Treatment of Depression in Adolescents: TADS Results and Future Directions

Bruce Waslick, MD;

Treatment of Depression in Adolescents: TADS Results and Future Directions

October 1, 2007

Bruce Waslick, MD

Publication

Article

Psychiatric TimesPsychiatric Times Vol 24 No 12

Volume 24

Issue 12

The Treatment for Adolescents With Depression Study (TADS) represents the single largest and potentially most informative clinical trial of adolescents with depressive illness. The overall aim of the study was to investigate the effectiveness of standard interventions for adolescent outpatients with moderate to severe depression. The rationale for the trial, the study design, the clinical characteristics of the sample recruited for the trial, the acute phase treatment results, and a series of secondary analyses of the acute treatment phase have been published in recent papers.^1-11 This article summarizes these published reports with the goal of guiding stakeholders in making informed decisions about treatment interventions for youths with depression.

The 4 stages of the study

Funded in 1998 by the NIMH, TADS, a multisite study, was designed to answer questions regarding the short-term and longer-term relative effectiveness and safety of the best evidence-supported interventions available for the treatment of depression in adolescents. The study consisted of 4 stages. In the 12-week acute trial phase (stage 1), 3 active treatments-cognitive-behavioral therapy (CBT), fluoxetine, and a combination of fluoxetine and CBT (COMB)-were compared with placebo.

Patients who were considered responders or partial responders to an active treatment (CBT, fluoxetine, or COMB) in the first stage could continue in their assigned treatment arm for 6 weeks of intensive (for partial responders) or maintenance treatment (for full responders) in stage 2. Treatment responders at the end of the second stage could continue in their assigned treatment arm for 18 weeks of additional maintenance level treatment in stage 3. At the end of stage 3, patients and families received appropriate treatment recommendations and exited the controlled part of the trial.

All adolescents recruited for the trial in all treatment arms, regardless of responder status, were encouraged to participate in all study assessments to track their clinical and functional outcomes. In stage 4, patients were followed naturalistically for a year after completing the initial treatment protocol.

Primary outcome measures included clinician-rated and self-reported depressive symptoms and clinician-rated global improvement in depressive illness since beginning treatment. Because of the importance of eliminating bias from the primary outcome measures, independent evaluators (IE) blinded to treatment assignment followed the adolescents and their families through the course of the trial and provided the primary effectiveness outcome data. The primary measures comparing safety and tolerance of the interventions were clinician-rated adverse-event measures.

Results to date

The data analyses published to date have focused on the baseline characteristics of the recruited sample and primary and secondary analyses of the data collected through stage 1. TADS data suggest that the investigators were successful in recruiting adolescents with moderate to severe depression.³ Some of the pertinent summary statistics of the baseline characteristics of the sample are presented in Table 1. For the most part, generally comparable subsamples were randomly assigned to the 3 active (CBT, fluoxetine, or COMB) and the comparison (placebo) interventions.³

Total number of subjects: 439

Average age: 15 (12 - 17)

Sex

Male: 54%

Female: 46%

Race

White: 75%

Black: 13%

Hispanic: 9%

Median family income: $75,000

In terms of the effectiveness of the 4 interventions studied, a clear and generally consistent story emerged after 12 weeks of treatment (end of stage 1).² The data on change in IE-rated or self-reported depressive symptoms and IE-rated improvement in depressive illness suggest that the best treatment results were seen with the COMB intervention (Table 2) [Table restricted. Please see print edition for content]. Results for COMB treatment were clearly superior to those for the CBT and placebo groups. Improvement in depressive symptoms and functional outcomes appeared to be better in patients who were treated with fluoxetine than in those who received CBT and placebo, but not as good as in patients who received COMB.

Overall response rates in the 4 groups after 12 weeks of treatment are presented in the Figure and suggest the superiority of the COMB intervention. Also of interest are the remission rates after 12 weeks of treatment. "Remission" was considered to reflect an excellent outcome in terms of decreased depressive symptoms compared with "response," which meant that the patient achieved at least a good response. As expected, remission rates at 12 weeks were lower than the response rates (COMB, 37%; fluoxetine, 23%; CBT, 16%; placebo, 17%) but were still significantly better with COMB treatment.⁹

In secondary analyses, COMB treatment not only improved depressive symptoms and depressive illness more than the other treatments, it also led to better change in functioning and quality of life.⁷

Safety outcomes suggested a somewhat different pattern in the primary analyses.² CBT was associated with very low rates of medically-related adverse events, psychiatric adverse events, or suicidality; the rates in this group were even lower than those in the placebo group. Whether this reflects actual protective effects against adverse events by CBT or methodological variance in assessing adverse events across conditions is not clear.

Adverse events

Fluoxetine treatment in either the fluoxetine or the COMB intervention was associated with a variety of medical and psychiatric adverse events compared with placebo (Table 3).² These adverse events are often encountered in clinical practice with SSRI therapy for adults or minors with depressive illness. TADS methodology recorded and reported only adverse events determined by the treating clinician to be at least moderately severe and causing at least some functional impairment, thus adverse events reported in TADS may underestimate the rate of all clinically relevant adverse events.

TABLE 3 Adverse events associated with fluoxetine in TADS

Adverse event category

Type of event observed

Nonpsychiatric

Headaches, GI symptoms, tremor

Psychiatric

Mania/hypomania, irritability, worsened depressed mood,agitation/restlessness, anxiety, sedation, insomnia

Suicidality

Suicidal thoughts, suicide attempts

Patients randomized to fluoxetine experienced more events suggestive of treatment-emergent suicidality during the first 12 weeks of treatment.² Overall, fewer suicidal events occurred in the group treated with COMB, suggesting that combining CBT with fluoxetine may mitigate the risk of emergent suicidal events associated with fluoxetine monotherapy.

Secondary analyses of suicidality included self-report measures of suicidal thinking, clinician-rated items on suicidality, and IE-rated suicidality measures and suggested that in general, suicidality improved in all groups in the trial, although improvement with COMB appeared to be the most robust.⁵ Secondary measures of suicidal worsening in the trial did not consistently suggest that fluoxetine was associated with treatment-emergent suicidality, leaving the issue somewhat unresolved. A cautious interpretation would suggest that in a relatively small number of cases, treatment of depressed adolescents with fluoxetine may be associated with treatment-emergent suicidality and that close monitoring and vigilance for this negative outcome is prudent.

Other issues

Other clinically relevant issues in treatment research with depressed adolescents include trying to better understand which subgroups of patients respond best to treatment in general and how various subgroups respond to the different treatment interventions. TADS explored these issues by conducting secondary analyses of predictors (variables measured before beginning treatment that predicted lower depressive symptom scores for the entire group of participants) and moderators (variables measured at baseline that predicted differential response to treatment through lower depressive symptom scores).⁶ Because the sample size was rather small for these types of analyses and these analyses were not the primary aims of the study, the results should be considered exploratory and hypothesis-generating rather than definitive (Table 4) [Table restricted. Please see print edition for content].

TADS was able to identify a limited number of demographic, clinical, and treatment expectancy variables that predicted a better outcome of depressive symptoms at 12 weeks. In general, younger patients benefited more from treatment than older adolescents.⁶ Sex, race, ethnicity, and socioeconomic status did not significantly influence treatment response. Analyses examining clinical variables suggested that adolescents who were less chronically depressed at baseline, who had less functional impairment before beginning treatment, who were less hopeless with less suicidal ideation, who had fewer melancholic features, and who had fewer comorbid diagnoses had better outcomes for depressive illness. Furthermore, patients who had higher expectations of treatment benefit before starting treatment tended to be less depressed at the end of 12 weeks.

Analyses of moderators produced some unexpected results.⁶ The findings suggest that CBT monotherapy, for unexplained reasons, may be an appropriate best practice intervention for patients with higher socioeconomic status. This is in marked contrast to the superiority of COMB treatment compared with CBT only in participants from lower socioeconomic status families.

In addition, COMB treatment results were not significantly different from those for fluoxetine monotherapy in the most severely depressed patients, whereas substantial differential benefit was seen in the less severely depressed patients.

Beyond socioeconomic status and depression severity at baseline, and despite analyses examining a variety of demographic and clinical factors, no other statistically significant moderators of treatment results were identified. With the caveat that TADS was not designed to definitively address moderator questions, TADS results suggest that COMB treatment is robustly superior to the other treatment interventions evaluated in most subgroups of patients.

Future directions

Although TADS has generated a great deal of important information to date, many questions remain regarding the treatment of adolescents with depression. Some questions will be addressed by forthcoming analyses of additional TADS data. For example, although acute treatment outcome is certainly important, little is known about efficacy, safety, and especially, functional status outcomes obtained with available treatment intervention for longer periods.

Future publication of TADS data will address the issue of longer-term outcomes. Two publications will be of special importance because they will report on additional aims of the TADS project. First, comparisons of depressive symptom outcome, improvement in depressive illness, and adverse events in the groups receiving active treatments (fluoxetine, CBT, or COMB) at week 36 will be examined in a forthcoming publication.¹² Second, data on the naturalistic observation of all patients retained for assessment 1 year after stopping study treatments are currently being analyzed and prepared for publication. Longer-term benefit and safety issues are of considerable importance to clinicians, patients, family members, and policy makers in deciding on treatment interventions, both at a clinical level and at a service-delivery level.

Unresolved questions

Other important questions raised by TADS probably will require further research with new study designs and cohorts. For example, one recent study (with some significant differences from TADS in design features, setting, and sample characteristics) was not able to replicate the TADS finding of the superiority of combining medication with psychotherapy versus monotherapy medication treatments in depressed adolescents,¹³ which highlights interesting unresolved questions. Can we conclude that care planning for adolescents with depression should set combined treatment with evidence-based psychotherapy and medication treatment as a standard? Is offering combination treatment to all adolescent patients presenting with depressive illness a cost-effective way to manage health care resources? Also, following up on TADS safety data findings, how do clinicians identify patients at risk for treatment-emergent suicidality during antidepressant treatment and what are the best management strategies should this occur during the course of treatment?

Conclusion

TADS was not designed to answer all questions about the safety and effectiveness of available treatment interventions for adolescents with depressive illness. It is, of course, unreasonable to expect wide-ranging definitive answers in any specialized area of medicine or psychology from one study, no matter how well the study is funded, designed, or executed. TADS has, however, probably served its primary purpose, which is to generate new knowledge, relevant to a variety of stakeholders, on which current consumer choices, clinical decision making, mental health policy decisions, and future research can be built.

References:

References

Treatment for Adolescents With Depression Study Team. Treatment for Adolescents With Depression Study (TADS): rationale, design, and methods.