Rare IgM Interference with Roche/Hitachi Modular Glucose and γ-Glutamyltransferase Methods in Heparin Samples

We report an uncommon interference from IgM paraproteins in the measurement of glucose in heparin-plasma samples from 2 patients (patients A and B) and of γ-glutamyltransferase (GGT) in heparin samples from 1 of these patients (patient A). The interferences occurred with Roche reagents on Hitachi Modular D and P systems (Roche Diagnostics) with a hexokinase method for glucose (Modular P 1876899 and Modular D 1929569; application no. 767, Ver. 2) and a Szasz method (Modular P 2016958 and Modular D 2017288; application no. 219) for GGT. The glucose results for multiple specimens from each patient were low, in some cases giving negative results, and only the negative results were marked with error flags. The samples were automatically reanalyzed (on dilution with Modular P and the same volume with Modular D), and the same errors occurred, again flagged with the same error flags. The GGT results from patient A were falsely increased and were marked with error flags (limit or linearity flags). Both patients were known to have Waldenstrom macroglobulinemia (IgM-κ type). We found a single published report (1) on interference attributable to paraproteins in serum samples for the Roche glucose (hexokinase) method. To the best of our knowledge, no information has been published regarding such interference for lithium heparin–plasma samples. Furthermore, a review of the literature indicated that photometric interference with the IFCC and Scandinavian GGT methods can be caused by turbidity attributable to complex formation between heparin and various plasma proteins(2). Similarly, IgM paraprotein interference problems have been observed with the Roche GGT (Szasz) method(3). Paraprotein interferences have also been documented with other Roche assays or reagent formulations, such as phosphate(4)(5)(6), bilirubin, and HDL(6). Additionally, IgM interference in glucose and phosphate measurements has been reported with serum samples on other analytical systems(7). The errors arising with the glucose estimation prompted us to investigate possible causes. The GGT results were included to highlight the error type for the wider readership. Lithium heparin and serum samples were collected from patients A and B, as part of routine clinical care, into Greiner Vacuette serum and lithium heparin tubes with separator gel (Interpath, Australia). In addition to analysis on the Roche Hitachi analyzer, glucose was measured on a Vitros 950 analyzer using a glucose oxidase method in a dry-slide format (Ortho Clinical Diagnostics) and a Rapidlab 865 blood gas analyzer using a glucose oxidase electrode (Bayer Diagnostics). GGT was also measured on the Vitros 950 analyzer using the IFCC formulation in a dry-slide format. We also analyzed serum (0.5 mL) after it had been added to a lithium heparin tube as well as heparin-plasma samples (0.5 mL) to which we added 10 μL of 10 g/L (1%) protamine sulfate (Aventis). In each case the interference occurred in heparin plasma and only on the Roche analyzer (Table 11 ). The effect could be reproduced by adding heparin to serum, and it was largely blocked in heparin plasma by addition of the heparin antagonist protamine sulfate. Addition of reagent 1 (R1) to plasma caused a marked increase in the rate of increase of light absorbance compared with that seen with the serum samples (not shown), consistent with formation of macromolecular complexes. As a result, when the absorbance of samples with added R1 was greater than the final absorbance after addition of reagent 2 (R2), the signal was negative. The degree of change in absorbance on addition of R2 showed very little difference between the plasma and serum samples. This suggests that the complex formation/precipitation was caused by lithium heparin in combination with R1, more specifically, that the R1 composition facilitated precipitation. Addition of R1 to plasma from patient A led to a substantial increase in absorbance. The complex formation then caused photometric interference with the rate reaction on addition of R2, leading to linearity and limit flags and falsely increased results. The original discrepancy between plasma and serum results suggested that the interference could have been caused by fibrinogen precipitation resulting from the action of heparin in the appropriate conditions created by the addition of R1, rather than heparin and the paraprotein. Mixing of the serum sample with lithium heparin produced the same type of interference, which confirmed that precipitation of the paraprotein was induced by lithium heparin. For patient B, increasing the ratio of lithium heparin to paraprotein also led to precipitant formation and an erroneously high (flagged) result. The addition of protamine sulfate (which counteracts heparin activity) to the plasma sample produced the same results as the serum sample and did not lead to error flags, thus confirming that the precipitant formation was attributable to the interaction of lithium heparin with the paraprotein. When we used serum samples, we observed no differences between the Roche and Vitros glucose or GGT results. Similarly, we observed no differences between heparin-plasma and serum samples analyzed on the Vitros system. The layered slide technology on the Vitros prevents precipitation penetrating through to the reaction layer. When we measured glucose, we observed no differences between the heparin-plasma or serum samples with the Rapidlab 865. The membrane on the glucose sensor also prevents precipitation from potentially causing reaction problems. In conclusion, 2 samples with IgM-κ paraproteins produced interference with the Roche glucose and GGT methods when collected as lithium heparin–plasma samples. Automatic reruns gave the same error flags. The best way to obtain accurate glucose and GGT results for these patients was to use a serum sample or a different measurement system for glucose analysis. Glucose and GGT results obtained with the Roche Hitachi Modular, Vitros 950, and Bayer Rapidlab 865 methods for simultaneously collected lithium heparin–plasma and serum samples from 2 patients. Lin, linearity error flag; LIMTL, limit error flag. Serum samples mixed with lithium heparin. Heparin plasma treated with 10 g/L (1%) protamine sulfate. Glucose and GGT results obtained with the Roche Hitachi Modular, Vitros 950, and Bayer Rapidlab 865 methods for simultaneously collected lithium heparin–plasma and serum samples from 2 patients. Lin, linearity error flag; LIMTL, limit error flag. Serum samples mixed with lithium heparin. Heparin plasma treated with 10 g/L (1%) protamine sulfate. We thank Barbara Mottram (Mater Laboratories Services, Brisbane, Australia) for assaying the samples on their Vitros 950 analytical system and Dr. Graham Jones (St Vincent’s Hospital, Sydney, Australia) for reviewing the manuscript.