Objectives. To determine the usefulness of the bilirubin-albumin (B:A) mola
r ratio (MR) and unbound bilirubin (UB) as compared with serum total biliru
bin (TB) in predicting bilirubin encephalopathy as assessed by auditory bra
instem responses (ABR) in infants of 28 to 32 weeks' gestational age.
Study Design. During a 2-year period, serial ABRs were obtained on 143 infa
nts of 28 to 32 weeks' gestational age during the first postnatal week. Wav
eforms were categorized on the basis of response replicability and the pres
ence of waves III and V. Wave V latencies were also serially analyzed when
measurable for individual infants. Maturation of the ABR was defined as abn
ormal when the waveform category worsened and/or latency increased during t
he study interval. Serum albumin was analyzed at 48 to 72 hours of age in a
ll patients. Serum TB was analyzed as clinically indicated. Aliquots of the
same samples were also analyzed for UB in a subset of infants.
Results. The mean peak TB concentration (10.1 +/- 1.7 mg/dL) for the 71 inf
ants with normal ABR maturation was not significantly different from the me
an peak TB (10.2 +/- 2.1 mg/dL) in the 24-hour period preceding the ABR's f
irst showing abnormal maturation in the other 55 infants. However, in infan
ts with UB analyzed, the mean peak UB (0.62 +/- 0.20 vs 0.40 +/- 0.15 mg/dL
) was significantly higher in the infants with abnormal maturation (n = 25)
than in infants with normal maturation (n = 20). The B:A MR results were e
quivocal. In the entire study population, there was no difference in B:A MR
between infants with normal versus abnormal ABR maturation. However, in th
e subset of infants in whom UB was measured, although TB was not different,
there was a significant difference in B:A MR. Based on receiver-operating
characteristic curves, a UB level of 0.5 mg/dL was the best discriminator w
ith a sensitivity of 70% and a specificity of 75%. The proportion of infant
s who had UB >0.5 mug/dL and UB less than or equal to0.5 mg/dL and who had
abnormal ABR, maturation was 0.81 and 0.33, respectively, with a significan
t difference in the incidence of transient bilirubin encephalopathy among t
hese 2 groups. The relative risk of abnormal ABR maturation with UB >0.5 mu
g/dL compared with UB less than or equal to0.05 mg/dL was 2.45 (95% confide
nce interval: 1.33-4.49).
Conclusions. UB is a more sensitive predictor than either serum bilirubin o
r B:A MR of abnormal ABR maturation, and hence transient bilirubin encephal
opathy in premature newborns with hyperbilirubinemia.