While hepatitis remains the greatest risk for recipients of banked blood, HIV infection has made more urgent the search for an alternative to donated blood. Artificial blood may soon be available, and a crucial step in its development was research on DNA by the Canadian scientist Michael Smith, who was last week awarded a share of the 1993 Nobel Prize for Chemistry.
The function of blood given in a transfusion is to maintain fluid pressure, and, above all, to carry oxygen to the tissues. The red cells are the oxygen transporters and thus usually the most crucial constituent of transfused blood.
Red cells carry on their surface the blood-type antigens that can lead to rejection if the wrong kind of blood is used; these cells have a brief shelf life, so that banked blood becomes useless after five weeks. And any remaining white cells and plasma may be be infected.
In the Sixties, a partial solution was found in perfluorocarbons (PFCs), synthetic substances with a high oxygen affinity. Mice submerged in PFCs were able to breathe the liquid and survive.
High oxygen affinity, however, is a mixed blessing. A blood substitute may be excellent at picking up oxygen in the lungs, but incapable of releasing it in the tissues. Human blood can achieve the best of both worlds, so recent research has focused on the prime ingredient of our blood - the intricate and extraordinary protein, haemoglobin.
One red cell contains 280 million haemoglobin molecules. Pure haemoglobin carries no information that might cause rejection and would seem an ideal candidate for artificial blood; but when it was tried in experiments 50 years ago, it caused acute kidney failure.
The problem is haemoglobin's natural instability when removed from its cell: it tends to split into two amino chains small enough to slip through the glomerular membrane in the kidney, clogging it up on the way to the bladder.
Free haemoglobin was also found to be reluctant to release its oxygen to the tissues. It emerged that each red blood cell had a crucial extra component, a molecule called DPG that encouraged the haems to unload their cargo; the trick, it seemed, was to reduce the oxygen affinity of the haemoglobin molecule while increasing its stability.
Ten years ago Dr Kiyoshi Nagai and fellow researchers at the Medical Research Council's Laboratory of Molecular Biology in Cambridge succeeded in fermenting human haemoglobin in bacterial cells. They had introduced the haemoglobin gene into the DNA of the bacterium E. coli, and 'tricked' it into producing large amounts of the protein.
Once they could make haemoglobin from scratch, the MRC scientists set about altering its properties. 'We joined two genes of the sub-units, and this solved the problem of molecules splitting,' Dr Nagai said. 'We also introduced a mutation, found in natural abnormal haemoglobin and known to reduce oxygen affinity, so that we could adjust it to normal levels.' The process stemmed directly from pioneering work on DNA by Professor Michael Smith of the University of British Columbia, Canada. 'We could not have done it without that work,' Dr Nagai said. The laboratory had overcome the last obstacles to creating artificial blood.
The MRC's partner, the American company Somatogen, is seeking approval for its product from the US Food and Drug Administration, and has begun clinical tests. Healthy volunteers have been given up to a pint of the artificial haemoglobin, known as rHb1.1. So far no toxic effects have been reported.
Other companies are adopting different approaches. Hemosol, of Canada, extracts haemoglobin from old red cells, then chemically bonds the molecules in twos and threes. The bonds reduce oxygen affinity to a manageable level by preventing the molecule from contorting in its usual way. Their product Hemolink is being tested on dogs.
The US company DNX, of Princeton, has a genetically engineered herd of pigs producing human haemoglobin, and Biopure of Boston uses cows' blood.
The transfusion market is worth more than dollars 5bn (pounds 3.3bn) in the developed world. It is estimated that this market could double if artificial blood were available.
Doctors are now seeing other uses for haemoglobin-based substitutes. Perhaps the best news could be for sufferers of sickle-cell anaemia. This painful and potentially lethal genetic disorder, which affects black people, results from defective haemoglobin distorting red blood cells into a sickle shape that clogs the blood vessels. With the tissues starved of oxygen, a patient suffers bouts of severe pain.
Haemotologists treating sufferers are hampered by the fact that, if they introduce more blood, thus increasing oxygen supply, they increase viscosity and the potential for dangerous blockages. Haemoglobin, about 300 million times as small as a red cell, can penetrate these clots and stop, or even reverse, a sickle-cell crisis.
Artificial blood could also be a superb agent for the delivery of drugs, many of which are more effective in the presence of oxygen. This is excellent news for cancer chemotherapists, whose treatments will not only be more powerful, but also will be carried with ample oxygen to the very heart of tumours, where red cells cannot reach.
None of these advances is likely to happen before 1997. FDA approval is a lengthy process, and more tests must be completed before licences are granted. In Britain, the artificial haemoglobin must be approved by the Drug Licensing Authority. This could happen within a year of it being approved in America.
The manufacturer will have to be able to compete with donated blood on price. It may irritate blood donors to hear that what they give freely soon after acquires a market price: about pounds 60 a litre in Britain and much more in America. If the artificial product does not compete, says one researcher, it will be just another amazing discovery with no practical purpose.