Children from better-off families in rural areas are at higher risk of developing cancer, probably because their isolation makes them more vulnerable to infection, scientists say.

A 25-year search for the causes of childhood cancer has found rates are up to 30 per cent higher in better-off rural counties of Britain than in poorer urban districts and tend to occur in clusters that would not be expected by chance.

The finding from the biggest study of childhood cancers in the world, based on more than 30,000 cases in Britain between 1969 and 1993, runs counter to experience with almost all other diseases, which are commoner in poorer populations.

It supports the theory that childhood leukaemia and other childhood cancers may be caused by viruses, and raises the hope that a vaccine against childhood cancer might be developed, the scientists said.

The report, by the Committee on Medical Aspects of Radiation in the Environment (Comare), in effect ruled out radiation from nuclear power stations and similar sources as a major cause of childhood cancers, although radiation and other environmental factors such as pollution could still play a small role, it said.

Professor Alex Elliott, the chairman, said it was the most important report in the committee's 20-year history. "If you believe theories that development of childhood cancer involves multiple steps and one of them is caused by a virus, if you can identify that virus it may be possible to develop a vaccine against it and break that causative chain," he added. There are about 1,500 cases of cancer in children under 15 each year, including 450 with leukaemia, which are commonest in social classes one and two. Cure rates have improved sharply and eight of 10 affected children now survive.

Childhood leukaemia, and possibly other childhood cancers, are thought to be caused by a "double whammy", a genetic defect in the womb, followed by a "second hit" that precipitates the disease.

Children of wealthier families kept apart from others in their early months in hygienic surroundings with little exposure to bacteria, may fail to develop strong immune systems necessary to ward off infections later in childhood, which may trigger cancer in the vulnerable.

Professor Elliott said: "If you're wealthy, you tend to live in a big house with more land around it, and have contact with fewer people. So it's theoretically believable that, if there is a viral component, you have less chance of coming across that particular virus." Childhood cancer was highest in areas with the lowest overcrowding and lowest where overcrowding was worst, lending more support to the viral theory, Professor Elliott said.

The study had demonstrated for the first time that, like childhood leukaemia, other childhood cancers, including bone tumours and soft tissue sarcomas, occurred in clusters. "This suggests their causative processes may be similar," the professor said.

The committee was set up in 1985 after a Yorkshire TV documentary, Windscale: The Nuclear Laundry to investigate claims of an excess of leukaemia cases around the nuclear reprocessing plant, which is now known as Sellafield .

No link was found but the committee recommended that the clusters at Sellafield and Dounreay should be investigated further.