Critical congenital heart defects (CCHD) are the most common congenital malformations and account for up to 10 percent of infant deaths each year. Universal screening with pulse oximetry screening has been recommended for more than a decade. The updated 2018 American Academy of Pediatrics (AAP) algorithm recommends testing in the (pre-ductal) right upper extremity and one of the feet, with a pulse ox of at least 95 percent in both extremities with under three percent difference between the two extremities required for a negative screen. The current algorithm allows for only one rescreen an hour after the first. The prior iteration of the algorithm was more liberal —- with passing results given to newborns with a pulse oximetry of at least 95 percent in either extremity with allowance for two rescreens. Studies have found variability in the rates of false positive results and raise concerns about the anxiety and cost of false positive screening, particularly with the newer, stricter algorithm.
Investigators at a large community hospital conducted a retrospective study from 2012 to 2020 to evaluate the test characteristics of pulse-ox screening based upon the standard and revised algorithms. After excluding those with prenatally diagnosed CCHD or symptoms, 65,403 newborns admitted to the well-baby nursery during this time were screened. Data from pulse-ox screening were extracted from a Department of Health database, which keeps records of all newborn screens. Authors conducted extensive chart review at the community hospital and surrounding tertiary care centers for those with an abnormal pulse-ox screening. Those with a normal pulse-ox but abnormal echocardiogram (obtained for other reasons) were counted as false negative results. To identify false negatives after discharge, researchers searched a large surveillance database for newborns with a normal pulse-ox screening and subsequent CCHD diagnosis.
Based on the standard algorithm, a total of 31 newborns had an abnormal pulse-ox screening (fail rate of 4.6 per 10,000). Of these, 12 had CCHD such as total anomalous pulmonary venous return or aortic coarctation, nine had non-CCHD, such as AV canal defect, and the remainder had a non-cardiac condition such as pneumonia. Through the registry data of the state, one false negative was identified over the eight-year period, a newborn found to have aortic coarctation at 31 days of life. This resulted in a specificity of 99.97 percent and a sensitivity of 92.3 percent for CCHD. If the revised algorithm had been applied, five additional newborns would have failed. Only one of these had a significant non-CCHD illness and the remainder would have been classified as normal, increasing the false-positive rate to 0.04 percent (4 per 10,000).
Pulse oximetry appears to meet all the criteria for a ‘good’ screening test. This relatively easy, cheap, and quick test appears to identify most cases of CCHD. Limitations to this study include incomplete ascertainment of false negative results as only a single state registry was searched for missed cases. The standard and revised versions of the pulse-ox screening algorithm identified CCHD at similar rates, suggesting that the revised algorithm may result in more false positive results without additional benefit.