Plasma exchange treatment involves replacing the infected plasma with fresh plasma to clear the E. coli toxins from the body.

The technique requires that up to 80% of the patient's blood is cleansed in a machine similar to that used for kidney dialysis.

The incidence of E. coli 0157 infection is increasing in the UK, particularly in Scotland, where the rate of infection is among the highest in the world.

E. coli poisoning can lead to life-threatening complications. The most serious are haemolytic uraemic syndrome (HUS), which causes kidney failure, and thrombotic thrombocytopenic purpura (TTP), which can lead to kidney failure but also affects the brain and the heart.

Adults who develop HUS or TTP induced by E. coli tend to be elderly and almost nine out of ten die.

Lanarkshire outbreak

Dr Andrew Todd believes plasma exchange is the best treatment

A team of Scottish scientists used the plasma exchange technique to treat 16 adult patients who were infected during a major outbreak of E. coli poisoning in Lanarkshire in 1996. All the patients were suffering from either HUS or TTP, and had an average age of 71.

Eleven of the 16 patients who were treated with plasma exchange survived, a survival rate of 69%.

The Scottish team, who reported their findings in The Lancet medical journal, compared their findings to those of an outbreak of E. coli 0157 in Canada in 1985 in which 11 of 12 patients with either HUS or TTP died. None of these patients were treated with plasma exchange.

The doctors say a larger scale clinical trial is required to test the effect of plasma exchange, but would be extremely difficult to set up.

Instead, they suggest a national register of patients who develop HUS or TTP is set up to monitor the impact of different treatments.

Dr Andrew Todd, of the Department of Infectious Diseases at Monklands Hospital, Airdrie, is one of the doctors who treated patients with plasma exchange.

He said plasma exchange had been tried in the past on a number of individual cases of HUS, but its use in cases of E. coli poisoning was controversial.

"We feel that we were justified in using the technique. We did not have anyone who developed serious problems from the technique and the number who survived was significantly higher than the expected mortality rate in that age group."

It is estimated that five per cent of people infected with E. coli 0157 develop HUS or TTP.