Interference of Iodinated Contrast Media in Serum Capillary Zone Electrophoresis

Automated capillary zone electrophoresis (e.g., on the Paragon 2000 CZE, Beckman Instruments) is an alternative to conventional electrophoretic methods for the separation of serum proteins (1)(2)(3). The technique uses ultraviolet (UV) detection at 214 nm for direct quantification of peptide bonds, whereas conventional methods detect the protein fractions by staining with various reagents. CZE has the advantage of improved precision and a faster turnaround (4). Recent reports indicate, however, that the Paragon CZE 2000 masks deficiencies of α1-antitrypsin (5) and that some radio-opaque agents simulate a monoclonal component in the α2-globulin fraction in the CZE electropherogram (6)(7). We encountered an interference from the contrast medium Omnitrast®, which produced an unexpected peak in the α2-zone of the CZE electropherogram. We have now investigated the effect of 12 iodinated radio-opaque media (commercially available in Spain) on instrumentation (Paragon CZE 2000) operated as described by Bossuyt et al. (6). When we used the desalting method described by Blessum et al. (7) to remove the observed interferences (see below), we found undesirable obstruction in Paragon system capillaries. Therefore, we investigated alternative desalting procedures. The studied contrast media were Bilisegrol®, Gastrografin®, Urografin®, Uroangiografin®, Omnitrast, and Ultravist® (all from Schering); Telebrix®, Xenetix®, and Hexabrix® (all from Guerbet); Iopamiro® and Iomeron® (both from Rovi); Optiray® (Mallinckroft); and Omnipaque® (Nycomed), whose chemical compositions appear in Table 1 . The addition of each opaque medium (11 μL of medium to 1 mL of serum) to a control serum to reach the expected concentration after bolus injection for radiographic examination (7.5 g/L) led to the appearance of an abnormal peak in the CZE. Such induced peaks had the same shapes and locations as the extra peaks found in the electropherograms of patients who were infused with contrast agents. The interference peak in the electropherograms was located at the prealbumin zone for meglumine iotroxate, at the anodal site of the α2-globulin fraction for meglumine amidotrizoate, in the middle of the α2-globulin fraction for ioxitalamic acid and iobitridol, at the cathodal site of the α2-globulin fraction for iohexol, at the anodal site of the β-globulin fraction for sodium-meglumine ioxaglate, and in the middle of the β-globulin fraction for ioversol. In each of these cases, no peak was discerned on agarose electrophoresis (Paragon SPE kit; Beckman). The interference caused by these low-molecular weight agents was removed by addition of activated charcoal (0.2 g to 1 mL of serum) to serum, agitation of the mixture in a vortex-mixer for 20 s, and subsequent centrifugation at 2000g for 5–10 min at 25 °C. Serum was carefully recovered without removing the charcoal residues from the bottom and walls of the tube. In some cases, a second or a third centrifugation stage was necessary to achieve complete clarity of serum. This procedure was completely useful, fast, and simple for all samples. Bossuyt et al. (6) suggested (with data not shown) that this interference is attributable to the absorbance at 214 nm of both resolved proteins and iodinated contrast media. UV spectral data (obtained with a 550 UV-Vis Perkin-Elmer spectrophotometer) showed that the iodinated contrast agents absorbed light in the far UV but with maxima between 237 and 244 nm in either physiological media or CZE buffer at pH 10. The locations of the interfering peaks in the electropherogram became increasingly cathodal with increasing wavelengths of UV absorbance maxima of the radio-opaque media (Table 1 ). After treatment of the contrast media with the signal reagent (which, according to a personal communication from the manufacturer, contains at least two different peptides), shifts of the UV absorbance maxima at ∼270 nm were observed; however, if the process was repeated with a mixture of the signal reagent and the borate buffer reagent (pH 10) with adequate dilution, the final absorbance was decreased at the wavelength at which detection occurs (208–220 nm). These experimental findings gave evidence of actual binding of the radio-opaque agents to peptides. Therefore, we must reasonably wonder if the appearance of interfering peaks in the electropherograms is only a coincidence of the unexpected elution times and absorbance spectra of the iodinated radio-opaque media with the stated elution times and absorbance spectra of the proteins (6) or a result of binding to the signal reagent. We conclude that blood must not be collected for protein electrophoresis by CZE from patients who have received contrast media during the preceding 2–6 days. This time period has been described for some radio-opaque agents by Bossuyt et al. (6) and should be established for new contrast media according to their elimination times from blood. In addition, in questionable samples for paraprotein with negative immunofixation, we recommend the desalting procedure as described. Location in the electropherogram and UV maxima for radio-opaque media. Location in the electropherogram and UV maxima for radio-opaque media. We thank Schering España, Guerbet, Nycomed Amersham, Bracco (Rovi), and Mallinckrodt Medical for providing the necessary radio-opaque media for this study.