Childhood Leukemia and Natural Background Radiation
Gerry M Kendall1, Mark P Little2, Richard Wakeford3, Kathryn J Bunch1, Jon CH Miles4, Tim J Vincent1, Jill R Meara5 and Michael FG Murphy1
1Childhood Cancer Research Group, University of Oxford; 2Radiation Epidemiology Branch, NCI, DHHS, NIH, DCEG,
Bethesda, USA; 3Institute of Population Health, University of Manchester; 4Formerly Public Health England, CRCE,
Didcot; 5Public Health England, CRCEH, Didcot
Introduction
Epidemiological studies of natural background ionizing radiation and childhood cancer must be Huge to have sufficient power to detect the very small radiation effects predicted by standard models. Interview-based case-control studies with sufficient study subjects are impracticable and would also be liable to various important biases. Record-based studies, with cases and controls drawn from existing registers, can be potentially large enough and can avoid obvious biases.
However, they generally lack direct measurements of radiation levels in subjects’ homes, and other relevant variables.
Background (Phase I Study)
A published record-based case-control study from Great Britain involving 27 447 cases and 36 793 controls investigated links between childhood cancer risk and two components of naturally occurring radiation: radon and gamma rays (including the directly ionizing component of cosmic rays). It detected a statistically significant association between indoor gamma ray exposures and the risk of childhood leukemia; the Excess Relative Risk (ERR) per unit dose was compatible with that extrapolated from higher dose studies, in particular those of the Japanese Atomic Bomb Survivors, and also with a UK study of cancer in children and young adults following diagnostic Computerized tomographic imaging. For gamma rays, relative risks for other childhood cancers, and for radon, relative risks for both leukemia and other cancers, were above one but not close to statistical significance.
Phase II Study
A criticism of Phase I was that it used small area averages as measures of gamma ray doses, based on a set of 2283 measurements. The study in preparation will use estimates of individual doses rather than small area averages, based on model extrapolations from a larger measurement set (n = 10 199). It will also be substantially larger including provisionally 54 462 cases and 69 992 controls. This study should allow radiation-related risks from Phase I to be refined.
View Full Abstract
38