Review covers 136 countries in US, Canada, UK, France, Germany, Japan and Spain.

Leukaemia rates in children and young people are elevated near nuclear facilities, but no clear explanation exists to explain the rise, according to a research review published in the July issue of European Journal of Cancer Care.

Researchers at the Medical University of South Carolina carried out a sophisticated meta-analysis of 17 research papers covering 136 nuclear sites in the UK, Canada, France, the USA, Germany, Japan and Spain.

They found that death rates for children up to the age of nine were elevated by between five and 24 per cent, depending on their proximity to nuclear facilities, and by two to 18 per cent in children and young people up to the age of 25.

Incidence rates were increased by 14 to 21 per cent in zero to nine year olds and seven to ten percent in zero to 25 year-olds.

"Childhood leukaemia is a rare disease and nuclear sites are commonly found in rural areas, which means that sample sizes tend to be small" says lead author Dr Peter J Baker.

"The advantage of carrying out a meta-analysis is that it enables us to draw together a number of studies that have employed common methods and draw wider conclusions."

Eight separate analyses were performed - including unadjusted, random and fixed effect models - and the figures they produced showed considerable consistency.

But the authors point out that dose-response studies they looked at - which describe how an organism is affected by different levels of exposure - did not show excess rates near nuclear facilities.

"Several difficulties arise when conducting dose-response studies in an epidemiological setting as they rely on a wide range of factors that are often hard to quantify" explains Dr Baker. "It is also possible that there are environmental issues involved that we don't yet understand.

"If the amount of exposure were too low to cause the excess risk, we would expect leukaemia rates to remain consistent before and after the start-up of a nuclear facility. However, our meta-analysis, consistently showed elevated illness and death rates for children and young people living near nuclear facilities."

The research review looked at studies carried out between 1984 and 1999, focusing on research that provided statistics for individual sites on children and young people aged from zero to 25.

Four studies covered the UK, with a further three covering just Scotland. Three covered France, two looked at Canada and there was one study each from the USA, Japan, Spain, the former East Germany and the former West Germany.

"Although our meta-analysis found consistently elevated rates of leukaemia near nuclear facilities, it is important to note that there are still many questions to be answered, not least about why these rates increase" concludes Dr Baker.

"Several hypotheses have been proposed to explain the excess of childhood leukaemia in the vicinity of nuclear facilities, including environmental exposure and parental exposure. Professor Kinlen from Oxford University has also put forward a hypothesis that viral transmission, caused by mixing populations in a new rural location, could be responsible.

"It is clear that further research is needed into this important subject."