Singleton pregnancies conceived with infertility treatments and the risk of neonatal and infant mortality.

OBJECTIVES

To examine the risks of neonatal and infant mortality in relation to infertility treatment and to quantify the extent to which preterm delivery mediates this relationship.

DESIGN

Cross-sectional study.

SETTING

United States, 2015-2018.

PATIENT(S): A total of 14,961,207 pregnancies resulting in a singleton live birth.

INTERVENTION(S): Any infertility treatment, including assisted reproductive technology and fertility-enhancing drugs.

MAIN OUTCOME MEASURE(S): Neonatal (<28 days) mortality. The effect measure, risk ratio (RR), and 95% confidence interval (CI) were derived from log-linear Poisson models. A causal mediation analysis of the relationship between infertility treatment and mortality associated with preterm delivery (<37 weeks) was performed. The effects of exposure misclassification and unmeasured confounding biases were assessed.

RESULT(S): Any infertility treatment was documented in 1.3% (n = 198,986) of pregnancies. Infertility treatment was associated with a 51% increased risk of neonatal mortality (RR 1.51, 95% CI 1.39-1.64), with a slightly higher risk for early neonatal mortality (RR 1.57, 95% CI 1.43-1.73) than late neonatal mortality (RR 1.33, 95% CI 1.11-1.58). These risks were similar for pregnancies conceived through assisted reproductive technology and fertility-enhancing drugs. The mediation analysis showed that 72% (95% CI 59-85) of the total effect of infertility treatment on neonatal mortality was mediated through preterm delivery. In a sensitivity analysis, following corrections for exposure misclassification and unmeasured confounding biases, these risks were higher for early, but not for late, neonatal mortality.

CONCLUSION(S): Pregnancies conceived with infertility treatment are associated with increased neonatal mortality, and this association is largely mediated through preterm delivery. However, given the substantial underreporting of infertility treatment, these associations must be cautiously interpreted.