Accuracy of Fetal Diagnoses of Congenital Heart Diseases
How accurate is fetal cardiac diagnosis, and how does this impact neonatal care and long-term management?
This was a retrospective single-center French study over a 10-year period that included 1,258 neonates diagnosed in utero to have congenital heart defects (CHDs) and 189 fetal autopsies following fetal death or pregnancy termination, comparing pre- and postnatal diagnoses. Discordant diagnoses were classified into three groups: group 1–completely different CHD; group 2–partially different CHD with a negative or positive impact on neonatal care (unforseen or overcalled ductal dependence; unexpected need for, or an overestimation of the need for, urgent neonatal intervention); and group 3–partially different diagnosis with no impact on neonatal care. Impact on long-term care strategy for discordant pre- and postnatal diagnoses was deemed positive if a biventricular (biV) repair was possible, the Aristotle complexity score was lower, or less interventions were necessary, contrary to the prenatal impression. Long-term strategy was deemed to be a negative impact when CHD severity was underestimated, resulting in unexpected compassionate care, a biV repair was not possible, or the number of interventions or the Aristotle scores were higher than prenatally predicted.
The median age at fetal CHD diagnosis was approximately 26 weeks (range 14-40 weeks). There was a pre- and postnatal diagnosis discrepancy of about 29% amongst live births, with about 11% classified as a group 1 or 2 neonatal discordance, and about 7% exhibiting a positive or negative impact on long-term care management strategy. For the available fetal autopsies, approximately 23% of the pre- and postnatal discrepancies would have impacted management, but about 32% of diagnosis variances were deemed minor. The most common types of anatomical discordant diagnoses included bicuspid aortic valves, anomalies of the systemic or pulmonary venous return, ventricular septal defects (VSDs), incorrect description of a univentricular heart (including malestimation of ventricular size), incorrect VSD location in double outlet right ventricle, malestimation of the degree of mitral stenosis or outflow obstruction, arch sidedness, and discontinuous pulmonary arteries.
The authors concluded that discrepancies in pre- and postnatal CHD diagnosis result in significant changes in neonatal and late care management.
It is known that fetal echocardiography is not perfect. What this thoughtful study brings to the table is information on how discrepancies in fetal CHD diagnosis impact outcomes. It is somewhat comforting that more variances belong to group 3 (with little or no impact on care) than groups 1 and 2. Practitioners of this craft continue to be humbled as the old adage that a ‘normal’ fetal echocardiogram does not guarantee a ‘normal’ heart still holds true to this day. The fact that about 23% of fetal autopsies have a diagnosis variance that would have impacted care should give us pause, especially when pregnancy termination is considered. What this study fails to address is the impact of false-negative results in the primary screening level, the missed diagnoses of significant CHD that eluded high-level evaluation because they were not referred for a formal fetal echocardiogram.
< Back to Listings