Reproducability of non-invasive bilirubin measurements

To determine the level of agreement between transcutaneous bilirubin (TcB) measurements when sequentially measured by one and two users.

A method comparison study design was used to compare two sequential TcB measurements (JM103, Drager Medical Inc., Telford, PA) obtained by the same user and then by two different users. Users were blinded to each others' results. Differences (bias) and limits of agreement (±2 precision or SD) between sequential measurements with the TcB by a single user and two users were calculated and graphed using the Bland-Altman method. Acceptable levels for bias and precision were set a priori at bias ≤±1.5 mg/dL and/or precision ≤±3.0 mg/dL.

A total of 34 newborn infants were studied, with total serum bilirubin (TSB) levels ranging from 2.7 to 8.1 mg/dL, averaging (±SD) 5.5 ± 1.8 mg/dL. Differences (bias) and the level of agreement (precision; SD) for two sequential noninvasive TcB measurements obtained by a single user and two different users (-0.2 ± 0.7 and -0.4 ± 0.7, respectively) were within the acceptable ranges.

We found good agreement between TcB measurements obtained sequentially by a single user and by two different users of the device, as well as between laboratory TSB values in low-risk newborn infants with normal bilirubin levels. These findings support the use of a noninvasive bilirubin meter to screen for hyperbilirubinemia, which could reduce the amount of blood obtained invasively from newborns.