Correlation between prenatal ultrasound and autopsy findings: a study of second-trimester abortions

Objective To evaluate discrepancies between sonographic and autopsy findings following termination of pregnancy (TOP) in the second trimester. Methods This retrospective report is based on 288 consecutive second-trimester abortions carried out because of fetal malformations diagnosed by ultrasound examination at a tertiary referral center. The correlation between the results from the ultrasound and autopsy examinations was evaluated. Results Autopsy was performed in 274 cases. In 160 of the 274 pregnancies (58.4%) there was full agreement between the two examination methods. Findings detected by autopsy (in addition to those leading to termination) were not observed by ultrasonograpby in 86 (31.4%) of the pregnancies; of the 64 malformations that occurred, 30 (46.9%) were judged as 'detectable'. In 27 (9.9%) pregnancies, observations made by ultrasound (in addition to those leading to termination) were not confirmed at autopsy. in one pregnancy, postmortem radiology examination - but not autopsy - confirmed the ultrasound observations. No pregnancies were terminated because of false positive ultrasound observations. The correlation between ultrasound and autopsy findings was evaluated by three investigators; the inter-rater agreement was high (kappa = 0.85). Conclusion Discrepancies between ultrasound and autopsy findings were observed in about 40% of the pregnancies. These discrepancies confirm the need for autopsy following TOP. Copyright (c) 2006 ISUOG. Published by John Wiley & Sons, Ltd.