You are here: Biographies / Eulogy for Donald F. Klein (1928-2019) by Barry Blackwell. / Thomas A. Ban’s Comments
Tuesday, 18.05.2021

Barry Blackwell: Eulogy for Donald F. Klein (1928-2019)

Thomas A. Ban’s comments


         In these comments I would like to provide some background information to Barry Blackwell’s eulogy of Don Klein, posted on our website December 5, 2019. To recapitulate briefly Barry’s account: In 1962, Don discovered that imipramine reduced the frequency of anxiety attacks in hospitalized psychiatric patients (Klein and Fink 1962). He followed up this lead and in 1964, on the basis of pharmacological responsiveness he referred to as “pharmacological dissection,” he separated within ”anxiety disorder” two pharmacologically distinctly responding syndromes, one characterized by “recurrent anxiety attacks.”  He coined the term “panic disorder” as a label for this population and it was adopted as an Axis I diagnosis in 1980 in the Third Edition of the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association, DSM-III (American Psychiatric Association 1980; Klein 1964)

         Klein was among the first, in the 1960s, to recognize the importance of psychiatric diagnosis in predicting drug effects (Klein 1967).   In 1969 he published his first text, Diagnosis and Treatment of Psychiatric Disorders, co-authored by John Davis (Klein and Davis 1969).

         Pursuing his research with pharmacological dissection further, in the 1970s he identified “endogenomorhic depression” (Klein 1974). Then, after extending the scope of his research to include children, he introduced imipramine in the treatment of “separation anxiety” (Klein and Gittelman-Klein 1978; Klein, Gittelman-Klein, Quitkin and Rifkin 1980.)

         In the 1990s, with consideration of findings that panic attacks could be triggered by an increase of carbon dioxide in arterial blood flow, Klein re-conceptualized anxiety by formulating his hypothesis that “spontaneous panics” are “false suffocation alarms” (Klein 1993; Papp, Klein and Gorman 1993; Papp, Martinez, Klein et al 1989).

         Finally, in 2010, his research team reported findings which were suggestive of the involvement of the endogenous opioid system in panic attacks (Preter, Lee, Petkova, Vemucci and Klein 2010).

         Don Klein was interviewed twice in the “oral history” project of the American College of Neuropsychopharmacology. The first interview was conducted by Leo E. Hollister, in San Juan, Puerto Rico, in 1994 and the second by John M. Davis, in Boca Raton, Florida, in 2007. Both interviews took place at the annual meeting of the College.

         In the following the edited transcripts of these interviews are presented.  

First interview of Donald F. Klein

Interviewed by Leo E. Hollister, San Juan, Puerto Rico, December 13, 1994

LH: I am Leo Hollister and I am interviewing Donald Klein. He has been president of this great organization and he’s been Mr. Panic Disorder for the last 30 years.  In fact, I’m normally against new diseases, syndromes and signs, but if somebody came up with the idea of a Klein Syndrome for Panic, I would be perfectly agreeable to it.  But, how did this all get started, Don?  What made you go into medicine, in general, and psychiatry, in particular?

DK: I wanted to be a scientist since childhood. No one in my family but my father fostered this by regular trips to the museum of Natural History and Planetarium. Actually, when I was in high school at the Bronx High School of Science, a great experience, I wanted to go into Chemistry as a research scientist. I had been fooling with chemistry sets since I was a kid.  I loved it and did well in Chemistry. And, then, when I got to college, I stumbled onto Freud. I got to college young; about 15 when I started. And I was wandering around the stacks of books in the library one day and found Freud’s books; he was talking about all the things I was interested in.  It was mostly sex and I figured this guy must have something going for him.  So, I really got interested and read a great deal of what he wrote; my desire at the time was to be a research psychoanalyst. And I understood that to be a psychoanalyst you have to be an MD. It struck me that was pretty foolish, but I didn’t mind becoming an MD.

LH: So, you choose to become a psychiatrist before you actually entered medical school.  You were going into psychiatry to be an analyst.

DK: Exactly, psychiatry was sort of a steppingstone to become an analyst.  Anyway, in 1948 I graduated from college, top of my class.  I was 18.  I couldn’t get into medical school, probably because the Vets had all come back then, but also anti-Semitism for medical school admission was very real. So, I spent a year in NYU graduate school in Biochemistry and Physiology, which was not a lost year as I learned how to use a library and some fundamentals of physiology and endocrinology. Finally, I got admitted to Long Island College of Medicine (LICM.) The only other acceptance was from Howard.  LICM was extremely clinical, generally considered a ‘baby catching’ trainee school. However, the clinicians were astute and critical, and the training to deal with patients has stood me in very good stead. In my senior year it became Downstate University.  Downstate made a budding hematologist out of me, because psychiatry was so lousy and hematology, as taught by Janet Watson, was really engaging. She was a pioneer in hemoglobin molecular structure and Sickle Cell anemia. I worked as a laboratory assistant to Norman Kretchmer who had his PhD and was in my MD class. We did paper chromatography when there were only two papers, winging it with Pyrex pie-plates. Norm went on to be the head of NICHD and remained a good friend, although he consistently referred to me as a “spook”. The psychoanalysts were terrible; they spent all their time reading to us from their textbooks, although we all knew how to read, and telling us if we had a question about anything that was our resistance or countertransference. But it didn’t turn me off completely. Anyway, I interned in the Public Health Service. It was in the Korean War and I hoped to spend two more years with the Public Health Service, taking care of tubercular Eskimos, rather than go to Korea. But I got fired at the end of my intern year, because Eisenhower had a reduction in force and the bottom half of the intern class was dropped. I was squarely in the bottom half, because I didn’t get along with them very well at all. I asked too many questions. However, their psychiatry rotation was mentored by two very bright psychoanalysts, Richard Silberstein and Milton Horowitz, who also thought questions were resistance, but were personally engaging and intellectually alive. This revived my psychoanalytic interests and I re-developed the misguided goal of being a research psychoanalyst. So, I was scurrying around. I had a wife and a kid at the time and felt lucky when I got a job as a first-year resident at the Creedmoor State Hospital, where they gave me a house and a gardener, a maid and, probably, a chauffeur.

LH: Who was in charge of psychiatry there at the time?

DK: Nobody. When I was a first-year resident, in 1953, Creedmoor was a 6,000-bed locked hospital and after I was given a book on the mental status, I was told that I should take care of my 300 patients upstairs. They also told me that the nurses would teach me how to do ECT but fortunately I knew that from my internship. That was all the training I got. It was a great experience, but also a great responsibility. The patients were fantastic.  We had no psychotropic drugs at all in 1953, except paraldehyde and amytal; all we had was ECT and nursing care.  And then, primarily because the draft was after me, I went back into the Public Health Service and spent two years at the Narcotic Hospital in Lexington, KY. That was a wonderful experience; that’s what’s turned me onto Psychopharmacology. 

LH: Now, you jumped out of the Public Health Service earlier, but, then, you went back into it?

DK: I called them up and I said, “Look, I’ve got a whole year of psychiatric residency, don’t you need me?”  And, they said, certainly, with a whole year of psychiatric residency, my goodness we certainly need you down in Lexington, KY to run the Admission and Withdrawal Service.  I didn’t know much about admission or withdrawal or narcotic addictions or anything at the time, but I went there and the guy in charge took me on rounds once. And from the next day I was on my own.  I had 70 beds to take care of using methadone withdrawal. It was a terrific place and I liked it a great deal.  I had complete misconceptions about what addiction and addicts were like.

LH: There were some giants working there at the time.

DK: Well, Abe Wikler who was probably the smartest guy I ever met, thinking deeply about psychiatry and pharmacology.  He wrote a book called The Relationship of Pharmacology to Psychiatry. During the period I was down there, I had the opportunity to discuss with him what the contents of the book and its layout should be.

LH: It’s a classic text.

DK: A classic book, which fell like a lead balloon, because it just came out after imipramine was introduced, so it had nothing about the antidepressants. And he missed the boat on lithium.  He felt that lithium was having its’ effects by toxicity. Yet because of his thinking and the discussion of how you go about studying drugs and relating them to psychiatry, the book may still be the best single volume in the field.  The book disappeared, so I don’t know anybody who knows about it anymore.

LH: Heavily underused, right?

DK: I tell people about it, especially my juniors.  Anyway, I got involved in Lexington with studies on reserpine, chlorpromazine, and LSD for a two-year period.  The LSD studies were being financed by a mysterious foundation. As we found out later, they were funded, by the CIA. The criteria for selecting subjects and the requirements for inclusion in the LSD studies were pretty clear. The people selected were from those with two, five-year federal prison terms, who considered themselves “stoned junkies” and were never going to recover. 

LH: A CIA front?

DK: The whole thing was due to brainwashing concerns, which was the glib explanation of the time. The Koreans had American pilots, who had been shot down, on TV saying they were capitalist stooges. You didn’t see the rifle pointing at them off screen, so it was assumed they had been brainwashed and LSD was the obvious culprit. Anyway, I made good friends with Wikler at that time although he was really put off when he understood I wanted to be a psychoanalyst. He thought that was not too smart.

LH: Heresy.

DK: But Lexington was actually an analytic hospital.  It was run just like Chestnut Lodge.  The head of the hospital was a training analyst and 50 patients were in intensive psychotherapy. They had a psychotherapist and an administrative therapist, who took care of all the grimy details, like parole. In retrospect, it was a completely bizarre setting.  I remember seeing at least one authentic miracle. I was in charge of a ward with 100 veterans who had been hospitalized continuously since World War I. They weren’t in the VA, because there wasn’t any VA for World War I veterans. They were under a thing called Executive Order and got to Lexington from the psychiatric unit at St. Elizabeth’s’ Hospital.

LH: But weren’t they drug-users?

DK: No, they were just plain army folks, who had gone psychotic during or just following World War I. Most of them sat around on benches and looked at the wall. They had excellent nursing care and did all sorts of interesting things with occupational therapy and psychotherapy. But whatever they got had not done any good to anyone. So, I decided to give them all Thorazine (chlorpromazine), 200 milligrams a day, which was a big dose then. One of them came up to me, after about six weeks, and said, “hey, Doc, when am I getting out of here?”  I never thought that could happen. It was really remarkable and made a big impression. I left shortly after and went back to Creedmoor, finished my residency and got into research. I was working first with a group of psychoanalysts, who were running an intensive psychoanalytic clinic devoted to six families with autistic children. These identical autistic twins walked around on their toes. I asked the supervising psychoanalyst how the mother had done that but was told this was resistance. That was somewhat disillusioning. When that boondoggle shut down, I worked for John Whittier who was unusual, an MD, PhD, psychoanalyst and veterinarian. We did one of the first controlled studies on mepazine, a drug that everyone said was terrific because it didn’t cause those terrible side effects and mental confusion like other phenothiazines. The only trouble was that it didn’t work.  It was the only phenothiazine taken off the market.  That experience reaffirmed that double blind, randomized, controlled studies were a pretty good thing to do. I went to the New York Psychoanalytic Institute about 1957, and that’s a long sad story, in itself.  Essentially, I burned out two analysts and they got rid of me. I ended up in 1959 at Hillside, working for Max Fink. Hillside was a psychoanalytic hospital, but Max Fink was a whole different character. He was a neurologist and psychiatrist who had psychoanalytic training and worked with Morris Bender. He was studying ECT and used it as an experimental treatment to be studied for its effects on brain function, rather than just as some sort of punishment. Also, he wanted to get involved in studying the new psychotropic drugs. And that’s how I got going.  Max was a complete nihilist.  He did not believe in diagnosis.  He thought there was no evidence for any of the diagnoses because people in reliability studies, making independent diagnoses, did a very bad job of it.

LH: Still do.

DK: Probably.  At least we have some inclusion/exclusion criteria now, but then we didn’t have anything.  So, to study the new drugs Max and I went to the head of the hospital, Lew Robbins who was an analyst, but very broadminded. He wanted to understand things. We said we’ve got to collect data on what these drugs do to people, and Lew agreed we should do it so I was the only person in the hospital who could write orders for medication. I would write orders for anybody but the residents had to first call me up, so I was able to ask why they were putting the patient on medication. I also interviewed the ward staff, resident and supervisor as well as the patient. Then I would see the patients every week until they were discharged. Anytime the residents wanted to change the dose or the drug, they had to call me up and tell me why.  That was the best learning experience I ever had, because I saw all sorts of things done that I never would have done, and  some worked and vice versa.  I did that for almost two years. During that period, I evolved this notion of pharmacological dissection. The idea came from an observation that there were distinct patterns of response to Tofranil (imipramine) and Thorazine (chlorpromazine). One of the patterns I came upon early was that there were patients who had what we now call panic disorder with agoraphobia. In those days, they called them schizophrenic although not delusional or hallucinated. But when patients with this pattern went on Thorazine, which we thought was an anti-anxiety drug, they got much worse.  That was very disappointing because at the time Menninger believed that anxiety caused everything. Thorazine was good for schizophrenia which goes with very bad anxiety, so it should have been good for lesser anxieties. But it wasn’t. When we got them, as a last resort, on Tofranil, it stopped their spontaneous panic attacks. I published that in the early 1960s, but nobody believed it.  They thought it was just some sort of crazy idea. First of all, Tofranil was an antidepressant and these people weren’t depressed. And, besides, Thorazine was an anti-anxiety drug, so why did it not work?  It struck everyone as very strange you would have an antidepressant drug knock out panic, the worst form of anxiety, and make it possible for patients who were afraid to leave a room, go out by themselves.  I did a long series of double-blind placebo-controlled studies which showed I was right. Imipramine stopped the panic attack and I developed a theory that agoraphobia was secondary to panic inciting anticipatory anxiety, followed by avoiding situations where you might get a panic attack and couldn’t get help or get out.

LH: Did you try any MAO inhibitors at that time?

DK: Yes, I did.  As a matter of fact, in our second paper we reported on 4 patients, who responded positively with MAO inhibitors.  But I was able to point out that what we had wasn’t a general antidepressant effect. These patients responded badly to ECT. So it wasn’t that they were just depressed in some peculiar way. For many years it was thought that all the antidepressants worked to block panic. Now we know there are antidepressants that don’t work.  I came up with this idea of pharmacological dissection, putting people together with a similar pattern of response to medication to see if there was something in their baseline state you could use diagnostically. That’s the way I’ve been thinking about refining diagnosis.

LH: It is a rather unusual way of making a diagnosis; choosing a drug that you think is good for that diagnosis. You would make the diagnosis after the fact?

DK: Exactly.  The first big study we did in this area of research is still one of the biggest studies done in a single place.  We took 300 patients and randomly assigned them to placebo, Thorazine or imipramine and tried to figure out what they responded to. That took me out of my antidiagnostic phase, because the best way I could make sense of the various drug response patterns was to recognize there were relevant diagnoses. But they weren’t the diagnoses like schizophrenia that everyone was using loosely. They had been described a long time ago as agoraphobics and even before that as secondary to panic attacks by Freud. But there were depressed people, who responded poorly to imipramine and I recognized those were the ones the English described as atypical depressions.

LH: That was Will Sargent?

DK: Sargent, Dally, West, and a whole group of English psychiatrists, who were very good observers, and recognized these peculiar depressions, which did not respond to ECT and tricyclic antidepressants but to monoamine oxidase inhibitors. I followed that up, years later, with Fred Quitkin, when I got to Columbia in 1976. It became the largest series of randomized placebo-controlled trials contrasting imipramine, phenelzine and placebo. Phenelzine worked by far the best. In the new DSM-IV, atypical depression is included as a parenthetical modifier.

LH: In those days, you were trying barbiturates first in the treatment of panic on the assumption those were considered antianxiety drugs, but these patients didn’t respond to barbiturates. Is that correct?

DK: Panic wouldn’t respond, but between panic attacks on barbiturates they would feel better.  But, then, a panic attack would come along and they’d start taking more barbiturates. They thought a panic attack was the outgrowth of their chronic anticipatory anxiety.  If anything, it was the other way around; it was the panic attack that was promoting the chronic anxiety.  So, a fair number of people who had a panic disorder ended up getting hooked on barbiturates and alcohol, which helped anticipatory anxiety but not panic attacks.

LH: In the 1980’s alprazolam came along and that seemed also to work in panic.

DK: There’s no question that alprazolam works in panic.  What I said was panic had to differ from anxiety because imipramine treated the panic and the person was left with chronic anxiety and phobic avoidance.  I didn’t say, although a lot of people thought I did, that generalized anxiety disorder would not respond to imipramine.  I simply said that chronic anxiety would take a long time to go away. We then showed that for people who only had specific phobia and anticipatory anxiety, but not agoraphobia, imipramine was no better than placebo. Now, the question was whether imipramine works in panic disorder and alprazolam in generalized anxiety disorder?  The answer seems to be that imipramine and alprazolam work in both. So that does confuse the issues, in terms of trying to get a neat dissection. There were some interesting findings regarding this in the Upjohn study when they compared imipramine and alprazolam. Two British psychiatrists did a cluster analysis of the patients’ description of their panic attacks and found that for those patients who had a lot of dyspnea, shortness of breath, imipramine worked better than alprazolam, but for those who didn’t have a lot of shortness of breath alprazolam worked better than imipramine.  So, it struck me that maybe there are different sorts of panic attacks. That is another type of pharmacological dissection. In generalized anxiety disorder, it takes 4 to 6 weeks for imipramine to work, and it works in doses of 80 or 90 mgs a day which aren’t good enough for panic disorder; patients with panic disorders need more than that. It’s still not clear to me whether the very high potency benzodiazepines, such as alprazolam, clonazepam and bromazepam, which are effective in panic disorders, are doing something different than the lower potency benzodiazepines. There’s only one study on diazepam in panic disorder, but in that study they ran the dose up to about 45 milligrams a day. The patients got somewhat better, but the panic measures were quite unclear. So, I think it’s still moot.

LH: 40 milligrams of diazepam would get you in the ballpark, on the basis of the comparative potency of diazepam and alprazolam.

DK: That’s true.

LH: So pharmacological dissection in psychopathology led to a new formulation for panic attack?

DK: Yes. And, then, when I got to Columbia, we started studying the psychophysiology of panic attacks. Back in the 1960s Pete Pitts discovered giving intravenous lactate to patients he called anxiety neurosis created a panic attack. But he got into a fight because it was argued that the tremor and feeling of paresthesias the patients got from lactate frightened them into a panic attack. It was argued non-specific stress produced the panic. What Pitt then did was to give the patients EDTA, a powerful calcium-chelating agent, which threw some into tetany, but the patients didn’t panic. But that got ignored and the general consensus was that lactate was doing nothing specific.  Then, an English psychiatrist, Desmond Kelly reported on 8 agoraphobics who panicked after lactate and, when he gave a MAO inhibitor to these patients, 5 out of the 8 got better. Then he gave lactate again and the 5 patients, who had gotten better on the MAO inhibitor, didn’t panic anymore. I said that’s more than conditioning so, when I got to Columbia, I set up an experiment with lactate and imipramine, showing that imipramine blocked the panic and even after the patients were taken off imipramine for a month they did not get panic if you gave them lactate again.

LH:  It’s kind of desensitization.

DK: It pushed the switches around.  I wasn’t sure how, because we brought back a number of them six months later. They were panic free for six months and they had not expected to get again a panic attack at all.  But what we found was that about 40 percent of them panicked.  Whatever imipramine does, I think it down regulates the suffocation alarm, goes away eventually.

LH: How did you get to this “suffocation alarm” hypothesis?

DK:  First of all, everyone assumed that panic is a sort of fear, which makes sense. But it doesn’t look like fear, because the outstanding feature of the panic attack is dyspnea that, depending on the seriousness of the attack, occurs in 70 to 90 percent of patients. The person says I can’t get a deep breath; I’ll run to the window; I’ll throw it open; I just can’t get a deep breath.  And, that’s not part of fear. There have been seven good studies of people who have been shot at in combat or jumping out of airplanes and they all report palpitations, trembling and sweating, but they don’t report dyspnea. The other thing that tipped us off that panic wasn’t fear was that when we took the blood levels of epinephrine, norepinephrine, cortisol, and ACTH, of panicking patients it was flat. There was no surge of these substances in panic attacks as you have with fear. So, we reported that lactate was suppressing the hypothalamic pituitary adrenal system, but got the same effects with inhaled carbon dioxide, which doesn’t give you an osmotic load. Scott Woods, at Yale, took patients, wired them up, put cannulas in, walked them into situations where they were likely get a panic attack, like a supermarket, and again found no cortisol surge during clinical panic.  Then I started to think isn’t it peculiar that the two powerful panicogens which don’t produce any increase in cortisol are lactate and carbon dioxide, substances which are intimately tied in with what happens to you if your respiration is compromised.  The surest sign that you’re not breathing enough is that your blood carbon dioxide is going up. And lactate is a remarkable substance that only comes from one place and it only goes to one place. When glucose is being burned, it goes through pyruvate on its way out as carbon dioxide but if you don’t have enough oxygen it gets shunted into lactate. So, you have two sure signs there that there’s something wrong with your respiration; carbon dioxide or lactate going up. And those two things induce panic. So, that’s what got us to develop the idea of a suffocation alarm system. We’ve been pursuing that idea and written about it extensively.  We’ve got a lot of good circumstantial evidence that situations where carbon dioxide is likely to increase are those where the amount of panic increases. When carbon dioxide is low, or kept low, panic is unlikely to happen. Childbirth is a situation which, according to all the psychological theories, should be very panicogenic. There are many internal sensations signaling danger. You are actually in danger. There is uncertainty, because you don’t know what’s going to happen next. In fact, patients with panic disorder never panic during childbirth, perhaps because people have the lowest blood carbon dioxide levels during childbirth.

LH: Because they are hyperventilating?

DK: Hyperventilating at a furious rate. I want to tell you one more story.  If this works out, I will be very pleased. What an experimenter ought to do when he or she develops a theory is to look for a place where the theory doesn’t stand up, because that’s a way to enrich the theory. So, my theory implies that anytime somebody is suffocating, asphyxiating, they ought to panic.  And, in general, that’s true, but there’s one big exception, which nobody had ever pointed out and this is carbon monoxide intoxication. When people asphyxiate with carbon monoxide, they just fade away.  If brought back before they die, they don’t tell you that they had a panic reaction.  People have been found in their cars in the garage, but nobody runs out of the garage, panicking.  So, that seems to be a hole in the theory.

LH: Carbon monoxide intoxication would be primarily oxygen deprivation.

DK: The carotid body is measuring both oxygen and carbon dioxide levels and if the oxygen level goes down, or the carbon dioxide level goes up, that stimulates the brain respiratory centers by the 9th and 10th nerve. So, the carotid body is a suffocation monitor.  Sol Snyder found that carbon monoxide was a neurotransmitter, so maybe it’s screwing up the alarm system.  Then I got a letter from Sol, saying they have shown that carbon monoxide is an inhibitory neurotransmitter in the carotid body. It gave me an idea for a terrific study which I was trying, unsuccessfully, to get through the IRB. The idea was to produce panic attacks with 7 percent carbon dioxide in panic patients and by mixing small amounts of carbon monoxide with carbon dioxide, the alarm system should be sabotaged and stop them from panicking. I think this would really be conclusive evidence for the theory. So, I thought we’ll find out, but never did.

LH: It would be an important study. It’s very interesting you mentioned that shortness of breath signifies panic. I remember talking with Mandel Cohen a few years back and we recognized that some of the patients, who had nocturnal panic attacks, would wake up in the middle of the night, gasping for breath and fool you.  Of course, they didn’t have a large heart or wet lungs so that is an aspect of fear and anxiety.

DK: Mandel Cohen was 40 years ahead of everybody.  He showed that carbon dioxide was a panicogen, but nobody picked up on it.  He went to World War II veterans and showed that they did not have dyspnea when they had fear on the battlefield.  He said back in the 1950s, that whatever the peculiar thing was that happened in “neurocirculatory asthenia” it was not fear.  He was very clear about that.  We invited him to give a lecture at Columbia where he did not endear himself by comparing the influx of European analysts to a swarm of locusts! He was a real pioneer.

LH: He certainly was.  From now on I suppose you are going to develop and test this hypothesis in every way you can think of.

DK: Right. We have another hypothesis we’re working on. We pointed out a long time ago that half the patients with panic disorder have a history of separation anxiety as kids; they remember they didn’t want to go to camp; they fought going to school; they stayed out of school; they wouldn’t go to sleepovers. With my wife, Rachel Klein, we have done a 20 year follow up on school phobic kids we treated, and the only thing that developed, in excess, was panic disorder in later life. So how do you bring together the two hypotheses, suffocation alarm and separation anxiety? Well, it has been shown in animals that endorphins decrease both separation anxiety and carbon dioxide sensitivity. So, it’s conceivable there is some link there. If you had a situation like periodic “endorphin deficit,” that in my opinion, would increase both separation anxiety and suffocation sensitivity.

LH: Has anybody ever looked at submarine crews?  You have a situation where people are threatened with suffocation or drowning.

DK: I went back to the work by Haldane, done in 1918, on submarine crews. What came out of that study is that carbon dioxide levels in subs are 2% and you can become a submariner only if you can adjust to that. We’re trying to experiment on the endorphin line and find out if you give lactate to normal subjects what happens. We already know that if you give naloxone to normals, very little happens.  Pickar gave a whacking dose of naloxone to normals and they got nervous and anxious, but that was about it, nothing terrible happened.  But I wonder what would happen if you gave lactate to subjects pretreated with naloxone, whether normal subjects given an endorphinergic deficit are lactate sensitive.  We’ll see how that works out.

LH: I don’t recall anybody placed on naltrexone develop panic.

DK: No, there is no report on any patient who developed panic after being put on naltrexone. There are reports on some peculiar episodic dysphorias with naltrexone, but they are not well documented.

LH: If you muck up a system as important as the endorphinergic system, it is likely something will happen.  We reported a number of years ago dysphoria produced by naltrexone that would probably explain why it’s so hard to get it accepted into clinical practice. People don’t feel particularly good on it. Well, you’ve made a career of pharmacological dissection and developing a systematic description of agoraphobia and testing its’ pathophysiology.  It will keep you busy for a while.

DK: I hope so.

LH:  Since 1976 your career has been at New York Psychiatric Institute.

DK: Right.  I’m Director of Research at the Psychiatric Institute, which is a nominal title.  I don’t have any real power, it’s just one of those titles.  I do have a very big department in Therapeutics that has an anxiety clinic, a depression service, a family study group, and a biological studies group.  They do a lot of work.

LH: That has been very productive in terms of publications.  Well, that’s an interesting career you’ve had and more to come, I think.

DK: I certainly hope so. It was also fun being involved with ACNP; it’s a very elite organization.  People, who are in it, are very smart successful people. One of the problems with being successful is that it makes you somewhat conservative, so you don’t want to rock the boat too much, because, after all, you’ve done all right. But there have been a number of developments recently that should shake us up in terms of how psychopharmacology research is going in coming years, both from the point of federal support and from the pharmaceutical industry. The ACNP could play some proactive roles there.  I hope it will do.

LH:  You’ve been a creative thinker in this line; what do you think the ACNP should do?

DK: The ACNP ought to try to formulize a relationship with the heads of the federal agencies, including the FDA, NIH and so forth, and meet with them regarding their agendas.  For instance, I’m head of a mental health clinical research center but I’m not at all certain whether mental health clinical research centers are viewed as a sensible way to spend money. I think psychiatry is in a relatively primitive state compared to internal medicine. They’re way ahead of us in objective measurements and physiological understanding. Are RO1s by independent investigators a really good way of funding research?

LH: That’s Rosalyn Yalow’s idea. You provide support to individuals rather than huge amounts of money to centers.

DK: For Rosalyn that makes great sense, but for psychiatry, we still need to get critical masses that can collaborate as experts in a variety of fields, because we’re nowhere near Rosalyn Yalow. For that reason, centers make sense in psychiatry. It would be interesting to have a discussion about that with someone like Dr. Harold Varmus.

LH: What do you think of the future of psychiatry, with everybody nipping at our heels and trying to get a piece of the turf?  You know psychologists will soon prescribe drugs.

DK: I’d be surprised frankly if that happens in the near future, because they’re not qualified; they don’t have any knowledge of medicine and drug interactions. The training they get doesn’t qualify them for it and malpractice lawyers would have a field day. So, I don’t think that’s going to happen. But the psychotherapy area will change. Psychiatrists who solely practice psychotherapy, will be people dedicated to surviving on low incomes, because the psychologists will undercut the psychiatrists, and the social workers will undercut the psychologists, and the psychiatric nurses will undercut the social workers. You won’t be able to make a living doing solely psychotherapy as an M.D.  Psychotherapy has a place in treatment but it’s a function that can be delegated, so it ought to be supervised, following proper diagnosis.  In my own practice we have psychotherapists, but they don’t work up the patient. The psychiatrist does that and dispenses the medication. Psychiatrists collaborate with them and do whatever else is needed; internal medicine with a psychosocial backup.   It makes sense to me.

LH: Undoubtedly, there’s going to be some big changes in the not too distant future.

DK: Right.

LH: I’m sure you’ll be part of the thinking about it.  You have always applied your agile mind to many issues, and I look forward to seeing what you do in the future.

DK:  Thank you, Leo.  I appreciate it.

Second interview of Donald F. Klein

Interviewed by John M. Davis, Boca Raton, Florida, December 12, 2007

JD: This is an interview with Donald Klein for the ACNP history project. I’m John Davis and I’d like to start out asking when you were born, a word or two about basic demographics and then get on to medical training.

DK: I was born on September 4, 1928 in New York City where I lived most of my life.  The big turning point was going to the Bronx High School of Science, a city run specialty school, where you had very intensive training and a remarkably smart staff.  My father, whose education only went to the high school level, was a very intelligent man; we used to go to museums forever. 

I wanted to be a scientist all my life. I wasn’t sure what kind of scientist, probably chemistry initially.  I went to college at fifteen and stumbled on Freud who was talking about things I was interested in at the time, sex and aggression, which made me want to be a psychoanalyst.  I found that to be a psychoanalyst you had to go to medical school so as to become, what turned out to be non-existent, a research psychoanalyst. I graduated Colby College in 1947, magna cum laude but.  I couldn’t get into medical school, although I had been number one in my class.  It was a combination of things.  It was the end of World War II and veterans were flooding back and they got priority, which was understandable. There was also a fair amount of anti-Semitism at the time. I tell everyone at Columbia that they turned me down twice.  I spent a year in graduate school at NYU, in biochemistry and endocrinology, which turned out to be extremely useful. It gave me a running start on what was necessary to be a systematic scientist. After that I was accepted to medical school, worked as a research laboratory  technician for my friend  Norman Kretchmer who went on to be Director of NICHD), and did a rotating internship in the US Public Health Service. I intended to stay in public health during the Korean War, but they terminated me after my internship. In July 1953 I ended up as a first-year psychiatry resident in Creedmoor State Hospital, which was a six thousand bed jail, with non-existent psychotropic medications.  I had no experience with psychiatric patients, but they put me in effective charge of the admitting unit and the male acute ward. I stayed there for a year and saw a lot of amazing psychopathology, things that people just don’t see anymore. Then I went back into the Public Health Service and, by good fortune, landed at the 1000 bed USPHS Hospital for Opiate Addiction in Lexington, Kentucky. The inmates were 70% black, 80% federal prisoners and 20% volunteers for treatment .I ran the 70-bed admission/methadone withdrawal unit. I had no experience in addiction but was put in charge of the Admission and Withdrawal Unit,

My training was going on rounds once with the departing director.

This jail was run by three supervising psychoanalysts who tried to turn it into a model, based on Chestnut Lodge. The 50 white female patients who were in psychotherapy had both a psychotherapist and an administrative psychiatrist, who dealt with realistic issues, things like parole. I had the good fortune to meet Abe Wikler and Harris Isbell, who were running the most advanced human experimentation program in the world, funded by the Public Health Service. 

The FDA had approved meperidine (Demerol) as a non- addicting narcotic because the strain of dogs they tried it on didn’t get addicted. But, humans did and that caused a tremendous scandal, a prefiguring of our current post marketing problems. They realized the only way to test possibly narcotic drugs was on human beings and species wise that’s probably correct. They figured that prisoner addicts, who had served ten years in prison and who volunteered, would be suitable subjects. It was a remarkable experience.  I had no formal research role but was able to participate in the early studies of LSD, reserpine and chlorpromazine.

I was put in charge of the psychotic unit which was for WWI veterans, who had become psychotic before there was a VA. They were on something called the Executive Order. They had been hospitalized forever. Although they had received excellent rehabilitation care with excellent nurses most were mute, incomprehensible or grossly paranoid. I decided to give them all chlorpromazine, 200 mg a day, which was a big dose then. One of them, after about six weeks, came to me and said, “Hey Doc, when am I getting out of here?” It was the first time he had said anything in thirty years; it was very convincing that chlorpromazine was not just a chemical straight jacket.

 I went back to Creedmoor, where I finished my residency and became a member of the Creedmoor Institute for Psychobiological Studies which was initially run by Arthur Sackler but had been thrown out by the Creedmoor Director on entirely political grounds.

I got involved in a variety of things.  For a year I was the Clinical Director of a psychoanalytically oriented clinic for six families with autistic kids. I was in charge of play therapy for a pair of identical autistic twins as well as the father’s group therapy; I asked my supervising analyst how the mother got the twins to walk on their toes, and he told me that was resistance.   Later, the remarkable Loretta Bender came to Creedmoor to run child studies. Eventually, the State closed our clinic down; our patients all got hospitalized despite their intensive care, and we went into geriatrics and early studies on anti-psychotics.  We picked, as our first antipsychotic, mepazine (Pacatal), because it had anecdotal reviews of being a terrific agent, .We did a double blind randomized “add on” study.  Patients who were already on anti-psychotic medication were randomized to either placebo or mepazine supplements. We found nothing.  The drug, as far as we could see, just did not work. We published the results and a couple of months later there was a large VA study published that had used phenobarbital as the placebo, mepazine and a couple of the other anti-psychotics. They found the other anti-psychotics were far better than phenobarbital, but mepazine was only marginally different from phenobarbital and much worse than other antipsychotics. The drug was withdrawn from the market; I think it’s the only phenothiazine ever withdrawn. This was way before the Kefauver Harrison act and the efficacy requirement for marketed drugs. We also did a large study on dicumarol vs. placebo vs. no treatment in hospitalized, demented patients aged over sixty that was published in the Journal of Chronic Diseases; it was one of our first papers. We found they lived longer on dicumarol, but their mental status didn’t get better.   We did a number of other studies of modest interest.

The big shift in my life came in 1959.  There was an opening at Hillside Hospital, a two hundred bed psychoanalytic hospital of the Federation of Jewish Philanthropies with a Research Department of Experimental Psychiatry. The hospital director was Lew Robbins, a training analyst from Topeka, an extremely nice man with an open mind, who said very early on that we didn’t understand these drugs but should try to figure them out.  I went to work for Max Fink, who ran the quite unique Department of Experimental Psychiatry. Hillside   was a non-academic hospital, affiliated with no medical school. But it was an unusual place because it had a research tradition. Max had done excellent work on ECT and was a terrific mentor, guide and relentless critic.  He sent back my first attempt at a paper about seventeen times, telling me to fix this, fix that and he was right. We got involved with extensive pilot studies of chlorpromazine and imipramine for two years doing what would now be called early phase two studies, We were able to study patients for optimum dosage, running the dose up and down over time, finding out how long the drugs took to work, how to deal with side effects, who the drugs worked or didn’t to work on. That was a Utopian opportunity; one of the biggest missing pieces in current psychopharmacology is adequate phase two studies.  One of the reason that we don’t have any longer these type of studies is that industry can’t use them as definitive for the FDA; they don’ count. Another reason is that drugs are rushed through so as to have maximum patent protected marketing monopoly.  Eventually we wrote two papers, one on about a 100 patients treated with chlorpromazine and another 100 treated with imipramine. We described the various patterns of response to the medications. I developed the notion of pharmacological dissection since we had no clear idea about how people were grouped diagnostically. When a group of patients had a similar positive or negative response to medication they could be categorized as pathophysiologically similar.

JD: How did you get that unusual idea?

DK: It was based on my experience with imipramine when it was not yet marketed. The relationship between drug houses and investigators was entirely different than it is now.  As I recall, Max Fink was able to deal with Geigy, so they gave us a lot of imipramine and said do something useful with it and let us know what you find out.  We had arranged with Lew Robbins that for this two hundred-bed hospital, with an average length of stay of ten months, that I or Max were the only people who could write orders for medication. The residents had to call us first, and say Mrs. Jones is schizophrenic and we want to put her on chlorpromazine, 200 milligrams a day. That gave me the invaluable opportunity to ask the resident, why are you doing it now?   After all, the patient has been here nine months and has been schizophrenic all along. The resident would say it was to lower anxiety which was interfering with psychotherapy. Then I would talk with the supervising analyst who told me the resident was doing a bad job in psychotherapy so had to fall back on medication. Questioning the ward staff revealed a rising level of discontent; anything would be better than aggressive outbursts. The patients would say they had given up on going home and were willing to try anything. I would prescribe the chlorpromazine and follow the patient up weekly, doing whatever the treating staff wanted me to.  It was the world’s best learning experience. Patients getting the same drug were having very different reactions.  Finally, at one point, we knew that imipramine was a good antidepressant, but it also had some funny antianxiety effects. We had a patient who was diagnosed schizophrenic and had responded badly to chlorpromazine, so we put him on imipramine. The patient said that meant the hospital had lost all hope or they wouldn’t put him on an experimental drug. I slowly raised the dose and after two or three weeks the patient was complaining incessantly that the drug was doing nothing for him.  The therapist didn’t think much was happening; the supervisor was certain that nothing was happening but then the ward staff tipped me off.  This patient had been there for ten months and   kept running to the nurses’ station saying he was dying.  The nurses would hold his hand and reassure him he was not, that it was just terrible anxiety and there was nothing wrong with his heart.  After ten minutes or so, the patient would wander away but would be back a few hours later again complaining he was dying.  Now the nurses said he hadn’t done this for the past week.  So, I went to the patient and said, “I understand you are feeling better”. He asked, “Who told you that”?   Well, I said, “the nurses did”, and he replied, "What do they know”?  So, I said, “Isn’t it true that you have been running to the nurses’ station for months, but you haven’t done that this week”? It really stunned the man, because he hadn’t thought about that at all. When I asked why he had stopped he said he finally learned they couldn’t help him.  I asked him how he had managed to learn that this week, He swiftly replied, “Well, you have to learn some time.” We were able to figure out later he was suffering from what I called a “three-layer cake.”  He got spontaneous panics, ran for help; developed tremendous chronic, anticipatory anxiety, and “phobic” avoidance, where he wouldn’t go anywhere unless help was easily available or without somebody with him, even in hospital. Now, this is recognized as agoraphobia. He was tying his family into knots. On imipramine the panics stopped, so running to the nursing station also stopped. However, the tremendous anticipatory anxiety had not stopped so the phobic avoidance continued. That took time and exposure before it stopped. At that point, I realized that anxiety wasn’t a single thing, that here we had three different kinds of anxiety dissected out by the medication. That’s where I got the idea of pharmacological dissection; where patients’ symptoms and behaviors, lumped under the same label, became distinguishable and patients who looked similar but had different drug responses must have different pathophysiologies.

We later did a double-blind placebo-controlled study, randomized regardless of diagnosis - diagnosis was terrible at the time - to placebo, imipramine, or chlorpromazine mixed with procyclidine, an anti-Parkinson drug.  The drugs were given in liquid form to prevent checking and patient knowledge of dose changes. We used a fixed-flexible dose aiming at 300 mg per day of imipramine or 1200 mg per day of chlorpromazine. These doses came from several years of pilot observations. We studied 150 patients and, then, did another 150 patients.  It’s one of the largest single site studies. We were able to systematically validate that these drugs work in patient dependent ways. We found, by the way, that chlorpromazine was an excellent antidepressant. A handful of other studies by Leo Hollister and people in Europe also showed that antipsychotics for severe depressives really worked. However, that fit none of the current theories at the time, so it fell off the therapeutic and cognitive table and was killed by the recognition of tardive dyskinesia.  The new brand of antipsychotics, that don’t have much EPS, came along and some have found them useful in affective disorder. Somewhere along the line you, John Davis, came into my life. In the 1960s we became friendly, talking with each other at meetings and you told me it would be a good idea if we wrote a textbook. 

JD: I figured nobody would read a textbook written by a resident, so I needed a good thinker and an experienced clinician, and it went back to the papers you were just talking about.

DK: That textbook John and I wrote was the first systematic textbook on Diagnosis and Drug Treatment of Psychiatric Disorders. We had a discussion, about diagnosis broken up into psychoses, affective disorders and neuroses. For each category we had a complete systematic review of the literature which John did. It’s still one of the best reviews in the literature. That sort of extremely detailed literature review has been replaced by meta-analysis which is much worse in every way.

JD: The value of meta-analysis is that clinicians realize when there isn’t an answer. That encourages the clinician to use his intuition and may empower the patient to state their preferences, because it identifies where there’s literature and where there isn’t.  Instead of the expert making it up, in our book, I put in the controlled studies and you put in the clinical wisdom, so we had both.

DK: It was a good book and I’m very pleased and proud we did that together. It came out about 1969.

JD: Yes.

DK: One of the funny stories I remember about the book was that after we worked four years on it I got a call from Williams and Wilkins who told me the book was bound and ready to go. The title of book was Diagnosis and Drug Treatment of Psychiatric Disorders, but they had left the word “drug” out from the cover completely. When they called from the publishing house and asked, whether leaving out the word would matter, I blew up, of course, and they had to rebind the books and put a strip on the back, which regularly peeled off. The book had a heavy emphasis on descriptive diagnosis and was a forerunner of the reawakening of interest in descriptive diagnosis.

JD: Let me ask you about discovering panic attack by using pharmacological dissection. You identified a new disease and also its treatment.

DK: The point of pharmacological dissection is that when you note an unusually effective treatment in some sub-group of people you ought to identify and describe that sub-group.  We also did that, later on, with atypical depression, which was an even a better story in some ways.  I also worked closely with my wife, Rachel Gittelman Klein, and did early studies on ADHD children.  She’s still doing controlled follow ups on that same group that we studied in the late 1960s. Those children are in their forties now; she’s going through her third wave of follow up and now is getting brain scans on them with Xavier Castellanos.  We have been able to follow-up particular interests for long periods of time; panic attack we have followed in terms of treatment, and lactate and carbon dioxide challenges. More recently we have considered a possible endogenous opioid deficiency. I just finished a study with Maurice Preter, trying to produce something like a panic attack in normal people. Lactate really doesn’t affect normals much but causes panic in panic disorder.  Based on our theory, we hypothesized that interfering with the opioidergic system before lactate infusion in normal subjects would produce panic like symptoms. So, we did a controlled study, randomizing subjects to naloxone prior to lactate, naloxone prior to saline and saline prior to lactate.  We have a paper in Biometrics, of all places, showing that it’s only when lactate was preceded by naloxone that we got very marked tidal volume increments  like those in spontaneous panic. So, it’s suggestive but not definitive evidence. The next step would be showing whether this reaction is specifically blocked by antipanic drugs.

The other thing we have done that is interesting is in the atypical depression area, in studies largely carried out by the late Fred Quitkin. The prototype atypically depressed patient has a temporarily responsive mood, tends to overeat and oversleep, is very rejection sensitive, and has enormous fatigue or may have just one of these features. Comparing a MAO inhibitor, phenelzine, versus a tricyclic antidepressant, imipramine, versus placebo, we found in about six different studies that MAO inhibitors really work.  The tricyclics are barely better than placebo and the MAO inhibitors are much better than the tricyclics. Jon Stewart went back over all the histories, just like we’ve gone back over the panic disorder histories, and found that in patients who had early onset, or chronic depression, tricyclics didn’t work at all; only the MAO inhibitors did. In people who had later onset or more intermittent depressions, even though they look like atypical depression, the tricyclics really worked.  That’s the line of pharmacological dissection we have been pursuing all along.

JD: It is unfortunate there’s not more of that work.

DK: Who is going to support it? NIMH stopped funding any placebo-controlled studies on marketed drugs.  The industry is also not interested on a couple of counts. First of all, they prefer, a broad syndrome, because that’s what the FDA approves and you have a much bigger market, so it’s counterproductive from their profit point of view to refine syndromes.  So, the two major funding sources, industry and NIMH, aren’t going to do this sort of thing.  I have been very fortunate working at the Psychiatric Institute because we have hard line New York State support. Fred Quitkin, Jon Stewart and Pat McGrath have worked with me for twenty-five years or more, so we are not at the mercy of the project grant system, which is the pride and joy of NIMH but has this terrible problem.  If you don’t get grants, you have to disperse your team.  So being able to do long term intensive work becomes impossible.

JD: I’d like to explore a couple of issues since this is a history interview. It’s important to paint the picture of what psychiatry was like when you started to do academic work. I don’t think people appreciate the atmosphere at the university departments of psychiatry dominated by psychoanalysts or the clinical problems in a state hospital with 6,000 patients and only 10 doctors.

DK: At the time I went to medical school from 1948 to 1952 I believe every chairman of psychiatry in the United States was a psychoanalyst. In a way, it was understandable; everybody likes to have a theory, and nobody was doing much in the way of outcome studies. Psychoanalysis has a charming theory, a lot of literary interests and did something that we can’t claim; it could cure by getting to the root of the matter and resolving the unconscious conflict, compared to the mere symptomatic effects of other psychotherapies and medication, which fostered infantile oral regression. So, it was terrific in terms of promise.  The state hospitals were getting bigger and bigger. A deal had been made between the states and the cities. The city was responsible for acute care up to thirty days, and after that it was the state hospitals’ problem.   The state hospitals had no after care system and no good treatments.  The wards were bedlam; the treatments we had were unmodified ECT, and some people used insulin coma. The prediction was that the expanding hospitals were going to bankrupt the state.  The turning point in New York State was 1959; there were 140,000 inpatients at that time.   We are down to a couple of thousand now, all due to chlorpromazine.  That happened under the title of “deinstitutionalization”; but of course, it wasn’t deinstitutionalization, it was trans-institutionalization, from the state to the city budget. The patients, but not the money, were dumped on the cities, the cities didn’t pick them up and that was a disaster. Instead of an open door we had a revolving door. There was a sudden rise in homelessness. This led to federally supported “community care” which was staffed by therapy minded psychologists. They had a psychiatrist who did little but sign bundles of prescriptions. The ex-hospital patients were not considered appropriate for psychotherapy so the community center clientele were largely non-psychotic. I don’t know if this story is true, but I was told that when Nate Kline was studying reserpine at Rockland State Hospital, his objective indicator was the decrease in the number of broken windows per week showing   patients were getting quieter. Psychoanalysis was still the rage, in terms of both medical schools and psychiatric practice. It promised cure and was the only game in town with a comprehensive clinical theory. I was part of that from 1957 through1961 as a candidate at the New York Psychoanalytic Institute. I was under the delusion that one could be a research psychoanalyst and that the Institute would welcome this. I quit more than a little disappointed with what happened. I could tell innumerable stories of how counterproductive psychoanalysis has been for psychiatry. One will suffice. When I told my analyst in 1959 that I was going to Hillside Hospital to study drugs in clinical trials, he said, “that is your sadism,” which had not been noticed in the past two years. Descriptive psychiatry was held in very low esteem because it wasn’t important.  The underlying conflicts were important, and you had to be a trained analyst to perceive them.  I was one of the first NIMH Career Mental Health investigators. In 1961, I went to a meeting of about 40 career investigators, and we went around the room saying what we were doing.  There was me and one other fellow, who studied whole human beings. We were all in analysis, which was the only way to understand the depths of whole people, everything else was superficial, so budding scientists were driven towards the laboratory. When you were a career investigator in those days, they gave you five thousand a year for your analysis.  That was part of my grant, but I was so refractory that I quit my analysis before I used the money.

JD: Did you have to give the money back?

DK: I called NIMH and said I’m not in analysis anymore, what to do with this five thousand bucks? They said keep it; we’re sure you’ll do something useful with it, which is not quite the situation now.

The other thing we were involved in that has made a big difference is the DSM.  I was on the original Task Force that Bob Spitzer put together. At the time it was considered an unimportant effort, so Bob could recruit a bunch of skeptical people.  We decided early on that the issue was reliable clinical communication about syndromes. Psychologists had shown that inter-rater diagnostic reliability in psychiatry was dismal. If a patient was called schizophrenic, you had no idea what he was like. Symptom description was highly reliable, but everybody had their own definitions of diagnosis. This became known as “criterion variance.” That descriptive diagnosis was unimportant was fed by the profusion of contradictory “schools” of etiological theory; Freudian, Jungian, Adlerian, Horneyan, Pavlovian, Cognitive, Behavioral, etc.  The DSM-III stand was that diagnosis should not depend on a particular theoretical presumption, as was the case with DSM-II. This was misleadingly referred to as “atheoretical.” A syndrome was understood as a polythetic category, with inclusion and exclusion criteria and a listing of the symptoms, with a minimum criterion set number. That number was due to clinical consensus alert to both false positives and false negatives while lacking useful systematic data. The text made it clear that these were not carved in stone and clinician judgment was overriding.  However, nobody read the text and the residents had to memorize the criteria.  DSM-III was basically expert clinical consensus, because we had very little in the way of data.  At various times psychoanalytic groups would complain that they were locked out and Bob put a couple of smart analysts on the Task Force, but they didn’t have a lot to say and soon left. The real trouble started when we got to neuroses. Many people think that DSM-III was some kind of anti-psychoanalytic cabal but that was not the case. We were trying to figure out how to deal with what was called neurosis. We had good exclusion criteria that neurotics were not hallucinating or delusional, but we didn’t have overall inclusion criteria, apart from theoretical dicta about unconscious conflicts. So, we ditched the term “neurosis” and turned to their common descriptive feature as Anxiety Disorders, suggested by Rachel Klein who was a DSM-III consultant. That caused a tremendous wave of resentment and the whole process almost got shut down by the American Psychiatric Association, who thought that we were taking away their bread and butter by saying that neuroses didn’t exist. But we didn’t say that, we said that neurosis was not a useful super-ordinate term. Eventually DSM-III was a totally unexpected, profitable hit. As I remember, Bob Spitzer got the job, after they offered it to Henry Brill who turned it down, saying he wasn’t interested and felt somewhat retired. Spitzer got the job because it was unimportant. The whole notion of diagnosis was just a nuisance and not really central to anybody’s concerns.  So, that was a tremendous surprise, although it had its pluses and minuses. It’s been helpful in improving the reliability of clinicians’ descriptive statements.  At least, you know pretty well what somebody means when they say a patient is schizophrenic. However, I think it has deflected clinicians away from taking detailed developmental histories because they’ve got those neat symptomatic criteria. DSM-III laid the groundwork for arguments about validity and underlying pathophysiology but many, including scientists, made the unwarranted assumption that these clearly heterogeneous syndromes could be handled as if they had a homogeneous etiology. Some think it served to constrict scientists to the DSM-III diagnoses, so if you applied for a grant or for FDA approval you had to use the DSM-III categories. That may be true, but it never bothered me. I came up with atypical depression, panic disorder and childhood asocial schizophrenia because we thought the DSM categories weren’t very good.

JD: From the historical point of view my recollection of academic psychiatry in the 1950s and ‘60s was that it was all psychoanalytic, that all mental illness was thought of as a defense against anxiety with only one etiology. 

DK: That was exactly what Karl Menninger said. He said that it was all one disease due to more and more anxiety. A psychosis was due to the fact that people had so much anxiety that their ego crumbled, and they regressed to the oral stage. So, when chlorpromazine came out it was unanimously mislabeled as a wonderful anti-anxiety agent.

JD: But, not accepted by the analysts. The analysts at that time thought you were not a good therapist if you had to use a medication.

DK: They said it was just a chemical straight jacket.  Even among doctors who saw chlorpromazine’s benefit, they thought of it as an anti-anxiety agent, because that was the conventional wisdom. When we applied chlorpromazine to ordinary non-psychotic anxious people, they should have been helped but they got worse, so that didn’t fit.  This affirmed my idea of pharmacological dissection; you just couldn’t combine psychoses with anxiety states.

JD: It was all a defense against anxiety. I remember the teacher of psychopharmacology at the University of Maryland said, use chlorpromazine against schizophrenia because that was the severest form of anxiety and use meprobamate against depression, because it wasn’t so severe.

DK: I do not think that was atypical.

JD: Thinking about different drugs for different syndromes was a complete shift in the way of thinking.

DK: Right. People become confused sometimes between pharmacological dissection and pharmacological amalgamation. They think if two conditions both respond to the same drug, it must be the same condition.

JD: No, not necessarily.

DK: Usually you can find substantial differences.  It has always struck me as strange and important that every antipsychotic that works in schizophrenia, also works for mania, and often severe depression. It doesn’t work the other way around.  There’s something similar and as well as different, concerning the pathophysiology of the major "functional” psychoses. I discussed this in our textbook, but it has been widely ignored.

JD: At the time you discovered panic attack disease it was not known as an entity.

DK: Not at all. It was buried in the anxiety neuroses.  In 1895 Freud wrote a paper on distinguishing anxiety neurosis from neurasthenia. It’s one of the best descriptive papers around.  He describes panic attacks exquisitely well, but he doesn’t see it as anything different from all the other forms of anxiety people complain about. People complain about fatigue, belly ache, diarrhea, dizziness, palpitations, etc. Panic is just one of the entities embedded in anxiety neuroses.  One of my arguments, which I got into trouble for years ago, with the conventional wisdom is, that panic attack is not fear.

J.D.: How do we know that? 

DK: Symptomatically, the vast majority of panic disorders that lead to agoraphobia are associated with acute air hunger or dyspnea, which is not a feature of acute fear.  That’s been known since World War II by studying people who have been wounded in battle, jumping out of planes or disassembling bombs. They have palpitations, sweating; and trembling, but they don’t have acute air hunger.  Also, we can experimentally generate panic attacks with IV lactate and carbon dioxide inhalation in the laboratory but only in patients with panic disorder, not in normal subjects or patients with other anxiety disorders or depression. .Strangely, there is no increase in output from the hypothalamic pituitary adrenal axis; ACTH, cortisol and catecholamine don’t go up.  What is going on? That’s not fear. These features are something different, which led to my “suffocation false alarm “theory.”

JD: It is interesting the role that Dr. Robbins had when he gave you the freedom to do the work with an open-ended mandate.

DK: Lew Robbins was a terrific guy; he was remarkable.  He had been clinical director and outpatient manager at the Menninger Clinic for seventeen years but when he approached Karl Menninger and asked to become a partner in the clinic, he was told that it was a family business. So, Lew came to Hillside intending to make it the Menninger’s of the East, but that was a flop.  One reason was because we had two bedded rooms that couldn’t attract the carriage trade.

JD: Can you make an analogy between your role in that hospital and the role of a specialty clinic in general medicine or surgery?  In the modern medical center, the hepatology service will see every case of liver disease and the same with any specialty service, so they get a lot of intensive exposure to their area of interest.

DK: That’s right.  We also opened up research clinics where we got into outpatient studies.  We opened up the first phobia clinic, a depression clinic and a schizophrenic aftercare clinic to do outpatient studies.  These were clinics run by scientists as a way to learn things.  Of

JD: In the old days everybody was amalgamated, as in psychoanalysis.

DK: Right.

JD: Which meant that everybody saw everything, and all the problems were counter transference.

DK: Right, but the patients also had realistic complaints about the way clinical services were run then. Clinical services now have their own headaches. The nice thing about Hillside was that you saw the patient for a long time, the average length of stay was ten months. The city paid for all indigent care for indefinite stays. That was unusual because the maximum length of stay in any of the city supported hospitals was ninety days.  Hillside had a special deal going.  When the city finally said it’s got to be cut down to ninety days, this created outrage. The city, using salami tactics, said half your beds can be indefinite, but half must be maximum ninety days.  I saw it as an opportunity for a study and Lew Robbins agreed. We took control of the admissions and randomized patients to the different length of stays. As it turned out one group almost uniformly stayed eighty-nine days while the other averaged six months.  They were evaluated coming, going out and six months after hospitalization; but length of stay didn’t make any difference.

JD: I want to go on to the follow up of panic attack disease.  You alluded to it, but it might be worth a couple more sentences. You had a discreet syndrome with a response to a specific drug so you looked at mechanisms and discovered it could be produced by lactate and CO2 and maybe a couple of other things.  How did you begin to work out the biochemical mechanism?

DK: Let’s take a big step back.  Mandel Cohen had shown that in neurocirculatory asthenia blood lactate levels were high. Perhaps lactate was pathogenic. Pete Pitts found, I believe at the  suggestion of Eli Robbins, that in what was called anxiety neurosis, intravenous lactate produced  anxiety attacks, in a good controlled study  That caused a wave of criticism suggesting  that the lactate caused bad feelings; the patients were scared by this until they panicked. Pitts, cleverly, gave the patients EDTA, which is a calcium chelating agent producing tetanic spasms. Despite the patient’s discomfort they didn’t have panic attacks. However, this stopped nobody from criticism. To this day, when talking about panic attacks, cognitive behavior therapists assert it’s on the basis of a person being scared of their internal sensations. It makes no sense. Anyway, we started to study lactate when we went to Columbia because we had a good physiological setup, which we didn’t have at Hillside.  We expected that the HPA Axis would kick in but we didn’t find that.  I went to our Chairman, Ed Sachar, who was a cortisol expert, and showed him our data. His response was, “What are you doing wrong?” which was understandable, because people still don’t believe it, although it is very well established now.  The next question was, could we block the lactate effect with anti-panic drugs, and it turned out we can. Also, IV lactate doesn’t regularly produce panic in other anxiety or depressive conditions. Drugs often thought to be anti-anxiety like propranolol (Inderal), or intravenous valium (diazepam), don’t block lactate induced panic, so we have something quite specific.  One of the theories at that time was that panics were caused by hyperventilation and we did note that during the lactate attack people were hyperventilating, at least they were breathing more deeply. The argument was that when you hyperventilated you blew off metabolically produced carbon dioxide (CO2) ending in a state of acute respiratory alkalosis, and for some reason that nobody quite understood, it caused panic.  Therefore, you were supposed to breathe into a paper bag, which brought your CO2 up. Our experimental question was how to get somebody to hyperventilate and not blow off carbon dioxide .We worked out a system where subjects were in a transparent, ventilated, tent at 5% CO2 because there is five percent in the lungs. So, they were in dynamic equilibrium, and could not blow CO2 off.  We had a computerized spirograph, so we could measure every breath going in and out. Patients and normal control subjects hyperventilated in room air or in CO2 carbon dioxide.  Our expectation was that if the hyperventilation theory was correct, the patient should panic when hyperventilating in room air, but should not panic when hyperventilating in carbon dioxide. We found exactly the opposite, which blew everybody’s mind. We were able to convince patients to go through both CO2 challenges and challenges with lactate; showing that the CO2 panic response was a sub-group of lactate panic response.  That was odd, because IV lactate produces metabolic alkalosis whereas breathing CO2 causes acute respiratory acidosis. Trying to put that all together was very difficult. And I haven’t mentioned yet the very high incidence of childhood separation anxiety in patients with panic disorder as something else to account for. Then, I went to Washington DC for two years as Senior Science Advisor to ADAMHA, invited by Fred Goodwin. It became a sabbatical, because I didn’t have much to do. Thinking about the whole thing I realized that lactate was a signal of compensation for inadequate glucose oxidation. By shunting pyruvate into lactate, the machinery kept running by anaerobic glycolysis. In the meantime, you build up lactate as an oxygen debt.   There has been a lot of argument lately about lactate’s role in CNS.  There was also evidence that in asphyxiation the first thing that happens to the brain is that lactate goes up. Also, CO2 accumulation is plainly a sign of potential asphyxiation. Further, panic patients had acute air hunger. The way I put it together was that we have many specific alarm systems, including a potential suffocation alarm system. Following Cannon’s work, people thought of fear and autonomic sympathetic arousal as the sole alarm system for danger. That was usually thought of in terms of predation type danger. However, that we may have different alarm systems keyed to different evolutionary dangers was an idea people didn’t believe. I hypothesized that “spontaneous” panic was due to a suffocation alarm system that differed from the predation alarm system.  To diverge for a second, we don’t usually think of thirst and hunger as alarm systems, but they are. But they are slow alarms caused by slowly increasing danger.  Deprived of water you get thirsty. You get thirsty and thirstier until finally all you can think about is getting water. It is the same way with food. However, with air deprivation the danger quickly mounts so the alarm signal has to be fast because two minutes without oxygen and you are brain damaged.

JD: And predator danger is fast.

DK: Right. And, there are a few other things that have to be fast.  Falling has to be fast.  Infants have a Moro grasp reflex if dropped six inches. The other way you can get an infant to cry without hurting it is just close its nostrils, no pinching, and they shriek, which may be because their mothers lying on top of them is a recurrent evolutionary danger. Once you get into the framework of multiple alarm systems, a lot of different things in psychopathology begin to resonate. 

JD: It’s also interesting from the point of view of social organization. Then, you started some of the first specialty clinics.

DK: Right; and we continue that at the Psychiatric Institute as the only way to learn.

JD: I think a lot of specialties are doing likewise. I mean OCD clinics, Tourette’s clinics, so people see more cases and begin to put things together.          

DK: That’s correct; otherwise, it’s all bits and pieces.

JD: In trying to work out innovative things you get new ideas and find some of the mechanisms for them. Is the NIMH supportive?

DK: Not particularly.  NIMH has gone very basic and, frankly, there appear to be non-clinical review groups for grants. I don’t think anybody on these peer review groups has ever seen an un-medicated schizophrenic. They are primarily researchers at an animal or cellular level. You get a lot of shocking reviews from people who don’t understand that working with humans in clinical trials is not the same as working with rats.  I think NIMH is struggling with a narrowly focused peer review system and I don’t know how they are going to correct that, because it has a lot of cachet.   I don’t think the RO1 system as presently managed is a good idea because they are hard to get and any lapse in funding destroys hard to develop teams. I don’t think that NIMH has been supportive of the effort to subdivide syndromes experimentally to detect specific pathophysiologies.

JD: Have you ever had a grant rejected because it was too innovative?

DK: I’ve had that happen.

JD: Why did they say it was rejected?

DK: Too ambitious is usually the word they use, or not enough pilot data. It’s a very clumsy situation. Somebody was just telling us the other day that among the people that get career awards only  fifty percent go on to an RO1.  And, those are the best of the lot.  Obviously, there ought to be some changes especially with regard to career building, not just training. This administration is very constricted for money. The next administration may be different. We’ll see.

JD: How close do you think you are getting to the mechanism of panic attack disease?

DK: I think the opioid dysfunction hypothesis is interesting.  I’ve got a speculation that there may be a defect at the Δ-opiate receptor level.  In our current studies of naloxone anteceding lactate, we’re using doses of naloxone that far exceed the dose for µ- blockade. What I have read is that the µ-knockout mice don’t feel morphine but are otherwise pretty okay. Whereas the Δ-knockout mice are apparently nervous wrecks and might specifically overreact to lactate and CO2.

JD: If you find the mechanism it would be interesting. You will have discovered the disease, the treatment and the mechanism. What you can say now, which is considerable, is you’ve discovered the disease, the treatment and a number of leads to the mechanism.  You have pinned down a good deal.

DK: I think we are a lot closer to a common pathophysiology than to a diffuse etiology. 

JD: It’s going in that direction. 

DK: I hope so.

JD: I know you have been nominated for the Nobel Prize.

DK: That doesn’t mean a lot. A lot of people get nominated. The Nobel Prize is likely when the scientist is doing the sort of thing that Eric Kandel does; wonderful work, but very basic.  That’s what they tend to support.        

JD: I hope you get it. I certainly regard your work as Nobel Prize caliber.

DK: Thank you.




American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. Third Edition. DSM-III. Washington: American Psychiatric Association; 1980. 

Ban TA, editor. An Oral History of Neuropsychopharmacology. The First Fifty Years. Peer Interviews. Brentwood. American College of Neuropsychopharmacology. 2011. 

Blackwell B, editor. Update. In: Ban TA, editor. An Oral History of Neuropsychopharmacology. The First Fifty Years. Peer Interviews. Volume 9. Brentwood American College of Neuropsychopharmacology; 2011. 

Klein DF. Delineation of two drug-responsive anxiety syndromes. Psychopharmacologia 1964; 5: 397-408. 

Klein DF. Importance of psychiatric diagnosis in prediction of clinical drug effects. Arch Gen Psychiatry 1967; 16: 118-26.  

Klein DF. Drug therapy a means of syndromal identification  and nosological revision. In: Cole JO, Freedman AM, Friedhoff AJ, editors. Psychopathology and Psychopharmacology. Baltimore and London: The Johns Hopkins University Press; 1973. p. 143-60.   

Klein DF. Endogenomorphic depression. A conceptual and terminological revision. Arch Gen Psychiatry 1974; 31: 447-51.  

Klein DF. Anxiety re-conceptualized Comprehensive Psychiatry 1980; 21: 411-27

Klein DF. Anxiety re-conceptualized; Gleaming from pharmacological dissection - early experience with imipramine and anxiety. In: Klein DF, Rabkin JG, editors. Anxiety: New Research & Changing Concepts. New York: Raven Press; 1981. p. 235-62.

Klein DF. False suffocation alarms, spontaneous panics, and related conditions; An integrative hypothesis. Arch Gen Psychiatry 1993; 50: 306-17.

Klein DF, interviewed by Leo E. Hollister. In: Levine J, editor. Psychopharmacology. In Ban TA, editor. An Oral History of Neuropsychopharmacology. Volume 4. Brentwood: American College of Neuropsychopharmacology; 2011, pp. 187-99.

Klein DF, interviewed by John M. Davis. In: Blackwell B, editor. Update. In: Ban TA, editor. An Oral History of Neuropsychopharmacology. Volume 9. Brentwood: American College of Neuropsychopharmacology; 2011, pp. 205 -29.

Klein DF, Davis JM.  Diagnosis and Drug Treatment of Psychiatric Disorders. Baltimore: Williams & Wilkins; 1969.

Klein DF, Fink M. Psychiatric reaction patterns to imipramine. Am J Psychiatry 1962; 119: 432-8.

Klein DF, Gittelman-Klein R. Drug treatment of separation anxiety and depressive illness in children. In: Mendlewicz J, editor. Advances in Biological Psychiatry. Basel: Karger; 1978. p. 50-60.

Klein DF, Gittelman R, Quitkin F, Rifkin A. Diagnosis and Drug Treatment of Psychiatric Disorders: Adults and Children. Baltimore: Williams & Wilkins; 1980.

Levine J, editor. In: Ban TA, editor. An Oral History of Neuropsychopharmacology. Volume 9. Brentwood: American College of Neuropsychopharmacology; 2011

Papp LA, Klein DF, Gorman JM. Carbon dioxide hypersensitivity, hyperventilation and panic disorder. Am J Psychiatry 1993; 150: 1149-57.

Papp LA, Martinez JM, Klein DF, Liebowitz MR, Fyer AJ, Hollander I, Gorman JM. Arterial blood changes in panic disorder and lactate-induced panic. Psychiatry Research 1989; 28: 171-80.

Preter M, Lee SH, Petkova E, Vannucci M, Kim S, Klein DF. Controlled cross-over study in normal subjects of naloxone-preceding-lactate infusions: respiratory and subjective responses; relationship to endogenous opioid system, suffocation false alarm theory and childhood parental loss. Psychol Med 2010; 6: 1-9.


December 12, 2019