Diagnostic accuracy of transcutaneous bilirubin measurement before, during, and after phototherapy in late preterm and term neonates: a prospective observational study

Background Neonatal hyperbilirubinemia is a common condition requiring assessment and treatment. Although total serum bilirubin (TSB) is the standard for bilirubin monitoring, repeated blood sampling causes discomfort and blood loss. Transcutaneous bilirubin (TcB) offers a non-invasive alternative; however, phototherapy may affect its accuracy. This study evaluated the diagnostic accuracy and agreement of TcB with TSB before, during, and after phototherapy in late preterm and term neonates. Methods This prospective observational study included 75 neonates (38 late preterm and 37 term infants). TcB was measured at the forehead and sternum, using protected skin sites during phototherapy. Corresponding TSB measurements were obtained at each time point. Correlation, mean bias, Bland–Altman limits of agreement, diagnostic performance, and receiver operating characteristic (ROC) curve analysis were assessed. Results Mean gestational age was 35.4 ± 0.8 weeks in late preterm and 39.3 ± 1.3 weeks in term neonates. Before phototherapy, the mean TcB–TSB difference was −1.2 ± 0.7 mg/dL in late preterm and −0.8 ± 0.6 mg/dL in term neonates. The average of forehead and sternum measurements showed strong correlation with TSB before phototherapy (r=0.91 and r=0.95, respectively; both p<0.001). Correlation decreased during phototherapy (r=0.73 and r=0.77) and improved afterwards (r=0.86 and r=0.90). Overall ROC analysis yielded areas under the curve of 0.93, 0.80, and 0.91 before, during, and after phototherapy, respectively. Optimal TcB cut-offs were 14.5, 14.0, and 14.5 mg/dL, respectively. Overall Bland–Altman bias was −1.0, −1.6, and −1.0 mg/dL, respectively. Conclusion TcB showed better diagnostic performance before and after phototherapy than during treatment. Protected-site measurements may assist non-invasive bilirubin monitoring; however, TSB confirmation is recommended when values approach treatment thresholds or clinical decisions depend on accurate bilirubin estimation.