Related Experiment Video
Updated: Jul 5, 2026
Establishment of an Embryo Implantation Model In Vitro
Published on: June 21, 2024
A multiple marker model to predict pregnancy viability when progesterone is indeterminate
Beth J Plante1, Jeffrey D Blume, Geralyn Lambert-Messerlian
1Department of Obstetrics and Gynecology, Women and Infants Hospital, Brown Medical School, USA. bethplante@gmail.com
Objective:
To develop a model to differentiate viable from nonviable pregnancies at a single first-trimester visit.
Study Design:
This prospective cohort study included 256 symptomatic women in the first trimester who presented to our urgent care unit in Providence, Rhode Island, between 2002 and 2004. Predictors of pregnancy viability were collected, including clinical information and serum samples for several biomarkers. Each predictor was evaluated alone and in conjunction with other variables using receiver operator characteristic curves. The cohort was separated into 2 subgroups based on whether the progesterone value was at the "extremes" (<5 ng/mL or >25 ng/mL) or in the "grey zone" (5-25 ng/mL).
Results:
Among single biomarkers, progesterone had the greatest diagnostic accuracy in predicting viability. Progesterone was highly accurate at the extremes (area under the curve [AUC] = 0.99) but less accurate in the grey zone (AUC = 0.71). A multiple marker model was developed to include progesterone for all patients, and human chorionic gonadotropin, ultrasound findings and symptoms for participants in the grey zone (AUC = 0.90).
Conclusion:
A multiple marker model predicted pregnancy viability in symptomatic women with overall accuracy of 90%.
