Tim Kendall Steve Pilling Craig J. Whittington, PhD Curr Opin Psychiatry 18(1):21-25, 2005.
Abstract and Introduction
Purpose of Review: This article summarizes recent developments in the use of selective serotonin reuptake inhibitors and atypical antidepressants to treat children and adolescents with depression at a time when their use in this context has generated considerable controversy and confusion for clinicians, patients and their families. Recent reports and recommendations from drug regulators in the UK and the US are discussed, alongside other reviews and recently published randomized controlled trials. Recent Findings: It is now widely accepted that these drugs increase the risk of suicide-related behaviours and although recently published trials have been more positive, a metaanalysis of published and unpublished trials has cast doubt about efficacy. The evidence for publication bias in the studies considered is also raised and the implications discussed. There is some evidence, however, that the combination of psychological treatment with fluoxetine may be both effective and protective against the increased risk of suicide-related behaviours, although problems over blinding suggest further research is needed to clarify this potentially positive combination treatment. Summary: Current evidence supports the conclusions of the UK drug regulator in warning against the use of all the newer antidepressants except fluoxetine in this age group, and alternative therapies should be sought in the first instance. Caution is needed in interpreting drug company sponsored trials given the evidence of selective reporting and publication bias. Combining fluoxetine with a psychological treatment such as cognitive-behavioural therapy is also worth considering.
Introduction This paper reviews the current evidence for the safety and efficacy of the selective serotonin reuptake inhibitors (SSRIs) and atypical antidepressants (venlafaxine, nefazodone and mirtazapine) in the treatment of major depressive disorder (MDD) in children and adolescents. This has been the subject of considerable controversy involving the pharmaceutical industry, the regulatory authorities and clinical and academic experts in both Europe and North America.[1*,2*] In the UK, the Medicines and Healthcare Products Regulatory Agency (MHRA) advised that the SSRIs citalopram, escitalopram, paroxetine, and sertraline, and the related antidepressant venlafaxine are contraindicated as new treatment in those under 18 years of age with depressive illness. The MHRA also advised that fluvoxamine should not be used to treat depression in the under 18s because of a lack of information on safety and efficacy. More recently, the Food and Drug Administration (FDA) has directed all antidepressant drug manufacturers to label their products with a ‘black-box’ warning about the increased risk of suicidality (suicidal thinking and behaviour).
Although there now appears to be general agreement between the MHRA and the FDA that the SSRIs and atypical antidepressants carry the risk of increasing suicidality in depressed children and adolescents, there is much less agreement as to whether or not they are efficacious as antidepressants in this context. Most recently published reviews and trials suggest that these drugs are associated with either a small additional improvement over and above placebo or have no added benefit. However, doubts about efficacy have been variously explained as the result of poor study design, poor quality of the evidence,[6*] insufficient data, differences in study methodology, or studying the wrong groups of children.
This uncertainty is further complicated by the failure to publish trials, sponsored by the pharmaceutical industry, which have negative findings. This has led to publication bias and distorting the evidence base upon which judgements about safety and efficacy are ordinarily made in clinical practice.[9**] The role of the pharmaceutical companies in withholding negative trial data has been the focus of considerable criticism, and has no doubt contributed towards the decision of the International Committee of Medical Journal Editors to now insist upon pre-trial registration of all trials as a condition of later publication of the findings.
In this article we will first briefly look at the background to the regulatory decisions, then examine the reviews and trials of SSRIs and atypical antidepressants in the treatment of childhood depression published since July 2003, with a particular focus upon the issue of efficacy.
Regulatory Background In December 2003, the MHRA, in an unprecedented move, released a detailed summary of an internal regulatory review, after earlier work had raised concerns about the safety of paroxetine and venlafaxine in children with depression.[13,14] The initial MHRA review reported data from 11 randomized, double-blind, placebo-controlled trials (two trials with fluoxetine, three with paroxetine, two with sertraline, two with citalopram, two with venlafaxine). Later, two trials with mirtazapine were added to the review. The MHRA concluded that for children and adolescents with MDD, efficacy had not been demonstrated in controlled clinical trials and that moreover the risk of suicidal behaviour and thinking was increased, suggesting unfavourable risk-benefit profiles for all the newer antidepressants except fluoxetine. In the US, the FDA undertook their own review of the SSRIs and atypical antidepressants. The FDA analysis confirmed that there was little evidence for efficacy in children and adolescents with MDD, again except for fluoxetine, but argued that the lack of evidence did not amount to positive proof of a lack of efficacy. Controversially, the FDA held back from reaching a final verdict on the basis that, in their view, the classification of suicide-related behaviours made in the original studies was confusing, making firm conclusions regarding safety unreliable. Instead, the FDA commissioned a re-classification of the original patient-level data by experts in suicidality at Columbia University.
The FDA concluded that this blinded classification process identified and corrected many misclassification errors, providing more accurate risk estimates. The data from the re-classification identified that in the MDD trials there was a 41% increase in the risk of suicidality (definitive suicidal behaviour/ideation) and a 134% increase in the risk of treatment-emergent hostility or agitation in those treated with SSRIs (not including the atypical antidepressants) when compared with placebo. The risk of suicidality increased to 71% when the atypical antidepressants were included in the metaanalysis. Importantly, the FDA also concluded that 60% of the events related to suicidality occurred in the four trials with the highest baseline proportion of patients with a history of suicide attempt/ideation. However, a subgroup analysis did not suggest a difference in risk between those with and without a history of suicide attempt or ideation. In addition, age group, gender, and premature discontinuation from the trial did not appear to act as risk modifiers.
Recent Reviews of SSRIs and Atypical Antidepressants In the past year, several reviews have examined the treatment of depression in children and adolescents. In a preliminary report published in January 2004, the American College of Neuropsychopharmacology reviewed both published and unpublished data from 15 randomized controlled trials (RCTs) (two trials with fluoxetine, two with citalopram, two with mirtazapine, two with nefazodone, three with paroxetine, two with sertraline, and two with venlafaxine), which includes all of those assessed by the MHRA. However, in contrast to the MHRA, the American College of Neuropsychopharmacology concluded that fluoxetine, sertraline, paroxetine, citalopram, and nefazodone were efficacious as antidepressants for children and adolescents. This conclusion was based on the argument that since efficacy had been demonstrated in at least one trial of each drug, the failure of other trials to confirm this could well be the result of differences in study methodology. Furthermore, the authors concluded that based on the available data from clinical trials, epidemiology and autopsy studies, there was no significant increase in the risk of suicidal behaviour in children and adolescents treated with SSRIs for depression. However, doubt has since been cast on the use of epidemiology studies that suggest a trend between increased prescribing and reduced suicide rates.[17-19]
In April 2004, Jureidini et al.[6*] reviewed data from seven published RCTs (three trials with fluoxetine, one with paroxetine, two with sertraline, and one with venlafaxine), comparing antidepressants with placebo in the treatment of depression in children and adolescents. This review raised serious doubts about the quality and accuracy of interpretation of the published trials, suggesting that claims about efficacy had been exaggerated and evidence of harm had been down-played by the trial authors. Jureidini et al. concluded that although efficacy had been demonstrated on some outcomes, the evidence did not support the view that antidepressants produced any major clinical benefit for this group.
A week later, Whittington et al.[9**] published a metaanalysis of the same 11 RCTs previously reviewed by the MHRA, comparing the published versus the unpublished data. Their analysis suggested that published efficacy data were generally more favourable than unpublished data and, when combined using metaanalysis, indicated clinically important benefit for fluoxetine only. The metaanalysis also supported the conclusions of the MHRA review regarding the increased risk of suicide-related behaviours associated with SSRIs and atypical antidepressants in this age group.
In a cost-effectiveness analysis comparing SSRIs with cognitive-behavioural therapy (CBT), Haby et al. utilized four RCTs (three with fluoxetine, one with paroxetine) that were published at the time of their analysis, and concluded that SSRIs have a lower total health benefit when compared with CBT. They also concluded, however, that the SSRIs examined were, nevertheless, cost-effective interventions as both first and second line treatment for MDD in children and adolescents. They also acknowledged the debate concerning the balance between risks and benefits of SSRIs, which arose after their research was conducted. Given that only four RCTs of fluoxetine and paroxetine were included in their analysis and the risk of harm was not taken into account, their results must be interpreted with caution.
Recent Published Trials of SSRIs Since July 2003, four randomized, placebo-controlled trials have been published to evaluate the safety and efficacy of SSRIs in children and adolescents with MDD: two trials of sertraline published as one report;[21*] one trial of citalopram;[22*] and one trial of fluoxetine and CBT.[23**] The trials of sertraline and citalopram[21*,22*] were trials incorporated into the reviews by the MHRA and the metaanalysis by Whittington et al.,[9**] both of which concluded that the risk-benefit profiles for both sertraline and citalopram were unfavourable.
In the two trials of sertraline, Wagner et al.[21*] assessed 189 patients (6-17 years old) with moderate to severe MDD and concluded that the drug produced a small but statistically significant advantage over placebo. Sertraline also produced a small but statistically significant increase in the chance of response (defined a priori as patients who had at least 40% decrease in the adjusted Children’s Depression Rating Scale-Revised (CDRS-R) total score) when compared with placebo. However, not reported by Wagner et al., but included in the MHRA review, and in the metaanalysis by Whittington et al.,[9**] were data on remission (defined as not meeting DSM-IV criteria for MDD) that suggested no benefit for sertraline above placebo. Moreover, two metaanalyses[9**,16] have suggested the potential for increased risk of suicide-related behaviour/ideation in those treated with sertraline when compared with placebo.
In the citalopram study, Wagner et al.[22*] assessed 178 patients (7-17 years old) with moderate to severe MDD, suggesting that citalopram produced a small but statistically significant advantage over placebo by the end of treatment. Citalopram also produced a small but nonsignificant increase in the chance of remission (defined as a CDRS-R score of =28). It is important to note that this mildly positive trial is one of two RCTs incorporated into a metaanalysis[9**] of previously unpublished trials of citalopram; the other, still unpublished trial was negative. Furthermore, when the data from these two trials were pooled, the risk-benefit profile of citalopram was unfavourable.[9**] The continued tendency for positive data to be published and the omission of negative trial results should maintain concern about selective reporting and publication bias.
The first stage of the largest community sampled trial, published by the Treatment for Adolescents with Depression Study Team,[23**] assessed the safety and effectiveness of fluoxetine, CBT, and their combination when compared with placebo in 439 young people (12-17 years old) with moderate to severe MDD. Response to treatment (defined as ‘much improved’ or ‘very much improved’ on the Clinical Global Impressions Improvement scale) suggested benefits favouring fluoxetine with CBT and fluoxetine alone over both CBT alone and placebo. Primary analysis of the CDRS-R indicated fluoxetine with CBT produced benefits over fluoxetine alone, CBT alone and placebo. Only a secondary analysis clearly demonstrated an advantage favouring fluoxetine alone over placebo.
Although an important addition to the data on SSRIs, the Treatment for Adolescents with Depression Study (TADS) is not without problems. Firstly, the suggestion that fluoxetine is superior to placebo in reducing depressive symptoms using a secondary analysis is undermined by the results derived from the primary analysis, which did not clearly demonstrate an advantage for fluoxetine. Secondly, the superiority of CBT combined with fluoxetine is undermined by the fact that those receiving the combined treatment were not blinded; in other words these trial participants knew they were receiving fluoxetine and not placebo, as well as CBT, leading to improvements that may well be the result of a more powerful placebo effect. Thirdly, harmful effects including suicide-related behaviour and ideation were more than twice as common in the fluoxetine group as in the placebo group, suggesting that fluoxetine may well suffer the same problems already identified in the side-effect profiles of other SSRIs.[9**,12,16] Importantly, combining CBT with fluoxetine appeared to reduce the risk of suicidality down to the level of placebo. And finally, the supposed independence of the 11 authors has recently been questioned as six of the authors have received funding from Eli Lilly and Co., the manufacturer of fluoxetine.
Conclusion There is increasing agreement that the SSRIs and atypical antidepressants are associated with an increased risk of suicide-related behaviour and ideation, at least in children and adolescents. Indeed, although fluoxetine has been thought to be free of this specific danger, data from the recent TADS study[23**] question that view. Perhaps adding to the warning from the MHRA regarding all the SSRIs except fluoxetine, we might suggest caution over the use of fluoxetine for similar reasons – a position now supported by the FDA.
There is also significant agreement that the research base for the efficacy of SSRIs and the atypical antidepressants in this age group is poor. This is partly the result of poor study designs, unsuitable recruiting methods, and mixing pre-pubertal and adolescent populations in some studies. The evidence base for efficacy is also problematic because depression is a large heterogeneous condition with considerable co-morbidity, especially in these younger age groups; and combined with a large placebo effect, valuable drug effects may not be easy to demonstrate. Indeed, the FDA have taken the view that for these and other reasons, the lack of efficacy of SSRIs and atypical antidepressants in trials should not be taken as evidence of an absent effect. However, evidence of selective reporting of drug company sponsored trials and publication bias[9**] and evidence of withholding of trial data by at least one drug company have cast doubt on this conclusion.
If pharmaceutical companies have been unable to show that any of the newer antidepressants, with the possible exception of fluoxetine, are efficacious as antidepressants for this age group, the possibility that these drugs are not in fact antidepressant should at least be considered. This, combined with the now accepted evidence for the increased risk of suicide-related behaviours in a group already at risk of suicide, should lead clinicians to consider safer alternative treatments. Nevertheless, additional independently funded research to clarify the seemingly positive interaction between fluoxetine and psychological therapies should be supported.
FAIR USE NOTICE: This may contain copyrighted (©) material the use of which has not always been specifically authorized by the copyright owner. Such material is made available for educational purposes, to advance understanding of human rights, democracy, scientific, moral, ethical, and social justice issues, etc. It is believed that this constitutes a ‘fair use’ of any such copyrighted material as provided for in Title 17 U.S.C. section 107 of the US Copyright Law. This material is distributed without profit.