Perinatal health in a cohort of children conceived after assisted reproduction in the UK: a population-based record-linkage study

Objective To compare the risk of hospitalisation for conditions originating in the perinatal period between children conceived via assisted reproductive technology and those that are naturally conceived, differentiating by treatment type.Study design, setting and participants Population-based record-linkage study of children born after assisted reproduction in the UK between 2002 and 2009 (n=44 618), their naturally conceived siblings (n=8462) and matched naturally conceived population (n=89 072) controls linked to their hospital inpatient records up to 31 March 2016.Primary and secondary outcome measures Robust estimates of the overall and cause-specific risk of hospital admission for adverse perinatal events and the comparison of outcomes by type of treatment.Results Over the study period, 17 132 (38.40%) children conceived via assisted reproduction and 30 306 (34.02%) and 1738 (20.54%) naturally conceived population and sibling controls, respectively, were admitted to the hospital for severe perinatal events. Compared with the population controls, singletons (Risk ratio (95% CI 1.30 (1.26, 1.34))) and twins (1.01 (0.99, 1.03)) conceived via assisted reproduction exhibited a higher risk of hospitalisation for any adverse perinatal event. However, no such increase was observed in the within-sibling analysis (0.97 (0.84, 1.12)). Similar patterns were seen for diagnoses related to length of gestation and fetal growth (vs population controls: 1.37 (1.29, 1.46); vs siblings: 1.17 (0.86, 1.60)); birth trauma (vs population controls: 1.23 (1.04, 1.44); vs siblings: 0.78 (0.47, 1.30)); respiratory and cardiovascular disorders (vs population controls: 1.28 (1.20, 1.38); vs siblings: 0.72 (0.53, 0.98)); infections (vs population controls: 1.30 (1.06, 1.59); vs siblings: 0,68 (0.24, 1.90)) and several other conditions. Associations were similar when comparing in vitro fertilisation to intracytoplasmic sperm injection and were higher when comparing fresh to frozen embryo transfers.Conclusion Children conceived via assisted reproduction showed modest increases in the risk of hospitalisations for severe perinatal events when compared with population controls, although these findings were attenuated in the sibling analyses. The imprecision of within-sibling analyses highlights the need for larger studies to explore potential causal effects.