Paxil in Adolescent Depression: What Are the Data?
Yesterday a couple of Brown University journalists showed up the mainstream press by writing a comparatively well researched and cogent article on a controversial 2001 paroxetine (Paxil; GSK) study in adolescent depression. The study's lead author, Brown psychiatry professor Martin Keller, is at the center of the controversy and the focus of charges from Australian psychiatrist Jon Jureidini that data in the GSK-funded study were manipulated to favor the antidepressant and downplay any related suicide risks.
However, what's missing* in this coverage (and the coverage of the coverage) is whether subsequent clinical trials support the use of paroxetine in adolescent depression. In other words, are the findings of Keller et al reproducible (notwithstanding allegations of data manipulation)?
Paroxetine is not indicated for the treatment of depression in adolescents and, in fact, its prescribing information (like that of other antidepressants) carries a black-box warning, which advises of the risks of suicidal ideation or acts in treated adolescents with psychiatric disorders.
A search of the PubMed database reveals 5 controlled clinical trials, in addition to the Keller study, that examined the use of paroxetine in adolescent depression. Three of these studies, like the Keller trial, compared paroxetine with placebo, and 2 of these studies (here and here) concluded that the antidepressant was no better than placebo, according to the Children's Depression Rating Scale-Revised or the Montgomery-Asberg Depression Rating Scale (MADRS), for the short-term (8- or 12-week) treatment of major depressive disorder in adolescents.
A post-hoc analysis of one of these studies found no difference between paroxetine- and placebo-treated patients in the rates of suicidal behavior or ideation. The second study found low rates of suicidal behavior or ideation with either treatment: paroxetine, 1.92% (2/104); placebo, 0.98% (1/102). (The third placebo-controlled paroxetine study, authored by investigators in Nova Scotia and a GSK employee, was conducted to assess the validity of the Kutcher Adolescent Depression Scale as a clinical-trial outcome measure.)
A clomipramine-controlled study, performed by the DEROXADO Study Group in France, found no significant efficacy differences between the 2 drugs at 8 weeks (primary outcome measures, Clinical Global Impression scale [CGI] and MADRS); however, side effects, particularly anticholinergic side effects, were significantly more frequent with clomipramine. The rates of "suicidal acts" with either drug were comparable: clomipramine, 12.1% (7/58); paroxetine, 12.7% (8/63).
Last, a study published this year in JAMA assessed the efficacy of switching to 1 of 3 SSRIs (paroxetine, citalopram [Celexa; Forest], or fluoxetine [Prozac; Eli Lilly]) or the SNRI venlafaxine (Effexor; Wyeth), with or without cognitive behavioral therapy (CBT) in 334 adolescents whose depression was refractory to a 2-month trial of an SSRI alone. At 12 weeks, responses rates (per the CGI-I) were relatively higher with either medication (SSRI or SNRI) when combined with CBT. The authors found no treatment differences in the rates of suicidal ideation or "harm-related" adverse events.
Consequently data regarding the clinical efficacy of paroxetine in adolescent depression are mixed (if you believe that they haven't been manipulated in their entirety), given the handful of studies that show either the drug's comparable efficacy to placebo or its comparable efficacy to other antidepressants (which may or may not depend on the outcome measure used). The data assessing the suicide risk (either ideation or act) with the use of paroxetine specifically are less murky, with little-to-no differences between the drug and placebo or other antidepressants.
CGI-I = Clinical Global Impression-Improvement scale; SNRI = serotonin-norepinephrine-reuptake inhibitor; SSRI = selective serotonin-reuptake inhibitor.
* Let's not even discuss whether it's more or less preferable to align your medical career with pharma, as in the case of Keller, or personal-injury lawyers, as in the case of Jureidini.
