Objective: The accuracy of serum beta-human chorionic gonadotropin levels as cutoff values for estimating gestational age was studied.
Material and methods: A database was created using information from previously performed research studies, which allowed entry of women both less than and greater than 49 days' gestation, involving medical abortion. Serum beta-human chorionic gonadotropin determinations and vaginal ultrasonography were performed in all studies before treatment. A total of 574 women had data available for analysis. A receiver operating characteristic curve was created to evaluate the predictive value of potential beta-human chorionic gonadotropin cutoff values for 42 and 49 days' gestation.
Results: Appropriate serum beta-human chorionic gonadotropin cutoff values for 42 and 49 days' gestation were 23,745 mIU/mL (sensitivity, 96%; specificity, 91%; positive predictive value, 68%; negative predictive value, 99%) and 71,160 mIU/mL (sensitivity, 95%; specificity, 62%; positive predictive value, 76%; negative predictive value, 91%), respectively. Under 42 days' gestation, the serum beta-human chorionic gonadotropin-time relationship appears to be linear, with a greater diversity of individual values after 42 days.
Conclusion: Serum beta-human chorionic gonadotropin values can be used with reasonable accuracy to screen for a gestational age up to 49 days' gestation.