When Case Reports Really Matter:
Why we might not see suicidality go up in antidepressant studies
(revised May 2004)

(skip to table of case reports)

Suicide is associated with depression -- obviously.  Every practicing clinician I know has found it equally obvious from experience that antidepressants can decrease depression.  So, wouldn't we expect to see a lower suicide rate in patients taking antidepressants?  This may be true:  for example, there actually was a reduction in suicide rates around the time when antidepressants became widely used in children.Olfson

Suppose the overall rate of suicide really does go down when people take antidepressants.  This might not show up in research studies for two reasons:  first, suicide is fortunately not that common, so in a small study, we might not see this benefit show up because there just aren't enough people attempting suicide in the control group. 

Secondly, these research studies routinely exclude patients with suicidal thinking.  After all, many of these patients are going to get a placebo.  We wouldn't want patients with suicidal depression to run the risk of being in the placebo group.  So, the studies being examined by the FDA are all on patients who weren't suicidal to start with. 

So, we shouldn't expect the possible benefit of antidepressants medications in reducing suicide attempts to show up in these studies.  And they don't, at least in the child studies reported to the FDA.

But what if we really had big studies, and they really showed a reduction in suicide rates for people taking antidepressants; would that prove that antidepressants really don't cause people to become suicidal?  That's a crucial question, because it clearly is the approach being taken by the FDA, e.g.  in this testimony at the 2004 hearing by their Dr. Laughren, in which he acknowledges then dismisses the case reports from families presented at that hearing:

I would like to thank the families who came forward this morning to talk about their very personal stories, both the families that talked about tragic outcomes and those who talked about children who appear to have been helped by medications.

It is very hard to do that, and I think it helps us to put all of this discussion in context, but a very important point, and this has been made several times, it is very difficult to assess causality based on individual cases.  That is true both of those cases where the outcome is tragic, but also true of the cases where the outcome is good.

For either of those, we have to turn to controlled trials, so my focus is going to be on the controlled trials.

But consider this scenario:  what if antidepressants really do work, and so lower suicide attempts; and at the same time, in some individuals, they also raise risk, somehow making a subset of patients become suicidal?  What will happen when the "focus is on the controlled trials"?   Might we miss a signal of increased suicidality because the overall group of patients is actually getting better and their suicide risk thus decreasing? 

This kind of thinking is not the way science usually works.  The scenario above, a hypothesis of sorts, is a violation of a rather golden rule in science (often called "Occam's Razor"):  the simplest explanation is usually correct.  The scenario is not the simplest explanation, which in this case would be that antidepressants do not increase suicide risk, period.

So, to practice good science, we should have a very good reason to violate the golden rule of simple explanations.  What would alert us to watch out for a more complicated situation?  The answer is "case reports".  What if we hear case reports of highly unusual suicidality that came on in highly unexpected places?  And what if we hear many such reports?  Can these, in this case, serve as "evidence"? 

Wait a minute, that is a violation of an even more holy scientific rule of reasoning.  Case reports are basically personal "testimonials".  They cannot serve as firm evidence for causality, because there is no way of controlling other factors that might account for the results seen.  Skeptics who want to know why I'd consider case reports as plausible evidence in this case, yet argue against that very reasoning when it comes to some new "miracle cure" someone's trying to sell, may wish to read a further explanation of  risk and reasoning.  

For whatever reason, you've read this far, so you might want to see some of those case reports.  One of most dramatic -- Mr. Cole --  is on the main page of this discussion. You can access more by going to the FDA hearing transcript (5th bullet) and scrolling through the "public testimony" section which begins on page 79.  However, to make this easier, in the table below you'll find cases which bear particular note.

These cases were selected from those presented at the hearing because in them the history is clear, or there were very unusual behaviors. Those marked in red have some additional basis to suspect causation (see Reasoning essay) -- the medication was not given for depression; or there was testimony of a clear change in behavior associated with the medication; or there is comment on changes after the medication was stopped.  

FDA 2004 Hearing Case Reports


Starts on Page

Unusual Finding
Tom Woodward 85 His daughter hung herself after 7 days on Zoloft.  She had "akathisia" within one day of starting it.  No prior suicidal ideation had been noted.
Mark Miller 87 Son hung himself after 7 days on Zoloft.  " He became agitated on the pills.  He did not sleep.  He did not eat.  He could not sit still." 
Corey Baadsgard 90 Given Paxil, then Effexor, for social anxiety disorder:  woke up after first day of full Effexor dose in juvenile detention center.  He couldn't remember it, but he had "taken my high-powered rifle that I use for hunting to my third period class, took 23 of my classmates and 1 teacher hostage."
Joyce Story 91 17 yr old son shot a woman 5 days after starting Zoloft.  Had severe depression but this episode was diagnosed as "a manic reaction to Zoloft".  The news media called him the All-American boy.  He is now serving life without parole. 
Jame Tierney 93 Jame was given Effexor for migraines. When the dose was raised to 150 mg/day, "I thought daily about suicide and hurting myself. I felt void of normal emotions. I was so belligerent, agitated, and filled with hate - hate for my family, my friends, and most of all myself.  Rage consumed me.  I felt trapped. I said and did things I had never done before and never would do now.  I had little control and little inhibition.  It was as if I was watching a movie and some villain was destroying all the relationships around me.  I spent my time alone and viciously fighting with my parents. They would ask what was wrong and what had happened to me.  I could not answer them because I did not know or understand myself.  I was terrified.

I had never had these feelings before I took Effexor, I have never had these feelings since I stopped taking the Effexor.."

Sara Bostock 101 Her daughter took Paxil for 2 weeks during which time she "worsened", then stabbed herself twice in the chest with a kitchen knife. 
Cynthia Brockman 104 Her son took Zoloft and experienced " uncontrollable fits of anger, pitches and voices setting him off, not wanting to be touched, feeling horrible all over his body, not being in reality... When doctors stopped all his drugs, all symptoms disappeared".
Todd Shivak 107 11 yr old son given Paxil.  "It is impossible to describe the pain and utter helplessness we all felt watching Michael suffer, watch him cry, take up weapons against us, and beg us to let him die.  How do you erase the picture of your child trying to run in front of a moving car?"
Terri Williams 127 Her 14 yr. old hung himself after month on Prozac, a week at full-dose (20 mg).   "His friends would later tell me they had noticed the same behavioral change.  He also showed a verbal aggression and short temper, which had not been present before."
Glenn McIntosh 130 His 12 yr old went on Paxil for sleep problems associated with a "mild seizure disorder".  "Didn't do well" so it was stopped without taper; Zoloft started a week later.  She had increasing suicidal ideation and "akathisia" and was hospitalized.  She later hung herself (apparently still on the antidepressant, though this is not clear; time interval unspecified).
Christopher Pittman 134 14 year old writes the FDA (read at the hearing by his father):

"I would like to tell you what happened to me, what the medication did to me and how it made me feel.

"When I was taking Zoloft, I took the lives of two people that I loved more than anything, my grandparents.  I went to the doctor and he gave me a sample pack of Zoloft. He told me to take 50 milligrams once in the morning and another 50 at night.

"I didn't notice a change in my behavior until I was completely off the medication.  It made me hate everyone.  The smallest things made me blow up, and I started getting into fights, which was not me.  I would usually avoid fights.  Before the medication, I had only been in two fights my whole life.  I just hated the whole world for no apparent reason.

"A week after the doctor gave me the sample packs, he increased my dosage to 200 milligrams a day.  Everything just kept getting worse.  Then, I snapped.  I took everything out on my grandparents who I loved so very much.

"When I was lying in my bed that night, I couldn't sleep because my voice in my head kept echoing through my mind telling me to kill them until I got up, got the gun, and I went upstairs and I pulled the trigger."

Gary Cheslek 142 His college-age son hung himself 5 days after starting Effexor, after Paxil was ineffective for original problem: insomnia. "Beneath him was his laptop computer and a glass of Coke.  It was as if some sudden impulse had made him do this."
Jeff Avery 167 His 16 year old:  In the morning, asked about a boating trip, wanted to make a cake for his mother; in the evening, "looks fine", then hung himself an hour after confirming the boating trip. He was on Zoloft, prescribed after he said he was "feeling kind of down, not his usual energetic self".
Amy Coburn 174

(Age unspecified): She explained that 3 weeks after starting Paxil her father committed suicide.

"I started going to a counselor to work through my grief, and I was put on Paxil, the same drug my father was on.

"I started acting differently, then very soon after I started having suicidal thoughts, mood swings, I was fighting with my friends, and the one thing my mom noticed is that I wouldn't talk about how I was feeling.  The only thing she could get out of me was "I am fine, leave me alone."

"Six weeks after I was put on the drug, I stayed home from school, wrote my good-bye letters, and swallowed a cupful of poisonous bathroom cleaner".


Prior to this new FDA warning,  the American College of Neuropsychopharmacology, which is an organization that has a bunch of really good brains, had issued its opinion.  They were quite firm that antidepressants do not cause suicidality.   (One might think that the list of pharmaceutical company affiliations, which can be found at the bottom of the paper by scrolling right to the bottom, is rather lengthy for physicians who are going to evaluate these data and offer an opinion?).  

But in their revisiting of this issue,  the 2004 FDA warning raises the possibility -- not the certainty -- that antidepressants "cause" suicidality.  They have emphasized that we can not yet reach a conclusion on this matter, but should be watchful for this phenomenon.  They appear to have gone from a relative disinterest in this issue (1991) to a new curiosity and concern.  We should all be curious as to what is driving this change.