The first inkling that doctors were facing an unprecedented medical crisis began with a call to the critical care unit at Northwick Park Hospital last Monday afternoon. Patients on the seventh floor were having strange reactions to a new drug, they were told. Some were having breathing difficulties, others were losing consciousness.

The news triggered alarm, but nothing more. The unit's six beds were already occupied by patients who were too ill to be moved. So a separate bay was taken over and the first patients rushed down. It was then that staff realised the nature of the calamity they were up against. Within minutes, the bay was filled with the writhing bodies of young men screaming for help.

Consultants were summoned from their homes. All were baffled. Even the team who had administered the test drugs could not explain why the men were slipping in and out of comas, why their heads had puffed up to twice their normal sizes, why some were vomiting, or why some were in such pain. Their only real comfort was provided by an experienced nurse who supervised the installation of drips and ventilators.

By the time Ganesh Suntharalingam, director of intensive care, arrived at the unit in the sprawling, windswept Northwick campus in Harrow, north-west London, the scene was horrific. Six healthy young men - including British Asians, an Australian, a New Zealander, and a South African - were now in the throes of massive inflammatory reactions.

The men, who had been contracted by the US drug testing company Parexel to test the anti-cancer drug TGN1412, were put on ventilators. Then some began showing signs of kidney failure. Blood filtration units were brought it to clean out their veins and arteries. Relatives were called and told to expect the worst. Only a miracle would see them through, one was told.

By any standards, it is a shocking story. How could a carefully controlled medical trial have produced such mayhem in such a short space of time? What exactly happened to those six healthy young men? Why did the rigorous rules laid down by the Medicines and Healthcare Products Regulatory Agency (MHRA) fail to halt last week's dreadful events? And what are the implications for the chaos triggered by the drug-trial disaster?

The MHRA has announced it has begun an investigation into the calamity of Parexel's trial, one that should provide answers to these questions. 'This was completely unexpected,' said Simon Gregor, spokesman for the MHRA. 'There are a number of possibilities which we have to consider: was there a manufacturing problem or some form of contamination or was there something about this product that could not have shown up at an earlier stage?' He predicted it would take several weeks, possibly months, before any inquiries would produce results.

Few expect a whitewash. This country is one of the world's biotechnology leaders but that reputation will be seriously tarnished if it is found there were significant regulatory errors involved in the medical disaster at Northwick Park.

Consider the simple issue of timing. It is now clear the doses of TGN1412 used in the trial were given with extraordinary rapidity, with an interval of only minutes separating each patient's injection. Normally, doses would be interspersed with waits of hours to give researchers time to study volunteers for side-effects. Indeed, sometimes doses are given at intervals of days, even weeks, as Dr Kate Law, of Cancer Research UK, said.

'We usually give one person a small dose. Then weeks later we give another volunteer an equally small dose,' she said. That precaution was clearly not enforced last week.

Within minutes of the last volunteer getting his TGN1412, the first injected recruit had begun to complain of a severe headache, backache, fever and pain. He tore his shirt off and yelled he was burning. 'An Asian guy next to me started screaming and his breathing went haywire as though he was having a terrible panic attack,' said one of the volunteers, Raste Khan, who had been given a placebo.

'They put an oxygen mask on him but he kept tearing it off. He was shouting "Doctor, doctor, please help me!" He started convulsing, shouting that he was getting shooting pains in his back. People were fainting and coming back to consciousness. It was terrifying. I kept expecting it to happen to me at any moment. But I felt fine and didn't know why.'

The whole affair 'is shocking because it is unprecedented,' said Simon Best, chair of the UK Bioindustry Association. 'If you think about the number of Phase I trials carried out in this country, you've never heard of anything having such a terrible impact.'

In fact, Phase 1 trials - which are carried out merely to determine a drug's toxicity, its efficacy being assessed at subsequent Phase 2 and 3 trials - can all too frequently sound the death knell for a new medicine. 'Between 20 to 30 per cent of drug developments go no further than this stage because of ill-effects suffered by volunteers,' Ray Noble, a UK medical ethicist, pointed out. However, no drug has ever seen such an abrupt termination as that of TGN1412.

Not that Parexel are newcomers to the business. The company made more than $100m last year carrying out clinical trials for researchers and has plenty of experience in the field, with more than 5,000 employees in 39 countries. All the doctors and nurses at its Northwick Park unit are employees and the company takes great care in getting participants to sign full consent forms and to be given medical examinations. The company also provides pool tables, digital TVs and computer games for volunteers.

And that leads us to another vexed issue: money. Khan, like the other volunteers taking part in the trial of TGN1412, had joined up because of the fee he would receive: £2,000. Last week, medical ethicists and trial experts raised serious worries about payments of this level. 'People should be recompensed for the time they take out of their lives to become involved in a trial but that is a different matter from paying inducements,' said Janet Derbyshire, of the Medical Research Council (MRC).

This point was backed by Ray Noble, a UK medical ethicist. 'People who are designing these trials have to make sure they do not offer so much money that young people simply ignore the boxes about their medical conditions in their consent forms in order to make sure they get the thousands of pounds they need to pay off their student loans.'

Indeed, it emerged last week that Parexel - which launched the trial on behalf of the German drug company TeGenero, the manufacturers of TGN1412 - had breached guidelines laid down by the Association of British Pharmaceutical Industries, the drug industry's main trade body, over the offering of inducements. 'Neither payment, nor the level thereof, should be mentioned in a public notice,' these state. Yet Parexel's web site for enrolling volunteers to its TGN1412 trials clearly states that recruits would be 'paid for your time and inconvenience'.

These issues are serious enough. However, of all the concerns raised by scientists last week, it was the decision by Parexel to use healthy volunteers, not cancer patients, as guinea pigs that has caused most unease. The issue takes us to the very heart of the crisis that has engulfed Northwick Park and to the nature of TGN1412, a drug known as a monoclonal antibody or mab.

Mabs were first developed in Britain by Cesar Milstein and George Köhler, who shared the 1984 Nobel Prize for Physiology for the achievement. They are artificial versions of natural chemicals called antibodies, proteins that are made in the body and which latch on to the surfaces of cells. Herceptin, the breast cancer drug, is a monoclonal antibody. It fits on to a site on a tumour cell, like a key sliding into a lock, initiating a chemical reaction inside the cell that causes it to stop dividing. Thus the spread of breast cancer is halted. It is this specificity, this ability to use mabs as tiny biological guided missiles that makes them so powerful, and dangerous.

Today, sales of antibody drugs now stand at more than £7bn a year and there are 17 of them on the market. Most work by stopping a cell from dividing. But one or two do the opposite. They stimulate a cell into action. And crucially, TGN1412 is one of these.

The drug targets a site called CD28 which lies on the surfaces of white blood cells called T-cells, key components of the body's immune system. In effect, the drug latches on to the T-cell and initiates a chemical reaction that should fire the body's immune system into action. Thus it was hoped TGN1412 could be used to treat a disease called B-cell chronic lymphatic leukaemia (B-CLL), and also auto-immune diseases which include rheumatoid arthritis and multiple sclerosis.

'The aim would be to kick the body's defences into a highly active state so that they could mount a more effective attack on the proliferating B-cells that were causing a patient's leukaemia,' said Dr David Chiswell, a monoclonal expert and former chairman of the BioIndustry Forum.

Patients who would be the target of such drugs have weakened immune systems. Healthy people do not. Hence the puzzle over the decision to use fit young people for the trial. Instead of prodding weakened bodily defences into action, TGN1412 let loose a fully armed missile, unleashing a chain reaction in which T-cells released cytokines - messengers that control the immune system - that in turn triggered other T-cells into releasing more cytokines. Doctors have a name for this deadly cascade. It is called a cytokine storm.

In the lungs, fluids and immune cells accumulate and can block off air passageways. Blood vessels become inflamed and organs are eaten up. The effect is often grotesque. Myfanwy Marshall said her partner, one of the trial's volunteers, had been left looking like 'the Elephant Man, completely puffed up,' his face and head had swollen horribly, and were now a 'weird purple and yellow colour'.

So why employ healthy volunteers who would have been susceptible to cytokine storms? For most experts, the decision seems questionable. For example, both Cancer Research UK and the MRC remained very doubtful about the decision. 'We would have been much more cautious,' said Derbyshire, a sentiment echoed by Law. 'I doubt if we would have given approval for the trial as it was carried out.'

Only the MHRA probe will reveal the truth, though The Observer can reveal that when TGN1412 was in pre-clinical studies in animals in 2004, TeGenero was at that stage considering using cancer patients for initial safety tests. Yet when they began trials last Monday, they picked eight fit young men. Why?

One answer, put forward by a researcher involved in early work on TGN1412, suggests the company thought toxicity data from animal studies of TGN1412, which included experiments on rabbits and monkeys, was so encouraging it decided to test the drug on healthy people straightaway, in order to look at the spectrum of potential side-effects.

Or the company may simply have encountered a problem in recruiting cancer patients, a process that takes up more more time than any other part of a clinical trial. Almost half of all trial delays result from difficulties in finding volunteers, and that can equate with a company losing millions of pounds in sales for a new drug. In 2004, 11 per cent of all newly diagnosed cancer patients agreed to participate in some kind of drug trial. However, as more new medical technologies come on stream, a bigger army of recruits will be needed to join in research programmes.

Whatever the answer, the issue of financial pressures that bear down on drug researchers from their boardrooms cannot be ignored in this grim tale. 'The people behind this company are reputable scientists,' said one UK immunologist last week. 'I can only think they may have got carried away and didn't see the apparent dangers in the design of this trial. But the first question I would have asked is this: how are you going to avoid a cytokine storm? You know this drug is going to trigger the T-cells into a burst of maximal activity. Is anyone surprised by what happened?

Yesterday Dr Thomas Hanke, TeGenero's chief scientific officer, defended the decision to use healthy volunteers. 'We had early evidence [from animal studies] that the effect of the drug depends on the status of the patient, so it dampens the immune response for patients with auto-immune disease and it boosts the response for the other group [with cancer]. We are so terribly sorry it went so wrong. We are devastated by the tragic events that have happened.'

Certainly, the story of TGN1412 spells out many warnings for the pharmaceutical industry and for the medical establishment. 'If nothing else, the reaction of these patients tells us one thing - that there are things going on at the CD28 site, the target for TGN1412, that we do not understand,' said monoclonal expert, David Glover, a former chief medical officer for Cambridge Antibody Technology.

'Any company who is researching on CD28 drugs today should stop what they are doing at once until we find out more about what is going on there.'

It is equally clear, added Glover, that in an era in which more and more new medicines are created by biological means, that new procedures be urgently introduced to prevent any more disasters.

TGN1412 is a humanised monoclonal antibody. In other words, it is made up mostly of human protein. Testing it in animals is therefore a tricky issue. 'What you need to do is understand what the issues are that you face and think carefully about them and find creative solutions.' Genetically engineered mice with a humanised immune system may be part of the answer, he added.

Another idea proposed by Glover is a new super-safe technique called 'microdosing' which was just starting to attract the attention of the pharmaceutical industry. This involves producing a blister on a patient's arm and exposing it to tiny amounts of the drug, rather than treating the whole person.

The blister fills with inflammatory fluid, containing cells whose reaction to the drug can be monitored. 'This allows you to get a handle on the effect before you expose the whole patient to it,' added Glover.

Certainly, action will be needed to stop any repetition of the tragedy that unfolded at Northwick Park. The sight of loved ones, horribly disfigured, will be indelibly fixed in the minds of friends and relatives who gathered by victims' bedsides.

Nor was their suffering helped by the behaviour of the hospital. At one point, those watching TVs in the patients' rooms listened to a hospital press statement which said they might die within hours. One relative, whose son is an affected volunteer, said friends and families were left panicking. 'One minute he [his son] was being told that in two days he would be better. But on the outside they are saying that there was little hope.'

Yesterday, two of the stricken volunteers had been taken off organ support after responding to treatment and the conditions of two others were reported to be improving. Four are now conscious and able to talk. But two - 21-year-old Ryan Wilson, of Highbury, north London and a 28-year-old assistant bar manager - remain in a critical condition.

Meanwhile, Parexel's recruitment website for its drug trials continues to urge volunteers to come forward. Yesterday it was still displaying photographs of happy young people playing pool and chess - while another showed a cheque for an undisclosed sum being handed over to a lucky recipient.

One man's experience of a clinical trial
Clint Witchalls

As someone who had been putting off having a vasectomy for years, I was intrigued to read about a clinical trial for a male 'pill'. Unfortunately, the trial was about to fold as not enough volunteers were stepping forward. I called the hospital and told them I'd like to volunteer.

The 'pill' was actually a progesterone implant supplemented with a testosterone injection. What I hadn't factored into the equation was that all clinical trials are randomised, double-blind and placebo controlled. That is, I might be given a placebo. The doctor explained that I would have to continue using whatever contraception I currently used.

This wasn't enough to deter me. For the previous couple of years, I'd suffered from severe hypochondria. On this trial a doctor would monitor my health very closely on a weekly basis. Everything from my bone density to the levels of C-reactive protein in my blood would be scrutinised. I signed the contract for an 18-month trial for which I would get paid £1,200.

Things went swimmingly for the first few months, but then my moods began to dip and rise. One moment I would be vulnerable and sensitive, the next aggressive and distracted. (I assumed the progesterone caused the former and the testosterone the latter.) Of course, there was a chance I'd been assigned to the control group and the symptoms were purely psychosomatic.

As the months progressed, my moods got worse. I made an appointment with the hospital to have the progesterone implant removed.The doctor said it didn't sound like the effect of the hormones. What I was describing was clinical depression. He put me on Prozac. Suspicious, I went to my GP for a second opinion.

'I agree with the diagnosis,' he said. 'You have a text-book case of clinical depression.' He went on to say that it's possible that the hormones triggered the depression.

The Prozac worked. I continued taking the pills until the progesterone implant was removed a couple of months later. Although the depression hasn't returned, 75 per cent of sufferers have a relapse within 10 years. I live in constant fear of my black dog returning.

I'm not angry with the doctors who ran the trial - I believe they acted in good faith - however, I'm surprised that no one thought to ask the volunteers about their mental health before allowing them to sign the consent form. What I know now is that hypochondriacs are 10 times more likely to suffer from depression than your average person. If I'd known that then, I wouldn't have signed.

In spite of what happened to me, I haven't lost faith in the process of clinical trials. If you want new drugs, you have to accept that they need to be tested on humans. Whenever you take a drug, you can't be 100 per cent sure it won't kill you. We can't live in a risk-free society, but we need to continually strive for a better understanding of what the risks are. Being human, we sometimes fall short.

· Clint Witchalls is writing a book about his experiences on the trial. It will be published by Rowohlt next year