Britain's blood transfusion service faces a crisis after new research revealed that techniques used to filter the human form of BSE - or "mad cow disease" - are far less effective than was previously thought.
That means that there is a danger of infection to those who receive blood each year from people who may have eaten BSE-infected food in the 1980s, when the disease was at its height.
Every year roughly three million units of blood are donated and used in the UK. Buying bloodfrom other countries would be expensive - but the National Blood Service (NBS), which runs the scheme in England and Wales, is faced with a serious problem because there is no test to show whether blood is infected with vCJD, the human form of BSE.
Although only 147 people have so far died from the disease in Britain, the recent revelation that one person had caught it from a blood transfusion shows that it could spread. Previously, only people with a particular genetic make-up, or genotype, restricted to about one-third of the UK, had caught it. But the genotype of the dead person is found in nearly half the population.
Now scientists believe anyone could be at risk. Professor James Ironside, from the National CJD Surveillance Unit in Edinburgh, said the transfusion finding "has major implications for future estimations of vCJD cases in the UK".
He added: "It's absolutely possible that there may be a new epidemic. I'm not in the business of scaremongering, but quite clearly the idea that this problem is on the way out is not the case at all."
Papers published today in the medical journal The Lancet separately detail how an unnamed person who died of a ruptured artery was subsequently found to be incubating "variant Creutzfeldt-Jakob Disease" (vCJD), the human form of BSE. Five years previously the patient, who was described only as "elderly", had received a blood transfusion from a person who later developed vCJD.
Another paper, by a team at the University of Maryland, suggests that removing white blood cells to reduce the risk of vCJD transmission - a method used in the UK since November 1999 - is less effective than previously believed. Tests found it only reduced the risk of infection by 40 per cent.
The NBS said it could not comment on the implications of this research. Professor Ironside said it was significant that the elderly patient did not die from vCJD, and the infection would never have been detected without a post-mortem examination. But he warned that had that patient given blood, the disease could have been passed on easily.
The possibility of vCJD passing into the human blood supply by such "asymptomatic" carriers has long worried scientists, who advised that while the risk from blood was theoretical, it was also serious.
Fears over vCJD contamination led the NBS to recall blood plasma products from 26 English distribution sites in 1997. Cases of vCJD peaked in 2000, when 28 deaths were reported. Since then the trend has been generally downward, with 17 cases in 2002 and 18 last year.