Causal inference of multiple interrelated exposures: Contributions of ischemic placental disease to preterm delivery

Ischemic placental disease (IPD) involves three obstetrical complications - preeclampsia, placental abruption, and fetal growth restriction - and is implicated in up to 55% of early deliveries. Despite substantial overlap in their etiology and epidemiologic risk profiles, existing studies have assessed their effects independently. Using data from the Consortium on Safe Labor (2002-2008; n = 203, 990), we examined the composite and separate effects of the IPD conditions on spontaneous and indicated preterm delivery (sPTD and iPTD). We systematically specified joint IPD probabilities, corresponding to combinations of hypothetical stochastic and deterministic interventions. Eliminating all IPD reduced sPTD (RR = 0.93; 95% CI = 0.92, 0.94) and iPTD (RR = 0.66; 95% CI = 0.64, 0.67) from observed levels. Eliminating preeclampsia alone reduced sPTD (RR = 0.94; 95% CI = 0.94, 0.95) and iPTD (RR = 0.72; 95% CI=0.71, 0.73). Partially reducing joint IPD co-occurrence among 58% of individuals who met criteria for high-risk, by magnitudes drawn from separate low-dose aspirin meta-analyses, yielded minimal reductions for sPTD (RR = 0.99; 95% CI=0.98, 0.99) and iPTD (RR = 0.89; 95%CI=0.88, 0.89). Preeclampsia drives IPD's association with both PTD subtypes; under the stated assumptions, hypothetically eliminating preeclampsia yielded the largest iPTD reduction.