A multiple marker model to predict pregnancy viability when progesterone is indeterminate

Area of Science:

• Obstetrics and Gynecology
• Reproductive Endocrinology
• Clinical Diagnostics

Background:

• Early pregnancy viability assessment is crucial for symptomatic women.
• Differentiating viable from nonviable pregnancies often requires multiple visits and tests.
• Accurate prediction at a single visit can improve patient management and reduce anxiety.

Purpose of the Study:

• To develop and validate a predictive model for pregnancy viability.
• To assess the diagnostic accuracy of single and multiple biomarkers.
• To identify key predictors for differentiating viable from nonviable first-trimester pregnancies.

Main Methods:

• Prospective cohort study of 256 symptomatic first-trimester women.
• Collection of clinical data, serum biomarkers (including progesterone and human chorionic gonadotropin), and ultrasound findings.
• Analysis using receiver operator characteristic curves to evaluate predictor accuracy.

Main Results:

• Progesterone demonstrated high accuracy in predicting viability, especially at extreme values (AUC=0.99).
• A multiple marker model incorporating progesterone, hCG, ultrasound, and symptoms achieved 90% accuracy (AUC=0.90) in the 'grey zone'.
• The model effectively differentiated viable from nonviable pregnancies in symptomatic women.

Conclusions:

• A multiple marker model provides accurate prediction of pregnancy viability in symptomatic women.
• This model can aid in timely diagnosis and management decisions during the first trimester.
• Integrating multiple biomarkers enhances diagnostic performance beyond single marker assessment.