug Administration (FDA) to develop the sui- als and procedures were designed to be clinically appli- cidality classification system used in their antidepressant cable, additional research will be necessary to better un- safety analysis. Subsequently, as part of an effort to help derstand how to disseminate the TADS findings and to execute the FDA suicidality classification mandates, Dr Pos- explore topics, such as sequential treatment, not ad- ner has had research support from Abbott Laboratories, AstraZeneca Pharmaceuticals, Bristol-Myers Squibb, Cepha-lon, Eisai Inc, Forest Laboratories, GlaxoSmithKline, Submitted for Publication: August 28, 2006; final revi-
Johnson & Johnson, Novartis, Organon USA, Sanofi- sion received December 23, 2006; accepted December 26, Aventis, Shire Pharmaceuticals, UCB Pharma, and Wy- Correspondence: John S. March, MD, MPH, Depart-
Funding/Support: The TADS is supported by contract
ment of Psychiatry, Duke University Medical Center, Box RFP-NIH-NIMH 98-DS-0008 from the NIMH to Duke 3527, Durham, NC 27710 (jsmarch@duke.edu).
University Medical Center (principal investigator, John The TADS Team Authors: The TADS is coordinated by
S. March, MD, MPH). Eli Lilly and Company provided the Department of Psychiatry and Behavioral Sciences and fluoxetine and matching placebo under an independent the Duke Clinical Research Institute at Duke University educational grant to Duke University.
Medical Center, Durham, North Carolina, in collabora- Disclaimer: The opinions and assertions contained in this
tion with the NIMH. The Coordinating Center principal report are the private views of the authors and are not to collaborators are John S. March, Susan Silva, Stephen be construed as official or as reflecting the views of the Petrycki, John Curry, Karen Wells, John Fairbank, Barbara NIMH, the National Institutes of Health, or the Depart- Burns, Marisa Domino, and Steven McNulty. The NIMH ment of Health and Human Services. Eli Lilly and Com- principal collaborators are Benedetto Vitiello and Joanne pany had no role in the design or implementation of the Severe. Principal investigators and coinvestigators from study, data analysis, or in the writing of this manu- the clinical sites are as follows: Charles Casat, Jeanette (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
Additional Information: The TADS protocol and all of
14. Kratochvil CJ, Simons A, Vitiello B, Walkup J, Emslie G, Rosenberg D, March the TADS manuals are available on the Internet at https: JS. A multisite psychotherapy and medication trial for depressed adolescents:background and benefits. Cogn Behav Pract. 2005;12(2):159-165.
//trialweb.dcri.duke.edu/tads/index.html.
Additional Contributions: The NIMH Program Staff par-
15. Rogers GM, Reinecke MA, Curry JF. Case formulation in TADS CBT. Cogn Be- ticipated in the design and implementation of the TADS, hav Pract. 2005;12(2):198-208.
in analysis of the data, and in preparing this report. The 16. Rohde P, Feeny NC, Robins M. Characteristics and components of the TADS CBT TADS scientific advisors (Susan Essock, PhD, Mount Si- approach. Cogn Behav Pract. 2005;12(2):186-197.
nai School of Medicine, New York, New York; Barbara 17. Simons AD, Rohde P, Kennard BD, Robins M. Relapse and recurrence preven- tion in the Treatment for Adolescents With Depression Study. Cogn Behav Pract.
Geller, MD, Washington University, St Louis, Missouri; Joel Greenhouse, PhD, Carnegie Mellon University, Pitts- 18. Sweeney M, Robins M, Ruberu M, Jones J. African-American and Latino fami- burgh, Pennsylvania; Robert Johnson, MD, New Jersey lies in TADS: recruitment and treatment considerations. Cogn Behav Pract. 2005; Medical School, Newark; James Leckman, MD, Yale Uni- versity, New Haven, Connecticut; Lydia Lewis, Depres- 19. Poznanski E, Mokros H. Children’s Depression Rating Scale–Revised (CDRS-R). sion and Bipolar Support Alliance, Chicago, Illinois; Sue 20. Guy W. ECDEU Assessment Manual for Psychopharmacology. 2nd ed. Wash- M. Marcus, PhD, Mount Sinai School of Medicine; and ington, DC: Dept of Health Education and Welfare; 1976. DHEW publication (ABM) Kevin Stark, PhD, University of Texas at Austin) con- tributed to the design and methods of the study. David 21. Reynolds WM. Professional Manual for the Reynolds Adolescent Depression Scale. Brent, MD, and Greg Clarke, PhD, provided CBT con- Odessa, FL: Psychological Assessment Resources Inc; 1987.
22. Reinecke MA, Ryan NE, DuBois DL. Cognitive-behavioral therapy of depression sultation. The Columbia Suicidality Classification Group and depressive symptoms during adolescence: a review and meta-analysis.
led by Dr Posner, including Maria O. Qquendo, MD, J Am Acad Child Adolesc Psychiatry. 1998;37(1):26-34.
Madelyn Gould, PhD, MPH, and Barbara Stanley, PhD, 23. Reynolds WM. Professional Manual for the Suicidal Ideation Questionnaire. Odessa, were contracted to independently code suicidal events.
FL: Psychological Assessment Resources Inc; 1987.
The members of the NIMH Data and Safety Monitoring 24. Brown H, Prescott R. Applied Mixed Models in Medicine. New York, NY: John Board monitored the progress of the study.
25. Koch GG, Gansky SA. Statistical considerations for multiplicity in confirmatory protocols. Drug Inf J. 1996;30:523-534.
26. Rosenthal R, Rosnow R, Rubin DB. Contrasts and Effect Sizes in Behavioral Research. New York, NY: Cambridge University Press; 2000.
27. Sackett D, Richardson W, Rosenberg W, Haynes B. Evidence-Based Medicine. 1. Essau C, Dobson K. Epidemiology of depressive disorders. In: Essau C, Peter- 2nd ed. New York, NY: Churchill Livingston Inc; 2000.
mann F, eds. Depressive Disorders in Children and Adolescents: Epidemiology, 28. Jensen PS. NIMH’s TADS: more than just a tad of progress? Cogn Behav Pract.
Course and Treatment. Northvale, NJ: Jason Aronson, Inc; 1999:69-103.
2. Angold A, Messer SC, Stangl D, Farmer EM, Costello EJ, Burns BJ. Perceived 29. Kennard B, Silva S, Vitiello B, Curry J, Kratochvil C, Simons A, Hughes J, Feeny parental burden and service use for child and adolescent psychiatric disorders.
N, Weller E, Sweeney M, Reinecke M, Pathak S, Ginsburg G, Emslie G, March J.
Am J Public Health. 1998;88(1):75-80.
TADS Team. Remission and residual symptoms after short-term treatment in 3. Shaffer D, Gould MS, Fisher P, Trautman P, Moreau D, Kleinman M, Flory M.
the Treatment of Adolescents with Depression Study (TADS). J Am Acad Child Psychiatric diagnosis in child and adolescent suicide. Arch Gen Psychiatry. 1996; Adolesc Psychiatry. 2006;45(12):1404-1411.
30. Vitiello B, Rohde P, Silva S, Wells K, Casat C, Waslick B, Simons A, Reinecke M, 4. Gould MS, King R, Greenwald S, Fisher P, Schwab-Stone M, Kramer R, Flisher Weller E, Kratochvil C, Walkup J, Pathak S, Robins M, March J. TADS Team. Func- AJ, Goodman S, Canino G, Shaffer D. Psychopathology associated with suicidal tioning and quality of life in the Treatment for Adolescents With Depression Study ideation and attempts among children and adolescents. J Am Acad Child Ado- (TADS). J Am Acad Child Adolesc Psychiatry. 2006;45(12):1419-1426.
lesc Psychiatry. 1998;37(9):915-923.
31. March J, Silva S, Vitiello B. TADS Team. The Treatment for Adolescents With 5. Treatment for Adolescents With Depression Study Team. Treatment for Adoles- Depression Study (TADS): methods and message at 12 weeks. J Am Acad Child cents With Depression Study (TADS): rationale, design, and methods. J Am Acad Adolesc Psychiatry. 2006;45(12):1393-1403.
Child Adolesc Psychiatry. 2003;42(5):531-542.
32. Melvin GA, Tonge BJ, King NJ, Heyne D, Gordon MS, Klimkeit E. A comparison 6. Treatment for Adolescents With Depression Study Team. The Treatment for Ado- of cognitive-behavioral therapy, sertraline, and their combination for adolescent lescents With Depression Study (TADS): demographic and clinical characteristics.
depression. J Am Acad Child Adolesc Psychiatry. 2006;45(10):1151-1161.
J Am Acad Child Adolesc Psychiatry. 2005;44(1):28-40.
33. Clarke G, Debar L, Lynch F, Powell J, Gale J, O’Connor E, Ludman E, Bush T, 7. March J, Silva S, Petrycki S, Curry J, Wells K, Fairbank J, Burns B, Domino M, Lin EH, Von Korff M, Hertert S. A randomized effectiveness trial of brief McNulty S, Vitiello B, Severe J; Treatment for Adolescents With Depression Study cognitive-behavioral therapy for depressed adolescents receiving antidepres- (TADS) Team. Fluoxetine, cognitive-behavioral therapy, and their combination sant medication. J Am Acad Child Adolesc Psychiatry. 2005;44(9):888-898.
for adolescents with depression: Treatment for Adolescents With Depression Study 34. Curry J, Rohde P, Simons A, Silva S, Vitiello B, Kratochvil C, Reinecke M, Feeny (TADS) randomized controlled trial. JAMA. 2004;292(7):807-820.
N, Wells K, Pathak S, Weller E, Rosenberg D, Kennard B, Robins M, Ginsburg G, 8. Hammad TA, Laughren T, Racoosin J. Suicidality in pediatric patients treated with March J; TADS Team. Predictors and moderators of acute treatment outcome in antidepressant drugs. Arch Gen Psychiatry. 2006;63(3):332-339.
the Treatment for Adolescents With Depression Study (TADS). J Am Acad Child 9. Emslie G, Kratochvil C, Vitiello B, Silva S, Mayes T, McNulty S, Weller E, Waslick Adolesc Psychiatry. 2006;45(12):1427-1439.
B, Casat C, Walkup J, Pathak S, Rohde P, Posner K, March J; Columbia Suicidality 35. Compton SN, March JS, Brent D, Albano AM, Weersing R, Curry J. Cognitive- Classification Group. TADS Team. Treatment for Adolescents with Depression Study: behavioral psychotherapy for anxiety and depressive disorders in children and safety results. J Am Acad Child Adolesc Psychiatry. 2006;45(12):1440-1455.
adolescents: an evidence-based medicine review. J Am Acad Child Adolesc 10. Vitiello B, Swedo S. Antidepressant medications in children. N Engl J Med. 2004; Psychiatry. 2004;43(8):930-959.
36. Whittington CJ, Kendall T, Fonagy P, Cottrell D, Cotgrove A, Boddington E.
11. Curry JF, Wells KC. Striving for effectiveness in the treatment of adolescent de- Selective serotonin reuptake inhibitors in childhood depression: systematic re- pression: cognitive behavior therapy for multisite community intervention. Cogn view of published versus unpublished data. Lancet. 2004;363(9418):1341- Behav Pract. 2005;12(2):177-185.
12. Ginsburg GS, Albano AM, Findling RL, Kratochvil C, Walkup J. Integrating cog- 37. Emslie GJ, Mayes TL, Ruberu M. Continuation and maintenance therapy of early- nitive behavioral therapy and pharmacotherapy in the treatment of adolescent onset major depressive disorder. Paediatr Drugs. 2005;7(4):203-217.
depression. Cogn Behav Pract. 2005;12(2):252-262.
38. Hollon SD, Garber J, Shelton RC. Treatment of depression in adolescents with 13. Kennard BD, Ginsburg GS, Feeny NC, Sweeney M, Zagurski R. Implementation cognitive behavior therapy and medications: a commentary on the TADS project.
challenges to TADS cognitive-behavioral therapy. Cogn Behav Pract. 2005; Cogn Behav Pract. 2005;12(2):149-155.
39. Hollon SD, DeRubeis RJ, Shelton RC, Amsterdam JD, Salomon RM, O’Reardon (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.
JP, Lovett ML, Young PR, Haman KL, Freeman BB, Gallop R. Prevention of re- adolescents: a placebo-controlled, randomized clinical trial. J Am Acad Child Ado- lapse following cognitive therapy vs medications in moderate to severe depression.
lesc Psychiatry. 2002;41(10):1205-1215.
Arch Gen Psychiatry. 2005;62(4):417-422.
49. Negron R, Piacentini J, Graae F, Davies M, Shaffer D. Microanalysis of adoles- 40. Hollon SD, Jarrett RB, Nierenberg AA, Thase ME, Trivedi M, Rush AJ. Psycho- cent suicide attempters and ideators during the acute suicidal episode. J Am Acad therapy and medication in the treatment of adult and geriatric depression: which Child Adolesc Psychiatry. 1997;36(11):1512-1519.
monotherapy or combined treatment? J Clin Psychiatry. 2005;66(4):455-468.
50. March J, Kratochvil C, Clarke G, Beardslee W, Derivan A, Emslie G, Green EP, 41. Weisz JR, McCarty CA, Valeri SM. Effects of psychotherapy for depression in Heiligenstein J, Hinshaw S, Hoagwood K, Jensen P, Lavori P, Leonard H, McNulty children and adolescents: a meta-analysis. Psychol Bull. 2006;132(1):132- J, Michaels MA, Mossholder A, Osher T, Petti T, Prentice E, Vitiello B, Wells .
AACAP 2002 research forum: placebo and alternatives to placebo in random- 42. Jureidini JN, Doecke CJ, Mansfield PR, Haby MM, Menkes DB, Tonkin AL. Effi- ized controlled trials in pediatric psychopharmacology. J Am Acad Child Ado- cacy and safety of antidepressants for children and adolescents. BMJ. 2004; lesc Psychiatry. 2004;43(8):1046-1056.
51. Birmaher B, Arbelaez C, Brent D. Course and outcome of child and adolescent major 43. Jureidini J, Tonkin A, Mansfield PR. TADS study raises concerns. BMJ. 2004;329 depressive disorder. Child Adolesc Psychiatr Clin N Am. 2002;11(3):619-637, x.
52. Birmaher B, Brent DA, Kolko D, Baugher M, Bridge J, Holder D, Iyengar S, Ulloa 44. Glass RM. Treatment of adolescents with major depression: contributions of a RE. Clinical outcome after short-term psychotherapy for adolescents with major major trial. JAMA. 2004;292(7):861-863.
depressive disorder. Arch Gen Psychiatry. 2000;57(1):29-36.
45. Apter A, Kronenberg S, Brent D. Turning darkness into light: a new landmark 53. Clarke GN, Rohde P, Lewinsohn PM, Hops H, Seeley JR. Cognitive-behavioral study on the treatment of adolescent depression: comments on the TADS study.
treatment of adolescent depression: efficacy of acute group treatment and booster Eur Child Adolesc Psychiatry. 2005;14(3):113-116.
46. Kratochvil C, Emslie G, Sliva S, McNulty S, Walkup J, Curry J, Reinecke M, Vi- sessions. J Am Acad Child Adolesc Psychiatry. 1999;38(3):272-279.
tiello B, Rohde P, Feeny N, Casat C, Pathak S, Weller E, May D, Mayes T, Robins 54. Antonuccio D, Burns D. Adolescents with depression [letter]. JAMA. 2004;292(21): M, March J; TADS Team. Acute time to response in the Treatment for Adoles- cents with Depression Study (TADS). J Am Acad Child Adolesc Psychiatry. 2006; 55. Mufson L, Dorta KP, Wickramaratne P, Nomura Y, Olfson M, Weissman MM.
A randomized effectiveness trial of interpersonal psychotherapy for depressed 47. Emslie GJ, Rush AJ, Weinberg WA, Kowatch RA, Hughes CW, Carmody T, Rin- adolescents. Arch Gen Psychiatry. 2004;61(6):577-584.
telmann J. A double-blind, randomized, placebo-controlled trial of fluoxetine in 56. National Institute for Health and Clinical Excellence. Depression in Children and children and adolescents with depression. Arch Gen Psychiatry. 1997;54(11): Young People. London, England: National Institute for Health and Clinical Excel- 48. Emslie GJ, Heiligenstein JH, Wagner KD, Hoog SL, Ernest DE, Brown E, Nilsson 57. Burns BJ. Children and evidence-based practice. Psychiatr Clin North Am. 2003; M, Jacobson JG. Fluoxetine for acute treatment of depression in children and (REPRINTED) ARCH GEN PSYCHIATRY/ VOL 64 (NO. 10), OCT 2007 2007 American Medical Association. All rights reserved.

Source: http://filer.case.edu/org/feenylab/documents/TADS.pdf