The diagnostic accuracy of fetal cardiac anomaly scans to postnatal cardiac defects

Abstract

Purpose: To evaluate the diagnostic accuracy of routine anomaly ultrasound, performed between 18 to 24 weeks of gestation, in detecting fetal cardiac defects. The antenatal cardiac defects were compared to the gold standard, which was defined as the postnatal echocardiography and the post-mortem findings from autopsy reports. Antenatal detection of congenital heart defects (CHDs) optimises antenatal care and improves perinatal and infant outcomes. This novel study evaluates the sensitivity and specificity of the national anomaly screening programme.

Objectives: The objectives included identifying: the number of anomaly scans performed and the number of births in a set timeframe, the number and type of cardiac anomalies detected antenatally on ultrasound and those detected postnatally. The maternal and fetal risk factors for CHDs were also evaluated. Through this data the commonest type of cardiac anomaly in this study was identified and the incidence of CHDs and individual defects was established.

Methodology: A hospital based retrospective, observational, quantitative and comparative research method was adopted. The cardiac defects detected on anomaly ultrasound scan, from the initiation of the anomaly screening programme in January 2016 up to August 2018 (with births occurring by January 2019), were compared to the postnatal echoes performed by the Paediatric cardiologist or to the post-mortem findings derived from autopsy results. Data was analysed using both descriptive and inferential statistics.

Results: A total of 29 cases of CHD were detected antenatally, from which 28 were confirmed postnatally with complete agreement in 20 cases. The sensitivity of anomaly scans in the detection of CHDs was observed to be 30.43% and the specificity of 99.98%. The positive predictive value (PPV) was observed at 96.55% and the negative predictive value (NPV) at 99.24%. The commonest diagnosed CHD was a ventricular septal defect (VSD) while the incidence of CHDs was 18.15 per 1000 live births. Conclusion: A moderate sensitivity was obtained with an excellent specificity, PPV and NPV.