Assisted reproductive technology and the risk of birth defects mediated by multifetal pregnancy: evidence from the China Birth Cohort Study.

Abstract

Both assisted reproductive technology and multifetal pregnancy have been reported to be associated with an increased risk of birth defects. However, it is unclear whether multifetal pregnancy is a mediator that increases the risk of birth defects in assisted reproductive technology-conceived pregnancies. To investigate the risk of birth defects associated with assisted reproductive technology, and to quantify the mediating effect of multifetal pregnancy between assisted reproductive technology and birth defects. From November 2017 to August 2021, 132,386 pregnant women from 50 study sites in 17 provinces of China were monitored from their first trimester until delivery (or miscarriage/termination). Birth defects were detected in both live births and stillbirths. Directed acyclic graph was drawn to understand the causal relationship between assisted reproductive technology and birth defects, and inverse probability of treatment weighting was used to adjust for confounders. Then, log-binomial generalized estimating equation was established in the inverse probability weighted cohort to estimate the adjusted relative risk of birth defects associated with assisted reproductive technology compared to natural conception, while addressing the cluster effect of study sites. Mediation analysis was performed to assess the mediating effect of multifetal pregnancy between assisted reproductive technology and birth defects. Number needed to harm was calculated for multifetal pregnancy vs singleton pregnancy among assisted reproductive technology-conceived pregnancies. A total of 126,035 naturally conceived and 5124 assisted reproductive technology-conceived pregnancies were included in the analysis. The incidence of birth defects in the assisted reproductive technology conception group (4.2%) was significantly higher than that in the natural conception group (2.5%), with an adjusted relative risk of 1.48 (95% confidence interval, 1.24-1.78). Multifetal pregnancies were also more common in the assisted reproductive technology conception group (21.4%) than the natural conception group (1.1%). Mediation through multifetal pregnancy accounted for 25% of the risk of birth defects associated with assisted reproductive technology. Notably, if assisted reproductive technology-conceived pregnant women could have singleton pregnancy rather than multifetal pregnancy, one case of birth defects could be potentially avoided for every 52 (95% confidence interval, 29-233) pregnancies. Multifetal pregnancy mediated the effect of assisted reproductive technology on birth defects. Therefore, when assisted reproductive technology treatment is needed, single embryo transfer should be considered under clinically feasible conditions for the prevention of birth defects.