the regulators’ recordlast verified 2026-09-23

the 1991 hearing, and the thirteen years before the label changed

a one-day meeting in september 1991. families in the morning, then the agency, the manufacturer, and the national institute of mental health. at the end of it a majority of the committee said there was no clear evidence and recommended no change to the label. the boxed warning arrived thirteen years later, and it did not come from the data in that room. this page is what each side had in hand on each date, from the fda’s own account, and the part of the later finding that the “what if” version leaves out.

what each side had in hand, by date

the third column is the one the “what if” version of this story skips. a warning has to be written from something. this is what there was to write it from, on each date.

whenwhat happenedwhat the evidence was
February 1990[2]six depressed patients, not recently suicidal, develop “intense, violent suicidal preoccupation” two to seven weeks into fluoxetine. a case series from harvard, published in the american journal of psychiatry.six people, no control group
1990 to 1991[1]the fda re-reads its spontaneous-report database. the manufacturer re-analyses its trials. families of people who died begin to organise. reports to the fda rise, and the fda later says the rise tracked the publicity.reports, which count what people send in
September 1991[3]a one-day advisory committee meeting. open session: mostly family members of suicide victims. fda: the number of reports, and how the pattern followed the teicher paper and press. nimh: depression itself carries the risk, and the data did not show the drugs added to it. manufacturer: 17 pooled trials, no signal.1,765 adults on the drug in controlled trials, 569 on placebo; suicidal acts 0.3 percent vs 0.2 percent
end of that day[1]“a majority of the committee concluded that there was no clear evidence of an increased risk of suicidality in association with the use of prozac, and they did not recommend any changes to prozac labeling with regard to this issue.”the fda’s own words, written in 2004
1997 and 2001[1]congress lets the fda grant six months of extra market exclusivity to companies that run requested studies in children. the antidepressant pediatric trials get run because of this. they are the trials that produce the warning.the fda memo: the trials “were done largely in response to” the 1997 and 2001 laws
September to October 2004[6]the fda pools 24 pediatric trials of nine drugs. suicidal thinking or behaviour in about 4 percent on drug vs 2 percent on placebo. no suicides in any trial. on 15 october the fda orders a boxed warning on every antidepressant for children and adolescents.4,582 patients; risk ratio 1.95; risk difference 2 in 100; zero completed suicides
2 May 2007[7]the fda extends the warning to adults 18 to 24 in the first one to two months of treatment, and in the same notice writes into the label that adults over 24 showed no increased risk and adults 65 and older showed a decreased one.the age gradient, from 372 adult trials and 99,231 people

the case against the committee

it is real, and it is not the one usually made. the committee had a case series of six, a rising pile of reports, a room full of grieving families, and one pooled analysis written by the manufacturer’s own employees.[3] it chose the manufacturer’s analysis. it was right about the adults in that analysis, and it had no way to know about the children, because nobody had run the trials. the fair charge is not that it ignored the evidence. it is that in 1991 there was no independent trial evidence, the agency did not require any, and the question sat for six years until a marketing-exclusivity law, not a safety rule, produced the trials that answered it.[1] the warning came from an incentive to sell more, in children, for longer.

the case for the committee

the pooled trials it heard were not nothing: 1,765 adults on fluoxetine against 569 on placebo, suicidal acts at 0.3 percent and 0.2 percent, and new suicidal ideation emerging in 1.2 percent on the drug against 2.6 percent on placebo.[3] eighteen years later the fda’s own analysis of 372 trials and 99,231 adults found the same thing for adults 25 to 64: odds of suicidal behaviour or ideation slightly below placebo, and for adults 65 and older about a third of placebo.[8] the signal that exists is in people under 25.[8] a 1991 label written from a case series and a reports count would have warned every adult about a risk the adult trials, then and since, do not show. the version of the label that eventually shipped in 2007 says both halves in one paragraph: increased risk under 25 in the first months, no increase over 24, a decrease over 65.[7]

what a 1991 warning would have said

this is the test the “what if” has to pass. the 2004 box says antidepressants increase suicidal thinking and behaviour in children and adolescents, that anyone considering one must weigh that against the clinical need, and that patients started on therapy should be watched closely.[4] every clause of it comes from pediatric trial data that did not exist in 1991. strip those out and what is left for a 1991 label is the monitoring sentence: watch people closely in the first weeks. that sentence could have been written in 1991, on the strength of the case series alone, and the fda said as much when it asked for exactly that language in march 2004, for adults and children alike, before its analysis was finished and while still noting that a causal role “has not been established.”[11] so the honest “what if” is narrower than the post: not a warning that the drugs cause suicide, which the adult data did not and does not support, but a monitoring instruction thirteen years early.

what the post says, and what we could check

the post that prompted this page makes four claims about the meeting. one source we can read, the fda’s own 2004 account, covers the meeting in two paragraphs. the transcript itself we have not read. here is where each claim stands.

claimstatus
the committee voted 10 to 0 that the evidence did not support a causal linkfrom the 1991 meeting transcript, which we have not read. the fda’s 2004 memo says “a majority of the committee concluded” there was no clear evidence, which is compatible with a unanimous vote on the question as put and also with a split. we use the memo’s words until the transcript is read.
conflict-of-interest waivers for members with financial ties to antidepressant manufacturers were read at the start of the meetingwaivers are a standard opening item at fda advisory meetings and are in the transcript; we have not read this one, so we do not say who held what. the 2004 memo does not mention them.
at the time, fluoxetine had the highest number of adverse-event reports in the fda’s reporting system, per an fda surveillance officialattributed to the transcript; not read. what the fda’s own 2004 account does say is that the number of reports was “very substantial” and that the pattern of reporting “was clearly linked to the publication of the teicher, et al, paper and other publicity.” a report count in a spontaneous system measures reporting, not rate.
the committee did not listen to the familiesnot a checkable claim. the memo records that the open session was several hours long and consisted mostly of family members. what is checkable is that the committee weighed those statements against 17 controlled trials and chose the trials.

the numbers behind the warning that did come

twenty-four trials, 4,582 children and adolescents, nine drugs, up to four months each.[6] suicidal thinking or behaviour in about 4 in 100 on drug and 2 in 100 on placebo, so two extra events per hundred treated.[4] risk ratio 1.95 across all drugs and indications, 1.66 for ssris in depression trials.[6] zero completed suicides in any arm of any trial.[5] the fda put the box on every antidepressant, including the ones with no signal, because “the currently available data are not adequate to exclude any single medication.”[5] the debate about whether the box did net harm, by cutting prescribing and monitoring together, began at once and had not ended a decade later.[9] that is a different page.

questions people ask

What happened at the 1991 FDA hearing on fluoxetine and suicide?

According to the FDA’s own 2004 briefing memo, the Psychopharmacologic Drugs Advisory Committee met for one day in September 1991. Family members of suicide victims spoke in the open session. The FDA presented its spontaneous reports and showed the reporting pattern followed the Teicher paper and the publicity around it. NIMH argued that depression itself carries suicide risk and the data did not show the drugs added to it. Eli Lilly presented pooled data from 17 controlled trials showing no signal. A majority of the committee concluded there was no clear evidence of increased risk and recommended no labeling change.

Why did the FDA wait until 2004 to add the black box warning?

Because the trials that produced it did not exist in 1991. The 2004 warning rests on 24 placebo-controlled trials in children and adolescents, which the FDA memo says were run largely because of pediatric-exclusivity incentives Congress created in 1997 and renewed in 2001. In those trials suicidal thinking or behaviour occurred in about 4 percent on drug vs 2 percent on placebo, with no completed suicides. The 1991 committee had adult trials in front of it, and those showed no difference in suicidal acts and less emergence of suicidal ideation on the drug.

Does the black box warning apply to adults?

To adults under 25, since May 2007, during the first one to two months of treatment. The same FDA notice put into the label that the data did not show an increased risk in adults over 24 and showed a decreased risk in adults 65 and older. The FDA’s 2009 analysis of 372 trials found the odds of suicidal behaviour or ideation fell about 2.6 percent for every year of age.

Should I stop my antidepressant because of this history?

No. The FDA’s 2007 notice says it in one sentence: people currently prescribed antidepressants should not stop taking them, and those with concerns should tell their prescriber. This page is the history for that conversation, not a substitute for it.

sources

  1. Laughren TP. Memorandum to members of the Psychopharmacologic Drugs Advisory Committee and the Pediatric Subcommittee of the Anti-Infective Drugs Advisory Committee: background on suicidality associated with antidepressant drug treatment. US FDA, 5 January 2004 (briefing document for the 2 February 2004 meeting, reissued as Tab 2 of the September 2004 briefing package). Pages 3 to 4 describe the September 1991 PDAC meeting: statements by family members in the open session; FDA’s update on “the very substantial number of spontaneous reports” and how “the pattern of reporting was clearly linked to the publication of the Teicher, et al, paper and other publicity”; NIMH’s case that depression “itself is associated with suicidality”; Lilly’s pooled analysis “revealing no signal”; and the close: “a majority of the committee concluded that there was no clear evidence of an increased risk of suicidality in association with the use of Prozac, and they did not recommend any changes to Prozac labeling with regard to this issue.” Pages 4 to 5: the pediatric trials “were done largely in response to” the pediatric-exclusivity provisions of the FDA Modernization Act of 1997 and the Best Pharmaceuticals for Children Act of 2001; 24 studies, 9 drugs, over 4,000 patients. https://web.archive.org/web/2006/http://www.fda.gov/ohrms/dockets/ac/04/briefing/2004-4065b1-04-Tab02-Laughren-Jan5.pdf
  2. Teicher MH, Glod C, Cole JO. Emergence of intense suicidal preoccupation during fluoxetine treatment. American Journal of Psychiatry 1990;147:207-210. Six depressed patients “free of recent serious suicidal ideation developed intense, violent suicidal preoccupation after 2-7 weeks of fluoxetine treatment.” A case series, no control group. PMID 2301661. https://doi.org/10.1176/ajp.147.2.207
  3. Beasley CM Jr, Dornseif BE, Bosomworth JC, et al. Fluoxetine and suicide: a meta-analysis of controlled trials of treatment for depression. BMJ 1991;303:685-692. Pooled data from 17 double-blind trials in major depression: fluoxetine n=1,765, tricyclic n=731, placebo n=569. Suicidal acts 0.3 percent on fluoxetine, 0.2 percent on placebo, 0.4 percent on tricyclics, no significant difference. Emergence of substantial suicidal ideation 1.2 percent on fluoxetine vs 2.6 percent on placebo. Improvement in suicidal ideation 72 percent vs 55 percent on placebo. The authors were Lilly employees. PMID 1833012. https://doi.org/10.1136/bmj.303.6804.685
  4. US FDA. Public Health Advisory: Suicidality in Children and Adolescents Being Treated With Antidepressant Medications. 15 October 2004. Archived copy. Combined analysis of 24 short-term placebo-controlled trials of nine antidepressants, “over 4400 patients”; “The average risk of such events on drug was 4%, twice the placebo risk of 2%. No suicides occurred in these trials.” Directs a boxed warning on all antidepressants and a patient medication guide; “consistent with the recommendations made to the Agency at a joint meeting… on September 13-14, 2004.” http://web.archive.org/web/2005/http://www.fda.gov/cder/drug/antidepressants/SSRIPHA200410.htm
  5. US FDA. News release P04-97: FDA Launches a Multi-Pronged Strategy to Strengthen Safeguards for Children Treated With Antidepressant Medications. 15 October 2004. Archived copy. Applies the boxed warning to the entire class “because the currently available data are not adequate to exclude any single medication”; lists the nine drugs; “In these studies, there was no reported case of a suicide.” http://web.archive.org/web/2005/http://www.fda.gov/bbs/topics/news/2004/NEW01124.html
  6. Hammad TA, Laughren T, Racoosin J. Suicidality in pediatric patients treated with antidepressant drugs. Archives of General Psychiatry 2006;63:332-339. The FDA analysis behind the 2004 warning: 4,582 patients in 24 trials; risk ratio 1.95 (95% CI 1.28 to 2.98) across all drugs and indications, 1.66 (1.02 to 2.68) for SSRIs in depression trials; risk difference 0.02; “There were no completed suicides in any of these trials.” PMID 16520440. https://doi.org/10.1001/archpsyc.63.3.332
  7. US FDA. News release P07-77: FDA Proposes New Warnings About Suicidal Thinking, Behavior in Young Adults Who Take Antidepressant Medications. 2 May 2007. Archived copy. Extends the boxed warning to ages 18 to 24 “during initial treatment (generally the first one to two months)”; the labeling also states “that scientific data did not show this increased risk in adults older than 24, and that adults ages 65 and older taking antidepressants have a decreased risk of suicidality”; “depression and certain other serious psychiatric disorders are themselves the most important causes of suicide”; “People currently prescribed antidepressant medications should not stop taking them.” http://web.archive.org/web/2007/http://www.fda.gov/bbs/topics/NEWS/2007/NEW01624.html
  8. 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. 372 placebo-controlled trials, 99,231 adults. Odds ratio for suicidal behaviour or ideation: 1.62 (0.97 to 2.71) under 25; 0.79 (0.64 to 0.98) at 25 to 64; 0.37 (0.18 to 0.76) at 65 and over. The odds ratio fell 2.6 percent per year of age. PMID 19671933. https://doi.org/10.1136/bmj.b2880
  9. Friedman RA. Antidepressants’ black-box warning: 10 years later. New England Journal of Medicine 2014;371:1666-1668. A perspective piece arguing the warning had costs; cited here only as the marker that the debate did not end in 2004. PMID 25354101. https://doi.org/10.1056/NEJMp1408480
  10. The fluoxetine legal and safety record, including the pediatric approval and the litigation history, on this site. https://www.veisund.com/topics/fluoxetine-legal-and-safety-record
  11. Laughren TP. Memorandum: Overview for September 13 & 14, 2004 Meeting of Psychopharmacological Drugs Advisory Committee and Pediatric Drugs Advisory Committee. US FDA, 16 August 2004 (Tab 1 of the September 2004 briefing package). Records that at the February 2004 meeting the committees reached “a consensus… that, whether or not any of these drugs could be shown more definitively to have a role in the induction of suicidality, it would be important to remind clinicians… to be alert to the emergence of suicidality,” that FDA issued a Public Health Advisory on 22 March 2004 asking companies to add that warning language, that it “applies both to adults and children,” and that it “notes that a causal role for antidepressants in inducing such behaviors has not been established.” https://web.archive.org/web/2006/http://www.fda.gov/ohrms/dockets/ac/04/briefing/2004-4065b1-03-tab01-laughren-aug16.pdf

related: fluoxetine: the legal and safety record · more treatment, same suicide rate · who pays for the fda

adam