There is a problem with what everyone knows about antidepressant drugs. It isn’t true. The whole idea of mental distress being caused simply by a chemical imbalance is “a myth,”, sold to us by the drug companies.
In the United States, 40 percent of the regulators’ wages are paid by the drug companies, and in Britain, it’s 100 percent. The rules they have written are designed to make it extraordinarily easy to get a drug approved.
“There was never any basis for it, ever. It was just marketing copy. At the time the drugs came out in the early 1990s, you couldn’t have got any decent expert to go on a platform and say, ‘Look, there’s a lowering of serotonin in the brains of people who are depressed’ There wasn’t ever any evidence for it.” It hasn’t been discredited, because “it didn’t ever get ‘credited.” We don’t know what a “chemically balanced” brain would look like. The effects of these drugs on depression itself are in reality tiny. No matter what chemical you tinker with, you get the same outcome. Antidepressants are little more than active placebos, drugs with very little specific therapeutic benefit, but with serious side effects.
What do the people taking these different drugs actually have in common? Only one thing: the belief that the drugs work, because you believe you are being looked after and offered a solution. Clever marketing over solid empirical evidence.
The serotonin theory “is a lie. I don’t think we should dress it up and say, ‘Oh, well, maybe there’s evidence to support that.’ There isn’t.” Most people on these drugs, after an initial kick, remain depressed or anxious. The belief that antidepressants can cure depression chemically is simply wrong.
The year after I swallowed my first antidepressant, Tipper Gore, the wife of Vice President Al Gore, explained to the newspaper USA Today why she had recently become depressed. “It was definitely a clinical depression, one that I was going to have to have help to overcome,” she said. “What I learned about is your brain needs a certain amount of serotonin and when you run out of that, it’s like running out of gas.” Tens of millions of people, including me, were being told the same thing.
When Irving Kirsch discovered that these serotonin boosting drugs were not having the effects that everyone was being sold, complete/nonfiltered FDA drug company study/research records show that the effects of these drugs on depression itself are in reality tiny, he began, to his surprise, to ask an even more basic question.
What’s the evidence, he began to wonder, that depression is caused primarily by an imbalance of serotonin, or any other chemical, in the brain? Where did it come from?
The serotonin story began, Irving learned, quite by accident in a tuberculosis ward in New York City in the clammy summer of 1952, when some patients began to dance uncontrollably down a hospital corridor. A new drug named Marsilid had come along that doctors thought might help TB patients. It turned out it didn’t have much effect on TB, but the doctors noticed it did something else entirely. They could hardly miss it. It made the patients gleefully, joyfully euphoric, some began to dance frenetically.
So it wasn’t long before somebody decided, perfectly logically, to try to give it to depressed people, and it seemed to have a similar effect on them, for a short time. Not long after that, other drugs came along that seemed to have similar effects (also for short periods), ones named Ipronid and Imipramine. So what, people started to ask, could these new drugs have in common? And whatever it was, could it hold the key to unlocking depression?
Nobody really knew where to look, and so for a decade the question hung in the air, tantalizing researchers. And then in 1965, a British doctor called Alec Coppen came up with a theory. What if, he asked, all these drugs were increasing levels of serotonin in the brain? If that were true, it would suggest that depression might be caused by low levels of serotonin.
“It’s hard to overstate just how far out on a limb these scientists were climbing,” Dr. Gary Greenberg, who has written the history of this period, explains. “They really had no idea what serotonin was doing in the brain.” To be fair to the scientists who first put forward the idea, he says, they put it forward tentatively, as a suggestion. One of them said it was “at best a reductionist simplification,” and said it couldn’t be shown to be true “on the basis of data currently available.”
But a few years later, in the 1970s, it was finally possible to start testing these theories. It was discovered that you can give people a chemical brew that lowers their serotonin levels. So if this theory was right, if low serotonin caused depression, what should happen? After taking this brew, people should become depressed. So they tried it. They gave people a drug to lower their serotonin levels and watched to see what would happen. And, unless they had already been taking powerful drugs they didn’t become depressed. In fact, in the vast majority of patients, it didn’t affect their mood at all.
I went to see one of the first scientists to study these new antidepressants in Britain, Professor David Healy, in his clinic in Bangor, a town in the north of Wales. He has written the most detailed history of antidepressants we have. When it comes to the idea that depression is caused by low serotonin, he told me: “There was never any basis for it, ever. It was just marketing copy. At the time the drugs came out in the early 1990s, you couldn’t have got any decent expert to go on a platform and say, ‘Look, there’s a lowering of serotonin in the brains of people who are depressed’ There wasn’t ever any evidence for it.” It hasn’t been discredited, he said, because “it didn’t ever get ‘credited,’ in a sense. There wasn’t ever a point in time when 50 percent of the field actually believed it.” In the biggest study of serotonin’s effects on humans, it found no direct relationships with depression. Professor Andrew Skull of Princeton has said attributing depression to low serotonin is “deeply misleading and unscientific.“
It had been useful in only one sense. When the drug companies wanted to sell antidepressants to people like me and Tipper Gore, it was a great metaphor. It’s easy to grasp, and it gives you the impression that what antidepressants do is restore you to a natural state, the kind of balance that everyone else enjoys.
Irving learned that once serotonin was abandoned by scientists (but certainly not by drug company PR teams) as an explanation for depression and anxiety, there was a shift in scientific research. Okay, they said: if it’s not low serotonin that’s causing depression and anxiety, then it must be the lack of some other chemical. It was still taken for granted that these problems are caused by a chemical imbalance in the brain, and antidepressants work by correcting that chemical imbalance. If one chemical turns out not to be the psychological killer, they must start searching for another one.
But Irving began to ask an awkward question. If depression and anxiety are caused by a chemical imbalance, and antidepressants work by fixing that imbalance, then you have to account for something odd that he kept finding. Antidepressant drugs that increase serotonin in the brain have the same modest effect, in clinical trials, as drugs that reduce serotonin in the brain. And they have the same effect as drugs that increase another chemical, norepinephrine. And they have the same effect as drugs that increase another chemical, dopamine. In other words, no matter what chemical you tinker with, you get the same outcome.
So Irving asked: What do the people taking these different drugs actually have in common? Only, he found, one thing: the belief that the drugs work. It works, Irving believes, largely for the same reason that John Haygarth’s wand worked: because you believe you are being looked after and offered a solution.
After twenty years researching this at the highest level, Irving has come to believe that the notion depression is caused by a chemical imbalance is just “an accident of history,” produced by scientists initially misreading what they were seeing, and then drug companies selling that misperception to the world to cash in.
And so, Irving says, the primary explanation for depression offered in our culture starts to fall apart. The idea you feel terrible because of a “chemical imbalance” was built on a series of mistakes and errors. It has come as close to being proved wrong, he told me, as you ever get in science. It’s lying broken on the floor, like a neurochemical Humpty Dumpty with a very sad smile.
I had traveled a long way with Irving on his journey but I stopped there, startled. Could this really be true? I am trained in the social sciences, which is the kind of evidence that I’ll be discussing in the rest of this book. I’m not trained in the kind of science he is a specialist in. I wondered if I was misunderstanding him, or if he was a scientific outlier. So I read all that I could, and I got as many other scientists to explain it to me as possible. “There’s no evidence that there’s a chemical imbalance” in depressed or anxious people’s brains, Professor Joanna Moncrieff, one of the leading experts on this question-explained to me bluntly in her office at the University College of London. The term doesn’t really make any sense, she said: we don’t know what a “chemically balanced” brain would look like. People are told that drugs like antidepressants restore a natural balance to your brain, she said, but it’s not true-they create an artificial state. The whole idea of mental distress being caused simply by a chemical imbalance is “a myth,” she has come to believe, sold to us by the drug companies.
The clinical psychologist Dr. Lucy Johnstone was more blunt still. “Almost everything you were told was bullshit,” she said to me over coffee. The serotonin theory “is a lie. I don’t think we should dress it up and say, ‘Oh, well, maybe there’s evidence to support that.’ There isn’t.”
Yet it seemed wildly implausible to me that something so huge, one of the most popular drugs in the world, taken by so many people all around me, could be so wrong. Obviously, there are protections against this happening: huge hurdles of scientific testing that have to take place before a drug gets to our bathroom cabinets. I felt as if I had just landed in a flight from JFK to LAX, only to be told that the plane had been flown by a monkey the whole way. Surely there are procedures in place to stop something like this from happening? How could these drugs have gotten through the procedures in place, if they were really as limited as this deeper research suggested?
I discussed this with one of the leading scientists in this field, Professor John Ioannidis, who the Atlantic Monthly has said “may be one of the most influential scientists alive.” He says it is not surprising that the drug companies could simply override the evidence and get the drugs to market anyway, because in fact it happens all the time. He talked me through how these antidepressants got from the development stage to my mouth.
It works like this: “The companies are often running their own trials on their own products,” he said. That means they set up the clinical trial, and they get to decide who gets to see any results. So “they are judging their own products. They’re involving all these poor researchers who have no other source of funding, and who have little control over how the results will be written up and presented.” Once the scientific evidence is gathered, it’s not even the scientists who write it up much of the time. “Typically, it’s the company people who write up the published scientific reports.”
This evidence then goes to the regulators, whose job is to decide whether to allow the drug onto the market. But in the United States, 40 percent of the regulators’ wages are paid by the drug companies, and in Britain, it’s 100 percent. When a society is trying to figure out which drug is safe to put on the market, there are meant to be two teams: the drug company making the case for it, and a referee working for us, the public, figuring out if it properly works. But Professor Ioannidis was telling me that in this match, the referee is paid by the drug company team, and that team almost always wins.
The rules they have written are designed to make it extraordinarily easy to get a drug approved. All you have to do is produce two trials, any time, anywhere in the world, that suggest some positive effect of the drug. If there are two, and there is some effect, that’s enough. So you could have a situation in which there are one thousand scientific trials, and 998 find the drug doesn’t work at all, and two find there is a tiny effect, and that means the drug will be making its way to your local pharmacy.
“I think that this is a field that is seriously sick,” Professor Ioannidis told me. “The field is just sick and bought and corrupted, and I can’t describe it otherwise.” I asked him how it made him feel to have learned all of this. “It’s depressing,” he said. That’s ironic, I replied. “But it’s not depressing,” he responded, “to the severe extent that I would take SSRIs [antidepressants].”
I tried to laugh, but it caught in my throat.
Some people said to Irving, so what? Okay, so say it’s a placebo effect. Whatever the reason, people still feel better. Why break the spell? He explained: the evidence from the clinical trials suggests that the antidepressant effects are largely a placebo, but the side effects are mostly the result of the chemicals themselves, and they can be very severe.
“Of course,” Irving says, there’s “weight gain.” I massively ballooned, and saw the weight fall off almost as soon as I stopped. “We know that SSRIs [the new type of antidepressants] in particular contribute to sexual dysfunction, and the rates for most SSRIs are around 75 percent of treatment-engendered sexual dysfunction,” he continued. Though it’s painful to talk about, this rang true for me, too. In the years I was taking Paxil, I found my genitals were a lot less sensitive, and it took a really long time to ejaculate. This made sex painful and it reduced the pleasure I took from it. It was only when I stopped taking the drug and I started having more pleasurable sex again that I remembered regular sex is one of the best natural antidepressants in the world.
“In young people, these chemical antidepressants increase the risk of suicide. There’s a new Swedish study showing that it increases the risk of violent criminal behavior,” Irving continued. “In older people it increases the risk of death from all causes, increases the risk of stroke. In everybody, it increases the risk of type 2 diabetes. In pregnant women, it increases the risk of miscarriage and of having children born with autism or physical deformities. So all of these things are known.” And if you start experiencing these effects, it can be hard to stop, about 20 percent of people experience serious withdrawal symptoms.
So, he says, “if you want to use something to get its placebo effect, at least use something that’s safe.” We could be giving people the herb St. John’s Wort, Irving says, and we’d have all the positive placebo effects and none of these drawbacks. Although, of course, St. John’s Wort isn’t patented by the drug companies, so nobody would be making much profit off it.
By this time, Irving was starting, he told me softly, to feel “guilty” for having pushed those pills for all those years.
In 1802, John Haygarth revealed the true story of the wands to the public. Some people are really recovering from their pain for a time, he explained, but it’s not because of the power in the wands. It’s because of the power in their minds. It was a placebo effect, and it likely wouldn’t last, because it wasn’t solving the underlying problem.
This message angered almost everyone? Some felt duped by the people who had sold the expensive wands in the first place, but many more felt furious with Haygarth himself, and said he was clearly talking rubbish. “The intelligence excited great commotions, accompanied by threats and abuse,” he wrote. “A counterdeclaration was to be signed by a great number of very respectable persons”, including some leading scientists of the day, explaining that the wand worked, and its powers were physical, and real.
Since Irving published his early results, and as he has built on them over the years, the reaction has been similar. Nobody denies that the drug companies’ own data, submitted to the FDA, shows that antidepressants have only a really small effect over and above placebo. Nobody denies that my own drug company admitted privately that the drug I was given, Paxil, was not going to work for people like me, and they had to make a payout in court for their deception.
But some scientists, a considerable number, do dispute many of Kirsch’s wider arguments. I wanted to study carefully what they say. I hoped the old story could still, somehow, be saved. I turned to a man who, more than anyone else alive, successfully sold antidepressants to the wider public, and he did it because he believed it: he never took a cent from the drug companies.
In the 1990s, Dr. Peter Kramer was watching as patient after patient walked into his therapy office in Rhode Island, transformed before his eyes after they were given the new antidepressant drugs. It’s not just that they seemed to have improved; they became, he argued, “better than well”, they had more resilience and energy than the average person. The book he wrote about this, Listening to Prozac, became the bestselling book ever about antidepressants. I read it soon after I started taking the drugs. I was sure the process Peter described so compellingly was happening to me. I wrote about it, and I made his case to the public in articles and interviews.
So when Irving started to present his evidence, Peter, by then a professor at Brown Medical School, was horrified. He started taking apart Irving’s critique of antidepressants at length, in public, both in books and in a series of charged public debates.
His first argument is that Irving is not giving antidepressants enough time. The clinical trials he has analyzed, almost all the ones submitted to the regulator, typically last for four to eight weeks. But that isn’t enough. It takes longer for these drugs to have a real effect.
This seemed to me to be an important objection. Irving thought so, too. So he looked to see if there were any drug trials that had lasted longer, to find their results. It turns out there were two, and in the first, the placebo did the same as the drug, and in the second, the placebo did better.
Peter then pointed to another mistake he believed Irving had made. The antidepressant trials that Irving is looking at lump together two groups: moderately depressed people and severely depressed people. Maybe these drugs don’t work much for moderately depressed people, Peter concedes, but they do work for severely depressed people. He’s seen it. So when Irving adds up an average for everyone, lumping together the mildly depressed and the severely depressed, the effect of the drugs looks small, but that’s only because he’s diluting the real effect, as surely as Coke will lose its flavor if you mix it with pints and pints of water.
Again, Irving thought this was a potentially important point, and one he was keen to understand, so he went back over the studies he had drawn his data from. He discovered that, with a single exception, he had looked only at studies of people classed as having very severe depression.
This then led Peter to turn to his most powerful argument. It’s the heart of his case against Irving and for antidepressants.
In 2012, Peter went to watch some clinical trials being conducted, in a medical center that looked like a beautiful glass cube, and gazed out over expensive houses.
When the company there wants to conduct trials into antidepressants, they have two headaches. They have to recruit volunteers who will swallow potentially dangerous pills over a sustained period of time, but they are restricted by law to paying only small amounts: between $40 and $75. At the same time, they have to find people who have very specific mental health disorders, for example, if you are doing a trial for depression, they have to have only depression and no other complicating factors. Given all that, it’s pretty difficult for them to find anyone who will take part, so they often turn to quite desperate people, and they have to offer other things to tempt them. Peter watched as poor people were bused in from across the city to be offered a gorgeous buffet of care they’d never normally receive at home, therapy, a whole community of people who’d listen to them, a warm place to be during the day, medication, and money that could double their poverty-level income.
As he watched this, he was struck by something. The people who turn up at this center have a strong incentive to pretend to have any condition they happen to be studying there, and the for-profit companies conducting the clinical trials have a strong incentive to pretend to believe them. Peter looked on as both sides seemed to be effectively bullshitting each other. When he saw people being asked to rate how well the drugs had worked, he thought they were often clearly just giving the interviewer whatever answer they wanted.
So Peter concluded that the results from clinical trials of antidepressants, all the data we have, are meaningless. That means Irving is building his conclusion that their effect is very small (at best) on a heap of garbage, Peter declared. The trials themselves are fraudulent.
It’s a devastating point, and Peter has proved it quite powerfully. But it puzzled Irving when he heard it, and it puzzled me. The leading scientific defender of antidepressants, Peter Kramer, is making the case for them by saying that the scientific evidence for them is junk.
When I spoke to Peter, I told him that if he is right (and I think he is), then that’s not a case for the drugs. It’s a case against them. It means that, by law, they should never have been brought to market.
When I started to ask about this, in a friendly tone Peter became quite irritable, and said even bad trials can yield usable results. He soon changed the subject. Given that he puts so much weight on what he’s seen with his own eyes, I asked Peter what he would say to the people who claimed that John Haygarth’s wand worked, because they, too, were just believing what they saw with their own eyes. He said that in cases like that, “the collection of experts isn’t as expert or as numerous as what we’re talking about here. I mean, this would be [an] orders-of-magnitude bigger scandal if these were [like] just bones wrapped in cloth.”
Shortly after, he said: “I think I want to cut off this conversation.”
Even Peter Kramer had one note of caution to offer about these drugs. He stressed to me that the evidence he has seen only makes the case for prescribing antidepressants for six to twenty weeks. Beyond that, he said, “I think that the evidence is thinner, and my dedication to the arguments is less as you get to long-term use. I mean, does anyone really know about what fourteen years of use does in terms of harm and benefit? I think the answer is we don’t really know.” I felt anxious as he said that, I had already told him that I used the drugs for almost that long. Perhaps because he sensed my anxiety, he added: “Although I do think we’ve been reasonably lucky. People like you come off and function.”
Very few scientists now defend the idea that depression is simply caused by low levels of serotonin, but the debate about whether chemical antidepressants work for some other reason we don’t fully understand, is still ongoing. There is no scientific consensus. Many distinguished scientists agree with Irving Kirsch; many agree with Peter Kramer. I wasn’t sure what to take away from all of this, until Irving led me to one last piece of evidence. I think it tells us the most important fact we need to know about chemical antidepressants.
In the late 1990s, a group of scientists wanted to test the effects of the new SSRI antidepressants in a situation that wasn’t just a lab, or a clinical trial. They wanted to look at what happens in a more everyday situation, so they set up something called the Star-D Trial. It was pretty simple. A normal patient goes to the doctor and explains he’s depressed. The doctor talks through the options with him, and if they both agree, he starts taking an antidepressant. At this point, the scientists conducting the trial start to monitor the patient. If the antidepressant doesn’t work for him, he’s given another one. If that one doesn’t work, he’s given another one, and on and on until he gets one that feels as though it works. This is how it works for most of us out there in the real world: a majority of people who get prescribed antidepressants try more than one, or try more than one dosage, until they find the effect they’re looking for.
And what the trial found is that the drugs worked. Some 67 percent of patients did feel better, just like I did in those first months.
But then they found something else. Within a year, half of the patients were fully depressed again. Only one in three of the people who stayed on the pills had a lasting, proper recovery from their depression. (And even that exaggerates the effect, since we know many of those people would have recovered naturally without the pills.)
It seemed like my story, played out line by line. I felt better at first; the effect wore off; I tried increasing the dose, and then that wore off, too. When I realized that antidepressants weren’t working for me any more, that no matter how much I jacked up the dose, the sadness would still seep back through, I assumed there was something wrong with me.
Now I was reading the Star-D Trial’s results, and I realized I was normal. My experience was straight from the textbook: far from being an outlier, I had the typical antidepressant experience.
This evidence has been followed up several times since, and the proportion of people on antidepressants who continue to be depressed is found to be between 65 and 80 percent.
To me, this seems like the most crucial piece of evidence about antidepressants of all: most people on these drugs, after an initial kick, remain depressed or anxious.
I want to stress, some reputable scientists still believe that these drugs genuinely work for a minority of people who take them, due to a real chemical effect. It’s possible. Chemical antidepressants may well be a partial solution for a minority of depressed and anxious people, I certainly don’t want to take away anything that’s giving relief to anyone. If you feel helped by them, and the positives outweigh the side effects, you should carry on. (And if you are going to stop taking them, then it’s essential that you don’t do it overnight, because you can experience severe physical withdrawal symptoms and a great deal of panic as a result. I gradually reduced my dose very slowly, over six months, in consultation with my doctor, to prevent this from happening.)
But it is impossible, in the face of this evidence, to say they are enough, for a big majority of depressed and anxious people.
I couldn’t deny it any longer: for the vast majority we clearly needed to find a different story about what is making us feel this way, and a different set of solutions. But what, asked myself, bewildered, could they be?
The Emperor’s New Drugs
Everyone knows that antidepressant drugs are miracles of modern medicine. Professor Irving Kirsch knew this as well as anyone. But, as he discovered during his research, there is a problem with what everyone knows about antidepressant drugs. It isn’t true.
How did antidepressant drugs gain their reputation as a magic bullet for depression? And why has it taken so long for the story to become public? Answering these questions takes us to the point where the lines between clinical research and marketing disappear altogether.
Using the Freedom of Information Act, Kirsch accessed clinical trials that were withheld, by drug companies, from the public and from the doctors who prescribe antidepressants. What he found, and what he documents here, promises to bring revolutionary change to the way our society perceives, and consumes, antidepressants.
The Emperor’s New Drugs exposes what we have failed to see before: depression is not caused by a chemical imbalance in the brain; antidepressants are significantly more dangerous than other forms of treatment and are only marginally more effective than placebos; and, there are other ways to combat depression, treatments that don’t only include the empty promise of the antidepressant prescription.
This is not a book about alternative medicine and its outlandish claims. This is a book about fantasy and wishful thinking in the heart of clinical medicine, about the seductions of myth, and the final stubbornness of facts.
Irving Kirsch is a lecturer in medicine at the Harvard Medical School and a professor of psychology at Plymouth University, as well as professor emeritus of psychology at the University of Hull, and the University of Connecticut. He has published eight books and numerous scientific articles on placebo effects, antidepressant medication, hypnosis, and suggestion. His work has appeared in Science, Science News, New Scientist, New York Times, Newsweek, and BBC Focus and many other leading magazines, newspapers, and television documentaries.
Like most people, I used to think that antidepressants worked. As a clinical psychologist, I referred depressed psychotherapy clients to psychiatric colleagues for the prescription of medication, believing that it might help. Sometimes the antidepressant seemed to work; sometimes it did not. When it did work, I assumed it was the active ingredient in the antidepressant that was helping my clients cope with their psychological condition.
According to drug companies, more than 80 per cent of depressed patients can be treated successfully by antidepressants. Claims like this made these medications one of the most widely prescribed class of prescription drugs in the world, with global sales that make it a $19-billion-a-year industry. Newspaper and magazine articles heralded antidepressants as miracle drugs that had changed the lives of millions of people. Depression, we were told, is an illness a disease of the brain that can be cured by medication. I was not so sure that depression was really an illness, but I did believe that the drugs worked and that they could be a helpful adjunct to psychotherapy for very severely depressed clients. That is why I referred these clients to psychiatrists who could prescribe antidepressants that the clients could take while continuing in psychotherapy to work on the psychological issues that had made them depressed.
But was it really the drug they were taking that made my clients feel better? Perhaps I should have suspected that the improvement they reported might not have been a drug effect. People obtain considerable benefits from many medications, but they also can experience symptom improvement just by knowing they are being treated. This is called the placebo effect. As a researcher at the University of Connecticut, I had been studying placebo effects for many years. I was well aware of the power of belief to alleviate depression, and I understood that this was an important part of any treatment, be it psychological or pharmacological. But I also believed that antidepressant drugs added something substantial over and beyond the placebo effect.
As I wrote in my first book, ‘comparisons of antidepressive medication with placebo pills indicate that the former has a greater effect, the existing data suggest a pharmacologically specific effect of imipramine on depression’. As a researcher, I trusted the data as it had been presented in the published literature. I believed that antidepressants like imipramine were highly effective drugs, and I referred to this as ‘the established superiority of imipramine over placebo treatment’.
When I began the research that I describe in this book, I was not particularly interested in investigating the effects of antidepressants. But I was definitely interested in investigating placebo effects wherever I could find them, and it seemed to me that depression was a perfect place to look. Why did I expect to find a large placebo effect in the treatment of depression? If you ask depressed people to tell you what the most depressing thing in their lives is, many answer that it is their depression. Clinical depression is a debilitating condition. People with severe depression feel unbearably sad and anxious, at times to the point of considering suicide as a way to relieve the burden. They may be racked with feelings of worthlessness and guilt. Many suffer from insomnia, whereas others sleep too much and find it difficult to get out of bed in the morning. Some have difficulty concentrating and have lost interest in all of the activities that previously brought pleasure and meaning into their lives. Worst of all, they feel hopeless about ever recovering from this terrible state, and this sense of hopelessness may lead them to feel that life is not worth living. In short, depression is depressing. John Teasdale, a leading researcher on depression at Oxford and Cambridge universities, labelled this phenomenon ‘depression about depression’ and claimed that effective treatments for depression work at least in part by altering the sense of hopelessness that comes from being depressed about one’s own depression?!
Whereas hopelessness is a central feature of depression, hope lies at the core of the placebo effect. Placebos instil hope in patients by promising them relief from their distress. Genuine medical treatments also instil hope, and this is the placebo component of their effectiveness.
When the promise of relief instils hope, it counters a fundamental attribute of depression. Indeed, it is difficult to imagine any treatment successfully treating depression without reducing the sense of hopelessness that depressed people feel. Conversely, any treatment that reduces hopelessness must also assuage depression. So a convincing placebo ought to relieve depression.
It was with that in mind that one of my postgraduate students, Guy Sapirstein, and I set out to investigate the placebo effect in depression, an investigation that I describe in the first chapter of this book, and that produced the first of a series of surprises that transformed my views about antidepressants and their role in the treatment of depression. In this book I invite you to share this journey in which I moved from acceptance to dissent, and finally to a thorough rejection of the conventional view of antidepressants.
The drug companies claimed and still maintain that the effectiveness of antidepressants has been proven in published clinical trials showing that the drugs are substantially better than placebos (dummy pills with no active ingredients at all). But the data that Sapirstein and I examined told a very different story. Although many depressed patients improve when given medication, so do many who are given a placebo, and the difference between the drug response and the placebo response is not all that great.
What the published studies really indicate is that most of the improvement shown by depressed people when they take antidepressants is due to the placebo effect.
Our finding that most of the effects of antidepressants could be explained as a placebo effect was only the first of a number of surprises that changed my views about antidepressants. Following up on this research, I learned that the published clinical trials we had analysed were not the only studies assessing the effectiveness of antidepressants. I discovered that approximately 40 per cent of the clinical trials conducted had been withheld from publication by the drug companies that had sponsored them. By and large, these were studies that had failed to show a significant benefit from taking the actual drug. When we analysed all of the data, those that had been published and those that had been suppressed my colleagues and I were led to the inescapable conclusion that antidepressants are little more than active placebos, drugs with very little specific therapeutic benefit, but with serious side effects. I describe these analyses and the reaction to them in Chapters 3 and 4.
How can this be? Before a new drug is put on the market, it is subjected to rigorous testing. The drug companies sponsor expensive clinical trials, in which some patients are given medication and others are given placebos. The drug is considered effective only if patients given the real drug improve significantly more than patients given the placebos. Reports of these trials are then sent out to medical journals, where they are subjected to rigorous peer review before they are published. They are also sent to regulatory agencies, like the Food and Drug Administration (FDA) in the US, the Medicines and Healthcare products Regulatory Agency (MHRA) in the UK and the European Medicine Agency (EMEA) in the EU. These regulatory agencies carefully review the data on safety and effectiveness, before deciding whether to approve the drugs for marketing. So there must be substantial evidence backing the effectiveness of any medication that has reached the market.
And yet I remain convinced that antidepressant drugs are not effective treatments and that the idea of depression as a chemical imbalance in the brain is a myth. When I began to write this book, my claim was more modest. I believed that the clinical effectiveness of antidepressants had not been proven for most of the millions of patients to whom they are prescribed, but I also acknowledged that they might be beneficial to at least a subset of depressed patients. During the process of putting all of the data together, those that I had analysed over the years and newer data that have just recently seen the light of day, I realized that the situation was even worse than I thought.
The belief that antidepressants can cure depression chemically is simply wrong.
In this book I will share with you the process by which I came to this conclusion and the scientific evidence on which it is based. This includes evidence that was known to the pharmaceutical companies and to regulatory agencies, but that was intentionally withheld from prescribing physicians, their patients and even from the National Institute for Health and Clinical Excellence (NICE) when it was drawing up treatment guidelines for the National Health Service (NHS).
My colleagues and I obtained some of these hidden data by using the Freedom of Information Act in the US. We analysed the data and submitted the results for peer review to medical and psychological journals, where they were then published. Our analyses have become the focus of a national and international debate, in which many doctors have changed their prescribing habits and others have reacted with anger and incredulity.
My intention in this book is to present the data in a plain and straightforward way, so that you will be able to decide for yourself whether my conclusions about antidepressants are justified.
The conventional view of depression is that it is caused by a chemical imbalance in the brain. The basis for this idea was the belief that antidepressant drugs were effective treatments. Our analysis showing that most if not all of the effects of these medications are really placebo effects challenges this widespread view of depression. In Chapter 4 I examine the chemical-imbalance theory. You may be surprised to learn that it is actually a rather controversial theory and that there is not much scientific evidence to support it. While writing this chapter I came to an even stronger conclusion. It is not just that there is not much supportive evidence; rather, there is a ton of data indicating that the chemical-imbalance theory is simply wrong.
The chemical effect of antidepressant drugs may be small or even non-existent, but these medications do produce a powerful placebo effect. In Chapters 5 and 6 I examine the placebo effect itself. I look at the myriad of effects that placebos have been shown to have and explore the theories of how these effects are produced. I explain how placebos are able to produce substantial relief from depression, almost as much as that produced by medication, and the implications that this has for the treatment of depression.
Finally, in Chapter 7, I describe some of the alternatives to medication for the treatment of depression and assess the evidence for their effectiveness. One of my aims is to provide essential scientifically grounded information for making informed choices between the various treatment options that are available.
Much of what I write in this book will seem controversial, but it is all thoroughly grounded on scientific evidence, evidence that I describe in detail in this book. Furthermore, as controversial as my conclusions seem, there has been a growing acceptance of them. NICE has acknowledged the failure of antidepressant treatment to provide clinically meaningful benefits to most depressed patients; the UK government has instituted plans for providing alternative treatments; and neuroscientists have noted the inability of the chemical-imbalance theory to explain depression. We seem to be on the cusp of a revolution in the way we understand and treat depression.
Learning the facts behind the myths about antidepressants has been, for me, a journey of discovery. It was a journey filled with shocks and surprises, surprises about how drugs are tested and how they are approved, what doctors are told and what is kept hidden from them, what regulatory agencies know and what they don’t want you to know, and the myth of depression as a brain disease. I would like to share that journey with you. Perhaps you will find it as surprising and shocking as I did. It is my hope that making this information public will foster changes in the way new drugs are tested and approved in the future, in the public availability of the data and in the treatment of depression.
Listening to Prozac, but Hearing Placebo
In 1995 Guy Sapirstein and I set out to assess the placebo effect in the treatment of depression. Instead of doing a brand-new study, we decided to pool the results of previous studies in which placebos had been used to treat depression and analyse them together. What we did is called a meta-analysis, and it is a common technique for making sense of the data when a large number of studies have been done to answer a particular question. It was once considered somewhat controversial, but meta-analyses are now common features in all of the leading medical journals. Indeed, it is hard to see how one could interpret the results of large numbers of studies without the aid of a meta-anaiysis.
In doing our meta-analysis, it was not enough to find studies in which depressed patients had been given placebos. We also needed to find studies in which depression had been tracked in patients who were not given any treatment at all. This was to make sure that any effect we found was really due to the administration of the placebo. To better understand the reason for this, imagine that you are investigating a new remedy for colds. If the patients are given the new medicine, they get better, if they are given placebos, they also get better. Seeing these data, you might be tempted to think that the improvement was a placebo effect. But people recover from colds even if you give them nothing at all. So when the patients in our imaginary study took a dummy pill and their colds got better, the improvement may have had nothing to do with the placebo effect. It might simply have been due to the passage of time and the fact that colds are short-lasting illnesses.
Spontaneous improvement is not limited to colds. It can also happen when people are depressed. Because people sometimes recover from bouts of depression with no treatment at all, seeing that a person has become less depressed after taking a placebo does not mean that the person has experienced a placebo effect. The improvement could have been due to any of a number of other factors. For example, people can get better because of positive changes in life circumstances, such as finding a job after a period of unemployment or meeting a new romantic partner. Improvement can also be facilitated by the loving support of friends and family. Sometimes a good friend can function as a surrogate therapist. In fact, a very influential book on psychotherapy bore the title Psychotherapy: The Purchase of Friendship. The author did not claim that psychotherapy was merely friendship, but the title does make the point that it can be very therapeutic to have a friend who is empathic and knows how to listen.
The point is that without comparing the effect of placebos against rates of spontaneous recovery, it is impossible to assess the placebo effect. Just as we have to control for the placebo effect to evaluate the effect of a drug, so too we have to control for the passage of time when assessing the placebo effect. The drug effect is the difference between what happens when people are given the active drug and what happens when they are given the placebo. Analogously, the placebo effect is the difference between what happens when people are given placebos and what happens when they are not treated at all.
It is rare for a study to focus on the placebo effect or on the effect of the simple passage of time, for that matter. So where were we to find our placebo data and no-treatment data? We found our placebo data in clinical studies of antidepressants, and our no-treatment data in clinical studies of psychotherapy. It is common to have no-treatment or wait-list control groups in studies of the effects of psychotherapy. These groups consist of patients who are not given any treatment at all during the course of the study, although they may be placed on a wait list and given treatment after the research is concluded.
For the purpose of our research, Sapirstein and I were not particularly interested in the effects of the antidepressants or psychotherapy. What we were interested in was the placebo effect. But since we had the treatment data to hand, we looked at them as well. And, as it turned out, it was the comparison of drug and placebo that proved to be the most interesting part of our study.
All told, we analysed 38 clinical trials involving more than 3,000 depressed patients. We looked at the average improvement during the course of the study in each of the four types of groups: drug, placebo, psychotherapy and no-treatment. I am going to use a graph here (Figure 1.1) to show what the data tell us. Although the text will have a couple more such charts, I am going to keep them to a minimum. But this is one that I think we need, to make the point clearly. What the graph shows is that there was substantial improvement in both the drug and psychotherapy groups. People got better when given either form of treatment, and the difference between the two was not significant. People also got better when given placebos, and here too the improvement was remarkably large, although not as great as the improvement following drugs or psychotherapy. In contrast, the patients who had not been given any treatment at all showed relatively little improvement.
The first thing to notice in this graph is the difference in improvement between patients given placebos and patients not given any treatment at all. This difference shows that most of the improvement in the placebo groups was produced by the fact that they had been given placebos. The reduction in depression that people experienced was not just caused by the passage of time, the natural course of depression or any of the other factors that might produce an improvement in untreated patients. It was a placebo effect. and it was powerful.
Figure 1.1. Average improvement on drug, psychotherapy, placebo and no treatment. ‘lmprovement’ refers to the reduction of symptoms on scales used to measure depression. The numbers are called ‘effect sizes’. They are commonly used when the results of different studies are pooled together. Typically, effect sizes of 0.5 are considered moderate, whereas effect sizes of 0.8 are considered large. So the graph shows that antidepressants, psychotherapy and placebos produce large changes in the symptoms of depression, but there was only a relatively small average improvement in people who were not given any treatment at all.
One thing to learn from these data is that doing nothing is not the best way to respond to depression. People should not just wait to recover spontaneously from clinical depression, nor should they be expected just to snap out of it. There may be some improvement that is associated with the simple passage of time, but compared to doing nothing at all, treatment even if it is just placebo treatment provides substantial benefit.
Sapirstein and I were not surprised to find that there was a powerful placebo effect in the treatment of depression. Actually, we were quite pleased. That was our hypothesis and our reason for doing the study. What did surprise us, however, was how small the difference was between the response to the drug and the response to the placebo. That difference is the drug effect. Although the drug effect in the published clinical trials that we had analysed was statistically significant, it was much smaller than we had anticipated. Much of the therapeutic response to the drug was due to the placebo effect.
The relatively small size of the drug effect was the first of a series of surprises that the antidepressant data had in store for us.
One way to understand the size of the drug effect is to think about it as only a part of the improvement that patients experience when taking medication. Part of the improvement might be spontaneous that is, it might have occurred without any treatment at all and part may be a placebo effect. What is left over after you subtract spontaneous improvement and the placebo effect is the drug effect. You can see in Figure 1.1 that improvement in patients who had been given a placebo was about 75 per cent of the response to the real medication. That means that only 25 per cent of the benefit of antidepressant treatment was really due to the chemical effect of the drug. It also means that 50 per cent of the improvement was a placebo effect. In other words, the placebo effect was twice as large as the drug effect.
The drug effect seemed rather small to us, considering that these medications had been heralded as a revolution in the treatment of depression, blockbuster drugs that have been prescribed to hundreds of millions of patients, with annual sales totalling billions of pounds: Sapirstein and I must have done something wrong in either collecting or analysing the data. But what? We spent months trying to figure it out.
ARE ALL DRUGS CREATED EQUAL? DOUBLEBLIND OR DOUBLE-TALK
One thing that occurred to us, when considering how surprisingly small the drug effect was in the clinical trials we had analysed, was that a number of different medications had been assessed in those studies. Perhaps some of them were effective, whereas others were not. If this were the case, we had underestimated the benefits of effective drugs by lumping them together with ineffective medications. So before we sent our paper out for review, we went back to the data and examined the type of drugs that had been administered in each of the clinical trials in our meta-analysis.
We found that some of these trials had assessed tricyclic antidepressants, an older type of medication that was the most commonly used antidepressant in the 1960s and 1970s. In other trials, the focus was on selective serotonin reuptake inhibitors (SSRIs) like Prozac (fluoxetine), the first of the ‘new-generation’ drugs that replaced tricyclics as the top-selling type of antidepressant. And there were other types of antidepressants investigated in these trials as well. When we reanalysed the data, examining the drug effect and the placebo effect for each type of medication separately, we found that the diversity of drugs had not affected the outcome of our analysis. In fact, the data were remarkably consistent much more so than is usually the case when one analyses different groups of data. Not only did all of these medications produce the same degree of improvement in depression, but also, in each case, only 25 per cent of the improvement was due to the effect of the drug. The rest could be explained by the passage of time and the placebo effect.
The lack of difference we found between one class of antidepressants and another is now a rather frequent finding in antidepressant research. The newer antidepressants (SSRIs, for example) are no more effective than the older medications. Their advantage is that their side effects are less troubling, so that patients are more likely to stay on them rather than discontinue treatment. Still, the consistency of the size of the drug effect was surprising. It was not just that the percentages were close; they were virtually identical. They ranged from 24 to 26 per cent. At the time I thought, ‘What a nice coincidence! It will look great in a PowerPoint slide when I am invited to speak on this topic.’ But since then I have been struck by similar instances in which the consistency of the data is remarkable, and it is part of what has transformed me from a doubter to a disbeliever. I will note similar consistencies as we encounter them in this book.
The consistency of the effects of different types of antidepressants meant that we had not underestimated the antidepressant drug effect by lumping together the effects of more effective and less effective drugs. But our re-examination of the data in our meta-analysis held another surprise for us. Some of the medications we had analysed were not antidepressants at all, even though they had been evaluated for their effects on depression. One was a barbiturate, a depressant that had been used as a sleeping aid, before being replaced by less dangerous medications. Another was a benzodiazepine a sedative that has largely replaced the more dangerous barbiturates. Yet another was a synthetic thyroid hormone that had been given to depressed patients who did not have a thyroid disorder. Although none of these drugs are considered antidepressants, their effects on depression were every bit as great as those of antidepressants and significantly better than placebos. Joanna Moncrieff, a psychiatrist at University College London, has since listed other drugs that have been shown to be as effective as medications for depression. These include antipsychotic drugs, stimulants and herbal remedies. Opiates are also better than placebos, but I have not seen them compared to antidepressants.
If sedatives, barbiturates, antipsychotic drugs, stimulants, opiates and thyroid medications all outperform inert placebos in the treatment of depression, does this mean that any active drug can function as an antidepressant? Apparently not. In September 1998 the pharmaceutical company Merck announced the discovery of a novel antidepressant with a completely different mode of action than other medications for depression. This new drug, which they later marketed under the trade name Emend for the prevention of nausea and vomiting due to chemotherapy, seemed to show considerable promise as an antidepressant in early clinical trials. Four months later the company announced its decision to pull the plug on the drug as a treatment for depression. The reason? It could not find a significant benefit for the active drug over placebos in subsequent clinical trials.
This was unfortunate for a number of reasons. One is that the announcement caused a 5 per cent drop in the value of the company’s stock. Another is that the drug had an important advantage over current antidepressants, it produced substantially fewer side effects. The relative lack of side effects had been one reason for the enthusiasm about Merck’s new antidepressant. However, it may also have been the reason for its subsequent failure in controlled clinical trials. It seems that easily noticeable side effects are needed to show antidepressant benefit for an active drug compared to a placebo.
. . .
THE EMPEROR’S NEW DRUGS
by Irving Kirsch
get it at Amazon.com