Objective To evaluate the specificity of the ultrasound diagnosis of fetal anomalies. Pregnancies which proceeded to termination and ultrasound-diagnosed fetal anomalies which were not offered termination were considered.

Design Prospective, region-wide study over three and a half years.

Setting Cases were identified through 25 ultrasound departments representing the 15 districts in the Yorkshire Region.

Subjects Pregnant women with an ultrasound-diagnosed fetal anomaly.

Main outcome measures Information obtained from the ultrasound report was compared with the outcomes determined by cytogenetics, postmortem or paediatric examination.

Results Of 2261 pregnancies with an ultrasound-diagnosed fetal anomaly 369, (16%) were terminated and 357 (97%) were followed by postmortem examination. Ultrasound findings exactly matched those of the postmortem or were accompanied by additional anomalies in 325 cases (91%). In 32 cases ultrasound findings were not confirmed by postmortem, but in 30 of these the decision to offer termination remained justified because the correct diagnosis was judged equally or more serious. Two (0.5%) were terminated for an anomaly which subsequently proved less severe than predicted on ultrasound. Ultrasound significantly over-or under-diagnosed a major fetal anomaly in 27 of the 1139 (2.4%) cases in which an anomaly was detected, but the pregnancy was not terminated.

Conclusion Termination of pregnancy was based on the correct prognosis in over 99.5 % of cases. This does not obviate the need for pathological examination of the fetus which changed or refined the diagnosis in 35 % of cases.