Routine screening of pregnant women for risk factors for preeclampsia (PE) is recommended. However, the detection rate using only maternal history is poor. In a study published in the mid-90’s, use of bilateral uterine artery Doppler (UtA Doppler) velocimetry for screening had high sensitivity, but poor specificity. Data from recent studies suggest that a better approach for PE screening includes integration of UtA Doppler with both maternal history and biochemical serum markers that have been associated with placental and endothelial dysfunction. However, the additive value of placental biomarkers measured during the first trimester to predict which women will subsequently develop PE has not been established.
This prospective cohort study was designed to estimate the ability of a model based on integration of maternal history, serum biomarkers, and UtA Doppler in the first trimester to predict hypertensive disorders of pregnancy. Information was obtained on maternal history, UtA Doppler, and for 4 serum biomarkers, PAPP-A, placental growth factor (PlGF), PP-13 and free β human chorionic gonadotropin (β-hCG). Two PE predictive models were developed for outcomes of gestational hypertension, PE, early-onset PE, and late-onset PE. With the first model, multivariate analysis included all maternal variables and markers—those that were statistically significant as well as those that were not. The second model was performed following a stepwise backward logistic regression and included only statistically significant factors (P < 0.05). For both models, sensitivity was calculated for fixed false-positive rates (FPRs) of 5%, 10%, 15%, and 20% from the receiver operating characteristic curves.
Of the 2118 women included in the analysis, 46 (2.2%) developed gestational hypertension, and 25 (1.2%) developed PE, including 12 (0.6%) with early-onset PE and 13 (0.6%) with late-onset PE. The detection rates for early-onset PE using serum PlGF, free β-hCG, and chronic hypertension as predictors of early-onset PE for a fixed FPR of 10% and 5% were 75% and 67%, respectively. Using the same PE predictive model, the combination of UtA pulsatility index (PI) with PlGF and chronic hypertension reached a sensitivity of 60% but with a 20% FPR.
These findings suggest that integration of maternal characteristics and first-trimester maternal serum biomarkers (free β-hCG and PlGF) may provide a useful screening test for early-onset PE. Although UtA PI is statistically significant in the overall PE model, it does not improve the detection rate of early-onset PE.