Prediction of Hyperbilirubinemia in Coombs Positive Infants
Lawrence M. Noble, Ivan L. Hand, Alan Shotkin, Donna Geiss Pediatrics, Albert Einstein Col Med/Jacobi Medical Center, Bronx, NY
BACKGROUND: There is growing concern on the association of early discharge of newborns from the hospital and the risk of hyperbilirubinemia and possible kernicterus. AAP guidelines limit early discharges to babies without hemolytic disease. However, most newborns who are Coombs positive never develop significant hemolysis and clinically significant hyperbilirubinemia. An easy test that could predict which Coombs positive infants are at risk would be of great importance. We hypothesized that a reticulocyte count, as a marker of hemolysis, could predict which babies would develop hyperbilirubinemia.
OBJECTIVE: The objective of this study was to determine the usefulness of a reticulocyte count in Coombs positive newbons in predicting those who are at risk for hyperbilirubinemia.
DESIGN/METHODS: 4,220 term or near-term healthy newborns consecutively admitted to the newborn nursery from August 1999 through November 2001 had a blood type and Coombs performed. Those who were Coombs positive were included in the study. As part of the medical care of the baby, a reticulocyte count and bilirubin values were obtained. Hyperbilirubinemia was defined as a peak bilirubin ≥15 and severe hyperbilirubinemia was defined as a peak bilirubin ≥17.
RESULTS: There were 155 Coombs positive infants(3.7%): 81(52%)with OA incompatability and 66(43%) OB. Mean reticulocyte count was significantly different between those with OA or OB incompatability (7.3% vs. 12.5%, p=.003), but there was no significant difference in bilirubin levels (9.1 vs. 9.8mg/dl). 13% of these newborns developed hyperbilirubinemia and 7% severe hyperbilirubinemia. The infants who developed hyperbilirubinemia or severe hyperbilirubinemia had a reticulocyte value greater than those who did not (13.2±7.0 vs. 8.7±6.4, p=.022, 15.9±7.3 vs. 8.7±6.2, p=.003, respectively). 30% of the babies had a reticulocyte count ≥11. As a screening test for hyperbilirubinemia, a reticulocyte value ≥11 had a sensitivity of 60%, specificity 77%, PPV of 45% and NPV of 86%. As a screen for severe hyperbilirubinemia, it had a sensitivity of 89%, specificity of 79%, PPV of 40% and NPV of 97%.
CONCLUSIONS: Although a reticulocyte count of ≥11 in Coombs positive newborns is a poor predictor of hyperbilirubinemia, a reticulocyte count of <11 may be important in predicting infants who are not at risk for significant hyperbilirubinemia (NPV=97%). These infants may be candidates for shortened hospital stays and less bilirubin determinations.