To the Editor:
We observed a patient with markedly increased free triiodothyronine (FT3) and free thyroxine (FT4) concentrations measured on a Vitros ECi analyzer (Ortho Clinical Diagnostics). The patient was a hospitalized 42-year-old woman with lupus erythematosus who appeared euthyroid and had normal thyroid-stimulating hormone (TSH) concentrations (Table 1 in the Data Supplement that accompanies the online version of this letter at http://www.clinchem.org/content/vol51/issue6/). Thyroid peroxidase antibody and thyroglobulin antibody were <0.3 kilounits/L, and rheumatoid factor was <6.0 kilounits/L. We suspected interference from heterophilic antibodies (1)(2), but our experiments suggested interference from antibodies to diiodothyronine (T2), T3, or their conjugates, as have been described (3)(4)(5). The procedures in this study were in accordance with the Helsinki Declaration of 1975 and the subsequent 1996 amendments.
In contrast to results of the Vitros ECi FT3 and FT4 assays, which use solid phases with T2- and T3-gelatin, respectively, and labeled sheep antibodies, FT3 and FT4 were normal by the Elecsys assays (Roche Diagnostics), which use biotinylated antibodies in a one-step method.
To the Vitros FT3II and FT4 assay wells we added 0.05 mL of serum from the patient or from controls (n = 5) and 0.1 mL of diluent [phosphate-buffered saline (PBS) containing, per liter, 2 mL of Tween 20 (Sigma) and 10 g of bovine globulin (Sigma)], and incubated them at 37 °C for 18 min. After washing each well, we added 0.125 mL of horseradish peroxidase-labeled goat anti-human IgG antibody conjugate (Chemicon International. Inc.), diluted 50 000-fold in the same diluent, and incubated the mixtures at 37 °C for 18 min. After each well was washed, the Vitros Signal Reagent was added to the well, and the luminescence was measured (ALOKA luminometer). The IgG fraction of serum samples from the patient and from five control individuals was purified with a MAbTrap™ Kit (Amersham Biosciences), and the IgG concentration was adjusted to 4.0 g/L. The purified IgG (0.08 mL) was incubated at 4 °C for 24 h with 0.08 mL of PBS alone or with T2, T3, or T4 (Sigma; at 4570, 42, and 768 nmol/L, respectively) dissolved in PBS. Each sample was mixed vigorously with 1.2 mL of polyethylene glycol (PEG; 125 g/L), centrifuged at 2800g for 30 min, aspirated, and washed with 1.2 mL of PEG (125 g/L). After the precipitates were dissolved in 0.001 mol/L hydrochloric acid (0.04 mL) and neutralized by equal amounts of 0.001 mol/L sodium hydroxide, T2 and T3 were measured by the FT3 assay and T4 by the FT4 assay. The concentrations of T2 were expressed as T3 concentrations. Each sample was analyzed in duplicate. The ability of the purified IgG to bind T2, T3, or T4 was defined as the difference between the FT3 or FT4 assay result and the respective blank value and is reported as the δT2, δT3, or δT4 value.
The ratios of FT3 and FT4 concentrations in PEG-treated samples (6) to those in untreated samples were significantly lower for the patient than for 37 other patients (Table 1 in the online Data Supplement); therefore, immunoglobulins in the patient’s serum interfered with both the FT3 and FT4 assays. FT3 and FT4 values in the mixtures of serum with sheep IgG, bovine globulin, and gelatin did not differ significantly from those in the mixtures of serum and PBS only, suggesting that heterophilic antibodies and anti-gelatin antibodies did not cause the high FT3 and FT4 values.
When we examined the patient’s IgG binding with Vitros FT3II and FT4 assay wells, the luminescence generated by the patient’s serum was higher than that of the 5 control individuals (Table 1 in the online Data Supplement). This suggested that the patient’s IgG bound to T2- and T3-gelatin.
The FT3 and FT4 concentrations in purified IgG and in treated samples of purified IgG (6) were below the lower detection limits of the Elecsys assays, suggesting an absence of T3 and T4 contamination in the IgG fractions. The patient’s δT2 and δT3 values were higher than those of the 5 control individuals, but the δT4 value was within 2 SD of the values for the 5 controls (Table 1 in the online Data Supplement). This finding implies that the patient’s IgG interacted with T2 and T3 but not with T4. The cross-reactivity of the anti-T3 antibody with T2 in the Vitros FT3II assay was very low, whereas the patient’s δT2 value was evidently higher than that of 5 control individuals. We conclude that T2 was bound to the patient’s IgG.
Because anti-gelatin antibodies in the patient’s serum were not recognized, we suggest that the interfering substance were antibodies to T2 and T3. As the interfering antibodies did not interfere with the Elecsys FT3 assay, the interfering antibody in the patient’s serum may recognize T2 and T3 conjugates used in the Vitros ECi FT3 and FT4 assays, as reported for a labeled-antibody assay (7). We suggest that this interference in the Vitros ECi FT3 and FT4 assays arose from antibodies to T2, T3, or their conjugates.
- © 2005 The American Association for Clinical Chemistry
In this issue
