Suicide by Antidepressant?

According to the New York Times ,

After analyzing data from clinical trials, GlaxoSmithKline has sent letters to doctors warning that its antidepressant drug Paxil appears to increase the risk of suicide attempts in some young adults.

We've known for a while that Selective Serotonin Reuptake Inhibitors (SSRI's) sometimes cause agitation and suicidal ideation in depressed adolescents, but this is the first time that SSRI's have been linked to the same behavior in young adults. It's important to note, however, that the study reports increased risk of suicidal ideation and suicide gestures. It did not identify an increased risk of completed suicide.

I can recall a few cases like this. Both the psychiatrists I worked with and myself were mystified. It was terrible to watch someone get good care and deteriorate so quickly. One client wound up in a state psychiatric hospital for about three months and was still quite agitated and bizarre after discharge. He was still on the same SSRI he was admitted on.

Nobody is too sure why these antidepressants, SSRI's such as Paxil, do this. The original thinking identified the psychology of depression. There are two different aspects to depression: the cognitive and the behavioral. Almost all depressed clients manifest cognitive changes, consisting of self criticism and pessimism. Behaviorally, many clients also manifest vegetative symptoms. They have little energy, they have difficulty getting out of bed, and they can't concentrate or organize their behavior.

These vegetative symptoms are actually protective. A depressed, suicidal, individual, who is also vegetative can't organize his or her behavior well enough to commit suicide. But with treatment, the vegetative symptoms may lift before the cognitive symptoms, leaving the client more capable of planning and carrying out a suicide. Every clinician, treating with drugs or psychotherapy, worries about this.

This is what we thought this was happening when clients on SSRI's became suicidal. Unfortunately, the data didn't support that. The agitation lasted too long and risk of suicide lasted even after the cognitive symptoms had improved.

I suspect that we are seeing something else. We've also known that SSRI's increase risk of mania in bipolar clients. Bipolar disorder , previously called manic-depression, consists of mood swings, from depressed (often with vegetative symptoms) to manic. The mania is marked by agitation, impulsiveness, irritability, grandiosity, and insomnia. Some forms of bipolar disorder manifest a milder form of mania, called hypomania. Hypomania, looks a lot like agitation with impulsivity. So there are three hypotheses:

First, some of these depressed clients, who react badly to SSRI's, may actually be bipolar. Bipolar disorder usually emerges in adolescence or young adulthood. It's not unusual for there to be a period where the client is diagnosed with recurrent depression. It's not until an unmistakable manic episode breaks through that we diagnose bipolar disorder. It's possible, then, that SSRI's may cause a manic or hypomanic episode in bipolar clients who who have been misdiagnosed with recurrent depression.

Second, most depression fluctuates. That is, a person who is clinically depressed may also have periods where he or she feels pretty good, or at least, not terrible. Then, there are other times when he or she feels suicidal and vegetative. I wonder if there is, at a biochemical level, some similarity between bipolar disorder and some recurrent depressions. To what extent does this similarity cause similar responses to SSRI's?

Finally, the adolescent brain still is developing the structures necessary for impulse control. I wonder, here, too, if the SSRI's differentially affect the adolescent brain. It is possible that these structures have not fully developed in the brain of the young adult, too.

In any case, until we can improve our understanding of this problem, we need to increase our monitoring of depressed patients. Physicians should never prescribe SSRI's or any other antidepressant to an unknown patient and send them away for 3 months.

Obviously, my preference would be referral for psychotherapy, but even that strategy is prone to problems. Most therapists I know today are overloaded. I often can't see someone for two or three weeks after the initial appointment. I have to schedule several appointments to keep seeing a client on a regular basis. This leaves the client unmonitored for a few weeks during the most critical period of starting medication. Fortunately, my employer has a crisis team, who, on my instructions, can call a client in crisis on a specified schedule to check on them. This study should encourage changes in practice for all practitioners.