We have carried out a review of the literature on the incidence and aetiology of bone tumours. There are a limited number of studies. One major set of analyses from Great Britain has found an unusual geographical pattern in the incidence of childhood osteosarcoma, but not Ewing’s sarcoma. This is interpreted as providing support for the involvement of a spatially varying environmental exposure for osteosarcoma. We have identified a small number of studies that specifically examine the possible role that fluoride or radium in drinking water may have on subsequent risk of developing a bone tumour. One study from Wisconsin found no evidence that radium in drinking water resulted in excess cases of osteosarcoma. However, two studies from Ontario, found that both increased incidence and increased mortality was associated with radium in drinking water. Two time trend analyses found no association between fluoride in drinking water and increased risk of osteosarcoma. However, one case-control study from the USA found an association between fluoride in drinking water and increased risk of osteosarcoma. There was a consistent association for males but not for females. There has been a lack of investigations examining the putative role that exposures in drinking water may have on risk of osteosarcoma in the UK. In some areas fluoride has been a natural constituent of the water supply. In other areas, fluoride has been added by the water supplier. Fluoridation of water supplies was first introduced in the Midlands and North East England during the 1960’s. The 2003 Water Act has extended fluoridation to all areas, with local policy determined by the health authority.

This new pilot study has two aims. The first aim is to examine geographical variation in the incidence of bone tumours in Great Britain. The second aim is to use the water supply areas to allocate fluoride levels and to analyse geographical variation in incidence in relation to fluoride levels. Three data sets will be mapped and statistically analysed: (i) 900 bone tumours in 0-49 year olds from the Northern and Yorkshire region diagnosed 1985-2005; (ii) 2198 bone tumours in 0-39 year olds from Great Britain diagnosed 1980-1994; and (iii) 1917 bone tumours in 12-24 year olds from Great Britain diagnosed 1979-1997. Analyses will be done for all cases together and also separately by sex, age-group and diagnostic group (osteosarcoma, Ewing’s sarcoma).

The first geographical analysis will require allocation of cases to census small areas. Corresponding population data will be obtained from the 1981, 1991 and 2001 UK censuses. The second geographical analysis will involve allocating a fluoride level to a water supply area. The fluoride may be naturally occurring or artificially added. The initial analyses will consider presence or absence of fluoride. The feasibility of obtaining more detailed fluoride measurements will be investigated. Risk estimates for bone sarcomas and Ewing’s sarcomas will be obtained and mapped. The investigators have extensive experience in this area of research. Risk estimates will be adjusted to allow for patterns that might occur due to deprivation and population density, since these may confuse the picture. Furthermore, the maps will be formally analysed for the presence of non-random patterns and incidence “hot-spots” using recent methodology. We will directly correlate and model the association between incidence of bone sarcomas and presence of fluoride in the water supply.

If it proves feasible to obtain and construct more accurate area-based data on level of fluoride these new measurements will be included in the exploratory statistical analyses. Findings from this study will be disseminated via presentations at scientific meetings and publications in peer-reviewed journals. We anticipate that this study will indicate the direction of future research concerning the aetiology of bone tumours.

Dr Richard McNally, Newcastle University