Defining a Strategy for Laboratory Evaluation with Expectant Management of Preeclampsia

Abstract

Objective The objective of this study was to compare the frequency and timing of laboratory abnormalities and evaluate optimal laboratory testing strategies in women with preeclampsia (PE) undergoing expectant management. Study Design Retrospective cohort study of women with inpatient expectant management of PE at ≥23 weeks at a tertiary center from 2015 to 2018 was conducted. Women ineligible for expectant management or with less than two laboratory sets (platelets, aspartate aminotransferase, and serum creatinine) before the decision to deliver were excluded. Women were categorized as per the American College of Obstetricians and Gynecologists' definitions by initial diagnosis: PE without severe features, superimposed preeclampsia (SiPE) without severe features, and their forms with severe features. The frequency and timing of laboratory abnormalities were compared across the four PE categories. Kaplan–Meier curves modeled time to a laboratory abnormality (event) with censoring for delivery and were compared using log-rank tests. Logistic regression analysis modeled the development of a laboratory abnormality as a function of testing time intervals (days) for each PE type. Receiver operating characteristic curves and areas under the curve (AUC) were calculated; optimal cut points were determined using the Liu method. Results Among 636 women who met inclusion criteria, laboratory abnormalities were uncommon (6.3%). The median time to a laboratory abnormality among all women was ≤10 days, time being shortest in women with PE with severe features. Time to laboratory abnormality development did not differ significantly between the four PE groups (p = 0.36). Laboratory assessment intervals were most predictive for PE and SiPE with severe features (AUC = 0.87, AUC = 0.72). Optimal cutoffs were every 4 days for PE without severe features, 2 days for PE with severe features, 8 days for SiPE without severe features, and 3 days for SiPE with severe features. Conclusion Most laboratory abnormalities in PE occur earlier and more frequently in those with severe features. Individual phenotypes should undergo serial evaluation based on this risk stratification. Key Points