Diagnostic accuracy of ultrasound above and below the beta-hCG discriminatory zone.

Abstract

To evaluate clinical use and accuracy of transvaginal ultrasound for diagnosing intrauterine pregnancies, spontaneous miscarriages, and ectopic pregnancies in women who present with beta-hCG concentrations above or below an established discriminatory zone. Ultrasound diagnosis at presentation was compared with final clinical diagnosis in 333 consecutive pregnant women who presented to an emergency department with vaginal bleeding or abdominal pain. The sensitivity, specificity, predictive value, and overall diagnostic accuracy of ultrasound were calculated. Transvaginal ultrasound was nondiagnostic in 59 (17.7%) of 333 subjects and 43 (67.2%) of 64 subjects with beta-hCG levels below 1500 mIU/mL at presentation. Preliminary ultrasound diagnoses were reported significantly more frequently when presenting beta-hCG levels were above 1500 mIU/mL (253 [94.1%] of 269 subjects) compared with levels below 1500 mIU/mL (21 [32.8%] of 64 subjects: P < .001; relative risk (RR) 3.4 [95% confidence interval (CI) 2.23, 5.18]). The proportion of accurate preliminary ultrasound diagnoses was significantly higher in subjects who presented with beta-hCG levels above 1500 mIU/mL (227 [91.5%] of 248 subjects) compared with levels below 1500 mIU/mL (18 [28.6%] of 63 subjects: P < .001; RR 2.9 [95% CI 2.04, 4.15]). Sensitivity of transvaginal ultrasound diagnosis of intrauterine pregnancy, spontaneous miscarriage, and ectopic pregnancy in women who presented with beta-hCG levels below 1500 mIU/mL was 33.3%, 28.2%, and 25.0%, respectively. The sensitivity, predictive value, and accuracy of transvaginal ultrasound for diagnosing complications of early pregnancy were poor when beta-hCG levels were below the discriminatory zone at presentation. Ultrasound impressions should be correlated with beta-hCG concentrations.