Diagnostic Properties of a Portable Point-of-Care Method to Measure Bilirubin and a Transcutaneous Bilirubinometer

Background: Recently, the Bilistick®, a point-of-care instrument to measure bilirubin levels, has been developed. It is fast and cheaper than transcutaneous bilirubin (TCB)-measuring devices, but data on diagnostic properties are scarce. Objective: This study aimed to compare the performance of the Bilistick® (BM-BS 1.0 - FW version 2.0.1) and the JM-105 bilirubinometer for measuring bilirubin. Method: This is a prospective study in infants born after ≥32 weeks' gestation, and/or a birth weight of ≥1,500 g, and a postnatal age ≤14 days in Surabaya, Indonesia. Bilirubin was measured with the Bilistick® System (BM-BS 1.0 - FW version 2.0.1), transcutaneously (TCB) with the JM-105 bilirubinometer, and in serum (TSB) with a routine laboratory technique. Mean differences and 95% limits of agreement (LOA) and correlations were calculated. Result: We enrolled 149 neonates and 126 had paired measurements of Bilistick® bilirubin, TCB, and TSB. Bilistick® failed in 16 (10.7%) infants. Mean Bilistick® bilirubin-TSB difference was -11 μmol/L (95% LOA: -101 to 79 μmol/L) and r = 0.738 (p < 0.001). Mean TCB-TSB difference was 26 μmol/L (95% LOA: -33 to 88) and r = 0.785 (p < 0.001). The sensitivity, specificity, PPV, and NPV for Bilistick® bilirubin for a TSB above treatment thresholds were 0.74, 0.84, 0.67, and 0.88, respectively, and for TCB 0.92, 0.64, 0.54, and 0.95, respectively. Conclusion: The Bilistick® System (BM-BS 1.0 - FW version 2.0.1) underestimates TSB, whereas TCB overestimates TSB in jaundiced Indonesian infants. Further improvement of Bilistick®'s diagnostic accuracy with less false-negative readings is essential to increase its use.