Chronic lymphocytic leukemia cells may interfere in a glycated hemoglobin assay based on fluorescence quenching.

Recently, we observed falsely low results by the Abbott IMx Glycated Hemoglobin assay apparently caused by a high number of leukocytes in patients with chronic lymphocytic leukemia. Because chronic lymphocytic leukemia and diabetes mellitus (monitored by glycohemoglobin measurements) occur more frequently in the elderly, we examined this interference further. The IMx Glycated Hemoglobin assay (1 ) is based on ion-capture technology. A glass-fiber matrix is precoated with a positively charged quaternary ammonium compound that, after hemolysis of the blood sample, electrostatically captures the negatively charged glycohemoglobin complex. A dihydroxyboronate/high-molecular weight polyacrylic acid affinity reagent specifically captures the glycohemoglobin, whereas the nonglycohemoglobin constituents are removed by a washing procedure. Glycohemoglobin is quantified by measurement of fluorescence quenching. In 20 consecutive blood samples from patients with known chronic lymphocytic leukemia, a total leukocyte count and a differential were done in a Coulter Counter STKS, and glycohemoglobin was measured simultaneously in the Abbott IMx (ion capture technology, fluorescence quenching), the Bayer DCA 2000 (immunologic latex agglutination), and the Bio-Rad DiaSTAT (low pressure liquid chromatography). The results, arranged in ascending leukocyte order, are shown in Table 1. Constant systematic differences were observed among the three methods. Compared with the two other methods, however, the Abbott IMx results seemed to be affected by high total leukocyte counts, with falsely decreased values at leukocyte concentrations in the range 67.9 3 10/L to 100 3 10/L. At higher leukocyte counts, asterisks appeared as an error warning. After the patients’ blood samples were mixed manually with the hemolyzing and other reagents for the IMx glycohemoglobin assay in the same proportions as are used in this analyzer, numerous intact leukocytes (lymphocytes) could be observed by microscopy. The draining velocities through the glass fiber matrix were comparable for solutions of the patients’ samples and for blood from healthy subjects. Hence, the length of time that solutions remained on the matrix during the reading does not appear to contribute to interference. None of these patients had a monoclonal gammopathy, some of which can react with quaternary amines (D. Bruns, personal communication, October 5, 1998). Aspiration of the buffy coat in one patient’s blood sample, which reduced the lymphocyte count from 154 3 10/L to 93.7 3 10/L, avoided the false lowering of the glycohemoglobin result. Hence, interference from intact leukocytes in the glycohemoglobin fluorescence quenching is the most likely explanation of the falsely low results obtained in these samples. As suggested by the results in the leukocyte range 67.9 3 10/L to 100 3 10/L, individual differences may exist in leukocyte (lymphocyte) resistance to the hemolytic agents used. Interference by high numbers of leukocytes in a fluorescence quenching assay for glycohemoglobin has not been reported. We suggest the possibility of such interference in other fluorescence quenching assays based on “hemolyzed” whole blood.