Improved sensitivity and specificity of a single measurement of serum progesterone over serial quantitative beta-human chorionic gonadotrophin in screening for ectopic pregnancy

Abstract

The sensitivity and specificity of a single serum progesterone measurement was compared against two beta-human chorionic gonadotrophin (HCG) measurements 48 h apart in screening for abnormal pregnancy, i.e. ectopic pregnancy, completed or incomplete abortion. Of 1120 patients in the first trimester presenting with a positive urinary pregnancy test, 116/1120 (10.4%) had an ectopic pregnancy, 755/1120 (67.4%) had ultrasonographically confirmed intra-uterine pregnancies, and 249/1120 (22.2%) had abnormal intra-uterine pregnancies documented as complete, incomplete or missed abortions. Of the ectopic pregnancies, 113/116 (97.4%) had a serum progesterone level less than 25 ng/ml while 516/755 (68.3%) viable intra-uterine pregnancies had a serum progesterone level greater than or equal to ng/ml. Of the 1120 patients screened, 402 (35.9%) had both a serum progesterone and two HCG measurements and were eligible for inclusion in this study. Setting a cut-off of 25 ng/ml, the sensitivity and specificity of a single serum progesterone measurement was then compared against two serial HCG measurements, utilizing receiver operating characteristic curves. This analysis demonstrated that a single serum progesterone measurement was significantly more sensitive (P less than 0.05) than two HCG measurements in screening for an abnormal pregnancy. In some patients, a single serum progesterone makes possible the diagnosis of ectopic pregnancy 2 days earlier than two HCG determinations because a second blood sample was not required. We conclude that a single serum progesterone measurement should be added to serial HCG determinations as a standard diagnostic screening test for ectopic pregnancy.