Thyrotoxicosis factitia masquerading as recurrent Graves' disease: endogenous antibody immunoassay interference,a pitfall for the unwary

Introduction

The high analytical sensitivity and specificity of immunoassays have greatly advanced the practice of medicine and particularly endocrinology. Antibody interferences in immunoassay are well documented, ^1–6 but not widely appreciated. Interfering antibodies may be heterophilic antibodies (polyreactive low affinity antibodies with no clear immunogen, e.g. rheumatoid factor), human anti-animal antibodies (specific high affinity antibodies produced against a specific animal immunogen) and autoantibodies. ^1–6 These antibodies behave unpredictably in immunoassays generating false-positive or false-negative results, which may adversely affect patient care. ^1,2 Although antibody interferences are relatively uncommon the vast numbers of immunoassays performed make them a significant clinical problem. ²

We present a case of Graves' disease complicated by thyrotoxicosis factitia. The diagnosis of thyrotoxicosis factitia was delayed because of thyroglobulin antibody interference in the thyroglobulin immunoassay resulting in a falsely high thyroglobulin result. Lack of recognition of this analytical interference erroneously excluded thyrotoxicosis factitia and misleadingly indicated endogenous hyperthyroidism.

Case report

In May 2003, a 36-year-old woman presented to her general practitioner with dyspnoea, palpitations, tremors and weight loss. Serum free thyroxine (FT4) 27.9 pmol/L (reference range 12.0–22.0) and serum thyroid-stimulating hormone (TSH) <0.01 mU/L (0.27–4.2) confirmed hyperthyroidism. She was started on thionamides but developed a rash with carbimazole and persistent vomiting with propylthiouracil. She was therefore referred to the endocrine services for further evaluation and treatment.

On examination, she was thyrotoxic. After discussion, she opted for treatment with radioactive iodine (RI) in preference to thyroid surgery. She was started on propanolol pending further investigation and definitive treatment. Investigations confirmed thyrotoxicosis (FT4 42.0 pmol/L; TSH <0.01 mU/L) and increased uptake on a thyroid radioiodine uptake (RIU) scan with positive thyrotrophin receptor antibodies (TRAb) were consistent with Graves' disease. In October 2003, she was administered 400 MBq of I¹³¹ iodine, but remained persistently thyrotoxic (FT4 28.6 pmol/L; free triiodothyronine [FT3] 10.5 pmol/L [3.95–6.80]; TSH <0.01 mU/L). In February 2004, a further 400 MBq of I¹³¹ iodine was therefore administered. Following RI, she became euthyroid (FT4 15.3 pmol/L; TSH 0.22 mU/L) and β-blockers were, therefore, stopped. Fifteen months following the second course of RI, she developed hypothyroidism (FT4 8.9 pmol/L; TSH 15.35 mU/L) and was commenced on thyroxine replacement therapy. Within two months, however, she relapsed into hyperthyroidism and remained thyrotoxic after stopping thyroxine replacement (FT4 30.6 pmol/L; FT3 24.9 pmol/L; TSH <0.01 mU/L). Propanlol was restarted and a third dose of RI planned. There was, however, absent uptake on a thyroid RIU raising the possibility of thyrotoxicosis factitia due to exogenous thyroid hormone ingestion. The patient and her partner, however, were adamant that she was not taking thyroid hormone, thyroid extract or herbal remedies. Serum thyroglobulin was therefore measured to confirm thyrotoxicosis factitia but was elevated excluding exogenous thyroid hormone ingestion and indicating endogenous hyperthyroidism (Table 1). Although very unlikely, thyroiditis and ectopic thyroxine production as causes of hyperthyroidism with absent uptake on thyroid RIU and elevated serum thyroglobulin were, respectively, excluded by the absence of discernable thyroid gland on ultrasound and absent RIU on a whole body RIU.

The discordance between imaging studies and biochemical investigations aroused suspicion of an interference in the thyroglobulin radioimmunoassay (RIA). Serial dilutions of the patient's serum sample gave non-linear thyroglobulin RIA results indicating probable interference in the RIA. Subsequent measurement of thyroglobulin using an immunometric assay (IMA) gave an undetectable value (Table 1) consistent with the radiological and clinical picture of thyrotoxicosis factitia and an interference in the thyroglobulin RIA. Thyroglobulin antibodies were elevated, and were the likely source of interference in the thyroglobulin RIA (Table 1).

The patient continued to deny ingestion of thyroid hormone or extract and her continuing thyrotoxicosis was controlled with β-blockers. In November 2006, she became clinically hypothyroid (FT4 2.0 pmol/L; TSH 90.9 mU/L). Propanolol was stopped and thyroxine replacement therapy commenced. She is currently euthyroid on 100 μg of thyroxine daily.

Discussion and conclusion

This case is unusual in several respects. Surprisingly, the development of thyrotoxicosis factitia following successful treatment of hyperthyroidism has only rarely been previously reported. ⁷ On presentation, the increased uptake on the thyroid RIU and positive TRAb were consistent with hyperthyroidism due to Graves' disease. We speculate that the patient started ingesting excessive amounts of thyroxine when she was started on thyroxine replacement for RAI-induced hypothyroidism.

In thyrotoxicosis factitia, exogenous thyroid hormone ingestion suppresses TSH concentrations. Serum FT4 and FT3 concentrations are elevated but serum FT4 may be low if the patient is ingesting only triiodothyronine. The relatively high FT3 in comparison with FT4, in this patient, did not exclude sole thyroxine ingestion but suggested the possibility of ingestion of thyroxine and triiodothyronine or thyroid extract but the patient denied these possibilities. Suppression of endogenous thyroid function by exogenous thyroid hormones reduces RIU on a thyroid RIU, leads to thyroid gland atrophy and decreases thyroglobulin secretion. Serum thyroglobulin should, therefore, be undetectable following excessive thyroid hormone ingestion thus differentiating thyrotoxicosis factitia from other causes of hyperthyroidism in which serum thyroglobulin is elevated. ⁸

Measurement of thyroglobulin is most frequently used as a tumour marker for the detection of thyroid cancer recurrence, usually in patients who have had a total thyroidectomy. Interference by thyroglobulin antibodies in thyroglobulin immunoassays is well known and usually demonstrated by a false increase in RIA but a false decrease in IMA, ^9–11 It has, therefore, been suggested that thyroglobulin and thyroglobulin antibodies be measured simultaneously. ¹² In this case, false-positive RIA thyroglobulin results without consideration of analytical interference despite the laboratory report indicating the possibility of interference by endogenous thyroglobulin antibodies led to clinical misdirection and unnecessary investigation.

Analytical interference in thyroglobulin assays is relatively common and well-recognized in the field of thyroid cancer ^10,11 and laboratory reports indicate this possibility. The problem may be less well recognized in other areas, since other antibody interferences are relatively uncommon with a reported prevalence of 0.05 to 2.0%. ² The very large volume of immunoassays performed, however, makes them a significant and underestimated problem. Clinicians should be vigilant to the possibility of immunoassay interferences (Box 1a) as these may lead to misdiagnosis and mismanagement of the patient. ^1,2,4 Laboratory personnel should also be alert to possible immunoassay interferences which depending on the immunoassay can then be identified by analytical strategies usually requiring measurement in a different immunoassay or following serial dilution in non-immune serum or before and after treatment with interference blocking agents (Box 1b). ^2,5,6 Effectiveness of these strategies is in demonstrating interference rather than excluding it. Non-parallelism on dilution and a difference in results after use of blocking antibody agent or using alternative methodology is good evidence of assay interference but the lack of a difference does not exclude interference. For the investigation of interference in thyroglobulin assays, the discrepancy between RIA and IMA is taken as the most accurate way of identifying interference, though it does not indicate which is the correct result. Recovery experiments and assessment of linearity may be suggestive but are not always conclusive. Treatment with ‘blocking tubes’ may be suitable to detect heterophilic antibodies, which cause false-negatives in IMA but may not be suitable for the detection of interference due to endogenous thyroglobulin antibodies. Measurement of endogenous thyroglobulin antibodies does not always predict assay interference.

In summary, unrecognized antibody interferences in immunoassays have the potential to cause patient harm and waste health-care resources. Although, it is the laboratory's duty to ensure the validity of the results it produces, ¹³ we suggest that good clinician-laboratory communication underpins effective clinical and laboratory strategies to detect erroneous laboratory results resulting from endogenous antibody interference in immunoassays.