Background Life-threatening congenital heart defects (CHD) often go undetected. This prospective study assessed the accuracy of pulse oximetry as a screening test for CHD.

Methods Pulse oximetry was performed in asymptomatic newborns >34 weeks gestation, prior to discharge from six maternity units. Those not achieving predetermined oxygen saturation thresholds underwent echocardiography. All other infants were followed up to age 12 months through interrogation of registries and clinical follow-up. The main outcome measure was detection of major CHD – subdivided into critical (death or intervention before 28 days), and serious (death or intervention between 1 and 12 months of age).

Findings 20 055 babies were screened and 53 had major CHD (24 critical, 29 serious). Pulse oximetry had a sensitivity of 75.0% (95%CI 53.3% to 90.2%) for critical cases and 49.1% (95%CI 35.1% to 63.2%) for all major CHD. In 35 cases CHD was already suspected following antenatal US, when these were excluded, pulse oximetry had a sensitivity of 58.3% (95%CI 27.7% to 84.8%) for critical cases and 28.6% (95%CI 14.6% to 46.3%) for all major CHD. False positive results occurred in 0.8% of babies (specificity: 99.2%, 95%CI 99.0% to 99.3%). However of the 169 false positives, there were six cases of significant (not major) CHD and 40 cases of illness requiring medical intervention.

Interpretation Pulse oximetry is a safe, feasible test which adds value to existing screening. It identifies cases of critical CHD which go undetected with antenatal ultrasound screening. The early detection of other pathologies is an additional advantage.