Developing and testing a drug from discovery to market takes ten years and costs more than £550 million. Few processes are more scrupulously designed or more carefully monitored, and while they often end in disappointment they seldom produce such dramatic and frightening effects as the trial of TGN 1412.

Drugs begin as bright ideas in the laboratory, and progress through screening tests that can screen more than a million compounds against a single target in a few weeks. The target might be, say, an enzyme known to play a key role in a disease, and the screening is designed to see if the compound inhibits or enhances the enzyme’s action.

TGN 1412 falls into a slightly different class. It is not a chemical but a biological agent, an antibody designed to lock on to a particular target in the immune system and modify its behaviour. Such “monoclonal antibodies” have shown huge promise in treating diseases such as rheumatoid arthritis.

TGN 1412 had shown “exquisite and unique ex vivo and in vivo T lymphocyte activating capacity and great therapeutic potential”, says TeGenero, its German manufacturer. In plain English, what this means is that in tests in culture dishes, using human cells, and in animals, it had performed as they had hoped.

Once a drug candidate has completed animal trials successfully it enters human trials. These usually consist of three phases, each phase involving increasing numbers of people.

Phase 1 trials are small, using only a handful of healthy volunteers. The purpose is to establish safety, and to assess what the most effective dose of the drug may be.

It was at this early stage that TGN 1412 demonstrated powerful and unexpected effects on the immune systems of the six healthy volunteers who took it. Phase 1 trials very seldom come up with shocks as profound as this, as the animal tests are designed to weed out any drugs likely to show toxic effects.

Phase 1 trials confer no benefits on those who take part in them. Sometimes volunteers are recruited among the altruistic, who simply want to help. But in this case, the volunteers were found through the website of Parexel, an international company based in Massachusetts and with offices in 36 countries around the world.

Parexel employs more than 5,000 people in a range of services to the drug industry, including the conduct of clinical trials. TeGenero is a tiny company, with only 15 employees, so was not equipped to do its own Phase 1 trial.

On its website yesterday, Parexel was advertising five clinical trials at its 36-bed unit at Northwick Park Hospital, including trial 68419, for a drug developed for the treatment of inflammatory diseases such as rheumatism. This was the TGN 1412 trial which was to go so dramatically wrong.

For this trial Parexel sought healthy males, aged between 18 and 40, who must not have suffered recently from flu, used any medication or recreational drugs, nor suffered from allergies. Other trials advertised at the same time were for two diabetes drugs, an antidepressant and a drug to treat asthma and chronic lung disease.

Because the trial offered them no therapeutic benefits, the volunteers could be paid for their time and inconvenience, the website explains. This payment is understood to have been £150 a day, for a commitment of around 15 days.

Almost all such Phase 1 trials are uneventful, and provide useful information about the effects and the pharmacokinetics of the drug — how quickly it is eliminated from the body, for example. For some reason so far unexplained, trial 68419 went horribly wrong.

Had all gone normally, the next stage would have been a Phase 2 trial, involving perhaps 200-400 people who suffer from arthritis. This would have been designed to test that TGN 1412 worked, and whether it had any side-effects.

Finally, before it could be marketed, it would have undergone a much larger Phase 3 trial, involving thousands of patients, which would have compared it with an established drug, and/or a placebo. This trial is designed to prove that the drug works in real life, and confers benefits that exceed any identifiable risks.

As things stand, TGN 1412 is never likely to get that far, unless it can be shown that the violent reactions were a consequence of contamination, or poor manufacturing methods. If they were caused by the drug itself, a lot of questions are likely to be asked about the animal testing phase and whether that held any hints of the disaster that was to follow.

'I won't do it again'

Five years ago, Jared Schiller, 25, took part in two separate medical trials for a blood-thinning drug and an anaesthetic drug. Although he reported dizziness during the second trial, he had been warned to expect the symptom and has since suffered no other ill effects.

Yesterday, he said he was unlikely to participate in further drug trials following the publicity surrounding the Parexel trial.

“I was a student so it was a very convenient way of making quite a lot of money in a very short space of time. I don't think I’d do it again.

“Although you are told that there are risks, it’s in the context that ‘these sort of things don't really happen’ . . . so you're kind of reassured. I thought they would have had all these things worked out. It’s worrying to think something like this could have happened.”

Times Recommends

• An officer and a gentle woman
• Dame Helen Mirren is worthy of the Epidaurus
• Foods that can help to prevent swine flu