OUP user menu

Intrinsic Factor Blocking Antibody Interference Is Not Detected in Five Automated Cobalamin Immunoassays

Stephen D. Merrigan, David T. Yang MD, Joely A. Straseski PhD
DOI: http://dx.doi.org/10.1309/AJCPPWHSB94FTOGU 702-705 First published online: 1 May 2014

Abstract

Objectives To systematically assess five automated cobalamin assays for intrinsic factor binding antibody (IFBA)–induced interference using pooled serum.

Methods Six pools created from IFBA-negative and IFBA-positive serum representing low, normal, and high cobalamin concentrations were analyzed before and after polyethylene glycol (PEG) precipitation of immunoglobulins on five cobalamin assays: the Centaur XP (Siemens Healthcare Diagnostics, Tarrytown, NY), IMMULITE 2000 (Siemens Healthcare Diagnostics), ARCHITECT i2000SR (Abbott Diagnostics, Abbott Park, IL), UniCel Dxl 800 (Beckman Coulter, Chaska, MN), and Modular E170 (Roche Diagnostics, Indianapolis, IN).

Results Cobalamin concentrations before and after PEG treatment were similar, almost all within a 30% total allowable error, with no difference in pattern between the IFBA-negative and IFBA-positive pools regardless of the cobalamin concentration.

Conclusions Our results suggest that, under optimal conditions, the five automated cobalamin assays assessed are not susceptible to IFBA-mediated interference.

Key Words:
  • Cobalamin
  • Vitamin B12
  • Interference
  • Intrinsic factor
  • Immunoassay

Several authors have recently reported false-normal or even elevated cobalamin (vitamin B12) results in cobalamin-deficient patients diagnosed with pernicious anemia.1,2 Autoantibodies in patients with pernicious anemia, intrinsic factor blocking antibody (IFBA) in particular, have the potential to bind reagent intrinsic factor in competitive-binding luminescence assays (CBLAs) for cobalamin. This binding blocks the binding site for cobalamin, prevents the formation of luminescent complexes, and thereby causes an overestimation of serum cobalamin concentrations. This result is of particular concern in Western populations, in whom pernicious anemia is a common cause of megaloblastic anemia, and 70% of these patients have detectable IFBA concentrations in their serum.3

In light of a voluntary recall in May 2012 of cobalamin reagent due to IFBA interference (Dimension Vista System; Siemens Healthcare Diagnostics, Tarrytown, NY), we investigated five automated cobalamin immunoassays to determine if they were similarly affected.

Materials and Methods

Establishing a Reference Platform

Serum from a patient with IFBA and untreated pernicious anemia that had previously generated false-normal cobalamin results on the Dimension Vista instrument1 was analyzed on the Centaur XP instrument (Siemens Healthcare Diagnostics). Results of an unmodified aliquot and an aliquot subjected to immunoglobulin precipitation (described below) run on the Centaur XP were compared with one another.

Protein Precipitation by Polyethylene Glycol

In total, 100 μL of serum was diluted with 100 μL of 25% polyethylene glycol (PEG) (w/v), vortexed, incubated at room temperature for 10 minutes, and centrifuged at 13,200g for 3 minutes. The supernatant was carefully removed for analysis without disturbing the precipitate.

Generation of Serum Pools

IFBA-positive serum, determined by enzyme-linked immunosorbent assay (ELISA) (QUANTA Lite Intrinsic Factor ELISA; INOVA Diagnostics, San Diego, CA), was split into three pools based on cobalamin concentrations as determined by the Centaur XP: those with high (>911 pg/mL [>672 pmol/L]), normal (210–911 pg/mL [155–672 pmol/L]), and low (<210 pg/mL [<155 pmol/L]) cobalamin concentrations. Matching IFBA-negative pools with high, normal, and low cobalamin concentrations were also generated. All sample pools were assayed by all five immunoassays immediately after creation. All studies involving the use of human specimens were approved by the University of Utah Institutional Review Board.

Evaluation of Instruments

Five automated cobalamin assays were assessed: the Centaur XP (vitamin B12, Siemens Healthcare Diagnostics) and IMMULITE 2000 (vitamin B12, Siemens Healthcare Diagnostics), ARCHITECT i2000SR (ARCHITECT B12, Abbott Diagnostics, Abbott Park, IL), UniCel Dxl 800 (vitamin B12, Beckman Coulter, Chaska, MN), and Modular E170 (vitamin B12, Roche Diagnostics, Indianapolis, IN). Cobalamin concentrations for each pool were tested in triplicate both before and after PEG precipitation of immunoglobulins. PEG-treated serum was tested for IFBA by ELISA to confirm removal of IFBAs.

Data Analysis

Comparison between untreated and PEG-treated pools was performed by establishing a range of total allowable error (TAE) (mean ± 30%) based on results of the untreated pool.4 All results of the PEG-treated pools were corrected by a 1:1 dilution factor.

Results

Establishing a Reference Platform

To establish a reasonable reference platform for creation of the serum pools, we analyzed the IFBA-positive (87.8 U) serum on the Centaur XP from a patient with untreated pernicious anemia that had previously generated false-normal cobalamin results on the Dimension Vista instrument.1 In contrast to results generated on the Dimension Vista, the Centaur XP reported similarly low cobalamin concentrations in the untreated and PEG-treated samples Figure 1, suggesting that the Centaur XP was not affected by the same protein-based interference that produced the spuriously elevated results reported by the Dimension Vista.

Figure 1

Spurious elevation of cobalamin concentrations from a patient with untreated pernicious anemia using the Dimension Vista (Siemens Healthcare Diagnostics, Tarrytown, NY) but not observed with the Centaur XP (Siemens Healthcare Diagnostics). Untreated (black bars) and polyethylene glycol (PEG)–treated (white bars) cobalamin results did not differ from each other using the Centaur XP. Conversely, spuriously elevated cobalamin concentration previously generated using the Dimension Vista was rectified after precipitation by PEG.1 To convert cobalamin values to picomoles per liter, multiply by 0.7378.

Characterization of the Serum Pools

Six serum pools, three prepared from IFBA-positive serum and three prepared from IFBA-negative serum, were created based on high, normal, and low cobalamin concentrations as determined by the Centaur XP (see Materials and Methods). The assembled pools were then tested for cobalamin concentrations on the Centaur XP and for IFBA by ELISA both before and after PEG treatment Table 1 and Table 2. The IFBA-negative pools were confirmed to indeed be IFBA negative, and PEG treatment did not affect cobalamin concentrations. The IFBA-positive pools were confirmed to be IFBA positive and showed a marked decrease in IFBA detection after PEG treatment, consistent with near-complete precipitation of IFBA by PEG. As expected for the reference instrument, there was little change in cobalamin concentrations in these IFBA-positive pools before and after PEG treatment.

View this table:
Table 1
View this table:
Table 2

Performance of Cobalamin Assays With IFBA-Positive and IFBA-Negative Serum Pools

Assay interference by IFBA was assessed on each of five commercial platforms by comparison of cobalamin concentrations measured before and after PEG treatment. It was assumed that spuriously elevated cobalamin concentrations due to IFBA interference would be corrected by PEG treatment, resulting in a difference between cobalamin concentrations measured before and after PEG treatment.1 This difference was assessed by establishing a TAE of 30% (mean ± 30%) based on the pre-PEG treatment cobalamin concentration.4 For the IFBA-positive and IFBA-negative low-cobalamin pools, all instruments generated similar TAE ranges except for the IMMULITE 2000, which was below the lower limit of detection (150 pg/mL [111 pmol/L]) for both pools Figure 2A. Cobalamin concentrations after PEG treatment were all within the TAE for both IFBA-negative and IFBA-positive pools except for the Centaur XP, which was slightly above the TAE for the IFBA-positive pool. However, both results on the Centaur XP would qualitatively be classified as cobalamin deficient (<120 pg/mL [<89 pmol/L]). More important, PEG-treated results from both the IFBA-negative and IFBA-positive pools showed a similar pattern compared with their respective TAEs, suggesting a lack of IFBA interference in all platforms. The same experiments performed on IFBA-positive and IFBA-negative normal cobalamin pools showed similar results, again suggesting a lack of IFBA interference Figure 2B. The mean difference in measureable concentration between the six PEG-treated and untreated pools using the five immunoassays was 16.9%. For the high-cobalamin pools, PEG-treated and untreated cobalamin concentrations exceeded the measurement range in all assays (data not shown).

Figure 2

Comparable cobalamin results for polyethylene glycol (PEG)–treated and untreated serum pools using five automate immunoassays: the Centaur XP (Siemens Healthcare Diagnostics, Tarrytown, NY), IMMULITE 2000 (Siemens Healthcare Diagnostics), ARCHITECT i2000SR (Abbott Diagnostics, Abbott Park, IL), UniCel Dxl 800 (Beckman Coulter, Chaska, MN), and Modular E170 (Roche Diagnostics, Indianapolis, IN). Average cobalamin results for PEG-treated pools are represented by a dash, and range of total allowable error (mean ± 30%) is represented by a bar. Similar results were observed with and without intrinsic factor binding antibody (IFBA) for low (A) and normal (B) cobalamin sample pools. PEG-treated low cobalamin/IFBA-positive pool results using the Centaur XP assay slightly exceeded the 30% range. Results are not shown for high-cobalamin pools since all PEG-treated and untreated cobalamin concentrations exceeded the measurement range of all assays. aValue was below the lower limit of detection (150 pg/mL). bValue was below the lower limit of detection (150 pg/mL); multiplication by PEG dilution factor results in less than 300 pg/mL. To convert cobalamin values to picomoles per liter, multiply by 0.7378.

Discussion

While one CBLA has been voluntarily recalled due to IFBA interference (Dimension Vista, May 2012), there is little information regarding the performance of other CBLA assays currently in use. To systematically assess a variety of CBLA assays, we created six pools of human IFBA-negative and IFBA-positive serum representing low, normal, and high cobalamin concentrations. Comparison of cobalamin concentrations before and after PEG treatment showed little difference (<30% TAE), and this finding held true in both IFBA-negative and IFBA-positive pools. Assuming PEG precipitation removed IFBA (confirmed by negative IFBA ELISA concentrations), these findings suggest a lack of IFBA interference at all cobalamin concentrations tested in the five CBLAs included in this study.

Footnotes

  • Cobalamin reagents were graciously provided by Abbott Diagnostics, Beckman Coulter, Roche Diagnostics, and Siemens Healthcare Diagnostics. Additional support was provided by the ARUP Institute for Clinical and Experimental Pathology.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
View Abstract