Grave Clinicopathologic Correlation: A Case of Hyperthyroxinemia

A family physician ordered a set of laboratory investigations to rule out treatable causes of fatigue and a sensation of coldness in a 16-year-old female who had been previously well. The investigations included a borderline low thyroid-stimulating hormone (TSH)3 [0.41 mU/L, reference interval (RI) 0.51–4.3 mU/L] and increased free thyroxine (fT4) measurement of 3.42 ng/dL (44 pmol/L) (RI 0.85–1.71 ng/dL). On the basis of the fT4 result, the family doctor repeated the tests and sent the patient to an endocrinologist for evaluation. The results on a sample collected 6 days later were similar, with a TSH of 0.49 mU/L and fT4 of 3.57 ng/dL (46 pmol/L). The endocrinologist assessed the patient 2 weeks later and found her to be clinically euthyroid with neither goiter nor Graves' ophthalmopathy. Nevertheless, the endocrinologist ordered repeated investigations of TSH and fT4 as well as free triiodothyronine (fT3) and markers of those autoimmune thyroid disorders, which could be associated with suppressed TSH and increased fT4. The new sample was collected 8 weeks after the original and results were as follows: TSH 0.68 mU/L, fT4 2.10 ng/dL (27 pmol/L), fT3 0.61 ng/dL (9.3 pmol/L) (RI 0.23–0.43 ng/dL), TSH receptor antibody (TSHRa) of 5 U/L (RI <1.8 U/L), and undetectable anti-thyroid peroxidase (anti-TPO) antibody. The clinical laboratory was consulted to assist in resolution of the discordance between the laboratory results, suggestive of Graves' disease, and the clinical assessment of a euthyroid state. The samples from the first 3 collections were analyzed by 2 different laboratories using the same commercial immunoassay methods for TSH, fT4, fT3, and TSHRa, respectively (Roche Diagnostics, electrochemiluminescence, Cobas® e601). The anti-TPO test was performed by immunoassay by a different manufacturer (Siemens IMMULITE® 2000 XPi). To investigate methodology-specific interference affecting TSH, fT4, and fT3 results, the analytes were measured at a third laboratory using an alternate methodology (Beckman Coulter, Access 2®). This third laboratory gave values of TSH, fT4, and fT3 within the respective reference intervals (Table 1). Because these subsequent test results on an alternate technology matched the clinical status of the patient, there was high suspicion that tests performed on the Roche platform were generating spurious results. The samples were retested on the Roche instrument after incubation with a heterophilic blocking tube (HBT) (Scantibodies) and a separate aliquot after incubation with streptavidin-coated magnetic beads (SB) from a Roche reagent kit. Pretreatment with either HBT or streptavidin-coated magnetic beads led to a marked decrease in free thyroid hormone results (Table 1). Incubation with HBT also normalized TSHRa; however, use of the streptavidin-coated magnetic beads slurry interfered with the TSHRa assay, as observed for both the patient and a control specimen. Although not considered reportable, the post-HBT results were no longer suggestive of hyperthyroidism and were therefore consistent with the clinical status of the patient. Thyroid hormone and thyroid autoantibody testing results by method, day of collection, and pretreatment with HBT or SB. fT4: multiply by 12.87 to convert to pmol/L. fT3: multiply by 15.36 to convert to pmol/L. NA, not applicable. Bold font indicates that the test result is outside of the listed test normal reference interval. Thyroid hormone and thyroid autoantibody testing results by method, day of collection, and pretreatment with HBT or SB. fT4: multiply by 12.87 to convert to pmol/L. fT3: multiply by 15.36 to convert to pmol/L. NA, not applicable. Bold font indicates that the test result is outside of the listed test normal reference interval. Immunoassay methods are prone to sporadic interferences that, in some cases, have led to serious misdirection of clinical care (1). In the case presented, immunoassay interference led to spurious laboratory results suggestive of Graves' disease. Fortunately, clinical care was not seriously misdirected because of the initiation of a joint clinicopathologic correlation between the laboratory professionals and the clinical endocrinologist. The strong clinical impression of a euthyroid 16-year-old female directed the investigation towards a search for laboratory error. Because the abnormal thyroid test results had been repeated over several weeks, the likelihood of pre- or postanalytical error was reduced and other causes (both analytical and physiological) were considered. Other possibilities under consideration, before the availability of the TSHRa measurement, included genetic anomalies resulting in euthyroid hyperthyroxinemia. True thyroid hormone pathology associated with subclinical findings that would progress and later manifest were considered less likely. The clinician considered in the differential diagnosis familial dysalbuminemic hyperthyroxinemia (Online Mendelian Inheritance in Man® [OMIM] no. 615999). This condition is associated with increased total thyroxine (T4) and, in some cases, with increased total triiodothyronine (T3). While equilibrium dialysis methods would show normal fT4 and fT3 concentrations, certain commercial immunoassay methods measure an increased concentration due to interference from altered thyroid hormone–binding proteins (2). This condition is inherited in an autosomal-dominant manner and testing of relatives should be considered to establish the pattern of inheritance. Bisalbuminemia is common in this condition and can be detected by serum protein electrophoresis. If these tests return noncontributory results, and familial dysalbuminemic hyperthyroxinemia remains in the differential diagnosis, further investigations include genetic sequencing of the albumin (ALB)4 gene (2) or mass spectrometric analysis of serum albumin (3) to confirm the presence of a variant albumin gene/protein. Similarly, thyroid hormone resistance (OMIM no. 188570) is another cause of euthyroid state with increased free thyroid hormones and an inappropriately normal or high TSH. Because this patient had a relatively low TSH and no accompanying goiter, this condition was deemed less probable. Nevertheless, further testing via sequencing of the thyroid hormone receptor beta (THRB) gene was considered; however, the increased TSHRa result suggested that the primary diagnosis was neither familial dysalbuminemic hyperthyroxinemia nor genetic thyroid hormone resistance. Accordingly, pursuit of a genetic diagnosis was abandoned, as was that of other, less clinically relevant considerations such as a TSH-secreting pituitary tumor, the recovery phase of nonthyroidal illness, or factitious hyperthyroxinemia. After the positive TSHRa test, 2 autoimmune thyroid conditions were considered: Graves' hyperthyroidism and the hyperthyroid phase of Hashimoto thyroiditis. However, the negative anti-TPO test was atypical for both conditions, as was the relatively unsuppressed TSH. More importantly, neither condition was given strong consideration on clinical grounds given the patient's euthyroid presentation. The lack of a unifying clinical explanation for the abnormal test results refocused the investigation on possible sources of analytical error. On review of the 3 markedly aberrant test results (fT4, fT3, TSHRa), it was noted that all were performed with the same methodology (competitive electrochemiluminescent immunoassay), and all had unexpectedly increased concentrations. In a competitive immunoassay, the signal (in this case electrochemiluminescence) is inversely proportional to the concentration of the analyte; thus, an inappropriately low assay signal results in a high concentration of the analyte of interest. In contrast, in a noncompetitive sandwich immunoassay, the signal is directly proportional to the concentration of the analyte of interest. In this case, the TSH measured by noncompetitive sandwich immunoassay, was considered potentially negatively biased owing to signal suppression. A first step to investigate this possibility was to determine if the potential interference was specifically related to the methodology of the Roche immunoassay platform. The TSH, fT4, and fT3 tests when performed on an alternate immunoassay platform (Table 1) were normal, in congruence with the clinical impression, suggesting that the “abnormal” Roche thyroid test results were analytical errors. With this impression, further validation of the Roche test results with sample dilution (for TSH) or equilibrium dialysis (for fT4 and fT3) was not pursued. Because the analytic error was method specific, the sources of signal suppression particular to the Roche methodology were reviewed. These sources include high-dose biotin supplementation (>5 mg/day) or endogenous antibodies including those binding to ruthenium or streptavidin (1, 2, 4–7). High-dose biotin therapy is used for a number of purposes including the treatment of deficiencies in select inborn errors of metabolism (e.g., biotinidase deficiency), as a nutritional supplement to enhance nail growth (8), and as an investigational therapy for multiple sclerosis (9). This patient was taking vitamin B12, vitamin C, and iron supplements, but not biotin. Because laboratory results for fT4, fT3, and TSHRa were significantly lower after HBT treatment (Table 1), while TSH had increased, endogenous antibody interference was suspected. Heterophile interferences causing suppression of the chemiluminescent signal from multiple immunoassays is theoretically possible. However, the pattern of results was more suggestive of antistreptavidin or antiruthenium antibodies targeting the Roche immunoassay electrochemiluminescent signaling mechanism in a generic manner. Two recent studies by Peltier et al. (1) and Rulander et al. (5) established that this particular pattern of interference with the Roche methodology, including TSH suppression and reciprocal elevation in thyroxine, could be induced by antistreptavidin antibodies. In the case of the Rulander study, as in our case, these interfering antibodies were also cleared by HBT preincubation as well as preincubation with streptavidin. Both the clinician and patient were informed of the potential for a manufacturer-specific analytical interference in test results, and subsequent laboratory testing on alternative platforms was advised. However, as clinical immunoassays frequently make use of streptavidin–biotin interactions to capture reagent antibodies, all future immunoassay results from this patient must be interpreted with attention to assay methodology and verified with ancillary testing (e.g., comparison of test results obtained with and without sample preincubation in HBT) in which there is potential for antistrepavidin antibody interference. In summary, immunoassay signal interference errors can lead to a laboratory thyroid hormone profile suggestive of Graves' disease including an abnormally high TSHRa in a clinically euthyroid individual. Awareness of this pattern by clinicians and laboratory professionals can allow for early recognition of this form of analytical error and prevent unnecessary and potentially harmful subsequent investigations and therapies. Consultation between laboratory chemists and clinical endocrinologists to perform joint clinicopathologic correlations is a valuable tool to prevent misdirected patient care. Low TSH, increased free thyroid hormones, and increased TSHRa are the characteristic laboratory profile of hyperthyroidism due to Graves' disease; however, overt signs of disease should be present with marked laboratory abnormalities. Interferences targeting manufacturer-specific substrates result in a specific pattern of laboratory error: falsely low concentration, and falsely high concentration, for noncompetitive and competitive immunoassays, respectively. thyroid-stimulating hormone reference interval free thyroxine free triiodothyronine TSH receptor antibody anti-thyroid peroxidase heterophilic blocking tube streptavidin-coated magnetic beads Online Mendelian Inheritance in Man®. albumin thyroid hormone receptor beta.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.