Advancing Sexual and Reproductive Health and Rights
Family Planning Perspectives
Volume 32, Number 2, March/April 2000

Infants Fathered by Men Who Are Exposed to Radiation in the Workplace Have an Elevated Risk of Being Stillborn

A man's workplace exposure to ionizing radiation, a known mutagen, may increase the risk that an infant he fathers will be stillborn, according to a study based on workers at a British nuclear reprocessing plant.1 The risk of stillbirth increased by 24% for each 100 millisieverts (mSv) of radiation that a man had been exposed to in his lifetime; it rose even further if the infant had a congenital anomaly, particularly a neural-tube defect.

The study is based on data linking records of all singleton births (live births and stillbirths) in 1950-1989 in the county of Cumbria to employee records from a nuclear plant there. Men's records included information that was routinely gathered for regulatory and radiological protection purposes on occupational exposure to radiation. The researchers conducted both a cohort analysis and a case-control study to examine the effects of a worker's radiation exposure on the risk of stillbirth. They focused on this outcome because the causes of stillbirth include congenital anomalies, which in turn may result from genetic factors.

Workers' lifetime exposure to ionizing radiation up to the time that their partners conceived varied widely (0.1-911 mSv), but generally was low. The median lifetime exposure was 30.1 mSv; only 1% of men had had a total dose of more than 350 mSv. The median exposure in the 90 days before conception was 1.7 mSv.

Over the study period, partners of radiation workers had 9,078 live births and 130 stillbirths. The resulting stillbirth rate (14.1 per 1,000 live births) was essentially the same as the rate among women in the county whose partners did not work at the nuclear plant (14.7 per 1,000). However, when the investigators standardized the data for year, social class and birth order, they found significantly more stillbirths among the offspring of radiation workers (130) than expected (115).2

For the cohort analysis, the investigators used logistic regression techniques to assess the risk of stillbirth, controlling for birth order and year of birth, and for father's social class and age. Results indicated that for every 100 mSv of a man's lifetime exposure to ionizing radiation, the odds that his infant would be stillborn increased by 24%; for every 10 mSv of exposure in the 90 days before conception, the odds of stillbirth rose by 86%. To assess whether the stillborn infant whose father had the highest lifetime exposure to radiation affected the results, the researchers recalculated the odds, excluding that birth; in this analysis, the increase in risk dropped to 17% and was no longer statistically significant. Additional calculations to examine the relationship between a man's total dose of radiation and the odds of stillbirth revealed that the risk increased gradually up to about 425 mSv of lifetime exposure and then rose steeply.

Beginning in 1961, information was available on the cause of death among stillborn infants. Analyses using these data revealed that each 100 mSv increased the risk of stillbirth resulting from congenital anomalies by 43%; the increased risk of stillbirth attributable to neural-tube defects was even higher--69%. The stillborn infant whose father had been exposed to the highest dose of radiation did not have a congenital anomaly and so did not influence these results.

In the case-control study, which used more accurate dose measures than the cohort study, the investigators used conditional logistic regression analysis to assess the odds of stillbirth versus live birth among partners of radiation workers. Results for lifetime exposure to radiation were similar to those from the cohort analysis: The risk of stillbirth increased by 30% for every 100 mSv of exposure, but when the stillbirth associated with the highest lifetime dose of radiation was excluded from the analysis, the risk was reduced and was no longer statistically significant. This analysis, however, showed no significant effect of exposure in the 90 days before conception; according to the researchers, the differences in results for short-term exposure may relate to differences in how exposure was measured in the cohort and case-control studies.

Finally, the researchers used a variety of approaches to estimate the number of stillbirths that could be attributed to the father's exposure to radiation. Results varied according to the estimation procedure, but suggested that of the 130 stillbirths occurring among workers' partners throughout the study period, 0-32 might (with 95% likelihood) have been caused by the effects of radiation.

The analysts observe that their study was "the largest and most comprehensive investigation of transgenerational effects in any workforce occupationally exposed to ionising radiation of the few to use prospective dose measurements for each individual." Furthermore, they note that "statistical analysis showed no evidence of unmeasured factors influencing stillbirth risk." Nevertheless, they acknowledge the possibility that some unmeasured risk factor confounded their results.

A commentary accompanying the study stresses the need to keep the possible risk of stillbirth associated with exposure to radiation "in perspective."3 Particularly, the author points out that with "no obvious mechanism" to explain this association, it cannot be assumed to be causal. Furthermore, she notes that both stillbirth rates and workers' exposure to radiation declined sharply over the years studied, and if the association is causal, the increased risk is lower than that associated with above-average maternal age.--D. Hollander


1. Parker L et al., Stillbirths among offspring of male radiation workers at Sellafield nuclear reprocessing plant, Lancet, 1999, 354(9180):1407-1414.

2. Parker L et al., reply to letters to the editor, Lancet, 2000, 355(9202):493.

3. Inskip H, Stillbirth and paternal preconceptional radiation exposure, commentary, Lancet, 1999, 354(9180): 1400-1401.