They said it was Safe!

Prozac is the late 20th century's miracle drug - a medicine for a world that wants simple answers to life's complexities. Happiness is pill-shaped. Depression is soluble. No more worries. No more wrestling with the gut-wrenching anxieties thrown up by the pressure to succeed and the fact of our mortality. No more wondering whether it would be nobler to end it all. Unsurprising, then, that Prozac has been received across the globe with quasi-religious fervour. More than 38 million people have taken it. And, unlike the old prescription tranquillisers such as Valium and Librium, Prozac is said to be safe. It is almost impossible to kill yourself with an overdose. That has been its biggest-selling pitch - Prozac is simple, legal and safe. GPs are handing it out to teenagers, even to young children, in increasing numbers.

But since its launch in January 1988 in the US, and in the UK shortly after, when Prozac was let loose on whole populations rather than on selected patients in clinical trials, there has been a spate of disturbing accounts of violence and suicide committed by people prescribed the drug by their doctors. Some 200 cases have come to court in the US. Victims and families of killers have sued the multi-national Eli Lilly, manufacturers of the world's most commercially successful drug. Until recently, not one case reached a verdict. Either it was dropped, or Lilly settled out of court, sometimes for millions of dollars - Lilly's defence has always been the same: blame the disease, not the drug. Depressed people get put on Prozac. Depressed people are often suicidal. Keep on taking the tablets.

But earlier this year, for the first time, Lilly came up against a family in the US who would not settle. The Forsyths wanted a hearing. Internal documents belonging to Lilly were produced in court. And although Lilly won the case - the jury decided it could not hold it responsible for Bill Forsyth Sr's death - it may have lost the argument, for those documents showed that Lilly knew as long as 20 years ago that Prozac can produce in some people a strange, agitated state of mind that can trigger in them an unstoppable urge to commit suicide or murder. Dr David Healy, a leading UK psychiatrist in the field of anti-depressant medicine and author of The Antidepressant Era, the only comprehensive history of such drugs, believes that Lilly is guilty of a failure to warn doctors and public of the terrible potential consequences for some people of taking Prozac. "Based on published data and on Lilly's internal documents, the only reasonable estimate for the number of people who have worldwide, because of Prozac, tried to kill themselves since it was introduced would be a quarter of a million - around 25,000 will have actually succeeded," says Healy. Terrifying things happen to a number of people within the first few weeks of taking the drug, says Healy. They become agitated, restless and anxious. Out of the blue, and completely out of character, they may try to kill themselves in extremely violent ways, and they may try to take others with them.

What happened to Bill Forsyth Sr is typical of some people's catastrophic reaction to the drug, which hits the susceptible within days of starting on it. (The first Prozac case to come to litigation concerned Joseph Wesbecker, a Louisville printer, who took several automatic weapons to work one day and killed eight and injured 16 of his colleagues before turning the gun on himself.) Forsyth was a man of certainties. He was a go-getter, the sort of run-of-the-mill success story that America rejoices in. For 40 years, he'd been in the car business in California, land of the freeway, owning a car-rental firm based at Los Angeles airport. When the airport needed his space for expansion in 1986, it bought him out for big bucks.

So Bill retired. He had plenty of money to spend on the leisure and pleasure he'd never had time to enjoy. He and his wife spent four more years in California, where they had brought up their two children, Susan and Bill Jr. Then they moved to Hawaii, where their son had made his home with his wife and children. But life soon began to jar for Bill Sr. He found it hard to reconcile the simple, hedonistic life of his son with his own dogged, lifetime pursuit of ambition and material goals. And he and June, his wife of 37 years, were falling out. They had built themselves a luxurious house on Maui, but were under each other's feet, unused to being constantly together. Bill walked away a couple of times, flying back to LA for some space. Then he and June went to a marriage-guidance counsellor. They successfully sorted out their relationship.

But in December 1992 Bill began to have panic attacks. His doctor prescribed medication, which worried him a little: many years earlier, the self-imposed pressures of his business had led to heavy drinking, and he had not touched a drop for a very long time, so did not like the idea of taking mind-altering drugs. Still, he was the sort of man who wanted to do what the doctor told him, so he took his medicine. But it didn't work. Let's try something else, said the doctor. A new drug, Prozac. Obediently, Bill Sr took his pills. The very next day he experienced the Prozac miracle. He felt wonderful. The clouds had cleared. Bill called his doctor to tell him he felt 200% better.

The next day, the doctor got another call. It was from Bill Jr to tell him that a horrible change had come over his father. Bill Sr himself, who had rarely been in hospital in his life, had urgently demanded to be admitted to a psychiatric hospital. He spent a week in the Castle Medical Center, on the neighbouring island of Oahu, where doctors continued to give him Prozac. On March 3 1993, after 11 days on Prozac, Bill Sr went home at his own request. Bill Jr went round for dinner. Bill Sr and June planned to go out whale-watching with their son the next day. When they didn't turn up as arranged, he went to the house. He found a scene of carnage: during the night or early in the morning, his father had stabbed his mother 15 times and had then placed a serrated kitchen knife on a stool and impaled himself on it. Bill Jr and Susan and were devastated and disbelieving - never in a thousand years would they have guessed that their father might one day murder their mother and then kill himself in so violent a fashion. As far as they were concerned, there could be only one answer - that Prozac was responsible.

In March this year, their suit against Eli Lilly finally came to trial in Honolulu, Hawaii. In the run-up to the trial, the Forsyth family's lawyers contacted Dr Healy at his home in Bangor, north Wales. It was not the first time he had been asked to look at a case against Lilly that alleged Prozac had caused balanced individuals with minor depression to become suicidal killers. Every time, Healy, who is director of the north Wales department of psychological medicine, had come to the conclusion that there was no case to answer, and at first wasn't inclined to wade through the Forsyth papers. But he was about to fly to the US, anyway, so he relented. Okay, he told the lawyers, let me see the files when I get there. Several boxes of documents arrived in his hotel room. This case was, to Healy, clearer than any of the previous ones. Bill Sr had no history of mental illness. He had never shown any suicidal leanings. What had happened on the last night of his life was totally unexpected and out of character. Nobody would have predicted it, and nobody could understand it. Healy became increasingly convinced that Prozac had sent Forsyth into a homicidal, suicidal frenzy. He agreed to become an expert witness for the family against Eli Lilly.

What Healy has learned during the litigation has surprised and worried him. He believes, as he always has, that Prozac is a useful anti-depressant. But there is now a mound of evidence that, in a minority of cases, it induces a strange and disturbing state of mind that can lead to violence and suicide. This state of mind is a recognised psychiatric phenomenon, called akathisia. Akathisia was described by the Forsyths' attorney, Andy Vickery, as a sort of jitteriness or feeling "wired", like the effects of drinking too much strong black coffee. But on Prozac, the experience can be far more severe, sometimes leading to an inability to keep still and to restless pacing up and down. Vickery told the jury that it was like the onset of seasickness within hours or days of feeling fantastic at the start of a cruise. You try to ignore it. You tell people you feel fine, hoping it will go away. You can't stay in one place - you go outside for air, then back in to try to keep still. Then, just when you think it's over, you race to the side and retch your guts up.

Akathisia caused by antipsychotic drugs has long been recognised as leading to suicidal and homicidal-suicidal feelings. But antipsychotics such as chlorpromazine, while sometimes inducing suicidal feelings, take away the will to do anything about it. Never - before Prozac - had it been associated with antidepressants, which apply no such brakes on action. So doctors would not expect to see it. Lilly had issued no warnings that it could occur, even though akathisia had been spotted in some patients during the clinical trials before Prozac was given its licence. Lilly's own internal documents show it was identified as early as 1978. On August 2 of that year, when only three trials were under way, minutes of a meeting of the Fluoxetine (Prozac) Project Team run thus: "There have been a fairly large number of reports of adverse reactions... Another depressed patient developed psychosis... Akathisia and restlessness were reported in some patients." A similar meeting 10 days earlier had noted that "some patients have converted from severe depression to agitation within a few days; in one case the agitation was marked and the patient had to be taken off [the] drug." The minutes further state that "in future studies the use of benzodiazepines to control the agitation will be permitted". So, from that point on, Lilly's trial subjects would be put on tranquillisers to get them over the akathisia experienced by some in the early days on the drug. Yet once Prozac was on the market, there was no warning to doctors that such action might be necessary.

Those who developed akathisia or who had any suicidal tendencies were excluded from the trial data on the basis that they would otherwise obscure the results of the drug's success in treating depression. Yet the German licensing authority, the Bundes Gesundheit Amt (BGA), on scrutinising the results, expressed concerns about the drug's safety. On May 25, 1984, according to Lilly's internal documents, a letter from the BGA stated: "During the treatment with the preparation [Prozac], 16 suicide attempts were made, two of these with success. As patients with a risk of suicide were excluded from the studies, it is probable that this high proportion can be attributed to an action of the preparation [Prozac]."

In January 1985, the Germans told Lilly that they would not license the drug, giving "suicidal risk" as one of the reasons for their decision. Lilly's scientists continued trying to persuade the BGA to grant a licence, but focused most of their efforts on the US. By August 1989, it was clear to Lilly that the BGA would demand that Prozac carry a warning to GPs to the effect that they should be aware of the risk of suicide unless they gave patients sedation along with their Prozac. Such a warning, stating that there was a "risk of suicide", finally went on the German package insert in 1992. It goes on: "For his/her own safety, the patient must be sufficiently observed, until the antidepressive effect of Fluctin [Prozac] sets in. Taking an additional sedative may be necessary."

During the licensing process in the US, however, Lilly did not tell the Food and Drugs Administration (FDA) of the German concerns. Indeed, the firm's papers disclose a long and successful battle against the idea that Prozac could induce violence or suicide. They suggest that Lilly had an explicit strategy to blame the disease and not the drug, and that some of Lilly's own scientists had reservations about this. One of them, John Heiligenstein, wrote in an internal memo on September 14, 1990: "We feel caution should be exercised in a statement that 'suicidality and hostile acts in patients taking Prozac reflect the patient's disorder and not a causal relationship to Prozac'. Post-marketing reports [reports from GPs of suicides and violence in patients on the drug] are increasingly fuzzy and we have assigned, 'Yes, reasonably related', on several reports."

This memo was written two years after Prozac was granted a licence in the US, and just months after the most dangerous challenge to Lilly's position so far. Earlier in 1990, Martin Teicher, Jonathhan Cole and Carol Gllod, who were linked to Harvard University, ppublishhed a study of six patients on Prozac. They had a history of depression, but all, while on the drug, became violently suicidal in a way that surprised themselves and their doctors. The report noted that suicidal thoughts occurred within days or weeks of going on Prozac, or of having the dosage increased beyond a certain level, and that such thoughts disappeared when the patient stopped taking the drug. But Lilly insisted that Prozac did not cause akathisia. For good measure, the company asserted that the link between akathisia and suicide is questionable.

Lilly's internal documents of the time show that it was going through a difficult period. Some of the public criticism of its blockbuster drug was coming from the UK. "Anything that happens in the UK can threaten this drug [Prozac] in the US and worldwide," ran an internal memo from Leigh Thompson, one of Lilly's chief scientists. "We are now expending enormous efforts fending off attacks because of 1) relationship to murder and 2) inducing suicidal ideation [suicidal behaviour]." Another memo from Thompson ran: "I am concerned about reports I get re UK attitude toward Prozac safety. Leber [Dr Paul Leber of the FDA] suggested a few minutes ago we use CSM [the British Committee on Safety of Medicines] database to compare Prozac aggression and suicidal ideation with other antidepressants in the UK. Although he is a fan of Prozac and believes a lot of this is garbage, he is clearly a political creature and will have to respond to pressures. I hope Patrick [probably a Lilly employee, but not identified fully in the memo] realises that Lilly can go down the tubes if we lose Prozac, and just one event in the UK can cost us that."

This was how high the stakes had become. Without Prozac, Lilly could "go down the tubes". A memo from the German office to Lilly's US headquarters in that November indicates that Lilly was keen to root out the word "suicide" altogether from its database record of side-effects experienced by patients on the drug: Claude Bouchy and Hans Weber in Germany were alarmed by suggestions from their US superiors that, when GPs reported a suicide attempt on Prozac to them, they should record it as "overdose" (even though it is not possible to kill yourself by overdosing on Prozac), and that a GP's report of "suicidal ideation" should be recorded as "depression" - "Hans has medical problems with these directions and I have great concerns about it," runs a memo from Bouchy to Thompson. "I do not think I could explain to the BGA, to a judge, to a reporter or even to my family why we would do this, especially on the sensitive issue of suicide and suicide ideation."

Something had to be done. Lilly finally agreed to undertake the study suggested by the FDA, and look at the suicide rate among UK patients on Prozac, but it didn't. Instead, the company put together a "meta-analysis" from the clinical trials before the drug had been licensed (meta-analysis pools all the data from all available trials, and looks for trends from that very large sample of patients). The object was to find out whether more people on Prozac had become suicidal than those given a placebo or other treatment without knowing it. Lilly's own scientists, led by Charles Beasley, did the work. Beasley's study was rejected by the New England Journal of Medicine, but the British Medical Journal accepted and published it in 1991. It had "the appearance of scientific rigour", says Dr Healy, but it is clear, he says, in the light of the documents that emerged in the Forsyth case, that the so-called meta-analysis had included only 3,065 patients out of around 27,000 involved in the trials and that it had also included data that the FDA had rejected during licensing. Among those excluded from Lilly's study were the 5% of patients who had shown akathisia-like symptoms during the clinical trials and had dropped out, and also the 13 or 15 suicides - "given the populations being studied and the numbers involved, there should have been no suicides", says Dr Healy. Nor was there any mention of the fact that a considerable number of patients had been put on benzodiazepines to suppress the very problem that Lilly was claiming did not occur. Nor did the study mention any suicides since the licensing of the drug, which by that time numbered some 198 in the US and 94 elsewhere.

On the basis of this material, and on Lilly's constant reiteration that depression and not the drug causes suicide, the FDA's psychopharmacological drugs advisory committee decided in September 1991 that there was "no credible evidence of a causal link between the use of antidepressant drugs, includ ing Prozac, and suicidality or violent behaviour". The FDA voted six-three against demanding a warning on the label, but agreed that "more research is needed to further explore all the potential implications of these reports, not only for Prozac but for other antidepressants as well. Some members also expressed concern that some physicians may fail to properly monitor patients being treated with antidepressants." However, none of those on the panel would have been aware of the limi-tations of the Beasley study, because they would not have seen Lilly's internal documents.

It is this FDA conclusion from nine years ago that Lilly now cites every time questions are raised about suicides, homicides and its best-selling drug. The company's spokesman in Indianapolis told me : "That is more important than an attorney's selective manipulation of data. You have to take a look at the patient population. In people with depression there is probably a 15 per cent suicide rate. There is no evidence that Prozac causes suicide." Lilly adds to this the evidence from three small studies that, Healy argues, are flawed. One, for instance, was a study of 654 anxious - not depressed - patients, of whom only 187 were on Prozac. According to Lilly's argument, none of these patients should have committed suicide, because they were not depressed - and yet one of those on Prozac did.

The agitated state of mind that Prozac brings on in a minority of people who take it (perhaps one in four) might not have mattered if it had been aimed only at the seriously clinically depressed in hospital, where they would be regularly observed and sedated if they showed signs of acute anxiety. But Prozac is not that sort of drug. Prozac has always been aimed at the general population - those with a less significant depression or anxiety which did not wreck their lives but simply made them more difficult. These people get Prozac from their GP. And that GP is not around to see what effect the drug has on their behaviour. Nor is the GP warned that there might be a problem.

In 1995, new evidence of Prozac's dangers emerged from just the type of study that the FDA had requested years earlier, although it was not carried out for that specific purpose. A Boston-based scientist, Herschel Jick, carried out a study of suicides in the UK among people who had been prescribed antidepressants by their GP. Jick compared the suicide rates on 10 different antidepressants, and found that far more killed themselves on Prozac than on other drugs. Jick's study found that there were 187 sui cides per 100,000 depressed patients per year on Prozac. Lilly argues, however, that suicide rates among people with depression run at about 600 per 100,000. But those figures, says Healy, apply only to hospital patients with acute depression. Among the depressed population in the community, the published studies show the suicide rate is only around 30 per 100,000.

So, on those figures, 157 people prescribed Prozac by their GP out of every 100,000 will kill themselves because of it. In fact, says Healy, the likelihood of someone committing suicide on Prozac prescribed by their GP during their first month of treatment is 10 times greater than if they were untreated, which is a level of risk approaching that of smokers' likelihood of developing lung cancer. Lilly says that 38 million people worldwide have taken Prozac. Given that number, says Healy, 25,000 will have killed themselves and a quarter of a million will have tried. In the UK, between 1994 and 1999, at least one million people have taken Prozac, which, claims Healy, must mean 1,000 UK suicides and 10,000 attempts.

Healy does not want to see Prozac withdrawn, however. He wants, instead, to see a clear warning on the label, so that GPs will know they must keep a close watch on their patients for the first few weeks of treatment, and to give patients a sedative if they appear agitated. Left to themselves, with no doctor to please, patients suffering from akathisia will usually give up on the drug - they just feel too bad to continue - but Lilly's guide to the treatment, and the standard GP advice, is to carry on taking the medicine. Once a patient is over the bad patch, it is argued, they will feel terrific (see box opposite). They may well feel terrific - or they may be dead.

Teenagers in the UK are now being given Prozac by their doctors on the assumption that it is safe. They will not necessarily be closely monitored, let alone taken off it if they start getting agitated. "I have been notified of four or five cases of kids in their teens who have committed suicide by hanging themselves within weeks of going on Prozac - one of them only 13," said Healy. "It used to be almost unheard of for teenagers to kill themselves. They might make gestures and might overdose, but they usually do not die."

The Forsyths are going to appeal their case. Vickery, Warner and Co, of Houston, Texas, the law firm that represented the family, cannot believe that they lost. "I was shocked and disappointed for months," says Andy Vickery. "In the final argument, I told the jury that their verdict could save lives. I'm now representing families of people who killed themselves after that verdict." The arguments Vickery made in Honolulu may now be used in the UK: proceedings have recently been issued in the first British Prozac case. In 1996, 10 days after starting on Prozac, Reginald Payne from Wadebridge in Cornwall smothered his wife, Sally, to death and then jumped off a 200ft cliff. If Healy is right, and so many people have died for want of a warning to GPs who prescribe Prozac, it is an indictment not just of Eli Lilly but of the clinical trials system itself. In spite of all the work involved in these trials, all the volunteers who take part in the hope of helping themselves and benefiting mankind, and all the millions that are spent, they prove only that a drug will not obviously harm you and that it has some effect on the medical condition. They do not satisfactorily detect the side-effects that patients may go on to suffer. And once the drug is licensed, the reporting of side-effects by GPs who hear about them from their patients is notoriously unreliable.

The difficulty with a drug such as Prozac, which works on the mind, is that patients may not spontaneously report problems. During the trials, for instance, only 5% reported sexual problems - it is now known that half of those on the drug may experience changes in sexual functioning. And how are they supposed to report a side-effect, such as akathisia, that they've probably never heard of? The answer to this side-effects problem, says Healy, is to draw up a checklist. Patients in trials should be asked if they are suffering from any of a range of possible side-effects. One study has shown that patients who are asked only to tell the doctor if they have a problem may underestimate the side-effects they suffer by a factor of six to one. "As things stand at present," says Healy, "individuals entering a company-sponsored trial risk rendering a disservice both to themselves, to their fellow patients and to the community at large. They need, therefore, to consider before entering such trials. Their relatives and friends need to consider before letting them enter such trials."

In fact, Healy maintains, patients taking part in clinical trials where the side-effects are not recorded through a checklist may be putting themselves in legal jeopardy. If they fail to tell the doctor of any problems they experience - perhaps because they do not understand what is happening to them at the time - they may damage their chances of any compensation if they later suffer harm. Healy would like the UK ethics committees, which have to approve all trial protocols, to insist on the checklist approach to the reporting of side-effects. If that were to happen in the UK, he argues, the practice would inevitably spread worldwide, since virtually all trials are now international.

Healy has identified other problems with the licensing system, too, where patients' demands encourage manufacturers to concentrate on finding single "blockbuster" drugs that may make them millions, but that may equally break a company. The stakes are that high. So high, in fact, that Healy wonders about the legal advice companies are getting - several tobacco corporations, for example, have been advised by their lawyers not to do research into the dangers of cigarettes for fear of increasing their legal liability.

Whatever is going on inside the huge pharmaceutical multinationals today, all that really matters is the depressed patient who goes to see his GP tomorrow. He may be very miserable, and he may need help. He may think that his job or his marriage or both, are on the skids. But he is not so ill with depression that he wants to kill himself. Prozac may well be the answer, but it may also be a final solution. A warning and some close watching could make all the difference to him between life and death.