Incidence and clinical predictors of hyper-direct bilirubinemia in preterm infants without underlying disease

Backgrounds: This study aimed to identify the incidence and clinical predictors of hyper-direct bilirubinemia (hyper-DB) in preterm infants without underlying diseases. Methods: We enrolled neonates born at <34 weeks of gestational age (GA) between 2019 and 2020. The incidence of hyper-DB was calculated, and neonates were categorized into hyper-DB and nonhyper-DB groups. Hyper-DB was defined as DB ≥ 1 mg/dL when total bilirubin (TB) was <5 mg/dL or DB ≥20% of TB when TB was ≥5 mg/dL during their neonatal intensive care unit stay. Clinical data regarding maternal and neonatal factors were compared using univariate and multivariate analyses, respectively. A receiver operating characteristic curve was generated and the threshold value of the GA was determined using the Youden index. Results: Hyper-DB was diagnosed in 16 of the 131 infants (12%). Eleven clinical factors, including GA, birth weight, absence of premature rupture of membranes (PROM), and incidence of neonatal persistent pulmonary hypertension (PPHN), were significantly different between the two groups (p < 0.05). Multivariate analyses showed that a shorter GA (odds ratio [OR]: 0.48), presence of PPHN (OR: 87.2), and absence of PROM (OR: 0.01) were independent clinical predictors of the development of hyper-DB. Using the Youden index, a cutoff value of 30 weeks for GA was determined as the threshold to manifest hyper-DB. Conclusions: We observed that 12% of preterm infants at <34 weeks' GA without an underlying disease developed hyper-DB. Low GA (less than 30 weeks), presence of PPHN, and absence of PROM were associated with the development of hyper-DB.