The Utility of Inpatient Rebound Bilirubin Levels in Infants Readmitted After Birth Hospitalization for Hyperbilirubinemia

DISCUSSION

To date, there is a limited amount of research addressing the potential clinical utility of early (<12 hours) rebound bilirubin levels.^9–11 In 1999, Del Vecchio et al⁹ analyzed 48 infants with rebound levels ranging from 6 to 8 hours and found a decrease in average rebound bilirubin level compared with level at cessation of phototherapy. A 2002 retrospective analysis by Al-Saedi¹⁰ evaluated 301 infants with rebound bilirubin level times averaging 8.3 ± 5.3 hours and found no significant increase in early rebound bilirubin levels. Our study is in line with this previous research, as we reported no significant increase in early inpatient rebound bilirubin levels drawn at a mean of 6.1 ± 2.4 hours compared with the level at phototherapy cessation. However, to our knowledge, our study is the first to directly compare patients with early rebound bilirubin levels versus those without rebound levels, specifically addressing differences in clinical outcomes such as readmission rates, length of stay, and discharge timing.

Early inpatient rebound bilirubin levels were checked in a majority of infants at our institution, and this practice did not successfully prevent readmission. However, within the rebound group, 5 (3.8%) of 130 patients underwent repeat phototherapy during the same admission. Of these 5 infants, 1 had a cephalohematoma, and 2 had ABO incompatibility with negative results on the Coombs test. Importantly, all of these patients had phototherapy halted at a level >14 mg/dL, with bilirubin levels of 14.7, 15.9, 16.7, 16.8, and 17.6 mg/dL at phototherapy cessation. In cases in which phototherapy was initially continued to reach the AAP recommendation of ≤14 mg/dL, all rebound bilirubin levels remained below the light level for recommended phototherapy.

In the entire cohort, children with bilirubin levels ≤14 mg/dL were much less likely to receive repeat phototherapy than those who had phototherapy discontinued at a level >14 mg/dL. In the 14 children who required repeat phototherapy, 12 had phototherapy discontinued at a level >14 mg/dL. There were only 2 patients who received repeat phototherapy despite adherence to the AAP guideline of lowering the bilirubin to ≤14 mg/dL. Interestingly, one of these patients (originally from the no-rebound group) had a rebound bilirubin level checked on a subsequent hospitalization, and this rebound level still failed to prevent a third hospitalization. He was eventually diagnosed with elliptocytosis. This finding suggests that children who require recurrent phototherapy despite a level at cessation of ≤14 mg/dL may warrant additional evaluation for hemolysis.

Patients in the rebound group had longer lengths of stay and later discharge times, exposing them to potential iatrogenic complications. These patients also received significantly less time under phototherapy than children who did not have a rebound level. Thus, the rebound group spent more time in the hospital but received less total therapy. Keeping these patients under the lights for additional time to lower bilirubin levels to ≤14 mg/dL was more likely to prevent readmission than monitoring them with a rebound bilirubin level.

Furthermore, there was an increase in cost associated with the rebound groups’ prolonged length of stay and extra laboratory testing. Our institution’s financial department provided estimates for hourly hospital charges of approximately $25.00 and $50.00 for serum bilirubin testing. Using these estimates, the rebound bilirubin group averaged approximately $150.00 extra per patient. This cost is further compounded when considering the improvements in patient flow associated with the earlier discharge times observed in the no-rebound group.

The present study’s main limitation was its retrospective design, which provides no concrete way to control for why individual attending physicians chose to check a rebound level in some infants but not in others. We tried to address this limitation by evaluating baseline characteristics that typically influence this decision, and we found no significant differences in these characteristics between the 2 groups. There was a statistically significant difference between bilirubin levels at the time of phototherapy initiation and cessation, implying that the rebound group’s jaundice was more severe than the no-rebound group. However, the significantly shorter period of time under phototherapy for the rebound group argued against this idea, as one would potentially expect patients with more severe conditions to require more phototherapy. In addition, the mean bilirubin levels at initiation of phototherapy did not approach exchange transfusion levels in either group.

Another significant limitation was the study’s relatively small sample size, particularly considering that the primary outcome of hospital readmission was such a rare occurrence. These circumstances made it difficult to obtain the power necessary to definitively detect subtle differences in readmission rates between the 2 groups, and our post hoc power analysis showed our study to be powered at 80% to detect a 12% difference in readmission rates. Before 2007, our institution’s policy was to obtain rebound bilirubin levels on all patients receiving phototherapy, which limited our ability to increase patient recruitment before this time period. After 2007, the hospitalist group adopted the policy of not obtaining rebound bilirubin levels, whereas the majority of community pediatricians continued to obtain rebound bilirubin levels. This clinical environment made the comparison of the 2 groups possible but again did not provide a large enough sample size to afford substantial power to detect subtle differences in an outcome with such a rare occurrence.