Study question Is protracted exposure to low doses of ionising radiation associated with an increased risk of solid cancer? Methods In this cohort study, 308 297 workers in the nuclear industry from France, the United Kingdom, and the United States with detailed monitoring data for external exposure to ionising radiation were linked to death registries. Excess relative rate per Gy of radiation dose for mortality from cancer was estimated. Follow-up encompassed 8.2 million person years. Of 66 632 known deaths by the end of follow-up, 17 957 were due to solid cancers. Study answer and limitations Results suggest a linear increase in the rate of cancer with increasing radiation exposure. The average cumulative colon dose estimated among exposed workers was 20.9 mGy (median 4.1 mGy). The estimated rate of mortality from all cancers excluding leukaemia increased with cumulative dose by 48% per Gy (90% confidence interval 20% to 79%), lagged by 10 years. Similar associations were seen for mortality from all solid cancers (47% (18% to 79%)), and within each country. The estimated association over the dose range of 0-100 mGy was similar in magnitude to that obtained over the entire dose range but less precise. Smoking and occupational asbestos exposure are potential confounders; however, exclusion of deaths from lung cancer and pleural cancer did not affect the estimated association. Despite substantial efforts to characterise the performance of the radiation dosimeters used, the possibility of measurement error remains. What this study adds The study provides a direct estimate of the association between protracted low dose exposure to ionising radiation and solid cancer mortality. Although high dose rate exposures are thought to be more dangerous than low dose rate exposures, the risk per unit of radiation dose for cancer among radiation workers was similar to estimates derived from studies of Japanese atomic bomb survivors. Quantifying the cancer risks associated with protracted radiation exposures can help strengthen the foundation for radiation protection standards. Funding, competing interests, data sharing Support from the US Centers for Disease Control and Prevention; Ministry of Health, Labour and Welfare of Japan; Institut de Radioprotection et de Sûreté Nucléaire; AREVA; Electricité de France; US National Institute for Occupational Safety and Health; US Department of Energy; and Public Health England. Data are maintained and kept at the International Agency for Research on Cancer.

Background:The nested case-control design can be a very efficient approach to an epidemiological investigation. In order to obtain unbiased estimates of relative risk, controls should be selected by incidence density sampling, which involves matching each case to a sample of those who are at risk at the time of case occurrence.Methods:This paper presents a simple computer program for incidence density sampling. This program was evaluated using data derived from a cohort study of mortality among workers employed in the nuclear weapons industry. Controls were selected for cases via incidence density sampling; an estimate of the exposure-mortality association was obtained via conditional logistic regression. After 100 iterations of this procedure, the average effect estimate was compared to the risk estimate obtained via proportional hazards regression. The same methods were used to evaluate a program for incidence density sampling that was proposed previously by Pearce in 1989.⁵Results:Relative risk estimates obtained from nested case-control analyses conducted using the incidence density sampling program reported in this paper are unbiased. In contrast, the program for incidence density sampling proposed by Pearce⁵tended to produce biased relative risk estimates; the magnitude of bias increased with increasing numbers of controls selected per case.Conclusions:The computer program described in this paper offers a simple approach to incidence density sampling for nested case-control analyses with exact matching on attained age and appropriate enumeration of the pool of eligible controls for each case. This method overcomes problems of bias inherent in a previously proposed program for incidence density sampling.

The relative excess risk due to interaction (RERI) provides a useful metric of departure from additivity of effects on a relative risk scale. In this paper, the authors show that RERI is identical to the product term in a linear odds ratio or a linear relative risk model. SAS and STATA codes are provided for fitting a linear odds ratio model that directly parameterizes RERI. In addition, this paper presents a method for obtaining likelihood-based 95% confidence bound estimates for RERI. The authors show that likelihood-based confidence intervals may differ substantially from the asymptotic confidence interval estimates advocated by previous authors. The approach presented in this paper should facilitate estimation of RERI and associated likelihood-based confidence bounds, by using standard statistical packages.

Epidemiologic studies routinely suffer from bias due to exposure measurement error. In this paper, the authors examine the effect of measurement error when the exposure variable of interest is constrained by a lower limit. This is an important consideration, since often in epidemiologic studies an exposure variable is constrained by a lower limit such as zero or a nonzero detection limit. In this paper, attenuation of exposure-disease associations is defined within the framework of a classical model of uncorrelated additive error. Then, the special case of nonlinearity due to the effect of a lower threshold is examined. A general model is developed to characterize the effect of random measurement error when there is a lower threshold for recorded values. Findings are illustrated under the assumption that the true exposure follows the lognormal and gamma distributions. The authors show that the direction and magnitude of bias in estimated exposure-response associations depends on the population distribution of the exposure, the magnitude of the recording threshold, the value assigned to below-threshold measurement results, and the variance in the measured exposure due to random measurement error.