Healy pp001-102

Introduction
The disease that has on several occasions nearly killed me does killtens of thousands of people every year: most are young, most die un-necessarily, and many are among the most imaginative and gifted thatwe as a society have.1 F E W C O U L D R E A D Kay Redfield Jamison’s 1995 book Unquiet Mindand fail to agree that manic-depressive disease is debilitating and dan-gerous. For skeptics, the underlying statistics and science are laid out inher monograph with Fred Goodwin, then director of the National Insti-tutes for Mental Health (NIMH), Manic-Depressive Illness.2 Any remain-ing skeptics would be likely to recant on hearing Jamison or other in-ternational figures talk in the flesh about their illness. But for romantics,the personal testimonies of experts like Jamison can cause problems be-cause these experts endorse treatments such as electroconvulsive ther-apy (ECT), lithium, and psychotropic drugs. To claim to be the survivorof a psychiatric treatment makes for glamorous television material, butJamison and others talk about being the survivors of a psychiatric ill-ness—one that wrecks families and careers and claims lives.
The passion of Jamison and others has brought new recognition for sufferers of mood disorders and has ensured funding for research thatsomeday will make a difference. Yet there is a fine line to tread. Just ascorporate culture can incorporate Levi’s jeans and counterculturallifestyles, so also the message of someone like Jamison can be corruptedand used to sell treatments to people who should not have them—andwhose lives may be put at risk by them.
As recently as 1990, there was seemingly little danger that depres- sion could be corporatized. But about the time the Internet began to be I N T RO D U C T I O N : B E F O R E P RO Z AC used frequently in universities, and exam results began to prove con-clusively that women on average did better academically than men,things changed. Readers who remember what it was like to live in aworld without the Internet will not so easily remember what it was liketo live in a world where few, if any, risked getting depression. Yet thiswas so only a few years ago—before Prozac.
Magazines, newspapers, and television programs are now full of stories about how extraordinarily common and seriously disabling de-pression is.3 It is billed as one of the greatest and most common afflic-tions of mankind, second only to heart disease as a cause of disability inthe world. This seems to mesh with Jamison’s message that this is a se-rious disease. But the story of depression hides at least three disturbingsurprises.
First, unlike manic-depressive illness, depression was all but un- recognized before the antidepressants; only about fifty to one hundredpeople per million were thought to suffer from what was then melan-cholia. Current estimates put that figure at one hundred thousand af-fected people per million.4 This is a thousandfold increase, despite theavailability of treatments supposed to cure this terrible affliction.
Second, the discovery of antidepressants was not recognized at the time as a medical breakthrough. Indeed, they were almost an embar-rassment to the companies who found them.
The third surprise is that it was recognized from the discovery of the very first antidepressants that they might lead to suicide.
All three surprises come together in the story of Caitlin Hurcombe, The policeman, eyes averted, handed me her suicide note. There was a damp teardrop directly blurring the word “love.” I thought,“Your tear hasn’t dried, Caitlin, how can you be dead?” It was a Monday in April 1998 when my 19-year-old daughter bal- anced herself on a piano stool, covered her head with a pillowcase andhanged herself with her pony’s lunge rope from a beam in the bed-room. The executioner in her head allowed her no room for error.
Since her devastating death I have become an expert on suicide, on bullying, on teenagers and depression, young people and violence,on suicide prevention. Searching for a “why.” All bereaved parentswho have lost a loved one to suicide will experience this to one de-gree or another, and the common wisdom is to find a way to “let go.” I N T RO D U C T I O N : B E F O R E P RO Z AC Letting go of what is both irreplaceable and unforgettable is rather atall order.
Caitlin was prescribed Prozac a few weeks before she killed herself.
The inquest revealed that there was no drug in her system other than atherapeutic dose of Prozac. We had gone through a very rackety Christ-mas during which Caitlin learned, not for the first time, what rottersboy friends can be. Many of her friends recommended that she tryProzac for her unhappiness, not least because “you lose weight on it.” As an American expat I have too many times been duped myself by the instant cures promised by popping pills. An honest description ofmy feelings about Caitlin taking Prozac would be “gently skeptical.” Iread her an alternative medical pamphlet on the virtues of St. John’sWort over Prozac. She still wanted the Prozac and she got it. Howcould millions of people the world over be wrong? Danger wasn’t anissue.
Her doctor said not to worry if she didn’t feel better immediately, because the drug took sometimes two weeks or more to “kick in.” Thechange in Caitlin’s behavior, however, was immediate and dramatic.
She danced into the kitchen after college after taking pill number one,saying she felt great. But her elation was short-lived. Her behaviorturned bizarre. Break-ins at college, fierce arguments with her tutor.
Ferocious poetry. Stealing a stereo from the house of a local man notcoincidentally called “psycho,” and saying she didn’t know why.
Nightmares of murdering me and of her own death. And finally on thelast weekend of her life, a wild distracted skipping of college, self-mu-tilation, and a drunken and sordid episode in a pub.
Caitlin’s life is no less precious because she gave it up. And Prozac is no less dangerous to young people just because millions swear by it.
Among the hundreds of treasured letters and cards of condolence, somany said, “It’s not your fault!” “Depression is a physical disease!”When first she died we tortured ourselves wondering how we couldhave overlooked her suffering. We thought she was just very good atmasking her pain. Caitlin’s counsellor concluded she was “normal,”“psychologically intact.” So now I know that she was a normal angst-ridden and unhappy girl-woman, finding life hard to cope with. AndI believe the Prozac was her executioner.5 Opinions will divide on whether Caitlin Hurcombe committed suicidebecause of Prozac or because she was depressed. After all, Kay Jamison I N T RO D U C T I O N : B E F O R E P RO Z AC could easily have committed suicide before her condition was recog-nized. But for those who have no doubt it was the illness that caused thesuicide, the story of the origins of the antidepressants offers a disturb-ing fact. Forty-five years before Caitlin Hurcombe hanged herself, itwas known that antidepressants could produce suicides in people notdepressed. In fact, we would not have had Prozac had this not been thecase.
During the first eighty years of the twentieth century, depression wasconsidered a rare disorder. The vast majority of nervous states wereseen as anxiety disorders. Individuals suffering from “nerves” werenoted to have increased heart rates, butterflies in their stomach or othergastrointestinal complaints, headaches, sweating, and breathlessness.
The influence of Freud and psychoanalysis did a great deal to popular-ize the existence of nerves of this kind and to shape the portrayal of thiscondition as an anxiety disorder that could be treated, whether bydrugs or by talking therapies.
Popular thinking was also influenced by the world wars, which brought home the idea that extreme environmental stress could pro-duce “nervous breakdowns.” The most common treatment of the con-dition was sedation. Opiates or alcohol had been the sedative of choicethrough the nineteenth century. Bromides and barbiturates replacedthese in the first half of the twentieth century. In the 1930s, dexamphet-amine and other stimulants were marketed for people who sufferedfrom chronic fatigue. The first of “mother’s little helpers,” Dexamyl,combining a stimulant and a sedative—dexamphetamine and amylo-barbitone—appeared in the 1950s,6 producing dramatic benefits for“nerves” that are still not easily explained.
People with “nerves” could access these treatments without much medical intervention, since most were available over the counter. Oncemedical help was sought, a prescription could be refilled regularlywithout further medical contact. As a result, it is harder to gauge thelevel of nervous problems in the first half of the twentieth century thannow, but it is likely that it was no lower and no higher than today.
The launching of meprobamate under the brand name Miltown in 1955 was a watershed.7 This drug was discovered by Frank Berger, a I N T RO D U C T I O N : B E F O R E P RO Z AC Czechoslovak who emigrated to the United States before the SecondWorld War. Drug discovery at that time had little input from academicmedicine, and Berger was working on muscle relaxants in the laborato-ries of Carter-Wallace when he discovered that meprobamate calmedlaboratory animals without unduly sedating them. Since muscular ten-sion was a feature of anxiety, muscular relaxation without sedationmight be good for anxiety states. Given that his new drug was less seda-tive than older drugs, Berger adopted a new term—tranquilizer.8 Miltown at first looked to be the ideal answer for everyday anxiety.
It produced a pleasant, relaxed feeling that “liberated” people fromtheir nerves, encouraging them to do things they would not have doneotherwise. Because it left many people feeling better than well, a Mil-town craze was born in the newspapers and on radio and the newlyemerging television.9 Librium and Valium followed Miltown in the early 1960s. These benzodiazepine drugs looked as effective as Miltown but were evenless likely to cause sedation or dependence. Hoffman La Roche bril-liantly marketed Librium and Valium, “helping” doctors to realize thata significant proportion of the physical complaints appearing in theiroffices might be manifestations of anxiety. What drove ulcers, if not anx-iety? No one knew what caused hypertension, but it sounded as thoughthese patients were bottling up their anxiety. Patients with asthma orother breathing conditions were often anxious, as were patients withheadaches. For all these conditions, it became widespread practice toprescribe a tranquilizer along with whatever other medication the pa-tient was taking. Physicians were encouraged to give Valium to collegestudents who were facing the stresses of their new environment.
Housewives were regularly prescribed Valium to cope with the stressesof new suburban lifestyles. Busy executives were taking it. Sales of Val-ium soared in the 1970s, until at one point it was the best-selling drugon the market.
Critics began to question the appropriateness of tranquilizing on such a mass scale.10 Would these new agents dull the natural competi-tiveness people need for survival in a hard world? Would people (and,indeed, nations) become less fit for survival if these drugs were over-used? Problems coping at college surely were not medical disorders. Thestudent revolutionaries of the late 1960s argued that it was the politicalsystem that was confusing and disorienting people, and the appropri-ate therapeutic intervention was to change the system rather than treat I N T RO D U C T I O N : B E F O R E P RO Z AC individuals. This general political charge took a specific shape when theBritish government took Roche to court on the issue of overpricing andwon.
The benzodiazepine story came to an end in the 1980s.11 But the un- doing of the benzodiazepines came not from overcharging or frommass prescribing to mask social ills. It came when the possibility wasraised, around the end of the 1970s, that these drugs that had been so re-lied upon might lead to dependence. This specter helped create thephenomenon of health news. Before 1980, it was unusual to see healthcoverage in major newspapers. In the 1980s, however, stories aboutbenzodiazepines began to appear regularly, contributing to a regularhealth section in newspapers. Talk shows such as Donahue and Oprah inAmerica and That’s Life in England, on which individuals discuss theirproblems, furthered these developments. Health was a natural subjectfor these shows, leading indirectly to the formation of coalitions of cam-paigning patients on matters such as benzodiazepine dependence.
The benzodiazepines were portrayed as addictive, despite protests from clinicians that this was not the case. Animal tests of abuse liability,the establishment argued, showed benzodiazepines did not have theabuse liability of heroin, cocaine, or other classic drugs of abuse. Ad-dicts might use “Benzos,” but they had little street value. In contrast,they had a clear therapeutic niche, and most patients were able to dis-continue benzodiazepines without incident. The establishment viewscarried little weight against a rising tide of discontent and the outliningof the concept of normal dose dependence by Malcolm Lader, a profes-sor of psychopharmacology at the Maudsley Hospital in London, andothers. Before the 1980s, addicts had been socially shunned as perpe-trators of their own downfall. In contrast, benzodiazepine “addicts”were seen as victims of a medico-pharmaceutical establishment, andblame was directed at medical practitioners and pharmaceutical com-pany executives. Doctors and drug companies became the villains ofthe piece.12 The benzodiazepines rapidly passed in Western public perception from remarkably safe medicines to one of the greatest dangers to mod-ern society. The extraordinary nature of this development can be seen insharper focus when contrasted with what happened in Japan.13Whether because of safer prescribing in lower doses, or because theJapanese are genetically less liable to dependence on benzodiazepines,or because the tranquilized state was more socially desirable in Japan I N T RO D U C T I O N : B E F O R E P RO Z AC than comparable states of mind in the West, there was never a problemwith benzodiazepine use in Japan. When the tranquilizer market col-lapsed in the West, it continued to grow in Japan, despite medical re-forms and a growing acceptance that benzodiazepines should not beprescribed for hypertension, ulcers, or similar conditions. In contrast,the Japanese antidepressant market remained a small one, with noProzac available as late as 2003.
An important, though rarely mentioned, element in this story was the role of other pharmaceutical companies in bringing the benzodi-azepines down.14 When concerns were first raised about the benzodi-azepines, another group of drugs active on serotonin was in clinical tri-als. This group of drugs, of which buspirone (BuSpar) was later the bestknown, was in market development, and as part of this, Mead Johnsonin America and Bristol-Myers Squibb in Britain were prepared to high-light the problems of the benzodiazepines as part of a campaign tomarket buspirone as the first non-dependence-producing anxiolytic.15 Bristol-Myers Squibb sponsored symposia at conferences, special supplements in journals, and other articles by experts. They providedan opportunity for speakers to address large audiences of primary-carepractitioners and psychiatrists, outlining the hazards of the benzodi-azepines. This type of exercise, treading a fine line between educationand marketing, is commonplace. It may do good. It may do harm. Inthis case it backfired. Physicians and others were skeptical of the idea ofa non-dependence-producing anxiolytic. The benzodiazepine crisis hadeducated physicians to expect that any drug that treated anxiety wouldin due course be found to cause dependence. Patient resistance to BuS-par paralleled this physician skepticism: Buspirone was not as pleasantto take as the benzodiazepines. It flopped. The tranquilizer era wasover, and an antidepressant era was about to start.
Conventionally, the antidepressant story starts in 1957 with the twindiscoveries of the tricyclic antidepressant imipramine, by RolandKuhn, and the monoamine oxidase inhibitor (MAOI) iproniazid, byNathan Kline. But the pharmaceutical companies involved, Geigy andRoche, had little interest in an antidepressant and did nothing to pro-mote either of these drugs. Very few clinicians in office practice seemed I N T RO D U C T I O N : B E F O R E P RO Z AC to encounter depression. When Merck launched another antidepres-sant, amitriptyline, in 1961, unlike Geigy and Roche it decided it neededto market depression as well as amitriptyline. Frank Ayd from Balti-more, the first investigator to detect the antidepressant properties ofamitriptyline, had published a 1961 book titled Recognizing the De-pressed Patient.17 Merck commissioned fifty thousand copies of the bookto distribute to general physicians and psychiatrists wherever the com-pound was being marketed. Amitriptyline quickly became the best sell-ing of this group of drugs, later called the tricyclic antidepressants(TCAs or tricyclics) because of their three-ringed molecular structure.
But despite Merck’s efforts, the antidepressants remained the poorcousins of the tranquilizers.18 During the 1960s and 1970s, senior figures in biological psychiatry such as Nathan Kline and Fred Goodwin in the United States and PaulKielholz in Switzerland argued that many patients diagnosed as anx-ious were in fact depressed, and the appropriate treatment was an anti-depressant rather than an anxiolytic. This vision led Kielholz in 1972,supported by Ciba-Geigy, to set up the first meeting of the Committeefor the Prevention and Treatment of Depression.19 The brief of the groupwas to establish what needed to be done to improve the recognition andtreatment of depressive disorders. In the United States, Paul Wenderand Don Klein set up a comparable National Foundation for DepressiveIllness in the late 1980s.
The research underpinning the new thinking about depression came from social psychiatrists. In 1966, Michael Shepherd in Londonpublished the first book to suggest primary-care physicians rather thanpsychiatrists might be seeing the vast bulk of nervous problems, andthat a great many of these problems could be viewed as depression.20This laid the groundwork for a host of studies in the 1980s surveyingdepression in the general population, which in turn formed the basis formarketing the SSRIs.21 Far from welcoming the marketing of the SSRIs,Shepherd regarded the consequences of his research falling into thehands of pharmaceutical companies as comparable to those of the sor-cerer leaving his apprentice alone in the workshop.22 Another development came with the creation of DSM-III. Follow- ing the introduction of the new psychotropic drugs, in the late 1960spsychiatry was faced with the development of antipsychiatry, whichquestioned the legitimacy of psychiatric diagnoses and practices. Psy-chologists posing as patients famously got themselves admitted to psy- I N T RO D U C T I O N : B E F O R E P RO Z AC chiatric hospitals, unsuspected by medical staff even though real pa-tients easily detected the fraud.23 The controversies swirling aroundpsychiatry led in 1980 to the introduction of operational criteria for psy-chiatric disorders in a revised diagnostic manual, DSM-III. In DSM,anxiety neurosis was broken down into a number of apparently differ-ent disorders—social phobia, generalized anxiety disorder, panic disor-der, post-traumatic stress disorder (PTSD), and obsessive-compulsivedisorder. In contrast, the depressive disorders were collapsed into onelarge category—major depressive disorder.
DSM-III’s new formulations emerged just as Kielholz’s Committee for the Prevention and Treatment of Depression laid the basis for na-tional campaigns such as Depression—Awareness, Recognition, andTreatment (DART), in the United States, and Defeat Depression, in theUnited Kingdom. Eli Lilly supported both campaigns. In the case ofDART, Lilly funds went into eight million brochures titled Depression:What You Need to Know and two hundred thousand posters.24 As LewJudd, the director of the U.S. National Institute of Mental Health(NIMH) in 1987, put it: “By making these materials on depressive ill-ness available, accessible in physicians’ offices all over the country, im-portant information is effectively reaching the public in settings whichencourage questions, discussion, treatment, or referral.” Campaignslike this can do great good—or they can be self-serving.
In the early 1990s, surveys by the Defeat Depression campaign found most people thought everyday depression was not the kind ofcondition that should be treated with pills.25 But DART and other na-tional campaigns were launched on the waves of an incoming tide. The1980s saw a dramatic increase in articles about depression in both med-ical journals and general-readership magazines.26 Those hostile to psy-chiatry may smell a conspiracy, but the real interpretation has probablyto do with a vacuum opening up. Both academic and lay media were re-porting benzodiazepine horror stories in contrast to the feel-good sto-ries of previous years. There was a vacancy for stories about a new feel-good drug. News that clinicians such as Nathan Kline and Fred Good-win had been trying to cultivate for years, previously choked of light bythe canopy of overhanging tranquilizer publicity, was given a chance togrow. But no one expected it would turn out quite so easy to changemedical perceptions.
The depression campaigns had a twofold strategy. One was to alert physicians and third-party payers in health care to the huge economic I N T RO D U C T I O N : B E F O R E P RO Z AC burdens of untreated depression. The campaigns were so successful inthis strategy that a decade later no one bats an eye at claims that de-pression is one of the greatest single health burdens on mankind.27 Butno one asks whether treatments that are supposed to make a differenceactually do produce benefits. There is plenty of evidence that antide-pressants can be shown to do something in the short term but almost noevidence that things turn out better in the long run, and there are manyreasons to worry that we might be making things worse. Somethingmust surely be going wrong if the frequency of depression apparentlyjumps a thousandfold since the introduction of the antidepressants.28 The second strategy involved a series of educational campaigns to show physicians how many cases of depression they were missing—toshame them into detecting and treating depression. The tragedy ofSylvia Plath’s suicide, for instance, was held up as something that couldhave been prevented with better recognition of depression.29 This newemphasis has almost certainly led to diagnosis of depression for manypeople who do not regard themselves as being depressed or in need oftreatment. In individual cases, this heightening of clinician sensitivityto depression may have saved lives; on a broader front, there is no evi-dence that mass detection and treatment of depression have lowerednational suicide or disability rates.30 The antidepressant story has a further important twist. The con- ventional wisdom as of the early 1980s was that, unlike tranquilizers,which were feel-good agents that delivered a relatively immediate pay-off, antidepressants took several weeks to work. Prescribers were edu-cated to tell patients they could even expect to feel worse for someweeks before they began to feel better. This strategy sent the messagethat these were not quick-fix pills but rather medications that really cor-rected the problem.
But this educational information becomes problematic when these drugs are made available on prescription only from physicians trainedto deliver just this message. Against a background of instructions thatthese pills might take many weeks to work, on the one hand, for manydoctors the idea that Prozac or other antidepressants might lead to sui-cidal tendencies or other severe problems during the early period “be-fore the pills begin working” seems a contradiction in terms. On theother hand, patients faced with a doctor who has not been educatedabout the potential hazards in the early phase of treatment risk beingtrapped by their relationship with their doctor. Where patients would I N T RO D U C T I O N : B E F O R E P RO Z AC probably discontinue an over-the-counter medication if it did not seemto suit them, regard for their doctor, who very likely will advise them tocontinue treatment, can convince many to continue with a treatmentthat might kill them.
The messages of the depression campaigns were based on treat- ment of people hospitalized for the types of severe depression de-scribed by Kay Redfield Jamison. For these depressions, it made senseto avoid recurrences, and long-term treatment seemed a good idea. Butthis was an entirely different kind of disorder from the stress and ad-justment reactions and adolescent turmoil for which the SSRIs increas-ingly came to be administered in the 1990s, disorders that last on aver-age less than three months.31 Furthermore, the economics of depressionput forward to justify mass detection and treatment of depression wereall worked out initially on the basis of the Jamison form of depression.
None of the assumptions of such models—for example, that treatmentwill lower suicide rates or improve quality of life—holds for the kindsof “depression” that exist in the wider society.
For many people, one of the key factors that make it reasonable to takean antidepressant is the belief that a lowering of the brain neurotrans-mitter serotonin has been demonstrated in depression. If there actuallywere a confirmed lowering of serotonin in depression, giving drugs thatraised serotonin levels might seem a good idea.
The presence in the brain of serotonin was first reported in 1954.32 This quickly led to the hypothesis that this monoamine neurotransmit-ter might play some role in nervous problems. One way to investigatethis possibility was to look at the levels of the main metabolite of sero-tonin in the cerebrospinal fluid that bathes the brain. In 1960, GeorgeAshcroft, working in Edinburgh, found that cerebrospinal 5HIAA lev-els, the metabolite of serotonin, in depressives appeared to be low, lead-ing to the first theory that serotonin might be low in cases of depres-sion.33 While Europe was convinced that serotonin was a key neurotrans- mitter in nervous disorders, North Americans were certain that norep-inephrine was more important. Julius Axelrod, working at the NationalInstitutes of Health (NIH), had discovered an uptake mechanism for I N T RO D U C T I O N : B E F O R E P RO Z AC norepinephrine in 1961 and later found that tricyclic antidepressantslike imipramine blocked this mechanism.34 Axelrod dismissed Euro-pean work on serotonin as unimportant—serotonin was just “a rem-nant of our marine past.”35 In 1965, the NIMH’s Joseph Schildkraut published his cate- cholamine hypothesis of depression.36 This was the key article of thenew biological psychiatry, its Interpretation of Dreams. Schildkraut putforward the idea that brain norepinephrine was lowered in depressionand that antidepressants acted to increase its levels. Whatever the sci-ence base for this claim, crucially, this was a new language that bothpracticing psychiatrists and patients could understand. Talk of sexualcomplexes yielded to a new patter: “You have a chemical imbalance;these pills will restore your brain to normal.” Magazines from Time andNewsweek to the New Republic and the National Enquirer could embracethis idea, crucial to the later success of Prozac. This key myth still flour-ishes in popular consciousness almost forty years later.
By 1970, however, Ashcroft had concluded that, whatever was wrong in depression, it was not lowered serotonin. More sensitive stud-ies had shown no lowering of serotonin.37 Indeed, no abnormality ofserotonin in depression has ever been demonstrated.38 By the early1970s, moreover, Ashcroft and colleagues had introduced receptors intopsychiatry, after which few psychopharmacologists could be heardtalking about a lowering of either serotonin or norepinephrine in de-pression.39 A gap opened up between the science base and public un-derstanding—a gap crucial to the later development of media talkabout lowered serotonin levels.
But if the idea of lowered serotonin causing depression doesn’t come from scientific studies, where does it come from? The marketingof the SSRIs has emphasized that they increase serotonin levels: Doesthis not mean they were designed to increase serotonin levels in pa-tients where it is low? The answers to these questions interlock in thestory of reserpine, one of the first “antidepressants,” which loweredserotonin but also led to suicide.
The psychopharmacological era began in 1952 when two drugs, reser-pine and chlorpromazine, transformed the prospects for mental illness.
I N T RO D U C T I O N : B E F O R E P RO Z AC Initially, there was no such thing as an antipsychotic or antidepressant,partly because both of these drugs looked like they might be useful forall nervous problems. Chlorpromazine sold in Europe under the tradename Largactil because it seemed to work for a large range of condi-tions. Reserpine sold as Serpasil, and its distinctive benefits probablyled F. F. Yonkmann, a scientist working at Ciba, to coin the term tran-quilizer. When meprobamate and, later, the benzodiazepines hijackedthis new term, reserpine and chlorpromazine became the major tran-quilizers.
Nathan Kline was the first to show that reserpine in relatively large doses could be useful for treating psychoses. There was every reason tobelieve that in lower doses it might be beneficial for treating anxiety ordepressive disorders. This led Michael Shepherd to undertake the firstmodern clinical trial in psychiatry, in which he compared reserpine anda placebo in a group of anxious depressives.40 Reserpine came out as anantidepressant.41 The results of Shepherd’s study, published in 1955,were better than later results for Prozac in similar patient groups. Butthe discovery of the antidepressant properties of reserpine vanishedwithout trace, because for Ciba, the makers of Serpasil, the idea of anantidepressant was a complication rather than an added benefit, just asit was for Geigy and Roche.
Shepherd’s 1955 clinical trial paper was the first of its kind. Also in 1955, another reserpine paper was the first paper in another new scien-tific literature. In this, Steve Brodie and colleagues at the NIH demon-strated that reserpine both sedated rabbits and lowered their brain sero-tonin levels.42 This was the first demonstration of a link between brainchemistry and behavior. Brodie’s study, published in Science, becamethe classic that established the field of neuroscience, and pharmacolo-gists from around the world visited him.
Within two years, other laboratories had shown that reserpine also lowered norepinephrine and dopamine levels. Unaware that reserpinehad been shown to be an antidepressant, Schildkraut made reserpine’scapacity to lower norepinephrine levels the key element in his theory ofdepression, on the basis of reserpine’s ability to trigger suicide.
The capacity of reserpine to cause suicide was not discovered within psychiatry. In the early 1950s, in addition to being used for nerv-ous problems, reserpine was used as an antihypertensive. A number ofphysicians noted that hypertensive patients taking reserpine reportedfeeling “better than well.” In 1954, Robert Wilkins, in a phrase that I N T RO D U C T I O N : B E F O R E P RO Z AC echoes the later Prozac story, claimed physicians were suggesting it wastime to give up psychotherapy, as reserpine was good psychotherapy inpill form.43 But starting in 1955, a series of reports emerged of hypertensive pa- tients becoming “depressed” on reserpine and committing suicide. Inthe pages of The Lancet, for example, immediately prior to Shepherd’sreport on the antidepressant effects of reserpine, papers from NewZealand and Australia report on reserpine-induced suicidality in hy-pertensive patients.44 Psychiatrists argued that chlorpromazine causedsimilar problems.45 Both drugs, it seemed, could make some patientsfeel flat and unmotivated or anxious; but in contrast to proper depres-sion, this problem cleared up once the drug was stopped or if the pa-tient was given a stimulant. Nevertheless, the notion that reserpinecaused depression crept into the literature.
The misidentification of the problems that reserpine caused may have been one of the greatest mistakes in psychopharmacology. In thefirst paper to describe hypertensive patients becoming suicidal on re-serpine, in 1955, Richard Achor and colleagues from the Mayo Clinic re-ported that reserpine caused some patients “increased tenseness, rest-lessness, insomnia and a feeling of being very uncomfortable.”46 RobertFaucett gave further details: “The first few doses frequently made themanxious and apprehensive. . . . They reported increased feelings ofstrangeness, verbalized by statements such as ‘I don’t feel like myself’. . . or ‘I’m afraid of some of the unusual impulses that I have.’”47 Ger-ald Sarwer-Foner described a subject who, on the first day of treatment,reacted with marked anxiety and weeping, and on the second day “feltso terrible with such marked panic at night that the medication wascancelled.”48 This is not depression. These quotes are the first descriptions of a phenomenon new to physicians and psychiatrists. In 1955, in an unfor-tunate choice of terms, two German-speaking psychiatrists, Hans Steckand Hans Haase, labeled the problem akathisia.49 Akathisia, which liter-ally refers to an inability to sit still, was a problem first noted followingbrain injuries and then in the encephalitis lethargica epidemic of 1918.
Encephalitis lethargica produced instant Parkinson’s disease in somepeople, rendering them mute and stuporous, but it made others restlessand agitated.50 Akathisia was later to become a recognized hazard of an-tipsychotic drugs, one associated with suicide and homicide.51 But I N T RO D U C T I O N : B E F O R E P RO Z AC awareness of what was involved for many years did not percolate out-side a specialized psychiatric community.
The word akathisia remains one of the least understood pieces of medical jargon in the psychiatric lexicon, and the condition has neverbeen properly investigated. The quotes from Faucett and others suggestthat the best translation into English is probably “turmoil” or “agita-tion,” rather than “restlessness.”52 When the problem began to show upin company trials of the SSRIs in the 1980s, it was coded under “agita-tion.” It is often said that a psychiatric drug can never be shown to cause suicide because patients are already at risk of suicide from the nervouscondition for which they are being treated. Reserpine demonstrates thesuperficial nature of this response. None of the people who committedsuicide on reserpine had a nervous problem—they were all hyperten-sive.
As we shall see, Prozac and other SSRIs can also cause suicide in in- dividuals who have no nervous conditions, primarily by inducing men-tal turmoil during the early stages of treatment. Patients who developakathisia on SSRIs are likely to be seen by primary-care physicians whohave not been trained to recognize the problem. When Prozac and otherSSRIs were first launched onto the market, there were no educationalprograms or warnings about this hazard.
One of the major differences between reserpine and Prozac is that patent laws in the 1950s meant there were up to twenty-six different re-serpine-containing compounds on the market. By the time Prozac ap-peared, the laws had changed so that only one company could own anew compound. Thus no company had the incentive to defend reser-pine as Lilly later had to defend Prozac. Arguably, because of this patentsituation the problems of reserpine could be recognized, whereas laterproblems with Prozac could not be acknowledged without potentiallyputting Lilly out of business. This situation has implications for patientstaking any drugs, for the psychiatrists and other physicians who treatthem, and for the business practices of the corporations producing thedrugs used in treatment.
The reserpine story contains all the seeds of a tragedy in the classi- cal sense, where some flaw within the protagonist brings the herodown. In the early days of the psychopharmacological era, academialooked down in disdain at the pharmaceutical industry, often refusing I N T RO D U C T I O N : B E F O R E P RO Z AC to allow anyone working in the industry to join academic societies, andrefusing to engage with the complex and messy problems posed bytherapeutics. To this day, the conflicts between academia and industryare portrayed as a corruption of academia by industry. But an alterna-tive is that there is a repeated failure of academics to engage with therange of scientific and moral issues involved in applied science. TheSSRI drugs are among the clearest products of applied science withinmedicine, and their use peppers the academia/industry interface withas yet unanswered questions. Who are the flawed heroes in thistragedy—the academics or industry? The scientific literature carries few full-page corrections, but in 1997 a“Correction” appeared in the journal Life Sciences.53 This referred to anarticle published in 1995 by David Wong, Frank Bymaster, and Eric En-gleman from Eli Lilly, titled “Prozac, the First Selective Serotonin Re-uptake Inhibitor and an Antidepressant Drug: Twenty Years since ItsFirst Publication.”54 The authors of the correction, Arvid Carlsson andDavid Wong, agreed that the 1995 article might have given the mis-leading impression that Prozac was the first of the SSRIs. Prozac wasnot the first SSRI, nor the first patented, nor the first in clinical trials, northe first launched.
The origin of the SSRIs lies in the 1960s, when Paul Kielholz became professor of psychiatry in Basel. Given the presence in Basel of themajor Swiss chemical companies, Kielholz was well placed to become aleading figure in the world of psychopharmacology. Depression was hisarea of interest, which he saw as underrecognized and poorly treated.
Kielholz believed more had to be done than simply teach physicians todetect depression and put patients on treatment. Different antidepres-sants did quite different things, he argued, and it was important to se-lect the right antidepressant for the individual patient.55 While the molecular structures of the early antidepressants looked very similar, leading to them collectively being called tricyclics, Kiel-holz argued that some of these tricyclics, such as desipramine, madepeople well by enhancing drive. Others, such as trimipramine, madepatients well through a sedative action. And yet others had some otheraction on mood or emotions that appeared to be important, which he I N T RO D U C T I O N : B E F O R E P RO Z AC found difficult to characterize. Clomipramine was the tricyclic that ap-peared to have more of this mysterious action than other antidepres-sants.
This was a striking observation that flew in the face of Schild- kraut’s catecholamine theory, which claimed that all antidepressantsessentially did the same thing—they inhibited norepinephrine reup-take. It took a new kind of player within psychopharmacology to re-solve this mismatch between clinical and laboratory observations—Arvid Carlsson. Carlsson, from Sweden, was one of the first neurosci-entists. Trained in Brodie’s laboratory, Carlsson had a string of earlyresearch successes to his credit, including participation in the keystudy demonstrating the existence of neurotransmitter pathways in thebrain. He had discovered dopamine in the brain and was among thefirst to suggest that it might be abnormal in Parkinson’s disease, workthat later led to a Nobel Prize.56 By this time he had put forward thefirst evidence in what later became the dopamine hypothesis of schiz-ophrenia.57 Faced with Kielholz’s claims, Carlsson saw a connection between the different effects of antidepressants Kielholz proposed and the ef-fects of these same drugs on various neurotransmitter systems. Thedrugs Kielholz claimed to be drive-enhancing had effects on the norep-inephrine system, whereas clomipramine in particular had effects onthe serotonin system.58 This led Carlsson to suggest developing drugsthat only inhibited the reuptake of serotonin. Such a drug might helpclarify the nature of this mysterious other action of antidepressants andmight also produce a useful agent for the treatment of depression.
But being able to detect an effect when a drug is acting on a brain system logically assumes that there is no abnormality in that system. Ifclomipramine corrected an abnormality in the serotonin system, all thatwould show up was normalization of the person being treated. Kiel-holz’s claim, in contrast, was that drugs active on the serotonin systemwere detectably doing something different from other drugs. If, in casesof depression, there were an abnormality of the serotonin system,which SSRIs corrected, these should be among the most potent drugs totreat depression; but (as later became clear) there is little evidence thatthey work in cases of moderate to severe depression. In contrast, ifKielholz and Carlsson were right, and SSRIs do something differentfrom other antidepressants—for instance, if they produce some kind ofanxiolysis—one might expect them to be useful across a range of mixed I N T RO D U C T I O N : B E F O R E P RO Z AC anxiety and depressive states, rather than just for depression. This is ex-actly what was found.
Following Kielholz’s lead, Carlsson set to work with Hanns Corrodiand Peder Berndtsson at Astra’s plant in Sweden to create an SSRI. Hetook the antihistamine chlorpheniramine and manipulated the mole-cule to come up with zimelidine.59 Carlsson applied for a patent onzimelidine as a selective serotonin reuptake inhibitor in Sweden, Bel-gium, and Britain on April 28, 1971. The first patent was published inMarch 1972. Prozac was patented in 1974.
Zimelidine went into clinical trials comparing it with the norepi- nephrine reuptake inhibitor desipramine. These trials were short andnot conducted on severe depression. The first results were presented in1980, and zimelidine was launched on the market in Europe as Zelmidin 1982. The first trials of Prozac in depression were not published until1985, and it was not launched until 1988. After vigorous promotion,Zelmid began to be prescribed widely.
Astra had signed a comarketing agreement with Merck to market Zelmid in the United States. Had this proceeded, there would probablynever have been a Prozac phenomenon. Merck was the largest pharma-ceutical company in the world and recognized as marketer par excel-lence. But just as the data on Zelmid was delivered to the U.S. Food andDrug Administration (FDA) in 1982, there were reports that Zelmidcould trigger a serious neurological disorder called Guillain-Barré syn-drome. This disorder, which could kill by paralyzing the respiratorymuscles, led to the immediate removal of Zelmid from the market.60 Astra had already begun the development of a derivative of Zelmid, called alaproclate, which was being investigated for treatmentof both depression and Alzheimer’s disease. But alaproclate causedliver problems in one strain of laboratory mice and was dropped.61Shortly thereafter, Astra introduced an innovative antipsychotic, re-moxipride, which seemed to have significantly fewer side effects thanolder agents. Several months after its launch, however, remoxipridewas reported to cause aplastic anemia in a small number of people, andit too was withdrawn.
I N T RO D U C T I O N : B E F O R E P RO Z AC In the face of these setbacks, Astra contemplated withdrawing from the research-based pharmaceutical market to focus on over-the-countermedicines. Around 1990, it was estimated that new FDA regulationsand other hurdles to drug development meant that the cost of bringinga drug to market had rocketed to $300 million.62 No company could eas-ily survive the loss of its flagship compounds. Astra kept going only be-cause they had the breakthrough antiulcer drug omeprazole, the first ofthe proton-pump inhibitors, on its way to becoming one of the best-sell-ing drugs on the market.63 Despite the revenues from omeprazole, Astrawas forced into a merger later in the decade. This indicates how big thedrug development stakes can be. A troublesome side effect early in thelife of a new compound can lead to the demise of a company. If the sideeffect can be portrayed as a feature of the disease being treated—for ex-ample, suicidality in the case of an antidepressant—what would a com-pany do if the alternative were to go out of business? This is the ethicaldilemma that has faced all the SSRI companies.
Zelmid’s history contains two points of special note. The first is that during clinical trials and postlaunch studies, patients taking Zelmidmade a greater number of suicide attempts than expected. No one knewwhat to make of this. Was it an artifact? The same trials indicated thatsome of the people who did best on Zelmid had been those most suici-dal to begin with.64 When Lilly later ran into serious difficulties gettinga license for Prozac in Germany, many outsiders were mystified thatGerman regulators came to a different conclusion about Prozac’s sui-cide risk than had regulators in America, although presented with sim-ilar data. The Germans, however, had prior exposure to zimelidine andfluvoxamine (Luvox), whereas Prozac was the first SSRI the FDA wasfaced with.
The second point to note concerns the patenting of SSRIs. Although Carlsson spent two years in correspondence with senior executives atLilly before the paper correcting the misleading impression as to whohad discovered the SSRIs was published, relatively selective serotoninreuptake inhibitors had, in fact, been on the market long before Carls-son’s work. To produce zimelidine, Carlsson and colleagues manipu-lated the structure of an existing antihistamine, chlorpheniramine, apotent serotonin reuptake inhibitor that has since been shown to sharemany properties of the SSRIs.65 For example, it is effective in treatinganxiety disorders and panic attacks. If companies or scientists had I N T RO D U C T I O N : B E F O R E P RO Z AC simply wanted to see what effects these new compounds might have,they needn’t have gone to the trouble to create new drugs. But becausezimelidine was a new molecule, Astra could take out a patent on it. Thepatent system offers the possibility of huge returns but brings responsi-bilities in exchange, which Astra acknowledged.
Another manipulation of the antihistamines produced indalpine,66 firstdeveloped by Gerard le Fur in a French company, Fournier Frères, laterpart of Rhône Poulenc. After clinical trials, indalpine was marketed inFrance and other European countries as Upstene, arriving just afterZelmid. It was greeted enthusiastically because it seemed to benefit pa-tients who hadn’t responded to other drugs.67 But indalpine ran into trouble. Clinical trials in other European countries suggested it might lower white blood cell counts.68 This hap-pens transiently with many psychotropic drugs and for the most part isnot a serious problem, although in rare cases, if undetected, it can provefatal. But to general astonishment, indalpine was removed from themarket.
French psychiatrists were extremely upset and lobbied the com- pany and the government. The experience of Pierre Lambert from Lyonin the south of France was typical. He and his colleagues had investi-gated more psychotropic drugs before launch than had any other groupin either Europe or America,69 and they believed they had seen dra-matic results with indalpine. The effectiveness of the drug was symbol-ized for them by the apparent recovery and post–indalpine removalsuicide of one of their patients. Chronically depressed, she had beentransformed by indalpine. When the drug was withdrawn from themarket, she relapsed. Nothing else appeared to make any difference.
She kept going in the hope that the drug might be restored to the mar-ket, but when it wasn’t, she committed suicide. Her suicide note askedthat instead of flowers, a headstone, or anything else at her funeral, acollection should be made for medical research. Her family later do-nated the suicide letter and the proceeds of the collection to research.
Indalpine had been born at a time when the development sequence for compounds had not yet been set in stone, and animal studies aimedat detecting problems could continue in parallel with studies in people.
I N T RO D U C T I O N : B E F O R E P RO Z AC In ongoing toxicology studies after the drug was launched, some of theanimals developed liver cancer. There is little reason to believe that thesame would have happened in humans. But by this stage, indalpinewas being prescribed far more widely than had been expected. It wastherefore a statistical certainty that some of those getting indalpinewould also develop liver cancers. In light of the toxicology data, whowas to know whether these cancers were simply coincidental or hadbeen triggered by the drug? These are difficult calculations for a com-pany to make.
Another factor influenced Rhône Poulenc. A range of what were first termed ecologist or Green groups and, later, pharmacovigilancegroups, some of which had been born in the antipsychiatry protests ofthe late 1960s, had emerged in Germany and other European countriesin the 1970s. The ecologists argued against physical therapies in psy-chiatry. Campaigning had already led to electroconvulsive therapy(ECT) being banned in a number of countries. Psychiatry was undersiege.70 The lowering of white cells that indalpine caused made it a firstdrug target for these new pharmacovigilantes.
Emboldened by the demise of indalpine, the new pharmacovigi- lantes set their sights on nomifensine, which in a number of cases hadtriggered hemolytic anemia.71 Nomifensine was withdrawn. Senior Eu-ropean psychiatrists were reeling at these developments. In the face ofthis onslaught from “fringe groups,” they dug in. This was not a matterof protecting the pharmaceutical industry but of losing useful tools.
There were then no other drugs like indalpine or nomifensine on themarket.
The next target for the pharmacovigilantes was the best-selling an- tidepressant in Europe, mianserin.72 Like indalpine, mianserin couldlower white cell counts. The drug’s manufacturer, Organon, receivedletters pointing out that mianserin could trigger this potentially fatalproblem. Roger Pinder, a senior scientist at Organon, supported bymany opinion leaders from psychiatry, responded that all the antide-pressants then available except mianserin could be lethal in overdose, sothat even if some people did die from white cell problems with mi-anserin, overall, fewer would die using it than any other antidepres-sant. The ecologists were not impressed—suicide, after all, was illegal.
(The possibility of accidental overdose was not addressed.) Organon’s defense worked across Europe except in Britain, where the Committee on Safety in Medicines wanted mianserin withdrawn I N T RO D U C T I O N : B E F O R E P RO Z AC from the market. This led to a series of legal cases; Organon made itclear they were prepared to take the matter all the way to the EuropeanCourt. This situation was unprecedented. In the ordinary course ofevents, a company faced with regulatory disapproval would simplycomply. Eventually Organon won, but the disputes led to a collapse ofmianserin sales in most countries.73 The Organon defense involved a risk-benefit calculation. When death from suicide is included in the equation, newer antidepressantsmay appear safer, although much more expensive and no more effec-tive than the older antidepressants. Safety in overdose became a keycard later played by Lilly in its arguments with the regulators over thesafety of Prozac. This argument was new to regulators, who were beingasked to contemplate a scenario equivalent to the pope being urged toallow condoms on the basis that they minimize the spread of AIDS. Un-like the pope, the regulators were faced with the dilemma that letting adrug with a known hazard remain on the market opened them up tolegal actions. This was uncharted territory.
The events surrounding indalpine and mianserin put in place a set of jigsaw pieces that gave glimpses of what was later to be the Prozacstory. Pharmacovigilance divisions within companies were pittedagainst a set of pharmacovigilantes—in Germany, the ecologists; inAmerica, the Church of Scientology. Suicide was to become an issue inthese debates. The defense of mianserin showed why companies wouldwant to mobilize a coalition of scientists, a network of “friends,” toargue the company’s case.74 The network of senior scientists that RogerPinder mobilized on Organon’s side later played a part in managing thecontroversy over Prozac in Europe. Lilly organized a similar network inthe United States. And finally, lawsuits had become a weapon. From themid-1980s, companies would need sophisticated legal advice on how toswim in these new waters.
The first SSRI to survive on the world market was fluvoxamine, whichwas developed in 1973 from another antihistamine.75 Duphar Labora-tories launched fluvoxamine in 1983 in Switzerland, and in the rest ofEurope between 1984 and 1986. But in Germany, fluvoxamine was heldup because in clinical trials there had been a higher number of suicides I N T RO D U C T I O N : B E F O R E P RO Z AC and suicide attempts on fluvoxamine than on the drugs with which itwas being compared. Duphar was asked to account for this.
Jenny Wakelin, working with the company, consulted with experts around Europe before coming up with the apparently clinching data.
When the trials were reanalyzed, focusing specifically on those mostsuicidal to begin with, it appeared that fluvoxamine was more likely toreduce suicidality than the comparator drugs imipramine andamitriptyline. The lesson many drew from this was that the apparentlyhigher rate of suicide attempts on fluvoxamine was a chance develop-ment—that, indeed, SSRIs might be even more antisuicidal than olderdrugs.76 The “experts” were learning how to handle the regulators onthis issue.
In the 1980s, the first patients to get a new antidepressant were pa- tients hospitalized with depression who seemed unresponsive to othertherapies—not a promising group on whom to try a new drug. Poor re-sponses in this group, along with severe nausea in many patients, led toearly clinical judgments that fluvoxamine was unlikely to make signif-icant inroads into the antidepressant market. It never did.
But another route to salvation opened up for fluvoxamine. By gen- eral agreement, clomipramine, first made in 1958, is the most powerfulantidepressant ever made.77 This tricyclic antidepressant, with actionson both the norepinephrine and serotonin systems, was the last of themajor tricyclic antidepressants to come to market. Many viewed it ini-tially as just another “me too” drug. The FDA was keen to discouragecopycat drugs and did not license clomipramine.
George Beaumont, a physician with Geigy UK, learned of reports that clomipramine might help treat obsessive-compulsive disorder(OCD).78 Setting out to establish a niche for clomipramine in the treat-ment of OCD, Beaumont organized a series of studies in the early 1970s,saw a positive response to the drug, and got clomipramine licensed forthe treatment of both depression and OCD in Britain.79 Then Judith Rapoport at the NIMH, in a randomized trial, gave children with OCD either desipramine, which has no effect on the sero-tonin system, or clomipramine. This study demonstrated conclusivelythat OCD responded to clomipramine but not to desipramine. Beforethen, the conventional wisdom had been that the benefits fromclomipramine in OCD were because an “antidepressant” likeclomipramine cleared up the mood disorder that accompanied OCD.
But desipramine was also an antidepressant, and Rapoport’s study had I N T RO D U C T I O N : B E F O R E P RO Z AC just demonstrated that it had no effects for OCD.80 There was somethingdistinctive about drugs that acted on the serotonin system.81 After publication of Rapoport’s results, and in particular following her runaway best-seller about OCD, The Boy Who Couldn’t Stop Washing,the scene quickly changed.82 Rapoport appeared on talk shows likeDonahue and Oprah; OCD, still thought to be a rare disorder, emergedfrom the shadows. Many who had suffered silently, concealing their rit-uals and intrusive thoughts for fear of ridicule or being thought insane,came forward for further studies and treatment.
Companies had regarded OCD as even less interesting than de- pression in the 1950s. But by the late 1980s, under the influence ofRapoport and the success of clomipramine, companies realized thatOCD was a market worth pursuing. Clomipramine was finally licensedin the United States in 1990 for the treatment of OCD rather than de-pression. Meanwhile, Duphar set up a marketing agreement with Up-john to develop fluvoxamine for OCD, and it made its way onto the U.S.
market under the brand name Luvox.83 Luvox was the low-profile SSRIuntil the killings at Columbine High School in Colorado, when it wasreported that one of the shooters was on Luvox for OCD.
CELEXA AND ZOLOFT—WAR BETWEEN THE SISTERS Citalopram and escitalopram were first produced in Europe by theLundbeck Pharmaceutical Company.84 In 1971, the company hired Klaus Bøgesø as a medicinal chemist.
Over the years, Bøgesø turned out to have a Midas touch at the game ofdrug hunting. The challenge facing him after his recruitment was toproduce a selective norepinephrine reuptake inhibitor. Like other com-panies at the time, Lundbeck had little interest in an SSRI.
Two potential antidepressants came out of Bøgesø’s first efforts— talopram and tasulopram. These were pressed into clinical trials. Bothturned out to increase energy levels, and both were linked to a numberof suicide attempts. This appeared to confirm one of the major theoriesof the time, also put forward by Paul Kielholz: that energy-increasing,or activating, antidepressants might lead to suicide. Lundbeck re-treated. Suicide was the greatest hazard of an antidepressant. But Kiel-holz’s views also suggested that nonactivating antidepressants wouldbe far less likely to trigger suicide, and SSRIs were less likely to be acti- I N T RO D U C T I O N : B E F O R E P RO Z AC vating. Following a lead from Carlsson, Bøgesø converted talopraminto citalopram, the most selective serotonin reuptake inhibitor to cometo the market.
The detour through talopram left Lundbeck behind its competitors.
Nevertheless, citalopram became the best-selling antidepressant in anumber of European countries. Lundbeck’s strategy of undercuttingthe cost of the other SSRIs and promoting citalopram as the most selec-tive SSRI, and therefore the one least likely to cause side effects,worked.
In the United States, the story was even more extraordinary. In Jan- uary 1998, the New Yorker featured an article by Andrew Solomon titled“Anatomy of Melancholy,” an account of the author’s own depres-sion.85 Within a month, Solomon received two thousand letters fromother depression sufferers. Clearly, he had struck a nerve. His articlewas anthologized in more than thirty books, and he became aspokesman in such forums as the American Psychiatric Association.86One of the striking points of the piece was his description of the effectsof Zoloft as being like drinking fifty-five cups of black coffee, with theeffects of Paxil being the equivalent of eleven cups of black coffee. Usersseemed well aware of this stimulating effect at a time when both man-ufacturers and clinicians were denying it.
After failing to negotiate a marketing agreement with Pfizer and then with Warner-Lambert, Lundbeck gave up on the U.S. market. Fi-nally, however, they entered into a licensing arrangement with ForrestLaboratories, a small pharmaceutical company run by a chief executivewho appeared confident this drug could run, even though it wouldhave to come from the back of the field. The chief executive wasHoward Solomon, Andrew Solomon’s father.87 Launched in September1998, a strategy of undercutting the price of other SSRIs and aggressivemarketing enabled Celexa to capture so large a market share that it be-came front-page news.88 Pfizer’s SSRI sertraline began life in 1977.89 Playing around with the nuclei of some of the original antipsychotic molecules, companychemists produced a series of norepinephrine reuptake inhibitors, ofwhich tametraline looked the most promising. When side effects haltedtametraline’s development in 1979, Willard Welch transformed it into aseries of serotonin reuptake inhibitors, one of which was sertraline.90 Like Lundbeck, Pfizer was behind the competition. Sertraline hit the market in North America in 1992 as Zoloft, and in Europe from 1990 I N T RO D U C T I O N : B E F O R E P RO Z AC through to 1993. Pfizer emphasized technical differences between it andthe other SSRIs, claiming that it had a better half-life, a cleaner break-down route, and a lower liability to interact with other compounds inthe body, and that therefore it was much safer than other SSRIs. Thismarketing strategy produced the appearance of science—lots of data—but very few of these data were clinically relevant. The approach wasgeared to making Zoloft appear “clean” compared with Prozac andPaxil.91 The other companies responded, leading to a “war between thesisters.” Pfizer had another approach to new drugs embodied in a program called CRAM—Central Research Assists Marketing—headed up by themarketing department.92 In the case of Zoloft, this program led Pfizer toestablish PRIME-MD, a research program supposedly aimed at collect-ing data about primary-care depression in order to educate primary-care physicians. This, of course, made it likely that many of these doc-tors would go on to treat patients who had been identified as depressed,and that Zoloft would be the first drug tried. Another research program,RHYTHMS, was aimed at studying patient education and compliance.
The immediate downside to such a program is that improving compli-ance with treatments not suited to particular patients may increasethese patients’ risk for suicide. The downside to the larger picture is thatthis type of market-driven science tends to displace real scientific effortsto answer clinical questions.
Celexa and Zoloft have been the backing chorus for the big two SSRIs,Prozac and Paxil. Paroxetine was developed in 1978 by Jorgen Buus-Lassen and colleagues in a small Danish company called Ferrosan.
Paroxetine was Buus-Lassen’s second SSRI. In 1975, he had producedfemoxetine, which appeared more effective than paroxetine but had onedisadvantage: it was not going to be a simple once-a-day pill.93 Ferrosan sold paroxetine to Beecham Pharmaceuticals in 1980, which later merged with SmithKline and French to become SmithKlineBeecham (SB). SB merged again at the turn of the millennium withGlaxo to become Glaxo-SmithKline (GSK), at that point the world’slargest pharmaceutical corporation. Ferrosan was acquired by Novo- I N T RO D U C T I O N : B E F O R E P RO Z AC Nordisk, which had little interest in psychiatry, and femoxetine diedfrom neglect.
Paroxetine nearly died of neglect as well. Beecham was considering shelving paroxetine because it appeared less effective in clinical trialsthan older antidepressants.94 A large study run by the Danish Universi-ties Antidepressant Group later confirmed this.95 This was at a timewhen the nonhospital depression market still appeared relatively small,and it was not obvious how a less effective antidepressant, even a saferone, could be expected to take a significant share in this market. As a re-sult of company ambivalence, the clinical development of paroxetinelagged behind that of Zelmid, Luvox, and later Prozac.
Paroxetine was licensed as Paxil in 1992 in the United States and as Seroxat in 1991 in the United Kingdom. Marketers within what wasnow SmithKline Beecham coined the acronym SSRI. Compared to theother serotonin reuptake inhibitors, paroxetine was supposedly the se-lective serotonin reuptake inhibitor.96 The name worked all too well. Itwas adopted for the entire group of compounds. Thus Paxil madeProzac and Zoloft into SSRIs.97 The idea of an SSRI conveys the impression of a clean and specific drug, freer from side effects than the nonselective tricyclics. However,selectivity meant different things to pharmacologists and clinicians. Forpharmacologists, an SSRI could act on every brain system except thenorepinephrine system, and SSRIs might in this sense be even dirtierdrugs than any of the tricyclics. Clinicians were misled if they thought“selective” meant that these drugs acted on only one brain site; but thiswas exactly what the marketing of these drugs implied to them.
Where Upjohn had targeted OCD to carve out a distinctive identity for Luvox, SmithKline targeted panic disorder, anxious depressions,generalized anxiety disorder, and, later, social phobia. The targeting ofPaxil for anxiety led to huge sales, making Paxil the closest rival toProzac. When GSK got a license to market Paxil for social phobia, itsstock rose: an antishyness pill was potentially a huge market.
Social phobia had, until the 1990s, been almost unknown in the Western world.98 First described by Isaac Marks in the 1960s at the In-stitute of Psychiatry in London, social phobia presented rarely to clin-ics. It would be a mistake, however, to think that SmithKline somehowinvented social phobia; in Japan and Korea, social phobias have alwaysbeen recognized. But there is an obvious overlap between social phobia I N T RO D U C T I O N : B E F O R E P RO Z AC and shyness. As a consequence, there is a real risk that legitimate effortsto market a treatment of benefit for a disabling medical condition willat the same time capture a significant number of people who are simplyshy—and who may be at more risk from the treatment than from theirshyness. Furthermore, the term social phobia apparently did not suit thebrave new world, and in the late 1990s it was replaced by social anxietydisorder, raising questions about the culture of psychiatry.99 While hugely successful in terms of sales, the targeting of Paxil for anxiety disorders contained a snag. Soon after its launch, primary-carephysicians and others began to describe patient dependence on SSRIsthrough adverse-event reporting systems.100 This happened first inGreat Britain.101 There was a much greater volume of reports for Paxilthan for other SSRIs. It was first suggested that these withdrawal effectsstemmed from the short half-life of the drug. It was also suggested thatthe fact that Paxil, more than other SSRIs, was used to treat patients whowere anxious meant that any problems had to do with the personalitiesof these patients, who were, after all, more likely than others to developphobias—so why not a withdrawal phobia? Amid reports of withdrawal symptoms in the mid-1990s, Lilly con- vened a panel of “opinion leaders” to discuss the phenomenon of whatwere termed “antidepressant discontinuation syndromes” rather thandependence problems.102 Prozac, given its very long half-life, seemedless likely to cause this problem than the other SSRIs.103 Lilly saw a mar-ket opportunity vis-à-vis Paxil and Zoloft, their closest competitors,and began to run advertisements about discontinuation syndromes.104 In so doing, Lilly pointed to a general problem with the SSRIs. A key reason for the development of SSRIs as antidepressants lay in thefact that clinicians suspected all tranquilizers would, in due course,produce dependence, just as the benzodiazepines and barbiturates haddone. In the early 1990s, the antidepressants were not associated withdependence. Clinicians felt comfortable denying the capacity of thesedrugs to produce addiction or dependence. When the Royal College ofPsychiatrists launched its Defeat Depression campaign in the 1992, pro-fessional polling organizations found most people thought antidepres-sants were likely to be addictive. The Royal College campaign went outof its way to emphasize that antidepressants were not addictive. Untilvery recently, the backs of Prozac packets contained an explicit state-ment: “Don’t worry about taking Prozac over a long period of time—Prozac is not addictive.” I N T RO D U C T I O N : B E F O R E P RO Z AC If by “addictive” we mean some property within a drug that will transform takers into junkies, likely to mortgage their livelihoods andfutures for an ongoing supply of drugs, then the SSRIs at present do notappear to be addictive. They do not lead to a life of crime or dissolu-tion.105 But this does not mean SSRIs don’t produce dependence. Manypeople experience grave difficulties in discontinuing treatment fromPaxil or Zoloft. In lay terms, you can just as easily become hooked toSSRIs as on benzodiazepines. For most of us, this is the meaning of ad-diction.
Far from the problem being simply one of dependence that emerges on withdrawal from the drugs, the SSRIs produce what may be moreappropriately termed “stress syndromes.” Unlike insulin or thyroidhormone, which are replacements for a deficiency, the SSRIs are alienchemicals and, as such, are brain stressors. The consequences of thisstress become apparent in some individuals on drug withdrawal, whenthe system attempts to regain equilibrium. But in others, the stressescan be visible during the course of treatment.
With the SSRIs, a problem called poop-out was noted early on.106 “Poop-out” refers to the drugs losing their effect; sometimes they can beprompted to work again by increasing the dose.107 This phenomenonfirst came to light in Internet chat rooms rather than through physiciansinformed by companies of the existence of a problem.108 Because com-panies denied the problem, they could not advise on the best means ofmanaging it. Clinicians were left to their own devices. This is hardly thekind of partnership supposed to characterize prescription-onlyarrangements.
The remainder of this book will deal with suicide on SSRIs, but a key feature of the Paxil story is that dependence on SSRIs is more likelyto bring this group of drugs into public disrepute than the issue of sui-cide. Suicide is something anyone contemplating using an SSRI findshard to envisage happening to them. But we can readily envisage get-ting hooked on a drug, and if this happens, we remain alive to complainabout it. Public anger is likely to be fueled by growing awareness thatassertions that people need to take SSRIs all their life, just as diabeticsneed to take insulin, were made against a background where the ca-pacity of Paxil to produce dependence had been noted by Beecham in aseries of healthy volunteer studies in the mid-1980s before the drugcame on the market, and the company had been warned by senior fig-ures in the field of this risk but, it would seem, chose to do nothing.109 I N T RO D U C T I O N : B E F O R E P RO Z AC The possibility of dependence takes the gloss off company claims thatmillions of us continue to take SSRIs successfully each year.
In the 1960s, Eli Lilly’s best-selling antidepressant was nortriptyline, anorepinephrine reuptake inhibitor. In late 1971, Bryan Molloy—achemist working in Indianapolis for what was then, in terms of psychi-atric drugs, a small pharmaceutical company—using another antihista-mine, diphenhydramine, as a starting point, came up with a group ofphenoxyphenyl-propylamines. Molloy was in this literal sense Prozac’screator. Of the fifty-seven phenoxyphenyl-propylamines he produced,the one given the code LY-94939, later called nisoxetine, turned out tohave the laboratory profile Lilly was interested in. It was a norepineph-rine reuptake inhibitor, and the company moved it into clinical trials.
Lilly had little interest in a serotonin reuptake inhibitor. But in line with standard practice at the time, the other compounds in the serieswere investigated. David Wong, a biochemist with little experience inpsychopharmacology, tested all of Molloy’s new series for serotonin re-uptake inhibiting properties. Several of them came out as serotonin in-hibitors, but LY-82816 stood out as the compound with the fewest ef-fects on the norepinephrine system. This compound, however, was dif-ficult to work with, as it couldn’t easily be dissolved, so it wasreformulated as a chloride salt, becoming LY-110140. At this point, workon LY-110140 was an academic exercise, meriting publication in a jour-nal, the first specifically about a serotonin reuptake-inhibiting drug.111On this basis, Wong is sometimes described as the discoverer of Prozac.
As reports of Zelmid’s progress came through, Frank Bymaster and Ray Fuller, pharmacologists with the company, looked at LY-110140’seffects on behavior. They screened it for antidepressant activity. Thebest known screening test involved trying to block the sedative effectsof reserpine on animals. All of the antidepressants then on the marketblocked reserpine-induced sedation. LY-110140 did not.112 Another test was a rat aggression model. If a drug made rats more aggressive and more likely to attack other rats, conventional wisdomhad it that such drugs had stimulant properties and might be useful inthe treatment of depression. LY-110140 increased aggression in rats.
I N T RO D U C T I O N : B E F O R E P RO Z AC Around 1975, therefore, Lilly had a compound with a relatively un- usual biochemical effect, and some poorly characterized behavioral ef-fects, but otherwise a mystery. Carlsson’s work suggested such a com-pound might be useful for treating nerves or depression, but most com-panies were adopting a wait-and-see attitude to these claims.
On September 11, 1975, LY-110140 was named fluoxetine. What was its future? There were a number of possibilities. Serotonin drugsshowed antihypertensive properties in animal models, and the marketfor pills to lower blood pressure was much larger than the market forantidepressants. Had fluoxetine shown any clear antihypertensive ac-tion in humans, there is little doubt it would have been developed as anantihypertensive. The “behavioral effects” would then have been writ-ten out of the script in the course of a market development program em-phasizing the rational engineering of a selective antihypertensive.
There were other lucrative possibilities. Early screening suggested that the drug might produce weight loss. An anti-obesity agent was cer-tain to make vastly more money than an antidepressant. The hint thatProzac had weight-reducing properties probably drove some of theearly mania, which later helped Prozac to hit the road running in 1989.
This idea was still a big part of the fluoxetine development program aslate as 1990, when the company hoped to license fluoxetine in a 60 mgpill under the trade name Loban for eating disorders. The vast amountsof money to be made in this market by drugs active on the serotonergicsystem made Redux headline news a few years later.113 A number of clinical investigators were invited by Lilly to a con- sultancy meeting in Britain in the late 1970s. One was Alec Coppen, aleading psychopharmacologist and one of the first advocates of theserotonin hypothesis of depression. They were presented with data ona range of Lilly compounds. Coppen recalls suggesting that fluoxetinemight be an antidepressant, only to be met with a reply that if fluoxe-tine was ever developed, there was little chance it would be for depres-sion.114 Lilly had good reasons to think this way, based on early clinical studies done with the drug. These were aimed at testing whether thedrug was tolerable and what its behavioral effects might be in humans.
The first trialist was Herbert Meltzer, who had a long-standing interestin the capacity of antipsychotics to cause side effects such as akathisiaand Parkinsonism.115 When one of his patients on fluoxetine developed I N T RO D U C T I O N : B E F O R E P RO Z AC muscle spasms (a dystonic reaction), Meltzer was certain the patienthad accidentally been given an antipsychotic; but then other patientsdeveloped akathisia and a range of other problems more commonly as-sociated with antipsychotics than antidepressants.116 Meltzer found little or no effect on depression. Other senior clini- cians found something similar.117 Most centers had patients who be-came agitated or akathisic. This led to recommendations from Lillymonitors that at least some patients be put on benzodiazepines whilethey were taking fluoxetine.118 A far-reaching implication of this is thatthere is probably no clinical trial in which fluoxetine on its own has beenshown to be an “antidepressant.” In this patient group, benzodi-azepines may well have been just as effective as fluoxetine itself.
In addition to trying to determine how tolerable fluoxetine was for patients, Lilly was trying in these early trials to establish whether therewere any conditions for which it might be a suitable treatment. Thecompany persuaded clinicians to try it out in patients with atypical psy-chotic disorders, as well as patients hospitalized with depressive disor-ders. It turned out to be ineffective in both groups: It made patients withpsychotic features worse, and it has never been shown to work for hos-pitalized depression cases. Development of fluoxetine was at a crisispoint.
Irwin Slater, a veteran of drug development within the company, was drafted to take over the clinical trials program. He tried fluoxetineout for pain syndromes, obesity, and other conditions, with no greatluck.119 Senior management were pushing to shelve the compound, butSlater and Fuller pointed out that zimelidine was almost through a clin-ical trials program for depression and fluvoxamine not far behind. Thehierarchy relented. A clinical trial program began chasing milder de-pressions. Louis Fabre, later investigated by Upjohn for “recruiting pa-tients from a half-way house for alcoholics,”120 was approached. Hegave fluoxetine to five patients; all responded. This turned the tide.
With fluoxetine rescued, the next step was to think about how to brand it. Lilly turned to Interbrand, who later claimed they invented the“discipline of naming” in the late 1970s.121 The success of the nameProzac changed how drugs were named. James Singer, who later left In-terbrand to set up his own NameBase/MediBrand, worked on the flu-oxetine project.122 Prior to Prozac, drugs had names that sounded sci-entific and referred in some way to the actual compound. Unlike Luvoxor Zelmid, for instance, which referred to the original pharmacological I N T RO D U C T I O N : B E F O R E P RO Z AC name, the name Prozac was seemingly designed to convey profession-alism (through its “pro-” element) and the ability of the medication totarget the right area for treatment (through its “-zac” element).
Prozac’s close brush with extinction may have had one long-lasting consequence. Many clinicians have wondered why Lilly didn’t bring inlow doses of Prozac. With a 5 mg dose, some of the problems thatemerged at higher doses of Prozac might have been minimized. Theconventional explanation was that Lilly had a brilliant marketing strat-egy, which involved selling one pill at one dose—something any foolcould give. In a later deposition, former marketer Richard Wood, whobecame chief executive of the company in the late 1980s, provided agreat deal of evidence for the sales-driven “one pill fits all” formula.123 But another possibility lies in the alarming early history of Prozac.
In an effort to make this drug work, the company pushed the dose upto 80 mg a day. In the mid-1980s, when FDA officials were finding it dif-ficult to be certain that even high doses of Prozac worked, a Lilly studydemonstrated that for the new mild depression market they were in-vestigating, 5 mg was as effective as 20 or 40 mg.124 Joachim Wernicke(later in charge of clinical trials in Lilly) e-mailed his colleagues: How much do we want to say about the 5 mg? I hedged a little, withthe thought that we may be able to show that 5 mg is not as good onall measures. Some day we will have to report it if we ever want to usethe information.125 In 1986, Stuart Montgomery reported that in milder depression, one pillper week was as good as one pill a day.126 This study vanished quietly,even though in 2001, after Prozac had lost its patent, Lilly would mar-ket a once-a-week form of treatment.
Lilly and Prozac were ultimately saved by two changes in the clinicaltrial world. In the United States, federal support for psychopharmacol-ogy research had all but shut down by the end of the 1970s. Three thingsdrove this:127 the financial crisis caused by the Vietnam War; the suspi-cion with which the Nixon administration viewed scientists; and esca-lating health service costs. Removal of federal support meant the end of I N T RO D U C T I O N : B E F O R E P RO Z AC independent clinical trials in psychopharmacology. Increasingly, formany investigators the only way to do research was to participate in in-dustry-run trials.
Industry paid clinical investigators up to $5,000 per patient entered in a study and looked with favor on investigators who could recruit pa-tients quickly. The faster the trials were completed and submitted to theFDA, the better the chances of registering a compound early in itspatent life, and the better the returns on the investment. Naturally, in-dustry would also look favorably on investigators who managed toproduce the right results. This led to a situation by the end of the 1980swhere some investigators were reporting on patients who didn’t exist,or on “professional” patients who might have been on several investi-gational compounds at much the same time. In some instances, indi-viduals other than clinicians conducted both recruitment and ratings.128 The second change occurred within the FDA. In 1981, Paul Leber, formerly a pathologist and then a psychiatrist, became disenchantedwith clinical work in New York and moved to a job in the Central Ner-vous System (CNS) division of the FDA.129 Quickly promoted to divi-sion head, Leber became a pivotal figure over the course of the next fif-teen years. His first innovation all but brought the house down. Look-ing at antidepressant trials conducted up until then, Leber made thepoint that in a trial where a new antidepressant is compared to an olderone and shown to be no worse than the older compound, where every-body assumed this meant that the new compound worked just as wellas the old compound, it was quite possible neither drug was working.
Proof that a new compound worked came only from trials against aplacebo.
There was uproar. Many clinical trial programs did not include placebo trials, setting development back several years, at considerablecost. Worse yet, several new compounds could not be shown to workwhen compared against a placebo. Mianserin, the best-selling antide-pressant in Europe, failed in U.S. placebo-controlled trials. This mayhave been because these trials were conducted in a too mildly de-pressed group, where it is difficult to demonstrate that any antidepres-sant is superior to placebo; but the prospect was raised that it might beincreasingly difficult for new antidepressants to be registered, owing todifficulties in showing that they were superior to placebo.
This was a real problem for the emerging SSRIs, which couldn’t be shown to work in hospital depression. Prozac accordingly had to make I N T RO D U C T I O N : B E F O R E P RO Z AC a tricky passage between the Scylla of serious depression—where itdidn’t work—and the Charybdis of mild depression—where mianserinhad come to grief. Leber’s reforms required that a new drug show evi-dence from two pivotal studies that it worked, and the majority of stud-ies performed should go the same way.130 The term pivotal study hadcrept into use as accepted code for a placebo-controlled study. As an in-creasing number of new antidepressants were SSRIs, and all took five toten years to develop, Leber, the FDA, and others must have been gen-uinely concerned that what appeared to be a necessary reform to thesystem might backfire. For a time, no new drugs made it to the U.S.
market, which probably contributed significantly to Prozac’s impact asthe first antidepressant to hit that market for what was, relativelyspeaking, a long time.
In the case of Prozac, there were three placebo-controlled studies.
Karl Rickels from Philadelphia conducted one; it demonstrated no ef-fect for Prozac. A second was a six-centered study, called Protocol 27,where Prozac was compared to imipramine and placebo. One of the in-vestigators here was Jay Cohn from Los Angeles. Dr. Cohn’s data werelater removed at the request of FDA, as the extremely favorable resultshe reported were at odds with the other data generated.131 When the re-maining studies were combined, Prozac was inferior to imipramine andbarely better than placebo on some scales, and no better than placebo onothers. Three of the six centers failed to show it worked better than aplacebo. In the final study, by Louis Fabre, there were only eleven com-pleters on Prozac, and the study period was effectively only four weeksin duration. This study came up with a positive result for Prozac. Withthe Fabre study, and counting Protocol 27 as one study, the score wastwo to one in favor of Prozac. If the component centers of the multicen-ter study are counted separately, the result was four centers in favor ofProzac and four against—hardly an overwhelming majority.132 The plan had been to launch Prozac in the United States in 1986.
The FDA finally approved it in late 1987, after over three years ofscrutiny during which the agency noted serious flaws in the designs ofits clinical trials.133 A pattern of approval of less effective antidepres-sants had begun. Since then, it has not been uncommon for new drugsto be presented to the FDA that are superior to placebo in perhaps onlytwo out of six trials.134 Instead of saying that, on balance, the new drugis simply not more effective than a placebo or is of such minimal ef-fectiveness that it is hardly worth permitting on the market, the FDA I N T RO D U C T I O N : B E F O R E P RO Z AC approach is to say that any trials comparing a new drug to an older an-tidepressant in which the older drug appeared no different to a placebowere failed trials. That is, the trial rather than the new drug has failed.
These ambiguities can be seen clearly in the license application for Zoloft, where only one of five studies indicated superiority overplacebo.135 A further study claimed superiority for Zoloft when somepatients who had their treatment discontinued “relapsed,” but in thiscase, what was termed “relapse” may have been withdrawal fromZoloft. Zoloft did worse than amitriptyline and failed in two hospitaldepression studies.136 As Paul Leber ended up putting it: How do we interpret . . . two positive results in the context of severalmore studies that fail to demonstrate that effect? I am not sure I havean answer to that but I am not sure that the law requires me to have ananswer to that—fortunately or unfortunately. That would mean, in asense, that the sponsor could just do studies until the cows come homeuntil he gets two of them that are statistically significant by chancealone, walks them out and says he has met the criteria.137 For those who believe that approval of a drug by the FDA means that itis in some sense good for you if taken properly, the situation is evenmore problematic than the above scenario might suggest. “Two positivestudies” doesn’t mean the drug works for depression in two studies. Itmeans there are two studies in which the drug can be shown to have aneffect in depression—can be shown to do something. Whether it is agood idea to take any of these drugs is not addressed. In other words,these trials do not offer evidence the drug works in the sense that mostpeople mean by the word works—that is, evidence that this drug clearsthe problem up.
Companies marketing their products do not have to reveal any- thing about the very weak evidence on which registration was based.
The new compound can be sold with all the glossy slogans of rationalengineering, hints of added benefits for weight loss or whatever, andcelebrity endorsements. Since the end of the 1980s, companies have nothad to bother with any questioning by independent investigators aboutjust how good their compound really is, as there are increasingly few in-dependent investigators left in psychopharmacology.138 Adopting the principle in trials for mild to moderate depression that a drug is an antidepressant if in some instances it is possible to I N T RO D U C T I O N : B E F O R E P RO Z AC show a difference to a placebo, it would almost certainly be possible toshow that stimulants such as dexamphetamine, methylphenidate, andeven nicotine are antidepressants. Using these rules, in a population ofmild to moderately depressed patients, the benzodiazepines wouldcome out as antidepressants. The fact that no one has done such trialsowes everything to a business calculation: These older drugs are offpatent, so no company stands to make money from them.
There is a key point to take from all this. We are accustomed to the notion that our regulators are looking after us, that they are acting insome sense as consumer watchdogs. But this is not their role. The roleof a regulator is to adjudicate on whether, for example, a yellow sub-stance in front of him or her meets minimal criteria for butter; to ensure,for example, that it is not lard injected with color. Regulators are notcalled upon to determine whether this butter is good butter or not, orwhether butter is good for your health. Consumer watchdogs must dothat. Within medicine, the physician is supposed to be the consumer’swatchdog—which, given that they rarely consume the product, makesfor an ambiguous and commercially unique situation.
The legal situation is even more complex. If the mandate of the reg- ulators is to let drugs onto the market that are proven to have an effect,what the FDA have been doing is defensible. If the brief, as per thewording of the statutes in most countries, is to license drugs that are ef-fective, then the case of the SSRIs is much less certain.
Prozac was launched in the United States and Canada in 1988. InBritain, a launch was planned for 1984 but did not materialize until latein 1989.139 At the time, Britain was seen as quick to approve drugs, whilethe FDA was widely criticized for taking much longer than other regu-lators. Prozac bucked this trend.
In line with a strategy developed in 1983, the early sales pitch in most countries emphasized that Prozac lacked the supposedly nastyside effects of the older tricyclic antidepressants but was as efficaciousas these drugs and came in a convenient once-daily dosage.140 Borrow-ing from the mianserin story, there was an emphasis on safety in over-dose. Market surveys before launch repeatedly asked people like mewhether the fact that this new drug did not cause weight gain would I N T RO D U C T I O N : B E F O R E P RO Z AC influence our prescribing. Right from the time of its launch in America,patients were lining up asking for Prozac by name, an experience newto American psychiatrists.
Over the course of the next few years, every company with an SSRI ran clinical trials comparing their drug with the older tricyclic antide-pressants. When all clinical trials were analyzed together, the SSRIswere no more effective in outpatient depressions than were the olderagents.141 As for the tolerability profiles of the SSRIs compared with theolder drugs, the dropout rate of patients from clinical trials was almostequal—it would take over thirty patients assigned to either set of drugsbefore there would be one less dropout in the SSRI group. This was thecase even though these trials were almost exclusively designed by SSRIcompanies, so that in over 30 percent of trials, the SSRI had been pitchedagainst the tricyclic generally thought to have the most side effects—amitriptyline.142 There was therefore an extraordinary contrast betweenthe marketing hype and the trials underpinning it. When these studieswere analyzed, the greatest predictor of outcome lay in the sponsorshipdetails of each study.143 Later in the 1990s, it became clear that a largenumber of trials with less favorable results for the SSRIs were simplynot reported, and that the results on quality-of-life scales used in manyof these trials were almost universally left unreported.
The trump card for the SSRIs has been that a greater number of pa- tients are likely to be put on and remain on what is thought to be a“therapeutic dose” of the drug than is likely with other agents. But evenhere, the puzzle is that no more than 40 percent of patients take theirdrugs for more than few weeks.144 Something goes wrong with theother 60 percent. This “something” tempts clinicians to blame patientsand tempts the “experts” in the field to blame the average clinician, whosupposedly hasn’t stressed to the patient the importance of remainingon treatment for six months or more. Nowhere in the literature is thereany concession to the possibility that SSRIs may not suit up to 60 per-cent of those put on them.
By the time Prozac got its license, the crisis over the benzodi- azepines had become severe. The psychiatric and primary-care worldswere receptive to the idea that behind every case of anxiety lay a case ofdepression. No one was inclined to question the idea that antidepres-sants were a more scientifically rational treatment for the nervous statespresenting in the community than anxiolytics, especially when no onethen expected an antidepressant to produce dependence. Add in the I N T RO D U C T I O N : B E F O R E P RO Z AC fact that, compared with older antidepressants, these new drugs weresafe in overdose, and there seemed to be no good reason not to prescribethem.
The plans had been to launch fluoxetine in Germany in 1984, but it took six more years for “Fluctin” to reach the market there. Probablyvery few people outside Eli Lilly knew of the German regulators’ viewon fluoxetine as of May 1984: Considering the benefit and the risk, we think this preparation totallyunsuitable for the treatment of depression.145

Source: http://www.nyupress.org/webchapters/0814736696intro.pdf