To verify the reliability and safety of transcutaneous bilirubin (TcB) measurements in patched skin areas in neonates born extremely preterm under phototherapy.

Sixty neonates (<30 weeks' gestation) receiving phototherapy were enrolled and TcB was measured via a second-generation transcutaneous bilirubinometer in patched skin areas (of at least 2.5 cm diameter). Total serum bilirubin (TSB), lactate, pH, hemoglobin, and skin temperature were measured within 10 minutes of the TcB assay. Clinicians were blinded to TcB values, and clinical decisions about phototherapy were made with the TSB measurement only.

TcB and TSB significantly were correlated (r = 0.84; P < .001), even after adjustment for hemoglobin, pH, lactate, gestational and postnatal age (standardized β = 0.8; P < .001; adjusted R² = 0.75), or treatment duration (standardized β = 0.8; P < .001; adjusted R² = 0.7). When the Bland-Altman analysis was used, TcB overestimated TSB at high values (mean difference TSB – TcB: −2.8 [2.4] mg/dL). If clinicians used the TcB only, no neonate would have had phototherapy stopped prematurely, and 21 (35%) would have continued phototherapy when it could have been stopped.

The correlation between TSB and TcB (measured in patched skin areas) was comparable with that obtained in more mature neonates, and it was not influenced by clinical variables or factors affecting skin bilirubin passage. TcB overestimated TSB, and this may expose infants born preterm to unnecessary phototherapy, although it could spare approximately 65% of TSB assays.