Revisiting the association between maternal and offspring preterm birth using a sibling design

BACKGROUND: Previous studies have reported an intergenerational association between maternal and offspring preterm birth (PTB) but the nature of the association remains unclear. We assessed the association between maternal and offspring preterm birth using a quasi-experimental sibling design and distinguishing between preterm birth types. METHODS: We conducted a retrospective intergenerational cohort study of 39,573 women born singleton in Manitoba, Canada (1980–2002) who gave birth to 79,198 singleton infants (1995–2016). To account for familial confounding we defined a subcohort of 1033 sisters with discordant PTB status who subsequently gave birth and compared offspring PTB rates between 2499 differentially exposed cousins using log-binomial fixed-effects generalized estimating equation models. PTB was defined as a delivery < 37 gestation weeks, divided into spontaneous and provider-initiated. RESULTS: In the population cohort, mothers born preterm were more likely to give birth preterm [Adjusted Relative Risk (ARR): 1.39; 95% Confidence Interval (CI): 1.25, 1.54] and very preterm birth [ARR: 1.76; 95% CI: 1.29, 2.41]. However, in the siblings cohort, the intergenerational association was not apparent among births to sisters with discordant PTB status [ARR: 1.02; 95% CI: 0.77, 1.34 for preterm birth and ARR: 0.88; 95% CI: 0.38, 2.02 for very preterm birth]. Mothers born at term with a sister born preterm had a similarly elevated risk of delivering a preterm infant (10%) than their preterm sisters. Intergenerational patterns were observed for spontaneous PTB but not for provider-initiated PTB. CONCLUSIONS: Our findings suggest that it is not the fact of having been born preterm that puts women at higher risk of delivering preterm, but the fact of having been born to a mother who ever delivered preterm. Consideration of a female family history of PTB may better identify women at higher risk of preterm delivery than relying on maternal preterm birth status alone. Further research may benefit from distinguishing preterm birth types.