The British Nuclear Fuels plc facility at Sellafield performs a range of nuclear-related activities. The site has been in operation since 1950 and has, in general, employed a stable work force, many of whom have accumulated relatively high occupational exposures to ionizing radiation. This paper compares the physical dosimetry with two biological end points for evaluating radiation exposure: fluorescence in situ hybridization with whole-chromosome painting probes to quantify stable chromosome aberrations (translocations and insertions), and glycophorin A (GPA) analysis of variant erythrocytes. For the cytogenetic analyses, 81 workers were evaluated in five dose categories, including 23 with minimal radiation exposure (< or = 50 mSv) and 58 with exposures ranging from 173 to 1108 mSv, all but 3 being > 500 mSv. In a univariate analysis, the mean stable chromosome aberration frequencies showed a significant increase with dose category (P = 0.032), and with cumulative dose when dose is treated as a continuous variable (P = 0.015). The slope of the dose response for stable aberrations is 0.79 +/- 0.22 aberrations per 100 cells per sievert (adjusted for smoking status), which is less than that observed among atomic bomb survivors, and suggests a dose and dose-rate effectiveness factor for chronic exposure of about 6. Analyses of the data for GPA N/O and N/N variants from 36 workers revealed no correlation with dose. Neither was there a correlation between the frequencies of N/O GPA variants and stable aberrations, although a weak negative association was observed between N/N variant frequency and stable aberrations (r = -0.38, P = 0.05). These results provide clear evidence for the accumulation of stable aberrations under conditions of chronic occupational exposure to ionizing radiation and show that stable chromosome aberrations are a more sensitive indicator for chronic radiation exposure than GPA variants. In comparison with human studies of brief exposure, chronic low-dose exposures appear substantially less effective for producing somatic effects as reflected by stable chromosome aberrations.

Three somatic mutation assays were evaluated in men exposed to low-dose, whole-body, ionizing radiation. Blood samples were obtained between 1992 and 1999 from 625 Russian Chernobyl cleanup workers and 182 Russian controls. The assays were chromosome translocations in lymphocytes detected by FISH, hypoxanthine phosphoribosyltransferase (HPRT) mutant frequency in lymphocytes by cloning, and flow cytometic assay for glycophorin A (GPA) variant frequency of both deletion (N/Ø) and recombination (N/N) events detected in erythrocytes. Over 30 exposure and lifestyle covariates were available from questionnaires. Among the covariates evaluated, some increased (e.g. age, smoking) and others decreased (e.g. date of sample) biomarker responses at a magnitude comparable to Chernobyl exposure. When adjusted for covariates, exposure at Chernobyl was a statistically significant factor for translocation frequency (increase of 30%, 95% CI of 10%-53%, P = 0.002) and HPRT mutant frequency (increase of 41%, 95% CI of 19%-66%, P < 0.001), but not for either GPA assay. The estimated average dose for the cleanup workers based on the average increase in translocations was 9.5 cGy. Translocation analysis is the preferred biomarker for low-dose radiation dosimetry given its sensitivity, relatively few covariates, and dose-response data. Based on this estimated dose, the risk of exposure-related cancer is expected to be low.

No abstract