La controverse sur les effets des faibles doses de rayonnements ionisants et la relation linéaire sans seuil

The ICRP and the BEIR VII reports recommend a linear no threshold (LNT) relationship for the estimation of cancer excess risk induced by ionising radiations (IR), but the 2005 report of Medicine and Science French Academies concludes that it leads to overestimate of risk for low and very low doses. The bases of LNT are challenged by recent biological and animal experimental studies which show that the defence against IR involves the cell microenvironment and the immunologic system. The defence mechanisms against low doses are different and comparatively more effective than for high doses. Cell death is predominant against low doses. DNA repairing is activated against high doses, in order to preserve tissue functions. These mechanisms provide for multicellular organisms an effective and low cost defence system. The differences between low and high doses defence mechanisms are obvious for alpha emitters which show several greys threshold effects. These differences result in an impairment of epidemiological studies which, for statistical power purpose, amalgamate high and low doses exposure data, since it would imply that cancer IR induction and defence mechanisms are similar in both cases. Low IR dose risk estimates should rely on specific epidemiological studies restricted to low dose exposures and taking precisely into account potential confounding factors. The preliminary synthesis of cohort studies for which low dose data (< 100 mSv) were available show no significant risk excess, neither for solid cancer nor for leukemias.