Historically, research into the effectiveness of antidepressants for depressed youths has been unimpressive. In the 1980s, Joaquim Puig-Antich studied the then-popular tricyclic antidepressants and found no evidence to support their effectiveness (Puig-Antich J, Arch Gen Psych 1987;44(1):81–89). To make matters worse for tricyclics, in the 1990s several studies hinted that these meds can cause sudden death in children (Popper CW et al., J Child Adolesc Psychopharmacol 1990;1(2):125–132). While the true level of hazard remains controversial, these reports led most child psychiatrists to abandon tricyclics in favor of SSRIs.

The history of SSRI use in children has been fraught with its own series of disappointments and controversies. In the early 2000s, controversy brewed about bias in pharmaceutical industry sponsored studies of SSRIs. The suppression of information regarding the increased risk for suicidal behavior in children treated with paroxetine (Paxil) (see Kondro W, Can Med Assoc J 2004;170(5):783) culminated in a black box warning for SSRIs in 2003, an FDA action that some have argued ultimately led to an increase in previously declining rates of suicide among youth in the U.S. In April 2004, The Lancet published a meta-analysis of studies evaluating SSRIs versus placebo in participantsfive to 18 years old. Taking into account previously suppressed studies, the authors posited that “risks could outweigh benefits of these drugs (except fluoxetine)” (Whittington CJ, et al., The Lancet 2004; 363(9418);1341–1345).

In this context, in 2004 child psychiatrists welcomed the non industry funded, and thus presumably unbiased, NIMH study from the Treatment for Adolescents with Depression Study (TADS) team, evaluating the efficacy of fluoxetine (Prozac) and cognitive behavioral therapy (CBT) in the treatment of depressed adolescents. In this large, placebo-controlled study, 439 depressed adolescents were randomly assigned to one of four treatment arms: fluoxetine alone, CBT alone, a combination of CBT and fluoxetine, or placebo. After 12 weeks, the rate of response to fluoxetine with CBT was 71 percent; fluoxetine alone 61 percent; CBT alone 43 percent; and placebo alone 35 percent (March JS et al., JAMA 2004;292(7):807– 820). While the study appeared to endorse the effectiveness of fluoxetine, it nevertheless had some limitations; for example, adolescents who had previously attempted suicide were excluded.

For this reason, the publication of three papers in the October 2009 issue of JAACAP (Vol. 48, Issue 10) from the NIMH funded Treatment of Adolescent Suicide Attempters (TASA) trial is significant. Indeed, the editorial by Garry Walter, MD, PhD, in the same issue borrows the language of the opera Turandot in its title, proclaiming: “Nessun Dorma (‘None Shall Sleep’)… At Least Not BeforeWe Digest Treatment of Adolescent Suicide Attempters (TASA).”

TASA was a six-month, open label trial (meaning both researchers and participants knew who got which treatment) that included 126 patients ages 12 to 18 with unipolar depression who had made a suicide attempt within 90 days of intake. Exclusion criteria included bipolar disorder, psychotic symptoms, substance dependence, and pervasive developmental disorder. Subjects were randomly assigned to one of three conditions: CBT, medication management, or the combination. The medication treatment was derived from the Texas Medication Algorithm, which suggests that clinicians begin by prescribing fluoxetine, citalopram (Celexa), or sertraline (Zoloft); followed in cases of nonresponse by an alternate SSRI; followed, if necessary, by an alternate class—venlafaxine (Effexor), duloxetine (Cymbalta), mirtazapine (Remeron), or bupropion (Wellbutrin). Although the study was initially designed as a three arm randomized trial, recruitment difficulties led to a shift, allowing participants either to be randomized (n=22), or to choose their preferred treatment (n=102). (Two participants dropped out before treatment assignments.) Ninety-three participants were in the combination therapy arm of the trial, while 17 had psychotherapy alone, and 14 had medication managment.

The first of the TASA articles addresses the prediction of suicidal events. Of the 124 enrolled participants, 24 experienced a suicidal event at some point during the six-month trial. While there were no completed suicides during the study, one completed suicide occurred after the study ended. There was no relationship between treatment assignment and suicide events. Risk factors for suicide attempts during the trial included higher self-rated depression, suicidal ideation, higher family income, greater number of previous suicide attempts, lower maximum lethality of previous attempt, history of sexual abuse and lower family cohesion (Brent DA, JAACAP 2009;48(10):987–996).

The second article looks at the course of depression during the treatment of these adolescents. The remission rates were 32 percent at week 12, and 50 percent at week 24. While these remission rates are similar to those reported in the TADS study, the two studies are not strictly comparable because the TADS study had a double blind design, whereas the TASA study did not. Typically, response and remission rates are higher in open label studies. Nonetheless, the authors note that their remission figures are consistent both with a previous smaller study in adolescents and with the adult data from the Sequenced Treatment Alternatives to Relieve Depression study (STAR*D) (Vitiello B et al., JAACAP 2009;48(10):997–1004).

The third of the three articles describes the manual based CBT-SP (cognitive behavioral therapy for suicide prevention) program that was used in the TASA study. The article lays out the theoretical background of the technique, with details about how to do a “chain analysis” of suicide attempts, safety planning, psychoeducation with family, addressing reasons for living and building hope, and other practical aspects of CBT-SP. The clarity of this section gives the non CBT trained clinician a good picture of this mode of treatment (Stanley B et al., JAACAP 2009;48(10):1005–1013).