Introduction: Total bilirubin tests are highly demanded in clinical laboratories. Since icteric index (I-index) has zero cost, we aimed to evaluate its clinical utility and cost-effectiveness to determine if total bilirubin is necessary to be tested. We took into account if haemolysis could interfere to icteric index determination.
Material and methods: Retrospectively we reviewed I-index results in two cohorts (43,372 and 8507 non-haemolysed and haemolysed samples, respectively). All determinations were done using Alinity c chemistry analysers (Abbott Diagnostics). Receiver operating characteristic (ROC) curve was used to determine the optimal index cut-off to discriminate between normal and abnormal bilirubin concentration (20.5 µmol/L).
Results: The ROC curve analysis suggested 21.4 µmol/L as the optimal I-index cut-off but differences in sensitivity and specificity were detected between patient derivation. For rejecting purpose, 15.4 µmol/L and 17.1 µmol/L I-index thresholds were selected based on patient derivation (inpatients and emergency room; and primary care and outpatients, respectively) with 97% sensitivity and 0.25% false negative results. Sensitivity was much lower in haemolysed samples. We selected 34.2 µmol/L I-index as threshold to detect hyperbilirubinemia with 99.7% specificity and 0.26% false positive results, independent of haemolysis. With the icteric index cut-offs proposed, we would save 66% of total bilirubin requested and analyse total bilirubin in around 2% of samples without total bilirubin requested.
Conclusions: This study supports the use of I-index to avoid bilirubin determination and to identify patients with hyperbilirubinemia. This work considers that the economic and test savings could help to increase the efficiency in clinical laboratories.
Keywords: bilirubin; haemolysis; hyperbilirubinemia; jaundice; receiver operating characteristic curve.
Croatian Society of Medical Biochemistry and Laboratory Medicine.