Detection of Monoclonal Proteins by Capillary Zone Electrophoresis: Comparison of 2 Multichannel Automated Systems

To the Editor: Capillary zone electrophoresis (CZE) is an alternative method for separation of serum proteins (1 ). Two dedicated and automated multichannel instruments are available, the Paragon 2000 (Beckman-Coulter) and the Capillarys (Sebia). The sensitivity and specificity of Paragon 2000 for detection of monoclonal proteins have been reported to be 95% (2–4) and 78% (4 ), respectively. The low reported specificity reflects the frequent occurrence of slight abnormalities in the electropherogram at the anodal part of the -globulin fraction (fibrinogen region) (4 ). Only 2 studies (5, 6) evaluated the Capillarys for detection of monoclonal proteins and reported a high sensitivity, and prospective studies in a routine clinical setting are lacking. We studied 597 consecutive samples submitted to the laboratory to test for the presence of a monoclonal protein or to reevaluate a known monoclonal protein. All samples were analyzed by the Capillarys (software version 5.2.1) and Paragon 2000 (software version 1.6.02) systems and by immunofixation, which was used as the reference standard. Immunofixation was performed on the Hydrasys Automate (Sebia) according to the manufacturer’s instructions, with the use of Hydragel 4 immunofixation gels. Immunofixation revealed the presence of a distinct monoclonal protein in 246 of the 597 samples. The distribution of the types of monoclonal proteins is given in Table 1. The sensitivity of Capillarys for detecting a monoclonal protein was 90% [95% confidence interval (CI), 85%–93%]. This sensitivity was comparable to that of the Paragon, which was 91.5% (95% CI, 87%–95%; P 0.5 by ). In 267 samples, immunofixation revealed no abnormalities. The specificity of the Capillarys was calculated to be 87% (95% CI, 83%–91%), higher than the specificity of the Paragon, which was 58% (95% CI, 52%– 66%; P 0.0001 by ). The lower specificity of the Paragon 2000 was attributable to the previously described (4 ) disturbed morphology at the fibrinogen position. If such disturbed morphology was not considered abnormal, then the specificity of the Paragon was 80% (95% CI, 75%– 88%;P 0.03 for comparison with Capillarys by ). In 84 samples, immunofixation was difficult to interpret (presence of M-protein not excluded, M-protein very faint or not identifiable, differentiation between monoclonal and polyclonal pattern questionable). The results for the Paragon 2000 and Capillarys for these samples are available from the authors. Quantification of the monoclonal protein by delimitation was possible in 87 samples. There was a good correlation between the Capillarys and Paragon (r 0.99; Pearson). Bland-Altman analysis revealed an intercept of 0.81 (95% CI, 0.558– 1.096) and a slope of 0.96 (95% CI, 0.95–0.98). As shown above, each system may fail to detect monoclonal proteins. For example, the Paragon 2000 failed to detect a monoclonal protein in a sample in which immunofixation revealed an IgM monoclonal protein. This monoclonal protein was clearly detected by the Capillarys. Conversely, the Capillarys failed to detect a significant IgM monoclonal protein and a IgG monoclonal protein. In both samples the CZE electropherograms showed a minor disturbance in morphology of the -region but not the obvious monoclonal peak that was observed with agarose gel electrophoresis. A remarkable finding on the Capillarys analysis, observed in both samples, was an unusually long separation between the albumin fraction and the 1-globulin fraction.