The warning in the box. Why some psychiatric medicines carry a suicide warning, where it came from, what twenty years of science shows, and how to stay safe when treatment starts.
Open the paperwork that comes with an antidepressant, and before the dosing or the side effects you'll find a black-bordered paragraph about suicidal thoughts. It frightens people, and it's supposed to get their attention. This post covers where that box came from, which medicines carry it and which don't, what the data behind it actually show, and what I think is going on when someone feels worse right after starting treatment.
Call or text 988 (US, 24/7). Text HOME to 741741. If you're in immediate danger, call 911 or go to the nearest emergency room. And if a new medicine is making you feel much worse, call the prescriber today. Please don't wait for the next appointment.
- Not every brain medicine carries the box. It's on every antidepressant (as a class, since 2004), on antipsychotics approved to treat depression, on esketamine1, and on three ADHD medicines whose own trials showed a signal: atomoxetine, viloxazine (Qelbree) and centanafadine (Simtriyo). Stimulants, guanfacine, clonidine, lithium and benzodiazepines don't carry a suicidality box. Some have boxes for other reasons2,3,4,5.
- The signal was real but small, and it was about thoughts, not deaths. In the pediatric trials, about 4% of young people on an antidepressant reported new or worse suicidal thoughts or behaviors, versus 2% on placebo. No one died by suicide in those trials6.
- Age is the story. The same signal fades in middle age and reverses after 65, where antidepressants were protective7.
- Benefit is much larger than harm. For every 1,000 young people treated, roughly 110 more improve from depression (and even more with anxiety), while about 7 more have a suicidal event8.
- The warning had side effects of its own. After it, depression diagnoses and antidepressant prescriptions for young people fell 20–50%, and several studies found suicide attempts rose9,10. Experts still argue about cause and effect11,12.
- Is it "just happening as they get better"? Partly, but not the way people usually mean. Risk is highest before treatment starts, in the crisis that brings someone in, and it falls afterward13,14. On top of that, a small group of mostly younger people react to the medicine with restlessness and agitation15. The popular "energy comes back before mood" idea is plausible but has little direct support16.
- The practical answer is closer watching, not avoidance: check in early, know the warning signs, move carefully with dose changes, and make a safety plan.
"Will this medicine make my kid suicidal?"
I hear some version of this question almost every week, from parents, from young adults and from other clinicians. It's a fair question. The answer I give is longer than a yes or no, because the box says two true things at once. A small number of young people do have new suicidal thoughts in the first weeks of treatment. And the illness being treated is far more dangerous than the medicine. This post is my attempt to show why both are true and what to do about it.
Which medicines carry it, and which don't
A boxed warning, sometimes called a black-box warning, is the FDA's strongest label warning. Many brain medicines have one, but the boxes are about different things: abuse and addiction for stimulants17, dependence and withdrawal for benzodiazepines18, and complex sleep behaviors for zolpidem and its relatives19. The suicidality box got onto labels in two ways:
- Class labeling. Since 2004, any medicine approved to treat depression carries it, whatever its mechanism2,20. That's why an antipsychotic like aripiprazole has it (it's approved as an add-on for depression) but haloperidol doesn't21. It's also why duloxetine carries it even when it's prescribed for nerve pain.
- A drug's own data. Atomoxetine, viloxazine and centanafadine aren't antidepressants. They got the box because their own placebo-controlled trials showed more suicidal thoughts on the drug3,4,5. Montelukast, an asthma and allergy medicine, got a broader neuropsychiatric box in 2020 from reports after it went on the market22.
Tap a status to filter. Each card says why the medicine is where it is.
Where it came from
The story starts with six case reports. In 1990 Martin Teicher and colleagues at Harvard described six patients who developed intense, new suicidal preoccupation while taking fluoxetine, which was then brand-new23. Case reports can't prove cause, but the description stuck: agitation, restlessness and a new, intrusive quality to the thoughts. Over the next decade the debate simmered. Then, in 2003, it boiled over. Pediatric trial data for paroxetine reached British regulators, who advised against it for depression in under-18s and soon extended that advice to every SSRI except fluoxetine24. The concern was partly the signal itself and partly that much of the pediatric trial data had never been published. A later independent reanalysis of the best-known of those trials, Study 329, found paroxetine neither effective nor safe in the teenagers studied, the opposite of how it had first been reported25.
The FDA commissioned an outside review, re-coded every event blind (more on that below), pooled the trials and, in October 2004, put the box on every antidepressant2,26,6. Tap through the timeline to see what came after.
Filter by era, then tap any entry to open it.
What the trials actually found
The FDA's 2004 analysis pooled 24 short-term placebo-controlled trials in about 4,400 children and teenagers with depression, OCD and anxiety. Suicidal thoughts or behaviors were reported by about 4% on medication and 2% on placebo, a risk ratio of about 1.95. There were no suicide deaths in any of the trials6. Two details are often left out:
- Most "events" were thoughts, reported spontaneously. They were things patients mentioned or staff noticed, not the answers to systematic questions. When the analysts looked at the suicide items on the depression rating scales, which every patient answered at every visit, they didn't find the same signal6.
- Defining an "event" was hard. The original company reports lumped together very different things. The FDA asked a Columbia University team to reclassify every case blind to treatment. That work became the C-CASA system, which grew into the Columbia scale used in trials today26. Since 2012 the FDA has expected new psychiatric drug trials to ask about suicidal thoughts at every visit, rather than waiting for someone to bring them up27.
Passive ("I wish I weren't here") through active thoughts of ending one's life.
Steps toward an attempt, without an attempt.
Self-injury with at least some intent to die.
Hurting oneself to cope or signal distress, with no intent to die.
Events that looked alarming in a report but weren't suicidal on review.
Couldn't be classified from the records.
Then came the adults. In 2009 the FDA analyzed 372 adult trials with about 100,000 participants. The same pattern appeared in the youngest patients, faded in middle age and reversed in older adults7. That is why the box covers people up to age 24, and why the label itself says antidepressants lowered the risk in people 65 and older20.
Benefit and risk on the same page
A warning tells you about one side of the scale. In 2007 Jeffrey Bridge and colleagues pooled 27 pediatric trials and weighed both sides. For depression, about 1 in 10 young people treated got better who wouldn't have on placebo. For OCD it was about 1 in 6, and for other anxiety disorders about 1 in 3. The added risk of a suicidal thought or behavior was about 0.7 percentage points, or roughly 1 in 143, and again with no deaths8. Here are those numbers as dots.
Pick a condition. Each dot is one young person.
From Bridge 2007, 27 pediatric trials: extra responders 11.0% (depression), 19.8% (OCD), 37.1% (other anxiety disorders). The extra suicidal events are the pooled 0.7% across conditions8. These are short-term trial averages, not a prediction for any one person.
What the warning did
A warning is an intervention, and interventions have side effects. When a 2024 systematic review gathered the best-designed studies of what happened after the FDA's warnings, the findings were sobering. Doctor visits for depression among young people dropped, antidepressant use fell by 20–50%, and fewer than 5% of young patients got the close monitoring the warning had called for. Several of the included studies found suicide attempts or suicides rising in the same years9. The largest single study found that after the warning and the media coverage around it, antidepressant use fell 31% among adolescents and 24% among young adults. Over the same period, suicide attempts treated in hospitals rose, while suicide deaths didn't change measurably10. An earlier analysis found youth suicide rates rising as SSRI prescribing fell, in both the US and the Netherlands28.
These studies are observational, and they can't fully prove the warning caused the harm. People who know this literature well disagree, including in back-to-back essays in the same 2014 issue of the New England Journal of Medicine12,11.
The case for keeping the box
- The signal showed up repeatedly across drugs and trials, and the dose of harm was clearly concentrated in the young6,7.
- Industry had sat on unfavorable pediatric data. Regulators had good reason to act on what they had25.
- The studies of what happened afterward are ecological. Prescribing, coding and suicide trends all had many other causes over those years11.
- A dissenting reanalysis of the FDA's adult trial database argues suicides and attempts were more common on antidepressants than the official analysis suggested29.
The case that it backfired
- It warned about thoughts reported in short trials, with no deaths, while the untreated illness carries a real risk of death6,8.
- When researchers asked every patient systematically, drug–placebo differences in suicidal thinking largely disappeared. In adults, improvement in depression brought suicidal thoughts down with it30.
- After the warning, treatment fell sharply, monitoring barely changed, and several studies found attempts rose9,10.
- The fear spilled over: families and doctors hesitated to treat at all, not just to prescribe12.
Where nearly everyone agrees
- The first weeks of treatment, and the weeks after a dose change, deserve closer attention, especially under age 25.
- Untreated depression is the bigger danger. The goal was always safer treatment, not less treatment.
So what's actually going on?
This is the question underneath most of the ones I get: if a medicine is meant to help, why would anyone feel more suicidal on it? Is it just happening along the way, as they get better? The honest answer is that several different things add up to the same statistic, and they have very different levels of evidence behind them. The picture below shows each one on the same timeline.
An illustrative picture, not data. Pick an explanation to see where it lives on the timeline.
People start treatment at their worst. In a study of about 70,000 people beginning depression treatment, suicide attempts were most common in the month before treatment started and fell steadily afterward. The pattern was the same whether people started medication or psychotherapy13. In 2026, a within-person study of more than 830,000 young people on ADHD medicines found the same shape: the highest risk came in the 60 days before a first prescription14. Some events in the first weeks are the tail of that crisis, not something the medicine caused.
Some people, more often children than teenagers and teenagers more than adults, react to serotonergic antidepressants with restlessness, jitteriness, insomnia, irritability or impulsivity. It usually starts early or after a dose increase, and it improves when the dose is lowered15. Akathisia is an intensely uncomfortable inner restlessness, and it has long been linked to sudden suicidal thoughts31. Teicher's original six cases read a lot like this23. This is the most plausible direct drug effect, and it's treatable once someone recognizes it.
For some young people, a first depression is the opening chapter of bipolar disorder. An antidepressant can tip them into a mixed state: the despair of depression with the energy and agitation of mania. It's one of the most dangerous combinations in psychiatry. This is why a careful history, including family history, matters before starting.
The classic explanation is that motivation and energy recover before hopelessness lifts, which opens a dangerous window. It's intuitive, and clinicians have said it for decades. But when researchers went looking, they found little evidence that risk rises as patients begin to recover. The data point to risk peaking before and at the start of treatment16,13. I treat it as a reason to stay alert, not as an established fact.
People in a drug arm have more side effects, and side effects mean more calls and more conversations, which means more chances to mention suicidal thoughts that were already there. When every patient was asked the same questions at every visit, the gap mostly closed6,30.
So, to the question: is it just happening along the way as someone gets better? Partly, yes. Much of what shows up early is the tail of the crisis that brought someone in, and it tends to ease as treatment takes hold. That's not the same as "getting better causes it." And it doesn't explain everything: for a minority of mostly younger people, the medicine itself seems to stir up restlessness and agitation that can feed suicidal thinking. That group is the reason the first weeks deserve extra attention.
Why young people?
No one has a proven answer, and I'd be suspicious of anyone who claims one. What we do know: activation side effects are more common the younger the patient15. First episodes of bipolar disorder often start in the teens and early twenties, before anyone knows the diagnosis. And adolescence is a period when the brain's emotional systems are running ahead of its braking systems, so impulsive responses to distress are more likely. Any of these could make the first weeks of treatment bumpier for a 16-year-old than for a 60-year-old. None of them has been shown to be the reason.
What about ADHD medicines?
Here the picture is more reassuring than the labels might suggest. The three medicines with a box got it from small numbers in their own trials. Atomoxetine: 5 of 1,357 children on the drug (0.4%) versus 0 of 851 on placebo, with one attempt and no deaths3. Qelbree: about 0.9% versus 0.4% in children4. Simtriyo: 2 attempts in 304 children aged 6–12 versus none in 153, which is why its box applies only to that age group5.
In the 2026 within-person study, the odds of suicidal behavior were somewhat higher during treatment than during the same young people's untreated time, but much lower than in the months just before treatment began14:
More detail on each ADHD medicine's label is in my post on new ADHD medications.
How to handle the risk
The box is a reminder to be careful. It isn't a reason to avoid treatment. Here's what careful looks like in practice.
If you're starting a medicine (or your child is)
- Know who to call, and when. Before leaving the appointment, find out how to reach the prescriber quickly and what counts as "call today."
- Watch hardest in the first four to eight weeks and after any dose change. That's when activation shows up, and it's also when the original crisis is still fading.
- Name the specific signs out loud, as a family: new restlessness, agitation, trouble sleeping, irritability, panic, acting on impulses, or a mood that's suddenly worse20. The tool below sorts them by urgency.
- Make a safety plan together, especially if suicidal thoughts were part of what brought you in. It takes about 20 minutes, and you can make one here33.
- Make the home safer during the early weeks. Store or remove anything that would make a bad night more dangerous. 988 or your clinician can walk you through how.
- Don't stop suddenly without talking to the prescriber. Abrupt stops can cause their own agitation and mood swings.
Tap anything you've noticed since starting or changing a medicine. The box shows the most urgent step. This is a guide, not a diagnosis. When in doubt, call.
Tap any signs above. If nothing's changed, keep going and keep checking in.
If you're the prescriber
- Screen before you start. Ask directly about suicidal thoughts (a brief structured screen helps), and take a personal and family history for bipolar disorder.
- Schedule the first contact early. A call or portal message at one to two weeks counts. The 2004 guidance called for weekly visits in the first month; in practice fewer than 1 in 20 young patients got anything close9. Something light that actually happens beats something ideal that doesn't.
- Start low in children, and treat activation as a dosing problem before you treat it as a new diagnosis15.
- Teach the warning signs to the family, not just the patient, and give them a written "call today" list.
- Pair medication with a safety plan and, when you can, therapy. Safety planning with follow-up calls cut suicidal behavior roughly in half after emergency visits in one large study33.
- Explain the box in plain language. Families who understand what it means are much less likely to stop treatment in a panic, and much more likely to call you.
Where I land
I think the FDA saw something real in 2004. In a small number of young people, starting an antidepressant is followed by new or worse suicidal thinking, often alongside restlessness and agitation. It deserves to be taken seriously. But the box compresses a nuanced finding into a frightening paragraph, and twenty years of follow-up suggests that fear cost more than it saved. The things that keep people safe are closer contact in the first weeks, families who know what to watch for, safety plans, and treatment that actually happens. When patients ask me about the box, I tell them the truth: yes, we'll watch closely, and here's exactly what we're watching for. What worries me more is the depression we're treating.
Other medicines come up in this conversation too. Antiseizure medicines carry a class warning about suicidal thoughts, based on a 2008 FDA analysis of 199 trials (about 0.43% vs 0.24%), but it isn't boxed34. Isotretinoin's box is about birth defects; its psychiatric warnings sit lower on the label35. Varenicline's box, added in 200936, was removed in 2016 after the 8,000-person EAGLES trial37,38. In January 2026 the FDA asked for the suicidality warning to be removed from GLP-1 weight-loss medicines after reviewing 91 trials and more than 2 million patients39. And lithium, far from carrying a suicide warning, is one of the few medicines shown to lower suicide risk in mood disorders40.
Sources
- Janssen Pharmaceuticals. SPRAVATO (esketamine) nasal spray, CIII. Full prescribing information, including boxed warning for sedation, dissociation, respiratory depression, abuse and misuse, and suicidal thoughts and behaviors.
- U.S. Food and Drug Administration. FDA launches a multi-pronged strategy to strengthen safeguards for children treated with antidepressant medications (boxed warning and Medication Guide). October 15, 2004; following the joint meeting of the Psychopharmacologic Drugs and Pediatric Advisory Committees, September 13–14, 2004.
- Eli Lilly and Company. STRATTERA (atomoxetine) capsules. Prescribing information, including the boxed warning on suicidal ideation added in 2005 (pooled pediatric trials: 0.4% vs 0%).
- Supernus Pharmaceuticals. QELBREE (viloxazine extended-release capsules), for oral use. Full prescribing information. Initial U.S. approval: 2021; revised 2025 (NDA 211964/S-013). Silver Spring, MD: U.S. Food and Drug Administration.
- Otsuka America Pharmaceutical. SIMTRIYO (centanafadine) extended-release capsules, for oral use. Full prescribing information. Initial U.S. approval: 2026. NDA 218145. Silver Spring, MD: U.S. Food and Drug Administration; 2026.
- Hammad TA, Laughren T, Racoosin J. Suicidality in pediatric patients treated with antidepressant drugs. Arch Gen Psychiatry. 2006;63(3):332–339.
- 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.
- Bridge JA, Iyengar S, Salary CB, et al. Clinical response and risk for reported suicidal ideation and suicide attempts in pediatric antidepressant treatment: a meta-analysis of randomized controlled trials. JAMA. 2007;297(15):1683–1696.
- Soumerai SB, Koppel R, Naci H, et al. Intended and unintended outcomes after FDA pediatric antidepressant warnings: a systematic review. Health Aff (Millwood). 2024;43(10). (11 studies screened from 1,841 reports, 2003–2022.)
- Lu CY, Zhang F, Lakoma MD, et al. Changes in antidepressant use by young people and suicidal behavior after FDA warnings and media coverage: quasi-experimental study. BMJ. 2014;348:g3596.
- Stone MB. The FDA warning on antidepressants and suicidality — why the controversy? N Engl J Med. 2014;371(18):1668–1671.
- Friedman RA. Antidepressants' black-box warning — 10 years later. N Engl J Med. 2014;371(18):1666–1668.
- Simon GE, Savarino J. Suicide attempts among patients starting depression treatment with medications or psychotherapy. Am J Psychiatry. 2007;164(7):1029–1034.
- Adams SM, Meraz R, O'Reilly LM, et al. Pharmacotherapies for attention-deficit/hyperactivity disorder and risk of suicidal behavior: a within-individual study of stimulants, atomoxetine, and alpha-2 agonists. Biol Psychiatry Glob Open Sci. 2026;6(3):100698.
- Luft MJ, Lamy M, DelBello MP, McNamara RK, Strawn JR. Antidepressant-induced activation in children and adolescents: risk, recognition and management. Curr Probl Pediatr Adolesc Health Care. 2018;48(2):50–62.
- Mittal V, Brown WA, Shorter E. Are patients with depression at heightened risk of suicide as they begin to recover? Psychiatr Serv. 2009;60(3):384–386.
- U.S. Food and Drug Administration. FDA updating warnings to improve safe use of prescription stimulants used to treat ADHD and other conditions. Drug Safety Communication, May 11, 2023.
- U.S. Food and Drug Administration. FDA requiring Boxed Warning updated to improve safe use of benzodiazepine drug class. Drug Safety Communication, September 23, 2020.
- U.S. Food and Drug Administration. FDA adds Boxed Warning for risk of serious injuries caused by complex sleep behaviors with certain prescription insomnia medicines (eszopiclone, zaleplon, zolpidem). Drug Safety Communication, April 30, 2019.
- U.S. Food and Drug Administration. Antidepressant use in children, adolescents, and adults: revisions to product labeling, extending the boxed warning to young adults aged 18–24 and adding that depression itself is associated with suicide risk. May 2, 2007. (Class labeling and Medication Guide: "Antidepressant medicines, depression and other serious mental illnesses, and suicidal thoughts or actions.")
- Otsuka America Pharmaceutical. ABILIFY (aripiprazole). Full prescribing information, boxed warning: increased mortality in elderly patients with dementia-related psychosis; and suicidal thoughts and behaviors with antidepressant drugs. The same class boxed warning appears on other antipsychotics approved to treat depression (e.g., quetiapine extended-release, brexpiprazole, lurasidone, cariprazine, lumateperone, olanzapine–fluoxetine).
- U.S. Food and Drug Administration. FDA requires Boxed Warning about serious mental health side effects for asthma and allergy drug montelukast (Singulair); advises restricting use for allergic rhinitis. Drug Safety Communication, March 4, 2020.
- Teicher MH, Glod C, Cole JO. Emergence of intense suicidal preoccupation during fluoxetine treatment. Am J Psychiatry. 1990;147(2):207–210.
- Medicines and Healthcare products Regulatory Agency (UK). Seroxat (paroxetine) should not be used to treat depressive illness in under-18s (June 10, 2003); Committee on Safety of Medicines advice extending this to other SSRIs except fluoxetine (December 10, 2003).
- Le Noury J, Nardo JM, Healy D, et al. Restoring Study 329: efficacy and harms of paroxetine and imipramine in treatment of major depression in adolescence. BMJ. 2015;351:h4320.
- Posner K, Oquendo MA, Gould M, Stanley B, Davies M. Columbia Classification Algorithm of Suicide Assessment (C-CASA): classification of suicidal events in the FDA's pediatric suicidal risk analysis of antidepressants. Am J Psychiatry. 2007;164(7):1035–1043.
- U.S. Food and Drug Administration. Guidance for industry: Suicidal ideation and behavior — prospective assessment of occurrence in clinical trials (draft guidance, revision 1). August 2012.
- Gibbons RD, Brown CH, Hur K, et al. Early evidence on the effects of regulators' suicidality warnings on SSRI prescriptions and suicide in children and adolescents. Am J Psychiatry. 2007;164(9):1356–1363.
- Hengartner MP, Plöderl M. Newer-generation antidepressants and suicide risk in randomized controlled trials: a re-analysis of the FDA database. Psychother Psychosom. 2019;88(4):247–248. (The dissenting analysis; see Kaminski JA, Bschor T, Psychother Psychosom 2020;89(1):58–59 for the response.)
- Gibbons RD, Brown CH, Hur K, Davis JM, Mann JJ. Suicidal thoughts and behavior with antidepressant treatment: reanalysis of the randomized placebo-controlled studies of fluoxetine and venlafaxine. Arch Gen Psychiatry. 2012;69(6):580–587.
- Hansen L. A critical review of akathisia, and its possible association with suicidal behaviour. Hum Psychopharmacol. 2001;16(7):495–505.
- Zhang L, Zhu N, et al. ADHD drug treatment and risk of suicidal behaviours, substance misuse, accidental injuries, transport accidents, and criminality: emulation of target trials. BMJ. Published August 13, 2025.
- Stanley B, Brown GK. Safety Planning Intervention: a brief intervention to mitigate suicide risk. Cogn Behav Pract. 2012;19(2):256–264.
- U.S. Food and Drug Administration. Information for healthcare professionals: suicidal behavior and ideation and antiepileptic drugs (January 31, 2008; updated December 16, 2008), based on a pooled analysis of 199 placebo-controlled trials of 11 drugs in 43,892 patients.
- Isotretinoin capsules. Prescribing information: boxed warning on embryo-fetal toxicity (iPLEDGE program); Warnings and Precautions on depression, psychosis and suicidal thoughts and behaviors.
- U.S. Food and Drug Administration. Boxed warning on serious neuropsychiatric events for varenicline (Chantix) and bupropion for smoking cessation (Zyban). July 1, 2009.
- Anthenelli RM, Benowitz NL, West R, et al. Neuropsychiatric safety and efficacy of varenicline, bupropion, and nicotine patch in smokers with and without psychiatric disorders (EAGLES): a double-blind, randomised, placebo-controlled clinical trial. Lancet. 2016;387(10037):2507–2520.
- U.S. Food and Drug Administration. FDA revises description of mental health side effects of the stop-smoking medicines Chantix (varenicline) and Zyban (bupropion) to reflect clinical trial findings; boxed warning removed. Drug Safety Communication, December 16, 2016.
- U.S. Food and Drug Administration. FDA requests removal of suicidal behavior and ideation warning from glucagon-like peptide-1 receptor agonist (GLP-1 RA) medications (Saxenda, Wegovy, Zepbound), based on a meta-analysis of 91 placebo-controlled trials (107,910 patients) and a claims cohort of 2.2 million users. Drug Safety Communication, January 13, 2026.
- Cipriani A, Hawton K, Stockton S, Geddes JR. Lithium in the prevention of suicide in mood disorders: updated systematic review and meta-analysis. BMJ. 2013;346:f3646.
- National Action Alliance for Suicide Prevention. Framework for Successful Messaging. suicidepreventionmessaging.org.
Written following the Recommendations for Reporting on Suicide and the Action Alliance Framework for Successful Messaging41. Educational only. This isn't medical advice and isn't a substitute for care. Please don't start, stop or change a medicine without talking to your prescriber.