Suicide is the third leading cause of death among persons 10 to 19 years of age, and up to 6 percent of children and adolescents are thought to have depression. A meta-analysis by Whittington and colleagues raises concern that the efficacy of selective serotonin reuptake inhibitors (SSRIs) in childhood depression could have been overestimated in the medical literature.

The authors identified more than 5,000 potentially relevant articles, but only five were randomized placebo-controlled trials of SSRIs. In two trials with 315 participants seven to 18 years of age who had major depression, it was found that fluoxetine was more likely than placebo to achieve remission by eight weeks. Fewer adverse effects were reported for fluoxetine than placebo, and discontinuations were similar in the two groups. When unpublished data were incorporated into the analysis, the risk-benefit profile remained favorable.

One published trial for paroxetine was identified. Of 180 participants 12 to 18 years of age with major depressive disorder who were treated for eight weeks, more patients in the treatment group met criteria for remission than patients in the placebo group, but other measures of efficacy were similar between the groups. Serious adverse events were reported in 11.8 percent of treated patients compared with 2.3 percent of placebo patients. When unpublished data were included in the analysis, the evidence suggested that paroxetine did not improve depressive symptoms or response but had an increased risk of serious adverse events, especially suicidal ideation or attempts.

Sertraline was studied in two published trials involving 376 participants six to 17 years of age with major depression. At 10 weeks, sertraline was more likely than placebo to bring about response but showed little difference in mean depressive scores. Serious adverse events were slightly more common among the sertraline group (3.7 percent compared with 3.3 percent of the placebo group). The addition of unpublished data on remission suggested that the risk-benefit ratio was unfavorable.

No published trials were identified on the use of citalopram in children, but two unpublished trials provided data from 422 patients seven to 18 years of age with major depressive disorder who were treated with citalopram. These data suggested that citalopram was unlikely to produce a clinically significant reduction in symptoms after eight to 12 weeks of therapy, but the rate of attempted suicide doubled (from 3.6 to 7.1 percent), and citalopram was associated with a small increase in adverse events.

The one small published trial (40 participants) and two unpublished studies (334 participants) of venlafaxine suggested that clinically significant improvement in symptoms was unlikely by six to eight weeks of therapy, and the risk of adverse events sufficient to cause discontinuation, including suicide, was about 10 percent.

The authors conclude that clinical guidelines on the use of SSRIs in children and adolescents have been influenced by the failure to publish results from all trials. When unpublished data are incorporated into evaluations, the risks could outweigh the benefits of treatment with some SSRIs.