Falsely Increased hCG in Patients with High Leukocyte Counts

Falsely increased plasma troponin (Tn) I and T results have been reported for several Tn methods, including the Beckman Coulter UniCel DxI AccuTnI assay (Beckman Coulter, Chaska, MN) (1–3). We and others have also observed falsely increased human chorionic gonadotropin (hCG) using the Beckman Coulter Access Total hCG on the UniCel DxI (1). These false increases were identified by the finding of substantially lower results following high-speed recentrifugation and reanalysis, suggesting a particulate interference. However, the sources of these false increases were not previously determined. In reviewing clinical data for patients from whom plasma samples were found to generate falsely high hCG results on the UniCel DxI in our laboratories, we found these patients had high blood leukocyte counts. Based on this finding, we investigated the interference of leukocytes in the Beckman Coulter Access Total hCG assay. This study was approved by the Emory Institutional Review Board. First, we measured hCG in leftover plasma samples from patients with known high blood leukocyte content of 25 10/ L. The thirteen randomly selected samples (10 male, 3 female) were inverted several times in the original lithium heparin plasma separator tubes (PSTs) (Becton Dickinson) to disturb any particulates settled during storage (up to 24 h refrigerated) and then measured for hCG on the Unicel DxI. Surprisingly, eleven (11/13) of these samples had markedly increased hCG concentrations exceeding the institutional reference interval of 5 mIU/mL (range: 10 – 256 mIU/ mL, mean: 106 mIU/mL) (Table 1). However, after transferring the plasma to aliquot tubes and recentrifugation, hCG decreased by an average of 93% in the 11 previously increased samples, all to 3 mIU/mL (Table 1). Recentrifugation and testing of the original PSTs resulted in less notable reductions in the interference (average decrease of 64%). False increases were not observed in eight PSTs collected from patients with low blood leukocyte content of 1 10/ L (Table 1). After finding a high incidence of falsely increased plasma hCG in the samples tested with high blood leukocyte content, we hypothesized that addition of leukocytes in known hCG-negative plasma would lead to false hCG increases. At three separate institutions, an hCG-negative plasma sample with undetectable hCG on the Unicel DxI was identified, transferred to an aliquot tube, spiked with a leukocyte-enriched buffy coat (125 L added to 700 L plasma), and returned to the original PST. The same buffy coat (leukocyte concentration of 40 10/ L), prepared from potassium EDTA preserved-blood, was used across institutions. The final spiked plasma leukocyte concentration of 6 10/uL approximated the cellular content found in PST plasma (after 10 inversions) from patients with high blood leukocyte counts. Untreated hCG-negative plasma served as controls. In one lab, an additional hCG naı̈ve sample was spiked with the same buffy coat in the same ratio but after incubating the buffy coat at 58 °C for 60 min. Both the spiked and control samples, in the original PSTs, were gently inverted 10 times, recentrifuged and measured for hCG on the Unicel DxI at the respective institution. All samples were also tested for hCG on either an Abbott Architect i2000 (Architect Total hCG, Abbott Diagnostics) or a Siemens Advia Centaur® Total hCG (Siemens Medical Solutions Diagnostics) for comparison. The control plasma samples were consistently negative throughout the experiment, as was the heattreated buffy coat-spiked sample. The three other spiked samples produced markedly increased hCG results on the Unicel DxI (range of positive results: 16 – 144 mIU/mL), but remained hCG negative ( 5 mIU/mL) by the Architect or Centaur. After transferring plasmas to aliquot tubes and recentrifuging, the Unicel DxI results for the spiked samples all decreased to 5 mIU/mL. The Beckman Coulter immunoassays are two-site immunoenzymatic assays, utilizing antiantigen antibody bound paramagnetic particles and antiantigen antibody: alkaline phosphatase (ALP) conjugate. The immobilized antibodyantigen-antibody:ALP complex is washed, ALP substrate is added, and the resulting amount of light production is measured. Increased plasma concentrations of ALP have previously been demonstrated as a source of positive interference in immunoassays which used an ALP conjugate (4 ). In our experiments, leukocyte ALP may be the source of the interference. This is supported by our data showing that the hCG interference was not observed in the sample spiked with heattreated buffy coat, suggesting a heat labile interference. ALP is not used in the Architect and Centaur assays, both of which were free from interference in our studies. Herein we identified clinically significant false increases of hCG in plasma of patients with high leukocyte counts and in leukocyteenriched plasma when tested on the Unicel DxI platform. Aliquoting and recentrifugation of the errant plasmas resulted in the most marked reduction in erroneous results. It is reasonable to assume that the more modest reduction observed after recentrifuging the Clinical Chemistry 59:7 000 – 000 (2013) Letters to the Editor