Nothing in all this necessarily contradicted anything that Donald Klein would later glean from his imipramine experiments. But that’s because in 1895 Freud still considered anxiety neurosis to be the product not of a “repressed idea” (which is what he believed underlay most psychopathology) but of a biological force. Anxiety neurosis, Freud theorized in these early writings, was the result of either genetic predisposition (a theory that modern molecular genetics supports) or some kind of pent-up physiological pressure—most notably, in Freud’s imagining, the pressure caused by thwarted sexual desire.
But in many of his subsequent writings (starting with Studies on Hysteria from around this same time), Freud asserted instead that attacks of anxiety—even ones that manifested themselves in acute physical symptoms—emanated from unresolved, and often unconscious, inner psychic conflicts. For nearly thirty years, Freud effectively abandoned the argument that anxiety attacks were a biological problem. He and his followers replaced anxiety neurosis with plain neurosis—a problem that had its basis in psychic discord, not in genes or biology. By the mid-twentieth century, the overwhelming psychiatric consensus was that anxiety was the result of a conflict between the desires of the id and the repressions of the superego—and that, furthermore, anxiety was the foundation of almost all mental illness, from schizophrenia to psychoneurotic depression. One of the main purposes of psychoanalysis—and of most forms of talk therapy—was to help the patient become aware of and contend with the underlying anxiety against which all of his various maladaptive “ego defenses” had been mustered. “The predominant American psychiatric theory was that all psychopathology was secondary to anxiety,” Klein later recalled, “which in turn was caused by intrapsychic conflict.”
But this didn’t square with what Klein was finding in his imipramine studies. If anxiety was the animating force behind all psychopathology, then why didn’t imipramine—which seemed to eliminate the panic suffered by patients with anxiety neurosis—help schizophrenics with their psychosis? Maybe, Klein ventured, not all mental illnesses lay on the spectrum of anxiety the way the Freudians thought they did.
The spectrum theory of anxiety held that what determined a mental illness’s severity was the intensity of the underlying anxiety: mild anxiety led to psychoneurosis and various neurotic behaviors; severe anxiety led to schizophrenia or manic depression. For many traditional Freudians, the settings—like bridges or elevators or airplanes—that tended to produce acute anxiety attacks had symbolic, and often sexual, significance that accounted for the anxiety they caused.
Balderdash, Klein said. Childhood trauma or sexual repression didn’t cause panic; a biological malfunction did.
Klein concluded that these attacks of paroxysmal anxiety—which he would come to call panic attacks—originated in a biological glitch that produced a suffocation alarm response, his term for the cascade of physiological activity that leads to, among other things, what feels like a spontaneous attack of overwhelming terror. Anytime someone starts to asphyxiate, internal physiological monitors detect the problem and send messages to the brain, causing intense arousal, gasping for breath, and an urge to flee—an adaptive survival mechanism. But some people, according to Klein’s false suffocation alarm theory, have defective monitors that occasionally fire even when the individual is getting enough oxygen. This causes the person to experience the physical symptoms that make up an anxiety attack. The source of panic is not psychic conflict but crossed physiological wires—wires that imipramine somehow untangled. Klein’s data suggested that imipramine eliminated spontaneous anxiety attacks in most patients who suffered them.
When Klein published an initial report on imipramine in The American Journal of Psychiatry in 1962, it was received, as he recalled, “like the proverbial lead balloon.” Subsequent articles over the next several years, in which he argued that panic anxiety was an illness distinct from chronic anxiety, were received with similar froideur. He was attacked from all sides, accused of being an apostate. But because imipramine seemed to cure panic anxiety without affecting feelings of general apprehension and neurosis, Klein remained convinced that panic anxiety had symptoms and physiological causes that were different in kind and not just in degree from other forms of anxiety.
Though he hadn’t set out to, Klein had achieved what’s known as the first pharmacological dissection: working backward from the effects of drug treatment, he had defined a new illness category, carving out panic anxiety from the more general anxiety that was supposed to underlie the Freudian neuroses.
Klein’s pharmacological dissection of anxiety met with enormous hostility from his colleagues. At a conference in 1980, just as the publication of the DSM-III was bringing panic disorder into existence, Klein’s presentation on how the suffocation alarm response caused panic anxiety was followed immediately by a lecture from John Nemiah, the longtime editor of the prestigious American Journal of Psychiatry. Rebutting Klein, Nemiah said that panic anxiety had nothing to do with a suffocation alarm response or problems of biological hardwiring but rather was “the reaction of the individual’s ego … to the threatened emergence into conscious awareness of unpleasant, forbidden, unwanted, frightening impulses, feelings, and thoughts.”
Though Klein’s theory has been to some degree officially adopted by American psychiatry since 1980, it remains controversial today. My own current therapist, Dr. W., a PhD psychologist who did his training in the 1960s, laments that Klein’s work led to a fundamental shift in how we think about mental illness, moving us from the dimensional model that prevailed through the age of the DSM-II to the categorical model that began with the publication of the DSM-III in 1980. According to the dimensional model, depression, neurosis, psychoneurosis, panic anxiety, general anxiety, social anxiety, obsessive-compulsive disorder, and so forth all exist along a spectrum emanating from the same roots in what Freud called intrapsychic conflicts (or Dr. W.’s “self-wounds”). According to the categorical model, as laid out in the DSM since its third edition, depression, panic anxiety, general anxiety, social anxiety, obsessive-compulsive disorder, and so forth are carved up into discrete categories based on distinctive symptom clusters that are believed to have different underlying biophysiological mechanisms.
Between 1962, when Klein published his first imipramine study, and 1980, when the DSM-III was published, the way psychiatry (and the culture generally) thought about anxiety underwent an enormous transformation. “It is hard to recall that fifteen or twenty years ago there was no such concept [as panic anxiety],” marveled Peter Kramer, a psychiatrist at Brown University, in his 1993 book, Listening to Prozac. “Neither in medical school nor psychiatry residency, both in the 1970s, did I ever meet a patient with ‘panic anxiety.’ ” Yet today panic disorder is a frequently diagnosed disease (some 18 percent of Americans are estimated to suffer from it), and “panic attack” has transcended the psychiatric clinic to become part of our lingua franca.
Panic disorder was the first psychiatric disease for which the determining factor in its creation was a drug reaction: imipramine cures panic; ipso facto panic disorder must exist. But this phenomenon—in which a drug effectively defined the syndrome for which it was prescribed—would soon recur.
DSM assigns each slice of craziness with a name and a number. Panic disorder, for example, is disease number 300.21, a diagnostic code.… But just because it has a name, is it actually a disease?
—DANIEL CARLAT, Unhinged: The Trouble with Psychiatry—a Doctor’s Revelations About a Profession in Crisis (2010)
An advertisement for an October 1956 public talk by Frank Berger, the inventor of Miltown, stated that tranquilizers were effective in treating high blood pressure, worry, jitters, “executive stomach,” “boss nerves,” and “housewife nerves.” None of these ailments were then, or are now, listed as illnesses in the American Psychiatric Association’s Diagnostic and Statistical Manual—which raises the question of whether Miltown prescriptions were aimed less at treating actual psychiatric disor
ders than at treating the age itself—at mitigating the effects of what Berger in this talk called “today’s pressure living.”
Every time new drug therapies come along, they raise the question of where the line between anxiety as psychiatric disorder and anxiety as a normal problem of living should get drawn. We see this again and again throughout the history of pharmacology: the rise of tranquilizers is followed by an increase in anxiety disorders diagnoses; the rise of antidepressants is followed by an increase in the rate of depression.
When the APA published the first edition of the DSM in the aftermath of World War II, the governing infrastructure of the profession was still Freudian: the first edition placed all disorders along a spectrum of anxiety. “The chief characteristic of [neurotic] disorders is ‘anxiety,’ ” the manual declared, “which may be directly felt and expressed or which may be unconsciously and automatically controlled by the utilization of various defense mechanisms.” The second edition, published in 1968, was even more explicitly psychoanalytic. When the APA decided in the 1970s that it was time for a third edition, the Freudians (who had dominated the task forces that wrote the first two editions) and the biological psychiatrists (who had gotten a boost from the recent findings of pharmacology research) prepared for a pitched battle.
The stakes were high. Doctors and therapists of different schools would see their professional fortunes rise or fall depending on whether the definitions of the diseases they specialized in were narrowed or expanded. Drug company profits would spike or plummet depending on whether the categories that were created could be targeted by—and could help secure FDA approval for—the medications they manufactured.
The publication of the DSM-III in 1980 represented at least a partial repudiation of Freudian concepts and a victory for biological psychiatry. (One medical historian called the DSM-III a “death thrust” for psychoanalysis.) Out went the neuroses, and in their place came the anxiety disorders: social anxiety disorder, generalized anxiety disorder, post-traumatic stress disorder, obsessive-compulsive disorder, and both panic disorder with agoraphobia and panic disorder without agoraphobia. Donald Klein’s pharmacological dissection of panic had prevailed.†
But in moving mental illness away from Freudianism and into the realm of medical diagnosis, the new DSM pathologized as “disordered” or “ill” many people who once would have been considered merely “neurotic.” This was a boon for the drug companies, which now had many more “sick” people for whom to develop and market medication. But did it benefit patients?
That’s a complicated question. On the one hand, the medicalization of depression and anxiety helped destigmatize conditions once regarded as shameful character weaknesses, and it allowed people to find relief (often pharmacological) from their misery. The number of people seeing depression or anxiety disorders as a health problem—as opposed to evidence of personal weakness—grew dramatically between 1980 and 2000, as Prozac and other SSRIs provided additional evidence for the idea of depression as a problem of chemical imbalance.‡ On the other hand, the expansion of medical categories of mental illness had the effect of drawing countless mentally healthy people into the nets of the pharmaceutical companies. Before the arrival of MAOIs and tricyclics in the late 1950s, depression (and its predecessors) was a rare diagnosis, given to only about 1 percent of the U.S. population. Today, by some official estimates, it’s a diagnosis given to up to 15 percent of us. Are we truly that much more depressed in 2011 than we were in 1960? Or have we defined depression and anxiety disorders too broadly, allowing the drug companies to bamboozle us (and our insurance companies) into paying for pills that treat diseases we didn’t know we had, diseases that didn’t exist before 1980?
The publication of each successive edition of the DSM has intended to convey the impression of science advancing. And, to be sure, the DSM-III, DSM-IV (which came out in 1994), and DSM-V (which came out in 2013) were more empirically grounded than the first two editions. They placed much less influence on etiology—that is, on the presumed causes of different illnesses—and much more on simple symptom description.§ But they were still political documents as much as scientific ones, representing the claims of one psychiatric school over another—and the professional interests of psychiatrists above everything else. “It is the task of the APA”—and therefore of the DSM—“to protect the earning power of psychiatrists,” Paul Fink, the vice president of the American Psychiatric Association, declared in 1986. Stuart Kirk and Herb Kutchins, social workers who have together written two books about the history of the DSM, say that the APA’s so-called bible is “a book of tentatively assembled agreements” that has led to the “pathologizing of everyday behaviors.”
When you probe more deeply into the process that produced the DSM-III, its pretensions to scientific rigor begin to seem rather strained. For starters, some of its new category distinctions appear awfully arbitrary. (Why does panic disorder require the presence of four symptoms, rather than three or five, from the list of thirteen? Why do symptoms have to persist for six months, and not five or seven, for an official diagnosis of social anxiety disorder?) The head of the DSM-III task force, Robert Spitzer, would concede years later that many of its decisions were made haphazardly. If a constituency lobbied hard enough for a disease, it tended to get incorporated—which helps explain why between its second and third editions the DSM grew from 100 to 494 pages and from 182 to 265 diagnoses.
David Sheehan worked on the DSM-III task force. One night in the mid-1970s, Sheehan recalls, a subset of the task force got together for dinner in Manhattan. “As the wine flowed,” Sheehan says, the committee members talked about how Donald Klein’s research showed that imipramine blocked anxiety attacks. This did seem to be pharmacological evidence of a panic disorder that was distinct from other kinds. As Sheehan puts it:
Panic disorder was born. And then the wine flowed some more, and the psychiatrists around the dinner table started talking about one of their colleagues who didn’t suffer from panic attacks but who worried all the time. How would we classify him? He’s just sort of generally anxious. Hey, how about ‘generalized anxiety disorder’? And then they toasted the christening of the disease with the next bottle of wine. And then for the next thirty years the world collected data on it.
Sheehan, a tall Irishman who today runs a psychiatric center in Florida, is regarded as something of an apostate by the profession. He cheerfully admits he is out to “sabotage the notion” that panic disorder is truly distinct from generalized anxiety disorder. So his jaundiced version of how generalized anxiety disorder was born should perhaps be viewed with some skepticism. But Sheehan, who has been studying and treating anxiety for decades, makes an important point: Once you create a new disease, it starts to take on a life of its own. Research studies accrete around it, and patients get diagnosed with it, and the concept saturates the psychiatric and popular culture. Generalized anxiety disorder, a disease conceived over a boozy dinner and written into the DSM with a fairly arbitrary set of criteria, has now had thousands of studies applied to it, and the FDA has approved multiple drugs for treating it. But what if, as Sheehan contends, there is no such thing as generalized anxiety disorder—at least not as a disease distinct from panic disorder or major depression?‖ If Sheehan is right, a large edifice of diagnosis, prescription, and academic study has been built upon something—generalized anxiety disorder—that is presumed to exist in nature but that in fact does not.
[At current rates of Valium use], the arrival of the millennium would coincide with the total tranquilization of America.
—“BENZODIAZEPINES: USE, OVERUSE, MISUSE, ABUSE,” EDITORIAL IN THE LANCET (MAY 19, 1973)
Even as Thorazine emptied the asylums and antidepressant prescriptions grew exponentially through the late 1950s, nothing approached the runaway commercial success of Miltown. Hence the instructions given to Leo Sternbach, a chemist at Hoffmann–La Roche in New Jersey: “Invent a new tranquilizer,” his bosses told him. So Sternbach though
t back to research he had done on heptoxdiazine-based dyes while a postdoctoral student in Poland in the 1930s. What would happen if he chemically modified them a little? He tested more than forty different variations on animals—but none seemed to have a tranquilizing effect. Hoffmann–La Roche abandoned the project. Sternbach was reassigned to work on antibiotics.
But one day in April 1957, a research assistant who was cleaning Sternbach’s lab came across a powder (official name: R0-5-090) that had been synthesized a year earlier but was never tested. Without hope, as Sternbach said later, he sent it over to the animal testers on May 7, his forty-ninth birthday. “We thought that the expected negative result would complete our work with this series of compounds and yield at least some publishable material. Little did we know that this was the start of a program which would keep us busy for many years.”
Happy birthday. Sternbach had, mostly by accident and nearly without realizing it, invented the first benzodiazepine: chlordiazepoxide, which would be given the trade name Librium (derived from “equilibrium”) and was the forerunner of Valium, Ativan, Klonopin, and Xanax, the dominant antianxiety medications of our age. Because of an error in his chemical process, R0-5-090 had a molecular structure different from the other forty compounds Sternbach had synthesized. (It had a benzene ring of six carbon atoms connected to a diazepine ring of five carbon atoms and two nitrogen atoms—thus “benzodiazepine.”) The director of pharmacological research at Hoffmann–La Roche tested the new substance on cats and mice and discovered, to his surprise, that although it was ten times more potent than Miltown, it did not notably impair the animals’ motor function. Time reported that keepers at the San Diego Zoo had tamed a wild lynx with Librium. A newspaper headline blared: “The Drug That Tames Tigers—What Will It Do for Nervous Women?”
My Age of Anxiety: Fear, Hope, Dread, and the Search for Peace of Mind Page 23