There has been considerable debate over the benefits and safety of antidepressive drugs in children and adolescents; regulatory bodies currently consider that use in this age group required caution. The FDA in the US carried out a major review of suicidal ideation and behaviour associated with paediatric use of antidepressives, as a result of which warnings were made mandatory in the product literature for these drugs. A previous meta-analysis has looked at the risk to benefit profile of these drugs for paediatric depressive illness but not other indications. As two large trials have been published since this was done, the authors of the current paper carried out a more up to date review and meta-analysis. Their objective was to determine newer antidepressive drugs' efficacy in major depressive disorder (MDD), obsessive-compulsive disorder (OCD), and non-OCD anxiety. They also looked at trial- and patient-level moderators of risk and benefit.

They carried out a fairly comprehensive search that included data submitted to US and UK Regulatory authorities. Eligible studies were randomised placebo-controlled trials of SSRI and other newer antidepressive drugs in children and adolescents up to age 18 with MDD, OCD, or non-OCD anxiety disorders; studies had to include efficacy data or data on suicidality (suicidal ideation or behaviour) for both active and placebo groups; where data on suicide ideation or behaviour was not included, this was sought from the FDA review or the original authors. Primary outcomes were efficacy, as the original study primary measure of response and change from baseline, and suicidal ideation or behaviour.

A total of 27 (n=5,310) eligible trials were identified, in MDD (15 trials, n=3,430), OCD (6 trials, n=718), and non-OCD anxiety disorders (6 trials, n=1,162). Most had data on response rates (in 25) and on suicidality (partial in 1); all included a usable measure of efficacy as a primary outcome. Analysis showed that antidepressives were associated with higher response rates in MDD (difference in response compared to placebo 11%, 95% CI 7.1 to 14.0%), OCD (difference 19.8%, 95% CI, 13.0% to 26.6%), and in non-OCD anxiety (difference 37.1%, 95% CI 22.5% to 51.7%). Over the pooled data, there was a small increase in risk of suicidal ideation or behaviour (risk difference 0.9%, 95% CI 0.1% to 1.3%; number needed to harm, 143, 95% CI, 77 to 1000). For individual indications, the differences in suicidality were not statistically significant. There were no completed suicides recorded. Data from the placebo groups indicated that the risk of suicidality was significantly greater in the MDD patients compared to the others.

The authors conclude that the meta-analysis shows evidence of efficacy for these drugs relative to placebo in all three conditions. Efficacy appears to be greatest in non-OCD anxiety, intermediate in OCD, and modest in MDD. Adolescents appear to respond better than children. There is evidence of a small increase in suicidality associated with use of the drugs, but the difference compared to placebo was less than 1%. They discuss the limitations of the analysis, which are primarily those of the underlying data. The overall number of trials and patients available is relatively small, which precludes the detection of all but major effects; all trials screened patients before randomisation to exclude those considered to be at highest suicide risk. Data on suicidality was also collected as part of adverse event reporting, not as a study outcome. Nevertheless, the analysis suggests that the benefits for these drugs outweigh the risks in patients similar to study populations.