The objectives of this study were to determine the optimal cutoff value (CV) and utility of a single serum beta human chorionic gonadotropin hormone (HCG) level in assessing the likelihood of ectopic pregnancy (ECP). A retrospective chart review was performed at an urban county hospital. The optimal CV was determined by comparing all available patients diagnosed with ECP and patients diagnosed with threatened abortion (TA) in the Emergency Department (ED) who subsequently delivered a baby at the same hospital, The utility of the test was assessed in a group of all known patients who presented to the ED with lower abdominal or pelvic pain or vaginal bleeding beta een 6 and 13 weeks estimated gestational age (EGA), There were 212 ECPs with measured HCGs, Only nine presented after 13 weeks EGA, The mean HCG value was 5,378 mIU/mL, There was no significant correlation between HCG and EGA. There were 132 TAs with measured HCGs who subsequently delivered. Only four TAs presented before 6 weeks EGA. Regarding the TA patients who presented between 6 and 13 weeks EGA, there were 81 with a mean HCG of 83,810 mIU/mL. Between 6 and 13 weeks EGA, HCG was not significantly correlated with EGA, The entire ECP group was compared with the TA group with 6-13 week EGA. A receiver operating characteristic curve with nine HCG cutoff values was constructed. Clinically the optimal CV was 40K. At this level, a test for the condition "absence of ECP" hall a specificity of 99%, and only two of 212 ECPs in our group would test false-positive with an HCG value over 40K. When the test was applied to a group of 175 women who presented with pain or bleeding between 6 and 13 weeks EGA, the sensitivity and positive predictive value for an intrauterine pregnancy were 24% and 99.9%, respectively. However, if this test was applied to all patients regardless of clinical findings, and then an ultrasound (U/S) algorithm was applied to the patients who had an HCG of less than 40K, it would predictably increase the rate of ECPs falsely identified as intrauterine pregnancies (IUPs) by 1%, regardless of the false-positive rate of the U/S algorithm itself. Given the potential morbidity and mortality of undiagnosed ectopic pregnancy, ultimately neither this test nor any other that is less than 100% specific for IUP can be recommended instead of an initial U/S evaluation for patients who present at risk for ECP. (C) 1999 Elsevier Science Inc.
