I hate to do this, because it’s old news. But then again, so is Peter Kramer’s argument in yesterday’s Times, in which he argues that the failure of antidepressants to beat placebo in clinical trials is not an indication that the drugs don’t work. Instead he blames it on the research.
Kramer’s problem is that he has to reconcile two apparently contradictory truths: that antidepressants don’t work (according to the FDA data) and that they do (according to his, and many others’, clinical experience, including mine).
The obvious , and scientific, way to deal with this conflict is to say that the evidence that they do work is anecdotal, and that the hard numbers–37 out of 74n trials failed to detect an advantage of drug over placebo, a lousy 2-Hamilton-point aggregate advantage, etc.–tell the real story. That;s the conclusion Kirsch draws and it leads him to say that even when they do work, antidepressants work by placebo effects (amplified by the side effects).
Kramer is right to say that Kirsch’s position is ideologically motivated. he wants to believe that drugs don’t work, so he listens to prozac and hears placebo, which to me simply indicates that he (Kirsch) hasn’t taken enough serotonergic drugs. But Kramer is wrong to say that the dismal results on the trials are an artifact of poor research design, at least in the way he imagines. I suppose the problem could be poor gatekeeping, as Kramer implies, and that the samp0le popuolation in the SSRI trials isn’t sick enough, but this is a pretty lame answer, especially when there’s another one that unties this knot.
It’s true that the Hamilton was standardized on hospitalized patients. it’s also true that it was never meant to be a diagnostic test. it was meant to measure improvement in a population already identified (by Hamilton himself in the old days, by something like the SCID now) with depression who were taking tricyclics (and, probably, MAOIs). The items, more than half of them neurovegetative signs and symptoms, are tailored to the drugs. But when the drugs changed, the tests didn’t. (Note that his is a purely commercial problem–the studies Kirsch looked at were conducted the way they were because they were the way to win the FDA approval game.) And it could be that the Hamilton is a a lousy tool for measuring the effects of SSRIs.
the conclusion that the SSRIs don;t work rests on the assumption that the Hamilton is both the best way to measure depression improvement and the best way to measure the effects of SSRIs; it also assumes that depression is the proper target for the drugs. I think all these assumptions are faulty. Of course, I think the whole idea of depression as a disease is vastly overrated. That’s my prejudice. But I also think the SSRIs do work, i.e., the chemical effect of tweaking serotonin metabolism is in part a change in consciousness. But it’s not one that the Hamilton is sensitive to. That’s why the drugs seem not to work. You can’t find feathers with a magnet.
to me, this outcome is the drug companies hoist on their own petard. They didn’t have to market SSRIs as antidepressants. That was a decision made for commercial reasons as much as scientific or medical and tough luck for them. Not that it matters. They got their indication and they laughed all the way to the bank.
What’s ironic about this is that the person who has done more to prove that antidepressants don’t simply relieve people’s depression is none other than Peter Kramer. But more important is the fact that this kerfuffle is in part due to a willful blindness. There are a few studies of the effects of SSRIs on personality, indicating that wahtever antidepressant effect they have is secondary to the way they change people’s personalities, but in general there’s not much incentive to do those studies. As a society, we’re pretty chary of drugs that change personality. Think about all those ads in which Pharma promises that the drugs won'[t change who you are. That would be a tough sell. It would also turn out that if what you want to do is to transform people’s personalities, there are other serotonergic drugs that do it much better and quicker (LSD, MDMA, etc.) and without daily intake, not to mention old-fashioned psychotherapy which at least doesn’t ruin your sex life.
Irving Kirsch did brilliant work but when it came to explaining the discrepancy regarding effectiveness for antidepressants, he left out a crucial piece of the puzzle — as has every other researcher into efficacy.
Coincidentally, it does have to do with the way depression is measured, mostly by multiple-choice instruments such as the HAM-D.
The problem is, these instruments cannot differentiate between the symptoms of antidepressant withdrawal and so-called depressive relapse.
The literature of antidepressant withdrawal identifies a 20%-80% occurrence rate. However, not one efficacy study involving discontinuation contains a protocol for identifying withdrawal symptoms. All study subjects displaying withdrawal symptoms were lumped together with those “relapsing,” thus bolstering statistics showing efficacy — where they should have added to statistics showing adverse effects.
It begs belief that in the entire history of antidepressant efficacy studies not one case of withdrawal syndrome occurred, yet that is what the record shows.
Going back to the beginning, none of these studies controlled for antidepressant withdrawal syndrome.
The placebo effect explains only part of the evidence that antidepressants “work.” The other part is withdrawal syndrome being reported in the wrong column — as a plus rather than a minus for these medications.
Antidepressant withdrawal syndrome is a confounding factor in statistics from all of the studies reviewed by Kirsch and others.
Very interesting. Is there time in an 8-10 week study for discontinuation effects to show up?
No doubt the withdrawal syndrome is underreported and underdiscussed. The few papers on the subject do all they can to avoid the use of “withdrawal” and “tolerance” and other words related to addiction. To use them would undermine the carefully constructed wall between antidepressants and recreational drugs. that’s bad enough, but the fact that this distinction also discourages further research into the question is even more distressing.
We don’t know what consciousness is yet so one can’t say “a change in consciousness”. This lack of clarity about mental processes that result in symptoms or distress is in part why no one takes psychiatry seriously. What process is the SSRIs changing that provides for the subjective and objective benefits?
I have not read any clinical trials, but for me, the exercise of measuring SSRI’s success rate in the treatment of depression is nothing but a commercial enterprise, and commercialization has no place in science.
To properly treat depression, one must utilize many forms of “treatment”. Talk therapy (which has multiple forms such as CBT, Psychoanalytical, Psychodynamic, Humanistic, Transpersonal…), change in diet, change in physical activity are all factors which will affect the sufferer’s improvement.
And what exactly constitutes improvement in regards to depression? Functionality? Abe Lincoln and Winston Churchill were plenty functional.