How did a collection of case reports lead to an untested drug being administered all over the world to people with serious bleeding, asks Ian Roberts
IN 1999, a leading medical journal published a remarkable story of survival against the odds. A 19-year-old Israeli soldier was admitted to hospital with a gunshot wound to his chest. The bullet had ripped through a major blood vessel causing torrential internal bleeding. Surgical intervention had failed and he was close to death. But then, 鈥渋n a desperate attempt to control the bleeding鈥�, he was given two intravenous doses of a drug called NovoSeven. Minutes later the bleeding stopped, allowing surgeons to repair the blood vessel. He survived ().
Such stories can be enormously persuasive. Even though most doctors know they are an unreliable source of information about drug efficacy, a compelling medical story can burn itself onto a doctor鈥檚 memory. Stories weave a simple yarn of causation between events, imposing order and banishing uncertainty. But in medicine, uncertainty is a reality that doctors have to work with.
Advertisement
NovoSeven, or recombinant factor VIIa, is a blood clotting agent used to treat haemophilia. It is not licensed for use in trauma patients. Yet this enormously expensive drug is now used in trauma rooms around the world, as well as by the US military to treat battlefield casualties.
How did this happen? To obtain a licence for the use of NovoSeven in trauma patients, manufacturer Novo Nordisk would have to carry out randomised controlled trials and present the results to drug regulators. The regulators insist on such trials because they are the best method we have of determining whether a treatment works.
A few small trials have been done, but as yet there is scant evidence that NovoSeven works in traumatic bleeding. A recent of all the evidence, prepared by the Cochrane Collaboration, concludes that there is no reliable evidence that NovoSeven is effective in traumatic bleeding.
Nevertheless, stories like that of the Israeli soldier continue to circulate. Bearing in mind that about 600,000 people bleed to death in hospitals every year, it is no surprise that other doctors have been inspired to take similar desperate measures. Doctors are legally entitled to use drugs in such 鈥渙ff label鈥� ways if they think it will help.
When the measures appear to work, doctors often tell the story as a case report in a medical journal 鈥� regardless of the fact that some of the patients would have survived anyway. In contrast, stories of patients who were given NovoSeven but died rarely see the light of day. Case reports have come to dominate the literature on NovoSeven, vastly outnumbering reports of randomised controlled trials.
鈥淲hen desperate measures work, doctors often tell the story as a case report in a journal鈥�
Some of these case reports are written by doctors who have a financial relationship with Novo Nordisk. For example, they may have been paid by the company as expert consultants.
Drug companies know that the credibility of any story depends on the authority and narrative ability of the storyteller, so they go to great lengths to find doctors who are willing to help them spread the word. One way they do this is to bring medical experts together to prepare guidelines on how to manage the condition for which the drug might be used.
For example, in 2005, a group of doctors convened to develop guidelines on the management of severe traumatic bleeding. Novo Nordisk paid for travel, hotel accommodation, meeting facilities, honoraria and the preparation of the guidelines. One of the recommendations in the guidelines was that NovoSeven should be considered if bleeding persists after standard treatment ().
By 2006, NovoSeven was being used in civilian trauma care around the world and also by both the to treat soldiers wounded in battle in Iraq and Afghanistan.
Then, in June 2008, announced that it had prematurely stopped its trial of NovoSeven for the treatment of bleeding in severe trauma. According to the company, this was because an interim analysis predicted 鈥渁 low likelihood of obtaining a positive trial outcome鈥�. The results of the trial have yet to be published in a peer-reviewed journal, though the company has posted a .
The synopsis shows that of 262 patients given NovoSeven, 12.2 per cent died. In contrast, 11.1 per cent of the 280 patients given a placebo died. This trial provides no new evidence that the drug is effective. The synopsis also noted that there was more unwanted clotting among people given NovoSeven. Unwanted clotting might increase the risk of heart attacks, stroke, clots in the lungs and thrombosis.
Until large randomised controlled trials have been completed, we cannot know whether NovoSeven does more good than harm in bleeding trauma patients. However, thanks to just such a trial funded by the UK government, of which I was chief investigator, we now have a safe and effective drug for severe bleeding in trauma: a cheap generic medicine called tranexamic acid ().
The British military has stopped using NovoSeven, and patience is wearing thin in the US. Earlier this year, The Baltimore Sun that federal investigators are exploring the US army鈥檚 use of NovoSeven. About the same time, at Novo Nordisk鈥檚 request, the US Food and Drug Administration added a warning to the label pointing out the risk of clotting when it is administered outside its licensed uses.
The moral of this particular medical story is clear. In the absence of evidence from randomised controlled trials we should remain sceptical about drug efficacy. Medical stories may be compelling, but they do not always give us the full picture.
