Results:
We analyzed 1,168,378 pregnancies in the UPDB and identified 32,823 cases (2.8%) of PA over the years 1939 – 2020 after excluding for missing data or ineligible pedigrees. Relatives included in the primary analysis were 313,823 FDRs, 559,116 SDRs, and 1,242,936 TDRs (Fig. 1). 16,542 (50.4%) of these cases were identified by the term “antepartum bleeding” without a known diagnosis of placenta previa. The majority of cases were identified based on birth certificate data (83.7%), with the remaining cases identified from medical records (13.4%), fetal death certificates (2.3%) and maternal death certificates (0.5%). Individuals affected by PA were more likely to smoke (8.6% vs. 6.4%,p -value <0.001), use drugs (1.2% vs. 0.5%,p -value <0.001), have pre-gestational diabetes (0.4% vs. 0.3%, p -value <0.001), hypertension (7.5% vs. 5.6%, p -value <0.001), or a uterine anomaly (0.2% vs. 0.1%, p -value <0.001), experience PROM (4.3% vs. 1.3%, p -value <0.001), have less than a high school education (8.7% vs. 6.1%, p -value <0.001) and were more likely to be Hispanic (9.5% vs. 8.2%, p -value <0.001) and less likely to be White (95.1% vs. 96.0%,p -value <0.001). See Table 1 for full details.
Table 2 represents results from the primary PA familial risk analysis. PA appears to be heritable, with a statistically significant risk transmitted to FDRs and SDRs. FDRs inherit an increased odds of PA with an aOR 1.18 (95% CI: 1.12 – 1.23, p <0.001). SDRs also inherit an increased odds of PA with an aOR 1.09 (95% CI: 1.06 – 1.13, p <0.001). TDRs do not appear to inherit a significant PA risk with an aOR 1.01 (95% CI: 0.99 – 1.03,p =0.428). When stratified by number of pregnancies affected by PA, FDRs of cases with two or more PAs inherit a more significant odds of PA with aOR 1.38 (95% CI: 1.17 – 1.63, p <0.001). The measures of association are also higher for SDRs and TDRs when their family member was affected by two or more PAs, with the estimation for SDRs nearing statistical significance (Fig. 2). Through a sensitivity analysis in which cases of PA obtained by the code “antepartum bleeding” were removed, the risk inherited by FDRs increased to an aOR 1.20 (95% CI: 1.12 – 1.27, p <0.001), while the risk for SDRs was similar at an aOR 1.09 (95% CI: 1.05 – 1.14, p<0.001). The association for TDRs did not reach statistical significance (Table 3). All relative-driven familial risk analyses were adjusted for by available demographic variables for relatives: age, year of pregnancy, race, ethnicity, education level, and parity.
For the ego-driven subgroup analysis, we analyzed 1,168,378 pregnancies in the UPDB and identified 53,072 cases (4.5%) of PA matched to 159,124 controls over the years 1939 – 2020 after excluding for missing data (Fig. S1). The majority of cases were identified based on birth certificate data (82.2%), with the remaining cases identified from medical records (14.9%), fetal death certificates (2.5%) and maternal death certificates (0.4%). Demographic differences between cases and controls were similar to the relative-driven method. See Table S1 for full details. Table 4 depicts clinical risk factors with estimated adjusted odds ratios associated with PA. Of these, PROM was most highly associated with PA with an OR 2.18 (95% CI: 1.80 – 2.64, p<0.001). The presence of a uterine anomaly was also highly associated with PA with an OR 2.17 (95% CI: 1.18 – 3.97,p =0.004).
Table 5 represents the results of the ego-driven familial risk analysis adjusting for demographic and potentially confounding clinical variables. If an individual has a first-degree family history of PA, their odds of developing PA is estimated at an aOR 1.16 (95% CI: 1.03 – 1.31, p =0.014). Using this alternative method of analysis, the odds estimated for an individual with second- and third-degree relatives affected by PA is not statistically significant.