OBJECTIVES: To determine the accuracy of ultrasound in the diagnosis of a tubal pregnancy in the absence of an obvious extra-uterine embryo.
METHODS: The systematic review was registered with PROSPERO (registration number: CRD42012003410). We searched MEDLINE, Embase and the Cochrane Library for citations (from database inception to July 2014). Studies were selected in a two-stage process and data extracted by two reviewers. Accuracy measures were calculated for each feature (i.e. empty uterus, pseudosac, adnexal mass and free fluid) alone and in various combinations. Individual study estimates were plotted in summary ROC and forest plots for examination of heterogeneity. Study quality was assessed using the QUADAS-2 checklist.
RESULTS: Thirty-one studies (5858 women) were selected from 19,959 citations. Following meta-analysis, an empty uterus was found to predict an ectopic pregnancy with a sensitivity of 32.4% (95% CI 19.7%-48.3%) and specificity of 93.3% (95% CI 84.6%-97.3%). The corresponding values for a pseudosac, adnexal mass and free fluid are: 3.3% (95% CI 1.9%-5.4%) and 95.0% (95%CI 92.2%-97.1%); 66.2% (95%CI 51.6%-78.2%) and 91.3% (95%CI 83.9%-95.5%); and 41.2% (95%CI 28.2%-55.5%) and 90.6% (95%CI 84.0%-94.7%) respectively.
CONCLUSION: When an obvious extra-uterine pregnancy is not present, the commonly used ultrasonographic features have poor sensitivity for identifying a tubal pregnancy. However, they have good specificity. We can therefore infer ultrasound features are more useful for 'ruling in' a tubal pregnancy than 'ruling out' one. The findings were somewhat limited by the quality of some of the studies included and the heterogeneity in the index test and reference standard.