Prediction of residual trophoblastic tissue in first-trimester abortions and low levels of human chorionic gonadotropin β-subunit

The clinical data and the histologic report of curettage specimens of 174 patients with first-trimester abortion and human chorionic gonadotropin β-subunit < 500 mIU/ml were reviewed. In only 22 (12.6%) were chorionic villi histologically reported. A multivariate analysis showed that increased log human chorionic gonadotropin β-subunit levels within the range of 20 to 500 mIU/ml, intrauterine tissue demonstrated by ultrasonography, higher gestational age, and active bleeding were significant predictive variables for the presence of villi in curettage specimens. A mathematical formula and a clinical scoring system for the prediction of residual villi in these cases were devised and were applied to the data of our study group. The sensitivity, false-negative rate, and specificity of the scoring system were 90.9%, 2.2%, and 59.2%, respectively; however, the false-positive rate was 75.6%. By use of this scoring system in patients with first-trimester abortion and low human chorionic gonadotropin β-subunit level, the presence of villi in curettage specimens can thus be predicted with a high sensitivity and as many as 60% of curettage procedures might be prevented.