An empirical approach was used to identify patterns of placental vascular lesions and their relation to preterm delivery (PTD) in the POUCH (Pregnancy Outcomes and Community Health) Study, which enrolled 3019 pregnant women (16-27 weeks' gestation) from 52 clinics among 5 Michigan cities. To date, 1052 sub-cohort placentas (238 preterm, 814 term) have been fully assessed. Pathology examinations were performed and items for exploratory factor analysis (EFA) were compiled based on a pathology diagnostic coding instrument. Six placental tissue `Factors' were derived from the EFA. They represented vascular lesions in the fetal compartment (Factors 1, 2) and maternal compartment (Factors 3 to 6), which closely paralleled the organizational structure in the pathology-based coding instrument. Factor-based scores were calculated by summing individual item values. Multinomial logistic models were performed with the outcome of PTD divided by weeks' gestation (<35, 35-36) and by clinical subtype: spontaneous labor, premature rupture of membranes (PPROM), and medically indicated (MI). Term delivery served as the referent category. Each factor-based score was dichotomized as explanatory variables, using the values that were closest to 75th percentile of the distribution as cutoff points. Weights were used to reflect the sub-cohort sampling scheme. In the model that included all six Factors, high factor scores were relevant for the following: Factor 5 (maternal vessel obstruction) was associated with MI deliveries at <35 weeks (adjusted odds ratio (AOR) = 3.8, 95% CI 1.7, 8.7); Factor 4 (evidence of bleeding) was associated with spontaneous and MI deliveries <35 weeks (AOR =2.4, 95% CI 1.0, 5.8; AOR =2.9, 95% CI 1.2, 7.0); Factor 3 (maternal vasculopathy) was associated with MI and PPROM <35 weeks (AOR = 4.4, 95% CI 2.0, 10; AOR= 2.9, 95% CI 1.2, 6.9).). These results suggest that EFA may be useful in identifying multiple PTD-related vascular pathways for further study.
Read
