History and Physical Examination to Estimate the Risk of Ectopic Pregnancy: Validation of a Clinical Prediction Model

Abstract

Study objective: To prospectively validate a clinical prediction model for ectopic pregnancy (EP). Methods: Prospective cohort with 14-month derivation and 12-month validation phases. All hemodynamically stable, first-trimester patients with abdominal pain or vaginal bleeding who presented to a military teaching hospital emergency department underwent follow-up until an outcome of intrauterine pregnancy (IUP) or EP was established. Patients were separated into the high-risk group, defined as having either peritoneal signs or definite cervical motion tenderness; intermediate-risk group, defined as the presence of pain or tenderness, other than midline cramping, plus absence of fetal heart tones, and absence of tissue visible at the cervical os; and low-risk group (neither high- nor intermediate-risk) using recursive partitioning. Results: Summarizing both phases, 915 patients had 845 (93%) IUPs and 70 (7.6%) EPs, with 18 (1.9%) lost to follow-up. The clinical prediction model classified 75 (8.2%) into the high-risk group (sensitivity 31%, 95% confidence interval [CI] 21% to 44%; specificity 94%, 95% CI 92% to 95%); and 644 (70%) in the intermediate-risk group (sensitivity 98%, 95% CI 89% to 100%; specificity 25%, 95% CI 22% to 29%). The remaining 196 (21%) patients who met neither high-risk nor intermediate-risk criteria were classified into the low-risk group. On the basis of EP prevalence of 7.7%, the risk of EP was less than 1% (95% CI 0% to 3%) for the low-risk group, 7% (95% CI 5% to 10%) for the intermediate-risk group, and 29% (95% CI 19% to 41%) for the high-risk group. Conclusion: This clinical prediction model is useful for estimating the risk of EP in first-trimester patients, particularly when ancillary testing is equivocal or not readily available. [Buckley RG, King KJ, Disney JD, Gorman JD, Klausen JH: History and physical examination to estimate the risk of ectopic pregnancy: Validation of a clinical prediction model. Ann Emerg Med November 1999;34:589-594.]