ABSTRACT
OBJECTIVE: To validate a combined algorithm for early prediction of pre-eclampsia (PE) in the Brazilian population. STUDY DESIGN: This is an unplanned secondary analysis of a cohort study. Consecutive singleton pregnancies undergoing first-trimester screening for PE involving examination of maternal characteristics, medical history, and biophysical markers were considered eligible. Women were classified as low-or high-risk using a cutoff of 1/200, but the individual risk was not used to dictate management, as aspirin prophylaxis was given to women based solely on clinical risk factors. Receiver-operating characteristics (ROC) curves for PE, preterm PE(PE < 37) and early 34(PE < 34) were constructed and detection rates(DR) and false-positive rates(FPR) were calculated, adjusting for the effect of aspirin. Propensity score analysis was utilized to account for possible confounding by indication. MAIN OUTCOME MEASURES: Screening performance and PE rates. RESULTS: Among 1695 women, 323(19.1%) were classified as high-risk for PE and 1372(80.9%) were considered low-risk. Aspirin use was registered in 62(3.7%) in the high-risk group and 33(1.9%) in the low-risk group. There were 164(9.7%) women who developed PE, including 41(2.4%) with PE < 37 and 18(1.1%) PE < 34.Subgroups with aspirin had higher incidence of PE, suggest confounding by indication. The algorithm had an AUC of 0.87, DR of 72% for PE < 34; an AUC of 0.8, DR of 59% for PE < 37, both with FPR of 18%. Accounting for effect of aspirin, we observed an improvement in DR of PE < 37 to 67%. CONCLUSION: Using combined predictive algorithm for preterm PE prediction is feasible in clinical practice in low/middle-income countries. Aspirin use needs to be accounted for when evaluating the performance of screening.
