Immunoassay Interference on Thyroid Function Tests During Treatment with Nivolumab

Introduction

Immune checkpoint inhibitors (ICIs) act by blocking specific molecules expressed in the tumor microenvironment (1). At present, the treatment of various advanced cancers involves treatment with ICIs that belong to three classes: cytotoxic T lymphocyte-associated protein 4 (CTLA-4) inhibitors (ipilimumab), programmed cell death protein-1 (PD-1) inhibitors (nivolumab and pembrolizumab), and programmed cell death-ligand 1 (PD-L1) inhibitors (atezolizumab and durvalumab) (2). In view of their immunomodulating properties, ICIs potentially have a unique toxicity profile. Several endocrinological adverse events of ICIs, including autoimmune thyroiditis, hypophysitis, primary adrenal insufficiency, and autoimmune diabetes mellitus, have been described.

A higher incidence of thyroid dysfunction, especially hyperthyroidism (∼3%), has been reported with the use of PD-1/PD-L1 inhibitors (1,3). In the vast majority of cases, hyperthyroidism presents with transient thyrotoxicosis secondary to a painless destructive thyroiditis that induces the release of preformed thyroid hormones. This thyrotoxicosis usually resolves spontaneously, and frequently evolves into hypothyroidism over several weeks and months. In contrast, ICI-related Graves' disease is an extremely rare condition (4). The differential diagnosis is usually straightforward and is crucial to the identification of an appropriate therapeutic approach.

Elevated levels of thyrotropin (TSH) autoantibodies, an elevated uptake on the thyroid scintigraphy, and increased thyroid vascularization on ultrasonography indicate Graves' disease. However, in some rare cases, differential diagnosis could prove challenging for clinicians.

We report the case of a patient who experienced severe ICI-related thyrotoxicosis during treatment for metastatic melanoma. The diagnosis was potentially complicated by a nivolumab-induced biochemical interference on the thyroid hormones assay.

Case Presentation

An 85-year-old man, diagnosed with a melanoma of the right knee and associated liver and lung metastases, presented to the hospital with severe thyrotoxicosis. The patient was prescribed nivolumab chemotherapy (240 mg every 2 weeks) by his oncologist, and baseline laboratory investigations conducted before the onset of anticancer therapy showed normal thyroid function tests (TFTs; Table 1, Assay 1).

A few days after receiving the second dose of nivolumab, the patient developed signs and symptoms of mild thyrotoxicosis, including mild palpitations, tachycardia, tremor, and anxiety. The patient's body weight remained normal. A repeat TFT investigation showed severe thyrotoxicosis (Table 1, Assay 2). The patient tested negative for anti-TSH-receptor autoantibodies, whereas thyroid ultrasonography showed a widely inhomogeneous echoic pattern, without nodules, and with normal vascularization. Based on the suggestions in evidence-based medicine (1), the patient was continued on ICI immunotherapy with a beta-blocker added on, and careful monitoring of TFTs.

Over a few weeks, the TSH levels normalized and rapidly increased, which indicated the classical evolution of painless thyroiditis into hypothyroidism. However, the free thyroxine (fT4) levels were inappropriately elevated and did not decrease as expected (Table 1, Assay 4a). We suspected analytical interference as the basis of this biochemical trend.

All samples were processed in the same laboratory using a Siemens ADVIA Centaur^® immunoassay (Siemens Healthcare Diagnostics, Inc., Newark, NJ). Therefore, we reanalyzed sample No. 4, which showed elevated TSH levels and inappropriately high fT4, at another laboratory and with a different immunoassay method (Roche Elecsys^® ECL; Roche Diagnostics GmbH, Mannheim, Germany). The results obtained from this assay (Table 1, Assay 4b) confirmed severe hypothyroidism with low fT4 levels. Therefore, we inferred there was a possible interference of ICIs with the fT4 levels.

To exclude a possible analytical interference induced by heterophilic antibodies, we processed sample No. 4 (with Siemens Centaur^® immunoassay) after treatment with heterophilic blocking tubes (HBT; Table 1, Assay 4c), which are clinically used for simple elimination of false-positive heterophilic interference in serum immunoassays. However, levels of free triiodothyronine (fT3) and fT4 were comparable with levels detected in Assay 4a, excluding the presence of heterophilic antibodies. Furthermore, normal TFTs obtained before PD1 inhibitor therapy supports the exclusion of a pre-existing interference on fT4 levels due to other endogenous factors.

Therefore, we suspected a possible nivolumab-associated interference on the fT4 assay. To verify our assumption, we subjected sample No. 4 to a polyethylene glycol (PEG) 6000 precipitation before assaying for fT4 levels (Siemens Centaur immunoassay; Table 1, Assay 4d). PEG precipitation is an easy simple method for the removal of macromolecules through the precipitation of interfering substances. The evaluation of the post-PEG precipitation sample revealed low fT4 levels.

Thus, hypothyroidism was confirmed, and the patient was started on levothyroxine (LT4) therapy, with monthly TFT monitoring using the Roche immunoassay, which is unaffected by nivolumab-induced interference, for therapy adjustment. Approximately 9 months after starting nivolumab therapy, the patient was advised treatment cessation by his oncologist. A month later, the TFTs were retested on a Siemens immunoassay (Table 1, Assay 5), and appropriate fT4 levels were observed in accordance with normal TSH levels on adequate LT4 replacement therapy. The documented trend in fT4 levels more realistically presents the possibility of nivolumab-induced endocrine interference.

Discussion

The uniqueness of this case is represented by the double effect related to the use of nivolumab: the initial well-known effect of a “destructive thyroiditis” that was characterized by thyrotoxicosis and rapid evolution to hypothyroidism, as well as a second effect, one that has not been reported so far, and is related to the biochemical interference of nivolumab in the fT4 immunoassay.

Several endogenous and exogenous factors have been reported to cause biochemical interferences with the thyroid function immunoassays, including macro-TSH, biotin, antistreptavidin antibodies, antiruthenium antibodies, antithyroid hormone autoantibodies (Ab-TH), human antimouse antibodies, and heterophilic antibodies (5). The Ab-TH (such as anti-T3 and anti-T4 antibodies) are the only antibodies known to interfere with TFTs (6) and are IgG-isotype autoantibodies that generate a polyclonal autoreactive response. Ab-TH are more prevalent in patients with autoimmune disorders (6) and have been proposed as an early biomarker of thyroid dysfunction during treatment with tyrosine kinase inhibitors and immunomodulatory drugs (7).

In one-step immunoassays, the patient's serum and the labeled hormone analogue are added to the reaction chamber, which competitively bind with the solid-phase antibody; the unbound material is washed away, and the bound analogue is measured (5). In fT4 immunoassays, anti-T4 autoantibodies bind to T4 analogues, which makes them unavailable for competitive binding. Thus, the signal is reduced and generates a falsely elevated hormone value. Theoretically, although only one-step immunoassays are affected by the interference of anti-TH autoantibodies (8), two-step immunoassays used in clinical practice may be susceptible to this interference.

To the best of our knowledge, there are no reports of biochemical interference caused by ICIs in the literature. Hypothetically, nivolumab, an IgG4 anti-PD-1 monoclonal antibody, should modulate its interference of the fT4 dosage through a mechanism akin to that reported for anti-fT4 autoantibodies. In fact, in our case, the interference detected on the one-step immunoassay was not confirmed on a two-step immunoassay. The absence of endogenous causes of biochemical interference can be validated by normal TFTs (obtained by a one-step immunoassay both before nivolumab therapy initiation and after cessation of anticancer treatment) and studies on post-HBT samples.

In conclusion, we report the possibility of a novel nivolumab-induced biochemical interference on assays of fT4 levels. In the present case, the hypothesis of a biochemical drug-induced interference is further supported by the disappearance of the interference after the withdrawal of nivolumab. Further studies are needed to prove the biochemical mechanisms of this interference, but it is important to consider the possible impact of ICI therapy on the biochemical evaluation of TFTs. Furthermore, these biochemical interferences can be misdiagnosed in the clinical setting due to the well-known effects of ICIs on thyroid function; therefore, nivolumab interference should always be considered in the presence of clinical or biochemical discrepancies in TFTs.

In particular, the interference on fT4 dosage can be obscured by a clinical suspicion of thyrotoxicosis. Therefore, it is mandatory to consider the relationship between fT3, fT4, and TSH levels in these cases, and discrepancies, if any, should be evaluated by repeat TFTs at another laboratory or by using a different immunoassay as well as the assessment of TFTs after stopping ICI anticancer therapy.