Subclassification of indeterminate pelvic ultrasonograms: stratifying the risk of ectopic pregnancy.

Abstract

To determine whether the subclassification of indeterminate ultrasound readings can identify patients who are at high, intermediate, or low risk for ectopic pregnancy. A retrospective review was made of consecutive ED patients presenting to an urban teaching hospital from August 1991 to December 1994 with abdominal pain and/or vaginal bleeding and a positive beta-hCG. Patients who had transvaginal ultrasonograms obtained during the ED visit that were read as indeterminate (no extrauterine findings of ectopic pregnancy and no intrauterine fetal pole or yolk sac) were eligible. Ultrasonograms were subclassified into 5 groups (empty uterus, nonspecific intrauterine fluid, echogenic debris within endometrial cavity, abnormal sac, normal sac) based on predetermined criteria. Patients were excluded if the final diagnosis could not be definitively determined. 248 patients were identified. 20 patients were excluded because a final diagnosis could not be determined. Patients with an empty uterus [25/94 = 27% (95% CI 18-36%)] had the highest frequency of ectopic pregnancy. Patients with nonspecific intrauterine fluid collections [4/30 = 13% (95% CI 4-31%)] had the next highest frequency of ectopic pregnancy. Patients with intrauterine echogenic debris [2/39 = 5% (95% CI 1-11%)], abnormal sacs [1/36 = 3% (95% CI 1-9%)], or normal-appearing sacs [0/29 = 0% (95% CI 0-8%)] had low frequencies of ectopic pregnancy. Subclassification of indeterminate ultrasound readings identifies patients at high, intermediate, or low risk for ectopic pregnancy and should improve the diagnostic accuracy of ultrasonography in patients at risk for ectopic pregnancy.