PREPARE: A Stepped-Wedge Cluster-Randomized Trial to Evaluate Whether Risk Stratification Can Reduce Preterm Deliveries among Patients with Suspected or Confirmed Preterm Preeclampsia

BACKGROUND: Early delivery in preterm preeclampsia may reduce the risks for the patient, but consequences of prematurity may be substantial for the baby. This trial evaluated whether the implementation of a risk stratification model could safely reduce prematurity. METHODS: This was a stepped-wedge cluster-randomized trial in seven clusters. Patients presenting with suspected or confirmed preeclampsia between 20+0and 36+6gestational weeks were considered eligible. At the start of the trial, all centers were allocated in the preintervention phase, and patients enrolled in this phase were managed according to local treatment guidance. Subsequently, every 4 months, 1 randomly allocated cluster transitioned to the intervention. Patients enrolled in the intervention phase had sFlt-1 (soluble fms-like tyrosine kinase-1)/PlGF (placental growth factor) ratio and preeclampsia integrated estimate of risk assessments performed. If sFlt-1/PlGF ≤38 and preeclampsia integrated estimate of risk <10%, patients were considered low risk and clinicians received recommendations to defer delivery. If sFlt-1/PlGF >38 and preeclampsia integrated estimate of risk ≥10%, patients were considered not low risk, and clinicians received recommendations to increase surveillance. The primary outcome was the proportion of patients with preterm preeclampsia delivered prematurely out of total deliveries. RESULTS: Between March 25, 2017 and December 24, 2019, 586 and 563 patients were analyzed in the intervention and usual care groups, respectively. The event rate was 1.09% in the intervention group, and 1.37% in the usual care group. After prespecified adjustments for variation between and within clusters over time, the adjusted risk ratio was 1.45 ([95% CI, 1.04-2.02]; P=0.029), indicating a higher risk of preterm deliveries in the intervention group. Post hoc analysis including calculation of risk differences did not show evidence of statistical differences. Abnormal sFlt-1/PlGF was associated with a higher rate of identifying preeclampsia with severe features. CONCLUSIONS: The introduction of an intervention based on biomarkers and clinical factors for risk stratification did not lead to reductions in preterm deliveries. Further training on the interpretation of disease severity in preeclampsia and the development of additional risk stratification is needed before adoption into clinical practice. REGISTRATION: URL: https://www.clinicaltrials.gov; Unique identifier: NCT03073317.