Alarming evidence from a new British study shows that the Prozac class of antidepressants can make healthy men, women and children with no history of depression feel suicidal.

The research undermines the claims of Eli Lilly, makers of Prozac, that people who kill themselves while on the tablets do so because of their depression, and that the disease, not the drug, is to blame for their suicide.

Its findings are particularly worrying because of the increasing numbers of people, including children, who are being given the drugs by their GP for mild depression, and who are not seriously clinically ill.

Prozac, the wonder pill of the 1980s and 1990s, became the biggest drug company blockbuster of all time, prescribed to more than 38m people around the world. It became a metaphor for late 20th century life and a cult in its own right, enshrined in a book called Prozac Nation.

What began as a medicine for the clinically depressed has been transformed by use and demand into a pill for minor ills.

But while the happy drug works for many people, in a significant number it can take them to the edge of despair. The study, conducted by David Healy, director of the North Wales Department of Psychological Medicine, reveals the real dangers for some of the SSRIs (selective serotonin reuptake inhibitors), as Prozac and its imitators are called.

Nightmare

It found that two out of 20 healthy volunteers on an antidepressant in the Prozac class called Lustral (or Zoloft in the USA) became dangerously suicidal, compared with none of them when they were put on an antidepressant of a different class called reboxetine.

One 30-year-old woman who took part had a nightmare about having her throat slit after one week and by the end of a fortnight, was suicidal. "She felt hopeless and alone. It seemed that all she could do was to follow a thought that had been planted in her brain from some alien force. She suddenly decided she should go and throw herself in front of a car, that this was the only answer.

"It was as if there was nothing out there apart from the car, which she was going to throw herself under. She didn't think of her partner or child," says the study, published in the journal Primary Care Psychiatry.

Later she completed a diary entry, describing herself as jumpy, anxious and suspicious. "Her mind was racing and spiralling out of control. Then it went blank except for the clear thought that she must kill herself violently by throwing herself beneath a car or a train."

Dr Healy says the results of the research should be a warning to GPs prescribing any SSRIs. "They may not all be equally the same," he told the Guardian. "But the risk holds for the whole of the group. Generally the findings would indicate that women and children and those who are least ill may be most at risk."

All the drugs have been licensed as both safe and efficacious on the basis of data from clinical trials. But Dr Healy believes that there are serious problems with the reporting of side-effects in these trials, and that this has allowed drugs to be handed out to millions around the globe without their true risks being understood.

Volunteers taking part in the early trials were never asked whether they experienced any suicidal feelings or the restless agitation which can be the precursor of a suicide attempt. If patients in later trials said they felt suicidal, it was recorded as part of their depression.

Dr Healey has written to the Medicines Control Agency, which licences medicines in the UK, expressing his concern and pointing out that he believes patients who today become suicidal on SSRIs are in a state of "legal jeopardy".

Firms are using data from trials that were not designed to look at suicidality to prove that their drugs could not have caused it. Until the system for reporting side-effects is changed, he questions whether anybody should take part in clinical trials.

The new study's findings have emerged at a time of acute embarrassment for Eli Lilly. Its patent on Prozac (fluoxetine), is soon to expire, but it recently bought the licence for a second version of the drug, called R-fluoxetine. The patent for the new drug, it has just been revealed in the US, states that R-fluoxetine is improvement on Prozac specifically as it is less likely to cause "suicidal thoughts and self-mutilation".

Eli Lilly argues that the patent was filed by the American scientist Martin Teicher and the company Sepracor which devised the new drug.

Mr Teicher, in 1990, was the first to warn that patients on Prozac were becoming suicidal, but Eli Lilly has always dismissed his study on the basis that those patients were suffering chronic depression.

Dr Healy, the UK's leading historian of antidepressant medication who has given evidence against Lilly in litigation in the US, has frequently taken issue with the major study commissioned by the company to persuade the US Food and Drugs Administration that Prozac carried no suicide risk.

Dr Healy has argued there has never been a prospective study and that the retrospective 1991 Beasley study, as it is known, included only a small selection of the trials that had taken place on Prozac.

An internal memo released by Eli Lilly during recent litigation appears to support Healy's argument.

In one of a series of memos, dated August 27, 1990, a UK-based clinician tells Eli Lilly management in Indianapolis that critics will be suspicious of the fact that not all the trials were included and concludes that it gives "the impression that the question of whether fluoxetine provokes suicidal thoughts or not has not been properly considered."

'Killing herself appeared irresistible'

Case A: 30-year-old woman
After a few days on sertraline, she experienced agitation and anxiety, racing thoughts and restlessness, says the study. "Over the first weekend she had a nightmare about having her throat slit so that it gaped open and she imagined she bled to death in the bed." Other versions of the nightmare recurred during the next two nights.

At the start of week two, she remained restless, withdrawn and preoccupied. By Wednesday she was tearful and did not seem herself. "She described swings of emotion, with misery predominating but she was not depressed.

"She was advised to stop taking the drug and agreed to do so. She did not stop. In retrospect, it was almost as if she could not stop herself from taking the tablets."

On Thursday, the study monitors stopped her medication but the effects persisted. "That night she was seriously suicidal... On the Friday she telephoned early in the morning, distressed and tearful after the previous night. Her conversation was garbled. She described almost going out and killing herself."

She described feeling hopeless and alone and becoming obsessed with the idea of throwing herself under a car or perhaps a train.

"This clear thought appeared irresistible and its appearance seemed to put an end to the anxiety. It was trance-like and only broken by a telephone call, which came when she was about to act on the basis of this idea." She remains very disturbed by what happened.

Case B: 28-year-old woman
Within a few days she noticed she had become snappy and more assertive but she was liable to mood swings from cheerfulness to withdrawal. She also reported feeling restless.

She described finding herself in a state where she did not think through the consequences of what she did or said. She did not appear to feel afraid.

Driving home in a car with her mother, she saw a group of teenage boys beside the road who were making obscene gestures. She stopped the car in the middle of moving traffic, went over to them and grabbed one of them, telling him if he did anything like that again she would 'deck' him." Her mother reported that she was extremely frightened.

On two consecutive nights, "while awake or lucidly dreaming" she spent a long period lying in her bed fantasising about hanging herself from a beam across the bedroom ceiling.

She was aware that these thoughts were accompanied by an abnormal lack of concern as to whether her partner, mother or any others might find her.

She is not aware of having comparable thoughts before. The reason she did nothing, she explained afterwards, was because she was a coward and had a vestigial concern about being found by her son. The episode repeated itself the following night.

"There was a strong feeling that while on the drug that in some way she was being controlled and that suicide might happen."