The popular press intermittently acts surprised that antidepressant medications actually have little scientific evidence supporting their efficacy. It’s old news, but it’s still important news and I’m glad for the recent reports. See: http://www.everydayhealth.com/news/did-studies-lack-key-data-on-link-between-antidepressants-youth-suicides/
Rita and I published an article about this in 1996. Below, I’ve pasted a pre-print excerpt from an article I published with Duncan Campbell in 2009 in the Journal of Contemporary Psychotherapy. It includes a brief summary of antidepressant medication research through 2008 or so. Check it out:
A Brief History and Analysis of Antidepressant Medication Treatment for Youth
Medication treatment for depressed youth has evolved over three relatively distinct periods. First, prior to 1987, small exploratory studies examined tricyclic antidepressant (TCAs) efficacy with young patients diagnosed with major depressive disorder (MDD). Second, from 1987-1994 there were a number of randomized, controlled trials (RCTs) of TCA efficacy; these efforts often employed double-blind procedures and inactive placebo controls. Third, since 1997, research efforts have primarily focused on evaluating selective serotonin reuptake inhibitor (SSRI) efficacy with RCTs.
Early Research: Pre-1987
In the early 1980s, psychiatric and pharmaceutical researchers began testing TCAs with youth. Early conclusions about the safety and efficacy of TCAs were generally optimistic (Klein, Jacobs, & Reinecke, 2007). This is a tendency that has been identified in the literature and it may be due to methodological limitations, confirmation bias or an allegiance to the medical model, or financial incentives associated with the pharmaceutical industry (Klein et al., 2007; Luborsky et al., 1999). For example, on the basis of existing studies and their very small double-blind trial with nine prepubertal children, Kashani and colleagues (1984) concluded that amitrityline was possibly efficacious for treating depression in children. Interestingly, the authors’ tentative claim was made despite the fact that no statistically significant effect was observed for amitriptyline and even though 11% of their sample “developed a hypomanic reaction while on the protocol” (p. 350).
RCTs with TCAs
From 1965 to 1994 there were 13 published RCTs evaluating TCA efficacy. Most of these studies were conducted from 1987 to 1994 (Fisher & Fisher, 1996; Sommers-Flanagan & Sommers-Flanagan, 1996). These RCTs confirmed the premature hopefulness of Kashani and colleagues’ early claims. Indeed, no study ever published showed that TCAs outperformed placebo in the treatment of youth depression (Hazell, 2000). More importantly, it is currently recognized that TCAs possess dangerous side effect profiles, while offering no demonstrable advantage over placebo in the treatment of youth depression (Hazell, 2000; Pellegrino, 1996).
In the mid-1990s there was considerable speculation about why TCAs were ineffective for treating youth. The primary hypothesis for involved the fact that children appear to have immature adrenergic synaptic systems. This possibility precipitated a more systematic inquiry of serotonergic medications.
RCTs with SSRIs
Using PsychInfo and PubMed searches combined with cross-referencing, we identified 12 published RCTs evaluating SSRI efficacy with 11 of these studies from 1997 to 2007. In total, these studies compared 1,223 SSRI treated patients to a similar number of placebo controls. On the basis of the researchers’ own efficacy criteria, six RCTs observed outcomes favoring medication over placebo, and six observed nonsignificant differences. Researchers described efficacious outcomes for fluoxetine (3 of 4 studies; G. J. Emslie et al., 2002; G. J. Emslie et al., 1997; Simeon, Dinicola, Ferguson, & Copping, 1990; Treatment for Adolescents With Depression Study (TADS) Team, US, 2004), paroxetine (1 of 3; Berard, Fong, Carpenter, Thomason, & Wilkinson, 2006; G. Emslie et al., 2006; M. B. Keller, 2001), sertraline (1 of 1; K. D. Wagner et al., 2003), and citalopram (1 of 1; K. D. Wagner et al., 2004). Neither of two studies observed efficacy for venlafaxine (G. J. Emslie, Findling, Yeung, Kunz, & Li, 2007; Mandoki, Tapia, Tapia, & Sumner, 1997), and the single escitalopram study returned negative results (K. D. Wagner, Jonas, Findling, Ventura, & Saikali, 2006).
Assessing a medication’s efficacy is a complex process with challenges that are difficult to address. We believe, however, that the six aforementioned RCTs favoring SSRIs suffered from methodological problems and issues that temper their positive conclusions. For example, (a) two of the three fluoxetine studies were characterized by unusually high and disproportionate discontinuation rates in the placebo conditions; (b) 11 of the 12 studies based their conclusions exclusively on a structured psychiatric interview; (c) despite simultaneous examination of several outcomes, no study used statistical adjustments for multiple comparisons; (d) placebo washouts and statistical approaches that advantage medications were nearly always employed (R. P. Greenberg, 2001); (e) no procedures were used to evaluate double-blind integrity (R. P. Greenberg & Fisher, 1997); and (f) despite documented inter-racial differences in medication metabolism and responsiveness, conclusions were generalized to all youth and inappropriately failed to account for racial/cultural specificity (Lin, Poland, & Nakasaki, 1993).
Side Effects and Adverse Events
In RCTs and other studies, patients treated with SSRIs experienced substantially more disturbing side effects and adverse events than those not treated with SSRIs. For example, in one of the most rigorous studies to date, the Treatment of Adolescents with Depression Study (TADS), 11.9% of the fluoxetine group evidenced harm-related adverse events (compared to 4.5% in the Cognitive Behavioral Therapy [CBT] group) and 21% experienced psychiatric adverse events (1% in the CBT group). Further, as the authors noted, “…suicidal crises and nonsuicidal self-harming behaviors were not uncommon and, with the caveat that the numbers were so small as to make statistical comparisons suspect, seemed possibly to be associated with fluoxetine treatment” (March et al., 2006; The TADS Team, 2007 p. 818; Treatment for Adolescents With Depression Study (TADS) Team, US, 2004).
Findings like these necessitate critical inspection of study results and should attenuate positive conclusions about medication safety. For example, Emslie et al.’s (1997) study of youth depression was the first ever to demonstrate superior outcome for an SSRI. In addition to the study’s numerous methodological problems, the authors noted that 6.3% of the fluoxetine patients (n = 3) developed manic symptoms. Although this percentage may sound small, extrapolation suggests that 6,250 of every 100,000 fluoxetine-treated youth might develop manic symptoms. Ultimately, despite data based solely on psychiatrist ratings and a placebo condition discontinuation rate approaching 46%, the authors concluded that fluoxetine “…is safe and effective in children and adolescents with MDD” (p. 1037). Moreover, the authors’ intent-to-treat analysis possibly conferred an advantage for the active drug group. In our opinion, this methodological problem and the mania data make it premature to conclude that fluoxetine is safe and effective in children.
Similarly, despite striking data that appear to demonstrate otherwise, authors of the single positive paroxetine study concluded that paroxetine is “safe and effective” for young patients (M. B. Keller et al., 2001). However, in their results section, the research team reported serious adverse effects, “…in 11 patients in the paroxetine group, 5 in the imipramine group, and 2 in the placebo group” (p. 769). More specifically, five adverse effects in the paroxetine group involved suicidal ideation or gestures. Despite these data, the researchers presented their results as evidence for the efficacy and safety of paroxetine treatment for adolescent depression. Because 12% of the paroxetine-treated adolescents experienced at least one adverse event and because 6% of these patients manifested increased suicidality or suicidal gestures (compared with zero in the imipramine and placebo groups), we believe the authors’ conclusion departs from the data in a significant and concerning way.
Shortly after publication of the Keller et al. (2001) study, regulatory agencies in France, Canada, and Great Britain restricted SSRI use among youth. In September of 2004, an expert panel of the U.S. Food and Drug Administration (FDA) followed suit and voted 25-0 in support of an SSRI-suicide link. Later, the panel voted 15-8 in favor of a ‘black box warning’ on SSRI medication labels. The warning states:
“Antidepressants increased the risk compared to placebo of suicidal thinking and behavior (suicidality) in children, adolescents, and young adults in short-term studies of Major Depressive Disorder (MDD) and other psychiatric disorders. Patients of all ages who are started on antidepressant therapy should be monitored appropriately and observed closely for clinical worsening, suicidality, or unusual changes in behavior. Families and caregivers should be advised of the need for close observation and communication with the prescriber.”
In 2006, the FDA extended its SSRI suicidality warning to adult patients aged 18-24 years (United States Food and Drug Administration, 2007).
Combination Medication and Psychotherapy Treatments
Many view the 2004 TADs study as a ‘state of the science’ comparison of SSRI medication (fluoxetine; FLU) with CBT and their combination (FLU + CBT). To date, it represents the largest placebo-controlled study comparing mono-therapy (FLU or CBT alone) with combination therapy. Not surprisingly, the TADs study has generated numerous publications and much controversy (Antonuccio & Burns, 2004; Diller, 2005; Weisz, McCarty, & Valeri, 2006).
To summarize, initial 12-week outcomes showed that 71% of FLU + CBT patients evidenced “much” or “very much” improvement on the on the CGI-Improvement item, a clinician-based assessment. FLU alone produced a similar outcome (60.6%), whereas the CBT alone (43.2) outcome did not differ significantly from placebo (34.8%). Based on these outcomes, several CBT researchers and practitioners criticized the specific CBT delivered to TADs participants. Brent (2006), for example, described TADS psychotherapy as a relatively “dense treatment, with multiple CBT strategies, each delivered at a relatively low dose” (p. 1463). In comparing the initial TADs CBT outcomes with previous and subsequent CBT studies, Weisz et al. (2006) suggested that the TADs CBT was weaker than most CBT interventions, for various reasons:
“the CBT ES (effect size) generated in TADS is not characteristic of most CBT or psychotherapy effects on youth depression; 20 of the 23 other CBT programs. . . showed larger ES than the TADS version of CBT, and the mean ES value across the non-TADS CBT programs. . . was 0.48, markedly higher than the -0.07 ES associated with the TADS CBT intervention” (p. 147).
To complicate issues further, follow-up data suggest that the TADs CBT evidenced delayed effectiveness, as it eventually “caught up” with FLU and CBT+FLU (The TADS Team, 2007). At week 18, for example, there were no statistically significant differences between CBT and FLU, and by week 36 there were no statistical differences among the three groups (CBT, FLU, and CBT + FLU) on primary outcome measures. Although the interventions including FLU might evidence a speedier antidepressant effect, these results suggest that CBT is equally effective over time.
The depression treatment literature frequently includes recommendations for combined interventions in order to maximize outcomes (Watanabe, Hunot, Omori, Churchill, & Furukawa, 2007). Unfortunately, however, little data exists to support these recommendations. In addition to TADs, the only other published RCT comparison of mono- and combination treatments for depressed adolescents reported partial remission rates of 71% for CBT, 33% for sertraline, and 47% for combination (Melvin et al., 2006). Medication group patients also evidenced significantly more adverse events and side effects. Although the researchers attributed the delayed response in the combination group to sertraline, they concluded with the puzzling statement that “CBT and sertraline are equally recommended for the treatment for adolescents with depression, each demonstrating an equivalent response” (Melvin et al., 2006 p. 1160).