Prediction of early pregnancy viability in the absence of an ultrasonically detectable embryo

Objective: To identify clinical, ultrasound and biochemical parameters that may allow prediction of pregnancy viability in women without a detectable embryo on ultrasound examination.

Methods: This was a prospective observational study of pregnant women with an ultrasound finding of a gestational sac measuring < 20 mm mean diameter without a visible embryo. Women's age, menstrual dates, clinical symptoms (pain and bleeding), mean gestational sac diameter and measurements of serum beta-human chorionic gonadotropin and progesterone were recorded in all cases. All women were managed expectantly until the pregnancy viability was established conclusively based on clinical and ultrasound findings. All parameters were tested by univariate analysis and then analyzed in a stepwise procedure to form a logistic regression model for predicting pregnancy viability.

Results: One hundred and eighteen (59%) women had a normal intrauterine pregnancy and 82 (41%) had a miscarriage. Stepwise analysis showed that three diagnostic parameters (maternal age, gestational sac diameter and serum progesterone) contributed significantly to the predictive power of the logistic model. With this model, at a cut-off value of 10% probability, the diagnosis of viable pregnancy was made with a sensitivity of 99.2% (95% CI, 95.8-99.97) and specificity of 70.7% (95% CI, 61.3-78.9).

Conclusion: The use of a logistic regression model allows prediction of pregnancy viability when an embryo cannot be visualized on ultrasound scan.