The data supporting the use of “antidepressants” in children and adolescents is largely unavailable. Academic publications give a different picture as regards benefits and harms to publications from regulatory other sources. Despite disagreements about the data driving use of these medicines, in practice “antidepressants” may now be the most commonly used drugs by adolescent girls, and children’s mental health services are attracting increasing attention.
This paper reviews the difficulties surrounding the data. It outlines a case for benefits (as well as risks) that would require physicians to exert a greater degree of professional autonomy than service managers might wish.
In 2016, Britain’s Health Minister stated children’s mental health services are the greatest failing of the National Health Service. Press coverage of his statement focussed on depression and noted that concerns were being raised even though despite substantial increases in funding children were attempting suicide while on waiting lists .
The selective serotonin reuptake inhibitors (SSRIs) were developed by Carlsson to allow the differences between catecholamine (energy enhancing) and serotonergic (cognitive/ anxiolytic) therapeutic principle to be distinguished in a way that was not possible with tricyclic antidepressants. At the launch of the SSRIs in the 1990s, there had been 15 randomized controlled trials (RCTs) of tricyclic and related antidepressants in children, all negative .
In 1997, an NIMH funded trial of fluoxetine in adolescent depression reported a benefit , as did a second RCT . These trials and a trial in OCD led regulators to approve claims fluoxetine could benefit adolescent depression and OCD. Subsequently trials for paroxetine  and sertraline , and citalopram reported benefits . Sertraline was also approved for OCD.
In 2004, doubts were raised about the benefits and safety of these drugs for children. The United States Food and Drug Administration (FDA) identified 15 paediatric antidepressant trials and designated the two fluoxetine studies and one citalopram study as positive and all paroxetine, sertraline, venlafaxine, nefazodone and mirtazapine trials as negative.
The FDA analysis of these trials reported a doubling of suicidal events (absolute increase of 1% ) on active treatment, and this led to a Black Box warning, as much perhaps because of the background lack of efficacy as for the data on suicidal events.
A recent analysis  of paediatric depressive and anxiety disorder RCTs, calculating effect sizes using scores on rating scales with continuous variables such as the Childrens’ Depression Rating Scale (CDRS), or CYBOCS, suggested there was evidence of a minor benefit in depressive and anxiety disorders.
We have categorised all depression trials undertaken as positive or negative on the basis of the findings on the a priori declared primary outcomes using Clinical Study Reports (CSRs), an FDA review of the issue in 2004 , FDA reviews of the fluoxetine paediatric application , and company reports on clinicaltrials.gov.
Taking this approach, all 20 trials conducted in children with depressive disorders between 1990 and 2006, almost 4000 children, were negative (see Table 1).
|329||N||Le Noury et al.||XS||275|
|511||N||Braconnier et al.||XS||125|
|1017||N||Laughren||XS||(1001 and 1017)|
|1990||N||Simeon et al.||XS||40|
|TADS||N||March et al.||XS (34v3)||439|
|Mandoki||N||Mandoki et al.||XS||40|
|Cit 18||N||Jureidini et al.||?||174|
Legend: N = negative; OL = open label; XS = excess.
The primary outcomes for the two fluoxetine trials commonly cited as positive, which provided the basis for approval of the drug, were negative . In seven trials in which fluoxetine has been compared to placebo, alone or in a design with duloxetine or venlafaxine, it has been negative on the primary outcome.
The citalopram trial claimed as positive  is not positive on its primary outcome if the results are analysed per protocol . A recent (es)-citalopram study report claims a benefit on the primary outcome measure on the basis of a 3.3 point difference between citalopram and placebo in the reduction from their respective baselines . This is not a clinically meaningful difference in a scale with a range of 96 points.
The data from these sources make it clear that there was an excess of suicidality on active treatment in all trials.
The present authors reanalysed Study 329 in which paroxetine was claimed to be effective and safe  and found that no method of analysis produced a positive result on the primary outcome . We also found more than double the number of suicidal events compared with the originally reported claims.
There are no grounds for thinking other paediatric antidepressant studies are exempt from the findings and lessons from the restoration of Study 329 and the re-examination of CIT-18, namely that when analysed strictly per protocol the basis for efficacy in depressive disorders claims for benefits shrink and the number of recorded harms grows.
Since 2006 there have been a further 15 depression studies involving over 6000 patients (Table 2). These have mainly involved duloxetine, (des)-venlafaxine, (levo-) milnacipran and vortioxetine, but have also included one (es)-citalopram and one paroxetine study. When considered in terms of primary outcome measures these depression studies have also been negative, with the possible exception of the (es)-citalopram study mentioned above.
Legend: N = negative; OL = open label; XS = excess.
While the depression trials were ongoing, a series of trials of the same drugs in OCD and other anxiety states were undertaken. A proportion of these trials appear positive on their primary outcome measures. Some studies with negative outcomes on the primary outcome measure remain unpublished. There is an excess of suicidal events on active treatment in these trials also.
3Use of antidepressants
The state of the evidence for the use of antidepressants in paediatric depression as of 2004 illustrates the greatest known divide between a set of Open Label studies claiming benefits (N = 70) and RCTs pointing consistently in the opposite direction. Given the RCT findings, one might have expected the then growing use of antidepressants in minors to have declined. But in the decade since 2004, while establishing the details of use is difficult as prescription data is only available on a commercial basis, it appears antidepressants are now among the most commonly prescribed drugs in adolescents, particularly in girls. There has been a 100-fold increase in Britain, and a recent CDC publication places the use of antidepressants among American adolescents potentially at 13% .
This increase in use likely stems from the positive view of efficacy in the published literature, along with an absence of harms. Even hints of efficacy will lead to prescriptions, when despite warnings in this age group, the academic literature makes little reference to harms or is dismissive of this possibility .
For instance, March and colleagues in the 2004 TADS study concluded: “The combination of fluoxetine with CBT offered the most favorable trade-off between benefit and risk for adolescents with major depressive disorder”. The TADS study gave rise to seven publications in major journals, none of which drew attention to the occurrence of 34 suicidal events on fluoxetine compared to 3 on placebo .
Challenge-dechallenge-rechallenge reports from the early 1990s in children as well as adults including a completed suicide in a child with OCD established conclusively that these drugs can cause suicide and the clinical trial data shows more children progress to a suicidal event than are prevented from a suicidal event in these 6–8 week trials.
There are other problems. All SSRIs affect QT intervals and are increasingly combined with other psychotropic agents which have effects on QT intervals .
In close to 100% of individuals who take an SSRI there is an immediate genital numbing. In a proportion of cases this endures semi-permanently after treatment stops, for which at present we have no remedies .
We do not know what proportion of adolescents who take antidepressants are unable to get off them, but current data suggests over 80% of antidepressant takers of all ages have been on treatment for more than a year. A significant proportion of these are likely to be continuing treatment because of difficulties in stopping.
Dependence is problematic in adolescents as SSRIs inhibit growth velocity, and cause weight gain, which if effectively permanent because of dependence may be problematic for self-image, in addition to increasing the risk of diabetes and other disorders.
Women enter the child-bearing years in adolescence. There is a significant evidence that SSRIs cause birth defects  and Autistic Spectrum Disorders  in addition to triggering miscarriages which are a risk factor for future mental health problems.
If primary care treatment goes wrong, subjects presenting to secondary services with complex pictures involving suicidality or disinhibition or antidepressant induced alcoholism, may end up diagnosed as bipolar and given inappropriate treatments often concomitantly.
Statements to the effect that the risk-benefit ratio of treatment remains favourable on the basis of clinical trials that last for only 6–8 weeks and fail to record the adverse effects of treatment properly, and where the data is inaccessible are problematic, particularly in the absence of evidence of efficacy. The notion of a risk benefit ratio for any drug is recent. Before 1990 the medical brief was to balance the harms from a drug against the harms of the condition.
4Deriving a benefit
There is another way to think about the clinical utility of these drugs. This lies in a return to the thinking that gave rise to SSRIs, which was an effort to make a specific therapeutic effect more visible. This therapeutic principle appears to be a form of anxiolysis referred to by many on treatment as emotional numbing, and at one point marketed as a serenic effect. This effect can be distinguished from the effects of antipsychotics, benzodiazepines and beta-blockers, which might also be termed anxiolytics.
It would seem highly likely that in appropriately designed trials, perhaps lasting no more than a week, there would be a clear and relatively immediate distinction between SSRIs and placebo on an outcome measure like a serenic effect.
No such trials have been done but recently Eriksson and colleagues have meta-analysed SSRI trials focusing on the depression item from the Hamilton Depression Rating Scale, on which they claim a significant change on medication compared to placebo .
This finding fits the view offered here in that one would expect an anxiolytic treatment to produce a more distinct difference on an item of the HDRS such as the depression or anxiety item than it does on the overall clinical condition. The view is also supported by the evidence these drugs produce a more distinct benefit in OCD and other anxiety disorders that in paediatric “depressive disorders”.
While there would be problems to overcome in acknowledging that a serenic effect is the therapeutic target of serotonin reuptake inhibitors, there is also much to gain. Designating a serenic effect as the primary outcome not just of trials but of therapy would help make clinical practice more rational. A patient, told that this is the effect desired, could contribute to care by reporting whether the treatment was having this effect. Where the desired effect is not happening, there would be an earlier opportunity to switch to a different therapeutic principle, reducing the likelihood of adverse reactions linked to the agitating effects these drugs can have in the short term and weight gain and other effects in the longer run.
If producing a serenic effect was the primary outcome of trials, the differences between the drugs and placebo would likely make claims that the effect of these drugs are all in the mind or the drugs don’t work redundant. The key point for debate and research instead would centre on establishing when it is reasonable to deploy such a therapeutic principle – in which age groups and for how long. The key point in clinical practice would centre on professional discretion rather than guideline or even indication mandated treatment approaches.
Conflict of interest
DH and JJ have been expert witnesses for plaintiffs in legal cases involving SSRI harms. JLN has no conflicting interests to declare.There is no external funding for this article from any source.
Statement re access
All materials are in the public domain but we will provide anything readers find difficult to access
Braconnier A , Le Coent R , Cohen D . Paroxetine v clomipramine in adolescents with severe major depression. Journal of American Academy of Child and Adolescent Psychopharmacology. 2006;42:22–9.
Campbell D . Jeremy Hunt says children’s mental health services are the NHS’ biggest failing. The Guardian October 20, 2016, https://www.theguardian.com/society/2016/oct/20/jeremy-hunt-promises-better-mental-health-services-children-adolescents
Emslie GJ , Rush AJ , Weinberg AW , Kowatch RA , Hughes CW , Carmody T . A double-blind, randomized placebo controlled trial of fluoxetine in depressed children and adolescents. Archives of General Psychiatry. 1997;54:1031–7.
Emslie GJ , Heligenstein JH , Wagner KD , Hoog SL , Ernest DE , Brown E , Nilsson M , Jacobson JG . Fluoxetine for acute treatment in children and adolescents: A placebo-controlled randomized clinical trial. J Am Acad Child Adoles Psychiatry. 2002;41:1205–15.
Emslie GJ , Ventura D , Korotzer A , Tourkodimitris S . Escitalopram in the treatment of adolescent depression: A randomized placebo-controlled multisite trial. J Am Acad Child Adolesc Psychiatry. 2009;48:721–9.
Healy D , Howe G , Mangin D . Sudden Cardiac Death and the Reverse Dodo Verdict. International J Risk & Safety in Medicine. 2014;26:71–9.
Healy D , Mangin D , Mintzes B . The ethics of randomized placebo controlled trials of antidepressants with pregnant women. Internat J of Risk and Safety in Medicine. 2010;22:7–16. DOI: 10.3233/JRS-2010-0487
Healy D , Le Noury J , Mangin D . Links between Serotonin Reuptake Inhibition Pregnancy & Neurodevelopmental Delay Spectrum Disorders: A Review of Epidemiological and Physiological Evidence. Int J Risk & Safety in Medicine. 2016;28:125–41. DOI: 10.3233/JRS-160726
Hieronymus F , Emilsson JF , Nilsson S , Eriksson E . Consistent superiority of selective serotonin reuptake inhibitors over placebo in reducing depressed mood in patients with major depression. Molecular Psychiatry. 2016;21:523–30.
Hogan C , Le Noury J , Healy D , Mangin D . One hundred and twenty cases of enduring sexual dysfunction following treatment. International J Risk & Safety in Medicine. 2014;26:109–16.
Hogberg G , Antonuccio D , Healy D . Suicidal Risk from TADS Study Was Higher than it First Appeared. Int J Risk & Safety in Medicine. 2015;27:85–91.
Jensen PS , Ryan ND , Prien R . Psychopharmacology of Child and Adolescent Major Depression: Present status and future directions. J Am Acad of Child and Adolescent Psychiatry. 1992;2:31–46.
Jureidini JN , Amsterdam JD , McHenry L . The citalopram CIT-MD-18 pediatric depression trial: Deconstruction of medical ghostwriting, data mischaracterization and academic malfeasance. International J of Risk & Safety in Medicine. 2016;28:33–43.
Keller MD , Ryan ND , Strober M , Klein RG , Kutcher SP , Birmaher B , Hagino OR , Koplewicz H , Carlsson GA , Clarke GN , Emslie GJ , Feinberg D , Geller B , Kusumakar V , Papatheodorou G , Sack WH , Sweeney M , Wagner KD , Weller E , Winters NC , Oakes R , McCafferty JP . Efficacy of paroxetine in the treatment of adolescent major depression: A randomized, controlled trial. J Am Acad Child Adolesc Psychiatry. 2001;40:762–72.
Laughren T . Memo: Background on suicidality associated with antidepressant drug treatment. CDER, Jan 5 2004.
Le Noury J , Nardo J , Jureidini J , Healy D , Raven M , Aba-Jaoude E , Tufanaru C . Restoring Study A randomised, controlled trial of the efficacy and harms of paroxetine and imipramine in the treatment of adolescent major depression. BMJ. 2015;351:h4320. DOI: 10.113 6/bmj.h4320
Locher C , Koechlin H , Zion SR , Werner C , Pine DS , Kirsch I , Kessler RC , Kossowsky J . Efficacy and Safety of Selective Serotonin Reuptake Inhibitors, Serotonin-Norepinephrine Reuptake Inhibitors, and Placebo for Common Psychiatric Disorders Among Children and Adolescents. A Systematic Review and Meta-analysis. . JAMA Psychiatry. 2017. doi: 10.1001/jamapsychiatry.2017.2432, Published online August 30, 2017.
March J , Silva S , Petrycki S . Treatment for Adolescents with Depression Study (TADS) Team. Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents with Depression Study (TADS) randomized controlled trial. JAMA. 2004;292:807–20.
Mandoki MW , Tapla MR , Tapla MA , Sumner GS , Parker JL . Venlafaxine in the treatment of children and adolescents with major depression. Psychopharmacology Bull. 1997;33:149–54.
Pratt LA , Brody DJ , Gu Q . Antidepressant Use Among Persons Aged 12 and Over: United States, 2011–2014. NCHS Data Brief No. 283, August 2017.
Mosholder A . NDA 18-936. Review and evaluation of clinical data. CDER. June 25th 2001.
Simeon JG , Dinicola VF , Ferguson HB , Copping W . Adolescent depression: A placebo-controlled fluoxetine treatment study and follow-up. Prog Neuropsychopharmacol Biol Psychiatry. 1990;14:791–5.
Wagner KD , Ambrosini P , Rynn M , Wohlberg C , Yang R , Greenbaum MS , Childress A , Donnelly C , Deas D . Efficacy of sertraline in the treatment of children and adolescents with major depressive disorder: Two randomized controlled trials. JAMA. 2003;290:1033–41.
Wagner KD , Robb AS , Findling RL , Jin J , Gutierrez MM , Heydorn WE . A randomized, placebo-controlled trial of citalopram for the treatment of major depression in children and adolescents. Am J Psychiatry. 2004;161:1079–83.