P0251 - Changes in Total and Free Bilirubin Following Treatment for Hyperbilirubinemia: Early Reports From Three Sites in Nigeria.

Background: Despite updates to practice guidelines regarding monitoring and treatment of hyperbilirubinemia acute bilirubin encephalopathy (ABE) and its consequence, kernicterus spectrum disorder, continues to occur. Total bilirubin (TB) is the standard measure of bilirubin despite free bilirubin (Bf) being the neurotoxic portion. In this study, we investigated whether Bf in the blood follows the same change as TB over the course of an infant’s treatment for hyperbilirubinemia and if either is predictive of clinical outcomes.

Objective: Correlate TB and Bf change over time to the clinical outcome of infants with severe hyperbilirubinemia.

Methods: 92 neonates ≥35weeks gestational age with evidence of hyperbilirubinemia and TB greater than 20mg/dl (342 µM/L) were recruited from three pediatric centers in northern Nigeria. Blood samples were collected with every clinically necessary blood draw. Clinical data, including bilirubin induced neurological dysfunction (BIND) scores to assess the severity of ABE, were collected. The serum was isolated and tested onsite for TB. Serum was then shipped to Children’s Mercy Hospital for testing of TB and other assays to assess integrity. Bf levels were measured by Fluoresprobe Sciences using a unbound bilirubin probe.

Results: Of the 92 neonates enrolled, 83 had at least 2 samples collected, and 69 had three or more. 33 infants had a BIND score ≥of 3, indicating significant ABE. Seven infants died. The ratio of Bf to TB trends down after treatment begins indicating that Bf values are dropping more quickly than TB values (Fig1A). Figure 1B. shows the spread of all the data points, including significant outliers. Figure 2A. demonstrates that the peak TB values are fairly linear across initial BIND scores while peak Bf values appear to trend upward with BIND scores. Interestingly, significant groups of outliers appear in the bottom and top BIND scores indicating possible subpopulations of extra resistant and extra sensitive. Error bars are as indicated in each figure. ANOVA with TUKEY post hoc comparison (*= p<0.05).

Conclusions: Preliminary results seem to indicate that Bf and TB respond differently to treatment, which supports previously reported data. The relationship between TB and Bf with BIND scores was not statistically significant, but several interesting patterns have begun to develop and appear to support the use of Bf measurements for predicting the severity of outcome over TB.