OBJECTIVE: We sought to assess the usefulness of clinical, biophysical, and biochemical tests in the prediction of preeclampsia.
DATA SOURCES: The sources of data we used to conduct this review included the computerized databases MEDLINE (1966 to February 2003), EMBASE, Popline, CINAHL, and LILACS plus reference lists, conference proceedings, textbooks, and contact with experts.
METHODS OF STUDY SELECTION: All cohort or cross-sectional studies reporting data on the relationship between a predictive test that was performed during pregnancy and the development of preeclampsia were eligible for inclusion, whereas case–control studies were excluded. Eighty-seven (211,369 women) of 7,191 potentially relevant articles met inclusion criteria. We evaluated the methodologic quality for each included study.
TABULATION, INTEGRATION, AND RESULTS: Using a standardized protocol, one reviewer selected and extracted data on study characteristics, quality, and accuracy. Data abstracted from each study were arranged in 2 × 2 tables to construct receiver operating characteristics plots (sensitivity against 1 − specificity) and pooled to produce summary likelihood ratios for positive and negative tests results. Moderate predictive accuracy of anticardiolipin antibodies, the presence of bilateral diastolic notches during Doppler ultrasonography, and urinary kallikrein were found in women at low risk of developing preeclampsia. Nevertheless, because the pretest probability of preeclampsia with a positive result was but minimally increased, the clinical use of these tests is limited. Other ultrasonography characteristics and the measurement of fetal and placental peptides showed low predictive accuracy. In populations that were deemed at high risk for preeclampsia, the use of Doppler ultrasonography had low predictive accuracy. No definitive conclusions were possible in the case of many other tests, because the number of studies that met the minimal inclusion criteria was limited.
CONCLUSION: As of 2004, there is no clinically useful screening test to predict the development of preeclampsia. Further prospective, longitudinal studies are needed.
LEVEL OF EVIDENCE: III