Clinical

Visual screening inadequate for detecting neonatal hyperbilirubinemia

Last Updated: 2008-06-27 12:52:56 -0400 (Reuters Health)

By David Douglas

NEW YORK (Reuters Health) - Visual assessment of jaundice is not reliable as a screening tool for detection of significant neonatal hyperbilirubinemia, Israeli researchers report in the June issue of Pediatrics.

"Although we are able to see that a baby is jaundiced, and have some ability to estimate the degree of jaundice by eye," lead investigator Dr. Arieh Riskin told Reuters Health, "for clinical practice and especially before discharge home from the nursery, this clinical estimation of jaundice is not reliable as the sole screening tool to detect those infants who may develop significant neonatal jaundice that may require treatment."

Dr. Riskin of Bnai-Zion Medical Center, Haifa and colleagues note that although total serum bilirubin (TSB) levels are useful in this regard, implementation of this approach and ancillary procedures can be costly and time consuming. To determine whether visual assessment might be a reliable alternative, the researchers examined the performance of 5 neonatologists and 17 nurses who made 3,532 observations in 1129 newborns.

Overall, the researchers found a good correlation between visual scoring and TSB levels. In fact, visual scoring gave a mean serum bilirubin level of 6.6 mg/dL compared to 6.7 mg/dL with TSB.

However, they say, other measures of agreement were poor. For example, 61.5% of the 109 babies deemed at high risk by TSB were wrongly classified as being at low risk by visual assessment.

Although there are few high risk cases (3.1%), such misclassification "is unacceptable for a screening test," the investigators stress.

Thus, they conclude that "pre-discharge screening for neonatal hyperbilirubinemia should be based on other modes, either universal blood sampling for TSB with the routine newborn screening or non-invasive transcutaneous bilirubinometry for more accurate evaluation of skin jaundice."

J Pediatr 2008;152:782-787.