| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Beckman Coulter Inc., Chaska, MN; Departments of 2 Laboratory Medicine and Pathology and 3 Medicine, Mayo Clinic, Rochester, MN
aaddress correspondence to this author at: Endocrine Laboratory, Hilton 730C, Mayo Clinic, 200 1st Street SW, Rochester, MN 55905; fax 507-284-9758, e-mail grebs{at}mayo.edu
The cobalamins, also referred to as vitamin B12, are a group of closely related enzymatic cofactors involved in the conversion of methylmalonyl-coenzyme A to succinyl-coenzyme A and in the synthesis of methionine from homocysteine (1)(2)(3). Vitamin B12 deficiency can lead to megaloblastic anemia and neurologic deficits (4). The latter may exist without anemia or precede it. Adequate replacement therapy will generally improve or cure cobalamin deficiency. Unfortunately, many other conditions, which require different interventions, can mimic the symptoms and signs of vitamin B12 deficiency (4). Moreover, even when cobalamin deficiency has been established, clinical improvement may require different dosages or routes of vitamin B12 replacement, depending on the underlying cause (3)(4). In particular, patients with pernicious anemia, possibly the commonest type of cobalamin deficiency in developed countries, require either massive doses of oral vitamin B12 or parenteral replacement therapy (3)(4)(5). The reason is that in pernicious anemia, patients suffer from gastric mucosal atrophy, which leads to diminished or absent gastric acid, pepsin, and intrinsic factor (IF) production. Because gastric acid, pepsin, and IF are required for liberation of cobalamin from binding proteins and efficient cobalamin absorption in the gut, respectively, vitamin B12 deficiency ensues (3)(4). Most of these patients have autoantibodies against gastric parietal cells or IF, with the latter being very specific but present in only 
50% of cases (1).
The aim of the work-up of patients with suspected vitamin B12 deficiency is therefore to confirm the presence of deficiency and to establish its most likely etiology. Many different testing algorithms have been proposed to achieve maximum sensitivity and specificity for both tasks, but almost all involve measurement of serum vitamin B12 and IF antibodies (IFABs) at some step (1)(2)(4)(6). Given that the prevalence of undiagnosed cobalamin deficiency approaches 2% in the elderly, both assays are high-volume tests (7). This represents no problem with regard to vitamin B12 measurements because there are many different automated assays available. To date, however, all IFAB tests have been manual immunoassays. In the following, we describe the development and validation of an automated IFAB assay.
The Beckman Access® vitamin B12 immunoassay first liberates protein-bound B12 in patient samples, followed by competition of vitamin B12 and a solid-phase monoclonal antibody against the IF-binding site for labeled IF. We realized that this assay format could be modified to allow detection of IFABs instead of vitamin B12. In the modified assay, IFABs in patient samples compete with monoclonal B12-binding site IFABs for enzyme-labeled IF, as illustrated in Fig 1
. We prevent interference from vitamin B12 by omitting the initial step of liberating protein-bound cobalamin and by adding a monoclonal mouse anti-vitamin B12 antibody to neutralize free cobalamin (McxB12; Sigma-Aldrich). This is important because patients are often given vitamin B12 before being tested for IFABs. Vitamin B12 concentrations >750 pmol/L (1000 ng/L) can saturate serum B12-binding proteins (8), leading to increased concentrations of free cobalamin and, potentially, false positive IFAB tests (in our practice, 50% of samples fall into this category) (9). The assay signal generated is inversely proportional to the concentration of IFABs in a patient sample and is expressed in antibody units (AU)/mL. One AU is defined as the maximum possible bound signal and should correspond to the absence of any IFABs in a patient’s serum. Results are reported in AU/mL and as negative, equivocal, or positive, based on cutoffs established during clinical validation (see below).
|
Assay validation included (a) intra- and interassay imprecision; (b) dilution linearity; (c) assay interference by hemoglobin, bilirubin, triglycerides, serum albumin, and vitamin B12; (d) establishment of healthy-population reference intervals; and (e) method comparison with an established manual RIA (DPC). The latter involved three separate steps. In the first, we established diagnostic cutoffs for abnormal samples by comparison with the DPC RIA. In the second step, we then validated the disease cutoffs in a second cohort of patients by comparing the respective diagnostic categorizations of patient samples by the automated IFAB assay with corresponding results obtained with the DPC RIA. In the final step, we compared the two assays with regard to their performance in samples from 87 patients with various autoimmune diseases (rheumatoid arthritis, type I diabetes, Graves disease, Hashimoto disease, and thyroid autoantibodies without hypothyroidism) or known presence of heterophile and human anti-mouse antibodies (HAMAs).
Intra- and interassay CVs were <8% across the clinically significant range (1.00 to 15.00 AU/mL). Serial dilution of positive samples showed a steady decrease in AU/mL. However, as is commonly seen with antibody assays, this was logarithmic, rather than linear. There was no interference from hemoglobin concentrations up to 5 g/L, bilirubin up to 200 mg/L, triglycerides up to 18 g/L, or serum albumin between 50 and 90 g/L. Free vitamin B12 interference studies, performed on samples to which purified cyanocobalamin had been added at concentrations of 1.1–392.6 pmol/L (1.5–530 ng/L), demonstrated that in the Access IFAB assay the minimum free cyanocobalamin concentration that caused a significant change in measured IFAB values compared with the same sample without added cyanocobalamin was 328.9 pmol/L (444 ng/L). The minimum concentration of free vitamin B12 added to a sample that caused similar interference in the DPC RIA was 172.6 pmol/L (233 ng/L).
The reference interval for a healthy population (n = 200; 50% male and 50% female) was 0.93–1.20 AU/mL.
The positive cutoff value for the Access IFAB Assay was determined by ROC analysis of 499 samples across a IFAB concentration range of 0.93–12.25 AU/mL, with the results of the DPC RIA used as the comparison method. The raw AU data showed a log (Access) to linear (DPC) relationship as depicted in Fig. 1
of the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol51/issue1/. The area under the ROC curve for the Access IFAB assay was 0.947. Values within the reference interval were categorized as negative. We selected 1.53 AU/mL as our positive cutoff, the point where the maximum sensitivity and specificity (87% for each) were obtained. Values between the upper limit of the reference interval (1.20 AU/mL) and 1.53 AU/mL were designated as equivocal. This classification mirrors that used in the DPC RIA, which also has an equivocal range between its upper reference interval limit and its positive diagnostic cutoff.
To validate the two diagnostic cutoffs for abnormal results, we then assayed, in a blinded fashion, 127 consecutive patient samples sent for diagnostic vitamin B12 testing with both the DPC RIA and the Access IFAB assay. The observed IFAB values ranged from 0.93 to 50.82 AU/mL in the Access IFAB assay. Results for each assay were classified by their respective cutoffs as negative, equivocal, or positive. The two diagnostic categorizations by the Access IFAB assay and the DPC RIA were compared by weighted 
testing (Table 1
). Diagnostic agreement between the assays was excellent. The overall weighted was 0.90 (95% confidence interval, 0.84–0.96), i.e., 90% agreement in classification of samples. The values for the individual categories were 0.87 (0.79–0.96) for negative IFAB results, 0.41 (0.09–0.73) for equivocal results, and 0.94 (0.87–1.00) for positive results. There was insufficient serum for further characterization of the two samples that were classified as negative by the Access IFAB assays but positive by the DPC RIA.
|
In the separate group of patients with autoimmune diseases or known presence of heterophile antibodies/HAMAs, the two assays showed 96.6% diagnostic agreement. All samples [24 with HAMAs, 25 with heterophilic antibodies, 5 from patients with diabetes, 5 from patients with Graves disease, and 14 from patients with other thyroid disorders (5 from patients with Hashimoto disease, 5 containing thyroid autoantibodies, and 4 from patients with hyperthyroidism of unspecified cause)] were negative in both methods tested. Of 14 rheumatoid arthritis samples, 5 were positive in the Access IFAB assay. Of these, two were also positive in the DPC RIA. The three discrepant samples were further characterized as IFAB positive by testing with the Genesis Diagnostics IFAb ELISA or by demonstrating inhibition by added IF in the Access IFAB assay.
We were able to develop an automated IFAB assay with relative ease by modifying an existing automated vitamin B12 assay. It appears that the new assay has diagnostic performance similar to that for the DPC RIA, which has performed well with regard to sensitivity and specificity in previous studies comparing different commercial and noncommercial IFAB assays (10). However, an optimized IFAB assay has been described (11), which may offer higher detection sensitivity than the DPC RIA and, by implication, possibly the Access IFAB assay. Similarly, detection of type II IFABs, which are not directed against the B12-binding pocket of IF, may improve diagnostic sensitivity (12). Detection of type II antibodies has not been characterized for the Access IFAB assay but may be possible because they could prevent the labeled IF from binding to the solid phase through steric hindrance.
Although there is ongoing debate about the diagnostic sensitivity of particular IFAB assays, they are all generally regarded as very specific (1)(6). Because of this specificity, a proportion of patients who are investigated for vitamin B12 deficiency will end up being tested for the presence of IFABs, regardless of the diagnostic cascade favored by the investigating physician(s) (1)(2). In referral laboratories, this may lead to substantial test volumes of labor-intensive manual RIAs, with all the associated problems. These range from an increased risk of sample mix-ups or analytical mishaps attributable to multiple manual sample-handling steps, to difficulties in maintaining consistent performance, to the need to handle and dispose of radioactivity, and to lower analytical precision and increased turnaround time compared with automated assays. The new automated IFAB assay addresses all of these problems while maintaining comparable diagnostic accuracy. Moreover, both vitamin B12 measurements and IFAB measurements can be performed on the same instrument, with the possibility to tag IFAB testing to individual vitamin B12 samples, conditional on the measured cobalamin concentration, a further improvement in workflow. We therefore believe that the availability of this new IFAB assay provides an opportunity for increased and more expeditious testing and diagnosis of pernicious anemia.
Acknowledgments
The Mayo Foundation received funds from Beckman Coulter for the validation of this IF antibody assay.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
