When results matter: reliable creatinine concentrations in hyperbilirubinemia patients.

Background Failure to report a creatinine concentration, especially in icteric patients who are eligible for a liver transplant, can result in a life-threatening situation. We assessed the influence of bilirubin interference on several creatinine assays and investigated ways to circumvent icteric interference without interfering with our normal automated sample logistics. Methods Using icteric patient sera (total bilirubin >255 μmol/L) we determined creatinine concentrations using an enzymatic and Jaffé assay (Roche Diagnostics) in both normal (i.e. undiluted) and decreased mode (i.e. diluted) as well as an enzyme-coupled amperometric assay on a Radiometer ABL837 FLEX analyzer. Creatinine concentrations from the five methods were compared with an in-house developed LC-MS/MS method. Passing and Bablok (proportional and constant bias) as well as difference plot parameters (bias and 95% limits of agreement [LoA]) were calculated. Interferograph-based regression analysis of the enzymatic and Jaffé results was used to investigate if such an approach could be used to report corrected creatinine concentrations in icteric patient sera. Results In icteric patient sera the enzyme-coupled amperometric assay was hardly influenced by icteric interference as shown by a difference plot bias of -1.5% (95% LoA -11.6 to +8.5%). The undiluted Jaffé method had a bias of -1.4% with a very broad 95% LoA (-35.1 to +32.2%) emphasizing the poor specificity of this method. The undiluted enzymatic method had the largest bias (-13.4%, 95% LoA -35.8 to +9.0%). Diluting sera in the enzymatic method did not improve the bias (-10.5%, 95% LoA -25.4 to 4.4%), while diluting the Jaffé method resulted in a bias increase (+11.4%, 95% LoA -14.7 to 37.5%). Using interferograph-based regression analysis we were able to reliably correct enzymatic creatinine concentrations in 97 out of 100 icteric patient sera. Conclusions Analytically, quantifying creatinine in icteric sera using the Radiometer ABL837 FLEX analyzer is the method of choice within our laboratory. However, not all laboratories are equipped with this method and even if available, the limited number of highly icteric patient sera makes this method costly. For those laboratories using the Roche enzymatic method, mathematically correcting an icteric creatinine concentration using an interferograph based on an LC-MS/MS reference method is a suitable alternative to report reliable creatinine results in icteric patients.