The accuracy of transcutaneous bilirubin to identify hyperbilirubinemia in jaundiced neonates

Objective To compare transcutaneous bilirubin (TcB) to total serum bilirubin (TSB) in jaundiced neonates by using a JM-103 bilirubinometer. Methods We conducted a retrospective study in jaundiced infants >= 35 weeks' gestation. Infants' and maternal information, TcB measurements, and TSB levels were collected. The correlation and consistency between TcB and TSB were analyzed by Pearson's correlation analysis and the Bland-Altman plot. Multivariate regression analysis was used to identify characteristics associated with the difference between TcB and TSB, underestimation by TcB and overestimation by TcB. Results A total of 787 paired samples were collected. There was a strong correlation between TcB and TSB with a Pearson correlation coefficient of 0.708 (p < .001). In preterm infants, outpatients and infants with high TSB levels, the TcB measurements were less precise. The TcB-TSB difference was significantly associated with gender, maternal hypertension disease, ABO hemolytic disease, and outpatient setting (p < .05). Only the outpatient setting had statistical significance in the underestimation of TSB >= 2 mg/dL. Conclusions Generally, TcB by a JM-103 bilirubinometer had a good correlation and consistency with the TSB values. However, small for gestational age, postnatal age of hour, ABO hemolytic disease, and outpatient setting may lead to inconsistencies between the TcB and TSB.