We read with great interest the article by Salinas et al 1 suggesting the use of the icteric index (II) as a front-line test for the preliminary identification of blood samples with abnormal total bilirubin (TB) concentrations. Following the authors' suggestion, we were encouraged to experimentally confirm their findings in our own hospital setting. Particularly, in our laboratory we use platforms (Cobas c 501 and Integra 800) and reagents of the same vendor (Roche Diagnostics) used by Salinas et al, and we think that our findings could contribute further evidence to their results, even if obtained with instruments of different series.We evaluated all the measurements of TB requested to our laboratory between January and December 2011 and their corresponding II values. TB concentrations and II were measured on Cobas c 501 using serum samples (routine) and on Integra 800 using lithium-heparin plasma samples (stat). Both systems use for TB the same diazo-based colorimetric assay, and for the II determination, an assay based on calculations of absorbance measurements of diluted samples. Statistical evaluation, including linear regression analysis, calculations of sensitivity, specificity, positive (LR+) and (LR−) negative likelihood ratios, positive predictive value (PPV) and negative predictive value (NPV) was performed for both serum and plasma samples, separately, by applying the same cut-off values used by Salinas et al, that is, 1.2 mg/dl (20.5 μmol/l) for TB concentrations and 2 for II. The analysis was Downloaded from *TB >1.2 mg/dl and icteric index <2. †TB ≤1.2 mg/dl and icteric index ≥2. LR+, positive likelihood ratio; LR−, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value.