Intended for healthcare professionals

CCBYNC Open access

Suicidality and aggression during antidepressant treatment: systematic review and meta-analyses based on clinical study reports

BMJ 2016; 352 doi: (Published 27 January 2016) Cite this as: BMJ 2016;352:i65

Re: Suicidality and aggression during antidepressant treatment: systematic review and meta-analyses based on clinical study reports

Dubicka et al. say that it is likely to harm young people when we point out that antidepressants increase their risk of suicide. They consider it harmful that some young people will not start taking antidepressant medication “because they believe it will do more harm than good.” What is harmful is when Dubicka et al. and other psychiatrists claim that these drugs protect against suicide despite the solid evidence we have for the contrary and the clear warnings from our drug agencies. Dubicka et al. say that “It is well established that under-treating depression in children and young people is linked to suicides” and refer to an observational study. There are many such studies, but they are all flawed (1).

Dubicka et al. say that it is incorrect when we stated that antidepressants doubled the risk of suicide in children and adolescents, and they argue that there is a “hugely important difference“ between suicides and suicidal behaviour. There isn’t. A suicide starts with a thought about suicide, which leads to preparations for suicide, a suicide attempt and suicide. It should surprise no one that the risk factors for serious suicide attempts are very similar to those for suicide (2,3).

The other arguments by Dubicka et al. are also invalid. There is no need for additional “contextual information” to interpret our results; it is not relevant whether or not the drugs we studied are being recommended by NICE, as the increased risk of suicide is a class effect, and as the BMJ is read outside the UK where other drugs are being used; and the fact that the most severely depressed and suicidal young people were often excluded from the trials is also irrelevant, as the suicidal effect of these drugs is also seen in healthy volunteers (1,4,5). We found that in at least 63% of the trials we analysed (which were performed both in children and in adults), the drug companies had excluded people at risk of committing suicide. If the companies had truly believed that their drugs could reduce suicides, they would not have excluded people at risk of suicide. Companies that believe they have a drug that can reduce heart attacks would surely not study it in young people who are not at risk of getting a heart attack.

Dubicka et al. recommend close monitoring when young people are treated with antidepressants. This is a fake fix. People cannot be monitored every minute and suicides can occur very suddenly when no one expects this could happen, e.g. just after the patient appeared to be happy and well (1).

Detlef Degner, who also works at a department of psychiatry, argues that the clinical meaningfulness of statistically significant results is uncertain. So are we supposed to accept that it is not clinically meaningful that SSRIs double the suicide risk in young people? He also claims that it is uncertain whether SSRIs can lead to suicide, citing grossly unreliable observational research, e.g. a study of trends in use of antidepressants and suicides (6). This study claimed that there was a clear protective effect from the drugs although it was clear by looking at the graphs that there wasn’t!

Marc Stone from the FDA misrepresents seriously not only his own work (7) but also a paper by one of his FDA colleagues, Thomas Laughren (8). We wrote that “The FDA did not consider the limitations of the trials that we identified and introduced some of their own - for example, by only counting events within 24 hours after the randomised phase was over.” With a hair-splitting argument, Stone asserts that if a person attempted suicide before the 24 hours were over, the FDA followed up on this, e.g. if a patient died some days later. He misses the whole point, which is that withdrawal symptoms after stopping the trial drug may come later than 24 hours and may lead to suicide (1).

Using FDA trial data, Laughren found 22 suicides in 22,062 patients randomised to antidepressants, or 10 per 10 000 (8), whereas in the large FDA meta-analysis he chaired five years later, there were only 5 suicides in 52,960 patients, or 1 per 10,000 (9). Stone asserts that the 22 suicides include those that occurred in “both placebo and treated groups and during open-label extensions and studies.” Stone’s assertion is misleading. Laughren wrote in his paper: “Suicide was defined as all deaths categorized by the investigator as suicide that occurred during the short-term phase of these trials or within 30 days of stopping assigned treatment.” As the follow-up was the same for the placebo group as for the drug group, it is perfectly legitimate to perform a simple test on the suicides. There were only two suicides in 8,692 patients on placebo, which Laughren interprets thus: “There is obviously no suggestion of an excess suicide risk in placebo-treated patients.” No, there sure¬ly isn’t, but we wonder why Laughren didn’t comment on the fact that there were four times as many suicides on antidepressants as on placebo (all ages included), which difference was statistically significant (P = 0.03, our calculation). Furthermore, some of the trials that were included in the 2006 FDA meta-analysis had reported far more suicides than the five the FDA reported (1), so it is not a question of open-label follow-up as Stone claims.

It is absolutely essential to include events up to 30 days after the randomised phase of the trials is over. In clinical practice, doses are often missed, which means that the patients may increase their risk of suicide because of withdrawal symptoms. Pfizer’s trials of sertraline given to adults illustrate this. The FDA’s meta-analysis from 2006 didn’t find an increase in suicide, suicide attempt or self-harm combined (9), relative risk = 0.87 (95% confidence interval 0.31 to 2.48; FDA’s Table 30), whereas Pfizer’s own meta-analysis suggested a halving of this risk, relative risk = 0.52 (0.17 to 1.59) when only events that occurred up to 24 hours were included (10). When Pfizer included events occurring up to 30 days after the randomised phase was over, there was an increase in these suicidality events of about 50%, relative risk = 1.47 (0.77 to 2.83) (10). A 2005 meta-analysis conducted by independent researchers using UK drug regulator data found a doubling in suicide or self-harm with sertraline used in adults and when events after 24 hours were included, relative risk = 2.14 (0.96 to 4.75, our calculation) (11). Like us, these researchers noted that the companies had underreported the suicide risk in their trials, and they also found that non-fatal self-harm and suicidality were seriously underreported compared to the reported suicides. There are many other reasons why the FDA seriously underreported suicides (1).

In my view, it should be forbidden to use antidepressants in children and young people, as they say they don’t help them in the placebo-controlled trials (1), whereas they are seriously harmed by them.

1. Gøtzsche PC. Deadly psychiatry and organised denial. Copenhagen: People’s Press; 2015.
2. Beautrais AL. Suicide and serious suicide attempts in youth: a multiple-group comparison study. Am J Psychiat 2003;160:1093–9.
3. Michel K. Suicide risk factors: a comparison of suicide attempters with suicide completers. Br J Psychiatry 1987;150:78-82.
4. Healy D. Let them eat Prozac. New York: New York University Press; 2004.
5. Bielefeldt A, Danborg PB, Gøtzsche PC. Systematic review of adverse effects of antidepressants in healthy volunteer studies. Cochrane Colloquium Vienna 2015; Oct 4.
6. Gusmão R, Quintão S, McDaid D, et al. Antidepressant utilization and suicide in Europe: an ecological multi-national study. PLoS One 2013;8:e66455.
7. Stone M, Laughren T, Jones ML, et al. Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. BMJ 2009;339:b2880.
8. Laughren TP. The scientific and ethical basis for placebo-controlled trials in depression and schizophrenia: an FDA perspective. Eur Psychiatry 2001;16:418-23.
9. Laughren TP. Overview for December 13 Meeting of Psychopharmacologic Drugs Advisory Committee (PDAC). 2006 Nov 16. Available online at: (accessed 22 October 2012).
10. Vanderburg DG, Batzar E, Fogel I, et al. A pooled analysis of suicidality in double-blind, placebo-controlled studies of sertraline in adults. J Clin Psychiatry 2009;70:674-83.
11. Gunnell D, Saperia J, Ashby D. Selective serotonin reuptake inhibitors (SSRIs) and suicide in adults: meta-analysis of drug company data from placebo controlled, randomised controlled trials submitted to the MHRA’s safety review. BMJ 2005;330:385.

Competing interests: No competing interests

02 February 2016
Peter C Gøtzsche
Nordic Cochrane Centre, Rigshospitalet, Denmark