In response to:

Talking Back to Prozac from the December 6, 2007 issue

To the Editors:

Regarding medication’s effects on the self, Frederick Crews writes [NYR, December 6, 2007], “Peter Kramer, without ironic intent, named this phenomenon ‘cosmetic psychopharmacology.'” Without ironic intent? Here is how I introduced the concept in Listening to Prozac. I had been writing about a troubled patient I called Tess, and now I asked readers to imagine someone less depressed who might nonetheless gain social competence by taking medicine. Using psychotherapy for the same end might be acceptable, I wrote,

But I wondered whether we were ready for “cosmetic psychopharmacology.” It was my musings about whether it would be kosher to medicate a patient like Tess in the absence of depression that led me to coin the phrase. Some people might prefer pharmacologic to psychologic self-actualization. Psychic steroids for mental gymnastics, medicinal attacks on the humors, antiwallflower compound—these might be hard to resist. Since you only live once, why not do it as a blonde? Why not as a peppy blonde?

I think fair-minded readers will agree that I was trying at once to capture a phenomenon and take distance from it—that is, that I was signaling ironic intent. This attitude is fully apparent in context; the paragraph continues with concerns about tweaking personality in the guise of treating depression:

Already, it seems to me, psychiatric diagnosis had been subject to a sort of “diagnostic bracket creep”—the expansion of categories to match the scope of relevant medications.

In the same vein, Crews writes:

Even Kramer, though, felt obliged to mention certain troubling effects of Prozac that were already coming to light in the early Nineties. These included, for some users, uncontrollable tremors, diminished sexual capacity, a growing tolerance that was leading to potentially noxious higher doses, and “suicidality,” or self-destructive tendencies cropping up in the early weeks of treatment.

Again: Even Kramer? Listing those side effects in 1993, I was relying on scattered case reports, along with my own and trusted colleagues’ clinical experience, and so was out ahead of the standard literature about the new antidepressants—hardly the posture of a reluctant critic. I noted these negative outcomes despite a warning posted in the book’s introduction. There, I told readers that since I would focus on theoretical issues related to the ethics of medication and personality change, I would give extensive attention neither to drug side effects nor to positive effects in the treatment of major mental illness.

Many of the concerns Crews shares, regarding the impulse to use drugs to enhance ordinary mood, were broached in Listening to Prozac. Later, I noticed that people were taking my worries about cosmetic psychopharmacology and applying them—in my view, mistakenly—to conventional uses of medication; it was this observation that led me to write Against Depression.

Peter D. Kramer

Clinical Professor of Psychiatry and Human Behavior

Brown University Medical School

Providence, Rhode Island

To the Editors:

Frederick Crews argues not only that depression severe enough to justify drug treatment is much rarer than the drug industry would have us believe but also that the serotonin deprivation hypothesis of depression has little to recommend it. The drug industry disagrees and would, no doubt, be delighted if everyone who had ever felt sad took antidepressant selective serotonin reuptake inhibitors (SSRIs) for the rest of their lives.

The truth lies between these positions. Enormous numbers of unrecognized severe depressives never receive psychiatric attention. These are mostly males, Thoreau’s “mass of men [leading] lives of quiet desperation.” As for the serotonin deprivation hypothesis: Crews’s opinion that this is like explaining headaches by aspirin deprivation lacks substance. Aspirin is only detectable in people who have taken aspirin. Serotonin is a normal constituent of every human brain where it is released from certain neuronal tracts to various sites of action. Evidence accumulating over half a century indicates that serotonin has a causal role in depressive illness and that the antidepressant properties of SSRIs are mediated by increasing its availability to specific brain sites (see reviews in Interactive Monoaminergic Disorders, edited by Paloma et al., 1999, by Cowen, Curzon, and Van Praag).

This does not imply that the brain is a kind of slot machine delivering happiness if serotonin is inserted, or as one British newspaper proclaimed when our royal family was undergoing multiple emotional crises, “Fergie and Di. It was the serotonin. The difference between happiness and misery.” Neither does it imply that all depressives respond to SSRIs, nor that SSRIs only act on sites mediating their therapeutic effect. Therapeutic benefit only occurs after the first few weeks of treatment during which initial period other sites become less responsive. During this initial period, some subjects experience harmful effects of the drugs which can be serious in a small fraction of patients.

Crews’s article raises important issues, e.g., the power of the drug companies, especially of their sales departments, and the temptations inherent in the high cost of drug development combined with the tremendous profits that can be gained from a few of the small fraction of drugs under development that reach the market. However, he is overpolemical and his cherry-picking for negative findings from a large and often confusing literature may produce more heat than light.


Gerald Curzon

Emeritus Professor of Neurochemistry

University of London

To the Editors:

Frederick Crews makes much of the notion that suicidality is a demonstrated “troubling effect” of antidepressants. He speaks of “SSRI-induced stabbings, shootings, and self-hangings” and claims that makers of these medications have “treated the victims of their recklessness with contempt.” But to speak of “effects,” “induced” behaviors, and “victims” here is to greatly overstep. There is some correlation between SSRIs and suicidality in the early weeks of treatment, but cause is different from correlation, as Crews takes pains to point out elsewhere in the review. To borrow his own analogy for conflating the two, saying that some suicides are caused by SSRIs is like saying that “headaches arise from aspirin deprivation.”

Pfizer’s failure to own up promptly to the correlation between Prozac and the incidence of suicide attempts is typical and lamentable behavior for a pharmaceutical company; that hardly needs saying. But to read Crews, one would think Big Pharma suppressed some magic data that proved their drugs alone can cause suicide. Psychiatric experts have attributed suicide attempts among SSRI patients to a combination of a preexisting low self-regard with the energy boost provided by the medication. I’m not qualified to definitely assess that hypothesis (perhaps no one is), but suicide seems far more likely to be in part a reflection of a natural mood rather than a straightforward product of medication alone, as talk of “SSRI-induced…self-hangings” strongly implies. We should remember that any patients who killed themselves on SSRIs were taking them because they were already deemed depressed.

Crews also writes, in laudably criticizing marketing practices, that “the pharmaceutical companies haven’t so much answered a need”—really?—“as turbocharged it.” I should think the importance of their doing the first outweighs by far the significance of the second.

Crews subscribes to the view that depression is overdiagnosed and antidepressants overprescribed, and that psychiatrists too often write prescriptions without an adequate familiarity with the life of the patient. The available evidence suggests that these beliefs are true, as is the notion that SSRIs have been tied to suicide. But all those claims are compatible with the widely recognized fact that antidepressants have helped a great many people and that, as John Michael Bostwick of the Mayo Clinic has written,

several recent, large nonindustry studies indicated that rates of suicide and suicidal behavior were actually reduced in children who used antidepressants, despite piteous anecdotal tales.

If Pfizer ought to have admitted to correlations inconvenient to them, a critic of Pfizer, already picking an easy target, ought to admit to these highly commendable correlations too, inconvenient as they might be to him.

Evan Hughes

New York City

Frederick Crews replies:

I stand corrected on the matter of Peter Kramer’s tone when introducing the term “cosmetic psychopharmacology.” As a social meditation, Listening to Prozac meant to identify and analyze a trend, not to endorse it. Unfortunately, hundreds of thousands of readers overlooked Kramer’s misgivings and rushed out to acquire Prozac so that they, too, like the small sample of patients described in the book, could become “better than well.” Psychologically healthy people thus exposed themselves to serious risks for a benefit that now appears to have been overrated.

Gerald Curzon’s point about an underdiagnosed and undertreated class of depressives is well taken. It doesn’t weigh against the historical fact that depression itself, as a disease entity, was vastly promoted by the companies that saw a market in it for their pills.

As a layman, I make no claim to familiarity with brain chemistry or medicine. I can only read what experts have written and compare it to the simplifications circulated by hucksters. Three concurring books that I cited emboldened me to dispute one such simplification: that depression is well known to be caused simply by a shortfall of serotonin in the brain. Dr. Curzon’s demurral from the journalistic “Fergie and Di” assertion would appear to put us in agreement on that point—but possibly not.

The physiological and medical facts that Dr. Curzon mentions, including the presence of serotonin in everyone’s brain and the therapeutic effect that extra serotonin can produce, are not in dispute. They certainly offer important clues for research. But they do not tell us whether a serotonin deficit causes depression or, rather, is itself a byproduct of other influences. The fact that a given agent, even a naturally occurring one, relieves a disorder doesn’t prove how that disorder came about. If Dr. Curzon believes otherwise, that is where we disagree.


It is entirely possible, as Evan Hughes suggests, that most suicide attempts and other destructive acts by a small minority of SSRI takers have resulted from an “energy boost” provided by the drug in combination with preexisting tendencies. Since, in such cases, the drug isn’t the only causative factor, Hughes declares that it hasn’t induced the behavior. This is, to put it mildly, a strange approach to suicide prevention. If all that’s required to send a depressed patient over the edge is a chemically supplied energy boost, shouldn’t a physician be cautious about prescribing it?

Hughes elsewhere shifts his ground and reverts to the drug companies’ early, now abandoned, claim that nothing more than a loose correlation, devoid of causal significance, links SSRIs to suicidality; depression itself, he says, is responsible for bad outcomes. As I related, that argument began to fail when word leaked out that, in clinical trials that Pfizer withheld from the FDA and again in a test by David Healy, undepressed patients experienced the same symptoms that had been blamed on depression. Moreover, Hughes ignores Eli Lilly’s eventual statement, in a later patent application, that its own Prozac can bring on “suicidal thoughts and self-mutilation.”

My article didn’t deny, as Hughes insinuates, that “antidepressants have helped a great many people.” Nor did it claim that SSRIs have induced more suicides among young people than they have prevented. On the contrary, I cited a recent article propounding the opposite conclusion—the same finding now advanced against me by Hughes. And I cited still another article indicating that genetic tests may someday render SSRIs safer by showing which patients would poorly tolerate a given drug. Neither of those references could have appeared in an article denying any utility to SSRIs.

Detailed evidence that SSRI makers have “treated the victims of their recklessness with contempt” can be found in Healy’s Let Them Eat Prozac, which Hughes shows no sign of having read.

This Issue

February 14, 2008