BACKGROUND: Data on the safety of inadvertent rubella vaccination in pregnancy is important for rubella vaccination programs aimed at preventing congenital rubella syndrome. METHODS: The association between monovalent rubella or combination vaccinations in or shortly before pregnancy and potential harm to the foetus was examined by conducting a systematic review and meta-analysis using fixed effect methods and simulation. RESULTS: Four cohort studies of inadvertently vaccinated and unvaccinated women were found, 15 cohorts of pregnant women who were rubella susceptible at time of inadvertent vaccination and 9 cohort studies with no information on susceptibility and case series. No case of vaccine associated congenital rubella syndrome (CRS) was identified. Cohort studies with an unvaccinated comparison group were limited in number and size, and based on these only a theoretical additional risk of 6 or more cases of CRS per 1000 vaccinated women (0% observed, upper 95% CI 0.6%) could be excluded. Based on cohorts of vaccinated rubella susceptible pregnant women a maximum theoretical risk of 1 CRS case in 1008 vaccinated women (0% observed, upper 95% CI 0.099%) was estimated. Asymptomatic rubella vaccine virus infection of the neonate was also noted (fixed effects estimate of risk overall 1.74%, 95% CI 1.21, 2.28). CONCLUSION: There is no evidence that CRS is caused by rubella-containing vaccines but transplacental vaccine virus infection can occur. CRS is effectively prevented by vaccination, thus the risk/benefit balance is unequivocally in favour of vaccination. The data confirm previous recommendations that inadvertent vaccination during pregnancy is not an indication for termination of pregnancy.