To the Editor,
A 9-year-old boy seen by a paediatrician underwent physical examination for recurrent abdominal pain and intermittent obstipation. The patient had no significant medical history and blood sampling was performed to exclude coeliac disease. Routine test results were normal except for thyroid function. Although not showing any symptoms, hyperthyroidism was suspected due to an elevated free thyroxine (FT4) value of 43.53 pmol/L (reference range [RR]: 12.50–21.50) concomitant with a normal thyroid-stimulating hormone (TSH) value of 1.64 mU/L (RR: 0.60–4.84). The patient was referred to an endocrinologist for a full thyroid work-up who repeated the phlebotomy. TSH, FT4, free triiodothyronine (FT3) and anti-TSH receptor antibody (anti-TSHR), were measured with our routine analyser: Cobas 8000® e602 module (Roche Diagnostics®, Basel, Switzerland). The TSH value was normal, 0.91 mU/L, whereas both FT4 and FT3 showed elevated serum concentrations of 36.42 pmol/L (RR: 12.50–21.50) and 10.53 pmol/L (RR: 3.88–8.02), respectively. Anti-TSHR was mildly elevated to a value of 2.02 U/L (RR: <1.75), which could be an indication of a rare autosomal gene defect that establishes resistance against thyroid hormones. Before initiating further investigations, the clinician contacted the clinical laboratory to rule out an analytical interference. We additionally screened for thyroglobulin and thyroid peroxidase antibodies with the ImmunoCAP 250 (ImmunoCAP, Uppsala, Sweden) to exclude any auto-immune disease of the thyroid . As both assays were negative, with values <222 kU/L and <26.0 kU/L, respectively, we challenged the sample with another platform: the Architect i2000SR (Abbott, Abbott Park, IL, USA). This analysis resulted in a value of 1.19 mU/L (RR: 0.70–4.17) for TSH, 14.32 pmol/L (RR: 11.48–17.60) for FT4 and 5.20 pmol/L (RR: 4.30–6.81) for FT3, all within normal range. Results are shown in Table 1.
Patient’s laboratory results of different serum samples on the Cobas 8000® and Architect i2000SR.
|Cobas 8000® e602 module||Architect i2000SR||Cobas 8000® e602 module|
|TSH, mU/L||FT4 II, pmol/L||FT3 III, pmol/L||Anti-TSHR, U/L||TSH, mU/L||FT4, pmol/L||FT3, pmol/L||FT4 III new assay, pmol/L|
|(RR: 0.60–4.84)||(RR: 12.50–21.50)||(RR: 3.88–8.02)||(RR: <1.75)||(RR: 0.70–4.17)||(RR: 11.48–17.60)||(RR: 4.30–6.81)||(RR: 12.80–20.40)|
|Third (follow-up) sample||0.85||22.33||6.44||ND||ND||ND||ND||ND|
Different methods were then used to characterise the interference on our routine analyser. We first excluded the presence of rheumatoid factors with the BN Prospec® (Siemens Healthcare Diagnostics, Marburg, Germany) (<11.5 U/mL) . We then decided to challenge the sample with a polyethylene glycol (PEG) precipitation for the possibility of macro complexes between immunoglobulins or between other proteins and free hormone . As previously described, equal volumes of serum and 25% PEG 6000 (w/w) (VWR, Lutterworth, UK) were added and mixed. After centrifugation, the supernatant was analysed . Upon comparison to five male control samples, the precipitation method resulted in a small increased recovery of TSH, and lower recoveries of FT4 and FT3. Sequential dilutions are frequently used to detect interferences, however, this method is controversial for the detection of free hormones due to an equilibrium between protein-bound and unbound fraction. As the PEG precipitation method also includes a dilution of the sample, we cannot use it as a sole indication of interference. Therefore, as a next step, we used heterophilic blocking agents (heterophilic blocking tubes [HBT], Scantibodies Laboratory, Inc., Santee, CA, USA) to detect presence of interfering antibodies . Following the manufacturer’s procedure, the TSH value was also slightly raised to a level of 1.18 mU/L, while levels of FT4 and FT3 dramatically dropped to values of 19.24 pmol/L and 6.08 pmol/L, respectively, which are within normal range. Results are shown in Table 2. We contacted Roche Diagnostics® and also provided them with a small volume of sample for further investigation.
Results of the interference experiments.
|TSH, mU/L||FT4 II, pmol/L||FT3 III, pmol/L||Recovery TSH [RR]||Recovery FT4 II [RR]||Recovery FT3 III [RR]||FT4 III new assay, pmol/L|
|PEG precipitation||0.38||16.24||4.68||83% [62%–77%]||89% [167%–183%]||89% [125%–172%]||ND|
|After streptavidin beads||ND||22.70||7.70||NA||NA||NA||19.30|
Reported values are as measured with the Cobas 8000® e602 module and multiplied by 2 for the PEG precipitation test. [RR]: recovery reference values calculated as the mean±2 SD of the analysis of 5 male control samples. PEG, polyethylene glycol; HBT, heterophilic blocking tubes; ND, not done; NA, not applicable.
Interferences have been previously related to an excess of biotin . As our patient did not take any drugs or dietary supplements, we excluded this as a primary cause. Immunoassays from Roche Diagnostics® are also known for interference due to anti-ruthenium antibodies . In the FT4 II and FT3 III generation assays, the ruthenium conjugate is replaced by a modified ruthenium molecule, which should lead to less false-positive values . As the use of HBT could eliminate the interference, we first suggested the presence of heterophilic antibodies. In a recent publication however, the presence of anti-streptavidin antibodies was shown with a positive HBT test . We therefore performed an additional experiment by incubating the patient’s sample for 1 h with streptavidin beads prepared from 5 mL of Cobas reagent (0.72 mg/mL in HEPES-bovine serum albumin buffer, pH 7.4; Roche) . After centrifugation, the supernatant was analysed and a significant decrease in FT4 and FT3 values were seen: 22.70 pmol/L and 7.70 pmol/L, respectively. TSH was not measured.
Roche Diagnostics® released a new FT4 III assay in September 2018. We were interested if the interference would still be present in this updated version. A frozen leftover from the original sample, and the post-streptavidin treatment, were tested. Measurement on the Cobas 8000® showed an FT4 value of 20.56 pmol/L (RR: 12.80–20.40) with the original serum and 19.30 pmol/L after treatment with streptavidin beads.
Both the second and third generation FT4 assays use the same monoclonal antibodies. We cannot exclude the presence of heterophilic antibodies, but we can hypothesise that a similar interference would have occurred. The Architect i2000SR immunoassay does not make use of a biotin-streptavidin complex and produced values within normal range. We therefore suggest that this was another case with anti-streptavidin antibodies , , . The manufacturer also confirmed the presence of an interference in the sample we sent for investigation, but unfortunately could not find its nature or did not want to disclose it. However, our results are similar to a recent publication where the FT4 III assay seems to be less prone to anti-streptavidin antibodies .
The circumstances in which the patient acquired the interfering antibodies are not clear. Peculiar is its transient existence. A follow-up blood sample obtained 6 months later showed only a slightly elevated FT4 value of 22.23 pmol/L, measured with the FT4 II assay, with normal FT3 and TSH concentrations of 6.44 pmol/L and 0.85 mU/L, respectively. This differs from some other cases where the interference has been shown to persist much longer (e.g. 12 months) . The presence of a monoclonal antibody in the original sample is however unlikely with normal serum immunoglobulin levels for IgG of 10.0 g/dL (RR: 5.9–15.5), IgA of 1.02 g/L (RR: 0.38–2.91) and IgM of 1.07 g/L (RR: 0.45–2.40).
In this study we show the presence of an antibody, presumably against streptavidin, in a 9-year-old boy. In our case no further investigations were performed nor were thyroid drugs prescribed. There were no major consequences due to the endocrinologist’s initial suspicion. We followed a classical protocol whenever such an interference is suspected. A first step is to analyse the serum with another platform as interferences depend on several factors, such as sample volume, sandwich or competitive immunoassay, one-step or two-step assay, wash or no-wash and measuring method . In our case we chose the Architect i2000SR which does not use a biotin-streptavidin immobilization system as the Roche assay does. The difference in severity of interference between TSH and the free hormones can be explained by the type of immunoassay, sandwich format versus competitive format. In sandwich immunoassays antibodies are present in excess, compared to competitive immunoassays where a fixed amount is used. Therefore, the former seems less vulnerable to the interference of anti-streptavidin antibodies.
However, it is not always necessary for the patient’s wellbeing to characterise the nature of an interference, it is important that it is documented in the clinical file and interferences keep being reported in the literature. This allows laboratory specialists and clinicians to keep being aware of such interferences. It also stimulates manufacturers to improve their immunoassays of which the new FT4 III assay is a good example.
This study was performed in accordance with ethical guidelines. Informed consent was obtained after the second sample for further investigations.
Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.
Gurnell M, Halsall DJ, Chatterjee VK. What should be done when thyroid function tests do not make sense? Clin Endocrinol (Oxf) 2011;74:673–8.
Norden AG, Jackson RA, Norden LE, Griffin AJ, Barnes MA, Little JA. Misleading results from immunoassays of serum free thyroxine in the presence of rheumatoid factor. Clin Chem 1997;43:957–62.
Zaninotto M, Tognon C, Venturini R, Betterle C, Plebani M. Interference in thyroid hormones with Roche immunoassays: an unfinished story [Letter]. Clin Chem Lab Med 2014;52:e269–70.
Van Der Watt G, Haarburger D, Berman P. Euthyroid patient with elevated serum free thyroxine. Clin Chem 2008;54:1239–41.
Bolstad N, Warren DJ, Nustad K. Heterophilic antibody interference in immunometric assays. Best Pract Res Clin Endocrinol Metab 2013;27:647–61.
Piketty ML, Prie D, Sedel F, Bernard D, Hercend C, Chanson P, et al. High-dose biotin therapy leading to false biochemical endocrine profiles: validation of a simple method to overcome biotin interference. Clin Chem Lab Med 2017;55:817–25.
Buijs MM, Gorgels JP, Endert E. Interference by antiruthenium antibodies in the Roche thyroid-stimulating hormone assay. Ann Clin Biochem 2011;48:276–81.
Favresse J, Lardinois B, Nassogne MC, Preumont V, Maiter D, Gruson D. Anti-streptavidin antibodies mimicking heterophilic antibodies in thyroid function tests. Clin Chem Lab Med 2018;56:e160–3.
Bayart JL, Favresse J, Melnik E, Lardinois B, Fillée C, Maiter D, et al. Erroneous thyroid and steroid hormones profile due to anti-streptavidin antibodies. Clin Chem Lab Med 2019;57:e255–8.
Ruth I, Mathieu E, Burniat A, Fage D, Cotton F, Wolff F. Interferences in free thyroxine concentration using the Roche analytical platform: improvement of the third generation? Clin Chem Lab Med 2020;58:e36–9.
Favresse J, Burlacu MC, Maiter D, Gruson D. Interferences with thyroid function immunoassays: clinical implications and detection algorithm. Endocr Rev 2018;39:830–50.