Abstract
Interference from heterophilic antibodies (often referred to as HAMA, although strictly this pertains only to antibodies to mouse immunoglobulins) in immunometric assays is well recognized, 1,2 but some variants have striking properties that makes them still worth reporting.
When interference is suspected, the measurement should be repeated after dilution, and/or assessed by an immunoassay of different origin using different antibodies.
The first case concerned a patient who was moderately suspected of thyroid disease (normal free T4 level of 17.1 pmol/L and elevated TSH 118 mIU/L, Roche Modular, normal values 1.5–3.5 mIU/L). The TSH diluted in a linear fashion up to 20 times. The measurement was repeated in another laboratory giving TSH 25.2 mIU/L (Abbott Architect, normal values 0.4–4.0 mIU/L) but on this occasion abnormal dilution behaviour was noted. After removal of IgG by Protein G affinity chromatography, as described previously, 3 a TSH of 2.7 mIU/L was found. The sample was distributed to other laboratories who tested for TSH with five further assay methods different from the previous two. In all cases spuriously elevated TSH values of variable apparent concentrations were found which seemed unrelated to the assay format (Table 1). It must be noted that some of the between-method differences were too small to raise suspicion. Dilution tests also gave variable results. Where tested (Perkin-Elmer DELFIA), dilution in mouse serum did not improve dilution behaviour, the use of Scantibodies ‘Blocking tubes’ was ineffective.
Apparent TSH levels in undiluted patient serum and after dilution by a factor 5 (occasionally 4 or 6)
First antibody is immobilized, second carries label. MM, MMα, MMβ, mouse monoclonal anti-wholeTSH, anti α-, anti β-TSH respectively; GP, goat polyclonal anti-whole TSH; SP, sheep polyclonal anti-whole TSH
*Averaged results from two labs
The second case concerns a female patient referred with dysphagia and a cystic thyroid gland. Neither signs of malignancy nor hypothyroidism were found. The TSH was greatly increased (48 mIU/L, non-linear dilution) with normal free T4 (13.2 pmol/L). Again, it was not possible to remove heterophilic antibodies using the Scantibodies ‘Blocking tubes.’ Immunoelectrophoresis with fluorescence detection using the Eu-labelled second antibody from the DELFIA TSH-assay showed an abnormal peak of the IgM-immunoglobulin class. After removal of the IgM antibodies, the remaining TSH had a concentration of 3.0 mIU/L.
We conclude that interference by heterophilic antibodies may escape detection if not extensively evaluated. Neither a normal dilution test nor similar results by two different assay methods exclude the presence of interference. Furthermore, screening for only IgG antibodies may give a false negative result, since interference can be due to IgM-antibodies as well. As interferences in TSH-assays are rare, 4 the above-mentioned cases show the importance of good communication between clinics and laboratories.